Oxygen activation at mononuclear nonheme iron centers: A superoxo perspective

Article abstract of DOI:10.1021/ic901891n

Dioxygen (O₂) activation by iron enzymes is responsible for many metabolically important transformations in biology. Often a high-valent iron oxo oxidant is proposed to form upon O₂ activation at a mononuclear nonheme iron center, presumably via intervening iron superoxo and iron peroxo species. While iron(IV) oxo intermediates have been trapped and characterized in enzymes and models, less is known of the putative iron(III) superoxo species. Utilizing a synthetic model for the 2-oxoglutarate-dependent monoiron enzymes, [(Tp^iPr2)Fe^II(O₂CC(O)CH₃)], we have obtained indirect evidence for the formation of the putative iron(III) superoxo species, which can undergo one-electron reduction, hydrogen-atom transfer, or conversion to an iron(IV) oxo species, depending on the reaction conditions. These results demonstrate the various roles that the iron(III) superoxo species can play in the course of O₂ activation at a nonheme iron center.

Full text of DOI:10.1021/ic901891n

3618 Inorg. Chem. 2010, 49, 3618–3628
DOI: 10.1021/ic901891n
Oxygen Activation at Mononuclear Nonheme Iron Centers: A Superoxo Perspective
Anusree Mukherjee,† Matthew A. Cranswick,† Mrinmoy Chakrabarti,‡ Tapan K. Paine,†,§ Kiyoshi Fujisawa,^
‡
,†
€
Eckard Munck, and Lawrence Que, Jr.*
†Department of Chemistry and Center for Metals in Biocatalysis, University of Minnesota, 207 Pleasant St. SE,
Minneapolis, Minnesota 55455, ^Department of Chemistry, Graduate School of Pure and Applied Sciences,
University of Tsukuba, Tsukuba 305-8571, Japan, and ‡Department of Chemistry, Carnegie Mellon University,
4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213. §Current address: Department of Inorganic Chemistry,
Indian Association for the Cultivation of Science, 2A & 2B Raja S. C. Mullick Road, Jadavpur, Kolkata 700032,
India.
Received September 24, 2009
Dioxygen (O2) activation by iron enzymes is responsible for many metabolically important transformations in biology.
Often a high-valent iron oxo oxidant is proposed to form upon O2 activation at a mononuclear nonheme iron center,
presumably via intervening iron superoxo and iron peroxo species. While iron(IV) oxo intermediates have been
trapped and characterized in enzymes and models, less is known of the putative iron(III) superoxo species. Utilizing a
synthetic model for the 2-oxoglutarate-dependent monoiron enzymes, [(TpiPr2)FeII(O2CC(O)CH3)], we have obtained
indirect evidence for the formation of the putative iron(III) superoxo species, which can undergo one-electron
reduction, hydrogen-atom transfer, or conversion to an iron(IV) oxo species, depending on the reaction conditions.
These results demonstrate the various roles that the iron(III) superoxo species can play in the course of O2 activation at
a nonheme iron center.
Introduction
illustrated in Scheme 1 for the proposed mechanism for the
2-oxoglutarate (2-OG)-dependent enzymes.6,7 The efforts of
Bollinger and Krebs have established that oxoiron(IV) inter-
mediates are, in fact, generated in the catalytic cycles of four
2-OG-dependent enzymes and that high-valent iron serves as
the oxidant for the key C-H bond cleavage step in many
reactions.8-10
In contrast, much less is known about the nature of the
FeO2 adducts that must initially form in the course of O2
activation. Often these species are considered merely as pass-
through points en route to the high-valent intermediates that
have been the focus of current efforts. However, recent
developments suggest that more attention should be paid to
some of these adducts because they may perform more
important roles than earlier construed.11
Many nonheme iron enzymes act by catalyzing the activa-
tion of dioxygen (O2) in many metabolically important
functions.1-5 Such functions include the hydroxylation of
methane in methanotrophs, the desaturation of fatty acids in
plants, DNA and RNA repair, the biosynthesis of β-lactam
antibiotics, and the sensing of hypoxia in mammalian cells to
signal the formation of blood vessels. As a consequence, this
subject has attracted the interest of bioinorganic chemists for
many years [see, for example, the special issue on oxygen
activation of Acc. Chem. Res. 2007, 40 (7), editedby Wonwoo
Nam]. At the simplest level, this process can be construed as a
series of electron-transfer steps with O2 progressing through
superoxo, peroxo, and oxo states concomitant with an
increase in the formal iron oxidation state. This notion is
Interestingly, a number of possible functions has been
proposed for the bound O2 in the mechanisms of these
nonheme iron enzymes.4 For Rieske dioxygenases like
naphthalene 1,2-dioxygenase (NDO), it serves to accept an
*To whom correspondence should be addressed. E-mail: larryque@
umn.edu.
(1) Solomon, E. I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.; Lee,
S.-K.; Lehnert, N.; Neese, F.; Skulan, A. J.; Yang, Y.-S.; Zhou, J. Chem. Rev.
2000, 100, 235–349.
(2) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. Rev. 2004,
104, 939–986.
ꢀ
(8) Krebs, C.; Galonic Fujimori, D.; Walsh, C. T.; Bollinger, J. M., Jr.
Acc. Chem. Res. 2007, 40, 484–492.
(9) Price, J. C.; Barr, E. W.; Hoffart, L. M.; Krebs, C.; Bollinger, J. M., Jr.
Biochemistry 2005, 44(22), 8138–8147.
(10) Galonic, D. P.; Barr, E. W.; Walsh, C. T.; Bollinger, J. M., Jr.; Krebs,
C. Nat. Chem. Biol. 2007, 3, 113–116.
(3) Hausinger, R. P. Crit. Rev. Biochem. Mol. Biol. 2004, 39, 21–68.
(4) Kovaleva, E. G.; Lipscomb, J. D. Nat. Chem. Biol. 2008, 4, 186–193.
(5) Loenarz, C.; Schofield, C. J. Nat. Chem. Biol. 2008, 4, 152–156.
€
(6) Hanauske-Abel, H. M.; Gunzler, V. J. Theor. Biol. 1982, 94, 421–455.
(7) Hanauske-Abel, H. M.; Popowicz, A. M. Curr. Med. Chem. 2003, 10,
1005–1019.
(11) Bollinger, J. M., Jr.; Krebs, C. Curr. Opin. Chem. Biol. 2007, 11, 151–
158.
r
pubs.acs.org/IC
Published on Web 04/12/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 8, 2010 3619
Scheme 1. General Mechanism for 2-OG-Dependent Iron Enzymes
Figure 1. Active-site structures of the O2 adducts of nonheme iron
enzymes that have been structurally characterized by X-ray crystallogra-
phy (orange, Fe; gray, C; red, O; blue, N; green, Cl). (A) End-on O2
adductofHAD fromRalstonia metalliduranscomplexed with an inhibitor
(1YFW.pdb). (B) Side-on O2 adduct NDO from Pseudomonas sp. in the
presence of the substrate analogue indole (1O7N.pdb). (C and D) Side-on
O2 adduct of the HPCD-4-nitrocatechol complex of Brevibacterium
fuscum found in subunit C and the subsequent alkylperoxo intermediate
found in subunits B and D, respectively (2IGA.pdb).
electron from reduced nicotinamide adenine dinucleotide via
a nearby Rieske iron-sulfur cluster to form an iron(III)
peroxo intermediate,12-14 completely analogous to its role in
the cytochrome P450 mechanism. For the large 2-OG-de-
pendent enzyme family, the bound O2 acts as a nucleophilic
superoxide and attacks the electrophilic keto carbon of
2-OG.6 For extradiol cleaving dioxygenases like homopro-
tocatechuate 2,3-dioxygenase (HPCD), the ligated O2 is
proposed to accept an electron from the adjacently bound
catecholate substrate using the metal center as a conduit to
generate an iron(II)/substrate radical/superoxo complex that
leads to an alkylperoxoiron(II) intermediate.15 For isopeni-
cillin N synthase (IPNS)11 and hydroxyethylphosphonate
dioxygenase (HEPD),16 the bound O2 is proposed to initiate
the four-electron oxidation of the substrate by abstracting a
hydrogen atom from the substrate; similar hydrogen-atom
abstractions by copper superoxo species are proposed for
dopamine hydroxylase and peptidyl-amidating monooxy-
genase.17,18 How an FeO2 adduct can be tuned to perform
such a variety of functionsisa fascinating question forfurther
investigation.
be electron paramagnetic resonance (EPR) active. Indeed,
rapid-freeze-quench samples of the reaction of a Mn-
HPCD-native substrate complex with O2 obtained 15 ms
after mixing gave rise to a short-lived S = 5/2 species in ∼5%
yield.19 Spectral analysis and comparison with the unreacted
Mn-HPCD-substrate complex led to its assignment as a
MnIII radical species, with the radical favored to be a super-
oxide. A later intermediate is also observed at higher con-
version and proposed to be a manganese(II) alkylperoxo
species that is further along on the reaction pathway. Analo-
gous intermediates presumably form with native Fe-HPCD
but will require alternative methods for detection.
The other superoxo intermediate is observed for myo-
inositol oxygenase (MIOX), the mammalian enzyme that
carries out the four-electron oxidation of myo-inositol to
D
-glucoronate by cleavage of the C1-C6 bond.21 X-ray
crystallography reveals MIOX to have a carboxylate-bridged
To date, spectroscopic evidence is available for only two
superoxo intermediates among iron enzymes. One example
derives from HPCD, but the superoxo intermediate is that
derived from the manganese-substituted enzyme.19 The man-
ganese substitution affords an equally active enzyme20 and
has the advantage that a MnIIO2 adduct would be expected to
diiron active site,22 which is shown to be in the FeIIFeIII state
23
€
by EPR and Mossbauer measurements. In the crystal
structure of the enzyme-substrate complex,22 myo-inositol
binds as a bidentate ligand to one iron, presumably the
iron(III) site, to activate the substrate. O2 is then proposed
to bind at the other iron, presumably the iron(II) site. Indeed,
a reversible O2 adduct can be trapped, and it exhibits an
S = 1/2 EPR signal that is proposed to arise from a diiron(III)
superoxo species.24 Because the decay of this species can be
retarded with the use of deuterated substrate, it is surmised
(12) Wolfe, M. D.; Parales, J. V.; Gibson, D. T.; Lipscomb, J. D. J. Biol.
Chem. 2001, 276, 1945–1953.
€
(13) Wolfe, M. D.; Altier, D. J.; Stubna, A.; Popescu, C. V.; Munck, E.;
Lipscomb, J. D. Biochemistry 2002, 41, 9611–9626.
(14) Wolfe, M. D.; Lipscomb, J. D. J. Biol. Chem. 2003, 278, 829–835.
(15) Kovaleva, E. G.; Lipscomb, J. D. Science 2007, 316, 453–457.
(16) Cicchillo, R. M.; Zhang, H.; Blodgett, J. A. V.; Whitteck, J. T.; Li,
G.; Nair, S. K.; van der Donk, W. A.; Metcalf, W. W. Nature 2009, 459, 871–
874.
(17) Evans, J. P.; Ahn, K.; Klinman, J. P. J. Biol. Chem. 2003, 278, 49691–
49698.
(18) Chen, P.; Solomon, E. I. J. Am. Chem. Soc. 2004, 126, 4991–5000.
(19) Gunderson, W. A.; Zatsman, A. I.; Emerson, J. P.; Farquhar, E. R.;
Que, L., Jr.; Lipscomb, J. D.; Hendrich, M. P. J. Am. Chem. Soc. 2008, 130,
14465–14467.
(20) Emerson, J. P.; Kovaleva, E. G.; Farquhar, E. R.; Lipscomb, J. D.;
Que, L., Jr. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 7347–7352.
(21) Bollinger, J. M., Jr.; Diao, Y.; Matthews, M. L.; Xing, G.; Krebs, C.
Dalton Trans. 2009, 905–914.
(22) Brown, P. M.; Caradoc-Davies, T. T.; Dickson, J. M. J.; Cooper,
G. J. S.; Loomes, K. M.; Baker, E. N. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
15032–15037.
(23) Xing, G.; Barr, E. W.; Diao, Y.; Hoffart, L. M.; Prabhu, K. S.; Arner,
R. J.; Reddy, C. C.; Krebs, C.; Bollinger, J. M., Jr. Biochemistry 2006, 45,
5402–5412.
(24) Xing, G.; Diao, Y.; Hoffart, L. M.; Barr, E. W.; Prabhu, K. S.; Arner,
R. J.; Reddy, C. C.; Krebs, C.; Bollinger, J. M., Jr. Proc. Natl. Acad. Sci.
U.S.A. 2006, 103, 6130–6135.

3620 Inorganic Chemistry, Vol. 49, No. 8, 2010
Mukherjee et al.
that the superoxo intermediate must be involved in the
cleavage of the labeled C-H bond, which likely represents
the first step in substrate oxidation.24 This proposed hydro-
gen-atom abstraction by the iron-bound O2 is, in fact,
analogous to the initial steps postulated in the mechanisms
of IPNS and HEPD briefly mentioned above.11,16
Scheme 2. Various Proposed Roles for the Initial FeO2 Adducts in
Reaction Pathways of 2-His-1-Carboxylate Iron Enzymes
X-ray crystallography has been used effectively in some
enzymes for visualizing the interaction of O2 with active-site
iron centers in this family of enzymes.15,25,26 An end-on-
bound O2 like those seen in heme-O2 adducts27 is found in
the structure of the 3-hydroxyanthranilate-3,4-dioxygenase
(HAD)-inhibitor-O2 complex (Figure 1A),26 while side-on-
bound O2 moieties have been observed in crystals of the
NDO-indole-O2 and the HPCD-4-nitrocatechol-O2
complexes (Figure 1B,C).15,25 A consideration of these struc-
tures, together with related mechanistic information, suggests
to us that the end-on-bound O2 in the HAD-inhibitor-O2
complex26 likely represents an early stage of the Fe-O2
interaction, while the side-on-bound O2 adducts correspond
to species wherein a subsequent electron-transfer step has
occurred.
For example, the NDO-indole-O2 adduct (Figure 1B) is
proposed to be an iron(III) peroxo species that derives from
the one-electron reduction of the initial FeO2 adduct by the
nearby reduced Rieske Fe2S2 cluster. The O-O bond dis-
25
˚
˚
tance is 1.46 A, and the Fe-O distances are 1.7 and 2.0 A,
consistent with an iron(III) η2-peroxo assignment.28,29 This
difference in the Fe-O bond lengths suggests that the more
distal oxygen atom may be protonated. The bound O2 is
found to be in an ideal geometry to attack the substrate CdC
bond to afford the cis-dihydrodiol product (Scheme 2). The
cis-dihydroxylation may occur via a concerted mechanism
where the O-O bond cleaves simultaneously with the for-
mation of the two C-O bonds or via a stepwise pathway
involving the initial O-O bond cleavage and subsequent
attack of the CdC bond by the resulting high-valent iron oxo
oxidant.30
catalytically active in crystallo. These results represent a
crystallographic tour de force by Kovaleva and Lipscomb15
and firmly establish the mechanistic notions previously
proposed for extradiol cleavage.31
The examples listed above demonstrate that much needs to
be clarified with respect to the chemistry of FeO2 adducts.
Biomimetic complexes can be useful in addressing some of
these questions because of their simpler structures and the
relative ease of ligand modification. However, the big chal-
lenge for synthetic bioinorganic chemists is to design model
complexes that mimic function and are amenable for
mechanistic dissection; only then can they help shed light
on the questions raised by the crystallographic results.
To date, there are only a handful of well-characterized
synthetic FeO2 adducts with nonheme ligand environments.
The three for which crystal structures are available all
represent 1,2-peroxo-bridged diiron(III) complexes obtained
from the reaction of O2 with iron(II) precursors.32-34 It is
plausible that these adducts form via the initial interaction of
O2 with one of the two iron atoms, but no evidence is
currently available for such a 1:1 FeO2 adduct in the course
of forming any of these three complexes.
There is strong crystallographic evidence for a different
electron-transfer event for the HPCD-4-nitrocatechol-O2
˚
adduct. In this case, the bound O2 has a 1.3 A O-O distance,
˚
the Fe-O distances average 2.5 A, and the bound 4-nitroca-
techolate ion has lost its aromaticity and become nonplanar
(Figure 1C and Scheme 2).15 These observations have been
interpreted to indicate electron transfer from the substrate to
the bound O2 via the iron(II) center, resulting in a
semiquinone-iron(II) superoxo complex. The two radical
moieties are then postulated to couple, forming an
alkylperoxoiron(II) intermediate (Figure 1D). Amazingly,
the species from this radical-radical coupling step is ob-
served in the two subunits adjacent to the subunit containing
the semiquinone-iron(II) superoxo structure in the four-
subunit unit cell. An enzyme-product complex is observed in
the fourth subunit, demonstrating that HPCD is, in fact,
However, in the case of a related diiron(II) precursor
[FeII2(μ-OH)2(6-Me3-TPA)2]2þ (Scheme 3), its oxygenation
appears to be more controlled to allow the transfer of one
electron at a time.35 Indeed, a superoxo intermediate can be
(25) Karlsson, A.; Parales, J. V.; Parales, R. E.; Gibson, D. T.; Eklund,
H.; Ramaswamy, S. Science 2003, 299, 1039–1042.
(26) Zhang, Y.; Colabroy, K. L.; Begley, T. P.; Ealick, S. E. Biochemistry
2005, 44, 7632–7643.
(27) Momenteau, M.; Reed, C. A. Chem. Rev. 1994, 94, 659–698.
(28) Roelfes, G.; Vrajmasu, V.; Chen, K.; Ho, R. Y. N.; Rohde, J.-U.;
Zondervan, C.; la Crois, R. M.; Schudde, E. P.; Lutz, M.; Spek, A. L.; Hage,
(31) Lipscomb, J. D.; Orville, A. M. Metal Ions in Biological Systems;
Sigel, H., Sigel, A., Eds.; Marcel Dekker: New York, 1992; Vol. 28, pp 243-298.
(32) Ookubo, T.; Sugimoto, H.; Nagayama, T.; Masuda, H.; Sato, T.;
Tanaka, K.; Maeda, Y.; Okawa, H.; Hayashi, Y.; Uehara, A.; Suzuki, M.
J. Am. Chem. Soc. 1996, 118, 701–702.
(33) Dong, Y.; Yan, S.; Young, V. G., Jr.; Que, L., Jr. Angew. Chem., Int.
Ed. Engl. 1996, 35, 618–620.
(34) Kim, K.; Lippard, S. J. J. Am. Chem. Soc. 1996, 118, 4914–4915.
(35) Shan, X.; Que, L., Jr. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5340–
5345.
€
R.; Feringa, B. L.; Munck, E.; Que, L., Jr. Inorg. Chem. 2003, 42, 2639–2653.
(29) Koehntop, K. D.; Rohde, J.-U.; Costas, M.; Que, L., Jr. Dalton
Trans. 2004, 3191–3198.
(30) Que, L., Jr. J. Biol. Inorg. Chem. 2004, 9, 684–690.

Article
Inorganic Chemistry, Vol. 49, No. 8, 2010 3621
Scheme 3. Ligands
Figure 2. UV-vis spectra in MeCN of
1 (black solid line), 2
(blue dashed-dotted line), the yellow solution obtained upon oxygenation
(red dashed line), and the solution obtained upon oxygenation in the
presence of THT (red solid line). Inset: corresponding magnified spectra
in the visible region.
stabilized at -80 ꢀC, which converts to the well-characterized
1,2-peroxo-bridged intermediate [FeIII2(μ-O)(μ-1,2-O2)-
(6-Me3-TPA)2]2þ upon warming to -60 ꢀC.35 The superoxo
intermediate is characterized by a ν(O-O) Raman mode
found at 1310 cm-1 that downshifts 71 cm-1 with 18O2. A
doublet pattern is observed for the ν(O-O) mode associated
with the 16O18O isotopomer in the mixed isotope Raman
experiment, supporting the notion that the superoxo moiety
is bound end-on. The ν(O-O) mode of 1310 cm-1 represents
the highest frequency yet observed for a metal-bound super-
oxo species. It is at least 100 cm-1 higher than those observed
for O2 adducts of heme proteins, synthetic iron porphyrins,
and biomimetic copper complexes.27,36 The smaller extent of
electron transfer from metal to bound O2, as reflected by the
higher ν(O-O) frequency, can be easily rationalized by the
fact that the iron centers in [FeII2(μ-OH)2(6-Me3-TPA)2]2þ
are the most Lewis acidic of the metal centers that give rise to
superoxo intermediates. Like their copper superoxo counter-
tparts,37-40 this intermediate is reactive even at -80 ꢀC and
can abstract a hydrogen atom from 2,4-di-tert-butylphenol,35
a reactivity that stands in stark contrast to the inability of the
corresponding peroxo intermediate to react with the same
phenol at -60 ꢀC.
surround and protect the diiron center.41 For these com-
plexes, O2 binding results in the formation of an EPR-silent
€
intermediate that exhibits two Mossbauer doublets corre-
sponding to an antiferromagnetically coupled FeIIFeIII cen-
ter. Unfortunately, in this case, no resonance Raman (rR)
data could be obtained.
One of the synthetic FeO2 adducts that have been char-
acterized crystallographically is [FeIII2(TpiPr2)2(O2CR)2-
(O2)],34 which derives from a mononuclear precursor
[FeII(TpiPr2)(O2CR)] first reported by Kitajima and co-work-
ers.42,43 By analogy to the established mechanism for forming
a (μ-1,2-peroxo)dicopper complex,44 an iron(III) superoxo
species would seem to be a highly plausible intermediate, but
no evidence for this complex has been reported. Related to
this complex are FeII(TpR2) complexes (R = Me or Ph) of
R-ketocarboxylates that serve as structural and functional
models for the 2-OG-dependent enzymes. These complexes
have been found to react with O2 and undergo oxidative
decarboxylation of the bound R-ketocarboxylate. This reac-
tion produces a presumed high-valent iron oxo oxidant cap-
able of intramolecular ligand hydroxylation, intermolecular
olefin epoxidation, or hydrocarbon dehydrogenation.45-49
Studies of the Fe(TpPh2) system showed that the rate of re-
action increased as a more electron-withdrawing substituent
was introduced into the R-ketocarboxylates. This behavior
was rationalized by Mehn et al. as indicating the rate-limiting
attack by a nucleophilic superoxide on the electrophilic
keto carbon.47 In this paper, we describe two complexes,
[FeII(TpiPr2)(O2CC(O)R)] (1, R = CH3; 2, R = Ph), and
The observed quantitative conversion of [FeII2(μ-OH)2-
(6-Me3-TPA)2]2þ to [FeIII2(μ-O)(μ-1,2-O2)(6-Me3-TPA)2]2þ
suggests that the intervening superoxo intermediate must
result from one-electron transfer from the diiron(II) center to
the bound O2 to form an FeIIFeIII superoxo adduct. As
expected, this species is EPR-silent, presumably because of
antiferromagnetic coupling. Unfortunately, it has not been
€
possible to ascertain the diiron oxidation state by Mossbauer
spectroscopy because the superoxo intermediate forms only
in CH2Cl2, a solvent that severely hampers the application of
this technique. Related chemistry has been reported for
diiron(II) complexes of carboxylates that are appended to
dendrimers, which provide a hydrophobic environment to
(42) Kitajima, N.; Fukui, H.; Moro-oka, Y.; Mizutani, Y.; Kitagawa, T.
J. Am. Chem. Soc. 1990, 112, 6402–6403.
(43) Kitajima, N.; Tamura, N.; Amagai, H.; Fukui, H.; Moro-oka, Y.;
Mizutani, Y.; Kitagawa, T.; Mathur, R.; Heerwegh, K.; Reed, C. A.;
Randall, C. R.; Que, L., Jr.; Tatsumi, K. J. Am. Chem. Soc. 1994, 116,
9071–9085.
(36) Cramer, C. J.; Tolman, W. B. Acc. Chem. Res. 2007, 40, 601–608.
€
(37) Wurtele, C.; Gaoutchenova, E.; Harms, K.; Holthausen, M. C.;
€
(44) Karlin, K. D.; Wei, N.; Jung, B.; Kaderli, S.; Niklaus, P.; Zuberbuhler,
Sundermeyer, J.; Schindler, S. Angew. Chem., Int. Ed. 2006, 45, 3867–3869.
(38) Maiti, D.; Fry, H. C.; Woertink, J. S.; Vance, M. A.; Solomon, E. I.;
Karlin, K. D. J. Am. Chem. Soc. 2007, 129, 264–265.
A. D. J. Am. Chem. Soc. 1993, 115, 9506–9514.
(45) Ha, E. H.; Ho, R. Y. N.; Kesiel, J. F.; Valentine, J. S. Inorg. Chem.
1995, 34, 2265–2266.
(46) Hegg, E. L.; Ho, R. Y. N.; Que, L., Jr. J. Am. Chem. Soc. 1999, 121,
1972–1973.
(47) Mehn, M. P.; Fujisawa, K.; Hegg, E. L.; Que, L., Jr. J. Am. Chem.
Soc. 2003, 125, 7828–7842.
€
(39) Maiti, D.; Lee, D.-H.; Gaoutchenova, K.; Wurtele, C.; Holthausen,
M. C.; Sarjeant, A. A. N.; Sundermeyer, J.; Schindler, S.; Karlin, K. D.
Angew. Chem., Int. Ed. 2008, 47, 82–85.
(40) Kunishita, A.; Kubo, M.; Sugimoto, H.; Ogura, T.; Sato, K.; Takui,
T.; Itoh, S. J. Am. Chem. Soc. 2009, 131, 2788–2789.
(41) Zhao, M.; Helms, B.; Slonkina, E.; Friedle, S.; Lee, D.; DuBois, J.;
(48) Fujisawa, K.; Tada, N.; Nishida, Y.; Miyashita, Y.; Okamoto, K.
Inorg. Chem. Commun. 2008, 11, 381–384.
ꢀ
€
Hedman, B.; Hodgson, K. O.; Frechet, J. M. J.; Lippard, S. J. J. Am. Chem.
(49) Mukherjee, A.; Martinho, M.; Bominaar, E. L.; Munck, E.; Que, L.,
Soc. 2008, 130, 4352–4363.
Jr. Angew. Chem., Int. Ed. 2009, 48, 1780–1783.

3622 Inorganic Chemistry, Vol. 49, No. 8, 2010
Mukherjee et al.
Figure 3. ORTEP plots for 1 and 2. Hydrogen atoms are omitted for clarity.
compare their reactions with O2 with those of the previously
characterized [FeII(TpiPr2)(O2CPh)] (3) to gain insight into
what role a superoxoiron(III) intermediate may play in O2
activation reactions of nonheme iron complexes.
Selected bond lengths are listed in Table 1 and com-
pared with those from related structures. The Fe-
0
Npyrazole bond lengths of 1 and 2 are typical of TpR,R
complexes with high-spin iron(II) centers.43,47,52 1H NMR
spectra of 1 and 2 support this conclusion as well (see the
Experimental Section). Both R-keto acids bind as mono-
anions, as indicated by the absence of counterions in 1
and 2. While the observed Fe-Oketo bond lengths a0re
comparable to those found for related [Fe(TpR,R )-
(R-ketocarboxylate)]complexes, theFe-Ocarboxylato bonds
Results and Discussion
Synthesis and Characterization of 1 and 2. The two
complexes 1 and 2 were synthesized by mixing equimolar
portions of iron(II) salt with KTpiPr2, followed by the
addition of sodium pyruvate (NaPRV) or sodium ben-
zoylformate (NaBF) (see Scheme 3 for ligand structures) .
Complex 1 exhibits a distinct red color in solution corre-
sponding to a visible band with an absorption maximum
at 500 nm along with two shoulders at 540 and 458 nm
(Figure 2). The corresponding transitions in 2 are red-
shifted with a maximum at 610 nm, giving rise to a blue
solution (Figure 2). These bands in the visible region are
assigned to metal-to-ligand charge-transfer (MLCT)
transitions of an iron(II) κ2-R-ketocarboxylate unit by
analogy to previous work.47,50 The MLCT assignment is
consistent with the observed red shift upon replacement
of the methyl group with a more electron-withdrawing
phenyl group on the R-ketocarboxylate.47
˚
are at least 0.14 A longer than those found for five-
coordinate complexes.
The observed nuclearities of 1 and 2 derive from a
difference in the solvents used for the preparation of the
complexes and their recrystallization, MeOH/CH2Cl2 for
1 and CH2Cl2/pentane for 2. These results are consistent
with those reported by Kitajima and co-workers for 3,
which crystallized as the five-coordinate complex in
pentane and as the six-coordinate solvate in MeCN.42,43
Indeed, FeTp complexes are often found to crystallize as
six-coordinate MeCN or pyrazole adducts.43,47,52 It
would thus appear likely that 2 is mononuclear in the
MeCN solution. In support, the λmax for 2 in toluene is
observed at 627 nm, somewhat red-shifted relative to that
in MeCN (610 nm), whereas λmax for 1 remains un-
changed in both of these solvents. The latter observation
suggests that a change in the solvent polarity does not
affect the position of the MLCT band in 1. The blue shift
observed for 2 in MeCN must have a different rationale
and can be understood by considering that 2 retains its
dimeric structure in the nonpolar toluene solvent but
becomes monomeric in the coordinating MeCN solvent.
Thus, the displacement of the carboxylate oxygen from
the other iron unit of dimeric 2 by a stronger-field MeCN
ligand to the iron(II) center would be expected to lower
the energy of the iron(II) dπ orbitals and increase the
energy gap for the iron(II)-to-R-ketocarboxylate MLCT
transition.
Diffraction-quality crystals could be obtained for both
1 and 2, and their crystal structures are shown in Figure 3.
The structure of 1 shows a six-coordinate iron center with
a face-capping tridentate TpiPr2, a bidentate PRV, and a
methanol solvate. The structure of 2 exhibits a similar
iron coordination environment but features a centrosym-
metric dinuclear complex where the methanol solvate in
the sixth coordination site of 1 is replaced by the carbonyl
oxygen atom of the carboxylate of the BF ligand on the
adjacent iron unit in 2. Complex 2 represents, in fact, the
first example of a complex with a κ3-bridging mode of BF.
The only other example of a κ3-bridging mode for an
R-keto acid is found in the structure of [FeII2(6-Me3-
TPA)2(phenylpyruvate)]2þ 51
In this latter case, the keto
.
group of the phenylpyruvate has become enolized and the
ligand is bound as a dianion, as indicated by a C2-C3
bond length corresponding to a double bond. In contrast,
the C2-C3 bond of the BF ligand of 2 has a C-C bond
Reactions of 1 and 2 with O2 at Room Temperature.
Bubbling O2 through the red solution of 1 in MeCN at
25 ꢀC resulted in the onset of immediate spectral changes
consisting of a loss of the MLCT band at 500 nm and its
replacement by an intense near-UV feature with an
˚
length consistent with a C-C single bond [1.475(3) A],
and the BF binds as the monoanionic keto tautomer.
(52) Hikichi, S.; Ogihara, T.; Fujisawa, K.; Kitajima, N.; Akita, M.;
Moro-oka, Y. Inorg. Chem. 1997, 36, 4539–4547.
(50) Chiou, Y.-M.; Que, L., Jr. J. Am. Chem. Soc. 1995, 117, 3999–4013.
(51) Paine, T. K.; Zheng, H.; Que, L., Jr. Inorg. Chem. 2005, 44, 474–476.

Article
Inorganic Chemistry, Vol. 49, No. 8, 2010 3623
Table 1. Comparison of Selected Bond Lengths of 1 and 2 with Those of Related Structures
1
2
[Fe(TpPh2)(PRV)]a
[Fe(TpPh2)(BF)]a
[Fe(TptBu,iPr)(BF)]b
Fe-Ocarboxylato
Fe-Oketo
2.1180(18)
2.2294(18)
2.181(2)
2.138(2)
2.139(2)
1.483(4)
2.1691(15), 2.1832(15)c
2.2514(15)
2.1954(17)
2.1041(17)
2.0919(17)
1.475(3)
1.967(5)
2.242(6)
2.163(6)
2.088(8)
2.106(8)
1.498(9)
1.968(4)
2.206(5)
2.188(5)
2.068(5)
2.086(5)
1.975(7)
2.261(7)
2.116(9)
2.173(9)
2.112(9)
Fe-N2
Fe-N4
Fe-N6
C2-C3R-keto acid
a Reference 47. b Reference 52. c Corresponding to r(Fe1-O30).
absorption maximum at 370 nm (ε = 6000 M-1 cm-1
)
yellow solutions with a strong acid. In the case of 2, the
quantitative conversion of benzoylformate to benzoate
was indicated by the disappearance of peaks character-
istic of the ortho protons of benzoylformate at 8.36 ppm
and the appearance of corresponding benzoate features at
8.14 ppm (Figure S3 in the Supporting Information).
More importantly for both 1 and 2, NMR features were
observed for two pyrazoles, namely, 3,5-diisopropylpyr-
azole (5.90 ppm for pyrazole-4-H) and 3-isopropenyl-5-
isopropylpyrazole (5.84 ppm for pyrazole-4-H and 5.3 ppm
for one of the vinyl C-H protons). These features were
found to be in a ratio of 2:1 (Figure 4), showing that one of
the three pyrazoles in the TpiPr2 ligand had been modified.
Both 3,5-diisopropylpyrazole and 3-isopropenyl-5-iso-
propylpyrazole were formed upon acid decomposition
of the TpiPr2* ligand. 3-Isopropenyl-5-isopropylpyrazole
was most likely derived from dehydration of hydroxy-
lated 3,5-diisopropylpyrazole because tertiary alcohols
are well-known to lose water easily upon reaction with
acid to generate alkenes. Notably, ESI-MS analysis of the
yellow product solution prior to acid decomposition
(Figure 4) did not show a prominent peak at m/z = 519
corresponding to the {(TpiPr2)FeII - 2H}þ ion, indicating
that 3-isopropenyl-5-isopropylpyrazole was not formed
during the initial oxygenation. These NMR results de-
monstrate the quantitative oxidative decarboxylation of
the bound R-keto acid and concomitant formation of an
oxidant that carries out the ligand oxidation, as prece-
dented in the oxygenation of [FeII(TpPh2)(R-keto-
carboxylate)] complexes.46,47,51 In the present study,
ligand hydroxylation appears to be quantitative.
over the course of 15 min (Figure 2). Electrospray ioniza-
tion mass spectrometry (ESI-MS) analysis suggested the
formation of an alkoxoiron(III) product, as manifested
by peaks at m/z = 536 and 688 that were observed along
with peaks at m/z = 521 and 673. The peak at m/z = 521
together with its isotope distribution pattern can be
assigned to the [FeII(TpiPr2)]þ ion. On the other hand,
the peak at m/z = 536 has a mass and an isotope
distribution pattern that correspond to an ion with the
composition of {[(TpiPr2)FeIII] þ O - H}þ(see Figure 4
and Figure S1 in the Supporting Information); that is, one
hydrogen atom has been replaced by an oxygen atom,
suggesting the hydroxylation of the TpiPr2 ligand in the
course of the reaction to form the oxidation product 4.
The most likely hydroxylation site is at one of the
3-isopropylmethine carbons because this is the weakest
C-H bond present in the ligand and is also in close
vicinity of the iron center. Hydroxylation of an isopropyl
group of the TpiPr2 ligand would introduce a tertiary
alcohol functionality near the iron center and lead to the
formation of a metal-alkoxide bond. We thus assign the
m/z = 536 peak as [Fe(TpiPr2*)]þ where TpiPr2* represents
the TpiPr2 ligand in which the 3ꢀ C-H bond of a
3-isopropyl group on the TpiPr2 ligand has been hydro-
xylated. The peaks at m/z = 673 and 688 correspond
respectively to the 3,5-diisopropylpyrazole adducts of
the ions with m/z = 521 and 536. Analogous attack of a
3-isopropyl C-H bond has previously been demon-
strated in the reactions of [MnII2(TpiPr2)2(OH)2] and
[CoI(TpiPr,Me)] with O2, with a crystal structure reported
for the alkoxomanganese product complex.53,54 The yel-
low chromophore of 4 thus arises from an alkoxo-to-
iron(III) charge-transfer transition, as observed for
To further characterize the oxidant responsible for
ligand hydroxylation, experiments were carried out in
the presence of potential substrates in an attempt to
intercept this oxidant intermolecularly. Such intercep-
tions by either dialkyl sulfides or hydrocarbons have been
documented in the oxygenation of [FeII(TpPh2)(BF)].49
When the oxygenation of 1 or 2 was performed in the
presence of 100 mM tetrahydrothiophene (THT), the
[FeIII(Py5)(OMe)]2þ (λmax, 337 nm; ε, 3600 M-1 cm-1
;
Py5 = 2,6-bis[bis(2-pyridyl)methoxymethane]pyridine)55
and [FeIII(N4Py)(OMe)]2þ (λmax, 360 nm; ε, 4000 M-1
cm-1; N4Py = N,N-bis(2-pyridylmethyl)bis(2-pyridyl)
methylamine).56
An additional corroboration for the ESI-MS results
was obtained from 1H NMR analysis of the organic
components of the yellow solutions derived from the
oxygenation of 1 and 2, extracted after treatment of the
intense
yellow
chromophore
associated
with
[Fe(TpiPr2*)]þ did not form. ESI-MS analysis of this
solution revealed only one major peak at m/z = 625,
whose mass and isotope distribution pattern matched
those for [FeII(TpiPr2){(OS(CH2)4}]þ (Figure S2 in the
Supporting Information). Tetramethylene sulfoxide was
confirmed to be the oxidation product by gas chroma-
tography (GC) and GC-MS analysis (80% yield with
naphthalene as an internal standard). Interestingly,
neither thioanisole nor diphenyl sulfide was capable of
intercepting the oxidant, in contrast to what was reported
in the oxygenation of [Fe(TpPh2)(BF)];49 instead, the 370-nm
chromophore obtained in the absence of any substrate
(53) Kitajima, N.; Osawa, M.; Tanaka, M.; Moro-oka, Y. J. Am. Chem.
Soc. 1991, 113, 8952–8953.
(54) Reinaud, O. M.; Theopold, K. H. J. Am. Chem. Soc. 1994, 116, 6979–
6980.
(55) Goldsmith, C. R.; Jonas, R. T.; Stack, T. D. P. J. Am. Chem. Soc.
2002, 124, 83–96.
(56) Roelfes, G.; Lubben, M.; Chen, K.; Ho, R. Y. N.; Meetsma, A.;
Genseberger, S.; Hermant, R. M.; Hage, R.; Mandal, S. K.; Young, V. G.,
Jr.; Zang, Y.; Kooijman, H.; Spek, A. L.; Que, L., Jr.; Feringa, B. L. Inorg.
Chem. 1999, 38, 1929–1936.

3624 Inorganic Chemistry, Vol. 49, No. 8, 2010
Mukherjee et al.
Figure 5. Oxygenation of 1 mM 1 (dashed line) at -40 ꢀC in MeCN
resulting in the formation of a green chromophore (solid line) that is
assigned to a (μ-1,2-peroxo)diiron(III) complex.
Figure 4. Top: ESI-MS spectrum of 1 after oxygenation but prior to
acid treatment. Bottom: 1H NMR spectrum in the 5.2-6.2 ppm region of
the pyrazoles in CDCl3 obtained from the reaction of 1 with O2 after acid
treatment to decompose the TpiPr2* ligand (f-marked peaks arise from
3-isopropenyl-5-isopropylpyrazole; the peak at 5.90 ppm is the 4-H of
3,5-diisopropylpyrazole).
was observed. Thus, 1 and 2 are only able to carry out oxo
transfer to THT. Both steric and electronic factors can be
used to rationalize the lack of reactivity toward thioani-
sole and diphenyl sulfide. Clearly, the rates of oxo trans-
fer to these potential substrates are smaller than the rate
of intramolecular C-H bond cleavage by the incipient
oxidant.
Reactions of 1 and 2 with O2 at -40 ꢀC. The reactivities
of 1 and 2 with O2 diverge at -40 ꢀC. While the reactivity
of 2 with O2 at -40 ꢀC mirrored that at 25 ꢀC, i.e.,
intramolecular ligand hydroxylation and intermolecular
sulfoxidation, this behavior did not apply to 1. Oxygena-
tion of 1 at -40 ꢀC in MeCN afforded a distinct chromo-
phore with λmax at 680 nm (ε ∼ 1600 M-1 cm-1/Fe) that
was stable for days at this temperature (Figure 5). This
green chromophore was very similar to those observed in
the oxygenation of 3 (λmax = 682 nm; ε = 1700 M-1
cm-1/Fe) and [FeII(TpiPr2)(O2CCH2Ph)] (λmax = 694 nm;
ε = 1300 M-1 cm-1/Fe) in toluene, which was shown by
spectroscopy and crystallography to be [Fe2(O2)(TpiPr2)2-
(O2CR)2], a complex with a (μ-1,2-peroxo)diiron(III)
core supported by two carboxylate bridges.34,43
Figure 6. rR spectra using λex = 647.1 nm of the green chromophore
from the oxygenation of 1 at -40 ꢀC in MeCN with 16O2 (top) and 18O2
(bottom). Solvent peaks are denoted as “S”, and the asterisk denotes a
laser plasma line.
complexes in the literature,2,57,58 only a handful derive from
O2 binding to a mononuclear iron complex.
In order to further corroborate the nature of the O2
adduct, the frozen MeCN solution of the green chromo-
€
phore was studied with Mossbauer spectroscopy. The
zero-field spectrum of Figure 7A exhibits two doublets.
One doublet [ΔEQ = 3.42(4) mm/s, δ = 1.21(2) mm/s,
40% of Fe] is attributable to a high-spin iron(II) complex
and very likely represents the starting material 1. The
second doublet (45% of Fe), outlined by the solid line, has
parameters [ΔEQ = 1.32(3) mm/s and δ = 0.65(2) mm/s]
very similar to those of the O2 adduct of [FeII-
(TpiPr2)(O2CCH2Ph)] characterized by Kim and Lip-
pard.34 Studies in strong applied magnetic fields (Figure
S4 in the Supporting Information) revealed that the sam-
ple contained also a mononuclear S = 5/2 iron(III) species
(10%). The spectra in Figures 7B and 7C, representing
the green chromophore, were obtained by subtracting
The rR spectrum of the green chromophore from the
oxygenation of 1 at -40 ꢀC was thus obtained to deter-
mine its similarity to the O2 adduct described by Kitajima
and co-workers.43 Excitation at 647.1 nm elicited two
resonance-enhanced vibrations at 889 and 424 cm-1
(Figure 6); these were respectively assigned to the
ν(O-O) and ν(Fe-O) modes of the (μ-1,2-peroxo)diiron-
(III) unit by their frequencies and 18O isotopic shifts,
which closely matched those reported for the Kitajima O2
adducts (Table 2). The isotopic shifts observed for these
vibrations, from 889 to 842 cm-1 and from 424 to 408 cm-1
,
are in accord with the isotopic shifts predicted for a
diatomic oscillator (838 and 405 cm-1, respectively),
supporting the assignment of the visible absorption band
as a peroxo-to-iron(III) charge-transfer transition (Figure 5).
While there are many examples of (μ-η1:η1-peroxo)diiron(III)
(57) Girerd, J.-J.; Banse, F.; Simaan, A. J. Struct. Bonding (Berlin) 2000,
97, 145–177.
(58) Fiedler, A. T.; Shan, X.; Mehn, M. P.; Kaizer, J.; Torelli, S.; Frisch,
J. R.; Kodera, M.; Que, L., Jr. J. Phys. Chem. A 2008, 112, 13037–13044.

Article
Inorganic Chemistry, Vol. 49, No. 8, 2010 3625
Table 2. Comparison of the Spectroscopic Properties of the Peroxo Intermediate Derived from 1 with Similar Complexes (18O Isotope Shifts Indicated in Parentheses)
δ, mm/s (ΔEQ, mm/s)
λ
max, nm (ε, M-1 cm-1
)
ν(O-O), cm-1
ν(Fe-O), cm-1
peroxo intermediate (obtained from 1 þ O2)
[FeIII2(μ-1,2-O2)(μ-O2CPh)2(TpiPr2)2]a
[FeIII2(μ-1,2-O2)(μ-O2CCH2Ph)2(TpiPr2)2]b
[FeIII2(μ-O)(μ-1,2-O2)(6-Me2BPP)]2þc
[FeIII2(μ-1,2-O2)(N-Et-HPTB)(OPPh3)2]2þd
[FeIII2(μ-O)(μ-1,2-O2)(6-Me-BQPA)]2þd
[FeIII2(μ-O)(μ-1,2-O2)(BQPA)]2þd
0.65 (1.32)
680 (2500)
682 (3450)
694 (2650)
644 (3000)
588 (1500)
not reported
620 (1000)
889 (-47)
876 (-48)
888 (-46)
847 (-33)
900 (-50)
853 (-45)
844 (-44)
424 (-16)
421 (-12)
415 (-11)
465 (-19)
471 (-16)
463 (-15)
464 (-17)
0.66 (1.40)
0.50 (1.31)
0.51 (0.80)
a Reference 43. b Reference 34. c Reference 62; 6-Me2BPP = N,N-bis(6-methyl-2-pyridylmethyl)-3-aminopropionate. d Reference 58; N-Et-HPTB =
anion of N,N,N0,N0-tetrakis(2-benzimidazolylmethyl)-2-hydroxy-1,3-diaminopropane; BQPA = bis(2-quinolylmethyl)-2-pyridylmethylamine.
Figure 8. Oxygenation of 1 mM [Fe(TpiPr2)PRV] (black line) at -40 ꢀC
in MeCN in the presence of 1 equiv of TBP-H in a 1-cm cuvette. The red
lines show the growth of TBP• features.
66 cm-1 determined from variable-temperature magnetic
susceptibility measurements on the O2 adduct of
[FeII(TpiPr2)(O2CPh)].43 Our studies thus formulate the
green chromophore as having an antiferromagnetically
coupled high-spin (μ-1,2-peroxo)diiron(III) unit, as
found for other O2 adducts of [FeII(TpiPr2)(O2CR)]
(Table 2).
The formation of the peroxo intermediate from 1
at -40 ꢀC was not affected by the presence of 100 mM
THT either before or after oxygenation. This suggests
that neither the peroxo intermediate itself nor any of its
possible precursors is capable of oxidizing THT. This lack
of reactivity is very similar to that observed by Lippard
and co-workers in their studies of (μ-1,2-peroxo)diiron-
(III) intermediates.63,64
However, the formation of the peroxo intermediate was
prevented by the presence of 2,4,6-tri-tert-butylphenol
(TBP-H). When the oxygenation of 1 (1 mM) was per-
formed in the presence of even only 1 equiv of TBP-H, the
green chromophore associated with the peroxo inter-
mediate did not form at all. Instead, sharp peaks char-
acteristic of TBP• were observed at 373 and 400 nm along
with a weak broad band at 620 nm (Figure 8), with
intensities corresponding to an 80% yield of the phenoxyl
radical. Under these conditions, phenoxyl radical forma-
tion occurred within 3 min, with an estimated first-order
€
Figure 7. Mossbauer spectra of the green solution obtained from the
oxygenation of 1 in MeCN at -40 ꢀC. (A) Spectrum recorded for B = 0.
The solid line outlines the contribution of the peroxo intermediate (45%
of Fe). (B and C) Spectra of the peroxo intermediate recorded in a parallel
field of B = 8.0 T at the temperatures indicated; features arising from the
iron(II) and mononuclear iron(III) species were removed as described in
the Supporting Information. Solid lines in parts B and C are spectral
simulations for J = 70 cm-1. Hyperfine parameters are A0/gnβn = -21 T,
ΔEQ = þ1.32 mm/s, η = 1, and δ = 0.65 mm/s for both sites; the
Hamiltonian is given in the Supporting Information.
from the raw data the contributions of the iron(II) and S =
5/2 iron(III) species (see the Supporting Information). The
solid line in Figure 7B is a spectral simulation based on the
assumption that the peroxo intermediate has a diamag-
netic ground state. The magnetic splitting at 100 K
(Figure 7C) is smaller than that at 4.2 K, a feature
observed for antiferromagnetically coupled dinuclear
iron(III) peroxo complexes with J values of around
70 cm-1 (in the H = JS1 S2 convention; S1 = S2 =
3
5/2). The observed diamagnetism in conjunction with the
observed exchange interactions demonstrates that the
green chromophore is a diiron(III) complex. Our spectral
simulation, performed with the 2Spin option of WMOSS,
yielded J = 70 ( 10 cm-1; details of the Mossbauer
€
method for determining J can be found in several papers.59-61
This J value is in good agreement with the value of
€
(59) Kauffmann, K.; Munck, E. In Spectroscopic Methods in Bioinorganic
(62) Zhang, X.; Furutachi, H.; Fujinami, S.; Nagatomo, S.; Maeda, Y.;
Watanabe, Y.; Kitagawa, T.; Suzuki, M. J. Am. Chem. Soc. 2005, 127, 826–
827.
(63) LeCloux, D. D.; Barrios, A. M.; Mizoguchi, T. J.; Lippard, S. J.
J. Am. Chem. Soc. 1998, 120, 9001–9014.
(64) LeCloux, D. D.; Barrios, A. M.; Lippard, S. J. Bioorg. Med. Chem.
1999, 7, 763–772.
Chemistry; Solomon, E. I., Hodgson, K. O., Eds.; American Chemical Society:
Washington, DC, 1998; pp 16-29.
(60) Krebs, C.; Bollinger, J. M., Jr.; Theil, E. C.; Huynh, B. H. J. Biol.
Inorg. Chem. 2002, 7, 863–869.
(61) Frisch, J. R.; Vu, V. V.; Martinho, M.; Munck, E.; Que, L., Jr. Inorg.
Chem. 2009, 48, 8325–8336.
€

3626 Inorganic Chemistry, Vol. 49, No. 8, 2010
Mukherjee et al.
Scheme 4. Reactions of [FeII(TpiPr2)X] Complexes with O2
rate constant of 2.1(4) ꢀ 10-2 s-1. This value did not
change significantly with the use of more equivalents of
TBP-H, suggesting that phenoxyl radical formation oc-
curs after the rate-determining step that generates the
oxidant.
Control experiments showed that the preformed
(μ-peroxo)diiron(III) species (1 mM in Fe) also decayed
in the presence of TBP-H at -40 ꢀC but with complex
kinetic behavior. Furthermore, the time scale for decay
was nearly 2 orders of magnitude longer (∼70-fold). This
much slower reaction cannot rationalize the effect of
TBP-H in completely preventing peroxo intermediate
formation. Instead, TBP-H must react more rapidly with
a precursor of the peroxo intermediate to prevent its
accumulation. We propose that this more reactive inter-
mediate is the initial O2 adduct of 1, the formation of
which represents the rate-determining step. Following
oxygenation of the iron(II) precursor in toluene but not in
MeCN.43 Kitajima and co-workers conjectured that the
likely coordination of MeCN to the metal may inhibit O2
binding to the coordinately saturated iron(II) center.
Clearly, the solvent MeCN does not play such an inhibi-
tory role in the case of 1 because a good yield of the
peroxo intermediate was observed in MeCN. In fact, only
a 10% yield of the peroxo intermediate was obtained for
the oxygenation of 1 at -40 ꢀC in toluene. The fact that
TBP-H fails to inhibit peroxo intermediate formation in
the case of 3 suggests that the lifetime of the putative
iron(III) superoxo intermediate is too short to allow its
interception by TBP-H. The subsequent reaction with the
residual iron(II) precursor must be strongly favored over
the reaction with TBP-H. Clearly, further investigation of
the reaction of 3 with O2 and its differences with 1 is
warranted.
heme precedents, the 1 O2 adduct can be formulated as
Mechanistic Framework. We have considered the accu-
mulated observations described above and postulate the
mechanistic framework illustrated in Scheme 4 to ratio-
nalize the different outcomes. We propose an iron(III)
superoxo species as the common intermediate that is
formed initially in the oxygenation of 1 and 2. This species
has a number of possible fates (Scheme 4). The bound
superoxide can act as a nucleophile and attack the
electrophilic keto carbon of the bidentate R-keto acid
to initiate oxidative decarboxylation, according to the
generally accepted mechanism for 2-OG-dependent iron
enzymes.6 Alternatively, the bound superoxide can accept
an electron from the residual iron(II) precursor to form
the diiron(III) peroxo intermediate. Third, the bound
3
an iron(III) superoxo species, which can abstract a hydro-
gen atom from phenol. Such hydrogen-atom abstractions
have been observed for other metal superoxo species.35,39
Interestingly, the oxygenation of [FeII(TpiPr2)(O2CPh)]
in the presence of as much as 100-fold excess of TBP-H
did not prevent the formation of the corresponding
peroxo intermediate, which suggests that TBP-H is un-
able to intercept the corresponding iron(III) superoxo
species in this case before it reacts with a residual iron(II)
precursor. However, there are some differences in the
conditions under which the respective peroxo intermedi-
ates of 1 and 3 can be observed. In the latter case,
formation of the peroxo intermediate was observed upon

Article
Inorganic Chemistry, Vol. 49, No. 8, 2010 3627
superoxide can abstract a hydrogen atom from added
phenol. All three outcomes have been observed in the case
of 1, depending on the reaction conditions.
become reduced by another equivalent of the iron(II)
complex to form a diiron(III) peroxo intermediate, or (C)
abstract a hydrogen atom from phenol. On the other
hand, 2 reacts with O2 and solely follows pathway A,
while 3 follows only pathway B because of the differences
in the electronic properties of the ancillary ligand. Iron-
(III) superoxo species have also been implicated recently
in the reactions of FeII(TMC) and FeII(N4Py) [TMC =
1,4,8,11-tetramethylcyclam; N4Py = N,N-bis(2-pyridyl-
methyl)bis(2-pyridyl)methylamine] with O2 in the pre-
sence of protons and reductants to form oxoiron(IV)
and hydroperoxoiron(III) species, respectively.63,64 These
reactions demonstrate the versatility of the iron(III)
superoxo moiety and support the mechanistic diversity
illustrated in Scheme 2 for the various FeO2 adducts that
must form in the catalytic cycles of nonheme iron oxi-
dases and oxygenases.
These three distinct reactions presumably have differ-
ent activation barriers. For 1, oxidative decarboxylation
is strongly favored at 25 ꢀC, and the putative iron(IV) oxo
oxidant produced therefrom is able to carry out intramo-
lecular attack of an isopropyl C-H bond on the Tp ligand
or intermolecular O-atom transfer to THT. In contrast,
oxidative decarboxylation does not readily occur at -40 ꢀC;
instead, a green chromophore identified to be a diiron-
(III) peroxo intermediate forms, analogous to that pre-
viously reported in the oxygenation of 3. However,
peroxo formation can be prevented by the presence of
the phenol TBP-H, which is oxidized to TBP•. This
interception suggests the involvement of a superoxo
species that is a precursor of the peroxo intermediate
but can also abstract a hydrogen atom from TBP-H.
The reaction energetics are clearly different for 2. In
this case, oxidative decarboxylation is favored at both
þ25 and -40 ꢀC and the corresponding peroxo inter-
mediate is not observed to form at the lower temperature.
This difference in reactivity from 1 suggests that the
intramolecular attack of the superoxo moiety on the
bound BF in 2 is much more facile at -40 ꢀC than
the corresponding attack on the bound PRV in 1, pre-
sumably resulting from the increased electrophilicity of
the BF carbonyl carbon because of the more electron-
withdrawing phenyl group. As a result, the superoxo
intermediate derived from 2 is too short-lived to allow
alternative intermolecular reaction pathways to occur.
Another factor to consider may be the relative basicities
of the carboxylate ligands used in this study, as reflected
by their aqueous pKa values. They increase in the order
benzoylformic acid (1.39),65 pyruvic acid (1.94),66 and
benzoic acid (4.19).67 The differing basicities of the bound
carboxylate ligands will modulate the reducing power of
the respective iron(II) centers and affect the equilibria
that govern the initial O2 binding step and the subsequent
reaction of the O2 adduct with a second molecule of
[FeII(TpiPr2)(O2CR)] to form the peroxo intermediate.
Thus, formation of the peroxo intermediate should be
most favorable for 3. Indeed, unlike in the oxygenation of
1 at -40 ꢀC, we found that the oxygenation of 3 cannot be
prevented by the presence of TBP-H. The fact that the
latter reaction occurs only in a toluene solvent introduces
another mechanistic variable to consider, and details of
the oxygenation mechanisms of 1 and 3 may differ.
Clearly, a more in-depth comparison of the oxygenation
chemistry of 1 and 3 will be informative.
Experimental Section
General Procedures. All reagents and solvents were purchased
from commercial sources and were used without further pur-
ification unless otherwise noted. Methanol was rigorously dried
by distillation from Mg(OMe)2 and degassed under N2 prior to
use. Anhydrous dichloromethane and acetonitrile were pur-
chased from Aldrich. The preparation and handling of air-
sensitive materials were carried out under an inert atmosphere
(N2) in a glovebox. Caution! Perchlorate salts are potentally
explosive and should be handled with care.
Synthesis of 1. To a mixture of ligand (0.126 g, 0.25 mmol) and
Fe(ClO4)2 (0.10 g, 0.25 mmol) in methanol (3 mL) was added
sodium pyruvate (0.029 g 0.25 mmol), resulting in an immediate
color change of the solution from colorless to red. This cloudy
solution was stirred for 1 h at room temperature. The solid was
then filtered and dried in vacuo. X-ray-quality crystals were
grown via the slow evaporation of a saturated CH2Cl2 solution
at -20 ꢀC. Anal. Calcd for 1, C31H53BFeN6O4 3CH2Cl2 (895.3 g/
3
mol): C, 45.61; H, 6.64; N, 9.39. Found: C, 45.54; H, 6.70; N, 9.30.
UV-vis [λmax, nm (ε, M-1 cm-1) in MeCN]: 458 (sh, 460), 500
(510), 540 (sh, 390). 1H NMR (benzene-d6, 300 MHz, 25 ꢀC): -7.8
(s, BH), -2.8 (s, CHMe2) 3.7 (s, CHMe2), 14.4 (s, CHMe2), 8.04 (s,
CHMe2), 63.7 (s, pz-4-H), 114.8 (s, O2CC(O)CH3).
Synthesis of 2. To a mixture of ligand (0.126 g, 0.25 mmol) and
Fe(OTf)2 2MeCN65 (0.10 g, 0.25 mmol) in methylene chloride
3
(3 mL) was added solid sodium benzoylformate (0.043 g 0.25 mmol).
The solution turned blue immediately and was stirred for 1 h
at room temperature. The solid was filtered and dried in vacuo.
X-ray-quality crystals were grown via the slow evaporation of a
saturated pentane solution at -20 ꢀC. Anal. Calcd for 2,
C35H51BFeN6O3 (670.47 g/mol): C, 62.70; H, 7.67; N, 12.53.
Found: C, 62.92; H, 7.83; N, 12.47. UV-vis [λmax, nm (ε, M-1
cm-1) in MeCN]: 553 (sh, 510), 610 (610). 1H NMR (benzene-d6,
300 MHz, 25 ꢀC): -8.1 (s, BH), -1.78 (s, CHMe2) 3.8
(s, CHMe2), 13.2 (s, CHMe2), 20.3 (o-BF), 16.7 (m-BF), 63.8
(s, pz-4-H).
In conclusion, we have described a set of three FeII-
(TpiPr2) complexes 1-3, where the ancillary carboxylate
or R-keto acid ligand significantly modulates the outcome
of the reaction. In all three cases, the initial formation of a
common iron(III) superoxo intermediate is inferred. Un-
der the appropriate conditions, this species derived from 1
can (A) act as a nucleophile to initiate the oxidative
decarboxylation of the bound R-ketocarboxylate, (B)
Identification of the Ligand Hydroxylation Product by NMR
Spectroscopy. 1 (5 mM) in 2 mL of MeCN was bubbled with O2
using a balloon, and the solution was stirred vigorously for 15 min.
To this solution was added ca. 2 mL of 1 M HCl with vigorous
stirring. To the above solution was added excess NH3(aq),
and organic products were extracted with diethyl ether. The
extract was dried over MgSO4, and to this solution was added
1,4-dimethoxybenzene as an internal standard. Evaporation of
the solvent resulted in colorless residue soluble in CDCl3 that
was analyzed by 1H NMR spectroscopy.
(65) Wheatley, M. S. Experientia 1956, 12, 339–340.
(66) Chiang, Y.; Kresge, A. J.; Pruszynski, P. J. Am. Chem. Soc. 1992, 114,
3103–3107.
(67) Lide, D. R. CRC Handbook of Chemistry and Physics, 73rd ed.; CRC
Press: Boca Raton, FL, 1992; pp 8-39.
Identification of the Acid Decarboxylation Product by NMR
Spectroscopy. 2 (5 mM) in 2 mL MeCN was bubbled with O2

3628 Inorganic Chemistry, Vol. 49, No. 8, 2010
Mukherjee et al.
using a balloon, and the solution was stirred vigorously for 15 min.
To this solution was added ca. 2 mL of 3 M HCl with vigorous
stirring. The organic layer was extracted with diethyl ether and
washed with distilled water. The extract was then dried over
MgSO4, and to this solution was added 1,4-dimethoxybenzene
as an internal standard. Evaporation of the solvent resulted in a
faintly colored residue soluble in CDCl3 that was analyzed by 1H
NMR spectroscopy.
Physical Methods. Elemental analyses were performed by
Atlantic Microlab, Inc. UV-vis spectra were recorded on a
HP 8452A diode-array spectrometer with samples maintained at
low temperature using a cryostat from Unisoku Scientific
Instruments, Osaka, Japan. ESI-MS studies were performed
by using s Bruker BioTOF mass spectrometer, and the data
processing was done by using Bruker Daltonics software. NMR
spectra were recorded on a Varian VXR-300 spectrometer at
room temperature. rR spectra were collected using an excitation
wavelength of 647.1 nm from a Spectra-Physics model 2060
krypton-ion laser and an Acton AM-506 monochromator
equipped with a Princeton LN/CCD data collection system.
Low-temperature spectra in CH3CN were obtained at 77 K
using a 135ꢀ backscattering geometry. Samples were frozen onto
a gold-plated copper coldfinger in thermal contact with a Dewar
flask containing liquid N2. Raman frequencies were calibrated
to indene prior to data collection. The monochromator slit
width was set for a band pass of 4 cm-1 for all spectra. rR
spectra were intensity corrected to the 773 cm-1 solvent peak.
Crystallography of 1 and 2. For both complexes, a crystal
(approximate dimensions 0.40 ꢀ 0.35 ꢀ 0.30 mm3) was placed
on the tip of a 0.1-mm-diameter glass capillary and mounted on
a CCD area detector diffractometer for data collection at 173(2)
K.66 A preliminary set of cell constants was calculated from
reflections harvested from three sets of 20 frames for 1 and four
sets of 30 frames for 2. These initial sets of frames were oriented
such that orthogonal wedges of reciprocal space were surveyed.
This produced initial orientation matrixes determined from
101 reflections for 1 and 324 reflections for 2. The data collection
was carried out using Mo KR radiation (graphite mono-
chromator) with a time frame of 20 s for 1 and 45 s for 2 and
a detector distance of 4.9 cm. A randomly oriented region of
reciprocal space was surveyed to the extent of one sphere and
€
Mossbauer spectra were recorded with two spectrometers, using
Janis Research Super-Varitemp Dewars that allowed studies in
applied magnetic fields up to 8.0 T in the temperature range
˚
˚
€
from 1.5 to 200 K. Mossbauer spectral simulations were per-
to a resolution of 0.82 A for 1 (0.80 A for 2). Four major sections
of frames were collected with 0.30ꢀ steps in ω at four different
φ settings and a detector position of -28ꢀ in 2θ. The intensity
data were corrected for absorption and decay (SADABS).67
Final cell constants were calculated from 2801 (for 1) and 2982
(for 2) strong reflections from the actual data collection
after integration (SAINT). Please refer to Table S1 in the
Supporting Information for additional crystal and refinement
information.
formed using the WMOSS software package (WEB Research,
Edina, MN). Isomer shifts are quoted relative to iron metal at
298 K.
Acknowledgment. This work was supported by the National
Institutes of Health (Grant GM-33162 to L.Q., Grant
EB-001475 to E.M., and postdoctoral fellowship ES017390-01
to M.A.C.) and by the Kurata Foundation (to K.F.). A.M. is
grateful for a dissertation fellowship from the University of
Minnesota Graduate School. We thank Dr. Victor G. Young,
Jr., and the X-ray Crystallographic Laboratory of University of
Minnesota for crystallographic assistance.
The structures were solved and refined using SHELXTL. The
space groups of P1 (for 1) and P2/c (for 2) were determined
based on systematic absences and intensity statistics. A direct
methods solution was calculated and provided most non-hydrogen
atoms from the E map. Full-matrix least-squares/difference
Fourier cycles that located the remaining non-hydrogen atoms
were performed. All non-hydrogen atoms were refined with
anisotropic displacement parameters. All hydrogen atoms were
placed in ideal positions and refined as riding atoms with relative
isotropic displacement parameters. The final full-matrix least-
squares refinement converged to R1 = 0.0473 for 1 and 0.0416
for 2 and wR2 = 0.1358 for 1 and 0.1100 for 2 (F2, all data).
Supporting Information Available: Crystallographic data in
CIF format for 1 and 2, a table comparing their parameters,
ESI-MS data along with simulations, 1H NMR spectra showing
€
the decarboxylation of benzoylformate, and Mossbauer spectra
of the monomeric high-spin FeIII species observed in the oxyge-
nation of 1 at -40 ꢀC. This material is available free of charge via
the Internet at http://pubs.acs.org.