Modeling the diiron centers of non-heme iron enzymes. Preparation of sterically hindered diiron(II) tetracarboxylate complexes and their reactions with dioxygen

Article abstract of DOI:10.1021/ja981216s

A series of diiron(II) complexes, [Fe₂(μ-L)(μ-O₂CR)(O₂CR)(N)₂], where L is a dinucleating bis(carboxylate) ligand based on m-xylylenediamine bis(Kemp's triacid imide) and N is a pyridine- or imidazole-derived ligand, were prepared as models for the carboxylate-bridged non-heme diiron cores of the O₂-activating enzymes, soluble methane monooxygenase hydroxylase (MMOH), and the R2 subunit of ribonucleotide reductase (RNR-R2). X-ray crystallographic studies revealed differences in the coordination geometry of the bridging monocarboxylate ligand, which shifts from monodentate to syn,syn-bidentate bonding modes. The extent of this carboxylate shift depends on both the steric bulk of the monocarboxylate and the basicity of the ancillary N-donor ligands. Exposure of these diiron(II) complexes to O₂ at - 77 °C in nonpolar solvents (CH₂Cl₂, THF, toluene) yielded deep blue solutions (λ(max) ? 580 nm, ε ? 1200 M^-1 cm^-1), consistent with the generation of diiron(III) peroxo species. This reaction was irreversible, and its stoichiometry was determined by manometry to be 1:1 in diiron(II) complex and O₂. The diiron(III) peroxo complexes exhibited oxygen isotope-sensitive resonance Raman bands at ~ 860 cm^-1, which are assigned to the O-O stretching frequency of a μ-1,2-peroxodiiron(III) core. ⁵⁷Fe Mossbauer spectroscopy confirmed the assignment of the diiron(III) oxidation level and indicated that the two iron sites have inequivalent environments (δ₁ ? 0.47 mm s^-1, ΔE(Q1) ? 0.88 mm s^-1; δ₂ ? 0.63 mm s^-1, ΔE(Q2) ? 1.20 mm s^-1). Kinetics experiments provided rate constants for the reaction and revealed it to be first order in both diiron(II) complex and O₂. The factors controlling the rate of formation of the blue species and its stability are discussed.

Full text of DOI:10.1021/ja981216s

J. Am. Chem. Soc. 1998, 120, 9001-9014
9001
Modeling the Diiron Centers of Non-Heme Iron Enzymes.
Preparation of Sterically Hindered Diiron(II) Tetracarboxylate
Complexes and Their Reactions with Dioxygen
Daniel D. LeCloux, Amy M. Barrios, Tadashi J. Mizoguchi, and Stephen J. Lippard*
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
ReceiVed April 10, 1998
Abstract: A series of diiron(II) complexes, [Fe2(µ-L)(µ-O2CR)(O2CR)(N)2], where L is a dinucleating bis-
(carboxylate) ligand based on m-xylylenediamine bis(Kemp’s triacid imide) and N is a pyridine- or imidazole-
derived ligand, were prepared as models for the carboxylate-bridged non-heme diiron cores of the O2-activating
enzymes, soluble methane monooxygenase hydroxylase (MMOH), and the R2 subunit of ribonucleotide reductase
(RNR-R2). X-ray crystallographic studies revealed differences in the coordination geometry of the bridging
monocarboxylate ligand, which shifts from monodentate to syn,syn-bidentate bonding modes. The extent of
this carboxylate shift depends on both the steric bulk of the monocarboxylate and the basicity of the ancillary
N-donor ligands. Exposure of these diiron(II) complexes to O2 at -77 °C in nonpolar solvents (CH2Cl2,
THF, toluene) yielded deep blue solutions (λmax ≈ 580 nm, ꢀ ≈ 1200 M-1 cm-1), consistent with the generation
of diiron(III) peroxo species. This reaction was irreversible, and its stoichiometry was determined by manometry
to be 1:1 in diiron(II) complex and O2. The diiron(III) peroxo complexes exhibited oxygen isotope-sensitive
resonance Raman bands at ∼860 cm-1, which are assigned to the O-O stretching frequency of a µ-1,2-
peroxodiiron(III) core. 57Fe Mo¨ssbauer spectroscopy confirmed the assignment of the diiron(III) oxidation
level and indicated that the two iron sites have inequivalent environments (δ1 ≈ 0.47 mm s-1, ∆EQ1 ≈ 0.88
mm s-1; δ2 ≈ 0.63 mm s-1, ∆EQ2 ≈ 1.20 mm s-1). Kinetics experiments provided rate constants for the
reaction and revealed it to be first order in both diiron(II) complex and O2. The factors controlling the rate
of formation of the blue species and its stability are discussed.
Introduction
O2 is activated such that one of its oxidizing equivalents
generates a tyrosyl radical. This radical accepts an electron from
the R1 subunit, the site of ribonucleotide reduction. Methane
monooxygenase (MMO) is used by obligate methanotrophic
bacteria to convert methane selectively to methanol. The soluble
form (sMMO) contains a carboxylate-bridged diiron active site
in the hydroxylase component (MMOH). This mixed-function
oxidase functions such that one of the oxygen atoms from
dioxygen is incorporated into methanol and the other into water
(eq 2). Figure 1 depicts the active-site structures of the reduced,
diiron(II) forms of RNR-R2, and MMOH, as deduced from
Carboxylate-bridged diiron proteins use dioxygen for func-
tions ranging from O2 transport to the selective oxidation of
organic substrates.1-5 This functional diversity is illustrated by
three of the most intensively studied proteins in this family,
hemerythrin, ribonucleotide reductase, and methane monooxy-
genase. Hemerythrin (Hr) is a dioxygen carrier in certain marine
invertebrates. The reduced protein contains a [Fe2(µ-OH)(µ-
O2CR)2(His)5]+ core and binds O2 in the form of a terminal
hydroperoxide ligand that forms a hydrogen bond to the oxo
bridge of the resulting (µ-oxo)diiron(III) unit (eq 1). Ribonucle-
CH4 + O2 + NADH + H+ f CH3OH + H2O + NAD+
(2)
X-ray crystal structural analyses.6-8
A detailed mechanistic picture of O2 activation in RNR-
R22,9,10 and sMMO2,4,9,10 is emerging as a consequence of kinetic
and spectroscopic studies which have identified transient
intermediates in their reaction cycles. For RNR-R2, the first
detectable intermediate forms and decays within 35 ms at 4 °C.11
On the basis of the similarity of its Mo¨ssbauer and UV-vis
spectra to those observed for a model complex and an analogous
(1)
otide reductase from Escherichia coli houses a carboxylate-
bridged diiron active site in the R2 subunit (RNR-R2), where
(1) Feig, A. L.; Lippard, S. J. Chem. ReV. 1994, 94, 759.
(2) Wallar, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625.
(3) Liu, K. E.; Lippard, S. J. In AdVances in Inorganic Chemistry; Sykes,
A. G., Ed.; Academic Press: San Diego, CA, 1995; Vol. 42, p 263.
(4) Valentine, A. M.; Lippard, S. J. J. Chem. Soc., Dalton Trans. 1997,
3925.
(6) Rosenzweig, A. C.; Brandstetter, H.; Whittington, D. A.; Nordlund,
P.; Lippard, S. J.; Frederick, C. A. Proteins 1997, 29, 141.
(7) Rosenzweig, A. C.; Nordlund, P.; Takahara, P. M.; Frederick, C.
A.; Lippard, S. J. Chem. Biol. 1995, 2, 409.
(8) Nordlund, P.; Eklund, H. Curr. Opin. Struct. Biol. 1995, 5, 758.
(9) Kurtz, D. M., Jr. JBIC 1997, 2, 159.
(5) Que, L., Jr. J. Chem. Soc., Dalton Trans. 1997, 3933.
(10) Edmondson, D. E.; Huynh, B. H. Inorg. Chim. Acta 1996, 252, 399.
S0002-7863(98)01216-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/21/1998

9002 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
include three crystallographically characterized diiron(III) per-
oxo complexes and a bis(µ-oxo)FeIIIFeIV species, the properties
of which are summarized in Table 1.18-22 The peroxo com-
plexes display relatively intense electronic transitions centered
at ∼600 nm (ꢀ ≈ 2000 M-1 cm-1) and O-O stretching
frequencies at ∼900 cm-1. X-ray structural studies revealed
the peroxide ligand to be coordinated symmetrically, in either
an eclipsed19,20 or a gauche18 µ-1,2 bridging mode.
Despite these advances, a model system has yet to be
described where a peroxo is derived from the reaction of a
diiron(II) precursor with O2, and which either converts to a high-
valent Q- or X-like species or performs hydrocarbon oxidation
chemistry. These shortcomings may stem from the fact that
the previously reported model systems do not reproduce the
ligand composition of the RNR-R2 and sMMO active sites; the
relatively soft, nitrogen-rich coordination environments of the
model complexes more closely resemble the active site of Hr.
The widely different fates of the peroxo species generated in
Hr, RNR-R2, and sMMO may be due in part to the ability of
the latter two systems, which have harder, more oxygen-rich
donor sets, to support the higher iron oxidation states required
for O-O bond cleavage. We were therefore interested in
investigating more carboxylate-rich synthetic models to deter-
mine whether they might more closely mimic the O2 binding,
O-O bond cleavage, and substrate oxidation properties of RNR
and sMMO.
Figure 1. Representations of the active-site structures of (A) reduced
RNR-R2 and (B) reduced MMOH (Hred).
intermediate in MMOH (vide infra), this species has been
tentatively assigned as a diiron(III) peroxo unit. Intermediate
X, which evolves subsequently, exhibits Mo¨ssbauer,12 Q-band
ENDOR,13 and EXAFS14 features consistent with an antiferro-
magnetically coupled FeIIIFeIV center having at least one oxo
bridge and a terminal water or hydroxide ligand. This high-
valent intermediate generates the active enzyme formally by
hydrogen atom abstraction from Tyr122 to afford an oxo-bridged
diiron(III) center and the tyrosyl radical. Intermediate X
presumably is the product of O-O bond cleavage, but it is
uncertain whether injection of an extra electron must accompany
O-O bond scission or whether an as yet unobserved dioxodi-
iron(IV) species might initially be generated.10
In pursuit of this goal, carboxylate-bridged diiron(II) com-
plexes employing the m-xylylenediamine bis(Kemp’s triacid
imide) (H2XDK) and related ligand systems (1-3) were
For MMOH, a metastable diiron(III) peroxo intermediate
(Hperoxo) has also been spectroscopically identified as an early
product of the O2 reaction chemistry,15 but it forms and decays
more slowly than its RNR-R2 analogue.11 Hperoxo converts to
a diiron(IV) species termed Q.15,16 This intermediate, or a
closely related derivative, is believed to oxidize the substrate,
since Q is the last species detected prior to methanol formation
and its decomposition is accelerated by the presence of
substrates. EXAFS studies of Q revealed parameters quite
similar to those of intermediate X in RNR-R2,17 but the data
do not provide unique structural information, particularly with
respect to coordination of oxygen ligands derived from O2.
prepared and their reactivity toward dioxygen explored.23
A
family of tris(µ-carboxylato)diiron(II) complexes with a coor-
dination environment analogous to that of the MMOH active
site has been obtained. When one such derivative was exposed
to O2 at low temperature in THF, the solution flashed deep
purple, but quickly decayed to a red-orange color in less than
15 s (eq 3). Monitoring this reaction by stopped-flow UV-vis
Considerable effort has been expended to prepare synthetic
models for the dioxygen intermediates described above for
RNR-R2 and sMMO. The goals have been to mimic their
spectroscopy, structural features, and ultimately their oxidation
chemistry.1,5 The most noteworthy systems reported to date
(11) Burdi, D.; Riggs-Gelasco, P. J.; Tong, W.; Willems, J.-P.; Lee, H.-
I.; Doan, P. E.; Sturgeon, B.; Shu, L.; Chen, S.; Edmonson, D. E.; Huynh,
B. H.; Que, L., Jr.; Hoffman, B. M.; Stubbe, J. Steenbock Symp. Proc.:
Biosynthesis and Function of Metal Clusters For Enzymes 1997, 25, 85.
(12) Sturgeon, B. E.; Burdi, D.; Chen, S.; Huynh, B. H.; Edmondson,
D. E.; Stubbe, J.; Hoffman, B. M. J. Am. Chem. Soc. 1996, 118, 7551.
(13) Willems, J.-P.; Lee, H.-I.; Burdi, D.; Doan, P. E.; Stubbe, J.;
Hoffman, B. M. J. Am. Chem. Soc. 1997, 119, 9816.
(14) Riggs-Gelasco, P. J.; Shu, L.; Chen, S.; Burdi, D.; Huynh, B. H.;
Que, L., Jr.; Stubbe, J. J. Am. Chem. Soc. 1998, 120, 849.
(15) Liu, K. E.; Valentine, A. M.; Wang, D.; Huynh, B. H.; Edmondson,
D. E.; Salifoglou, A.; Lippard, S. J. J. Am. Chem. Soc. 1995, 117, 10174.
(16) Lee, S.-K.; Nesheim, J. C.; Lipscomb, J. D. J. Biol. Chem. 1993,
268, 21569.
spectroscopy revealed the buildup and decay over 10 s of an
intermediate with an optical band centered at 650 nm. This
(18) Kim, K.; Lippard, S. J. J. Am. Chem. Soc. 1996, 118, 4914.
(19) 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.
(20) Dong, Y.; Yan, S.; Young, V. G., Jr.; Que, L., Jr. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 618.
(21) Dong, Y.; Zang, Y.; Shu, L.; Wilkinson, E. C.; Que, L., Jr.;
Kauffmann, K.; Mu¨nck, E. J. Am. Chem. Soc. 1997, 119, 12683.
(22) Que, L., Jr.; Dong, Y. Acc. Chem. Res. 1996, 29, 190.
(23) Herold, S.; Lippard, S. J. J. Am. Chem. Soc. 1997, 119, 145.
(17) Shu, L.; Nesheim, J. C.; Kauffmann, K.; Mu¨nck, E.; Lipscomb, J.
D.; Que, L., Jr. Science 1997, 275, 515.

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9003
Table 1. Summary of Data Acquired for Selected RNR-R2 and sMMO Model Complexes
a Not reported. b Not detected.
CPhCy),29 and [FeTp3,5-iPr (O2CCH2Ph)]18 were prepared according to
literature procedures. The thallium(I) salts of the carboxylic acid
ligands were obtained as described, or by analogous methods,28,30 and
characterized by 1H NMR spectroscopy. The ligands 1-methylimidazole
(Me-Im) and 1-butylimidazole (Bu-Im) were degassed and stored over
3-Å molecular sieves for several days prior to use. Dioxygen (99.994%,
BOC Gases) was dried by passing the gas stream through a column of
Drierite. All other reagents were procured from commercial sources
and used as received unless otherwise noted. Benzene, toluene, pentane,
THF, and Et2O were distilled from sodium benzophenone ketyl under
nitrogen. Dichloromethane, 1,2-dichloroethane, and pyridine were
distilled from CaH2 under nitrogen; CD2Cl2 was degassed, passed
through activated basic alumina, and stored over 3-Å molecular sieves
prior to use. All air-sensitive manipulations were carried out either in
a nitrogen-filled Vacuum Atmospheres drybox or by standard Schlenk
line techniques at room temperature unless otherwise noted.
2
chromophore was tentatively assigned to a diiron(III) peroxo
species, the inherent instability of which precluded further
characterization or functional reactivity studies.
In the present work we have extended this chemistry by
synthesizing sterically more hindered ligands with the aim of
stabilizing the incipient peroxo intermediate. This strategy was
adopted following previous work from our laboratory on the
thermal decomposition of well-defined diiron(III) peroxo com-
plexes.24,25 Kinetic studies revealed the decay to have a second-
order dependence on the concentration of the peroxo species,
resulting in extrusion of 0.5 equiv of O2 and cleanly yielding
oxo-bridged diiron(III) derivatives. Since bimolecular decay
pathways are typically retarded by implementing more sterically
demanding ligands, we modified our complexes accordingly.
The synthesis and reactivity of these new diiron(II) compounds,
their reaction with O2 to form relatively more stable peroxo
adducts, the physical characterization of these species, and
studies of the dependence of their rates of formation on the
nature of both the N- and O-donor ligands form the basis for
the present report. A detailed account of the reactivity of the
peroxo complexes toward electrophilic, nucleophilic, and hy-
drocarbon substrates will be published separately.26
[Fe(XDK)(py)2] (4). To a rapidly stirred suspension of Tl2(XDK)
(2.00 g, 2.02 mmol) and pyridine (1.64 mL, 20.2 mmol) in THF (15
mL) was added a solution of FeBr2 (446 mg, 2.02 mmol) in THF (5
mL). After 6 h, the precipitated TlBr was removed by filtration through
Celite, and the filtrate was evaporated to dryness. The crude residue
was washed with pentane (20 mL) to remove any residual pyridine,
isolated by filtration, and dried fully in vacuo to afford 4 as a light
yellow powder (1.44 g, 90%). This crude material was used for
subsequent reactions without further purification. Recrystallization from
CH2Cl2/pyridine/Et2O afforded yellow blocks which allowed 4 to be
characterized by X-ray crystallography. IR (KBr) 3068, 2961, 2929,
Experimental Section
General Considerations. The compounds H2XDK (1),27 H2PXDK
(2),23 H2BXDK (3),28 1-phenylcyclohexanecarboxylic acid (HO2-
(27) Rebek, J., Jr; Marshall, L.; Wolak, R.; Parris, K.; Killoran, M.;
Askew, B.; Nemeth, D.; Islam, N. J. Am. Chem. Soc. 1985, 107, 7476.
(28) LeCloux, D. D.; Lippard, S. J. Inorg. Chem. 1997, 36, 4035.
(29) Calderon, S. N.; Newman, A. H.; Tortella, F. C. J. Med. Chem.
1991, 34, 3159.
(30) Doyle, G.; Eriksen, K. A.; Modrick, M.; Ansell, G. Organometallics
1982, 1, 1613.
(24) Feig, A. L.; Masschelein, A.; Bakac, A.; Lippard, S. J. J. Am. Chem.
Soc. 1997, 119, 334.
(25) Feig, A. L.; Becker, M.; Schindler, S.; van Eldik, R.; Lippard, S. J.
Inorg. Chem. 1996, 35, 2590.
(26) LeCloux, D. D.; Lippard, S. J. Manuscript in preparation.

9004 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
1735, 1696, 1602, 1404, 1463, 1358, 1334, 1218, 1195, 1087, 1039,
958, 835, 760, 626 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1))
364 (580). Anal. Calcd for C42H48N4O8Fe: C, 63.64; H, 6.10; N, 7.07.
Found: C, 61.33; H, 6.26; N, 6.60. Since an X-ray structure
determination revealed 4 to be the desired material, the disagreement
with the expected carbon analysis was not pursued further.
[Fe(BXDK)(py)2] (5). This compound was prepared as a light
yellow powder from Tl2(BXDK) (2.33 g, 1.61 mmol), pyridine (560
µL, 6.92 mmol), and FeBr2 (354 mg, 1.61 mmol) by a procedure
analogous to that used to obtain 4 (1.55 g, 77%). The crude material
was used for all subsequent reactions. Recrystallization from CH2Cl2/
Et2O provided 5 as opaque yellow needles. IR (KBr) 3062, 3027, 2921,
2857, 1734, 1693, 1603, 1495, 1446, 1352, 1191, 1040, 935, 761, 700,
593, 526 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 361 (690).
Anal. Calcd for C78H72N4O8Fe: C, 74.99; H, 5.81; N, 4.48. Found:
C, 74.45; H, 5.80; N, 4.34.
[Fe2(µ-BXDK)(µ-O2Ct-Bu)(O2Ct-Bu)(py)2] (6). To a stirred solu-
tion of TlO2Ct-Bu (147 mg, 0.480 mmol) and pyridine (116 µL, 1.44
mmol) in THF (3 mL) was added FeBr2 (53 mg, 0.24 mmol) in THF
(1 mL). After 10 min this mixture was added in one portion to a stirred
solution of 5 (300 mg, 0.240 mmol) in CH2Cl2 (3 mL). The resulting
mixture was allowed to stir for 0.5 h, after which time the precipitated
TlBr was removed by filtration through Celite, and the volatile
components were removed in vacuo. Recrystallization from CH2Cl2/
Et2O afforded 6 as yellow blocks (265 mg, 73%) which allowed the
compound to be characterized by X-ray crystallography. IR (KBr)
3062, 3028, 2959, 2923, 2863, 1734, 1693, 1602, 1541, 1486, 1407,
1365, 1221, 1189, 1042, 937, 763, 701, 608, 528 cm-1; UV-vis (CH2-
Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 355 (920). Anal. Calcd for C88.5H91N4O12-
Fe2Cl (6‚0.5CH2Cl2): C, 68.59; H, 5.92; N, 3.61. Found: C, 68.45;
H, 6.06; N, 3.48. One lattice CH2Cl2 molecule identified by X-ray
crystallography was apparently not removed completely by pumping
on the sample in vacuo for 1 d. When this procedure was repeated
with heating, the sample decomposed.
[Fe2(µ-XDK)(µ-O2CPhCy)(O2CPhCy)(py)2] (7). This compound
was prepared from Tl(O2CPhCy) (308 mg, 0.757 mmol), pyridine (184
µL, 2.27 mmol), FeBr2 (83 mg, 0.378 mmol), and 4 (300 mg, 0.378
mmol), as described for 6. Recrystallization from benzene/pentane
afforded 7 as pale yellow blocks which were characterized by X-ray
crystallography (291 mg, 61%). IR (KBr) 3063, 2960, 2928, 2862,
1737, 1699, 1614, 1448, 1405, 1355, 1221, 1187, 1069, 957, 851, 760,
698, 637 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 354 (840).
Anal. Calcd for C68H78N4O12Fe2: C, 65.08; H, 6.26; N, 4.46. Found:
C, 64.09; H, 6.63; N, 4.08.
[Fe2(µ-BXDK)(µ-O2CCH2t-Bu)(O2CCH2t-Bu)(py)2] (11). This
compound was prepared from TlO2CCH2t-Bu (77 mg, 0.24 mmol),
pyridine (60 µL, 0.72 mmol), FeBr2 (26 mg, 0.12 mmol), and 5 (150
mg, 0.120 mmol) as described for 6. Recrystallization from THF/
pentane provided 11 as yellow blocks (130 mg, 71%). IR (KBr) 3062,
3028, 2952, 2866, 1733, 1693, 1603, 1548, 1447, 1364, 1190, 1041,
936, 744, 701, 528 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1))
358 (960). Anal. Calcd for C90H94N4O12Fe2: C, 70.40; H, 6.17; N,
3.65. Found: C, 70.20; H, 6.60; N, 3.26.
[Fe2(µ-BXDK)(µ-O2CPh)(O2CPh)(py)2] (12). This compound was
prepared from TlO2CPh (78 mg, 0.24 mmol), pyridine (60 µL, 0.72
mmol), FeBr2 (26 mg, 0.12 mmol), and 5 (150 mg, 0.120 mmol) as
described for 6. Recrystallization from THF/CH2Cl2/pentane provided
12 as yellow blocks (132 mg, 71%) which appeared suitable for X-ray
crystallography. IR (KBr) 3061, 3028, 2952, 2920, 2865, 1733, 1694,
1604, 1543, 1495, 1447, 1364, 1189, 1069, 1042, 937, 853, 762, 720,
701, 588, 528 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 351
(1400). Anal. Calcd for C93H84N4O12Fe2Cl2 (12‚CH2Cl2): C, 68.43;
H, 5.19; N, 3.43. Found: C, 68.84; H, 5.21; N, 3.13.
[Fe2(µ-BXDK)(µ-O2CPhCy)(O2CPhCy)(py)2] (13). This com-
pound was prepared from TlO2CPhCy (391 mg, 0.960 mmol), pyridine
(233 µL, 2.88 mmol), FeBr2 (106 mg, 0.480 mmol), and 5 (600 mg,
0.480 mmol) as described for 6. Recrystallization from CH2Cl2/THF/
pentane provided 13 as yellow blocks (623 mg, 76%) which appeared
suitable for X-ray crystallography. IR (KBr) 3061, 3028, 2920, 2856,
1735, 1696, 1605, 1494, 1452, 1398, 1366, 1191, 1073, 1042, 864,
762, 699, 526 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 354
(940). Anal. Calcd for C104H102N4O12Fe2: C, 72.98; H, 6.01; N, 3.27.
Found: C, 72.73; H, 6.09; N, 3.02.
[Fe2(µ-BXDK)(µ-O2CAr)(O2CAr)(py)2] (14). This compound was
prepared from TlO2CAr (108 mg, 0.240 mmol), pyridine (60 µL, 0.72
mmol), FeBr2 (26 mg, 0.12 mmol), and 5 (150 mg, 0.120 mmol) as
described for 6. Recrystallization from benzene/pentane provided 14
as yellow blocks (144 mg, 67%). IR (KBr) 3068, 3028, 2960, 2928,
2868, 1733, 1693, 1605, 1462, 1406, 1362, 1190, 1070, 937, 762, 701,
730, 528 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 352 (870).
Anal. Calcd for C110H118N4O12Fe2: C, 73.41; H, 6.61; N, 3.11.
Found: C, 72.96; H, 7.10; N, 3.03.
[Fe2(µ-XDK)(µ-O2CPhCy)(O2CPhCy)(Me-Im)2] (15). To a stirred
suspension of Tl2(XDK) (420 mg, 0.425 mmol) and Me-Im (71 µL,
0.89 mmol) in THF (5 mL) was added FeBr2 (187 mg, 0.851 mmol) in
THF (2 mL) followed by TlO2CPhCy (347 mg, 0.851 mmol) in THF
(2 mL). After 12 h the mixture was allowed to settle and the majority
of the precipitated TlBr was removed by filtration through Celite. The
resulting cloudy filtrate was evaporated to dryness and taken up in
benzene. Filtration through Celite afforded a clear filtrate. The Celite/
TlBr filter cakes were extracted twice with THF (5 mL) to obtain
additional product. The combined filtrate was evaporated to dryness
to yield a white residue. This material was suspended in pentane (10
mL), and the resulting white powder was isolated by filtration.
Recrystallization from THF/pentane afforded 15 as colorless needles
(386 mg, 72%). IR (KBr) 2959, 2929, 2859, 1734, 1695, 1609, 1535,
[Fe2(µ-XDK)(µ-O2CAr)(O2CAr)(py)2], O2CAr ) 2,4,6-Triiso-
propylbenzoate (8). This compound was prepared from TlO2CAr (342
mg, 0.757 mmol), pyridine (184 µL, 2.27 mmol), FeBr2 (83 mg, 0.378
mmol), and 4 (300 mg, 0.378 mmol) as described for 6. Recrystalli-
zation from THF/pentane provided 8 as yellow blocks (285 mg, 56%).
IR (KBr) 2960, 2870, 1736, 1699, 1611, 1536, 1464, 1412, 1357, 1220,
1192, 1069, 958, 841, 760, 698, 511 cm-1; UV-vis (CH2Cl2) (λmax
,
nm (ꢀ, M-1 cm-1)) 353 (790). Anal. Calcd for C74H94N4O12Fe2: C,
66.17; H, 7.05; N, 4.17. Found: C, 66.17; H, 7.01; N, 3.80.
1462, 1405, 1357, 1228, 1192, 1087, 958, 852, 762, 697, 628 cm-1
.
Anal. Calcd for C66H80N6O12Fe2: C, 62.86; H, 6.39; N, 6.66. Found:
C, 62.66; H, 6.79; N, 6.00.
[Fe2(µ-XDK)(µ-O2CAr)(O2CAr)(3-Fpy)2] (9). This compound was
prepared from TlO2CAr (342 mg, 0.757 mmol), 3-fluoropyridine (157
µL, 1.82 mmol), FeBr2 (83 mg, 0.378 mmol), and 4 (300 mg, 0.378
mmol) as described for 6. Recrystallization from THF/pentane provided
9 as yellow blocks (246 mg, 48%). IR (KBr) 2962, 2931, 2870, 1735,
[Fe2(µ-PXDK)(µ-O2CPhCy)(O2CPhCy)(Bu-Im)2] (16). This com-
pound was prepared from Tl2(PXDK) (788 mg, 0.682 mmol), Bu-Im
(184 µL, 1.40 mmol), FeBr2 (300 mg, 1.36 mmol), and TlO2CPhCy
(556 mg, 1.36 mmol) as described for 15. Recrystallization from Et2O/
pentane at -30 °C provided 16 as colorless needles (737 mg, 71%).
Recrystallization from a highly dilute pentane solution over several
days at room temperature yielded colorless blocks which allowed the
compound to be characterized by X-ray crystallography. IR (KBr)
3127, 2959, 2872, 1734, 1693, 1612, 1526, 1456, 1371, 1246, 1187,
1109, 948, 828, 763, 697, 660, 574 cm-1. Anal. Calcd for C84H116N6O12-
Fe2: C, 66.66; H, 7.72; N, 5.55. Found: C, 65.13; H, 7.81; N, 5.14.
Collection and Reduction of X-ray Data. Data were collected on
a Siemens CCD X-ray diffraction system controlled by a Pentium-
based PC running the SMART software package.31 All crystals were
1697, 1608, 1463, 1409, 1359, 1233, 1197, 1101, 959, 761, 537 cm-1
;
UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 361 (790). Anal. Calcd
for C74H92N4O12Fe2F2: C, 64.44; H, 6.72; N, 4.06. Found: C, 63.95;
H, 6.96; N, 3.53.
[Fe2(µ-BXDK)(µ-O2Ci-Pr)(O2Ci-Pr)(py)2] (10). This compound
was prepared from TlO2Ci-Pr (70 mg, 0.24 mmol), pyridine (60 µL,
0.72 mmol), FeBr2 (26 mg, 0.12 mmol), and 5 (150 mg, 0.120 mmol)
as described for 6. Recrystallization from THF/pentane provided 10
as yellow needles (135 mg, 76%). IR (KBr) 3062, 3028, 2965, 2923,
2869, 1733, 1693, 1602, 1546, 1495, 1446, 1364, 1190, 1034, 850,
763, 701, 528 cm-1; UV-vis (CH2Cl2) (λmax, nm (ꢀ, M-1 cm-1)) 357
(920). Anal. Calcd for C86H86N4O12Fe2: C, 69.82; H, 5.86; N, 3.79.
Found: C, 69.34; H, 6.05; N, 3.50.
(31) SMART; Version 4.0, Siemens Industrial Automation, Inc.: Madison,
WI, 1995.

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9005
Table 2. Summary of X-ray Crystallographic Data
6‚CH2Cl2
7‚2.5PhH
12‚2CH2Cl2‚0.5THF
13‚3PhH
16
formula
fw
space group
a, Å
b, Å
c, Å
C89H92N4O12Fe2Cl2
1592.27
P1h
C83H93N4O12Fe2
1450.38
P1h
C96H90N4O12.5Fe2Cl4
1753.22
P1h
C122H120N4O12Fe2
1945.92
P21/n
21.1197(11)
17.9106(9)
27.8373(14)
C84H116N6O12Fe2
1513.53
P1h
13.1316(2)
17.1922(1)
19.0528(1)
96.764(1)
104.042(1)
96.712(1)
4096.30(7)
2
14.7993(3)
15.5674(3)
18.0666(3)
88.8330(10)
81.5630(10)
70.7380(10)
3885.02(13)
2
14.0207(2)
16.78920(10)
19.3547(3)
96.1947(9)
102.8972(8)
96.2151(9)
4373.79(10)
2
13.2699(2)
14.09470(10)
23.0607(4)
97.1800(10)
97.3300(10)
106.4840(10)
4042.74(10)
2
R, deg
â, deg
γ, deg
V, Å3
Z
103.6300(10)
10233.4(9)
4
1.263
-85
0.349
0.791-1.000
3-57
40431
14536
9120
F
calcd, g cm-3
T, °C
1.291
-85
0.483
1.266
-85
0.437
1.331
-85
0.518
1.243
-85
0.422
µ(Mo KR), mm-1
transmission coeff
2θ limits, deg
total no. of data
no. of unique data
obsd dataa
0.857-1.000
0.788-1.000
0.858-1.000
0.871-1.000
3-46
3-57
3-57
3-57
17736
22684
25335
25104
11368
9221
16152
13216
17952
13809
17409
12887
no. of parameters
982
4.48
10.95
0.656, -0.506
897
4.59
12.94
1.046, -0.385
1025
8.81
25.18
1.247, -1.533
1231
7.52
15.56
0.752, -0.361
937
5.18
12.99
1.056, -0.597
R (%)b
wR2 (%)c
max, min peaks, e Å-3
a Observation criterion: I > 2σ(I). b R ) ∑||Fo| - |Fc||/∑|Fo|. c wR2 ) {∑[w(Fo - Fc2)2]/∑[w(Fo )2]}1/2
.
2
2
Table 3. Manometric Data on Dioxygen Uptake by 13, 16, and
mounted on the tips of glass fibers with Paratone-N (Exxon) and cooled
to low temperature (∼-85 °C). The general procedures for data
collection and reduction follow those reported previously.32 All
structures were solved by use of the direct methods programs SIR-
9233 or XS, part of the TEXSAN34 and SHELXTL35 program packages,
respectively. Structure refinements were carried out with XL, part of
the SHELXTL program package.35 All non-hydrogen atoms were
located and their positions refined by a series of least-squares cycles
and Fourier syntheses. Hydrogen atoms were assigned idealized
positions and given a thermal parameter 1.2 times the thermal parameter
of the carbon atom to which each was attached. Empirical absorption
corrections were calculated and applied for each structure with use of
the SADABS program.35
a
[Fe(Tp3,5-iPr )(O2CCH2Ph)]
2
O2 Uptake (equiv)
peroxo
formation
peroxo
decay
complex
solvent
[IrCl(CO)(PPh3)2]
CH2Cl2
CH2Cl2
THF
Et2O
THF
1.0
13
1.0 ( 0.1
0.55 ( 0.1
0.45 ( 0.1
0.9 ( 0.1
1.0 ( 0.1
0.7 ( 0.1
[Fe(Tp3,5-iPr )(O2CCH2Ph)]
0.3 ( 0.1
2
[Fe(Tp3,5-iPr )(O2CCH2Ph)]
0
2
16
16
1.0 ( 0.1
0
Et2O
a All O2 uptake values are in equivalents normalized to the amount
of O2 taken up by an equimolar quantity of [IrCl(CO)(PPh3)2] under
identical conditions. T ) -77 °C for the formation and 22 °C for the
decay. In each case, following decay, the system was brought back to
equilibrium at -77 °C for the final pressure reading.
The structure of 7 contains a lattice benzene molecule that was
refined at partial occupancy. Two of the three lattice benzene molecules
in the structure of 13 have large thermal parameters. The structure of
12 contains a lattice THF molecule disordered over two positions, where
an inversion center passes through the center of a carbon-carbon bond.
It was refined as a cyclopentane molecule with one of the carbon atoms
at full occupancy, and the occupancy of the other three carbon atoms
was distributed over two positions per atom and refined. The butyl
group of one of the Bu-Im ligands in the structure of 16 is disordered.
The occupancy of the carbon atom â to the imidazole nitrogen atom
was distributed over two positions and refined. The other carbon atoms
of the butyl group were refined at full occupancy.
HP8452A diode array spectrophotometer with a custom manufactured
immersion dewar, equipped with a 3-mL cell (1-cm path length) with
quartz windows connected to a 14/20 female joint via a 10-cm glass
tube. A temperature of -77 °C was maintained with a dry ice/acetone
bath and monitored with a Sensortek Model BAT-12 thermocouple
thermometer. Conductivity measurements were carried out in THF or
CH2Cl2 with a Fisher Scientific conductivity bridge, model 9-326,
outfitted with a platinum black electrode. Manometric experiments
were performed according to a literature procedure,36 using [IrCl(CO)-
Pertinent crystallographic information for each complex including
refinement residuals is available in Table 2. All bond distances and
angles, as well as atomic coordinates and equivalent isotropic displace-
ment parameters, are provided in Tables S1-S25 (Supporting Informa-
tion, CIF format).
18
(PPh3)2] or [FeTp3,5-iPr (O2CCH2Ph)] as authentic standards, which
2
form 1:1 and 2:1 complex:O2 adducts, respectively (Table 3).
EPR Spectroscopy. Spectra were recorded at 4.5 K on a Bruker
Model 300 ESP X-band spectrometer operating at 9.47 GHz. Liquid
helium temperatures were maintained with an Oxford Instruments EPR
900 cryostat. For all measurements, 1 mM frozen solutions were
prepared in 2-methyltetrahydrofuran.
Physical Measurements. NMR spectra were recorded on a Varian
Unity-300 NMR spectrometer. 1H NMR chemical shifts are reported
versus tetramethylsilane and referenced to the residual solvent peak.
FTIR spectra were recorded on a BioRad FTS-135 FTIR spectrometer.
Room-temperature UV-vis spectra were recorded on a Cary 1E
spectrophotometer. Low-temperature spectra were recorded on a
Resonance Raman Spectroscopy. Raman data were acquired by
using a Coherent Innova 90 Kr+ laser with an excitation wavelength
of 647.1 nm and 65 mW of power. A 0.6-m single monochromator
(1200 grooves/nm grating), with an entrance slit of 100 µm, and a liquid
nitrogen-cooled CCD detector (Princeton Instruments) were used in a
standard backscattering configuration. A holographic notch filter
(Kaiser Optical Systems) was employed to attenuate Rayleigh scattering.
(32) Feig, A. L.; Bautista, M. T.; Lippard, S. J. Inorg. Chem. 1996, 35,
6892.
(33) Burla, M. C.; Camalli, M.; Cascarano, G.; Giacovazzo, C.; Polidori,
G.; Spagna, R.; Viterbo, D. J. Appl. Crystallogr. 1989, 22, 389.
(34) TEXSAN: Single-Crystal Structure Analysis Software; Molecular
Structure Corporation: The Woodlands, TX, 1995.
(35) SHELXTL: Structure Analysis Program, 5.0; Siemens Industrial
Automation, Inc.: Madison, WI, 1995.
(36) Mahapatra, S.; Halfen, J. A.; Wilkinson, E. C.; Pan, G.; Wang, X.;
Young, V. G., Jr.; Cramer, C. J.; Que, L., Jr.; Tolman, W. B. J. Am. Chem.
Soc. 1996, 118, 11555.

9006 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
Figure 2. ORTEP diagrams with 50% thermal ellipsoids (clockwise from top left): [Fe2(µ-BXDK)(µ-O2Ct-Bu)(O2Ct-Bu)(py)2] (6), [Fe2(µ-XDK)(µ-
O2CPhCy)(O2CPhCy)(py)2] (7), [Fe2(µ-BXDK)(µ-O2CPh)(O2CPh)(py)2] (12), [Fe2(µ-BXDK)(µ-O2CPhCy)(O2CPhCy)(py)2] (13), and [Fe2(PXDK)(µ-
O2CPhCy)(O2CPhCy)(Bu-Im)2] (16).
Spectra of all samples were collected in toluene or CH2Cl2 solution at
-77 °C with the same custom-manufactured dewar used to acquire
low-temperature UV-vis spectra. Solutions were made as concentrated
as possible, ∼20 mM in the best cases, to ensure an optimal signal-
to-noise ratio. A total of 1000 scans, each with a 3-s exposure time,
were collected for each sample. Raman shifts were calibrated with
toluene or CH2Cl2 as an internal standard. The data were processed
by using CSMA software (Princeton Instruments, Version 2.4A) on a
Gateway 2000 computer, and the resulting ASCII files were processed
with Kaleidagraph (Abelbeck Software). Spectra are displayed in
Figure 6, and O-O stretching frequencies are listed in Table 4.
Mo1ssbauer Spectroscopy. Solid samples were prepared by sus-
pending ∼0.04 mol of powdered material in Apiezon N grease and
then packing the mixture into a nylon sample holder. Frozen solution
samples of diferrous complexes 13 and 15 were prepared in THF in a
drybox. In each case, 1 mL of an ∼40 mM solution was syringed into
a nylon sample holder, which was then sealed with Apiezon N grease
and rapidly frozen in liquid nitrogen. The peroxo derivative of 13 was
prepared in a round-bottom flask at -77 °C by exposing 1 mL of a
THF solution of 13 (40 mM) to excess O2 (1 atm) for 2 h. The sample
was quickly transferred (<5 s) into a nylon sample holder, immersed
in liquid nitrogen, by means of a glass syringe equipped with a 4-cm
stainless steel needle that had been precooled to -77 °C. All data
were collected at 77 K with a 57Co source in a Rh matrix at room
temperature and fit to Lorentzian line shapes. Spectra are illustrated
in Figure 7, and parameters derived from the fits are listed in Table 5.
UV-Vis Kinetics. Experiments were carried out by using the
custom dewar mentioned above. Reduced samples in freshly distilled
THF solution, typically 0.5 mM, were loaded in a drybox, and the dewar
was fitted with a rubber septum. The apparatus was brought out, placed
under a positive N2 pressure, and cooled to -77 °C with a dry ice/
acetone bath. The samples were subjected to an O2 purge at
atmospheric pressure for 5.0 min at a flow rate of 10 standard cubic
centimeters per minute (SCCM), maintained with a Matheson Dyna-
Blender Model 8280 oxygen/nitrogen constant flow mixing system.
Purging for this time length and at this flow rate resulted in highly
reproducible rate constants, thus ensuring that the entire sample solution
and vessel headspace had been completely purged with the desired gas
mixture. For dioxygen-dependent kinetic studies, the mole fraction of
O2 was controlled by diluting pure O2 with N2, while retaining a total
flow rate of 10 SCCM. Dioxygen concentrations were calculated by
using eq 4, where X ) mole fraction of O2, L ) Ostwald coefficient
-1
RT
P(g)LV°(1)
X )
+ 1
(4)
(
)
(0.245 for THF), P(g) ) partial pressure, and V°(1) ) molar volume.37
All experiments were carried out under pseudo-first-order dioxygen
concentrations, ranging from 3.07 to 15.4 mM, while keeping the initial
concentration of diferrous complex constant. Absorbance values were
measured at 600 nm every 2-10 s, depending on the rate at which the
reaction proceeded. Reactions were followed to at least 4 half-lives
in all cases. Data were analyzed with Kaleidagraph (Abelbeck
Software) and fit to model-dependent equations by using nonlinear least-
squares regression. In all cases the data were fit unambiguously to a
(37) Oxygen and Ozone; Battino, R., Ed.; Pergamon Press: Oxford, U.K.,
1981; Vol. 7.

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9007
Figure 5. Representative UV-vis spectra for the reaction of diferrous
complexes 6-15 with O2 at -77 °C in THF: (A) [Fe2(µ-BXDK)(µ-
O2CPhCy)(O2CPhCy)(py)2] (13) and (B) [Fe2(µ-PXDK)(µ-O2CPhCy)-
(O2CPhCy)(Bu-Im)2] (16).
Figure 3. A schematic illustration of the effect of ligand substituent
steric repulsion on the orientation of the bridging ancillary carboxylate,
including the relevant metrical parameters which account for this effect.
Figure 6. Resonance Raman spectra obtained with 647-nm excitation
at -77 °C on fluid solutions in toluene of (A) [Fe2(16O2)(µ-PXDK)(O2-
CPhCy)2(Bu-Im)2] (top) and [Fe2(18O2)(µ-PXDK)(O2CPhCy)2(Bu-Im)2]
(bottom). Peaks marked with an asterisk are due to solvent. Solvent-
subtracted spectra normalized to the 788-cm-1 toluene peak are shown
in part B.
Figure 4. Space-filling representations of [Fe2(µ-PXDK)(µ-O2Ct-Bu)-
(O2Ct-Bu)(Me-Im)2] from ref 23 (top) and [Fe2(µ-BXDK)(µ-O2Ct-Bu)-
(O2Ct-Bu)(py)2] (6) (bottom), looking down the Fe‚‚‚Fe vector.
A∞ is the absorbance at infinite time, and kobs (≡kψ) is the observed
pseudo-first-order rate constant. All plots of ln(kobs) vs ln([O2]) were
linear with a slope close to 1.0, implying that an overall second-order
rate law applies, eq 6. Second-order rate constants, listed in Table 6,
were derived from the slopes of kobs vs [O2] plots. Standard errors in
these rates are based on linear regression analyses of the slopes and
are reported at the 95% confidence level. Plots of A (at 600 nm) vs t,
kψ vs [O2], and ln(kψ) vs ln([O2]) for complexes 6, 7, 13, and 14 are
first-order exponential buildup, defined in eq 5, where At is the
At ) A∞ + (∆A)(1 - exp(-kobst))
rate ) k[FeII2][O2]
(5)
(6)
absorbance at time t, ∆A is the change in absorbance during the reaction,

9008 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
Table 4. Summary of Resonance Raman Data for the Peroxo
Adducts
Results and Discussion
Preparation and Characterization of [Fe(XDK)(py)2] (4)
and [Fe(BXDK)(py)2] (5). To develop an efficient synthetic
route to a family of diiron(II) complexes incorporating XDK
and related ligands, as well as ancillary carboxylates of variable
steric demands, access to large amounts of pure mononuclear
precursors was desired. Treatment of Tl2(XDK) or Tl2(BXDK)
with 1 equiv of iron(II) bromide and excess pyridine provided
yellow bis(pyridine) adducts 4 and 5, respectively, in multigram
quantities and high yield (eq 7). An X-ray structural analysis
of the XDK derivative revealed the iron atom to be coordinated
ν(16O-16O)
(18O2)
compound
solvent
[Fe2(O2)(µ-BXDK)(O2Ct-Bu)2(py)2]
[Fe2(O2)(µ-XDK)(O2CPhCy)2(py)2]
[Fe2(O2)(µ-XDK)(O2CAr)2(py)2]
[Fe2(O2)(µ-BXDK)(O2CPhCy)2(py)2]
[Fe2(O2)(µ-BXDK)(O2CPhCy)2(py)2]
[Fe2(O2)(µ-BXDK)(O2CAr)2(py)2]
[Fe2(O2)(µ-PXDK)(O2CPhCy)2(Bu-Im)2]
toluene
toluene
toluene
toluene
CH2Cl2
toluene
toluene
864
864
864
864
861 (811)
864
864 (818)
(7)
by three carboxylate oxygen and two pyridine nitrogen atoms.38
The structure is essentially identical to that determined for the
Zn(II) analogue,39 with one long Fe-OXDK interaction at
2.326(5) Å and the other two at an average distance of
2.080(3) Å. In contrast, the analogous reaction with 1-methyl-
or 1-butylimidazole led to an intractable mixture of products,
even when exactly 2 equiv were used.
Preparation and Characterization of Dinuclear [Fe2(µ-
BXDK)(µ-O2Ct-Bu)(O2Ct-Bu)(py)2] (6), [Fe2(µ-XDK)(µ-
O2CPhCy)(O2CPhCy)(py)2] (7), [Fe2(µ-XDK)(µ-O2CAr)-
(O2CAr)(py)2] (8), [Fe2(µ-XDK)(µ-O2CAr)(O2CAr)(3-Fpy)2]
(9), [Fe2(µ-BXDK)(µ-O2Ci-Pr)(O2Ci-Pr)(py)2] (10), [Fe2(µ-
BXDK)(µ-O2CCH2t-Bu)(O2CCH2t-Bu)(py)2] (11), [Fe2(µ-
BXDK)(µ-O2CPh)(O2CPh)(py)2] (12), [Fe2(µ-BXDK)(µ-
O2CPhCy)(O2CPhCy)(py)2] (13), [Fe2(µ-BXDK)(µ-O2CAr)-
(O2CAr)(py)2] (14), [Fe2(µ-XDK)(µ-O2CPhCy)(O2CPhCy)-
(Me-Im)2] (15), and [Fe2(µ-PXDK)(µ-O2CPhCy)(O2CPhCy)-
(Bu-Im)2] (16). The diferrous bis(pyridine) complexes 6-14
were conveniently prepared from 4 or 5 and “[Fe(O2CR)2(py)n]”,
the latter being generated in situ from iron(II) bromide, the
thallium(I) salts of the corresponding ancillary carboxylate, and
excess pyridine (eq 8). The 3-fluoropyridine derivative 9 and
imidazole adducts 15 and 16 were prepared in one-pot reactions
from the thallium carboxylates, 2 equiv of iron(II) bromide, and
2 equiv of 3-fluoropyridine or 1-alkylimidazole. In the synthesis
of 15 and 16, removal of solid TlBr was aided by allowing the
reaction mixtures to stand overnight prior to filtration through
Figure 7. Zero-field Mo¨ssbauer spectra (experimental data (•),
calculated fit (-)) recorded at 77 K for [Fe2(µ-BXDK)(µ-O2CPhCy)-
(O2CPhCy)(py)2] (13) in the solid state (A), in THF solution (B), and
after exposure to dioxygen in THF solution (C). The diferrous
component in spectrum C was approximated as a single quadrupole
doublet.
provided in Figures 8 and S6-S8 (see Supporting Information for
Figures S6-S8), respectively.
Table 5. Zero-Field Mo¨ssbauer Parameters for [Fe2(µ-BXDK)(µ-O2CPhCy)(O2CPhCy)(py)2] (13),
[Fe2(µ-XDK)(µ-O2CPhCy)(O2CPhCy)(Me-Im)2] (15), [Fe2(O2)(µ-BXDK)-(O2CPhCy)2(py)2], Oxyhemerythrin, and MMOHperoxo
complex
T (K)
δ (mm s-1
)
∆EQ (mm s-1
)
Γ (mm s-1
)
ref
13 (powder)
13 (THF)
15 (THF)
77
1.06
1.28
1.10
1.34
1.14
1.36
0.47
0.63
0.50
0.51
0.66
2.56
3.01
2.85
2.87
2.98
2.94
0.88
1.20
2.02
1.01
1.51
0.37
0.37
0.48
0.48
0.45
0.45
0.40
0.40
∼0.3
∼0.3
0.27
a
77
77
77
77
4.2
a
a
[Fe2(O2)(µ-BXDK)(O2CPhCy)2(py)2]
a
Oxyhemerythrin
43
15
MMOHperoxo
a This work.

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9009
Table 6. Correlation of Rate Constant of Peroxodiiron(III) Complex Formation at -77 °C, with Tilt Angle (τ) Measured in the Solid State
complex
parameter
6
7
13
14
k (M-1 s-1
τ (deg)a
)
0.0398 ( 0.0009
73
0.045 ( 0.003
71
0.16 ( 0.01
66
0.115 ( 0.005
b
a See Figure 4 for a definition of τ. b An X-ray structure was determined for 14, but it was of insufficient quality for an accurate measurement
of bond distances and angles.
(8)
Celite. Further extractions of the resulting Celite/TlBr filter
cake were required to recover substantial quantities of additional
product. All complexes were obtained in high yield and purity
following recrystallization at room temperature. Despite re-
peated attempts in each case, elemental analyses for samples
of 4, 7, and 16 deviated by 1-2% from the calculated carbon
values. It is unclear whether these complexes have poor
combustion properties or contain minor quantities of impurities,
but other closely related derivatives gave satisfactory analyses.
Structural diagrams of 4-16 are provided in Scheme 1.
The IR spectra of 7-16 displayed C-O stretches between
1602 and 1614 cm-1 and another centered at ∼1693 cm-1
,
which was more intense and broadened compared to that in
either the dithallium(I) or mononuclear iron(II) precursors. Both
features are attributed to the two additional carboxylate ligands
in the products. The UV-vis spectra of the pyridine adducts
4-14 all exhibited an absorbance at ∼360 nm (ꢀ ) 500-1000
M-1 cm-1), tentatively assigned as a metal-to-ligand charge-
transfer band, which was absent in the colorless 1-alkylimidazole
derivatives. Complexes 4-16 are all nonelectrolytes, and most
are soluble in solvents as nonpolar as THF. Complex 16, with
its highly saturated hydrocarbon periphery, is readily soluble
in diethyl ether and cyclopentane.
The ability to acquire X-ray crystal structures of 6, 7, 12,
13, and 16 has enabled us to ascertain how each ligand
component, with its particular steric and electronic properties,
might influence the iron coordination geometries and, ultimately,
dioxygen reactivity of the complexes. ORTEP and space-filling
diagrams of these compounds are displayed in Figures 2 and 4,
respectively, and selected bond distances and angles are listed
in Figure 3. Each structure consists of one six-coordinate and
one four- or five-coordinate iron atom. The number of donor
atoms on the coordinatively unsaturated iron center (Fe2)
depends on the choice of N-donor ligand and the stereochemical
demands of the XDK and ancillary carboxylates. The Fe2 atoms
in 6 and 12 are best described as distorted trigonal bipyramidal,
Figure 8. Kinetic data for the reaction of [Fe2(µ-BXDK)(µ-O2Ct-Bu)-
(O2Ct-Bu)(py)2] (6) with O2 in THF at -77 °C: (A) plot of A600 vs t
with [6]0 ) 0.73 mM and [O2] ) 15.4 mM and a fit to a first-order
exponential buildup, (B) plot of ln(kψ) vs ln([O2]), and (C) plot of kψ
vs [O2].
and those in 7, 13, and 16 as distorted tetrahedral. The other
iron atom (Fe1) adopts a pseudo-octahedral environment
comprising two carboxylate oxygen atoms of a syn,syn-bidentate
bridging XDK ligand, two symmetric chelating ancillary car-
boxylate oxygen atoms, and a nitrogen atom from the N-donor
ligand.
(38) X-ray data for 4 crystallized from CH2Cl2/Et2O: P21/n, a )
12.3185(2) Å, b ) 16.2489(2) Å, c ) 20.4920(1) Å, â ) 103.77(1)°, V )
3983.92(8) Å3, Z ) 4, T ) 188 K, R ) 0.0677, wR2 ) 0.1372. This structure
is not described in further detail here because it is nearly identical to that
of the Zn(II) analogue (see ref 39) .
The most significant difference in these structures is the
orientation of the ancillary bridging carboxylate ligand relative
(39) Tanase, T.; Yun, J. W.; Lippard, S. J. Inorg. Chem. 1995, 34, 4220.

9010 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
Scheme 1
to the rest of the complex framework. Six metrical parameters
change significantly within the series, Fe2-O402, Fe2-O401,
Fe1-O402-Fe2, Fe1-O402-C401, Fe2-O401-C401, and
Fe2-O402-C401. These changes are related to the previously
defined tilt angle, τ, of the monodentate bridging carboxylate
ligand, as illustrated in Figure 3.40 Within the subset of pyridine
complexes 6, 7, 12, and 13, repulsive interactions between
substituents on the ancillary carboxylate ligands decrease τ. In
12 and 13, τ is similarly affected by a steric interaction between
these substituents and the benzyl groups on BXDK. As the
Fe2-O402 bridge distance (B) lengthens, the Fe2-O401
dangling distance (D) shortens, resulting in little alteration of
the Fe1-O402 distance (A) across the series. Similarly, an
increase in the Fe2-O401-C401 angle (γ) is accompanied by
a decrease in the Fe1-O402-Fe2 (θ), Fe1-O402-C401 (R),
and Fe2-O402-C401 (â) angles. On the basis of the values
of these parameters, the order of increasing steric repulsion
between carboxylate substituents for this subset of pyridine
complexes is 12 < 6 < 7 < 13, corresponding to ∆τ ) -7°.
By contrast, τ is much smaller for the 1-butylimidazole adduct
16 compared to all of the pyridine complexes. We attribute
this property to the more basic character of the 1-alkylimidazole
donor compared to pyridine, a factor that increases the electron
density at Fe2, thereby completely abrogating the Fe2-O402
interaction. Hence, for 16, the steric and electronic factors have
the combined effect of inducing a nearly complete carboxylate
shift40 from the monodentate bridging mode observed for the
pyridine adducts to a syn,syn-bidentate bridging mode.
Reactions of the Diferrous Complexes with O2 To Form
Diiron(III) Peroxo Complexes. (a) UV-Vis Studies. Previ-
ous work revealed the generation of a short-lived intermediate
when [Fe2(µ-PXDK)(µ-O2Ct-Bu)(O2Ct-Bu)(Me-Im)2] was ex-
posed to excess O2 at -77 °C in THF solution.23 In a first
attempt to obtain a more stable derivative, we implemented the
more sterically demanding BXDK ligand, while holding the
other steric properties constant, and changed the N-donor ligand
(40) Rardin, R. L.; Tolman, W. B.; Lippard, S. J. New J. Chem. 1991,
15, 417.

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9011
to pyridine. Space-filling diagrams of the previously reported
complex and 6 are displayed in Figure 4. From these repre-
sentations, it is apparent that the BXDK benzyl substituents
proximal to its carboxylate oxygen donor atoms shield the iron
coordination sites to a greater extent than do the PXDK propyl
groups. Moreover, the benzyl substituents would tend to direct
the approach of the O2 molecule to the unsaturated metal center
along the Fe-Fe vector.
Exposure of 6 to O2 under conditions identical to those
previously reported23 resulted in a much slower color change
to deep midnight blue. The buildup of a broad absorption
centered at ∼580 nm (ꢀ ≈ 1200 M-1 cm-1) was monitored by
UV-vis spectroscopy over the course of 2-3 h. This feature
is characteristic of a peroxo-iron(III) ligand-to-metal charge-
transfer transition.1,5 The oxygenation reaction is irreversible,
since application of high vacuum/Ar purge cycles to the solution
at -77 °C resulted in no diminution in the intensity of the
charge-transfer band. Warming above -65 °C under excess
O2 or Ar, or under high vacuum, however, led to decomposition
within 5 min, as judged by a color change to light brown. A
similar decomposition was noted when the solution was allowed
to stand at -77 °C for 12-24 h.
The other diiron(II) tetracarboxylate complexes were screened
for their ability to form stable peroxo complexes. The O2
reactions either resulted in rapid decomposition to form Fe(III)
products or afforded an intermediate very similar to that
observed with 6 (Figure 5). For derivatives containing BXDK
and pyridine, ancillary carboxylate ligands smaller than pivalate
(10-12) reacted with O2 at a similar rate, but none of these
reactions resulted in buildup of the 580-nm charge-transfer band.
The sterically less demanding XDK ligand formed stable peroxo
adducts when significantly more hindered ancillary carboxylates
were employed, such as in compounds 7 and 8. These same
hindered carboxylates afforded similarly stable peroxo adducts
with 3-fluoropyridine as in 9 or 1-alkylimidazole donors as in
15 and 16 (Figure 5B). The rate of peroxo formation was
considerably faster for the 1-alkylimidazole complexes com-
pared to all the pyridine analogues, however, with 15 and 16
reacting with O2 to completion within seconds compared to
6-14, which required 0.25-2 h, as judged by UV-vis
spectroscopy. The energy and intensity of the peroxo-to-iron-
(III) charge-transfer bands in these derivatives were the same
as for the pyridine analogues. This result is not surprising
because other diiron(III) peroxo complexes with widely varying
ligand sets exhibit nearly identical transitions.1,5 Moreover, none
of the peroxo derivatives released O2 at -77 °C, and all had
similarly narrow ranges of thermal stability, irrespective of their
rates of formation.
elsewhere.26 Taken together, the results from these steric,
electronic, and solvent polarity studies demonstrate that peroxo
adducts can be stabilized in this ligand environment provided
that the carboxylate substituents afford sufficient steric bulk.
Characterization of the Diiron(III) Peroxo Complexes. (a)
EPR and 1H NMR Spectroscopy. The peroxo species
described above are EPR silent at 4.5 K, thereby implying
antiferromagnetically coupled, S ) 0 ground states. As with
other diiron(III) peroxides, including the Hperoxo intermediate
of MMOH,15 such coupling is probably mediated by a bridging
peroxide ligand.41,42 The failure to observe a 1H NMR spectrum
for the peroxo adducts in CD2Cl2 at -80 °C, however, suggests
the presence of low-lying paramagnetic excited states which
can be thermally accessed and which induce fast nuclear spin
relaxation of the ligand protons. In contrast, a paramagnetically
1
shifted but reasonably sharp H NMR spectrum was observed
for a diiron(III) peroxo complex having a sterically hindered
hydrotris(pyrazolyl)borate ligand.41
(b) Resonance Raman Spectroscopy. Raman spectra were
acquired for peroxo adducts of 6-8, 13, 14, and 16 in fluid
solution at -77 °C. The results of these experiments are
summarized in Table 4, and the spectra acquired for the peroxo
derivative of 16 are displayed in Figure 6. The pyridine and
1-butylimidazole derivatives all exhibit a ν(16O-16O) band at
864 cm-1 in toluene solution. When 18O2 was used in the
synthesis, this band shifted to 818 cm-1, in good agreement
with the theoretically predicted value of 814 cm-1 for a diatomic
O-O harmonic oscillator. The ν(16O-16O) peak shifts slightly
to 861 cm-1 for the peroxide complex of 13 in CH2Cl2 solution.
An Fe-O stretching frequency could not be assigned in toluene
or CH2Cl2 due to the presence of peaks from the complex and
solvent which obscured this region of the spectrum. The ν-
(16O-16O) frequencies of 861-864 cm-1 are 15-40 cm-1 lower
in energy than reported for crystallographically characterized
(µ-1,2-peroxo)diiron(III) complexes,1,5 but 50-60 cm-1 higher
in energy than those observed in mononuclear η2-peroxo-
iron(III) units.1 We therefore tentatively assign µ-1,2 coordina-
tion modes for the peroxo derivatives of 13 and 16.
Raman spectroscopy was also used to investigate the pos-
sibility that the peroxide ligand might exchange with free O2.
Such a process could be either dissociative (eq 9), implying
reversible O2 binding, or associative (eq 10). The former
18O2
[FeIII2(18O2)(µ-L)(O2CR)2(N)2] y-
z
16O2
[FeII (µ-L)(O2CR)2(N)2] y+
z
2
[FeIII2(16O2)(µ-L)(O2CR)2(N)2] (9)
16O2
[FeIII2(18O2)(µ-L)(O2CR)2(N)2] y+
z
(b) Manometric Studies. Dioxygen uptake was monitored
by manometry for the pyridine adduct 13 in CH2Cl2 and for
the 1-butylimidazole adduct 16 in THF and Et2O (Table 3).
Compounds 13 and 16 consumed 1.0 ( 0.1 equiv of O2,
consistent with formation of discrete diiron(III) peroxide species.
When these adducts were allowed to decay by warming them
to room temperature, additional O2 consumption was observed
in manometric experiments carried out in CH2Cl2 or THF,
whereas no further gas uptake occurred for diethyl ether
solutions. This behavior contrasts with that previously reported
for peroxo-bridged diiron(III) alkoxo complexes,25 where 0.5
equiv of O2 was released upon peroxo decay, suggesting that,
in the systems described here, either the peroxide ligand is
retained in the decomposition products or that ligand or solvent
oxidation occurs during peroxo decay. A detailed analysis of
this chemistry has been undertaken and will be reported
18O2
[FeIII2(18O2)(16O2)(µ-L)(O2CR)2(N)2] y-
z
[FeIII2(16O2)(µ-L)(O2CR)2(N)2] (10)
mechanism is unlikely because O2 was not liberated when a
solution of the peroxo adduct was placed under a dynamic
vacuum. A sample of the [18O]peroxo derivative of 16 was
prepared by exposing a 20 mM solution in toluene to ∼0.25
atm of excess 18O2 for 1 h, and resonance Raman spectra were
recorded prior to and following a vigorous, 5-min 16O2 purge.
(41) 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.
(42) Dong, Y.; Me´nage, S.; Brennan, B. A.; Elgren, T. E.; Jang, H. G.;
Pearce, L. L.; Que, L., Jr. J. Am. Chem. Soc. 1993, 115, 1851.

9012 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
The ν(16O-16O) feature appeared in the spectrum immediately
following the purge, but not at the expense of the intensity of
the ν(18O-18O) band. Moreover, the ν(16O-16O) peak contin-
ued to grow in intensity when the solution was allowed to stir
under 1 atm of 16O2 over 2 h, but again the ν(18O-18O) band
did not diminish. These results indicate that oxygenation of
the reduced complex is slow in the concentrated solutions of
16 required for the Raman experiment and that peroxide does
not exchange with free O2, at least over several hours.
(c) Mo1ssbauer Spectroscopy. Zero-field Mo¨ssbauer studies
were carried out to gain additional information concerning the
environment around the iron centers of the diiron(II) and
peroxodiiron(III) complexes. Spectra overlaid with theoretical
fits to the data are presented in Figure 7 for 13 and its peroxo
dinuclear species containing inequivalent iron centers or two
distinct dinuclear species with equivalent iron centers within
the same molecule. Since the generation of a 1:1 ratio of two
different diiron(III) peroxo isomers is unlikely to occur upon
addition of O2 to a solution of 13, we conclude that the former
description best accounts for the nature of this peroxo complex.
Kinetic Studies of the Reaction of the Diferrous Com-
plexes with O2. (a) Correlation of Kinetic Parameters with
Complex Stereochemistry. The UV-vis, manometric, EPR,
resonance Raman, and Mo¨ssbauer spectroscopic data collectively
point to the formation of discrete (µ-1,2-peroxo)diiron(III)
complexes with inequivalent metal centers. An investigation
of the kinetics of peroxo formation was undertaken, using time-
resolved UV-vis spectroscopy to determine the reaction order
with respect to the diferrous precursor complex and O2.
Moreover, since X-ray structural studies revealed that the
coordination number of one of the two iron atoms in the di-
iron(II) complexes could be modulated by intramolecular steric
factors (vide supra), we were interested to determine whether
such an effect might alter the kinetics of oxygenation.
derivative. The derived parameters, along with those for
15
oxyhemerythrin43 and Hperoxo
,
are summarized in Table 5.
The Mo¨ssbauer spectrum of powdered 13 (Figure 7A)
indicates the presence of two distinct high-spin Fe(II) centers,
consistent with its X-ray crystal structure (Figure 2). Valence
bond theory dictates that the lower-coordinate Fe2 center of 13
will have greater s-orbital character relative to the pseudooc-
tahedral Fe1. Because these two Fe(II) centers are coordinated
by similar weak-field ligands, the quadrupole doublet with the
smaller isomer shift (1.06 mm s-1) is assigned to Fe2 and the
one with the larger isomer shift (1.28 mm s-1) to Fe1. Such
an assignment of the two quadrupole doublets to specific iron
centers in 13 requires for consistency that the previously
reported23 Mo¨ssbauer parameters of less hindered analogues be
interchanged. In a separate issue, although one of the sterically
less hindered analogues of 13 exhibited nearly identical Mo¨ss-
bauer parameters in both the powdered and frozen solution
states,23 the spectrum of 13 appears different when measured
in a frozen THF solution (Figure 7B). The latter spectrum,
however, is essentially identical to the frozen solution Mo¨ssbauer
spectrum of 15, suggesting that 13 and 15 share very similar
structures in solution. The integrated intensities of the two
overlapping quadrupole doublets are equal, thereby indicating
that the dinuclear structure remains intact in solution.
Kinetics experiments were carried out at -77 °C on the
pyridine adducts 6, 7, 13, and 14. Data were collected under
pseudo-first-order conditions, with the diiron complex as the
limiting reagent. Dioxygen concentrations were varied over a
range of 3.1-15.4 mM. The time dependence of the buildup
of the peroxo adduct was monitored at 600 nm (Figures 8A
and S6-S8). The kinetic trace could be fit to first-order
exponential curves for each complex. Plots of ln(kψ) vs ln[O2]
afforded straight lines with slopes near 1.0, indicating that the
oxygenation reaction is first order in O2 (Figures 8B and S6-
S8). Second-order rate constants were obtained for the four
iron complexes from the slopes derived from plots of kψ vs [O2]
(Figures 8C and S6-S8). These rate constants are provided in
Table 6 along with metrical data from the crystal structure
determinations of 6, 7, and 13 describing the orientation of each
bridging ancillary carboxylate. Activation parameters could not
be obtained because the peroxo adducts decompose rapidly
above -65 °C.
The Mo¨ssbauer spectrum of the peroxo derivative of 13 in
frozen THF solution is displayed in Figure 7C. A moderate
amount (∼20%) of the diiron(II) starting material is detected
in this sample, as judged by a comparison of the position of
the absorption feature at ∼2.8 mm s-1 with that observed for
13 (Figure 7B). This result is in accord with the resonance
Raman experiment described above, where it was determined
that oxygenation of very concentrated (20 mM) diiron(II)
samples is slow relative to the rate in more dilute solutions,
such as that used for the UV-vis studies. Given the thermal
instability of the peroxo complexes, which decompose signifi-
cantly within 12-24 h at -77 °C, and the fact that the
spectroscopic properties of the reduced complexes are well-
known, we chose not to allow any of these concentrated
reactions to proceed beyond ∼2 h.
The first-order dependence in each reagent suggests that the
rate-limiting step in these oxygenation reactions is a bimolecular
collision between the diiron(II) complex and dioxygen. Within
the series, the rate constants increase modestly, over a 4-fold
range, with increasing steric demands of the carboxylate
substituents. This kinetic trend correlates well with lengthening
of the Fe2-O402 bond and decrease in τ as a function of the
same steric parameters and may indicate that, in the transition
state, cleavage of this bond is required to open a coordination
site for O2 binding to Fe2. This event corresponds to a rate-
limiting carboxylate shift from monodentate bridging to syn,syn-
bidentate bridging. Such structural/kinetic correlations must be
viewed with caution, however, since the Mo¨ssbauer studies on
the diiron(II) complexes reveal that the solid and solution
structures may not be the same. Nevertheless, the trend does
suggest that bond cleavage is involved in the rate determining
step and argues against a mechanism where binding of the O2
molecule to an iron center is impeded by the sterically
encumbering carboxylate substituents. Such a mechanism was
proposed for the oxygenation of a series of hydrotris(pyrazolyl)-
borate iron(II) carboxylate complexes, where exceedingly bulky
carboxylate substituents completely shut down the reaction.41
By contrast, space-filling models of 6-16 reveal that the four-
or five-coordinate iron center is unobstructed, irrespective of
the carboxylate substituents (see Figure 4, for example).
Figure 7C portrays the experimental data with the diiron(II)
component subtracted out, together with a fit to two quadrupole
doublets of equal integrated intensity. The isomer shift (0.48
and 0.63 mm s-1) and quadrupole splitting (0.88 and 1.20 mm
s-1) values are both within the range for high-spin Fe(III),44
and resemble those determined for the phenoxo-bridged per-
oxodiiron(III) complex in Table 1.19 These results indicate that
the peroxo complex of 13 is composed of either a single
(43) Clark, P. E.; Webb, J. Biochemistry 1981, 20, 4628.
(44) Dickson, D. P. E.; Berry, F. J. In Mo¨ssbauer Spectroscopy;
Cambridge University Press: Cambridge, U.K., 1986.
The 1-alkylimidazole adducts 15 and 16 react to completion

Models for Diiron Protein Centers
J. Am. Chem. Soc., Vol. 120, No. 35, 1998 9013
Scheme 2
with O2 within the time of mixing of the two reagents (<1 s),
much more rapidly than the oxygenation rates observed for the
pyridine complexes. Neither complex could be studied by the
conventional methods employed here. The less sterically
hindered analogue [Fe2(µ-PXDK)(µ-O2Ct-Bu)(O2Ct-Bu)-
(Me-Im)2] was previously analyzed by stopped-flow UV-vis
Comparison to Results on MMOH and RNR-R2. Several
of the findings in the present modeling study are of potential
relevance to chemistry that takes place at the carboxylate-bridged
diiron active sites in MMOH and RNR-R2. Recent experiments
reveal that the two enzymes can access common oxygen
intermediates, the fate of which is exquisitely regulated accord-
ing to their different oxidative functions. In one study, an RNR-
R2 mutant (D84E) was prepared in which an aspartate residue
in the active site was replaced with glutamate, affording an
MMOH-like coordination environment.45 With this modest
change in the flexibility of a single carboxylate ligand, the D84E
mutant remarkably afforded a diiron(III) peroxo intermediate
spectroscopy under very similar experimental conditions: [FeII ]
2
) 0.8 mM (after mixing) in THF, T ) -76 °C, excess O2,23
revealing a pseudo-first-order rate constant of 300 s-1. This
value is more than a factor of 105 larger than the pseudo-first-
order rate constants observed for the pyridine adducts. Such
dramatically enhanced rates could be due in part to the minimal
bonding interaction between the Fe2 and O402 in this complex.
Because changes in this parameter have only a modest effect
on the rate constants, however, we attribute most of the rate
difference instead to the greater basicity of the 1-alkylimidazole
ligand compared with pyridine. Such an electronic factor should
lead to a significantly more negative midpoint potential for the
FeIII/II couple with 1-alkylimidazole, thereby enhancing the
thermodynamic driving force of the O2 addition reaction and
accelerating the rate.
that formed and decayed at rates comparable to those of Hperoxo
.
The product contained a stoichiometric quantity of a stable
tyrosyl radical, most likely formed via intermediate X. These
results lend credence to the assignment of the initial O2
intermediate in native RNR-R2 as a diiron(III) peroxo species
and reveal that the stability of peroxo intermediates in RNR-
R2 and MMOH can be regulated by tuning the active-site
ligands.
The interplay between the stereochemical demands of the
carboxylate ligands, the extent of carboxylate shifting, and the
ability to form a peroxo adduct illustrates one possible way in
which the D84E mutant of RNR-R2 mutant might be able to
access the diiron(III) peroxo intermediate not seen with the
native protein.45 The profound dependence of the rate of
peroxodiiron(III) formation on the electronic properties of the
N-donor ligand provides additional insight into how the protein
cores might be tuned by the surrounding environment. The
X-ray structure of MMOH revealed hydrogen bonding of the
two histidine ligands coordinated to the dinuclear iron center
(His147 and His246, Figure 1) to nearby aspartate residues in
the four-helix bundle housing the active site of the enzyme.46
Attenuation of this hydrogen bonding interaction, for example
by the binding of protein B to the hydroxylase, could dramati-
cally alter the rate of conversion of the diiron(II) form of the
enzyme upon exposure to O2, as observed in changing from
pyridine to imidazole ligands in the present model studies.
Further direct studies of the enzymes are required to test these
(b) Proposed Mechanism for Peroxo Fomation. From
these considerations we propose (Scheme 2) a working mech-
anism for the formation of the peroxo complex. Although
peroxo buildup exhibits clean first-order dependence in each
reaction component, the kinetics do not reveal whether the rate-
limiting step leads directly to formation of the peroxo species.
Very fast electron-transfer and/or ligand rearrangement steps
may occur after O2 binding, prior to final peroxo assembly. If
formation of the peroxo complex occurs rapidly after the rate-
limiting step, possibilities for the latter would be formation of
a diiron(II)-O2 adduct, an FeIIFeIII superoxide complex, or
isomerization of an intermediate complex. Two possibilities
are represented by structures a and b in Scheme 2. For the
final structure, the spectroscopic data are consistent with a
bridged peroxo arrangement having inequivalent iron sites. To
accommodate this feature, the mirror plane and rotational
symmetry of the diiron-XDK framework must be eliminated.
Structures c-e incorporate these properties through alterations
in the orientation of the peroxide (c) or the other ligands (d
and e). We consider these or closely related isomers to be viable
candidates for the structures of the peroxo complexes presented
here.
(45) Bollinger, J. M., Jr.; Krebs, C.; Vicol, A.; Chen, S.; Ley, B. A.;
Edmondson, D. E.; Huynh, B. H. J. Am. Chem. Soc. 1998, 120, 1094.
(46) Rosenzweig, A. C.; Frederick, C. A.; Lippard, S. J.; Nordlund, P.
Nature 1993, 366, 537.

9014 J. Am. Chem. Soc., Vol. 120, No. 35, 1998
LeCloux et al.
suggestions. Several deficiencies in the models remain to be
addressed by future work. Included are the fact that the Hperoxo
intermediate has iron(III) centers that are indistinguishable by
Mo¨ssbauer spectroscopy and that the present model complexes
do not convert to a Q-type intermediate upon warming.26 The
restrictive “double-bridging” nature of the XDK ligand family
may be responsible in part for the latter shortcoming and is
currently being addressed.
rate is affected much more dramatically by the basicity of the
N-donor ligand, however, with a factor of >105 difference in
the pseudo-first-order rate constants for 1-alkylimidazole versus
pyridine ligation. These studies provide the most accurate
compositional models of the Hperoxo intermediate reported to
date, and their steric and electronic tunability have allowed us
to address several of the many factors which affect their
formation, structure, and stability.
Conclusions
Acknowledgment. This work was supported by a grant from
the National Science Foundation and AKZO Corp. We thank
Prof. W. M. Reiff (Northeastern University) for allowing us to
use his Mo¨ssbauer spectrometer and A. M. Valentine and D. P.
Steinhuebel for experimental assistance and many helpful
discussions.
A family of peroxo-bridged diiron(III) complexes has been
prepared which matches exactly the combination of carboxylate
and imidazole donors of the Hperoxo intermediate in the catalytic
cycle of sMMO. A variety of methods, including UV-vis,
EPR, resonance Raman, and Mo¨ssbauer spectroscopy, have been
employed to elucidate their structures. The peroxo species have
inequivalent iron coordination environments, a feature that is
attributed either to an asymmetric peroxide bridge, a dispro-
portionate arrangement of the ancillary carboxylate and nitrogen
donor ligands about each iron center, or a combination of the
two. Kinetic studies revealed the rate of peroxo formation to
be first order in diiron(II) complex and O2, implying a rate-
determining bimolecular collision between reactants. The rate
of peroxo formation varies inversely with the steric demands
of the carboxylate substituents, implying that a carboxylate shift
is associated with the rate-determining step. The oxygenation
Supporting Information Available: Figures S1-S5, dis-
playing fully labeled ORTEP diagrams for each reported
structure, and Figures S6-S8, displaying kinetic data for the
oxygenation of complexes 7, 13, and 14 (9 pages, print/PDF).
X-ray crystallographic files, in CIF format, for complexes 6, 7,
12, 13, and 16, are available on the Internet only. See any
current masthead page for ordering information and Web access
instructions.
JA981216S