Effect of the axial ligand on the reactivity of the oxoiron(IV) porphyrin π-cation radical complex: Higher stabilization of the product state relative to the reactant state

Article abstract of DOI:10.1021/ic3006597

The proximal heme axial ligand plays an important role in tuning the reactivity of oxoiron(IV) porphyrin π-cation radical species (compound I) in enzymatic and catalytic oxygenation reactions. To reveal the essence of the axial ligand effect on the reactivity, we investigated it from a thermodynamic viewpoint. Compound I model complexes, (TMP^+a?€￠) Fe^IVO(L) (where TMP is 5,10,15,20-tetramesitylporphyrin and TMP ^+? is its π-cation radical), can be provided with altered reactivity by changing the identity of the axial ligand, but the reactivity is not correlated with spectroscopic data (ν(Fe=O), redox potential, and so on) of (TMP^+?)Fe^IVO(L). Surprisingly, a clear correlation was found between the reactivity of (TMP^+?)Fe^IVO(L) and the Fe^II/Fe^III redox potential of (TMP)Fe ^IIIL, the final reaction product. This suggests that the thermodynamic stability of (TMP)Fe^IIIL is involved in the mechanism of the axial ligand effect. Axial ligand-exchange experiments and theoretical calculations demonstrate a linear free-energy relationship, in which the axial ligand modulates the reaction free energy by changing the thermodynamic stability of (TMP)Fe^III(L) to a greater extent than (TMP ^+?)Fe^IVO(L). The linear free energy relationship could be found for a wide range of anionic axial ligands and for various types of reactions, such as epoxidation, demethylation, and hydrogen abstraction reactions. The essence of the axial ligand effect is neither the electron donor ability of the axial ligand nor the electron affinity of compound I, but the binding ability of the axial ligand (the stabilization by the axial ligand). An axial ligand that binds more strongly makes (TMP)Fe^III(L) more stable and (TMP^+?)Fe^IVO(L) more reactive. All results indicate that the axial ligand controls the reactivity of compound I (the stability of the transition state) by the stability of the ground state of the final reaction product and not by compound I itself.

Full text of DOI:10.1021/ic3006597

Article
pubs.acs.org/IC
Effect of the Axial Ligand on the Reactivity of the Oxoiron(IV)
Porphyrin π-Cation Radical Complex: Higher Stabilization of the
Product State Relative to the Reactant State
Akihiro Takahashi,†,‡ Daisuke Yamaki,∥ Kenichiro Ikemura,§ Takuya Kurahashi,†,‡ Takashi Ogura,§
Masahiko Hada,∥ and Hiroshi Fujii*,†,‡
†Institute for Molecular Science and Okazaki Institute for Integrative Bioscience, National Institutes of Natural Sciences, Myodaiji,
Okazaki 444-8787, Japan
‡Department of Functional Molecular Science, The Graduate University for Advanced Studies (SOKENDAI), Myodaiji, Okazaki
444-8787, Japan
∥Department of Chemistry, Graduate School of Science and Engineering, Tokyo Metropolitan University, 1-1 Minami-Osawa,
Hachioji, Tokyo 192-0397, Japan
§Department of Life Science and Picobiology Institute, Graduate School of Life Science, University of Hyogo, Koto, Kamigori, Ako,
Hyogo 678-1297, Japan
S
* Supporting Information
ABSTRACT: The proximal heme axial ligand plays an important
role in tuning the reactivity of oxoiron(IV) porphyrin π-cation
radical species (compound I) in enzymatic and catalytic oxygen-
ation reactions. To reveal the essence of the axial ligand effect on
the reactivity, we investigated it from a thermodynamic viewpoint.
Compound I model complexes, (TMP+•)FeIVO(L) (where TMP is
5,10,15,20-tetramesitylporphyrin and TMP+• is its π-cation radical),
can be provided with altered reactivity by changing the identity of
the axial ligand, but the reactivity is not correlated with
spectroscopic data (ν(FeO), redox potential, and so on) of
(TMP+•)FeIVO(L). Surprisingly, a clear correlation was found
between the reactivity of (TMP+•)FeIVO(L) and the FeII/FeIII redox potential of (TMP)FeIIIL, the final reaction product. This
suggests that the thermodynamic stability of (TMP)FeIIIL is involved in the mechanism of the axial ligand effect. Axial ligand-
exchange experiments and theoretical calculations demonstrate a linear free-energy relationship, in which the axial ligand
modulates the reaction free energy by changing the thermodynamic stability of (TMP)FeIII(L) to a greater extent than
(TMP+•)FeIVO(L). The linear free energy relationship could be found for a wide range of anionic axial ligands and for various
types of reactions, such as epoxidation, demethylation, and hydrogen abstraction reactions. The essence of the axial ligand effect
is neither the electron donor ability of the axial ligand nor the electron affinity of compound I, but the binding ability of the axial
ligand (the stabilization by the axial ligand). An axial ligand that binds more strongly makes (TMP)FeIII(L) more stable and
(TMP+•)FeIVO(L) more reactive. All results indicate that the axial ligand controls the reactivity of compound I (the stability of
the transition state) by the stability of the ground state of the final reaction product and not by compound I itself.
compound I. In particular, it has been suggested that the
proximal cysteinate (thiolate) ligand of cytochrome P450
increases the oxidizing power of compound I for hydroxylation
of unactivated C−H bonds.12
To reveal the functional role of the axial ligand, its effect on
the reactivity of high-valent iron oxo species has been studied
systematically in synthetic enzyme model complexes. Pre-
viously, a pronounced axial ligand effect on the reaction rate of
styrene epoxidation was found by investigating synthetic
compound I model complexes with various axial anion ligands,
INTRODUCTION
■
High-valent iron oxo species have been identified as the key
reaction intermediates in the catalytic cycles of oxygen
activating iron enzymes, as well as synthetic oxygenation
catalysts.1−5 In heme enzymes, an oxoiron(IV) porphyrin π-
cation radical species known as compound I has been
characterized in the catalytic cycles of peroxidases, catalases,
and cytochrome P450s.6−10 These heme enzymes have the
same heme (iron protoporphyrin IX), but different heme
proximal (axial) ligands (histidine in peroxidases, tyrosine in
catalases, and cysteine in cytochrome P450s). The axial ligands
are highly conserved in these heme enzymes.1,6,11 Therefore, it
has been believed that the axial ligand tunes the reactivity of
Received: March 30, 2012
Published: June 20, 2012
© 2012 American Chemical Society
7296
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
such as chloride, fluoride, and acetate, among others.13
Recently, the axial ligand effect was studied in detail using
other synthetic compound I model systems, and by performing
density functional theory (DFT) calculations.14−19 These
studies confirmed that the axial ligand affects the reactivity of
compound I. The axial ligand effect on the reactivity was
examined to explain it by its electron donor effect and/or the
electron affinity of compound I. As the axial ligand becomes a
better electron donor, it strengthens the FeO-H bond,
increasing hydrogen abstraction activity, and it weakens the
FeO bond, enhancing oxo-transfer reaction. Although the
electron donor effect can explain a narrow range of the axial
ligand effect such as p-substituent effect of axial ligand,15c it
cannot be applicable to a wide range of the axial ligand effect. In
fact, the electron donor effect of the axial ligand obviously
contradicts the previous experimental result.13 The FeO
bond strength, ν(FeO), of compound I model complex did
not correlate with the reaction rate constant.13 Moreover,
reactivity of compound I model complexes having imidazole
and phenolate axial ligands did not correlate with their electron
donor effects.20 In addition, a recent report on the redox
potential of oxoiron(IV) porphyrin complexes showed that the
axial ligand does not change the redox potential (electron
affinity) of compound I.21 These results clearly indicate that
there is a more essential factor than these factors.
TFA, and NO3, was evaluated in the reactions with cyclooctene
(epoxidation), 1,4-cyclohexadiene (hydrogen abstraction), and
N,N-dimethyl-p-nitroaniline (demethylation) (Figure 1). The
final species of (TMP+•)FeIVO(L) after reactions with these
substrates were the corresponding iron(III) porphyrin com-
plexes, (TMP)FeIII(L). The reactivity data for these reactions
are summarized in Supporting Information, Figure S1 and
Table 1. All of these reactivity data clearly show a pronounced
axial ligand effect, and the reactivity of (TMP+•)FeIVO(L)
increases according to the order of L = NO3 < TFA < Ac < Cl <
Hc < Bz < F for each of the three substrates. Summaries of the
spectroscopic and electrochemical characterization of
(TMP+•)FeIVO(L) and (TMP)FeIII(L) are provided in
Supporting Information, Figures S2−S8 and in Table 1. All
data for (TMP+•)FeIVO(L) and (TMP)FeIII(L) are consistent
with oxoiron(IV) porphyrin π-cation radical complexes in the
quartet ground state (S = 3/2: ferromagnetic coupling of
iron(IV) S = 1 and porphyrin π-cation radical S = 1/2) and
iron(III) porphyrin complexes in the sextet ground state (S =5/
2: iron(III) high-spin), respectively.22,23
To explore the potential factors controlling the reactivity of
(TMP+•)FeIVO(L), we examined the correlation between the
reaction rate constants and the spectroscopic parameters shown
in Table 1. If the axial ligand activates compound I model
complex, the reaction rate constants of (TMP+•)FeIVO(L) are
expected to be correlated with particular spectroscopic
parameters of (TMP+•)FeIVO(L), such as FeO bond
To reveal the essence of the axial ligand effect on the
reactivity of compound I, we have investigated the axial ligand
effect with compound I model complexes, (TMP+•)FeIVO(L),
where TMP+• is the π-cation radical of 5,10,15,20-tetramesi-
tylporphyrinate (TMP) and L is selected from the following
anionic axial ligands: nitrate (NO3), trifluoroacetate (TFA),
acetate (Ac), chloride (Cl), fluoride (F), benzoate (Bz), and
hydrocinnamate (Hc) (Figure 1). In this study, we identified a
1
strength: ν(FeO), redox potential, H NMR shift, and EPR
g-parameters of (TMP+•)FeIVO(L).15−17 However, we could
not identify any significant correlations with these spectro-
scopic parameters. Although previous studies proposed the
electron donor ability of the axial ligand and electron affinity of
(TMP+•)FeIVO(L) as a key of the axial ligand effect, this study
reveals that these factors are not the essence of the axial ligand
effect. Interestingly, as shown in Figure 2, we identified a
significant correlation between the reaction rate constants of
(TMP+•)FeIVO(L) and the redox potential of FeII/FeIII redox
process of (TMP)FeIII(L): the final porphyrin species after the
reaction. Other spectroscopic parameters of (TMP)FeIII(L) in
Table 1 were not correlated with the reactivity data of
(TMP+•)FeIVO(L).
The E1/2 values for the FeII/FeIII redox couple of (TMP)-
FeIII(L), which is governed by the Nernst equation, depends on
the relative interactions of both [(TMP)FeIII]+ and (TMP)FeII
with the axial ligand (L).24 The binding constant of the anionic
axial ligand (L) to the ferric porphyrin complex is much larger
than that to the ferrous porphyrin complex because of the
overall positive charge of the ferric porphyrin complex (the
porphyrin itself has a −2 charge while ferric iron has a +3
charge). In fact, the binding constant of fluoride, which binds
strongly to the ferrous porphyrin complex, was estimated to be
only ∼600.24 Therefore, the E1/2 value for the FeII/FeIII redox
couple of (TMP)FeIII(L) is mainly controlled by the binding
constant of (TMP)FeIII(L). The binding constant indicates the
thermodynamic stability of (TMP)FeIII(L) relative to [(TMP)-
FeIII]+. Thus, the correlations observed in Figure 2 suggest that
the reactivity of (TMP+•)FeIVO(L) is correlated with the
thermodynamic stability of (TMP)FeIII(L).
Figure 1. Structures of (TMP+•)FeIVO(L) and (TMP)FeIII(L).
significant correlation between the reaction rate constants of
(TMP+•)FeIVO(L) and the redox potential of FeII/FeIII redox
process of (TMP)FeIII(L): the final heme species after the
reaction. On the other hand, we could not find a correlation
between the reaction rate constant and spectroscopic data of
(TMP+•)FeIVO(L). The observed correlation suggests a unique
idea that the axial ligand controls the reactivity of compound I
by modulating thermodynamic stability of the iron(III)
porphyrin species, not by that of compound I itself. This idea
was confirmed by the axial ligand exchange experiments for
(TMP+•)FeIVO(L) and (TMP)FeIII(L). On the basis of the
present results including the DFT calculations, we discuss the
essence of the axial ligand effect on the reactivity of compound
I.
Axial Ligand Effect on the Free Energy of Reaction. To
investigate the axial ligand effect on thermodynamic stability,
we estimated the thermodynamic stabilities of (TMP+•)FeIVO-
(L) and (TMP)FeIII(L) in ligand exchange reactions. Figure 3
RESULTS
■
Electronic Structure and Reactivity. The reactivity of
1
(TMP+•)FeIVO(L), where L is selected from F, Cl, Bz, Ac, Hc,
shows the H NMR spectral change which occurs during a
7297
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
Table 1. Spectroscopic and Reactivity Data for (TMP+•)FeIVO(L) and (TMP)FeIII(L) in Dichloromethane at 213 K
L
NO3
TFA
Ac
Cl
Hc
Bz
F
(TMP+•)FeIVO(L)
absorption/nm
403, 666
406, 666
409, 668
411, 668
413, 668
802
410, 670
804
409, 666
a
a
b
Raman: νFeO/cm−1
EPR: g-value (E/D)
821
816
800
804 (801)
4.25, 3.75,
1.99 (0.040)
801 (807)
4.44 3.52,
1.98 (0.075)
4.45, 3.52,
1.98 (0.075)
4.45, 3.51,
1.97 (0.075)
4.46, 3.50,
1.97 (0.075)
4.45, 3.55,
1.98 (0.075)
4.44 3.52,
1.98 (0.075)
1H NMR
: py-H −17.3
−13.5
58.7, 59.7
0.96
−7.5
65.7
1.04
−5.8
54.8
0.97
−7.6
61.9, 62.7
nd
−8.7
59.4, 61.0
0.99
−11.5
63.4, 63.9
0.96
/ppm from TMS
: m-H
64.0, 64.4
0.95
redox potential/V vs
c
SCE
(TMP)FeIII(L)
absorption/nm
EPR: g-value
411, 512, 578,
693
413, 509, 579, 414, 502, 567, 418, 509, 576, 415, 503, 566, 414, 503, 569, 414, 484, 548,
693
675
696
676
678
615
6.2, 5.7, 2.0 7.9, 6.0, 2.0
3.8, 1.8
6.0, 2.0
6.0, 2.0
6.0, 2.0
6.0, 2.0 6.1, 5.9, 6.0, 2.0
2.0
1H NMR
: py-H 104.7
: m-H 21.5, 19.4
105.3
110.0
113.0
110.3
110.1
112.3
/ppm from TMS
redox potential/V vs SCE
19.3, 17.1
−0.32
17.4, 15.5
−0.46
20.5, 18.0
−0.45
17.5, 15.5
−0.48
17.6, 15.6
−0.51
15.7, 14.0
−0.63
−0.18
Second-Order Reaction Rate Constant at 213 K (k2: M−1 s−1
)
cyclooctene (×10−2
N,N-dimethyl-p-
nitroaniline
)
1.5 0.1
1.6 0.1
4.7 0.1
3.2 0.1
5.0 0.1
6.3 0.1
9.1 0.2
6.4 0.1
14.9 0.1
12.6 0.1
20.1 0.3
14.9 0.2
29.4 4.4
29.7 2.0
1,4-cyclohexadiene
3.6 0.1
5.9 0.1
11.9 0.1
12.6 0.1
16.4 0.2
19.8 0.1
24.0 0.1
Thermodynamic Parameters at 213 K (kJ/mol)
Δ(ΔGEX
)
0.0
7.7
11.5
22.6
18.2
20.3
34.1
ΔG⧧
59.0 0.1
50.7 0.1
49.3 0.1
57.0 0.1
49.5 0.1
48.4 0.1
56.9 0.1
48.3 0.1
47.2 0.1
55.8 0.1
48.3 0.1
47.1 0.1
54.9 0.1
47.1 0.1
46.6 0.1
54.4 0.1
46.8 0.1
46.3 0.1
53.7 0.3
45.6 0.1
45.9 0.1
Cyclooctene
ΔG⧧
ΔG⧧
N,N‑Dimethyl‑p‑nitroaniline
1,4‑Cyclohexadiene
a
b
c
Ref 37. Ref 38. Ref 21.
Figure 2. Plot of ln of the reaction rate constants (k2) of the reaction
of (TMP+•)FeIVO(L) with cyclooctene (black circle), N,N-dimethyl-p-
nitroaniline (red triangle), and 1,4-cyclohexadiene (blue square) vs the
redox potential for FeII/FeIII redox process of (TMP)FeIII(L).
Figure 3. 1H NMR spectral change for the titration of (TMP+•)-
FeIVO(NO3) with n-Bu4N(Cl) in CD2Cl2 at −60 °C. (a) (TMP+•)-
FeIVO(NO3), (b) (a) + 0.5 equiv of n-Bu4N(Cl), (c) (a) + 1.0 equiv of
n-Bu4N(Cl), (d) (a) + 1.5 equiv of n-Bu4N(Cl).
titration of (TMP+•)FeIVO(NO3) with tetra-n-butylammonium
chloride (n-Bu41N(Cl)) at −60 °C. Upon addition of n-
Bu4N(Cl), the H NMR spectrum of (TMP+•)FeIVO(NO3)
changes to that of (TMP+•)FeIVO(Cl) because of axial ligand
exchange according to the following equation:
β protons in the 1H NMR spectra. The equilibrium constant for
the ligand exchange reaction was estimated to be K = 6
1
(TMP+•)FeIVO(L1) + nBu4N(L2)
(see Experimental Section). The difference in thermodynamic
stability between (TMP+•)FeIVO(NO3) and (TMP+•)FeIVO-
(Cl) for eq 1 was estimated to be (ΔGEX) = −3.2 0.3 kJ/mol
at 213 K from the relationship, ΔG = −RT ln K.25,26 We
performed similar experiments for (TMP+•)FeIVO(L) with the
remaining ligands. The results are summarized in Supporting
⇄ (TMP+•)FeIVO(L2) + nBu4N(L1)
(1)
where L1 and L2 are nitrate and chloride, respectively. The ratio
of (TMP+•)FeIVO(NO3) to (TMP+•)FeIVO(Cl) was deter-
mined by peak areas of their mesityl meta protons and pyrrole-
7298
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
Information, Table S1 and thermodynamic stabilities of
(TMP+•)FeIVO(L) relative to (TMP+•)FeIVO(NO3),
(ΔGEX)FeO, are shown in Figure 4.
linear behavior is found (Figure 5). Since the logarithm (ln) of
the reaction rate constant indicates the free energy of activation
Figure 5. Plot of ln of the reaction rate constants (k2) of
(TMP+•)FeIVO(L) with cyclooctene (black circle), N,N-dimethyl-p-
nitroaniline (red triangle), and 1,4-cyclohexadiene (blue square) vs the
ΔΔG value estimated from ligand exchange.
(ΔG⧧), the observed linear correlation is consistent with the
Brønsted−Evans−Polanyi relationship, eq 4, which directly
relates the change in the free energy of activation, ΔΔG⧧, to the
corresponding change of free energy of its reaction, ΔΔGr, via a
constant factor α (0 < α <1).27−30
Figure 4. Thermodynamic stabilities (kJ/mol) of (TMP+•)FeIVO(L)
(left) and (TMP)FeIII(L) (right) relative to (TMP+•)FeIVO(NO3) and
(TMP)FeIII(NO3), respectively.
Similarly, ligand exchange reactions for (TMP)FeIII(L) were
also examined with various tetra-n-butylammonium salts as
follows.
ΔΔG⧧ = βΔΔGr
(4)
From the slope of the line in Figure 5, we calculated the α to be
0.15 for the cyclooctene epoxidation reaction. An α value close
to 0 indicates an early transition state that is structurally similar
to the initial reactant state, while an α value close to 1 indicates
a late transition state which is structurally similar to the product
state. The estimated α value indicates an early transition state
for the cyclooctene epoxidation reaction of (TMP+•)FeIVO(L).
The α value estimated in this study was very close to that
(0.18) for an epoxidation reaction of norbornene with
oxochromium(V) porphyrin complex.31 This suggests that the
transition state structure of the epoxidation reaction of
(TMP+•)FeIVO(L) is similar to that of oxochromium(V)
porphyrin complex, which has been proposed as a charge-
transfer complex.31
(TMP)FeIII(L1) + nBu4N(L2)
⇄ (TMP)FeIII(L2) + nBu4N(L1)
(2)
The results are summarized in Supporting Information,
Table S2, and the thermodynamic stability values of (TMP)-
FeIII(L) relative to (TMP)FeIII(NO3), (ΔGEX)Fe(III), are also
summarized in Figure 4. It is obvious that the axial ligand effect
on thermodynamic stability for (TMP)FeIII(L) is more
significant than that for (TMP+•)FeIVO(L).
From these results, we estimated the changes in free energy
of the transition from (TMP+•)FeIVO(L) to (TMP)FeIII(L)
with eq 3 derived from eq 2 − eq 1, where L2 is nitrate.
Consistent linear relationships were also found for other
oxidation reactions of (TMP+•)FeIVO(L). As shown in Figure
5, the logarithm of the reaction rate constants of the reactions
of (TMP+•)FeIVO(L) with 1,4-cyclohexadiene and N,N-
dimethyl-p-nitroaniline show a consistent linear relationship
with the Δ(ΔGEX) values. The α values for 1,4-cyclohexadiene
and N,N-dimethyl-p-nitroaniline were estimated to be 0.10 and
0.15, respectively. The α value for N,N-dimethyl-p-nitroaniline
was the same as that for cyclooctene, but the value for 1,4-
cyclohexadiene was found to be smaller.
Quantum Chemical Calculations. To further study the
axial ligand effect, we performed DFT calculations for
epoxidation reactions of cyclooctene with (TMP+•)FeIVO(L).
To simplify the systems for calculations, we replaced the meso-
mesityl group of the TMP ligand with a hydrogen atom and
performed a calculation for the porphine (Por) complex. The
(TMP+•)FeIVO(NO3) + (TMP)FeIII(L1)
⇄ (TMP+•)FeIVO(L1) + (TMP)FeIII(NO3)
(3)
In eq 3, n-Bu4N(L1) and n-Bu4N(NO3), appearing in eqs 1
and 2, are canceled by the subtraction. Obviously, the ΔG value
for eq 3, Δ(ΔGEX), indicates the difference of the change in free
energy between the transitions from (TMP+•)FeIVO(L1) to
(TMP)FeIII(L1) and from (TMP+•)FeIVO(NO3) to (TMP)-
FeIII(NO3), and can be easily calculated from the difference
between the (ΔGEX)FeO and (ΔGEX)Fe(III), shown in Figure 4.
We calculated Δ(ΔGEX) values for all axial ligands, and the
estimated Δ(ΔGEX) values are listed in Table 1.
When the logarithm (ln) values of the reaction rate constants
for the cyclooctene epoxidation of (TMP+•)FeIVO(L) are
plotted against their estimated Δ(ΔGEX) values, a consistent
7299
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
optimized structures and bond lengths for (Por+•)FeIVO(L)
and (Por)FeIII(L), where L = F, Cl, Bz, Ac, TFA, and NO3, are
shown in Supporting Information, Figures S9 and S10 and
Table S3. Figure 6 shows the energy profiles of the four lowest-
the spectroscopic characterization of (TMP+•)FeIVO(L) and
(TMP)FeIII(L) and previous DFT calculations of other
compound I model complexes.15b,d,17c,18c
Table 2 shows energies of these spin states relative to the
ground state of (Por+•)FeIVO(L) before and after cyclooctene
epoxidation. The energies of these spin states of (Por+•)FeIVO-
(L) and (Por)FeIII(L) were changed by the axial ligand effect.
The axial ligand effect on the energies of the spin states for
(Por)FeIII(L) were found to be more pronounced than those
for (Por+•)FeIVO(L). This is consistent with the ligand
exchange experiments. The energies of the sextet ground
state of (Por)FeIII(L) in Table 2 indicate an enthalpy change
(ΔHDFT) during the cyclooctene epoxidation reaction
promoted by (Por+•)FeIVO(L). The negative ΔHDFT values
indicate exothermic reactions. The ΔHDFT values are changed
by the axial ligands and increase according to the order of L =
TFA < NO3 < Bz < Ac < Cl < F. This order is not identical to
the order of reaction rate constants and Δ(ΔGEX), but is
generally similar.
To compare the axial ligand effect on (Por+•)FeIVO(L) and
(Por)FeIII(L), we calculated the binding energies (ΔEbound) of
the axial ligands (Table 2). The estimated ΔEbound values from
the calculations are fairly large because of the absence of
solvation effects with respect to the free axial anion ligand (the
calculations were made in the gas phase). For both (Por+•)-
FeIVO(L) and (Por)FeIII(L), the ΔEbound value increases
according to the order of L = NO3 < TFA < Cl < Bz < Ac <
F. The ΔEbound values for (Por)FeIII(L) are larger than those for
(Por+•)FeIVO(L) and the axial ligand effect on the ΔEbound
value for (Por)FeIII(L) is more significant than that for
(Por+•)FeIVO(L).
Figure 6. Energy profile for the reaction of (Por+•)FeIVO(F) with
cyclooctene. The structure and states on the left side are for
(Por+•)FeIVO(L) and cyclooctene, and the structure and states on the
right side are for (Por)FeIII(L) and cyclooctene oxide.
lying spin states of (Por+•)FeIVO(L) and (Por)FeIII(L). The
ground states for all of the (Por+•)FeIVO(L) complexes are in
the quartet (S = 3/2) state, in which the two electron spins in
iron dπ (dxz and dyz) orbitals are ferromagnetically coupled with
the porphyrin π-cation radical spin in the a2u orbital. The
doublet (S =1/2) states, in which the iron dπ electron spins
antiferromagnetically couple with the a2u radical spin, are
slightly higher than the ground quartet states. There are two
excited states, the sextet (S = 5/2) and the quartet (S = 3/2)
states, in which four unpaired electrons in iron d-orbitals (dxy,
DISCUSSION
■
Linear Free Energy Relationship for the Axial Ligand
Effect. The observed linear correlation between ln of the
reaction rate constants of (TMP+•)FeIVO(L) and the change of
free energy from (TMP+•)FeIVO(L) to (TMP)FeIII(L) for
various anionic axial ligands indicate that the axial ligand
controls the reactivity of (TMP+•)FeIVO(L) by modulating the
free energy of the overall reaction. This linear free energy
relationship can be understood by examining the two-
dimensional representation of the crossing potential energy
surfaces: curve crossing diagram (Figure 7).30,32 Curve V1
represents the potential energy of the one-dimensional system
before the reaction (the reactant state), and curve V2 follows
that of the product state. The crossing point of the two
potential curves is regarded as the transition state, and the
energy from the local minimum of the reactant state (curve V1)
2
2
dxz, dyz, dx −y ) interact with an unpaired electron in the a2u
orbital. The ground states of (P)FeIII(L) are the sextet (high-
spin S = 5/2) states with unpaired electrons in all five d orbitals.
There were two excited quartet (intermediate-spin S = 3/2)
2
states with unpaired electrons in the dxz, dyz, dz orbitals and dxz,
2
2
dyz, dx −y orbitals. The doublet (low-spin S = 1/2) states have
the highest energies of all states of (Por)FeIII(L) complexes.
The results of the present DFT calculations are consistent with
Table 2. Energies (kJ/mol) of the Four Lowest-Lying Spin States of (Por+•)FeIVO(L) and (Por)FeIII(L) Relative to the Ground
States of (Por+•)FeIVO(L)
(Por+•)FeIVO(L)
quartet2
(Por)FeIII(L)
a
b
L
doublet
quartet
sextet
Ebound
doublet
quartet
quartet2
sextet
Ebound
NO3
TFA
Cl
0.6
0.7
0.6
1.3
0.9
0.9
0.0
0.0
0.0
0.0
0.0
0.0
67.1
66.4
64.7
66.7
65.6
57.0
62.6
61.7
61.4
62.2
61.2
52.3
447.6
457.2
466.4
517.6
478.1
799.2
−107.5
−98.7
−91.6
−107.9
−106.6
−97.2
−163.9
−163.2
−172.8
−169.5
−169.5
−186.6
−110.0
−105.7
−119.9
−116.2
−115.1
−184.1 (0)
489.5 (−41.9/0)
−182.8 (−1.2)
−200.4 (16.3)
−198.1 (14.0)
−196.9 (12.8)
−218.4 (34.3)
497.9 (−40.7/ −1.2)
524.7 (−58.3/16.3)
573.5 (−55.9/14.0)
532.8 (−54.7/12.8)
875.4 (−76.2/34.3)
Ac
Bz
F
−125.8
b
a
The numbers in parentheses are the difference between (Por)FeIII(NO3) and (Por)FeIII(L). The numbers in parentheses are the difference of
binding energies from (Por+•)FeIVO(L) to (Por)FeIII(L) and the change in the difference from the nitrate complex.
7300
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
related to this mechanism. The positive charge of [(TMP+•)-
FeIVO]+ is delocalized over the heme plane because of the
porphyrin π-cation radical character. In contrast, the positive
charge of [(TMP)FeIII]+ resides mainly on the heme iron
center. Therefore, the binding interaction of an anionic axial
ligand with [(TMP)FeIII]+ is stronger than the corresponding
binding to [(TMP+•)FeIVO]+. These factors lead to stabiliza-
tion of [(TMP)FeIII]+ by the anionic axial ligand more than
[(TMP+•)FeIVO]+. More generally, as the reaction stage shifts
from (TMP+•)FeIVO(L) to (TMP)FeIII(L), the binding of the
axial ligand becomes stronger and the stabilization by the axial
ligand becomes larger because the iron−oxygen bond length
becomes longer (Figure 8). As a result, the transition state is
Figure 7. Curve crossing diagram of potential-energy surfaces of the
reactant, (TMP+•)FeIVO(L), and product, (TMP)FeIII(L), states. V1
and V2 indicate the potential curves of (TMP+•)FeIVO(L) and
(TMP)FeIII(L), respectively. TS is the transition state of the reaction.
(a) Change of stability of (TMP+•)FeIVO(L) and (b) change of
stability of (TMP)FeIII(L).
to the crossing point represents the activation energy of the
reaction. The diagram shown in Figure 7 is too simple to
describe a reaction pathway of (TMP+•)FeIVO(L) with a
substrate because many previous studies have proposed
stepwise mechanisms via radical intermediates.18,21,32−34 We
use this simplified diagram to easily understand the observed
linear free energy relationship, and then expand the diagram
into the stepwise mechanism. When the reactant state is
destabilized or the product state is stabilized by the axial ligand
effect, the ΔGr value becomes larger, the transition state moves
to the reactant state, and the activation energy becomes smaller.
In contrast, when the reactant state is stabilized or the product
state is destabilized by the axial ligand, the ΔGr becomes
smaller, the transition state moves to the product state, and the
activation energy becomes larger. As a result, the activation
energy (ln of reaction rate constants) shows a linear correlation
with the free energy of the overall reaction.
Obviously, the activation energy and reaction free energy are
altered by the stabilities of the reactant and product states. The
answer to the following question should be determined: which
state controls these energies? As shown in Figure 4, the axial
ligand affects the stability of both (TMP+•)FeIVO(L) and
(TMP)FeIII(L); however, the axial ligand effect alters the
stability of (TMP)FeIII(L) much more significantly than the
stability of (TMP+•)FeIVO(L). This is further confirmed by the
present DFT calculations (Table 2). The activation energy of
the reaction (the reactivity of (TMP+•)FeIVO(L)) is governed
by the stability of the product state, (TMP)FeIII(L), more than
that of reactant state, (TMP+•)FeIVO(L).
The mechanism we have identified where reactivity of
(TMP+•)FeIVO(L) is controlled by the thermodynamic stability
of (TMP)FeIII(L) is very unique. An essential point to consider
is why the axial ligand changes the stability of (TMP)FeIII(L)
much more than that of (TMP+•)FeIVO(L). This can be
explained by the trans-ligand effect of the axial ligand.
(TMP+•)FeIVO(L) is a six-coordinate complex, but (TMP)-
FeIII(L) is a five-coordinate complex. The binding of the axial
ligand (L) to [(TMP+•)FeIVO]+ becomes weaker than the
binding of (L) to [(TMP)FeIII]+ because of the trans-ligand
effect provided by the strong oxo ligand. Moreover, the
porphyrin π-cation radical character of compound I also may be
Figure 8. Cartoon of the axial ligand effect on energy profiles along the
reaction of (TMP+•)FeIVO(L) with a substrate in a stepwise
mechanism. The black lines are the cases for the weak binding axial
ligand and the red line is that for strong ones.
stabilized more than (TMP+•)FeIVO(L) by the axial ligand, and
the activation energy becomes smaller as the binding of the
axial ligand becomes stronger. This is a key point of the axial
ligand effect. Previously, the axial ligand effect had been
believed to represent the electron-donor effect to compound I
and an intermediate and the electron affinity of compound
I.15−18 However, as shown here, the essence of the axial ligand
effect is provided by the extent of the stabilization of the
product state by the binding of the axial ligand. An axial ligand
that binds stronger makes compound I more reactive.
The diagram shown in Figure 7 represents the reaction
surface for the direct (concerted) oxygen transfer to a substrate.
Since the previous studies proposed that reactions of
(TMP+•)FeIVO(L) with substrates proceed stepwise via radical
intermediates,18,21,32,34 we discuss the axial ligand effect on the
reactivity for the stepwise mechanism. In the stepwise
mechanism, a radical intermediate is more stable than the
transition state for the direct oxygen transfer and the potential
curve for the radical intermediate cuts through the higher-
energy ridge for the direct oxygen transfer and splits the
process into two steps (Figure 9). Obviously, the transition
state of the first step (TS1) is controlled by the stabilities of
(TMP+•)FeIVO(L) and the radical intermediate, but it does not
directly connect with (TMP)FeIII(L), and the transition state of
the second step (TS2) is controlled by the stabilities of the
radical intermediate and (TMP)FeIII(L). As discussed above,
the axial ligand effect becomes stronger in the order of
7301
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
Figure 9. Curve crossing diagram of potential-energy surfaces from (TMP+•)FeIVO(L) to (TMP)FeIII(L) for a stepwise mechanism. V1 and V2
indicate the potential curves of (TMP+•)FeIVO(L) and (TMP)FeIII(L), respectively. VRI shows the potential curve of a radical intermediate (R.I.).
E−1 is the activation energy from R.I. to (TMP+•)FeIVO(L) and E2 is the activation energy from R.I. to (TMP)FeIII(L). (a) E1 ≈ E2, (b) E1 ≪ E2, and
(c) E1 ≫ E2.
(TMP+•)FeIVO(L) < TS1 < the radical intermediate < TS2 <
(TMP)FeIII(L) because the iron−oxygen bond length increases
in this order (Figure 8).
or (Por)FeOH(L), this study suggests that it is controlled by
the stability of (Por)FeOH(L). However, the calculated
BDEO−H values do not correlate with the experimental results
for the reactivity of (TMP+•)FeIVO(L) reported here and in a
previous paper.20 The order (L = Cl < TFA ≈ Ac < F) of the
reactivity estimated from calculated BDEO−H values is different
from the experimental results (TFA < Ac < Cl < F).18c In
addition, (TMP+•)FeIVO(L) has been reported to be very
reactive when L is imidazole,20 but it is expected to be very
inactive from BDEO−H calculations.18a,b Solvent effect may
affect the stability of the radical intermediate and (TMP)-
FeIII(L), leading to the change of the reactivity between in gas
phase and in solution.
The linear free energy relationship was observed not only for
epoxidation of cyclooctene, but also for demethylation of N,N-
dimethyl-p-nitroaniline and hydrogen abstraction from 1,4-
cyclohexadiene. Although the reaction mechanism and
transition state of these reactions are different from those of
the epoxidation reaction, the same mechanism is also applicable
to the axial ligand effect on the reactivity for these reactions.
In this study, the linear free energy relationship was found for
various anionic axial ligands. As reported previously,20 a drastic
axial ligand effect has been found for neutral axial ligands such
as imidazole. The present mechanism for the anioic axial ligand
effect would also be applicable to the neutral axial ligand effect.
However, the correlation line for the neutral axial ligand would
differ from the line for the anionic axial ligand shown here
because the Fe-L bond character is different between the
neutral ligand and the anionic ligand.
Structure and Electronic State of Transition State. As
discussed above, the small α values for the reactions of
(TMP+•)FeIVO(L) with cyclooctene, N,N-dimethyl-p-nitroani-
line, and 1,4-cyclohexadiene indicate the existence of early
transition states that are structurally close to the initial reactant
state. Therefore, in the transition state, the oxo ligand still binds
to the iron center and weakly interacts with the substrate.
Previous studies proposed reaction mechanisms for epoxidation
of cyclooctene and a demethylation reaction of dimethylaniline
from the kinetic analysis, isotope experiments, and redox
potentials of oxoiron(IV) porphyrin π-cation radical com-
plexes.21,33 In an epoxidation reaction, weak orbital interactions
between cyclooctene and the FeO moiety of the oxoiron(IV)
porphyrin π-cation radical complex may be formed at the
transition state, and the electron transfer may occur in the bond
formation process between the oxo-ligand of FeO and CC
moiety of olefin. On the other hand, the demethylation reaction
by (TMP+•)FeIVO(L) was proposed to proceed via electron
transfer followed by H-atom transfer which competes with back
There are three possible situations in the stepwise
mechanism (Figure 9). When the activation energy (E−1)
from the radical intermediate to (TMP+•)FeIVO(L) is close to
that (E2) from the radical intermediate to (TMP)FeIII(L)
(Figure 9a), the stability of (TMP)FeIII(L) controls the
reactivity of (TMP+•)FeIVO(L) because it changes the
activation energy E2. This is confirmed by a kinetic analysis
for the stepwise mechanism, in which the overall reaction rate
(k) can be derived as k = k1/(1 + k−1/k2) using the steady-state
approximation, where k1 is the reaction rate from (TMP+•)-
FeIVO(L) to the radical intermediate, k−1 is the reaction rate of
the back reaction, and k2 is the reaction rate from the radical
intermediate to (TMP)FeIII(L). When E−1 becomes much
smaller than E2 (Figure 9b, k−1 ≫ k2), the overall reaction rate
can be simplified as k = k1k2/k−1. This is the case where the
potential curve of the radical intermediate shifts to that of
(TMP+•)FeIVO(L) or (TMP)FeIII(L) is much destabilized
more than (TMP+•)FeIVO(L). TS2 is the transition state of the
overall reaction. Therefore, the stability of (TMP)FeIII(L)
controls the reactivity of (TMP+•)FeIVO(L) even in this
situation. On the other hand, when E−1 becomes much larger
than E2 (Figure 9c, k−1 ≪ k2), the overall reaction rate can be
simplified as k = k1. This is the case where the potential curve of
the radical intermediate shifts to that of (TMP)FeIII(L) or
(TMP)FeIII(L) is much more stabilized than (TMP+•)FeIVO-
(L). TS1 is the transition state of the overall reaction. The
stability of (TMP)FeIII(L) does not directly control the
reactivity of (TMP+•)FeIVO(L) in this situation. If the reaction
occurs in this situation, TS1 must be controlled by the stability
of the radical intermediate because a good correlation of the
reaction rate with the stability of (TMP+•)FeIVO(L) could not
be observed. This is consistent with the axial ligand effect
discussed here (Figure 8). Moreover, the observed correlation
of the reactivity of (TMP+•)FeIVO(L) with the stability of
(TMP)FeIII(L) suggests the strong correlation of the radical
intermediate with (TMP)FeIII(L). When the spin state of the
radical intermediate is the same as that of (TMP)FeIII(L), the
diagram of Figure 9c can be simplified as Figure 7b and the
axial ligand effect would be interpreted with the stability of
(TMP)FeIII(L).
Previous computational studies linked the axial ligand effect
with the bond dissociation energy (BDEO−H) of the O−H bond
in (Por)FeOH(L).18,19 This is also a linear free energy
relationship and can be interpreted by Figure 7. While
BDEO−H is changed by the stability of (Por+•)FeIVO(L) and/
7302
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
electron transfer. For N,N-dimethyl-p-nitroaniline, the H-atom
of the methyl group of the formed amine radical would weakly
interact with the FeO moiety. The small α value is consistent
with the proposed transition states.
present compound I model complexes. In fact, we recently
showed that the reactivity of compound I model complexes
with imidazole and phenolate axial ligands is 100−400-fold
higher than that of the compound I model with nitrate as the
axial ligand.20 The very large negative shift of the redox
potential for the FeII/FeIII redox process for iron(III) porphyrin
thiolate complexes (−1.19 V vs SCE) also indicates that the
resting state is very stable.36 This suggests that compound I of
cytochrome P450 is much more reactive than compound I of
peroxidases, catalases, and synthetic heme complexes. The
mechanism by which compound I gains high reactivity by
stabilizing the resting state, and not by destabilizing the
compound I state, is expected to be convenient for the
oxygenation reactions of heme enzymes because destabilization
of compound I might lead to inactivation of the enzymes
because of unfavorable side reactions involving oxidation of
amino acid residues of the protein in the vicinity of the heme
active site.
The α value for the hydrogen abstraction of (TMP+•)FeIVO-
(L) is smaller than the corresponding values for the epoxidation
and demethylation reactions, and may suggest the existence of
an early transition state. In the transition state, the FeO
moiety of (TMP+•)FeIVO(L) would interact with the proton
abstracted from 1,4-cyclohexadiene. However, the hydrogen
atom tunneling effect must be considered for interpretation of
the α value for the hydrogen abstraction reaction. A hydrogen
abstraction reaction with a large ΔG⧧ value, such as for
(TMP+•)FeIVO(NO3), may proceed via not only a normal
reaction pathway but also via a hydrogen atom tunneling
pathway. This leads to a decrease in the α value.
The ground states of (TMP+•)FeIVO(L) and (TMP)FeIII(L)
are quartet (S =3/2) and sextet (S = 5/2) states, respectively.
Thus, the spin state must be switched from quartet to sextet in
the reaction of (TMP+•)FeIVO(L). It is likely that this occurs
near the transition state. Since the energy of the transition state
is governed by the stability of the ground state of the product,
as discussed above, the energy of the sextet state is a key to
determining its reactivity. The sextet state has not been
extensively studied as a transition state in previous theoretical
studies.35 In most cases, the ground (quartet) state and the
lowest excited (doublet) states have been examined in efforts to
explain the reactivity of compound I, and the doublet ground
state was proposed for the transition state.17,18 The present
DFT calculations showed that the energies of the sextet states
are 50−60 kJ/mol above the quartet (ground) states for
(TMP+•)FeIVO(L), and these energies are close to the ΔG⧧
values determined at 213 K (Table 1 and 2). With the substrate
coming close to and interacting with (TMP+•)FeIVO(L), the
FeO bond of (TMP+•)FeIVO(L) becomes weaker and
longer, and the energy of the quartet (ground) state increases.
At last, the ground state is switched over from the quartet state
to the sextet state near the transition state. After switching to
the sextet state, the reaction would proceed along the energy
surface of the sextet state and reach the sextet state of the final
product of (TMP)FeIII(L).
EXPERIMENTAL SECTION
■
Instrumentation. UV−visible absorption spectra were recorded
on an Agilent 8453 spectrometer (Agilent Technologies) equipped
with a USP-203 low-temperature chamber (UNISOKU). H NMR
1
spectra were measured on a Lambda-500 spectrometer (JEOL).
Chemical shifts were referenced to the residual peak of dichloro-
methane, (5.32 ppm). The concentrations of NMR samples were 1−3
mM. EPR spectra were recorded in a quartz cell (d = 5 mm) on an
E500 continuous wave X-band spectrometer (Bruker) with an ESR910
helium-flow cryostat (Oxford Instruments). Gas chromatography−
mass spectroscopy (GC-MS) analysis was performed on a QP-5000
GC-MS system (Shimadzu) equipped with a capillary gas chromato-
graph (GC-17A, CBP5-M25-025 capillary column). Cyclic voltammo-
grams and differential pulse voltammograms were measured with an
ALS612A electrochemical analyzer in degassed dichloromethane
−
containing 0.1 M nBu4N+ClO4 (TBAP) as a supporting electrolyte.
A glassy carbon electrode was used as the working electrode and a
platinum-wire electrode was employed as the counter electrode. The
potentials were recorded with respect to a saturated calomel electrode
(SCE) as a reference electrode. Resonance Raman scattering was
excited at 406.7 nm from a Kr+ laser (Spectra Physics, BeamLok
2060), dispersed by a single polychromator (Ritsu Oyo Kogaku, MC-
1000DG), and detected with a liquid-nitrogen cooled CCD detector
(Roper Scientific, LNCCD-1100-PB). Laser power at the sample was
20 mW. The mechanical slit width was 200 μm. Sample concentrations
(∼1 mM) were prepared in an NMR tube (O.D. = 5 mm) at −80 °C
and set in a custom-designed glass cell with optical windows, which
was precooled at −80 °C with cold dry nitrogen gas. Resonance
Raman spectra were measured with 135° backscattering geometry with
rapid spinning of the sample tube.
Materials. Anhydrous dichloromethane was obtained commercially
and stored in the presence of 4 Å molecular sieves. Cyclooctene and
1,4-cyclohexadiene were purchased commercially and used after
passing through activated alumina immediately prior to use to remove
polar impurities. Other chemicals were purchased commercially and
used without further purification. meso-Tetramesitylporphyrin
(TMPH2) was prepared according to a previously published
procedure.39 (TMP)FeIIICl was prepared by insertion of iron into
TMPH2 with FeCl2 and sodium acetate in acetic acid, and purified
with a silica gel column using CH2Cl2/CH3OH as an eluent.40
(TMP)FeIIICl was passed through a basic alumina (10% water)
column with dichloromethane as an eluent to produce (TMP)-
FeIIIOH.21 The (TMP)FeIII(L) complexes were obtained from
reactions of (TMP)FeIII(OH) with corresponding acids (F: hydro-
fluoric acid, Ac: acetic acid, Bz: benzoic acid, TFA: trifluoroacetic acid,
Hc: hydrocinnamic acid, and NO3: nitric acid) in toluene solution and
purified by recrystallization from ether/n-hexane.41
The stabilization of the sextet state of (TMP)FeIII(L) by the
axial ligand effect is also manifested in the energy of the sextet
state of (TMP+•)FeIVO(L). The calculated energy gap between
the quartet ground state and the sextet state which is changed
by the axial ligand correlates with the ΔH⧧ value estimated for
cyclooctene epoxidation for (TMP+•)FeIVO(L) in this study
(Supporting Information, Table S4). The small energy gap
between the sextet state of (TMP+•)FeIVO(L) and the ΔG⧧
value is reasonable with respect to the early transition state
suggested from the present α values determined in this study.
Axial Ligand Effect in Heme Enzymes. The protein axial
ligands of peroxidase, catalases and cytochrome P450 are
histidine (imidazole), tyrosine (phenolate), and cystein
(thiolate), respectively.1,6,11 Since the five-coordinate ferric
heme complexes of these heme enzymes are in the high spin (S
= 5/2) state, as they are in the present model study, it is
reasonable to expect that an axial ligand effect exists in these
heme enzymes which is similar to that of the present model
system. Since the fluoride axial ligand, the strongest ligand
employed in this study, can be easily replaced by imidazole,
phenolate, and thiolate ligands, the compound I intermediates
of these heme enzymes would be more reactive than the
7303
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
Kinetics and Product Analysis. (TMP+•)FeIVO(L) (100 μM) in
dichloromethane was prepared in a 1 cm quartz cuvette in a low-
temperature chamber in a UV−visible absorption spectrometer by
bubbling O3 gas at low temperature (−30 to −80 °C). After
generation of (TMP+•)FeIVO(L), excess O3 gas was removed by
bubbling Ar gas. An excess (10−1000 equiv) of cyclooctene, 1,4-
cyclohexadiene, or N,N-dimethyl-p-nitroaniline was then added to the
solution with vigorous stirring, and the reactions were monitored by
the absorption spectral change at constant time intervals. The reaction
rate constants were determined by a curve fit of the time dependence
of the absorbance of the peaks in the vicinity of 500 and 667 nm. The
absorption spectral change followed the first-order kinetics in the
presence of the large excess of substrates, and the second-order
reaction rate constants (k2) were determined from the linear
dependence of the pseudo first-order rate constants on the
concentrations of these substrates (Supporting Information, Figure
S1). Activation of enthalpy (ΔH⧧) and activation of entropy (ΔS⧧)
values for the cyclooctene epoxidation reactions by (TMP+•)FeIVO(L)
were estimated from the Eyring plots (Supporting Information, Table
S4). The free energy of activation (ΔG⧧) values were calculated from
the second-order reaction rate constants with the relationship, ΔG⧧ =
−RT ln(hk2/kT).42 The yield of cyclooctene oxide was determined by
GC-MS using undecane as an internal standard. After completion of
the reactions, 10 equiv of tetra-n-butylammonium iodide (n-Bu4N+I−)
was added to the reaction solutions at low temperature. The reaction
mixture was taken from the low-temperature chamber and warmed to
room temperature. After addition of undecane, the reaction mixture
was analyzed by GC-MS. Yields of cyclooctene oxide were determined
from a standard calibration curve prepared with authentic cyclooctene
oxide.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figures S1−S10 and Table S1−S4. Complete ref 43. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +81-564-59-5578. E-mail: hiro@ims.ac.jp.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was supported by grants from JSPS (Grant-in-Aid
for Scientific Research, Grant No. 22350030 to H.F. and No.
22018026 to T.O.) and MEXT (Global COE Program).
REFERENCES
■
(1) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. Rev.
1996, 96, 2841−2887.
(2) Krebs, C.; Fujimori, D. G.; Walsh, C. T.; Bollinger, J. M., Jr. Acc.
Chem. Res. 2007, 40, 484−492.
(3) Groves, J. T.; Han, Y.-Z. In Cytochrome P450: Structure,
Mechanism, and Biochemistry; Ortiz de Montellano, P. R., Ed.; Plenum:
New York, 1995; pp 3−48.
(4) Watanabe, Y.; Fujii, H. Struct. Bonding (Berlin, Ger.) 2000, 97,
61−89.
Ligand Exchange. The relative energies for (TMP+•)FeIVO(L)
and (TMP)FeIII(L) were estimated through the ligand-exchange
reactions. Ligand-exchange reactions of (TMP+•)FeIVO(L1) or
(5) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. Rev.
2004, 104, 939−986.
1
FeIII(TMP)(L1) with nBu4N(L2) were monitored by using H NMR
(6) Dunford, H. B. Heme Peroxidases; Wiley-VCH: New York, 1999.
(7) Schonbaum, G. R.; Chance, B. In The Enzymes; Boyer, P. D., Ed.;
Academic Press: New York, 1979; Vol. 13, pp 363−408.
(8) Gouet, P.; Jouve, H.-M.; Williams, P. A.; Andersson, I.;
Andreoletti, P.; Nussaume, L.; Hajdu, J. Nat. Struct. Biol. 1996, 3,
951−956.
spectra. Samples of (TMP+•)FeIVO(L1) in CD2Cl2 for ligand-exchange
reactions were prepared by ozone oxidation reactions in NMR tubes.
To remove trace quantities of HCl, CD2Cl2 was passed through
activated alumina just prior to use. nBu4N(L2) compounds in CD2Cl2
were slowly titrated in the NMR tubes until the majority of the
(TMP+•)FeIVO(L1) complex was converted to (TMP+•)FeIVO(L2).
The addition of nBu4N(L2) showed a spectral change, indicating the
progression of the ligand-exchange reaction without reduction of the
porphyrin π-cation radical.
(9) Schlichting, I.; Berendzen, J.; Chu, K.; Stock, A. M.; Marves, S. A.;
Benson, D. E.; Sweet, R. M.; Ringe, D.; Petsko, G. A.; Sligar, S. G.
Science 2000, 287, 1615−1622.
(10) Rittle, J.; Green, M. T. Science 2010, 330, 933−937.
(11) Zamocky, M.; Furtmuller, P. G.; Obinger, C. Antioxid. Redox
̈
The equilibrium constant, K, for the ligand-exchange reactions,
(TMP+•)FeIVO(L1) + nBu4N(L2) ⇄ (TMP+•)FeIVO(L2) + nBu4N-
(L1), is defined as
Signaling 2008, 10, 1527−48.
(12) Ohno, T.; Suzuki, N.; Dokoh, T.; Urano, Y.; Kikuchi, K.;
Hirobe, M.; Higuchi, T.; Nagano, T. J. Inorg. Biochem. 2000, 82, 123−
125.
K = ([(TMP+•)FeIVO(L2)][nBu4N(L1)])
/([(TMP+•)FeIVO(L1)][nBu4N(L2)])
= α2/{(1 − α)(n − α)}
(13) Gross, Z. J. Biol. Inorg. Chem. 1996, 1, 368−371.
(14) Higuchi, T.; Uzu, S.; Hirobe, M. J. Am. Chem. Soc. 1990, 112,
7051−7052.
(15) (a) Song, W. J.; Ryu, Y. O.; Song, R.; Nam, W. J. Biol. Inorg.
Chem. 2005, 10, 294−304. (b) Kamachi, T.; Kouno, T.; Nam, W.;
Yoshizawa, K. J. Inorg. Biochem. 2006, 101, 1464−1472. (c) Kang, Y.;
Chen, H.; Jeong, Y. J.; Lai, W.; Bae, E. H.; Shaik, S.; Nam, W. Chem.
Eur. J. 2009, 15, 10039−10046. (d) de Visser, S. P.; Tahsini, L.; Nam,
W. Chem.Eur. J. 2009, 15, 5577−5587.
where α denotes the molar fraction of (TMP+•)FeIVO(L2) and n is the
molar ratio of added nBu4N(L2) over total heme. The α value for each
titration experiment was determined from the peak areas for m-H, o-
Me, p-Me, and pyrrole-H signals of (TMP+•)FeIVO(L1) and
(TMP+•)FeIVO(L2). The K values were determined by performing a
least-squares curve fit of the plot of α against n. The equilibrium
constants (K) for (TMP)FeIII(L) for the ligand-exchange reactions,
(TMP)FeIII(L1) + nBu4N(L2) ⇄ (TMP)FeIII(L2) + nBu4N(L1), were
also estimated using the same method.
Quantum Chemical Calculations. The quantum chemical
calculation was performed using the Gaussian09 program package.43
The 6-31+G* basis set was used for the Fe atom. The 6-31G* basis set
was used for the N atoms of the porphyrin and the atoms of the axial
ligands, and the 6-31G basis set was used for the other atoms. The
optimized structures and energies of the spin states were calculated
using an unrestricted B3LYP density functional method.
(16) Rydberg, P.; Sigfridsson, E.; Ryde, U. J. Biol. Inorg. Chem. 2004,
203−223.
(17) (a) Kumar, D.; de Visser, S. P.; Sharma, P. K.; Derat, E.; Shaik,
S. J. Biol. Inorg. Chem. 2005, 10, 181−189. (b) de Visser, S. P. J. Biol.
Inorg. Chem. 2006, 11, 168−178. (c) Renyue, W.; de Visser, S. P. J.
Inorg. Biochem. 2007, 101, 1464−1472.
(18) (a) de Visser, S. P. J. Am. Chem. Soc. 2010, 132, 1087−1097.
(b) Kumar, D.; Karamzadeh, B.; Sastry, G. N.; de Visser, S. P. J. Am.
Chem. Soc. 2010, 132, 7656−7667. (c) Kumar, D.; Sastry, G. N.; de
Visser, S. P. J. Phys. Chem. B 2012, 116, 718−730.
(19) Prokop, K. A.; de Visser, S. P.; Goldberg, D. P. Angew. Chem.,
Int. Ed. 2010, 49, 5091−5095.
7304
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305

Inorganic Chemistry
Article
(20) Takahashi, A.; Kurahashi, T.; Fujii, H. Inorg. Chem. 2009, 48,
2614−2625.
(21) Takahashi, A.; Kurahashi, T.; Fujii, H. Inorg. Chem. 2011, 50,
6922−6928.
(22) Fujii, H. Coord. Chem. Rev. 2002, 226, 51−60.
(23) Scheidt, W. R.; Gouterman, M. In Iron Porphyrins; Lever, A. B.
P., Gray, H. B., Eds.; Addison-Wesley: London, U.K., 1983; Part I, pp
89−139.
(24) Kadish, K. M.; Rhodes, R. K. Inorg. Chem. 1983, 22, 1090−1094.
(25) Jones, R. W.; Staley, R. H. J. Am. Chem. Soc. 1982, 104, 2296−
2300.
(26) Operti, L.; Tews, E. C.; Freiser, B. S. J. Am. Chem. Soc. 1988,
110, 3847−3853.
(27) Brønsted, J. N. Chem. Rev. 1928, 5, 231−338.
(28) Eyring, H.; Polanyi, M. Z. Z. Phys. Chem., Abt. B 1931, 12, 279.
(29) Evans, M. G.; Polanyi, M. Trans. Faraday Soc. 1938, 34, 11−24.
(30) van Santen, R. A.; Neurock, M.; Shetty, S. G. Chem. Rev. 2010,
110, 2005−2048.
(31) Garrison, J. M.; Ostovic, D.; Bruice, T. C. J. Am. Chem. Soc.
1989, 111, 4960−4966.
(32) Shaik, S.; Lai, W.; Chen, H.; Wang, Y. Acc. Chem. Rev. 2010, 42,
1154−1165.
(33) Goto, Y.; Watanabe, Y.; Fukuzumi, S.; Jones, J. P.; Dinnocenzo,
J. P. J. Am. Chem. Soc. 1998, 120, 10762−10763.
(34) (a) Groves, J. T.; McClusky, G. A. J. Am. Chem. Soc. 1976, 98,
859−861.
(35) (a) Schoneboom, J. C.; Neese, F.; Thiel, W. J. Am. Chem. Soc.
̈
2005, 127, 5840−5853. (b) Hirao, H.; Kumar, D.; Thiel, W.; Shaik, S.
J. Am. Chem. Soc. 2005, 127, 13007−13018.
(36) Tani, F.; Matsu-ura, M.; Nakayama, S.; Ichimura, M.; Nakamura,
N.; Naruta, Y. J. Am. Chem. Soc. 2001, 123, 1133−1142.
(37) Czarnecki, K.; Nimri, S.; Gross, Z.; Proniewicz, L. M.; Kincaid, J.
R. J. Am. Chem. Soc. 1996, 118, 2929−2935.
(38) Paeng, K.-J.; Shin, D.-H.; Paeng, I. R. Inorg. Chem. 1997, 36,
2901−2903.
(39) Lindsey, J.; Wagner, R. J. Org. Chem. 1989, 54, 828−836.
(40) Fujii, H. J. Am. Chem. Soc. 1993, 115, 4641−4648.
(41) Fuhrhop, J.-H.; Smith, K. M. In Porphyrins and Metal-
loporphyrins; Smith, K. M., Ed.; Elsevier: Amsterdam, The Nether-
lands, 1975; pp 757−869.
(42) Eyring, H. J. Chem. Phys. 1935, 3, 107−115.
(43) Frisch, M. J. et al. Gaussian09, Revision A.1; Gaussian, Inc.:
Wallingford, CT, 2009.
7305
dx.doi.org/10.1021/ic3006597 | Inorg. Chem. 2012, 51, 7296−7305