Toward functional carboxylate-bridged diiron protein mimics: Achieving structural stability and conformational flexibility using a macrocylic ligand framework

Article abstract of DOI:10.1021/ja2021312

A dinucleating macrocycle, H₂PIM, containing phenoxylimine metal-binding units has been prepared. Reaction of H₂PIM with [Fe₂(Mes)₄] (Mes = 2,4,6-trimethylphenyl) and sterically hindered carboxylic acids, Ph

Full text of DOI:10.1021/ja2021312

ARTICLE
pubs.acs.org/JACS
Toward Functional Carboxylate-Bridged Diiron Protein Mimics:
Achieving Structural Stability and Conformational Flexibility Using
a Macrocylic Ligand Framework
Loi H. Do and Stephen J. Lippard*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
S Supporting Information
b
ABSTRACT: A dinucleating macrocycle, H2PIM, containing
phenoxylimine metal-binding units has been prepared. Reaction
of H2PIM with [Fe2(Mes)4] (Mes = 2,4,6-trimethylphenyl) and
sterically hindered carboxylic acids, Ph3CCO2H or ArTolCO2H
(2,6-bis(p-tolyl)benzoic acid), afforded complexes [Fe2(PIM)-
(Ph3CCO2)2] (1) and [Fe2(PIM)(ArTolCO2)2] (2), respec-
tively. X-ray diffraction studies revealed that these diiron(II)
complexes closely mimic the active site structures of the
hydroxylase components of bacterial multicomponent mono-
oxygenases (BMMs), particularly the syn disposition of the
nitrogen donor atoms and the bridging μ-η1η2 and μ-η1η1 modes of the carboxylate ligands at the diiron(II) centers. Cyclic
voltammograms of 1 and 2 displayed quasi-reversible redox couples at þ16 and þ108 mV vs ferrocene/ferrocenium, respectively.
Treatment of 2 with silver perchlorate afforded a silver(I)/iron(III) heterodimetallic complex, [Fe2(μ-OH)2(ClO4)2(PIM)-
(ArTolCO2)Ag] (3), which was structurally and spectroscopically characterized. Complexes 1 and 2 both react rapidly with
dioxygen. Oxygenation of 1 afforded a (μ-hydroxo)diiron(III) complex [Fe2(μ-OH)(PIM)(Ph3CCO2)3] (4), a hexa(μ-hydroxo)-
tetrairon(III) complex [Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] (5), and an unidentified iron(III) species. Oxygenation of 2 exclusively
formed di(carboxylato)diiron(III) compounds, a testimony to the role of the macrocylic ligand in preserving the dinuclear iron
center under oxidizing conditions. X-ray crystallographic and 57Fe M€ossbauer spectroscopic investigations indicated that 2 reacts
with dioxygen to give a mixture of (μ-oxo)diiron(III) [Fe2(μ-O)(PIM)(ArTolCO2)2] (6) and di(μ-hydroxo)diiron(III)
[Fe2(μ-OH)2(PIM)(ArTolCO2)2] (7) units in the same crystal lattice. Compounds 6 and 7 spontaneously convert to a
tetrairon(III) complex, [Fe4(μ-OH)6(PIM)2(ArTolCO2)2] (8), when treated with excess H2O.
’ INTRODUCTION
and geometric arrangement of ligands at the diiron(II) center,
the model complex must be structurally stable yet conformationally
flexible upon reaction with O2 and other substrates. In the
enzyme, this balance is achieved by a rigid active site structure
that enforces the dinuclearity of the metal centers but allows
shifting of the carboxylate ligations to accommodate entry of
guest molecules.11,29 The most common strategy for assembling
discrete diiron compounds is to react Fe(II) or Fe(III) salts with
simple polydentate ligands.25,30 Although some of these com-
plexes are excellent structural and spectroscopic models for the
biological cofactor,31ꢀ34 they either exhibit poor dioxygen re-
activity or do not give rise to reactive oxygenated species.35,36
Given the extent of today’s synthetic toolbox, it should be possible to
prepare more sophisticated synthetic compounds that can mimic
various functional aspects of the biological diiron unit.26
Bacterial multicomponent monooxygenases (BMMs) are a
remarkable family of enzymes that catalyze the oxidation
of aliphatic and aromatic hydrocarbons using the naturally
abundant molecule O2.1,2 They include soluble methane
monooxygenase (sMMO),3 toluene/o-xylene monooxygenase
(ToMO),4ꢀ6 phenol hydroxylase (PH),7 and alkene monoox-
ygenase (AMO),8 among others.9,10 Extensive structural
studies revealed that the active sites of these enzymes contain a
non-heme diiron center located within a hydrophobic four-
helix bundle. The reduced form of sMMO hydroxylase
(sMMOHred) has a diiron(II) core coordinated by four carbox-
ylate amino acid side chains, two bridging and two terminal, and
two imidazole groups syn with respect to the FeꢀFe vector
(Scheme 1).11ꢀ13 This carboxylate-bridged diiron motif is ubi-
quitous in biology and is believed to be an essential component in
several mammalian proteins.14ꢀ16 Understanding the chemistry
that occurs at these bioinorganic cofactors has far-reaching
implications for human health as well as for environmental
remediation17ꢀ19 and chemical synthesis.20ꢀ23
In a significant step toward achieving this objective, we
employed a de novo approach to construct a diiron model
complex that mimics features of the BMM active sites. The
“primary” ligand must support a diiron core in multiple oxidation
Developing a functional model of the BMM active site is a
Received: March 8, 2011
Published: June 17, 2011
formidable challenge.24ꢀ28 In addition to matching the identity
r
2011 American Chemical Society
10568
dx.doi.org/10.1021/ja2021312 J. Am. Chem. Soc. 2011, 133, 10568–10581
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Reaction of the Diiron(II) Complex of Soluble
Methane Monooxygenase Hydroxylase (sMMOHred) with
Dioxygen and Saturated Hydrocarbons (RH) To Give Alco-
hols (ROH) and the Resting Diiron(III) State of the Protein
(sMMOHox)a
procedures for compounds AꢀD are provided in the SI. The com-
pounds 2,6-bis(p-tolyl)benzoic acid41 (ArTolCO2H, also referred to as
terphenylcarboxylic acid for simplicity) and [Fe2(Mes)4]42 (Mes =
2,4,6-trimethylphenyl) were prepared as reported. All 57Fe-enriched
compounds were prepared exactly as described for the unenriched
analogues, except that [57Fe2(Mes)4] was used as the starting material.
All air-sensitive manipulations were performed using standard Schlenk
techniques or under a nitrogen atmosphere inside an MBraun drybox.
Solvents were saturated with argon and purified by passage through two
columns of activated alumina. Dioxygen gas used in these experiments
was obtained from a high-purity gas cylinder (Airgas) and passed
through a 10-in. column of activated alumina before use.
General Physical Methods. NMR spectra were recorded on
500 MHz Varian Mercury spectrometers, and chemical shifts for 1H and
13C (proton-decoupled) NMR spectra were referenced to residual
solvent. 1H NMR spectral data of paramagnetic compounds were
obtained by widening the sweep window (þ100 to ꢀ30 ppm) and
collecting for longer acquisition times (∼1024 scans). IR spectra were
recorded on a ThermoNicolet Avatar 360 spectrophotometer with the
OMNIC software. Absorption spectra were recorded on a Cary 50
spectrophotometer using 6Q Spectrosil quartz cuvettes (Starna) with
1 cm path lengths. X-band EPR spectra were recorded at 5 K on a Bruker
EMX spectrometer. Electrochemical measurements were performed
with a VersaSTAT3 Princeton Applied Research potentiostat running
the V3-Studio electrochemical analysis software. A three-electrode setup
was employed comprising a platinum working electrode, a platinum wire
auxiliary electrode, and a 0.1 M Ag/AgNO3 solution in acetonitrile as the
reference electrode. Tetra-n-butylammonium hexafluorophosphate
(0.1 M) was used as the supporting electrolyte. Electrochemical potentials
are referenced externally to the ferrocene/ferrocenium couple at 0.00 V.
X-ray Data Collection and Refinement. Single crystals were
mounted in Paratone oil using 30 μm aperture MiTeGen MicroMounts
(Ithaca, NY) and frozen under a 100 K KRYO-FLEX nitrogen cold
stream. Data were collected on a Bruker SMART APEX CCD X-ray dif-
fractometer with Mo KR radiation (λ = 0.71073 Å) controlled by the
APEX 2 (v. 2010.1-2) software package. Data reduction was performed
using SAINT and empirical absorption corrections were applied using
SADABS.43 The structures were solved by Patterson methods with
refinement by full-matrix least squares based on F2 using the SHELXTL-
97 software package44 and checked for higher symmetry by the PLATON
software.45 All non-hydrogen atoms were located and refined anisotro-
pically. Hydrogen atoms were fixed to idealized positions unless other-
wise noted and given thermal parameters equal to either 1.5 (methyl
hydrogen atoms) or 1.2 (non-methyl hydrogen atoms) times the
thermal parameters of the atoms to which they are attached. Additional
X-ray crystallographic details are provided in the Supporting Informa-
a sMMOHox can be reduced back to sMMOHred by acquiring two
electrons from a reductase protein (sMMOR). The active site structures
of sMMOHred and sMMOHox are depicted.
Chart 1. Syn N-Donor Ligands Containing Mixed N,O Metal
Binding Unitsa
aCompounds H2LMe,Ph and H2BIPS spontaneously assemble into bis-
(ligand) diiron complexes in the presence of base and iron(II) salts. The
macrocyclic variant H2PIM was designed to have a flexible ether linkage
to prevent ligand interdigitation upon metal complexation.
states, enforce syn N-stereochemistry of nitrogen donors, have two
anionic oxygen atom donors, allow binding of two external
carboxylates to the iron centers, have an internal cavity to form
a quadrilateral Fe2(μ-O)2 core, and be readily synthesized in
gram quantities. Our initial studies using covalently linked N-
donor ligands were promising,37,38 but the diiron(II) complexes
prepared either were too kinetically labile or contained two
interdigitated syn N-donor units rather than one.39,40 We over-
came the previously encountered difficulties by designing a new
macrocylic ligand, H2PIM, containing phenoxylimine metal
binding groups (Chart 1). We used H2PIM and sterically
hindered carboxylates to synthesize structural models of the
carboxylate-bridged diiron active sites of the BMMs and related
enzymes. We show that PIM2ꢀ, the doubly deprotonated form of
H2PIM, has all the characteristics that were engineered into the
ligand design and imparts stability to the diiron unit, even
following reactions with O2. Although a functional protein mimic
has not yet been achieved, this work demonstrates that more
advanced diiron model complexes can be prepared using tailor-
made macrocyclic ligand platforms, thus providing a new con-
ceptual framework with which to guide future biomimetic
studies.
tion (SI).
57
€
Fe Mossbauer Spectroscopy. M€ossbauer spectra were re-
corded on an MSI spectrometer (WEB Research Co.) with a 57Co
source in a Rh matrix maintained at room temperature. Solid samples
were prepared by suspension of the complex (∼5ꢀ40 mg, depending on
whether the complex is enriched in 57Fe) in Apiezon M grease and
placed in a nylon sample holder. Solution samples were prepared by
freezing 400 μL of the complex (∼20 mM) in a nylon sample cup and
sealing with a screw cap. Samples containing natural abundance iron
were measured over the course of ∼5 d, whereas samples that are
enriched in 57Fe were collected over ∼12 h. Data were acquired at 80 K,
and isomer shift (δ) values are reported with respect to metallic iron that
was used for velocity calibration at room temperature. Spectra were fit to
Lorentzian lines using the WMOSS plot-and-fit program.
Syntheses. H2PIM. Compound D (1.70 g, 3.65 mmol) was dis-
solved in 10 mL of dichloromethane and combined with 3,30-diamino-
diphenylsulfone (0.900 g, 3.65 mmol) in 700 mL of dry acetonitrile. About
1 mL of trifluoroacetic acid was added, and the mixture was refluxed for
’ EXPERIMENTAL SECTION
Materials and Methods. Reagents obtained from Strem, Aldrich
Chemical Co., and Alfa Aesar were used as received. Synthetic
10569
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
6 h. Over the course of ∼1 h, a large amount of bright yellow-orange
material formed. The solid was isolated by filtration and washed with
diethyl ether to afford analytically pure product (1.80 g, 72%). 1H NMR
(CDCl3, 500 MHz): δ 13.16 (s, 2H), 8.63 (s, 2H), 7.89 (d, J = 9.5 Hz,
2H), 7.80 (m, 2H), 7.73 (s, 2H), 7.67 (t, J = 8.0 Hz, 2H), 7.44 (m, 4H),
7.39 (m, 4H), 7.28 (m, 2H), 7.21 (s, 2H), 4.69 (s, 4H), 2.36 (s, 6H) ppm.
13C NMR (CDCl3, 125 MHz): δ 164.78, 156.69, 149.82, 143.08, 138.09,
137.72, 136.08, 132.20, 130.72, 130.32, 129.57, 128.69, 128.51, 128.30,
127.06, 125.30, 123.73, 123.59, 118.85, 73.03, 20.61 ppm. IR (KBr): ν
2852, 1620, 1579, 1472, 1454, 1427, 1326, 1302, 1211, 1148, 1092, 886,
783, 700, 689, 620, 524 cmꢀ1. DART-MS(ꢀ) = 677.2121 (calcd =
677.2116 [MꢀH]ꢀ). Mp = 313ꢀ315 °C.
479 (sh, 3300 Mꢀ1 cmꢀ1), 600 (2300 Mꢀ1 cmꢀ1). Anal. Calcd for
Fe2C63H51AgCl2N2O17S (CH2Cl2) (3 CH2Cl2): C, 50.72; H, 3.52;
3
3
N, 1.85; Cl, 9.36. Found: C, 51.04; H, 3.64; N, 1.77; Cl, 9.23. Mp
(decomp) ≈ 270 °C. M€ossbauer (THF): δ = 0.49(2) mm/s, ΔEQ
=
1.38(2) mm/s, ΓL/R = 0.56(2) mm/s.
[Fe2(μ-OH)(PIM)(Ph3CCO2)3] (4). In an anaerobic drybox, solid
[Fe2(PIM)(Ph3CCO2)2] (1) (30 mg, 22 μmol), triphenylacetic acid
(13 mg, 44 μmol), and triethylamine (6.0 μL) were combined in 2.0 mL
of benzene. The reaction vial was sealed with a septum, brought outside
of the glovebox, and bubbled with dioxygen for 5 min. When exposed to
O2, the clear red solution quickly turned dark brown. Upon vapor dif-
fusion of pentane into the benzene solution, several large X-ray diffrac-
tion quality crystals were obtained. The product was isolated by filtration
(8 mg, 22%). 1H NMR (CDCl3, 500 MHz): δ 51.28, 44.24, 33.02, 29.70,
23.90, 6.60, ꢀ4.52, ꢀ7.85 ppm. IR (KBr): ν 3438, 3056, 3032, 2919,
1582, 1544, 1491, 1477, 1446, 1405, 1373, 1331, 1205, 1150, 1035,
827, 789, 744, 699, 677, 642, 604, 547, 523 cmꢀ1. UVꢀvis (CH2Cl2):
λmax = 284 (108 000 Mꢀ1 cmꢀ1), 372 (38 000 Mꢀ1 cmꢀ1), and 570
[Fe2(PIM)(Ph3CCO2)2] (1). In an anaerobic drybox, solid H2PIM
(100 mg, 147 μmol) and triphenylacetic acid (85.0 mg, 295 μmol) were
dissolved in 2.0 mL of tetrahydrofuran. A 1.0 mL solution of tetrahydrofuran
containing [Fe2(Mes)4] (86.0 mg, 147 μmol) was added to the reaction
vessel, and the mixture was stirred for 1 h. The dark red solution was
evaporated to dryness, and the residue was redissolved in benzene. The
solution was filtered through a glass wool plug and layered with pentane.
After ∼12 h, a large amount of red crystals had formed that were suitable
for X-ray diffraction analysis. The solid material was isolated by filtration
and washed with pentane to give the desired diiron complex (171 mg,
86%). 1H NMR (C6D6, 500 MHz): δ 78.37, 74.97, 48.27, 22.78,
8.20ꢀ1.19, ꢀ5.32, ꢀ6.20, ꢀ10.37, ꢀ13.49, ꢀ21.13 ppm. IR (KBr):
ν 3056, 3031, 2918, 2850, 1578, 1545, 1490, 1444, 1418, 1378, 1322,
1304, 1284, 1200, 1175, 1152, 1097, 1085, 1036, 985, 790, 745, 699, 673,
547 cmꢀ1. UVꢀvis (CH2Cl2): λmax = 290 (36 300 Mꢀ1 cmꢀ1), 410
(10 300 Mꢀ1 cmꢀ1) nm. Anal. Calcd for Fe2C102H78N2O12S (C6H6)-
3
(NC6H15)2 (4 C6H6 (NEt3)2): C, 73.99; H, 5.90; N, 2.88. Found: C,
3
3
74.40; H, 5.66; N, 3.22. Mp = 214ꢀ217 °C. M€ossbauer (polycrystalline,
apiezon M grease): δ = 0.52(2) mm/s, ΔEQ = 0.95(2) mm/s, ΓL/R
=
0.38(2) mm/s.
[Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] (5). In an anaerobic drybox, solid
[Fe2(PIM)(Ph3CCO2)2] (1) (40 mg, 29 μmol) was combined with
2.0 mL of benzene in a septum-sealed reaction vessel. Dioxygen was
bubbled through the mixture for 5 min. The solution was concentrated
to approximately half its volume and then left alone to crystallize over
∼1 d by slow evaporation. This method resulted in selective crystal-
lization of the desired tetranuclear [Fe4(μ-OH)6(PIM)2(Ph3CCO2)2]
(5) (hexagonal prisms) instead of the dinuclear [Fe2(μ-OH)(PIM)-
(Ph3CCO2)3] (4) (rectangular blocks). Crystals obtained from this
method were suitable for X-ray diffraction studies. The desired product
was isolated by filtration (10 mg, 32%). 1H NMR (CDCl3, 500 MHz):
δ 49.97, 43.69, 31.24, 15.60, 8.71, 5.94, 5.31, 4.14, ꢀ7.12 ppm. IR (NaCl):
ν 3554, 3445, 3064, 2961, 2921, 2850, 1580, 1542, 1381, 1340, 1324, 1260,
1238, 1204, 1151, 1101, 1080, 1017, 984, 800, 789, 747, 738, 705, 692, 671,
633, 528, 416 cmꢀ1. UVꢀvis (CH2Cl2): λmax = 280 (84 200 Mꢀ1 cmꢀ1),
370 (25 000 Mꢀ1 cmꢀ1), and 540 (8060 Mꢀ1 cmꢀ1) nm. Anal. Calcd for
Fe4C124H100N4O20S2 (5): C, 66.09; H, 4.47; N, 2.49. Found: C, 66.09;
H, 4.50; N, 2.14. Mp (decomp) > 300 °C. M€ossbauer (polycrystalline,
(16 000 Mꢀ1 cmꢀ1) nm. Anal. Calcd for Fe2C82H62N2O9S (C4H8O)
3
(1 THF): C, 71.97; H, 4.92; N, 1.95. Found: C, 71.80; H, 4.94; N, 2.22.
3
Mp (decomp) = 302 °C. M€ossbauer (polycrystalline, apiezon M grease):
δ1 = 1.18(2) mm/s, ΔEQ1 = 2.33(2) mm/s, ΓL/R(1) = 0.38(2) mm/s,
Site 1 Area = 53%; δ2 = 0.97(2) mm/s, ΔEQ2 = 2.25(2) mm/s, ΓL/R(2)
=
0.35(2) mm/s, Site 2 Area = 47%.
[Fe2(PIM)(ArTolCO2)2] (2). The synthesis of [Fe2(PIM)(ArTolCO2)2]
was performed as described for [Fe2(PIM)(Ph3CCO2)2] (1), except
that ArTolCO2H was used instead of Ph3CCO2H. The product crystal-
lized upon slow diffusion of pentane into a solution of the complex in
dichloromethane and was isolated as a red powder when dried (152 mg,
74%). X-ray diffraction quality crystals were readily obtained by slow
diffusion of pentane into a dichloromethane solution of the complex. 1H
NMR (CD2Cl2, 500 MHz): δ 76.92, 73.82, 49.00, 24.68, 7.86, 7.31, 6.22,
2.41, 1.37, 0.96, ꢀ0.41, ꢀ1.81, ꢀ5.72, ꢀ9.04, ꢀ19.60 ppm. IR (KBr):
ν 3052, 3024, 2912, 2851, 1614, 1577, 1533, 1445, 1413, 1381, 1302,
1286, 1202, 1149, 1073, 819, 791, 712, 695, 542 cmꢀ1. UVꢀvis (CH2Cl2):
λmax = 290 (36 700 Mꢀ1 cmꢀ1), 418 (14 000 Mꢀ1 cmꢀ1) nm. Anal.
Calcd for Fe2C84H66N2O9S (C4H8O) (2 THF): C, 72.23; H, 5.10; N,
apiezon M grease): δ = 0.51(2) mm/s, ΔEQ = 1.06(2) mm/s, ΓL/R
=
0.40(2) mm/s.
[Fe2(μ-O)(PIM)(ArTolCO2)2] (6)/[Fe2(μ-OH)2(PIM)(ArTolCO2)2] (7).
In an anaerobic drybox, [Fe2(PIM)(ArTolCO2)2] (2) (200 mg, 144 μmol)
was dissolved in 2 mL of tetrahydrofuran in a 20 mL vial. The vial was
sealed with a rubber septum and brought outside of the drybox. Dioxygen
was bubbled through the reaction vial for 5 min, at which time the red
solution became a dark brown. The tetrahydrofuran solvent was removed in
vacuo, and the resulting brown solid (117 mg, 58%) was isolated and
stored inside the drybox to prevent further reaction with ambient
moisture. Single crystals suitable for X-ray diffraction analysis were
obtained by slow diffusion of diethyl ether into a solution of the complex
3
3
1.91. Found: C, 71.72; H, 4.92; N, 2.17. Mp (decomp) = 240 °C.
M€ossbauer (polycrystalline, apiezon M grease): δ1 = 1.10(2) mm/s,
ΔEQ1 = 2.04(2) mm/s, ΓL/R(1) = 0.38(2) mm/s, Site 1 Area = 63%; δ2 =
0.95(2) mm/s, ΔEQ2 = 2.02(2) mm/s, ΓL/R(2) = 0.32(2) mm/s, Site 2
Area = 37%.
[Fe2(μ-OH)2(ClO4)2(PIM)(ArTolCO2)Ag] (3). In an anaerobic drybox,
[Fe2(PIM)(ArTolCO2)2] (2) (19 mg, 14 μmol) and silver perchlorate
(10 mg, 48 μmol) were combined in 2 mL of dichloromethane and
stirred for 1 h. The reaction mixture was filtered through a glass wool
plug to remove a black precipitate. The solution was concentrated to half
its volume and layered with 0.5 mL of pentane. After ∼14 h, a dark solid
formed on the bottom of the reaction vial. This product was isolated by
filtration, yielding a dark crystalline solid (15 mg, 76%). Crystals suitable
for X-ray diffraction studies were obtained by slow diffusion of pentane
into a solution of the complex in dichloromethane. IR (KBr): ν 3432,
2919, 2851, 1579, 1546, 1521, 1472, 1446, 1382, 1337, 1306, 1285,
1200, 1181, 1152, 1122, 1020, 796, 603, 548, 530 cmꢀ1. UVꢀvis
(CH2Cl2): λmax = 290 (32 300 Mꢀ1 cmꢀ1), 375 (11 100 Mꢀ1 cmꢀ1),
1
in acetonitrile. H NMR (CDCl3, 500 MHz): δ 26.30, 24.44, 21.84,
13.52, 15.92, 14.77, 11.56, 10.08, 9.58, 8.22, 5.44, 4.86, 4.26, 2.16, 1.80,
1.25, 0.88, ꢀ7.11 ppm. IR (KBr): ν 3448, 3021, 2918, 2856, 1613, 1582,
1545, 1515, 1473, 1448, 1383, 1336, 1304, 1286, 1205, 1152, 818, 791,
766, 700, 546, 531 cmꢀ1. UVꢀvis (CH2Cl2): λmax = 380 (sh), 473,
600 nm. Anal. Calcd for Fe2C84H68N2O11S (CH3CN) (6 CH3CN):
3
3
C, 70.45; H, 4.88; N, 2.87. Anal. Calcd for Fe2C84H66N2O10S
3
(CH3CN) (7 CH3CN): C, 71.32; H, 4.80; N, 2.90. Found: C,
3
69.89; H, 4.76; N, 2.59. As discussed in the text, compounds 6
and 7 exist as a mixture in the solid-state. Mp (decomp) > 300 °C.
M€ossbauer (polycrystalline, apiezon M grease): δ1 = 0.47(2) mm/s,
10570
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
ΔEQ1 = 1.52(2) mm/s, ΓL/R(1) = 0.36(2) mm/s, Site 1 Area = 21%;
δ2 = 0.50(2) mm/s, ΔEQ2 = 0.97(2) mm/s, ΓL/R(2) = 0.48(2) mm/s,
Site 2 Area = 79%.
Scheme 2a
[Fe4(μ-OH)6(PIM)2(ArTolCO2)2] (8). Solid [Fe2(PIM)(ArTolCO2)2]
(2) (200 mg, 144 μmol) was dissolved in 2.0 mL of benzene and stirred
under dioxygen for 5 min, giving a dark brown solution. About 10 μL of
water was then added to the reaction mixture, and the solution was
stirred for an additional 10 min and then filtered through a glass wool
plug before the filtrate was evaporated to dryness. The remaining solid
was isolated as a pale brown material (∼120 mg, 73%). 1H NMR (DMF-d7,
500 MHz): δ 72.80, 45.00, 26.24, 14.48, 13.26, 10.40, 7.30ꢀ0.12, ꢀ3.16,
ꢀ4.65, ꢀ11.36 ppm. IR (KBr): ν 3432, 3019, 2919, 2857, 1614, 1584,
1543, 1516, 1473, 1450, 1385, 1324, 1304, 1286, 1206, 1151, 791, 703,
586, 528 cmꢀ1. UVꢀvis (CH2Cl2): λmax = 360 (81 700 Mꢀ1 cmꢀ1), 550
(17 100 Mꢀ1 cmꢀ1) nm. Anal. Calcd for Fe4C126H104N4O20S2
3
(C6H6)2 (8 (C6H6)2): C, 67.99; H, 4.80; N, 2.30. Found: C, 67.92;
3
H, 4.80; N, 2.43. Mp (decomp) > 300 °C. M€ossbauer (THF): δ =
0.49(2) mm/s, ΔEQ = 0.97(2) mm/s, ΓL/R = 0.44(2) mm/s.
’ RESULTS AND DISCUSSION
a Reagents and conditions: i. a) NaH, dry THF, b) 3-bromobenzylbro-
mide. ii. a) Aryl zinc reagent: 2-(2-bromo-4-methylphenoxy)-tetrahy-
dro-2H-pyran, n-butyllithium, ZnCl2, THF, b) Pd(PPh3)4. iii. oxalic
acid, THF/MeOH (1:1), 50 °C. iv. anhydrous MgCl2, paraformalde-
hyde, NEt3, CH3CN, reflux. v. 3,30-diaminodiphenylsulfone, trifluoroa-
cetic acid, dry CH3CN.
Ligand Design and Synthesis. Our early attempts to prepare
more accurate structural mimics of carboxylate-bridged diiron
protein active sites led to the design of dinucleating ligands that
enforce the syn stereochemistry of nitrogen donor atoms.37,38
Studies using 1,2-bis(3-ethynyl-8-carboxylatequinoline)-4,5-diethyl-
benzene ethyl ester (Et2BCQEBEt) afforded a tri(μ-carboxylato)-
(syn N-donor)diiron(II) complex, which produced only an in-
tractable mixture upon exposure to dioxygen. Other syn N-donor
variants, such as H2LMe,Ph and H2BIPS (Chart 1), primarily afforded
bis(ligand)diiron complexes formed through interdigitation of two
dinucleating ligands when complexed with iron(II).39,40
catalyzed Negishi cross-coupling procedure was employed.
After silica gel column chromatography, B was isolated as a
colorless oil. Next, the tetrahydropyran protecting group of B was
removed by treatment with oxalic acid in tetrahydrofuran/methanol
to afford C as a white solid. Ortho formylation of C was achieved by
treatment with anhydrous magnesium chloride, paraformaldehyde,
and triethylamine to give D as a yellow oil. To obtain H2PIM, D was
condensed with 3,30-diaminodiphenylsulfone, generating the de-
sired product as a bright yellow-orange solid in good yield. The
synthetic route for H2PIM is amenable to scale-up, and the final
ligand has been obtained in multigram quantities.
To improve upon previous ligand designs, the macrocyclic
compound H2PIM was conceived (Chart 1). By linking the
phenyl groups of H2BIPS with a three-atom chain, H2PIM still
enforces syn arrangement of the nitrogen donors but has a more
preorganized structure. We used phenolate and imine moieties
instead of carboxylate and imidazole groups, respectively, be-
cause they confer both synthetic and functional advantages. Phenol
is a versatile building block for constructing complex molecular
architectures because its aromatic ring can be readily attached to
other substituents.46 Although the phenolate anion (pKa of
phenol ≈ 12ꢀ19 in DMSO) is more basic than a carboxylate
(pKa of carboxylic acid ≈ 9ꢀ13 in DMSO),47,48 more electron-
rich oxygen donors may stabilize iron in higher oxidation states,
such as occurs in the diiron(IV) unit of intermediate Q.49 Alter-
natively, the basicity of the phenolate groups can be lowered to
better match the donor strength of Glu/Asp side chains by
introducing electron-withdrawing substituents on PIM2ꢀ. Tun-
ing the pKa value of the phenolate may also eliminate the pos-
sibility of forming ligand-centered radicals,50 a potential barrier
to accessing high-valent diiron species. An additional advantage
of the phenolate ligand is that its iron complexes display visible
absorption bands that provide a spectroscopic handle for study-
ing otherwise optically silent species. Finally, imine groups are
valuable in the H2PIM framework because they form through
Schiff base condensation reactions that are efficient in producing
large macrocyclic structures with syn nitrogen donor atoms.51,52
The H2PIM ligand was prepared as indicated in Scheme 2. The
dibenzyl ether linker A was obtained by reaction of 3-bromo-
benzyl alcohol with sodium hydride, followed by refluxing with
3-bromobenzyl bromide. To prepare compound B, a palladium-
Assembly of Diiron(II) Complexes. The phenolate-to-iron
charge-transfer band is a useful spectroscopic probe for quanti-
tating metalꢀligand binding.40,53,54 A previous study demon-
strated that the non-macrocyclic syn N-donors L2ꢀ and BIPS2ꢀ
,
the doubly deprotonated forms of H2LMe,Ph and H2BIPS (Chart 1),
respectively, react with iron in a 1:1 ratio to form [Fe2(syn N-
donor)2] species. To test whether PIM2ꢀ exhibits different metal
binding behavior, iron(II) titration experiments were conducted.
When H2PIM was treated with 2 equiv of sodium hexamethyldisi-
lazide (NaHMDS) in tetrahydrofuran, optical bands at 240, 280,
and 420 nm appeared, corresponding to formation of PIM2ꢀ
.
Addition of iron(II) triflate to the PIM2ꢀ solution decreased the
intensity of the 240 and 420 nm bands, with concomitant absorption
increases at 280 and 375 nm (SI, Figure S1A). A plot of the
absorbance changes at 375 nm revealed an Fe-to-PIM2ꢀ ratio of
2:1 (Figure S1B), consistent with binding of two iron atoms by
each macrocycle to give [Fe2(PIM)(SO3CF3)2]. To examine
further the nature of this diiron species, titration experiments
were conducted using external carboxylates. Addition of sodium
triphenylacetate to a solution containing [Fe2(PIM)(SO3CF3)2]
generated new absorption features at 410 and 600 nm (Figure
S1C). A plot of the absorbance changes at 410 nm indicated that
carboxylates readily displace the triflate anions (Figure S1D).
10571
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Figure 1. Stick figure representation of the X-ray crystal structure of
[Fe2(PIM)(Ph3CCO2)2] (1). Solvent molecules and hydrogen atoms
are omitted for clarity. Color scheme: iron, orange; carbon, gray; nitrogen,
blue; oxygen, red; sulfur, yellow. Selected bond distances (Å) and angles
(deg): Fe(1)ꢀFe(2) = 3.6100(6); Fe(1)ꢀN(1) = 2.049(2); Fe(1)ꢀO-
(2) = 1.878(2); Fe(1)ꢀO(7) = 1.999(2); Fe(1)ꢀO(8) = 2.016(2);
Fe(2)ꢀN(2) = 2.039(2); Fe(2)ꢀO(1) = 1.892(2); Fe(2)ꢀO(7) =
2.254(2); Fe(2)ꢀO(6) = 2.153(2); Fe(2)ꢀO(9) = 2.024(2); O(2)ꢀ
Fe(1)ꢀN(1) = 92.49(8); O(1)ꢀFe(2)ꢀN(2) = 92.38(8); O(7)ꢀ
Fe(1)ꢀO(8) = 103.72(8); O(7)ꢀFe(2)ꢀO(9) = 89.14(7). A thermal
ellipsoid ORTEP diagram of 1 is provided in SI Figure S13.
Figure 3. Depiction of the X-ray crystal structures of the diiron sites of
sMMOHred (top, left) and sMMOHox (bottom, left). For structural
comparison, the synthetic complexes that mimic each protein state are
shown on its right. Some relevant bond lengths are provided; the
distances shown for 1 and 2 are averaged over the two complexes.
Color scheme: iron, orange; nitrogen, blue; oxygen, red; carbon, gray.
carboxylic acid, triphenylacetic acid or terphenylcarboxylic acid.
A diiron(II) unit containing triphenylacetate was crystallized by
slow diffusion of pentane into a solution of the compound in
benzene. X-ray diffraction studies revealed a dinuclear complex
havingthemolecularformula[Fe2(PIM)(Ph3CCO2)2](1,Figure1,
SI Table S1). The PIM2ꢀ ligand coordinates to two iron atoms
that are bridged by two triphenylacetates, one of which is bound
in an η1,η1-1,3 coordination mode and the other in an η1,η2-1,3
arrangement. The FeꢀO(carboxylate) distances range from
2.00 to 2.25 Å. The four- and five-coordinate iron atoms are
separated by 3.61 Å, with average FeꢀO(phenolate) and Feꢀ
N(imine) distances of 1.88 and 2.04 Å, respectivelyꢀ.
When the terphenylcarboxylate anion ArTolCO2 was em-
ployed in an analogous synthetic procedure, crystallization by
slow diffusion of pentane into a saturated dichloromethane
solution of the compound yielded [Fe2(PIM)(ArTolCO2)2] (2,
Figure 2, Table S1). The X-ray diffraction analysis shows that the
dinuclear core in 2 closely resembles that in 1. The iron atoms in
2 are bridged by two terphenylcarboxylate ligands, with Feꢀ
O(carboxylate) distances of 2.00ꢀ2.34 Å. This diiron unit is
supported by the syn N-donor macrocycle, resulting in Feꢀ
O(phenolate) and FeꢀN(imine) distances of 1.89 and 2.04 Å,
respectively. The FeꢀFe distance in 2 is 3.61 Å, the same as in 1.
To a first approximation, the structural parameters of the
primary coordination spheres of 1 and 2 compare favorably to
those of the diiron core in sMMOHred (Figure 3). The most
notable difference is in the FeꢀFe bond distance, 3.43 Å in the
protein and 3.61 Å in the model compounds. Differences in
metalꢀmetal distance can arise from several factors, such as the
donor strength of supporting ligands, constraints of the ligand
backbone, and orientation of the bridging carboxylates.
Figure 2. Stick figure representation of the X-ray crystal structure of
[Fe2(PIM)(ArTolCO2)2] (2). Solvent molecules and hydrogen atoms
are omitted for clarity. Color scheme: iron, orange; carbon, gray;
nitrogen, blue; oxygen, red; sulfur, yellow. Selected bond distances
(Å) andangles (deg): Fe(1)ꢀFe(2) = 3.607(1); Fe(1)ꢀO(1) =1.888(4);
Fe(1)ꢀN(1) = 2.037(4); Fe(1)ꢀO(6) = 1.998(3); Fe(1)ꢀO(8) =
2.028(3); Fe(2)ꢀO(2) = 1.894(3); Fe(2)ꢀN(2) = 2.045(4); Fe(2)ꢀ
O(7) = 2.048(4); Fe(2)ꢀO(8) = 2.092(3); Fe(2)ꢀO(9) = 2.342(4);
O(1)ꢀFe(1)ꢀN(1) = 93.1(2); O(2)ꢀFe(2)ꢀN(2) = 91.8(2); O(6)ꢀ
Fe(1)ꢀO(8) = 102.1(1); O(7)ꢀFe(2)ꢀO(8) = 89.5(1). A thermal
ellipsoid ORTEP diagram of 2 is provided in SI Figure S14.
Compounds 1 and 2 were further characterized by several
spectroscopic methods (Table 1). As shown in Figure 4, they
have strong absorption bands in the UVꢀvisible region, with
λmax values of approximately 290 and 410 nm. The higher energy
band is attributed to a πꢀπ* ligand transition, whereas the lower
These data suggested that H2PIM supports a diiron(II) structure
with carboxylate ligands.
Preparative-scale reactions were performed to metalate
H2PIM by treatment with [Fe2(Mes)4] and a sterically hindered
10572
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Table 1. Characterization Data for 1ꢀ8
opticala
M€ossbauerb
δ(Fe2)
X-ray
λ, nm
δ(Fe1)
ΔEQ(Fe1)
Γ(Fe1)
Area(Fe1)
[%]
ΔEQ(Fe2)
Γ(Fe2)
Area(Fe2)
[%]
FeꢀFe
(Å)
[ε, Mꢀ1 cmꢀ1
]
[mm/s]
[mm/s]
[mm/s]
[mm/s]
[mm/s]
[mm/s]
1
2
3
290 (36 300)
410 (16 000)
290 (36 700)
418 (14 000)
290 (32 300)
375(11 100)
479 (3300, sh)
600 (2300)
284 (108 000)
372 (38 000)
570 (10 300)
370 (25 000)
540 (8060)
380 (sh)
1.18
2.33
0.38
53
0.97
2.25
0.35
47
3.61
3.61
3.01
1.10
2.04
0.38
63
0.95
1.15*
—
2.02
1.75*
—
0.32
0.49*
—
37
1.23*
0.49*
2.58*
1.38*
0.57*
0.56*
59*
100*
41*
—
4
0.52
0.51
0.95
1.06
0.38
0.40
100
100
—
—
—
—
—
—
—
—
3.49
5
3.06
3.78c
3.44
6/7
0.47
1.52
0.36
21
0.50
0.48d
0.49*
—
0.97
1.03d
0.78*
—
0.48
0.34d
0.44*
—
79
473
0.46d
0.51*
0.49*
1.59d
1.24*
0.97*
0.58d
0.62*
0.44*
85d
61*
100*
15d
36*
—
600
8
360 (81 700)
550 (17 100)
—
a Absorption spectra were recorded in dichloromethane. b M€ossbauer spectra were acquired at 80 K. Polycrystalline samples were prepared by mixing
with Apiezon M grease, and solution samples, which are marked with an asterisk (*), were prepared in tetrahydrofuran. c This value corresponds to the
FeꢀFe distance between two diiron units. d The polycrystalline sample of 6/7 was dried under vacuum at 150 °C for 24 h.
parameters δ1 = 1.10(2) mm/s, ΔEQ1 = 2.04(2) mm/s, δ2 =
0.95(2) mm/s, and ΔEQ2 = 2.02(2) mm/s (Table 1). The isomer
shift values of 1 and 2 (g∼1.0 mm/s) are typical of high-spin
iron(II) complexes, and the quadruople splitting parameters are
as expected for iron coordinated by O,N-donors.30,58 A compar-
ison of the M€ossbauer parameters of 1 and 2 to those reported for
sMMOHred indicates that there are relatively minor electronic
differences between the synthetic and biological complexes.
Although the X-ray structure of sMMOHred revealed a diiron
core with two different iron environments, the M€ossbauer data
were fit to a single-quadrupole doublet (δ1 = 1.3 mm/s, ΔEQ1
=
3.0 mm/s).59 The primary difference between 1 and 2 compared
to sMMOHred is the identity of several donor groups, phenolates
in place of carboxylates and imines in place of imidazoles. By
modifying the electronic properties of the PIM2ꢀ or carboxylate
ligands in 1 and 2, however, a better match of the M€ossbauer
parameters to those of the biological cofactor may be achieved.
To further evaluate the electronic structure, an EPR spectrum
of 1 in 2-methyltetrahydrofuran was recorded at 5 K (SI Figure
S3). The spectrum exhibits a broad signal with g ≈ 15, similar to
that recorded for sMMOHred, in which two S = 2 iron(II) centers
are weakly ferromagnetically coupled.60 Although typically lower
than 15, mononuclear high-spin iron(II) compounds can also
exhibit high g values,61,62 but we nonetheless attribute this signal
to a diiron(II) species. Compelling evidence for such an assign-
ment comes from (1) iron titration experiments, described in the
preceding section, indicating that 1 is dinuclear in solution, and
(2) the solid-state structure of 1, which reveals an FeꢀFe
separation short enough to support magnetic exchange coupling.
The EPR spectrum of 2 was not measured, but because its
Figure 4. Absorption spectra of 1 and 2 in dichloromethane. Both
compounds exhibit optical bands at ∼290 and 410 nm.
energy band is most likely due to ligand-to-metal charge transfer.55,56
In addition, there is a pronounced shoulder at ∼540 nm.
The zero-field M€ossbauer spectrum of polycrystalline 1 at
80 K displays two quadrupole doublets, with parameters δ1 =
1.18(2) mm/s, ΔEQ1 = 2.33(2) mm/s, δ2 = 0.97(2) mm/s, and
ΔEQ2 = 2.25(2) mm/s (Figure 5, Table 1). Each iron site
accounts for ∼50% of the total iron content in the sample.57
The M€ossbauer spectrum of polycrystalline 2 recorded at 80 K
also contains two quadrupole doublets, which were fit with
10573
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Figure 7. Cyclic voltammograms of 1.0 mM solutions of 1 (A) and 2
(B) in dichloromethane at various scan rates. Tetra-n-butylammonium
hexafluorophosphate (0.2 M) was used as the supporting electrolyte. All
data were obtained using a platinum working electrode, and electro-
chemical potentials are referenced externally to the ferrocene/ferroce-
nium couple. Quasi-reversible redox couples at þ16 and þ108 mV were
measured for 1 and 2, respectively, at a scan rate of 100 mV/s.
Figure 5. Zero-field 57Fe M€ossbauer spectrum of a polycrystalline
sample of [Fe2(PIM)(Ph3CCO2)] (1) at 80 K. The raw data (black
hash lines) were fit to two distinct iron sites (green trace). The single-site
fits are shown as blue and red traces using the following parameters: δ1 =
1.18(2) mm/s, ΔEQ1 = 2.33(2) mm/s, ΓL/R(1) = 0.38(2) mm/s; δ2 =
0.97(2) mm/s, ΔEQ2 = 2.25(2) mm/s, ΓL/R(2) = 0.35(2) mm/s.
for 2 correspond to the PIM2ꢀ ligand. Because dichloromethane
is a non-coordinating solvent, the solution structures of 1 and 2 in
CD21Cl2 are probably similar to those in the solid state. When
the H NMR spectrum of 2 was recorded in acetonitrile-d3
(Figure 6B), significant shifts in the proton peaks were observed
by comparison to the spectrum acquired in dichloromethane-d2.
Most notably, new resonances at 69.33, 55.70, 48.28, 45.03,
42.85, 39.54, 38.55, and 35.55 ppm were present. Because
acetonitrile is coordinating, it is possible that solvent molecules
bind to the iron centers and/or displace the carboxylate ligands.
These data suggest that the solution structures of 1 and 2 depend
on the coordinating abilities of the solvent.
Redox Chemistry of Diiron(II) Complexes. A characteristic
property of carboxylate-bridged diiron proteins is their tendency
to undergo single-electron-transfer reactions.60 For example,
treatment of sMMOHox, which contains a diiron(III) unit, with
sodium dithionite or γ radiation from a Co-60 source led to
formation of a mixed-valent diiron(II,III) species (Hmv). Because
1 and 2 are close structural models of sMMOHred, their electro-
chemical properties are of interest.
A cyclic voltammogram (CV) of 1 measured in dichloro-
methane using a platinum working electrode and (n-Bu4N)PF6
as supporting electrolyte revealed two electrochemical events
(Figure 7A and SI Figure S4A). At a scan rate of 100 mV/s, a
quasi-reversible redox couple at þ16 mV, with a relatively large
peak-to-peak separation of þ210 mV, and an irreversible oxida-
tion at þ840 mV were observed. The CV of 2 was recorded
under identical conditions (Figures 7B and S4B) and displayed a
similar voltammogram, showing a quasi-reversible oxidation at
þ108 mV (ΔEp = þ198 mV) and an irreversible oxidation at
þ600 mV. It is uncertain whether the quasi-reversible waves
involve a one- or two-electron process and whether the large
peak-to-peak separation is due to structural reorganization or
some other chemical process. Ligand-centered redox chemistry,
observed in some phenolate iron complexes,65ꢀ67 is also a pos-
sibility that cannot be ruled out at this time.
Figure 6. 1H NMR spectra (500 MHz) of ∼5 mM [Fe2(PIM)-
(ArTolCO2)2] (2) in solutions of (A) dichloromethane-d2 and (B)
acetonitrile-d3 at room temperature.
UVꢀvisible and M€ossbauer spectra are similar to those of 1, it
most likely shares the same electronic structure.
The paramagnetism of 1 and 2 was verified by NMR spec-
troscopy. 1H NMR spectra of 1 (SI Figure S2) and 2 (Figure 6A)
in dichloromethane-d2 display resonances ranging from approxi-
mately 80 to ꢀ20 ppm. Although these paramagnetically shifted
resonances cannot be assigned without more detailed studies,63,64 the
peaks at 78.37, 74.97, 48.27, 22.78, ꢀ5.32, ꢀ10.37, ꢀ21.13 ppm for
1 and at 76.92, 73.82, 49.00, 24.68, ꢀ5.72, ꢀ9.04, ꢀ19.60 ppm
The redox chemistry of complex 2 was further explored by
using chemical oxidants (Scheme 3). Because the Agþ/Ag0 redox
couple has an E1/2 value of þ650 mV (vs ferrocene/ferrocenium) in
dichloromethane,68 silver(I) reagents were expected to oxidize 2
(E1/2 = þ108 mV). Treatment of 2 with ∼1 equiv of AgClO4 in
10574
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Scheme 3. Summary of Reaction Products Characterized in This Studya
a Full representation of the PIM2ꢀ ligand is omitted for clarity; only its phenolate oxygen and imine nitrogen atoms are depicted. See Chart 1 for the
structure of H2PIM, the doubly protonated form of PIM2ꢀ
.
dichloromethane gradually converted the initial dark red solution
to a pale yellow brown solution. After the solution was stirred for
about 1 h, a black precipitate, presumed to be silver metal, for-
med. This insoluble material was removed by filtration, and the
absorption spectrum of the filtrate revealed formation of a new
product that exhibited optical bands at 290, 375, 479 (sh), and
600 nm (SI Figure S5). The filtrate was concentrated, and
pentane was introduced by diffusion, affording a small number
of X-ray diffraction quality crystals. Structural analysis of these
crystals revealed the product to be [Fe2(μ-OH)2(ClO4)2(PIM)-
(ArTolCO2)Ag] (3), containing a di(μ-hydroxo)diiron(III) unit
and a silver(I) ion. Because the synthesis of 3 was performed
under anhydrous anaerobic conditions, the most likely source of
the OHꢀ groups is trace water in the reaction mixture. Using 3.5
equiv of AgClO4, instead of 1.0 equiv, provided a higher yield of
3 (see Experimental Section).
The octahedral iron atoms (FeꢀFe = 3.01 Å) in 3 are bridged
by two hydroxide ions (FeꢀOave = 2.02 Å) and one terphenyl-
carboxylate ligand (FeꢀOave = 2.04 Å) (Figure 8). The hydrogen
atoms of the hydroxides were not located, but the protonation
state could be confidently assigned on the basis of the FeꢀO
bond distances. Whereas bridging hydroxides afford FeꢀO dis-
tances of 1.9ꢀ2.1 Å, bridging oxides have FeꢀO bond lengths in
the 1.7ꢀ1.8 Å range.69,70 The phenolate (FeꢀOave = 1.86 Å) and
imine (FeꢀNave = 2.12 Å) donors of PIM2ꢀ and a perchlorate
anion (FeꢀOave = 2.13 Å) make up the rest of the iron coordination
sphere. A silver(I) ion is disordered over two positions within the
interior of the macrocycle. Ag(1A) interacts with a hydroxide
ligand (AgꢀO(6) = 2.25 Å), the ether oxygen of PIM2ꢀ
(AgꢀO(5) = 2.36 Å), and a CdC π bond of the terphenylcar-
boxylate moiety (AgꢀC(60) = 2.49 Å, AgꢀC(61) = 2.74 Å).
Ag(1B), a minority species, has close contacts with a hydroxide
ligand (AgꢀO(6) = 2.33 Å), the ether oxygen of PIM2ꢀ
(AgꢀO(5) = 2.40 Å), and one of the aromatic carbon atoms
(AgꢀC(60) = 2.69 Å) (SI Figure S15). The AgꢀC distances are
within the 2.36ꢀ2.77 Å limit reported for other silver(I)ꢀaryl
interactions.71ꢀ73 A bond valence sum (BVS) analysis,74,75 an
empirical quantity used to determine the oxidation state of metal
ions in coordination compounds on the basis of crystallographi-
cally determined metalꢀligand distances, returned values of 3.12
and 3.07 for Fe(1) and Fe(2), respectively (SI Table S2), indicating
that the iron centers in 3 are both in the þ3 oxidation state.
To provide further evidence for this assignment, the M€ossbauer
spectrum of a frozen solution of 3 in tetrahydrofuran was recorded
at 80 K (Table 1). The data could be satisfactorily fit to a single
quadrupole doublet, with δ = 0.49(2) mm/s and ΔEQ = 1.38(2)
mm/s, parameters typical of iron(III) complexes having mixed
oxygen and nitrogen donor groups. The single quadrupole doublet
in the M€ossbauer spectrum is consistent with the X-ray crystal
structure of 3, which has two chemically equivalent iron centers.
The reason for the broad line width (Γ = 0.56 mm/s, Table 1) is
not immediately clear, but it may be due to the presence of a
second closely related diiron(III) species in the sample or slow
nuclear relaxation rate of the iron atoms at 80 K.76
Because chemical oxidation of 2 using AgClO4 led to metal
binding by the perchlorate anion, the oxidation was repeated
using the silver salt of a less coordinating anion, AgSbF6. Reaction of
2 with 1 equiv of AgSbF6 in dichloromethane yielded a hetero-
geneous dark brown solution over the course of ∼1 h. After removal
of a black solid, the UVꢀvis spectrum of the filtrate revealed
bands at 300, 374, and 605 nm (SI Figure S6), suggesting for-
mation of a new species. Single crystals of the product were grown
by slow diffusion of diethyl ether into an acetonitrile solution
containing the material. X-ray diffraction studies revealed a tetra-
nuclear complex, [Fe4(μ-X)4(μ-Y)2(PIM)2(ArTolCO2)2], where
the identity of atoms X and Y is probably Fꢀ. We have not been
able to fully solve this structure owing to unresolved disorder in
the crystal. On the basis of its optical spectrum, however, we can
rule out [Fe4(μ-OH)6(PIM)2(ArTolCO2)2] (8) as a possible
product (vide infra).
10575
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Figure 9. Absorption spectra of 1 þ O2 (black trace), 4 (blue trace),
and 5 (red trace) in dichloromethane.
Figure 8. Stick figure representation of the X-ray crystal structure of
[Fe2(μ-OH)2(ClO4)2(PIM)(ArTolCO2)Ag] (3, top). An isolated view
of the heterometallic core is shown in the diagram below. The silver ion
is disordered over two positions; the major component (Ag(1A), ∼91%
occupancy) is depicted. Solvent molecules and hydrogen atoms are
omitted for clarity. Color scheme: iron, orange; carbon, gray; nitrogen,
blue; oxygen, red; sulfur, yellow; green, chlorine; purple, silver. Selected
bond distances (Å) and angles (deg): Fe(1)ꢀFe(2) = 3.0119(8);
Fe(1)ꢀO(1) = 1.847(3); Fe(1)ꢀN(1) = 2.119(4); Fe(1)ꢀO(6) =
2.048(3); Fe(1)ꢀO(7) = 1.996(3); Fe(1)ꢀO(8) = 2.030(3); Fe-
(1)ꢀO(13) = 2.124(3); Fe(2)ꢀO(2) = 1.866(3); Fe(2)ꢀN(2) =
2.128(4); Fe(2)ꢀO(6) = 2.043(3); Fe(2)ꢀO(7) = 1.974(3); Fe-
(2)ꢀO(9) = 2.048(3); Fe(2)ꢀO(15) = 2.136(3); Ag(1A)ꢀO(6) =
2.260(3); Ag(1A)ꢀO(5) = 2.414(3); Ag(1A)ꢀC(60) = 2.349(5);
Ag(1A)ꢀC(61) = 2.596(5); O(1)ꢀFe(1)ꢀN(1) = 88.5(1); O(2)ꢀFe-
(2)ꢀN(2) = 88.8(1); Fe(1)ꢀO(6)ꢀFe(2) = 94.8(1); Fe(1)ꢀO-
(7)ꢀFe(2) = 98.6(1); O(5)ꢀAg(1A)ꢀO(6) = 109.5(1); O(5)ꢀ
Ag(1A)ꢀC(60) = 100.4(1); O(6)ꢀAg(1A)ꢀC(61) = 117.4(1). See
SI Figure S15 for a depiction of Ag(1B): Ag(1B)ꢀO(6) = 2.335(6),
Ag(1B)ꢀO(5) = 2.404(6), Ag(1B)ꢀC(60) = 2.687(8), O(5)ꢀAg-
(1B)ꢀO(6) = 107.3(2), O(6)ꢀAg(1B)ꢀC(60) = 124.7(3), O(5)ꢀAg-
(1B)ꢀC(60) = 91.7(2). A thermal ellipsoid ORTEP diagram of 3 is
provided in Figure S15.
Figure 10. 1H NMR spectra (500 MHz) of (A) [Fe2(PIM)(Ph3C-
CO2)2] (1) þ O2, (B) [Fe2(μ-OH)(PIM)(Ph3CCO2)3] (4), and (C)
[Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] (5). Comparison of spectrum A to
spectra B and C reveals that reaction of 1 with O2 leads to formation of 4
and 5, in addition to at least one other product. The peaks that most
likely arise from 4 and 5 in spectrum A are marked with red and green
lines, respectively. Not all of the peaks in A corresponding to 4 and 5
were assigned because of uncertainty due to the broadness and overlap
of some resonances. All spectra were recorded at room temperature with
diiron complex concentrations of ∼5ꢀ10 mM in CDCl3. Relative peak
heights are not normalized between spectra.
Reactivity with O2. To determine whether 1 and 2 exhibit the
same functional activity as the diiron cores in the BMM proteins,3
their reactivity with dioxygen was investigated (Scheme 3).
Exposure of a dichloromethane solution of 1 to dioxygen at room
temperature led to an instantaneous color change and the ap-
pearance of new optical bands at 370 and 570 nm (Figure 9, black
trace). Facile reactivity of 1 with O2 was confirmed by NMR
spectroscopy. Injecting O2 into a septum-sealed NMR tube con-
taining 1 in dichloromethane-d2 afforded the 1H NMR spectrum
shown in Figure 10A. The numerous proton resonances in the
spectrum, ranging from approximately 80 to ꢀ10 ppm, suggest
the formation of multiple species in the reaction mixture.
To characterize the reaction product(s), single crystals of the
material were grown by slow diffusion of pentane into a benzene
solution containing the dark red-brown solid. After several days, a
10576
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Figure 11. Stick figure representation of the X-ray crystal structure of
[Fe2(μ-OH)(PIM)(Ph3CCO2)3] (4, top). An isolated view of the
diiron core is shown in the diagram below. Solvent molecules and
hydrogen atoms are omitted for clarity. Color scheme: iron, orange;
carbon, gray; nitrogen, blue; oxygen, red; sulfur, yellow. Selected bond
distances (Å) and angles (deg): Fe(1)ꢀFe(2) = 3.4873(5); Fe(1)ꢀ
O(1) = 1.890(2); Fe(1)ꢀN(1) = 2.115(2); Fe(1)ꢀO(6) = 1.951(2);
Fe(1)ꢀO(10) = 2.064(2); Fe(1)ꢀO(20) = 2.075(2); Fe(1)ꢀO(21) =
2.133(2); Fe(2)ꢀO(2) = 1.872(2); Fe(2)ꢀN(2) = 2.136(2); Fe-
(2)ꢀO(6) = 1.968(2); Fe(2)ꢀO(11) = 1.980(2); Fe(2)ꢀO(30) =
2.141(2); Fe(2)ꢀO(31) = 2.094(2); O(1)ꢀFe(1)ꢀN(1) = 87.42(7);
O(2)ꢀFe(2)ꢀN(2) = 87.47(7); Fe(1)ꢀO(6)ꢀFe(2) = 125.69(9);
O(6)ꢀFe(1)ꢀO(10) = 88.28(7); O(6)ꢀFe(2)ꢀO(11) = 88.89(7). A
thermal ellipsoid ORTEP diagram of 4 is provided in SI Figure S16.
Figure 12. Stick figure representation of the X-ray crystal structure of
[Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] (5, top). The center of the complex
is located on a crystallographic inversion center. An isolated view of the
tetrairon core is depicted in the diagram below. Solvent molecules and
hydrogen atoms are omitted for clarity. Color scheme: iron, orange;
carbon, gray; nitrogen, blue; oxygen, red; sulfur, yellow. Selected bond
distances (Å) and angles (deg): Fe(1)ꢀFe(2) = 3.0560(5); Fe(1)ꢀ
Fe(2A) = 3.780 (standard deviation not calculated); Fe(1)ꢀO(1) =
1.912(2); Fe(1)ꢀN(1) = 2.172(2); Fe(1)ꢀO(6) = 2.042(2); Fe(1)ꢀ
O(8) = 2.030(2); Fe(1)ꢀO(9) = 1.998(2); Fe(1)ꢀO(10) = 1.978(2);
Fe(2)ꢀO(2) = 1.910(2); Fe(2)ꢀN(2) = 2.229(2); Fe(2)ꢀO(7) =
2.073(2); Fe(2)ꢀO(8) = 2.029(2); Fe(2)ꢀO(9) = 2.011(2); N(1)ꢀ
Fe(1)ꢀO(1) = 83.38(8); N(2)ꢀFe(2)ꢀO(2) = 85.42(8); Fe(1)ꢀ
O(8)ꢀFe(2) = 97.70(9); Fe(1)ꢀO(9)ꢀFe(2) = 99.35(9); Fe(1)ꢀ
O(10)ꢀFe(1A) = 146.5(1). A thermal ellipsoid ORETP diagram of 5 is
provided in SI Figure S17.
mixture of dark brown rectangular blocks and dark brown hexagonal
prisms was obtained in addition to an amorphous solid. X-ray
diffraction analysis of the rectangular-shaped crystals showed that
the compound has a (μ-hydroxo)diiron(III) core and the molec-
ular formula [Fe2(μ-OH)(PIM)(Ph3CCO2)3] (4, Figure 11,
Table S1). The distorted octahedral iron centers are separated
by 3.49 Å, with Fe(1)ꢀO(6) and Fe(2)ꢀO(6) bond lengths of
1.95 and 1.97 Å, respectively. The hydrogen atom on O(6), the
bridging hydroxide ion, was located from a difference Fourier
map. The diiron core is supported by the PIM2ꢀ ligand, with
average FeꢀO(phenoxyl) and FeꢀN(imine) distances of 1.31
and 2.12 Å, respectively, and three triphenylacetate groups. Two
of the carboxylates coordinate to the iron atoms in a terminal
bidentate mode, giving average FeꢀO distances of 2.11 Å, and
the remaining carboxylate bridges the diiron core, with an average
FeꢀO bond length of 2.02 Å.
di(μ-hydroxo)(μ-triphenylacetato)diiron(III) units are linked
by two bridging hydroxide ions (Figure 12, Table S1). The
tetrairon(III) unit is located on a crystallographic inversion center.
The diiron subunit bound by the PIM2ꢀ ligand has an FeꢀFe
distance of 3.06 Å and is bridged by two hydroxide ligands
(FeꢀOave = 2.02 Å). The separation of iron atoms between two
[Fe2PIM] monomers is 3.78 Å, and the linkage involves only
one hydroxo bridge (FeꢀOave = 1.98 Å). The average Feꢀ
O(phenoxyl) and FeꢀN(imine) bond lengths were refined to be
approximately 1.91 and 2.20 Å, respectively. Finally, two bridging
X-ray diffraction studies of the hexagonal prisms, isolated
from the 1/O2 reaction mixture, revealed a tetranuclear
[Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] complex (5), in which two
10577
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
carboxylate groups (FeꢀOave = 2.06 Å) cap the tetranuclear
cluster at opposite ends.
about 39% of total iron in the 1/O2 sample, has not yet been
determined.
Analytically pure samples of 4 and 5 were prepared for
spectroscopic characterization. The dinuclear compound 4 was
synthesized by combining triphenylacetic acid, triethylamine,
and 1 in benzene, followed by exposure to dioxygen. Vapor diffusion
of pentane into the benzene reaction solution over the course of
∼24 h afforded dark brown crystals. X-ray diffraction analysis of
these crystals confirmed the desired (μ-hydroxo)diiron(III)
species 4. The UVꢀvisible spectrum of 4 exhibits absorption
bands at 370 and 570 nm (Figure 9, blue trace). The higher
energy band is most likely a ligand πꢀπ* transition, whereas the
visible band is most likely a hydroxo-to-iron(III) charge transfer.70
The dioxygen reactivity of the diiron(II) compound 2, which
contains sterically demanding terphenylcarboxylates rather than
triphenylacetates, was also examined (Scheme 3). Exposing a
dichloromethane solution of 2 to O2 led to an immediate color
change from bright red to dark brown. The absorption spectrum
of dioxygen-treated 2 has features at approximately 380, 473, and
1
600 nm (SI Figure S10, blue trace). The H NMR spectrum
revealed new resonances between 26 and ꢀ8 ppm (SI Figure
1
S11). Unlike the H NMR spectrum of 1/O2 (Figure 10A),
which shows multiple chemical species in solution, the data sug-
gest that reaction of 2 with O2 forms fewer products.
1
The H NMR spectrum of 4 was recorded in chloroform-d1
To determine the composition of the 2/O2 product, prepara-
tive-scale reactions were performed to obtain sufficient material
for characterization. Single crystals of oxygenated 2 were grown
by slow diffusion of diethyl ether into a solution of the complex in
acetonitrile. An X-ray structural investigation revealed that each
single crystal contains a mixture of two species, a (μ-oxo)di-
iron(III) [Fe2(μ-O)(PIM)(ArTolCO2)2] (6) compound and a di(μ-
hydroxo)diiron(III) [Fe2(OH)2(PIM)(ArTolCO2)2] (7) com-
pound (Figure 13, Table S1). This result was verified by analyzing
several crystals that were independently prepared by the same
method. The structures of 6 and 7 are identical except for the
nature of their bridging oxygen atoms, which were modeled with
positional disorder and partial occupancies assigned to atoms
O(6) [in 6] and O(100)/O(101) [in 7]. Data refinement
converged with 6 having an occupancy of 76% and 7 of 24%.
The iron atoms in both structures are separated by 3.44 Å. Com-
plex 6 is designated as a (μ-oxo)diiron(III) species because of its
short Fe(1)ꢀO(6) and Fe(2)ꢀO(6) distances of 1.74 and 1.75
Å, respectively.70 In contrast, 7 is a di(μ-hydroxo)diiron(III)
complex with the longer FeꢀO(100) and FeꢀO(101) bond
length of 2.10 Å (averaged), more typical of a dinuclear structure
with bridging hydroxide ligands. The iron atoms of 6 and 7 are
coordinated by the PIM2ꢀ ligand, with average FeꢀO(phenoxyl)
distances of 1.89 Å and average FeꢀN(imine) distances of 2.11
Å. Furthermore, each iron site contains a terminal terphenylcar-
boxylate (FeꢀOave = 2.11 Å). The iron atoms in 6 and 7 have
bond valence sums of 3.0 and 2.8, respectively (SI Table S2),
indicating that the iron centers are in the þ3 oxidation state.
Although three triphenylacetate ligands are accommodated by
the PIM2ꢀ platform in 4 (Figure 14A), only two terphenylcar-
boxylates could occupy the same space in 6 (Figure 14B). These
observations underscore the importance of considering both
shape and size when selecting an appropriate carboxylate for
synthetic modeling studies; a fan-shaped terphenylcarboxylate is
preferable over the cone-shaped triphenylacetate for producing
dinuclear oxygenated products using the PIM2ꢀ ligand framework.
The structure of 7 closely mimics that of the oxidized core of
sMMOHox,11 which also contains a di(μ-hydroxo)diiron(III)
unit (Figure 3). We attribute the longer FeꢀFe distance in 7,
compared to sMMOHox, to the lack of a bridging carboxylate.
(Figure 10B). As expected for a paramagnetic compound, the
spectrum shows several broad resonances, ranging from approxi-
mately 70 to ꢀ10 ppm. Assignment of the metal oxidation states
in 4 as iron(III) is supported by zero-field M€ossbauer spectro-
scopic measurements. Polycrystalline 57Fe-enriched 4 displays a
single quadrupole doublet in the M€ossbauer spectrum, with
parameters of δ = 0.52(2) mm/s and ΔEQ = 0.95(2) mm/s
(SI Figure S9A, Table 1). Although the two iron atoms in 4 are
not chemically equivalent, similarities in their coordination geome-
try and donor groups make them indistinguishable by zero-field
M€ossbauer spectroscopy.
The tetrairon(III) complex 5 was also fully characterized. The
compound was isolated as brown crystals by slow evaporation of
a benzene solution containing 1 and dioxygen under ambient
conditions. To ensure that the bulk product did not contain 4,
several representative crystals were analyzed by X-ray crystal-
lography. In all cases, tetranuclear 5 was obtained. The absorp-
tion spectrum has peaks with λmax at 370 and 540 nm (Figure 9,
red trace). The visible band, which is blue-shifted by approxi-
mately 30 nm from the 570 nm absorption feature in 4, is most
likely a ligand-to-metal transition involving the diironꢀhydroxo
1
unit.70 The H NMR spectrum of 5 in chloroform-d1 exhibits
paramagnetically broadened peaks at ∼50, 44, 31, 15, and
1
ꢀ8 ppm (Figure 10C). The simpler H NMR spectrum of 5,
compared to that of 4, is consistent with its higher molecular
symmetry (pseudo-C2h for 5 versus C1 for 4). The zero-field
M€ossbauer spectrum of 57Fe-enriched 5 was fit with parameters
that are distinct from those of 4, giving δ = 0.51(2) mm/s and
ΔEQ = 1.06(2) mm/s (Figure S9B, Table 1). Because the four
iron sites in 5 are chemically equivalent, the single-quadrupole
doublet observed in the M€ossbauer spectrum is consistent with
its X-ray structure.
Comparison of the 1H NMR spectrum of 1/O2 (Figure 10A)
with the spectra of 4 (Figure 10B) and 5 (Figure 10C) reveals
that not all of the peaks are accounted for, particularly the broad
resonances at ∼77, 65, 54, 27, 22, and 20 ppm. To determine the
number of iron-containing species generated upon reaction of 1
with dioxygen, the crude 1/O2 solid was examined by M€ossbauer
spectroscopy (SI Figure S8). The M€ossbauer data display three
overlapping quadrupole doublets, with the following character-
istics: δ1 = 0.50(2) mm/s, ΔEQ1 = 0.78(2) mm/s, Area 1 = 24%;
δ2 = 0.51(2) mm/s, ΔEQ2 = 1.12(2) mm/s, Area 2 = 37%; δ3 =
0.52(2) mm/s, ΔEQ3 = 1.59(2) mm/s, Area 3 = 39%. The three
sites have nearly identical isomer shift values but differ in their
quadrupole splitting parameters. On the basis of their similarities
to the M€ossbauer parameters of the crystallographically char-
acterized species, sites 1 and 2 are ascribed to compounds 4 and
5, respectively. The nature of the third site, which comprises
ꢀ
More importantly, however, switching of a bridging ArTolCO2
in 2 to a terminal position in 7 reproduces the redox-dependent
carboxylate shift observed in the BMM proteins.29
The presence of both (μ-oxo)diiron(III) and di(μ-hydro-
xo)diiron(III) species in the reaction product of 2 with dioxygen
was further supported by zero-field M€ossbauer measurements.
Polycrystalline 2/O2 gave a M€ossbauer spectrum that was best fit
to two quadrupole doublets, δ1 = 0.47(2) mm/s, ΔEQ1 = 1.52(2)
mm/s, δ2 = 0.50(2) mm/s, and ΔEQ2 = 0.97(2) mm/s.
10578
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
Figure 13. Stick figure representation of the X-ray crystal structure of [Fe2(μ-O)(PIM)(ArTolCO2)2] (6, A) and [Fe2(μ-OH)2(PIM)(ArTolCO2)2]
(7, B) . The two complexes occur in a single crystal and differ only in their (μ-oxo)diiron(III) and di(μ-hydroxo)diiron(III) cores, respectively. The ratio
of 6:7 was determined to be 76:24 and is similar in two other independently prepared crystals. Solvent molecules and hydrogen atoms are omitted for
clarity. Color scheme: iron, orange; carbon, gray; nitrogen, blue; oxygen, red; sulfur, yellow. Selected bond distances (Å) and angles (deg): Fe(1)ꢀFe(2) =
3.436(1); Fe(1)ꢀO(1) = 1.880(4); Fe(1)ꢀN(1) = 2.103(5); Fe(1)ꢀO(6) = 1.745(5); Fe(1)ꢀO(7) = 2.086(4); Fe(1)ꢀO(8) = 2.133(4);
Fe(1)ꢀO(100) = 2.12(2); Fe(1)ꢀO(101) = 2.09(2); Fe(2)ꢀO(2) = 1.896(4); Fe(2)ꢀN(2) = 2.122(5); Fe(2)ꢀO(6) = 1.751(5); Fe(2)ꢀO(9A) =
2.05(2); Fe(2)ꢀO(10A) = 2.03(2); Fe(2)ꢀO(100) = 2.10(2); Fe(2)ꢀO(101) = 2.08(2); O(1)ꢀFe(1)ꢀN(1) = 87.1(2); O(2)ꢀFe(2)ꢀN(2) =
87.9(2); Fe(1)ꢀO(6)ꢀFe(2) = 158.8(3); Fe(1)ꢀO(100)ꢀFe(2) = 109.0(9); Fe(1)ꢀO(101)ꢀFe(2) = 111.0(1). A thermal ellipsoid ORTEP
diagram of 6/7 is provided in SI Figure S18.
Because site 1 has a larger ΔEQ value than site 2, the ratio for 6:7
in tetrahydrofuran is 61:36. The varying percentages of 6 and 7,
as determined from X-ray crystallographic and M€ossbauer spec-
troscopic measurements, are probably due to the different
amounts of H2O present in each 2/O2 sample. Complex 6
probably forms directly from reaction of 2 with dioxygen and
converts to 7 by subsequent reaction with water (Scheme 3).
Such interconversion between (μ-oxo)dimetallic and di(μ-hy-
droxo)dimetallic units has been previously observed. M€ossbauer
spectroscopic studies revealed that drying of a [FeIII2(μ-OH)2]
complex resulted in extrusion of water to give a [FeIII2(μ-O)]
species.77 More recently, crystal-to-crystal conversion of a (μ-
Figure 14. Hybid stick and space-filling diagrams of [Fe2(μ-OH)-
(PIM)(Ph3CCO2)3] (4, A) and [Fe2(μ-O)(PIM)(ArTolCO2)2] (6, B).
The aromatic rings of the carboxylate ligands are representated as
spheres (carbon = green, hydrogen = white) to emphasize their shape
and volume. The location of the carboxylate unit, either Ph3CCO2ꢀ or
ArTolCO2ꢀ, is indicated by a white arrow. The PIM2ꢀ framework is
displayed in stick form (carbon/hydrogen, gray; nitrogen, blue; oxygen,
red; sulfur, yellow) with the iron atoms shown as orange spheres.
oxo)divanadium polyoxometalate cluster to a di(μ-hydroxo)di-
vandium analogue was discovered following exposure of the
starting complex to water vapor.78 These studies suggest that
similar processes may occur in the active sites of non-heme diiron
proteins as well.
Because 7 differs from 6 by a single water molecule, we explored
whether addition of H2O could convert the (μ-oxo)diiron(III)
species 6 to the di(μ-hydroxo)diiron(III) complex 7. When a
dichloromethane solution of 6/7 was treated with 10 μL of H2O,
∼15 equiv relative to the diiron complexes, the brown solution
instantaneously turned pale red. New optical bands at 360 and
550 nm were observed (Figure S10, red trace), which are reminis-
cent of those displayed by the tetrairon complex [Fe4(μ-OH)6-
(PIM)2(Ph3CCO2)2] (5, Figure 9, red trace). To obtain further
spectroscopic characterization, a sample containing 57Fe-en-
riched 6/7 and 15 equiv of H2O in tetrahydrofuran was studied
by M€ossbauer spectroscopy at 80 K (Figure S12C). The spec-
trum was fit to a single quadrupole doublet, with δ = 0.49(2)
mm/s and ΔEQ = 0.97(2) mm/s. These parameters are nearly
identical to those obtained for 5 (Table 1). Although we are
unable to crystallize this material, the spectroscopic data suggest
that the compound has the molecular formula [Fe4(μ-OH)6-
(PIM)2(ArTolCO2)2] (8).
Sites 1 and 2 were refined with areas of 21% and 79%, respectively.
Because (μ-oxo)diiron(III) species typically have larger ΔEQ
values (>1.0 mm/s) than (μ-hydroxo)diiron(III) complexes
(<1.0 mm/s), site 1 is attributed to complex 6 and site 2 to
complex 7. When the 2/O2 solid reaction product was dried
under vacuum at 150 °C for 24 h, M€ossbauer measurements
yielded a 6:7 ratio of 85:15 (Table 1), which is similar to the 6:7
ratio of 76:24 determined by X-ray crystallography (Figure 13),
suggesting that 6 can be obtained from 7 by extrusion of H2O.
To examine the 6:7 ratio in solution, an 57Fe-enriched sample
of the 2/O2 solid dissolved in tetrahydrofuran was studied by
M€ossbauer spectroscopy at 80 K, yielding the following para-
meters: δ1 = 0.51(2) mm/s, ΔEQ1 = 1.24(2) mm/s, δ2 = 0.49(2)
mm/s, and ΔEQ2 = 0.78(2) mm/s, with sites 1 and 2 having
occupancies of 61% and 36%, respectively (SI Figure S12B).
10579
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
The tendency for iron complexes to aggregate into clusters is
well documented in the porphryin as well as the synthetic diiron
literature.79,81 Unlike the BMMs, which encapsulate their diiron
cofactors within a protein matrix to avoid unwanted side reac-
tions, small-molecule mimics do not have such exquisite steric
protection. By installing bulkier groups around the PIM2ꢀ ligand
periphery, it should be possible to prevent formation of poly-
nuclear species that lead to a dead-end in modeling studies.
Having a “protected” ligand platform may also allow assembly of
diiron(II) models using less sterically demanding carboxylates;
this modification will reduce the steric congestion at the diiron
core and may facilitate more facile reaction with external guests,
such as O2 and other substrates.
’ ACKNOWLEDGMENT
The work was supported by a grant GM032134 from the
National Institute of General Medical Sciences. The authors
thank Dr. Daniela Buccella for X-ray crystallographic assistance.
’ REFERENCES
(1) Feig, A. L.; Lippard, S. J. Chem. Rev. 1994, 94, 759–805.
(2) Wallar, B. J.; Lipscomb, J. D. Chem. Rev. 1996, 96, 2625–2658.
(3) Merkx, M.; Kopp, D. A.; Sazinsky, M. H.; Blazyk, J. L.; M€uller, J.;
Lippard, S. J. Angew. Chem., Int. Ed. 2001, 40, 2782–2807.
(4) Arenghi, F. L. G.; Berlanda, D.; Galli, E.; Sello, G.; Barbieri, P.
Appl. Environ. Microbiol. 2001, 67, 3304–3308.
(5) Cafaro, V.; Izzo, V.; Scognamiglio, R.; Notomista, E.; Capasso,
P.; Casbarra, A.; Pucci, P.; Di Donato, A. Appl. Environ. Microbiol. 2004,
70, 2211–2219.
’ SUMMARY AND CONCLUSIONS
(6) Cafaro, V.; Scognamiglio, R.; Viggiani, A.; Izzo, V.; Passaro, I.;
Notomista, E.; Dal Piaz, F.; Amoresano, A.; Casbarra, A.; Pucci, P.; Di
Donato, A. Eur. J. Biochem. 2002, 269, 5689–5699.
(7) Cadieux, E.; Vrajmasu, V.; Achim, C.; Powlowski, J.; M€unck, E.
Biochemistry 2002, 41, 10680–10691.
(8) Gallagher, S. C.; Cammack, R.; Dalton, H. Eur. J. Biochem. 1997,
247, 635–641.
(9) Hamamura, N.; Storfa, R. T.; Semprini, L.; Arp, D. J. Appl.
Environ. Microbiol. 1999, 65, 4586–4593.
(10) Kotani, T.; Yamamoto, T.; Yurimoto, H.; Sakai, Y.; Kato, N.
J. Bacteriol. 2003, 185, 7120–7128.
(11) Rosenzweig, A. C.; Nordlund, P.; Takahara, P. M.; Frederick,
C. A.; Lippard, S. J. Chem. Biol. 1995, 2, 409–418.
(12) Rosenzweig, A. C.; Frederick, C. A.; Lippard, S. J.; Nordlund, P.
Nature 1993, 366, 537–543.
(13) Rosenzweig, A. C.; Lippard, S. J. Acc. Chem. Res. 1994,
27, 229–236.
We designed a new macrocycle, H2PIM, that can be prepared
by a convenient multistep procedure in gram quantities. Reac-
tions of H2PIM with [Fe2(Mes)4] in the presence of external
carboxylates afford diiron(II) complexes that are excellent mod-
els for the reduced forms of the BMM active sites. These diiron(II)
compounds react rapidly with dioxygen to give a variety of
iron(III)-containing products, including (μ-hydroxo)diiron(III),
(μ-oxo)diiron(III), and di(μ-hydroxo)diiron(III) complexes of
direct relevance to the oxidized forms of the BMM diiron centers.
Hexa(μ-hydroxo)tetrairon(III) complexes were also obtained.
The composition of the oxygenation products is controlled in
part by the nature of the external carboxylate ligand and the
presence of water. These results demonstrate that the PIM2ꢀ
ligand is sufficiently preorganized to support a diiron core but, at
the same time, sufficiently flexible to allow structural rearrange-
ments at the metal center; this delicate balance has been difficult
to achieve in prior diiron modeling chemistry.
The importance of this work is the demonstration that a
macrocyclic framework can support carboxylate-bridged diiron
cores in multiple oxidation states, enforce syn N-donor stereo-
chemistry of two nitrogen ligands, supply two anionic oxygen atom
donors, allow binding of two external carboxylates to the iron
centers, have an internal cavity to form a quadrilateral Fe2(μ-O)2
core, and be readily synthesized in gram quantities. To the best of
our knowledge, no other ligand platform satisfies all of these
design elements. As such, this study represents a significant step
toward the ultimate goal of devising functional mimics of
carboxylate-bridged diiron centers that occur at the active sites
of many biologically essential metalloenzymes.
(14) Nordlund, P.; Reichard, P. Annu. Rev. Biochem. 2006, 75,
681–706.
(15) Vu, V. V.; Emerson, J. P.; Martinho, M.; Kim, Y. S.; M€unck, E.;
Park, M. H.; Que, L., Jr. Proc. Natl. Acad. Sci. U.S.A. 2009, 106,
14814–14819.
(16) Behan, R. K.; Lippard, S. J. Biochemistry 2010, 49, 9679–9681.
(17) Oldenhuis, R.; Vink, R. L. J.; Janssen, D. B.; Witholt, B. Appl.
Environ. Microbiol. 1989, 55, 2819–2826.
(18) Fox, B. G.; Borneman, J. G.; Wackett, L. P.; Lipscomb, J. D.
Biochemistry 1990, 29, 6419–6427.
(19) Canada, K. A.; Iwashita, S.; Shim, H.; Wood, T. K. J. Bacteriol.
2002, 184, 344–349.
(20) Tao, Y.; Fishman, A.; Bentley, W. E.; Wood, T. K. J. Bacteriol.
2004, 186, 4705–4713.
(21) Fishman, A.; Tao, Y.; Rui, L.; Wood, T. K. J. Biol. Chem. 2005,
280, 506–514.
(22) Cafaro, V.; Notomista, E.; Capasso, P.; Di Donato, A. Appl.
Environ. Microbiol. 2005, 71, 4736–4743.
(23) Cafaro, V.; Notomista, E.; Capasso, P.; Di Donato, A. Appl.
Environ. Microbiol. 2005, 71, 4744–4750.
’ ASSOCIATED CONTENT
(24) Du Bois, J.; Mizoguchi, T. J.; Lippard, S. J. Coord. Chem. Rev.
2000, 200ꢀ202, 443–485.
S
Supporting Information. Additional spectroscopic and
b
(25) Fontecave, M.; Mꢀenage, S.; Duboc-Toia, C. Coord. Chem. Rev.
1998, 178ꢀ180, 1555–1572.
electrochemical data, X-ray data collection and refinement de-
tails, a bond valence sum table, and partially labeled thermal
ellipsoid diagrams of 1ꢀ7, as well as the corresponding CIF files.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(26) Que, L., Jr.; Tolman, W. B. Nature 2008, 455, 333–340.
(27) Friedle, S.; Reisner, E.; Lippard, S. J. Chem. Soc. Rev. 2010, 39,
2768–2779.
(28) Siewert, I.; Limberg, C. Chem.—Eur. J. 2009, 15, 10316–10328.
(29) Rardin, R. L.; Tolman, W. B.; Lippard, S. J. New J. Chem. 1991,
15, 417–430.
(30) Tshuva, E. Y.; Lippard, S. J. Chem. Rev. 2004, 104, 987–1012.
(31) Armstrong, W. H.; Spool, A.; Papaefthymiou, G. C.; Frankel,
R. B.; Lippard, S. J. J. Am. Chem. Soc. 1984, 106, 3653–3667.
’ AUTHOR INFORMATION
Corresponding Author
lippard@mit.edu
10580
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581

Journal of the American Chemical Society
ARTICLE
(32) Dong, Y.; Yan, S.; Young, V. G., Jr.; Que, L., Jr. Angew. Chem.,
Int. Ed. 1996, 35, 618–620.
(33) Kim, K.; Lippard, S. J. J. Am. Chem. Soc. 1996, 118, 4914–4915.
(34) 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.
(65) Adam, B.; Bill, E.; Bothe, E.; Goerdt, B.; Haselhorst, G.;
Hildenbrand, K.; Sokolowski, A.; Steenken, S.; Weyherm€uller, T.;
Wieghardt, K. Chem.—Eur. J. 1997, 3, 308–319.
(66) Roy, N.; Sproules, S.; Weyherm€uller, T.; Wieghardt, K. Inorg.
Chem. 2009, 48, 3783–3791.
(67) Strautmann, J. B. H.; Freiherr von Richthofen, C.-G.; George,
S. D.; Bothe, E.; Bill, E.; Glaser, T. Chem Commun. 2009, 2637–2639.
(68) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.
(69) Sanders-Loehr, J.; Wheeler, W. D.; Shiemke, A. K.; Averill,
B. A.; Loehr, T. M. J. Am. Chem. Soc. 1989, 111, 8084–8093.
(70) Kurtz, D. M., Jr. Chem. Rev. 1990, 90, 585–606.
(71) Kang, H. C.; Hanson, A. W.; Eaton, B.; Boekelheide, V. J. Am.
Chem. Soc. 1985, 107, 1979–1985.
(72) Xu, W.; Puddephatt, R. J.; Muir, K. W.; Torabi, A. A. Organo-
metallics 1994, 13, 3054–3062.
(35) LeCloux, D. D.; Barrios, A. M.; Lippard, S. J. Bioorg. Med. Chem.
1999, 7, 763–772.
(36) Dong, Y.; Mꢀenage, S.; Brennan, B. A.; Elgren, T. E.; Jang, H. G.;
Pearce, L. L.; Que, L., Jr. J. Am. Chem. Soc. 1993, 115, 1851–1859.
(37) Kuzelka, J.; Farrell, J. R.; Lippard, S. J. Inorg. Chem. 2003,
42, 8652–8662.
(38) Kodanko, J. J.; Morys, A. J.; Lippard, S. J. Org. Lett. 2005, 7,
4585–4588.
(39) Kodanko, J. J.; Lippard, S. J. Inorg. Chim. Acta 2008, 361,
894–900.
(73) Ning, G. L.; Wu, L. P.; Sugimoto, K.; Munakata, M.; Kuroda-
Sowa, T.; Maekawa, M. J. Chem. Soc., Dalton Trans. 1999, 2529–2536.
(74) Altermatt, D.; Brown, I. D. Acta Crystallogr. 1985, B41, 240–244.
(75) Thorp, H. H. Inorg. Chem. 1992, 31, 1585–1588.
(76) Drago, R. S. In Physical Methods For Chemists; Surfside Scien-
tific Publishers: Gainsville, FL, 1992; pp 626ꢀ649.
(40) Do, L. H.; Lippard, S. J. Inorg. Chem. 2009, 48, 10708–10719.
(41) Chen, C.-T.; Siegel, J. S. J. Am. Chem. Soc. 1994, 116, 5959–5960.
(42) Klose, A.; Solari, E.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C.; Re,
N. J. Am. Chem. Soc. 1994, 116, 9123–9135.
(43) Sheldrick, G. M. SADABS, Area-Detector Absorption Corrrec-
tion; University of G€ottingen: G€ottingen, Germany, 2001.
(44) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, A64, 112–122.
(45) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool;
Utrecht University: Utrecht, The Netherlands, 2000.
(46) Ambrosi, G.; Formica, M.; Fusi, V.; Giorgi, L.; Micheloni, M.
Coord. Chem. Rev. 2008, 252, 1121–1152.
(47) Bordwell, F. G.; Algrim, D. J. Org. Chem. 1976, 41, 2507–2508.
(48) Bordwell, F. G.; McCallum, R. J.; Olmstead, W. N. J. Org. Chem.
1984, 49, 1424–1427.
(49) Shu, L.; Nesheim, J. C.; Kauffmann, K.; M€unck, E.; Lipscomb,
J. D.; Que, L., Jr. Science 1997, 275, 515–518.
(77) Boudalis, A. K.; Clemente-Juan, J. M.; Dahan, F.; Psycharis, V.;
Raptopoulou, C. P.; Donnadieu, B.; Sanakis, Y.; Tuchagues, J.-P. Inorg.
Chem. 2008, 47, 11314–11323.
(78) Uehara, K.; Mizuno, N. J. Am. Chem. Soc. 2011, 133, 1622–1625.
(79) Collman, J. P.; Gagne, R. R.; Reed, C. A.; Halbert, T. R.; Lang,
G.; Robinson, W. T. J. Am. Chem. Soc. 1975, 97, 1427–1439.
(80) Lippard, S. J. Angew. Chem., Int. Ed. 1988, 27, 344–361.
(81) Chen, Q.; Lynch, J. B.; Gomez-Romero, P.; Ben-Hussein, A.;
Jameson, G. B.; O'Connor, C. J.; Que, L., Jr. Inorg. Chem. 1988,
27, 2673–2681.
(50) Kimura, S.; Bill, E.; Bothe, E.; Weyherm€uller, T.; Weighardt, K.
J. Am. Chem. Soc. 2001, 123, 6025–6039.
(51) Sessler, J. L.; Tomat, E.; Mody, T. D.; Lynch, V. M.; Veauthier,
J. M.; Mirsaidov, U.; Markert, J. T. Inorg. Chem. 2005, 44, 2125–2127.
(52) Tomat, E.; Cuesta, L.; Lynch, V. M.; Sessler, J. L. Inorg. Chem.
2007, 46, 6224–6226.
(53) Borovik, A. S.; Papaefthymiou, V.; Taylor, L. F.; Anderson,
O. P.; Que, L., Jr. J. Am. Chem. Soc. 1989, 111, 6183–6195.
(54) Cox, D. D.; Benkovic, S. J.; Bloom, L. M.; Bradley, F. C.;
Nelson, M. J.; Que, L. Jr.; Wallick, D. E. J. Am. Chem. Soc. 1988, 110,
2026–2032.
(55) Koch, S. A.; Millar, M. J. Am. Chem. Soc. 1982, 104, 5255–5257.
(56) Pyrz, J. W.; Roe, A. L.; Stern, L. J.; Que, L., Jr. J. Am. Chem. Soc.
1985, 107, 614–620.
(57) Although there is an ambiguity in fitting M€ossbauer data
containing overlapping quadrupole doublets with equal intensities, we
favor the current model because the chemically distinct four- and five-
coordinate iron atoms in 1 should give rise to different δ values. We have
not excluded the alternate model, however, which would afford similar δ
but different ΔEQ parameters.
(58) Yoon, S.; Lippard, S. J. J. Am. Chem. Soc. 2005, 127, 8386–8397.
(59) 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–
10185.
(60) DeWitt, J. G.; Bentsen, J. G.; Rosenzweig, A. C.; Hedman, B.;
Green, J.; Pilkington, S.; Papaefthymiou, G. C.; Dalton, H.; Hodgson,
K. O.; Lippard, S. J. J. Am. Chem. Soc. 1991, 113, 9219–9235.
(61) Stoian, S. A.; Smith, J. M.; Holland, P. L.; M€unck, E.; Bominaar,
E. L. Inorg. Chem. 2008, 47, 8687–8695.
(62) Hu, C.; Sulok, C. D.; Paulat, F.; Lehnert, N.; Twigg, A. I.;
Hendrich, M. P.; Schulz, C. E.; Scheidt, W. R. J. Am. Chem. Soc. 2010,
132, 3737–3750.
(63) Bertini, I.; Luchinat, C. In Physical Methods For Chemists; Drago,
R. S., Ed.; Surfside Scientific Publishers: Gainsville, FL, 1992; pp 500ꢀ558.
(64) Iggo, J. A. NMR Spectroscopy in Inorganic Chemistry; Oxford
University Press, Inc.: New York, 1999.
10581
dx.doi.org/10.1021/ja2021312 |J. Am. Chem. Soc. 2011, 133, 10568–10581