Roles of carboxylate donors in O-O bond scission of peroxodi-iron(iii) to high-spin oxodi-iron(iv) with a new carboxylate-containing dinucleating ligand

Article abstract of DOI:10.1039/c3sc51541a

Dioxygen activation proceeds via O-O bond scission of peroxodi-iron(iii) to high-spin oxodi-iron(iv) in soluble methane mono-oxygenase (sMMO). Recently, we have shown that reversible O-O bond scission of peroxodi-iron(iii) to high-spin oxodi-iron(iv) is a

Full text of DOI:10.1039/c3sc51541a

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Roles of carboxylate donors in O–O bond scission
of peroxodi-iron(III) to high-spin oxodi-iron(IV) with
a new carboxylate-containing dinucleating ligand†
a
Cite this: Chem. Sci., 2014, 5, 2282
Masahito Kodera, Tomokazu Tsuji,a Tomohiro Yasunaga,a Yuka Kawahara,a
*
Tomoya Hirano,a Yutaka Hitomi,a Takashi Nomura,b Takashi Ogura,b
Yoshio Kobayashi,c P. K. Sajith,d Yoshihito Shiotad and Kazunari Yoshizawad
Dioxygen activation proceeds via O–O bond scission of peroxodi-iron(III) to high-spin oxodi-iron(IV) in soluble
methane mono-oxygenase (sMMO). Recently, we have shown that reversible O–O bond scission of
peroxodi-iron(III) to high-spin oxodi-iron(IV) is attained with a bis-tpa type dinucleating ligand, 6-hpa. In this
study, a new carboxylate-containing dinucleating ligand, 1,2-bis[2-(N-2-pyridylmethyl-N-glycinylmethyl)-
6-pyridyl]ethane (H2BPG2E) and its m-oxodiaquadi-iron(III) complexes [Fe2(m-O)(H2O)2(BPG2E)]X2 [X ¼ ClO4
(2a) or OTf (2b)] were synthesized to mimic a common carboxylate-rich coordination environment in
O2-activating non-heme di-iron enzymes including sMMO. The crystal structures of 2a and 2b revealed
that BPG2E prefers a syn-diaqua binding mode. 2b catalyzed the epoxidation of alkenes with H2O2. A new
purple species was formed upon reaction of 2b with H2O2, and characterized by the elemental analysis
and spectral and kinetic studies. These clearly showed that the purple species was a m-oxo-m-peroxodi-
iron(III), and converted to high-spin m-oxodioxodi-iron(IV) via rate-determining reversible O–O bond
scission. In comparison of BPG2E with 6-hpa, it is shown that the carboxylate donor stabilizes the Fe–O–
O–Fe structure of the peroxo complex due to the structural effect to retard O–O bond scission. This may
shed light on the roles of carboxylate donors in the dioxygen activation of non-heme di-iron enzymes.
Received 1st June 2013
Accepted 12th February 2014
DOI: 10.1039/c3sc51541a
www.rsc.org/chemicalscience
site. However, the reactivity and spectral features of the inter-
mediates vary according to the enzyme. In sMMO, the peroxo
Introduction
intermediate P is converted to the high-spin oxodi-iron(IV) active
species Q.2,6 In TMO and ToMO, the peroxo species are
Peroxodi-iron(III) and high-spin oxodi-iron(IV) are formed as key
intermediates in dioxygen activation of O2-activating non-heme
di-iron enzymes,1 including soluble methane mono-oxygenase
(sMMO),2 toluene/o-xylene mono-oxygenase (ToMO),3 toluene
mono-oxygenase (TMO),4 deoxyhypusine hydroxylase (DOHH).5
The di-iron centers are bound at a common carboxylate-rich
unstable, showing almost no electronic absorption feature with
3,7
¨
unique Mossbauer parameters. In DOHH, the peroxodi-
iron(III) is a resting form, stable without substrate, and upon
addition of substrate, rapidly oxidizes it.5 The carboxylate donor
may control the reactivity of key intermediates, but the roles in
the dioxygen activation and the substrate oxidation have not
been claried yet.
aDepartment of Molecular Chemistry and Biochemistry, Doshisha University, Tatara
Miyakotani 1-3, Kyotanabe Kyoto 610-0321, Japan. E-mail: mkodera@mail.
doshisha.ac.jp; Fax: +81-774-65-6848; Tel: +81-774-65-6652
bDepartment of Life Science, University of Hyogo, Kouto 2-1, Ako-gun Kamigori-cho,
Hyogo 678-1297, Japan. E-mail: ogura@sci.u-hyogo.ac.jp; Fax: +81-791-58-0181;
Tel: +81-791-58-0181
cGraduate School of Informatics and Engineering, The University of Electro-
Communications, Chofugaoka 1-5-1, Chofu, Tokyo 182-8585, Japan. E-mail:
kyoshio@pc.uec.ac.jp; Fax: +81-42-443-5555; Tel: +81-42-443-5555
dInstitute for Materials Chemistry and Engineering, Kyushu University, Fukuoka 819-
0395, Japan. E-mail: kazunari@ms.ifoc.kyushu-u.ac.jp; Fax: +81-92-802-2529; Tel:
+81-92-802-2528
Some peroxodi-iron(III) complexes with carboxylate-containing
ligands mimicking carboxylate-rich sites have been reported.8 Que
reported the formation of a peroxodi-iron(III) with 2,6-di-mesi-
tylbenzoate.9 With a less sterically hindered ligand, 2,6-di-p-tol-
ylbenzoate, Lippard detected di-m-oxodi-iron(III)(IV).10 These
species, however, are so unstable that it is difficult to clarify the
O2-activation mechanism. Suzuki determined the crystal struc-
tures of m-oxo-m-peroxodi-iron(III) of a carboxylate-containing
ligand 6-Me2-BPP [Fe2(m-O)(m-O2)(6-Me2-BPP)2] and its m-hydroxo
form, both of which, however, do not undergo the O–O bond
scission and show no reactivity for substrate oxidation.11 Lippard
reported generation of a peroxo species upon reaction of m-oxo-
diaquadi-iron(III) of a carboxylate-containing dinucleating ligand,
H2BPG2DEV, [Fe2(m-O)(H2O)2(BPG2DEV)](ClO4)2 with H2O2, which
† Electronic supplementary information (ESI) available: X-ray analysis data of 2a
and 2b. ESI MS spectra of 2b and 18O-labeled 2b. Isotope peaks of the CSI MS
18
¨
spectrum for fully O-labeled 3. Moossbauer spectra of 2a and the decomposed
product of 3. Oxidation of alkene with H2O2 catalyzed by 2b. CCDC
978448–978449. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c3sc51541a
2282 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
shows low epoxidation activity.12 Thus, the formation of high-spin comparison of the BPG2E complexes with the corresponding 6-
di-iron(IV) and the efficient catalytic oxidation of hydrocarbons hpa complexes.
have never been attained by peroxo complexes of carboxylate-
containing ligands.
Results and discussion
Recently, we have shown that efficient catalytic epoxidation of
various alkenes with H2O2 (ref. 13a) and spectroscopic charac-
terization of peroxodi-iron(III) and high-spin oxodi-iron(IV)
The BPG2E ligand specically stabilizes m-oxodi-iron(III) core in a
solution as shown by the ESI MS spectrum of 2b where the
generated via reversible O–O bond scission in the H2O2 activa-
parent peak due to {[Fe2O(BPG2E)]OTf}+ appears at m/z 815 as a
tion13b are attained using a m-oxodiaquadi-iron(III) complex of a
major one (Fig. S6†). The crystal structures of 2a and 2b were
bis-tpa dinucleating ligand 6-hpa [Fe2(m-O)(H2O)2(6-hpa)](ClO4)4
determined by X-ray analysis with high-resolution quality (see
(1), where 6-hpa is 1,2-bis[2-[bis(2-pyridylmethyl)aminomethyl]-
6-pyridyl]ethane (see Scheme 1(A)). This is a promising func-
tional model capable of reproducing the conversion of peroxo
intermediate P to the active species Q, high-spin oxodi-iron(IV),
Experimental section and ESI†), where a m-oxodiaquadi-iron(III)
core is encapsulated by BPG2E as shown in the ORTEP view of 2a
(Fig. 1). The ORTEP view of 2b is almost identical to that of 2a
(Fig. S5†). The Fe(III) ions adopt distorted octahedral geometry.
The diaqua moiety is in a syn-binding mode with the average
Fe–Oaq bond distances 2.064 and 2.056 A and the Oaq/Oaq
in the catalytic O2-activation cycle of sMMO.
High-spin (S ¼ 2) oxoiron(IV) is postulated as an active species
˚
in nonheme oxygenases,14 and some high-spin (S ¼ 2) oxoiron(IV)
˚
distances 4.025 and 3.926 A for 2a and 2b, respectively. These
complexes have been reported as model compounds.15 Most of
distances are slightly different between 2a and 2b due to the
different hydrogen bonding structures of the diaqua ligands (see
Fig. S1 and S3†). The syn-binding mode of the diaqua moiety is
them, however, are less reactive to external substrate due to the
steric hindrance of the ligands, except for the high reactivity of a
high-spin oxodi-iron(III)(IV) mixed valence complex with a tpa-
common in all m-oxodiaquadi-iron(III) complexes with various
type mononucleating ligand reported by Que et al.16 The catalytic
carboxylate-containing ligands reported so far.12,17 Contrarily,
activity of high-spin oxodi-iron(III)(IV), however, has not been
with tpa-type polypyridine ligands, diaqua moiety generally
shown. Therefore, 6-hpa is the only ligand, known so far, capable
adopts an anti-binding mode in m-oxodiaquadi-iron(III)
of attaining not only the reversible O–O bond scission of per-
complexes, including 1.13a Since the pyridyl (py) group is larger
oxodi-iron(III) to high spin oxodi-iron(IV) but also efficient catal-
ysis of the epoxidation of various alkenes.
A similar dinucleating ligand where the pyridyl groups of 6-
hpa are substituted with the carboxylate groups may be useful
for examining the roles of carboxylate-rich coordination envi-
ronment of O2-activating non-heme di-iron enzymes. So, we
synthesized a new carboxylate-containing dinucleating ligand,
than aqua, two aqua/py pairs occupy syn-positions of m-oxodia-
quadi-iron core in 1 to prevent steric repulsion between py
groups, leading to the anti-diaqua binding mode. Meanwhile,
carboxylate/py pairs can occupy syn-positions of m-oxodiaquadi-
iron core in 2a and 2b since carboxylate group is less sterically
hindered than py, resulting in the syn-diaqua binding mode.
Thus, the size of the pendant carboxyl and py groups seems to x
1,2-bis[2-(N-2-pyrid-ylmethyl-N-glycinylmethyl)-6-pyridyl]ethane
the binding mode of the diaqua moiety. The size of pendant
(H2BPG2E) (see Scheme 1(B)). In the H2BPG2E ligand, two tet-
groups may decisively affect the stability of peroxo complexes of
radentate N,N-bis(2-pyridylmethyl)glycine (BPG) moieties are
BPG2E and 6-hpa to control the O–O bond scission.
connected by a structurally exible –CH2CH2– tether at the
2b catalyzed the epoxidation of alkenes with H2O2 in the
6-positions of pyridyl groups. The –CH2CH2– tether is common
presence of Et3N (2 eq.) in MeCN, where trans-b-methylstyrene
in 6-hpa and H2BPG2E, but two pyridyl groups are substituted
with two carboxyl groups in H2BPG2E. H2BPG2E forms m-oxo-
diaquadi-iron(III) complexes [Fe2(m-O)(H2O)2(BPG2E)]X2 [X ¼
ClO4 (2a) or OTf (2b)] that efficiently catalyzes epoxidation of
various alkenes with H2O2. Upon reaction of 2b with H2O2, a
new purple species was formed and isolated at low tempera-
tures. The purple species was characterized spectroscopically.
Here, we report the synthesis, crystal structures, and H2O2
activation of di-iron complexes with BPG2E. The roles of the
carboxylate donor in the O–O bond scission will be discussed in
Fig. 1 ORTEP view (70% probability) of the cationic portion of 2a.
Selected bond distances [A˚] and angle [ꢀ]: Fe1/Fe2 3.568, Fe1–O1
1.7869(13), Fe1–O2 2.0217(12), Fe1–O4 2.0657(14), Fe1–N1 2.2632(16),
Fe1–N2 2.1471(13), Fe1–N3 2.1833(12), Fe2–O1 1.7932(13), Fe2–O5
2.0105(10), Fe2–O7 2.0619(12), Fe2–N4 2.2656(16), Fe2–N5
Scheme 1 Chemical structures of (A) 6-hpa and (B) H2BPG2E.
2.1291(15), Fe2–N6 2.1837(14); Fe1–O1–Fe2 170.43(6).
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 2282–2292 | 2283

View Article Online
Chemical Science
Edge Article
was converted to the epoxide in 99% yield based on H2O2 used.
The turnover number of 2b exceeded 120. 2b showed catalase
activity at temperatures higher than 0 ꢀC, and both the yield of
epoxide and the turnover number were reduced upon addition
of H2O. Therefore, H2O2 was slowly added using a syringe pump
to keep the concentration of free H2O2 low in the catalytic
reaction. The detailed reaction conditions are described in the
Experimental section. It is noted that 1,2-cis-diol, oen
produced in the alkene oxidation catalyzed by monoiron
complexes with H2O2, was not detected in the reaction of 2b.
When cis-b-methylstyrene and cis-cyclooctene were used as
substrates, the yields of the epoxide were 43 and 15% based on
H2O2 used, respectively, and for the former substrate, cis- and
trans-epoxides were produced in 6 and 37% yield, respectively.
Moreover, when the reaction using trans- and cis-b-methylstyr-
ene as a substrate was carried out under O2-atmosphere benz-
aldehyde was produced in the yield of 98 and 40% based on the
H2O2 used, respectively (see Table S11†). These results sug-
gested that a cation-radical intermediate was formed via one-
electron oxidation of alkene. The cation radical may undergo a
conguration change to form trans-epoxide from cis-alkene in
the subsequent reactions, while under O2 atmosphere it may
afford benzaldehyde as a multi-electron oxidation product. On
the basis of these results it is expected that an active species
capable of oxidizing the alkene is formed upon reaction of 2b
with H2O2. Therefore, 2b can be a promising functional sMMO
model to study the roles of the carboxylate donor in the dioxy-
gen activation and substrate oxidation. However, it took ten-fold
longer time to add H2O2 in the epoxidation catalyzed by 2b than
by 1. This is due to the catalase activity and very slow H2O2
activation by 2b.
Fig. 2 Electronic absorption spectral change in the formation of the
intermediate 3 upon addition of 3 equiv. of Et3N and 100 equiv. of
H2O2 to 2b (0.5 mM) in MeCN–H2O (10 : 1, v/v) at ꢁ10 ꢀC under N2.
Each spectrum is recorded at 5 second interval.
To detect the intermediates in the activation of H2O2, we
carried out spectral pursuit for the reaction of 2b with H2O2.
Upon addition of an excess amount of H2O2 to 2b in the pres-
ence of Et3N (3 eq.) in MeCN–H2O (10 : 1, v/v) at ꢁ10 ꢀC, a dark
purple species 3 was generated, and showed absorption bands
at 452 nm (3 ¼ 1420 Mꢁ1 cmꢁ1), 546 (1300), and 700 (300) (see
Fig. 2). These data are similar to 462 nm (3 ¼ 1100 Mꢁ1 cmꢁ1),
577 (1500), and 750 (200) of [Fe2(m-O)(m-O2)(6-Me2-BPP)2],11
suggesting that 3 has a similar m-oxo-m-peroxodi-iron(III) core.
Cold-spray ionization (CSI) MS spectrum of 3 generated as
shown above in MeCN–H2O (10 : 1, v/v) shows a strong major
peak at m/z 699, corresponding to {H[Fe2O3(BPG2E)]}+ as shown
in Fig. 3 upper. When 3 was 18O-labeled with H218O2, the major
peak was shied to m/z 701 and 703 (see Fig. 3 upper). The shi
by four mass units demonstrated that two 18O-atoms were
incorporated into 3 from H218O2, suggesting the existence of a
peroxo moiety in 3. This is consistent with the presence of a m-
oxo-m-peroxodi-iron(III) core in 3 suggested by the electronic
spectrum. Meanwhile, the two mass unit shis may be
reasonably explained if the peroxo moiety of 3 undergoes O–O
Fig. 3 Upper: cold-spray ionization (CSI) MS spectra of 3 in MeCN–
H2O (10 : 1, v/v) showed a major peak at m/z 699 corresponding to {H
[Fe2O3(BPG2E)]}+. The isotope patterns of the major peak observed
bond scission under the CSI MS conditions to generate m-oxo- with (a) H216O2 and (b) H218O2 are shown in the insets. The red lines are
theoretical isotope patterns. Lower: CSI MS spectra of 18O3–3
dioxodi-iron(IV) 4, and one of the terminal O-atoms of the O]
obtained upon reaction of 18O–2b with H218O2 in MeCN–H218O
Fe(IV)–O–Fe(IV)]O moiety is exchanged with H2O. Moreover,
(10 : 1, v/v) showed a major peak at m/z 727 corresponding to {Na
18O-labeled 2b was prepared from 2b in MeCN–H218O (10 : 1, v/
[Fe2O3(BPG2E)]}+. The isotope pattern of the major peak with theo-
v) as shown in the Experimental section, and 86% of 18O-
retical red lines calculated for a 18 : 82 mixture of 18O2-labeled 3 and
incorporation into the m-oxo O-atom is estimated from the 18O3-labeled 3 is shown.
2284 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
intensity ratio of the isotope peaks as shown in Fig. S7.† Fully
labeled 18O3–3 was prepared upon addition of 50 eq. of H218O2
in MeCN to a solution of 18O-labeled 2b in the presence of 3 eq.
18
of Et3N in MeCN–H2 O (10 : 1, v/v) at ꢁ10 C. The H218O2 in
ꢀ
MeCN was contaminated by a small amount of H216O, which
was less than 10% of H218O used. The CSI MS spectrum of fully
labeled 18O3–3 shows major peak corresponding to {Na
[Fe2O3(BPG2E)]}+ with the isotope peaks at m/z 721, 723, 725,
and 727 in the intensity ratio of 1 : 7 : 21 : 71, respectively (see
Fig. 3 lower). The fact that the major isotope peak appeared at
727, shied by six mass units from 721, clearly showed that
three O-atoms of m-oxo and m-peroxo in 3 were 18O-labeled. The
ratio between 18O2–3 and 18O3–3 estimated from the intensity of
isotope peaks is 18 : 82, close to the 14 : 86, ratio between 2b
and 18O-labeled 2b. The slight difference may be due to the
exchange of the terminal O-atoms of 18O]Fe(IV)–18O–
Fe(IV)]18O of 4 with H216O. However, such O-atom exchange
was not observed in the corresponding 6-hpa complexes.13b
Thus, the carboxylate donor of BPG2E strongly affects the
reactivity of the Fe(IV)]O. The anionic character of the
carboxylate donor may decrease the Lewis acidity of the Fe(IV) to
make the release of terminal O-atoms easier. This may promote
the exchange of the terminal O-atoms with H2O. As a similar
effect, it has been reported that the carboxylate donor weakens
the Fe–Operoxo bond in [Fe2(m-O)(m-O2)(6-Me2-BPP)2].11,18
Upon addition of an excess amount of cold Et2O to a solution
of 3 generated as shown above, 3 was isolated as solid, and the
isolated yield was 60% based on 2b used. The elemental anal-
ysis of 3 agreed well with a formula [Fe2O3(BPG2E)]$8H2O. On
the basis of the elemental analysis and spectral data, 3 is
consistent with its assignment as m-oxo-m-peroxodi-iron(III)
[Fe2(m-O)(m-O2)(BPG2E)], and based on alkene epoxidation and
O-atom exchange upon 18O-labelling in the CSI MS measure-
ments, it is suggested that 3 is converted to m-oxodioxodi-
iron(IV) [Fe2(m-O)(O)2(BPG2E)] (4) via O–O bond scission. As
Fig. 4 Zero field Mossbauer spectra of the isolated solid, recorded by
¨
raising the temperatures at (A) 23, (B) 100, (C) 200, (D) 250, and (E) 293
K, and by recooling the same sample at (F) 17 K. The black line is the
least-square fitting to the raw data, and the red (component a), blue
(component b), and green (component c) lines are the deconvolution
spectra corresponding to m-oxo-m-peroxodi-iron(III) 3, m-oxodioxodi-
iron(IV) 4, and di-iron(II) complexes.
293 K the component a (red spectrum) was decreased with the
increase of b (blue spectrum) as shown in Fig. 4(B)–(E). The
ratios of a : b estimated from the deconvolution spectra are
85 : 15, 76 : 24, 59 : 41, and 45 : 55 at 100, 200, 250, and 293 K,
respectively. The d and DEQ values of a at 100 K, 0.392(4) and
1.650(2) mm sꢁ1, are close to 0.50 and 1.46 mm sꢁ1 of m-oxo-m-
peroxodi-iron(III) [Fe2(m-O)(m-O2)(6-Me2-BPP)2] at 80 K,11 and to
0.53–0.68 and 1.51–1.90 mm sꢁ1 of peroxodi-iron(III) interme-
diate P in sMMO (ref. 2) and their model compounds.19 Thus,
component a may be assigned to m-oxo-m-peroxodi-iron(III) 3
with a symmetric structure. These values, however, are largely
¨
shown below, the conversion of 3 to 4 is observed by Mossbauer
and IR spectral measurements using the isolated solid. The
presumed transformation of 2 to 3 and 4 is shown in Scheme 2.
¨
The zero-eld Mossbauer spectra of the isolated dark purple
species were recorded at some different temperatures raised
from 23 to 295 K, then recorded again using the same sample by
recooling to 17 K as shown in Fig. 4. The spectrum at 23 K
mainly consists of two quadrupole doublets with d ¼ 0.481(2)
mm sꢁ1, DEQ ¼ 1.657(2) mm sꢁ1 and d ¼ 0.20(3), DEQ ¼ 0.40(6),
as shown by the deconvolution spectra composed of compo-
nents a (red spectrum) and b (blue spectrum) in Fig. 4(A),
respectively. The intensity ratio of a and b is almost 90 : 10 at 23
K. When the temperature was increased to 100, 200, 250, and
¨
different from Mossbauer parameters d and DEQ ¼ 0.54 and
0.58 mm sꢁ1 for a peroxodi-iron(III) form of ToMOH,7 and 0.66
and 0.58 mm sꢁ1 for a peroxo complex generated from [Fe2(m-
O)(H2O)2(BPG2DEV)](ClO4)2.12 The d and DEQ values of the
component b at 24–293 K, 0.131(3)–0.26(2) and 0.40(6)–0.55(3)
mm sꢁ1, are almost in the range of d ¼ 0.14–0.21 and
DEQ ¼ 0.53–0.68 mm sꢁ1 of high-spin oxodi-iron(IV) in the
intermediate Q of sMMO.2,20 High-spin (S ¼ 2) oxoiron(IV)
complexes reported so far show the d and DEQ values around 0.1
and 0.5 mm sꢁ1, respectively,15 but low-spin (S ¼ 1) oxoiron(IV)
complexes have relatively higher DEQ values of 1–2 mm sꢁ1
.
21
Thus, component b may be assigned to high-spin oxodi-iron(IV)
similar to the intermediate Q and the synthetic high-spin
Scheme 2 Transformation of 2 to m-oxo-m-peroxodi-iron(III) 3 and
m-oxodioxodi-iron(IV) 4.
¨
oxoiron(IV) complexes. The Mossbauer spectra of starting
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 2282–2292 | 2285

View Article Online
Chemical Science
Edge Article
material 2b and decomposed product of 3 at 77 K show quad- (b) 18O-labeled 2b/H216O2, (c) 2b/H218O2 and (d) 18O-labeled 2b/
rupole doublets with d ¼ 0.461(2) mm sꢁ1, DEQ ¼ 1.670(2) mm H218O2 are shown in Fig. 5(a)–(d), respectively. The spectra of (a)
sꢁ1 and d ¼ 0.453(2) mm sꢁ1, DEQ ¼ 1.664(2) mm sꢁ1, respec- and (b) show a clear band at 835 cmꢁ1 (Fig. 5(a) and (b)). This
tively (Fig. S8†). Since these data are temperature-independent band is in the range of 822–919 cmꢁ1 reported as O–O
and clearly different from d ¼ 0.24(2) mm sꢁ1, DEQ ¼ 0.48(4) stretching vibration of peroxodi-iron(III) complexes.18,22 The
mm sꢁ1 of the component b at 50 K, the components a and b band shied to 784 cmꢁ1 by 18O-labelling with H218O2 as shown
can be distinguished from the starting material and decom- in the spectra of (c) and (d) (Fig. 5(c) and (d)). The observed
posed product. In conjunction with the elemental analysis and isotope shi by 51 cmꢁ1 is close to 48 cmꢁ1 expected based on a
the electronic absorption and CSI MS spectra, conclusively, the harmonic O–O oscillator model, but much larger than 37 cmꢁ1
components a and b are assigned to m-oxo-m-peroxodi-iron(III) 3 expected for a stretching vibration of the terminal Fe(IV)]O.
and high-spin m-oxodioxodi-iron(IV) 4, shown in Scheme 2, Thus, the band at 835 cmꢁ1 is assigned to the O–O stretching
respectively.
vibration of the m-oxo-m-peroxodi-iron(III) 3, and the O–O moiety
¨
Moreover, when the Mossbauer spectrum was recorded upon comes from hydrogen peroxide but not from the coordinated
recooling the same sample to 17 K the increase of component a H2O molecules. It was reported that the stretching vibration of a
and the concomitant decrease of component b were observed, terminal Fe(IV)]O in Fe(III)–O–Fe(IV)]O is shied by 18O-
where the ratio of a and b is 82 : 18 with a small amount, 7%, of labeling of the m-oxo bridge as shown by the bands at 840, 835,
a quadrupole doublet assignable to di-iron(II) as shown in 797, and 794 cmꢁ1 for Fe(III)–16O–Fe(IV)]16O, Fe(III)–18O–
Fig. 4(F). The spectrum thus obtained is completely equal to Fe(IV)]16O, Fe(III)–16O–Fe(IV)]18O, and Fe(III)–18O–Fe(IV)]18O,
that obtained by the rst measurement at 23 K. Thus, the respectively.23 If the band at 835 cmꢁ1 comes from the Fe(IV)]O
temperature-dependent Mossbauer measurements clearly it must be affected by the 18O-m-oxo labeling. As shown in Fig. 5,
¨
showed that 3 and 4 are interconverted each other. To the best however, it was not changed at all by the 18O-m-oxo labeling.
of our knowledge, this is the rst example of reversible Therefore, the band at 835 cmꢁ1 nally is assigned to the O–O
conversion of peroxodi-iron(III) to high-spin oxodi-iron(IV) with a stretching vibration of 3, and apparently, the Fe(IV)]O of 4 is
carboxylate-containing ligand. This reversibility is attained not detected in the RR spectra. Unfortunately, the RR band of
because of the bistability of 3 and 4. The exible –CH2CH2– the stretching vibration of Fe–Operoxo bond in the peroxodi-iron
tether in the BPG2E ligand may enable the bistability of perox- moiety of 3 was not obtained with 407 and 607 nm excitations.
odi-iron(III) and high spin oxodi-iron(IV) states as it was observed
in the reversible O–O bond scission of the corresponding di-iron in 4 at 810 cmꢁ1, using the isolated solid on the IR spectra at
complexes with a 6-hpa ligand.13b
room temperature. The band was decreased with the decom-
Finally, we observed a stretching vibration band of Fe(IV)]O
¨
On the basis of the Mossbauer data in this work and those position of 4. This may take place upon reaction with H2O2
reported for the di-iron complexes with 6-hpa, the ratios of the included as an impurity in the isolated solid due to the high
peroxodi-iron(III) to the oxodi-iron(IV) at 295 K are 45 : 55 and catalase activity of 4, and the rate seems to depend on the
15 : 85 for the BPG2E and the 6-hpa complexes,13b respectively. amount of H2O2 included. Aer thorough evacuation of H2O2
These data show that BPG2E relatively stabilizes the peroxo state
rather than the oxodi-iron(IV) state as compared with 6-hpa.
This may be explained by the structural effect of the pendant
carboxylate and pyridyl groups shown in the crystal structures of
2a and 1, where BPG2E stabilizes a syn-binding mode, and 6-hpa
does an anti-binding mode. Since it is rational that the peroxo
moiety of 3 takes a syn-like binding mode similar to that of
[Fe2(m-O)(m-O2)(6-Me2-BPP)2],11 BPG2E may stabilize the syn-type
structure of the Fe–O–O–Fe moiety in 3. Conversely, it turns out
that 6-hpa destabilizes the syn-like binding mode of Fe–O–O–Fe
moiety to promote the O–O bond scission. This suggests that
the O–O bond scission of peroxodi-iron(III) to oxodi-iron(IV) can
be controlled by the size of the pendant groups.
The peroxo moiety of the m-oxo-m-peroxodi-iron(III) 3 was
detected by resonance Raman (RR) measurements. To examine
the origin of the peroxo O–O atoms in 3 four different experi-
ments were carried out where 2b and 18O-labeled 2b were used
for the reaction with H216O2 and H218O2 in aqueous MeCN at
ꢁ30 ꢀC. The 18O-labeled 2b was prepared by treating 2b with an Fig. 5 Resonance Raman spectra of 3 in MeCN at ꢁ30 ꢀC obtained
excess amount of H218O in MeCN under N2 at 70 ꢀC. The
with excitation at 607 nm. The samples were prepared by treating 2b
or 18O-labeled 2b with 500 equiv. of H216O2 or H218O2 ((a) 2b/H216O2,
incorporation of one 18O-atom was monitored by the ESI MS
(b) 18O-labeled 2b/H216O2, (c) 2b/H218O2, and (d) 18O-labeled 2b/
spectra, and 86% of 2b was converted to 18O-labeled 2b in 30
min (see Fig. S7†). The RR spectra obtained with 607 nm exci-
H218O2) in the presence of 2 equiv. of Et3N. The numbers shown in the
spectra are the corresponding vibration bands of 3. S and * means the
tation for four different samples prepared with (a) 2b/H216O2, solvent band and H2O2, respectively.
2286 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
from the isolated solid under highly reduced pressure less than with a 6-hpa ligand. This is a clear electronic effect of the
10ꢁ5 Torr for more than three weeks at ꢁ40 ꢀC, the isolated carboxylate donor of BPG2E that relatively decreases the Lewis
solid gave the clear band at 810 cmꢁ1, which was kept intact for acidity of Fe(IV) as compared with the pyridine donor of 6-hpa.
several minutes and gradually decreased over an hour as fol- This is consistent with the easy O-atom exchange of Fe(IV)]O in
lowed by the IR spectral change (Fig. 6(A)/(B)). When 4 was 4 described above in the CSI MS measurements.
labelled with H218O2, the band at 810 cmꢁ1 disappeared and a
Decomposition of 3, prepared with 2b (5.0 ꢂ 10ꢁ4 M) and
band appeared at 774 cmꢁ1 (Fig. 6(C)), which decreased with H2O2 (1.2 eq.) in the presence of Et3N in MeCN–H2O (10 : 1, v/v),
time. The observed isotope shi 36 cmꢁ1 is equal to the theo- was monitored at 546 nm, and obeyed rst-order kinetics with
retical value 36 cmꢁ1, estimated based on a harmonic Fe–O k ¼ 1.5 ꢂ 10ꢁ3 sꢁ1 at ꢁ10 ꢀC as shown in Fig. 7. Upon addition
oscillator model.13b Thus, conversion of 3 to 4 via O–O bond of trans-b-methylstyrene as a substrate in the concentration
scission is shown by the IR as well as the Mossbauer measure- range of 2.5 ꢂ 10ꢁ2 to 1.0 ꢂ 10ꢁ1 M, the decay rate was not
¨
ments. The value 810 cmꢁ1 is in a range of 798–842 cmꢁ1 accelerated at all as shown in Fig. 8. These results demonstrated
reported for Fe(IV)]O, and slightly weaker than 820 cmꢁ1 that 3 is not the direct oxidant, and accordingly it is likely that
observed by the RR spectra of m-oxodioxodi-iron(IV) complex the O–O bond scission of 3 to 4 is the rate-determining step.
Therefore, it is difficult to accumulate 4 in a solution. This is
consistent with the fact that the Fe(IV)]O of 4 could not be
detected in a solution by the RR spectra. We carried out
Arrhenius plot for the thermal decomposition of 3 to estimate
the activation energy DEa (see Experimental section for the
details of the kinetic studies and the plot in Fig. S9†). The
observed DEa value is 22.8 kcal molꢁ1, close to 21.9 kcal molꢁ1
estimated by DFT calculation vide infra (Fig. 9).
The O–O bond scission of 3 to 4 is the rate-determining step,
but was detected by the isotope-labeling experiments in the CSI
MS measurements as an exchange of the O-atom of Fe(IV)]O
with H2O as shown above. The mass spectrometric conditions
are unusual in that high electric potential is charged to ionize
sample. Under such conditions, the O–O bond scission may be
accelerated to give 4, which may have a lifetime necessary to be
detected. The reason may be because 4 is isolated from H2O2 in
gaseous atmosphere under the CSI MS conditions. When using
the isolated solid thoroughly dried, the reversible conversion of
¨
3 to 4 was shown by the temperature-dependent Mossbauer
spectra, and the stretching vibration band of Fe(IV)]O in 4 was
detected in the IR spectra. These results commonly indicate that
4 can be detected spectroscopically when it is free from H2O2.
To rationalize the mechanism of the reversible conversion of
3 to 4 as proposed in Scheme 2 as well as to corroborate the
¨
temperature dependence of the Mossbauer spectra, we carried
Fig. 6 IR spectra of the isolated solid recorded at room temperature; Fig. 7 Electronic absorption spectral change upon decomposition of
(A) red: immediately after sample charge, black: 60 min later, (B) 3 (5.0 ꢂ 10ꢁ4 M) in MeCN–H2O (10 : 1, v/v) at ꢁ10 ꢀC (insert: time trace
difference spectrum of (A), (C) red: 18O-labelled, black: non-labelled.
of absorbance at 546 nm).
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 2282–2292 | 2287

View Article Online
Chemical Science
Edge Article
˚
distance is elongated by 0.406 A. As is evident from the
¨
Mossbauer spectra, the O–O bond cleavage of 3 occurs at rela-
tively high temperature because of the larger activation barrier
needed for this process and the maximum population of 4 is
observed at the highest temperature 293 K. The recooling of the
sample 4 from 295 K leads to the rebinding of the oxygen atoms
and hence complex 3 is thermodynamically and kinetically
obtained. Thus, the DFT studies nicely support the experi-
mental ndings of the reversible O–O bond scission.
Although in a solution, 4 can not be detected at all because
O–O bond scission of 3 is the rate-determining step, O–O bond
scission smoothly proceeds for the peroxo complex with 6-hpa
ligand as previously reported.13b Moreover, it took ten-fold
longer time for addition of H2O2 by syringe-pump to gain a good
yield for the epoxidation catalyzed by 2b than that by 1.13a This
may be caused by both the slower O–O bond scission of 3 and
the higher catalase activity of 4. The BPG2E ligand may stabilize
the syn-binding mode of 3 and 4 to retard the O–O bond scis-
sion of 3 and to enhance the catalase activity of 4 leading to the
non-productive decomposition. As described for the crystal
Fig. 8 Plot of kobs vs. [S], the substrate concentration, in the reaction
of 3 (0.5 mM) with trans-b-methylstyrene (with 0 eq., 50 eq., 200 eq.,
400 eq.) in MeCN–H2O (10 : 1, v/v) at ꢁ10 ꢀC.
out DFT calculations. In the optimized di-iron structures of 3 structures of 2a and 2b, the syn-binding mode is caused by the
and 4 in Fig. 9, the Fe centers are in octahedral environments in sterically less hindered pendant carboxylate group of BPG2E.
˚
which the Fe–Fe distances are 3.093 and 3.466 A, respectively. Therefore, it is suggested that the O–O bond scission of the
Calculated spin densities at the two Fe centers are (4.11 and peroxo species and the catalase activity of m-oxodioxodi-iron(IV)
ꢁ4.14), and (3.25 and ꢁ3.20) respectively for complexes 3 and 4, species are strongly affected by the size of the pendant groups to
due to the antiferromagnetic spin coupling at the Fe centers. In determine the overall rate of the catalytic reaction.
complex 3 the O2 displays a bridging bond in a m-h1:h1-O2
The importance of the size of the carboxylate donor has been
˚
fashion with a O–O distance of 1.404 A whereas in 4 the O–O pointed out for the rst time in comparison between BPG2E and
˚
distance is 3.751 A, indicating that the O–O bonding interaction 6-hpa in this study. In the biological relevance, it is suggested
is absent in the m-oxodioxodi-iron(IV) complex. The computed that the carboxylate donor may play an important role in the
energy diagram for the O–O bond scission of 3 to 4 at 298 K is stability of peroxo intermediates in the catalytic cycles of
depicted in Fig. 9. The reaction via transition state TS3–4 is sMMO, TMO, ToMO and DOHH. Since a carboxylate donor is
slightly endothermic by 0.8 kcal molꢁ1 with an activation sterically less hindered, there is no steric repulsion between the
barrier of 21.9 kcal molꢁ1. In this transition state the O–O donor groups located at the syn-position, and the Fe–O–O–Fe of
Fig. 9 Calculated energy diagram for the O–O bond scission of 3 to 4. Selected bond distances in A˚ are given along with the relative energies in
kcal molꢁ1
.
2288 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
peroxo intermediate can easily take the most stable structure. of H2O2 in MeCN was prepared by concentration of 40%
This may lead to the high thermal stability of the peroxo aqueous H2O2 and addition of MeCN, and determined by redox
intermediate as in DOHH. The stability of the peroxo interme- titration with KMnO4. Et3N was puried by distillation from
diate can be controlled by the conformation change, where it sodium, and kept on NaOH. A solution of Et3N in MeCN was
may bring about the change of the syn-binding mode to the determined by titration.
anti-binding mode to facilitate the O–O bond scission. This
aspect may give a new insight into the roles of the carboxylate
donors in the dioxygen activation of the non-heme di-iron
enzymes.
Measurements
Elemental analyses (C, H, and N) were carried out on a Per-
kin-Elmer Elemental Analyzer 2400 II. UV-vis absorption
Conclusions
spectrum was recorded on an Agilent 8543 UV spectroscopy
A carboxylate-containing dinucleating ligand H2BPG2E and its system with a Unisoku thermostated cell holder designed for
m-oxodiaquadi-iron(III) complexes [Fe2(m-O)(H2O)2(BPG2E)]X2 low-temperature measurements. Cold spray ionization (CSI)
(X ¼ ClO4 (2a) and TfO (2b)) were synthesized to mimic di-iron mass spectra were obtained on a JEOL JMS-T100CS spec-
centers surrounded by a common carboxylate-rich coordination trometer in MeCN solution at low temperatures. GLC analysis
environment of the O2-activating non-heme di-iron enzymes. 2b was performed on a Shimadzu GC-2014 gas chromatography
efficiently catalyzed the epoxidation of various alkenes with equipped with a GL Science InertCap capillary column (60 m
H2O2. A dark purple species was formed as an intermediate ꢂ 0.25 mm). Infrared (IR) spectra were recorded on a SHI-
upon reaction of 2b with H2O2, and isolated upon addition of an MADZU Single Reection HATR IRAffinity-1 MIRacle 10.
excess amount of cold Et2O. The purple species was assigned to 1H-NMR spectra were recorded on a JEOL ECA-500RX spec-
the m-oxo-m-peroxodi-iron(III) [Fe2III(m-O)(m-O2)(BPG2E)] (3) on trometer using Me4Si as an internal standard. Mossbauer
¨
the basis of the electronic absorption, CSI MS, and RR spectra. spectra were measured at Nishina Center, The Institute for
¨
¨
Moreover, on the basis of Mossbauer and IR spectra obtained by Physical and Chemical Research (RIKEN). The Mosbauer
using the isolated solid, it was shown that 3 was converted to spectra of 2b and the decomposed product of the peroxo
the high-spin m-oxodioxodi-iron(IV) [Fe2IV(m-O)(O)2(BPG2E)] (4). complex were measured at Kyoto University Research Reactor
The kinetic studies and the substrate oxidation showed that the Institute. The radioactive source was 57Co(Rh). Isomer shis
active species is generated via the rate-determining O–O bond were reported relative metallic iron foil. Resonance Raman
scission. This is the rst example for the reversible O–O bond (RR) scattering was excited at 407 and 607 nm using an
scission of peroxodi-iron(III) to high spin oxodi-iron(IV) with a Ar-dye-laser and detected with CCD detector (Princeton
carboxylate-containing ligand. Comparisons of the structures Instruments) attached to
a single polychrometer. All
ꢀ
and reactivity of peroxo complexes with BPG2E and 6-hpa clar- measurements were at ꢁ30 C using a spinning cell. Raman
ied roles of the carboxylate donors in the O–O bond scission. shis were calibrated with indene, and the accuracy of the
With 6-hpa, the O–O bond scission of peroxodi-iron(III) to high peak positions of the Raman bands was ꢃ1 cmꢁ1
.
spin oxodi-iron(IV) smoothly proceeded. On the contrary, BPG2E
Structure determination of single crystals. The crystal
stabilizes the peroxodi-iron(III) to retard the O–O bond scission. structures of 2a and 2b were determined using a Rigaku R-AXIS
This remarkable effect was reasonably explained by the size RAPID diffractometer using multi-layer mirror monochromated
ꢀ
difference in the carboxylate and pyridine donors of BPG2E and Mo-Ka radiation. The data were collected at ꢁ180 ꢃ 1 C to a
6-hpa, respectively. Although the electronic effect of carboxylate maximum 2q value of 54.9ꢀ. A total of 44 oscillation images were
donor has not been claried for the O–O bond scission yet, it collected. A sweep of data was done using u scans from 130.0 to
was shown that carboxylate donor of BPG2E weakens the 190.0ꢀ in 5.00ꢀ step, at c ¼ 45.0ꢀ and f ¼ 0.0ꢀ. The exposure rate
Fe(IV)]O bond on the basis of the 10 cmꢁ1 low frequency shi was 60.0 [seconds per degree]. A second sweep was performed
of stretching vibration of Fe(IV)]O of 4 and easy exchange of the using u scans from 0.0 to 160.0ꢀ in 5.00ꢀ step, at c ¼ 45.0ꢀ and
O-atom as compared with the pyridine donor of 6-hpa. This may f ¼ 180.0ꢀ. The exposure rate was 60.0 [seconds per degree]. The
suggest that in the non-heme di-iron enzymes, similarly, the crystal-to-detector distance was 127.40 mm for 2a and 127.00
carboxylate donors may sterically stabilize the peroxodi-iron(III) mm for 2b. Readout was performed in the 0.100 mm pixel
state, and therefore, the generation of high spin oxodi-iron(IV) mode. Of the 19 147 and 20 161 reections were collected for 2a
via the O–O bond scission may require a specic conforma- and 2b, where 8757 and 9218 were unique (Rint ¼ 0.0180 and
tional change at the di-iron center.
0.0303); equivalent reections were merged, respectively. An
empirical absorption correction was applied which resulted in
transmission factors ranging from 0.736 to 0.916 for 2a and
from 0.682 to 0.894 for 2b. The data were corrected for Lorentz
and polarization effects.
Experimental
Materials
ꢀ
All ordinary reagents and solvents were purchased and used as
Crystal data for 2a$3H2O. C30H40Cl2Fe2N6O18, triclinic, P1,
˚
received unless otherwise noted. MeCN was dried over P2O5 and Z ¼ 2, a ¼ 11.4730(4), b ¼ 12.9171(4), c ¼ 14.4931(4) A,
distilled. Alkenes used in this study were puried by distillation a ¼ 113.070(8)ꢀ, b ¼ 100.094(7)ꢀ, g ¼ 90.982(6)ꢀ, V ¼ 1936.75(17)
3
and treatment with alumina column just before use. A solution A , m(MoKa) ¼ 9.705 cmꢁ1, Dc ¼ 1.638 g cmꢁ3, R1 ¼ 0.0281,
˚
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 2282–2292 | 2289

View Article Online
Chemical Science
Edge Article
wR2 ¼ 0.0757, for 8757 unique reections, 599 variables, and dried in vacuo. The solid residue was dissolved in dry EtOH,
and the insoluble solid was removed by ltration. The ltrate
GOF ¼ 1.060.
ꢀ
Crystal data for 2b$2H2O. C32H38F6Fe2N6O15S2, triclinic, P1, was evaporated, To the residue were added water and
˚
Z ¼ 2, a ¼ 11.5827(3), b ¼ 13.4954(3), c ¼ 14.5372(3) A, a ¼ Fe(OTf)2$2MeCN 113 mg (0.26 mmol). The mixture was stirred
3
ꢀ
ꢀ
ꢀ
˚
114.729(8) , b ¼ 99.954(7) , g ¼ 90.485(6) , V ¼ 2024.60(17) A , for 24 h under O2. The pink-colored 2b was precipitated and
m(MoKa) ¼ 9.227 cmꢁ1, Dc ¼ 1.700 g cmꢁ3, R1 ¼ 0.0257, wR2 ¼ recrystallized from MeCN–H2O to give crystals suitable for X-ray
0.0814, for 9218 unique reections, 592 variables, GOF ¼ 1.098. structure analysis (78 mg, 60%). Elemental analysis of 2b$2H2O
(%) calcd for C32H34F6Fe2N6O15S2: C 37.23, H 3.32, N 8.14;
found: C 37.22, H 3.65, N 8.16. IR: n ¼ 1609, 1568 (pyridine ring),
631 cmꢁ1 (Fe–O–Fe). ESI MS: m/z 815 [M ꢁ 2H2O ꢁ TfO]+.
Synthesis
1,2-Bis[2-(2-pyridylmethyl)aminomethyl-6-pyridyl]ethane-
[Fe2(m-18O)(H218O)2(BPG2E)](TfO)2 (18O-labeled 2b). [Fe2(m-
N,N0-diacetic acid diethyl ester. A mixture of 1,2-bis(6-bromo- O)(H2O)2(BPG2E)](TfO)2 (2b) 30 mg (30 mmol) was dissolved in
methyl-2-pyridyl)ethane dihydrobromide 1.06 g (2.0 mmol), dry MeCN under N2. 18O-labeled water (H218O) 200 mL
2-pyridylmethyl glycine ethyl ester 0.777 g (4.0 mmol) and (11 mmol) was added to the solution, and the solution was
Na2CO3 2.12 g (20 mmol) in MeCN (15 mL) was stirred at room stirred at 70 ꢀC. The reaction was monitored by the ESI MS
temperature for 24 h. Then the Glay precipitation was removed spectra, and 86% of m-oxo O-atom was labeled with 18O-atom in
from the mixture. Water was added to the resulting solution and 30 min (see Fig. S7†). The resulting solution was evaporated,
the solution was extracted with CHCl3. The extracts were dried and the pink-colored 18O-labeled 2b was obtained. ESI MS: m/z
over Na2SO4 and evaporated. The crude product was isolated as 817 [M ꢁ 2H218O ꢁ TfO]+.
red-brown oil and puried by alumina column chromatography
Preparation and isolation of 3. To a solution of 2b (83.7 mg
(solvent: CHCl3). The product was isolated as yellow oil (1.07 g, 83.7 mmol) in MeCN (1 mL) was added 2.0 equiv. of Et3N
1
ꢀ
90%). H-NMR (Me4Si, in CDCl3): d¼ 8.52 (d, 1H, py0-3), 7.65 (120 mM) at ꢁ40 C. To the solution was added 500 equiv. of
(t, 1H, py0-4), 7.60 (d, 1H, py-5), 7.50 (t, 1H, py-4), 7.37 (d, 1H, py0- H2O2 (10 M), and stirred for 5 min. The solution turned dark
6), 7.14 (t, H, py0-5), 6.95 (d, 1H, py-3), 4.17(q, 2H, –CH2CH3), purple. To the solution was added 15 mL of Et2O at ꢁ40 ꢀC, then
3.99 (s, 2H, –CH2N–), 3.96(s, 2H, –CH2N–), 3.45 (s, 2H, purple solid precipitated. The supernatant was decanted off,
–CH2COO–), 3.17(s, 2H, –CH2CH2–), 1.25 ppm (t, 3H, –CH2CH3). and the precipitate was washed with Et2O several times at
ESI MS: m/z 619 [M + Na]+.
ꢁ40 ꢀC. The purple solid was dried in vacuo. The isolated solid is
2-Bis[2-(2-pyridylmethyl)aminomethyl-6-pyridyl]ethane- stable at low temperature, and not changed several days at room
N,N0-diacetic acid$nHCl (H2BPG2E$nHCl). To a mixture of THF temperature under dark in the absence of organic compounds
(10 mL) and conc. HCl (1.5 mL) was added 1,2-bis[2-(2-pyr- potentially acting as reductant. Yield 35.1 mg (60%). Anal. calcd
idylmethyl)aminomethyl-6-pyridyl]ethane-N,N0-diacetic
acid for C30H46N6O15Fe2: C, 42.77; H, 5.50; N, 9.98%. Found: C,
diethyl ester 0.195 g (0.33 mmol). The mixture was stirred 42.63; H, 4.90; N, 9.82%. The isolated solid was used for various
¨
overnight at room temperature, and white solid precipitated. spectral measurements including the Mossbauer spectra.
The white solid was collected by ltration, and washed with
THF. The solid was excellent purity based on NMR without
Reaction conditions for the epoxidation of alkenes
1
further purication (0.2 g, 89%). H-NMR (Me4Si, in D2O): d ¼
A solution of 1 mL of 2b (1.0 ꢂ 10ꢁ3 M) in MeCN was placed in a
ask equipped with a three-way cock, and quickly degassed by
several cycles of evacuation and relling with N2 gas. To the
solution were added a solution of Et3N (2.0 equiv.), 1 mmol of
alkene, and 2 mL of nitrobenzene, then a solution of H2O2
(10 mmol) in MeCN (1.0 mL) was added over 5 h by using a
syringe pump under N2 atmosphere with stirring at 25 ꢀC. Aer
further 30 min, a fraction of the solution was analyzed and
assayed by GC. cis-Cyclooctene and cis-b-methylstyrene were
also used as substrates instead of trans-b-methylstyrene, under
the same reaction conditions. Moreover, different reaction
conditions were examined as shown in Table S11.†
8.72 (d, 1H, py0-3), 8.52 (t, H, py0-5), 8.39 (t, 1H, py0-4), 8.03 (d,
1H, py0-6), 7.95 (t, 1H, py-4), 7.89 (d, 1H, py-5), 7.81 (d, 1H, py-3),
4.47 (s, 2H, –CH2N–), 4.44 (s, 2H, –CH2N–), 3.71 (s, 2H,
–CH2COO–), 3.57 ppm (s, 2H, –CH2CH2–). ESI MS: m/z 541 [M +
H]+, 563 [M + Na]+.
[Fe2(m-O)(H2O)2(BPG2E)](ClO4)2 (2a). An aqueous solution of
NaOH was added to H2BPG2E$4HCl 89 mg (0.13 mmol) in water
to be adjusted to pH 5. The resulting solution was evaporated
and dried in vacuo. The solid residue was dissolved in dry EtOH,
and the insoluble solid was removed by ltration. The ltrate
was evaporated. To the residue were added water and
Fe(ClO4)3$7H2O 125 mg (0.26 mmol). The pink-colored 2a was
precipitated and recrystallized from MeCN–H2O to give crystals
suitable for X-ray structure analysis (47 mg, 40%). Elemental
General procedures for kinetic studies
analysis of 2a$3H2O (%) calcd for C30H36Cl2Fe2N6O18: C 37.88, For the kinetic measurements, 3 was freshly prepared just
H 3.81, N 8.83; found: C 38.27, H 4.11, N 8.94. IR: n ¼ 1609, 1570 before the kinetic measurements. A typical method used for the
(pyridine ring), 1096 (ClO4), 621 cmꢁ1 (Fe–O–Fe). ESI MS: m/z kinetic measurements was as follows. A solution of 2 mL of 2b
765 [M ꢁ 2H2O ꢁ ClO4]+.
(5.0 ꢂ 10ꢁ4 M) in MeCN–H2O (10 : 1, v/v) was placed in a quartz-
[Fe2(m-O)(H2O)2(BPG2E)](TfO)2 (2b). An aqueous solution of cell equipped with a three-way cock, and quickly degassed by
NaOH was added to H2BPG2E$4HCl 89 mg (0.13 mmol) in water several cycles of evacuation and relling with N2 gas. The
to be adjusted to pH 5. The resulting solution was evaporated temperature was maintained at ꢁ10 ꢃ 0.2 ꢀC during the
2290 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
measurements. To the solution were added a solution of Et3N
(3.0 equiv.) in MeCN and aer stirring for 2 min a solution of
H2O2 (1.2 equiv.) in MeCN. The spectral change was recorded in
the range of 400–900 nm. The decrease of 3 was monitored at
several different wavelengths. The rst-order rate constants
(kobs) were obtained from ts of ꢁln(At ꢁ AN/A0 ꢁ AN) vs. time.
The spontaneous decomposition rate of 3 was obtained in the
absence of a substrate as shown above. Here, trans-b-methyl-
styrene was used as a substrate, and the decomposition rate of 3
was measured in the presence of the substrate at the several
different concentrations in the range of 25–200 ꢂ 10ꢁ3 M. For
the measurements of the activation energy of decomposition of
3, the amount of H2O2 was reduced from 1.2 to 0.8 eq. to
6 B. J. Brazeau, R. N. Austin, C. Tarr, J. T. Groves and
J. D. Lipscomb, J. Am. Chem. Soc., 2001, 123,
11831.
´
7 (a) L. J. Murray, S. G. Naik, D. O. Ortillo, R. Garcıa-Serres,
J. K. Lee, B. H. Huynh and S. J. Lippard, J. Am. Chem. Soc.,
2007, 129, 14500; (b) A. D. Bochevarov, J. Li, W. J. Song,
R. A. Friesner and S. J. Lippard, J. Am. Chem. Soc., 2011,
133, 7384.
8 (a) I. Siewert and C. Limberg, Chem. - Eur. J., 2009, 15, 10316;
(b) S. Friedle, E. Reisner and S. J. Lippard, Chem. Soc. Rev.,
2010, 39, 2768.
9 J. R. Hagadorn, L. Que and W. B. Tolman, J. Am. Chem. Soc.,
1998, 120, 13531.
decrease the concentration of free H2O2 in the reaction mixture 10 E. C. Carson and S. J. Lippard, J. Am. Chem. Soc., 2004, 126,
because catalase activity of 3 strongly affected the rate of 3412.
decomposition at higher temperatures.ꢀ The measurements 11 X. Zhang, H. Furutachi, S. Fujinami, S. Nagatomo, Y. Maeda,
were carried out at ꢁ10, ꢁ8, ꢁ6, and ꢁ4 C.
Y. Watanabe, T. Kitagawa and M. Suzuki, J. Am. Chem. Soc.,
2005, 127, 826.
¨
12 S. Friedle, J. J. Kodanko, A. T. Morys, T. Hayashi, P. Moenne-
Quantum mechanical calculations
Loccoz and S. J. Lippard, J. Am. Chem. Soc., 2009, 131,
14508.
All the DFT calculations were carried out at the spin-unre-
stricted PBE1KCIS/6-31G* level of theory24 implemented in the
Gaussian09 (ref. 25) suite of programs. The initial geometries
were retrieved from the X-ray crystal structure. The stationary
states were conrmed by frequency analysis and the transition
state was characterized by one imaginary frequency and their
relative motion towards the reactant and product side. The zero-
point energy (ZPE) correction was also applied to the energy of
all the reported structures.
13 (a) M. Kodera, M. Itoh, K. Kano, T. Funabiki and M. Reglier,
Angew. Chem., Int. Ed., 2005, 44, 7104; (b) M. Kodera,
Y. Kawahara, Y. Hitomi, T. Nomura, T. Ogura and
Y. Kobayashi, J. Am. Chem. Soc., 2012, 134, 13236.
14 (a) J. C. Price, E. W. Barr, B. Tirupati, J. M. Bollinger and
C. Krebs, Biochemistry, 2003, 42, 7497; (b) C. Krebs,
´
D. Galonic Fujimori, C. T. Walsh and J. M. Bollinger, Acc.
Chem. Res., 2007, 40, 484.
15 (a) J. England, M. Martinho, E. R. Farquhar, J. R. Frisch,
¨
E. L. Bominaar, E. Munck and L. Que, Angew. Chem., Int.
Acknowledgements
Ed., 2009, 48, 3622; (b) J. England, Y. Guo, K. M. Van
This work was supported by a Grant-in-Aid for Scientic Research
B (no. 21350037) from the Ministry of Education, Culture, Sports,
Science and Technology, Japan and by “Creating Research Center
for Advanced Molecular Biochemistry”, Strategic Development of
Research Infrastructure for Private Universities, the Ministry of
Education, Culture, Sports, Science and Technology, (MEXT)
Japan. The author appreciates to Dr Hiroyasu Sato at Rigaku
Corporation for determination of crystal structures of 2a and 2b,
and to Dr Shinji Kitao at Kyoto University Research Reactor
Heuvelen, M. A. Cranswick, G. T. Rohde, E. L. Bominaar,
¨
E. Munck and L. Que, J. Am. Chem. Soc., 2011, 133, 11880;
(c) D. C. Lacy, R. Gupta, K. L. Stone, J. Greaves, J. W. Ziller,
M. P. Hendrich and A. S. Borovik, J. Am. Chem. Soc., 2010,
132, 12188; (d) J. P. Bigi, W. H. Harman, B. Lassalle-Kaiser,
D. M. Robles, T. A. Stich, J. Yano, R. D. Britt and
C. J. Chang, J. Am. Chem. Soc., 2012, 134, 1536.
16 G. Xue, A. Pokutsa and L. Que, J. Am. Chem. Soc., 2011, 133,
16657.
¨
Institute for measurements of part of the Mossbauer spectra.
17 A. Hazell, K. B. Jensen, C. J. McKenzie and H. Tolund,
J. Chem. Soc., Dalton Trans., 1993, 3249.
18 A. T. Fiedler, X. Shan, M. P. Mehn, J. Kaizer, S. Torelli,
J. R. Frisch, M. Kodera and L. Que Jr, J. Phys. Chem. A,
2008, 112, 13037.
Notes and references
1 D. M. Kurtz Jr, JBIC, J. Biol. Inorg. Chem., 1997, 2, 159.
¨
2 (a) L. Shu, J. C. Nesheim, K. Kauffmann, E. Munck, 19 (a) K. Kim and S. J. Lippard, J. Am. Chem. Soc., 1996, 118,
J. D. Lipscomb and L. Que, Science, 1997, 275, 515; (b)
C. E. Tinberg and S. J. Lippard, Acc. Chem. Res., 2011, 44, 280.
3 (a) W. J. Song, M. S. McCormick, R. K. Behan, M. H. Sazinsky,
W. Jiang, J. Lin, C. Krebs and S. J. Lippard, J. Am. Chem. Soc.,
4914; (b) Y. Dong, Y. Zang, L. Shu, E. C. Wilkinson, L. Que
¨
Jr, K. Kauffmann and E. Munck, J. Am. Chem. Soc., 1997,
119, 12683; (c) M. Kodera, M. Itoh, K. Kano and
T. Funabiki, Bull. Chem. Soc. Jpn., 2006, 79, 252.
2010, 132, 13582; (b) W. J. Song and S. J. Lippard, 20 A. M. Valentine, P. Tavares, A. S. Pereira, R. Davydov,
Biochemistry, 2011, 50, 5391.
4 L. J. Bailey and B. G. Fox, Biochemistry, 2009, 48, 8932.
C. Krebs, B. M. Hoffman, D. E. Edmondson, B. H. Huynh
and S. J. Lippard, J. Am. Chem. Soc., 1998, 120, 2190.
¨
5 V. V. Vu, J. P. Emerson, M. Martinho, Y. S. Kim, E. Munck, 21 (a) M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Chem.
M. H. Park and L. Que, Proc. Natl. Acad. Sci. U. S. A., 2009,
106, 14814.
Rev., 2004, 104, 939; (b) G. Xue, D. Wang, R. De Hont,
¨
A. T. Fiedler, X. Shan, E. Munck and L. Que, Proc. Natl.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 2282–2292 | 2291

View Article Online
Chemical Science
Edge Article
¨
Acad. Sci. U. S. A., 2007, 104, 20713; (c) G. Xue, A. T. Fiedler, 23 H. Zheng, S. J. Yoo, E. Munck and L. Que Jr, J. Am. Chem. Soc.,
¨
M. Martinho, E. Munck and L. Que, Proc. Natl. Acad. Sci. U. S.
2000, 122, 3789.
A., 2008, 105, 20615.
22 (a) F. A. Chavez, R. Y. N. Ho, M. Pink, V. G. Young Jr,
24 (a) Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput., 2005,
1, 415; (b) R. Ditcheld, W. J. Hehre and J. A. Pople, J. Chem.
Phys., 1971, 54, 724; (c) W. J. Hehre, R. Ditcheld and
J. A. Pople, J. Chem. Phys., 1972, 56, 2257; (d)
V. A. Rassolov, J. A. Pople, M. A. Ratner and T. L. Windus,
J. Chem. Phys., 1998, 109, 1223.
¨
S. V. Kryatov, E. V. Rybak-Akimova, H. Andres, E. Munck,
L. Que Jr and W. B. Tolman, Angew. Chem., Int. Ed.,
2002, 41, 149; (b) M. Yamashita, H. Furutachi, T. Tosha,
S. Fujinami, W. Saito, Y. Maeda, K. Takahashi,
K. Tanaka, T. Kitagawa and M. Suzuki, J. Am. Chem. Soc., 25 M. J. Frisch, et al., Gaussian09, Revision C.01, Gaussian Inc.,
2007, 129, 2.
Wallingford, CT, 2010.
2292 | Chem. Sci., 2014, 5, 2282–2292
This journal is © The Royal Society of Chemistry 2014