Initial synthesis and structure of an all-ferrous
analogue of the fully reduced [Fe4S4]0 cluster
of the nitrogenase iron protein
Thomas A. Scott†, Curtis P. Berlinguette‡, Richard H. Holm‡§, and Hong-Cai Zhou†§
†Department of Chemistry and Biochemistry, Miami University, Oxford, OH 45056; and ‡Department of Chemistry and Chemical Biology, Harvard University,
Cambridge, MA 02138
The synthetic cubane-type iron–sulfur clusters [Fe4S4(SR)4]z form a
four-member electron transfer series (z ⴝ 3ⴚ, 2ⴚ, 1ⴚ, and 0), all
members of which except that with z ⴝ 0 have been isolated and
characterized. They serve as accurate analogues of protein-bound
[Fe4S4(SCys)4]z redox centers, which, in terms of core oxidation
states, exhibit the redox couples [Fe4S4]3ⴙ/2ⴙ and [Fe4S4]2ⴙ/1ⴙ.
Clusters with the all-ferrous core [Fe4S4]0 have never been isolated
because of their oxidative sensitivity. Recent work on the Fe
protein of Azotobacter vinelandii nitrogenase has demonstrated
the formation of the all-ferrous state upon reaction with a strong
reductant. Treatment of the cyanide cluster [Fe4S4(CN)4]3ⴚ with
K[Ph2CO] in acetonitrile兾tetrahydrofuran affords the all-ferrous
cluster [Fe4S4(CN)4]4ⴚ, isolated as the Bu4Nⴙ salt. The x-ray structure demonstrates retention of a cubane-type structure with idealized D2d symmetry. The Mössbauer spectrum unambiguously
demonstrates the [Fe4S4]0 oxidation state. Bond distances, core
volumes, 57Fe isomer shifts, and visible absorption spectra make
evident the high degree of structural and electronic similarity with
the fully reduced Fe protein. The attribute of cyanide ligation
causes positive [Fe4S4]2ⴙ/1ⴙ and [Fe4S4]1ⴙ/0 redox potential shifts,
facilitating the initial isolation of an analogue of the [Fe4S4]0
protein site.
iron–sulfur clusters 兩 synthetic analogue 兩 all-ferrous cluster
A
n extensive set of cubane-type clusters [Fe4S4(SR)4]3⫺,2⫺,1⫺
(R ⫽ alkyl or aryl) has been prepared whose members serve
as accurate analogues of the [Fe4S4(SCys)4] redox sites in ferredoxins and other iron–sulfur proteins (1, 2). In terms of core
oxidation states, these clusters form the four-member electron
transfer series 1 throughout which the cubane [Fe4(␮3-S)4] core
[Fe4S4]0 º [Fe4S4]1⫹ º [Fe4S4]2⫹ º [Fe4S4]3⫹
[1]
is retained. The existence of this series was recognized by
voltammetry at the beginning of iron–sulfur analogue chemistry
(3). Because of the dependence of redox potential on substituent
R of thiolate clusters and other ligands L in [Fe4S4L4]z clusters,
the entire series has not been traversed electrochemically or
synthetically at constant R or L, except for R ⫽ Ph in an unusual
toluene兾(Bu 4 N)(BF 4 ) electrochemical phase (4). Many
[Fe4S4]2⫹,1⫹ clusters (2) and one [Fe4S4]3⫹ species (5) have been
isolated and subjected to detailed electronic and structural
characterization. The [Fe4S4]2⫹/1⫹ couple (E⬘o ⬇ ⫺0.4 to ⫺0.7 V)
is operative in many ferredoxins and other proteins, whereas the
[Fe4S4]3⫹/2⫹ couple (E⬘o ⬇ 0.1–0.4 V) occurs primarily in the
smaller group of ‘‘high-potential’’ proteins. The all-ferrous
[Fe4S4]0 state was not recognized to be potentially physiologically relevant until relatively recently. Further, numerous attempts to isolate all-ferrous [Fe4S4]0 species that are members of
the synthetic series 1 failed because of their pronounced suscep(2⫹/1⫹)
tibility toward oxidation, as indicated by the potentials E1/2
⫽
(1⫹/0)
⫺1.05 V and, especially, E1/2 ⫽ ⫺1.70 V vs. standard calomel
electrode for [Fe4S4(SPh)4]2⫺ in acetonitrile (6). However, after
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0504258102
the initial assertion by Watt and Reddy (7) in 1994, spectroscopic
data for the Fe protein of Azotobacter vinelandii reduced by
TiIII(citrate) convincingly demonstrated attainment of the
[Fe4S4]0 state (8–11). This protein is the electron donor to the
catalytic MoFe protein of the nitrogenase complex (12). For this
cluster, E⬘o(2⫹/1⫹) ⬇ ⫺0.31 V (7) and E⬘o(1⫹/0) ⬇ ⫺0.79 V (13).
Retention of the cubane-type [Fe4S4(SCys)4] structure of the
oxidized Fe protein with small dimensional changes has been
established crystallographically (14). Evidence has been presented for a two-electron transfer by the fully reduced Fe protein
in nitrogenase catalysis (15). Here, we demonstrate that the
heretofore elusive all-ferrous state in a cubane-type cluster can
be isolated in substance.
Materials and Methods
Preparation of (Bu4N)4[Fe4S4(CN)4]. Method A. All operations were
carried out under strictly anaerobic and anhydrous conditions.
To a mixture of 0.15 g (0.11 mmol) of [Fe4S4(PPri3)4](BPh4) (16),
0.5 ml of PPri3, and 1 ml of tetrahydrofuran (THF), K[Ph2CO]
was added [freshly prepared from 5.7 mg (0.15 mmol) of
potassium and 35 mg (0.19 mmol) of Ph2CO in 2 ml of THF]. The
reaction mixture was filtered into a solution of 0.13 g (0.48 mmol)
of (Bu4N)(CN) in 1 ml of THF. The solid that separated was
collected and washed with ether and THF to afford the product
as 0.13 g (75%) of a red microcrystalline solid.
Method B. A solution of 0.20 g (0.17 mmol) of
(Bu4N)3[Fe4S4(CN)4] (17) and 0.042 g (0.15 mmol) of (Bu4N)Cl
in 2 ml of acetonitrile was treated with K[Ph2CO] [freshly
prepared from 7.0 mg (0.18 mmol) of potassium and 46 mg (0.25
mmol) of Ph2CO in 2 ml of THF]. The reaction mixture was
stirred for 30 min and was filtered through a short Celite column.
Vapor diffusion of ether into the deep red filtrate over 16 h
followed by drying in vacuo gave the bulk product as 0.17 g
(69%) of red plate-like crystals. Analysis indicated the composition as (Bu4N)4[Fe4S4(CN)4]䡠1⁄2MeCN䡠KCl. Calculated for
C69H145.5ClFe4KN8.5S4 was the following: C, 54.53 (54.31); H,
9.58 (9.79); S, 8.43 (8.08). The two products displayed identical
Mössbauer spectroscopic properties.
X-Ray Structure Determination. A red crystal obtained from the
reaction mixture of method B before the drying step was
crystallographically analyzed on an Apex D8 CCD diffractometer (Bruker, Billerica, MA) at 213 K. Diffraction data were
consistent with the composition (Bu4N)4[Fe4S4(CN)4]䡠2MeCN
(Mr ⫽ 1507.66) in orthorhombic space group Pccn. The cell
constants a ⫽ 31.295 (2) Å, b ⫽ 11.9272 (9) Å, c ⫽ 23.630 (2) Å,
Abbreviation: THF, tetrahydrofuran.
Data deposition: The atomic coordinates have been deposited in the Cambridge Structural
Database, Cambridge Crystallographic Data Centre, Cambridge CB2 1EZ, United Kingdom
(CSD reference no. 271929).
§To
whom correspondence may be addressed. E-mail: holm@chemistry.harvard.edu or
zhouh@muohio.edu.
© 2005 by The National Academy of Sciences of the USA
PNAS 兩 July 12, 2005 兩 vol. 102 兩 no. 28 兩 9741–9744
CHEMISTRY
Contributed by Richard H. Holm, May 23, 2005
Fig. 2. Structure of [Fe4S4(CN)4]4⫺ showing 50% probability ellipsoids and
atom labeling scheme; primed and unprimed atoms are related by a twofold
axis.
Fig. 1.
Synthesis of [Fe4S4(CN)4]4⫺.
and V ⫽ 8820 (1) Å3 were retrieved (with SMART) and refined
(with SAINT) on all observed reflections. The structure was
solved by direct methods using SHELXL-97 (16, 17) (2␪max ⫽ 50.16°
and ␮ ⫽ 0.780 cm⫺1) and was refined to the standard agreement
factors R1 ⫽ 0.057, wR2 ⫽ 0.1425, and goodness-of-fit ⫽ 0.994.
Other Physical Measurements. Cyclic voltammograms were determined at a scan rate of 100 mV兾s by using a Model 610A
electrochemical analyzer (CH Instruments, Austin, TX) with a
glassy carbon working electrode, 0.1 M (Bu4N)(PF6) supporting
electrode, and a Ag兾AgCl reference electrode. Potentials were
corrected to the standard calomel electrode (SCE) (ESCE ⫽
EAg/AgCl ⫺ 0.04 V) to allow comparison with reported potentials
of synthetic iron–sulfur clusters. Mössbauer spectra were recorded with a constant acceleration spectrometer. Data were
analyzed with WMOSS software (WEB Research, Edina, MN);
isomer shifts are referenced to iron metal at room temperature.
Results and Discussion
Synthesis. The red all-ferrous cluster [Fe4S4(CN)4]4⫺ has been
prepared by the two methods in Fig. 1 and isolated as the Bu4N⫹
salt. Reduction of the previously reported cubane-type [Fe4S4]1⫹
clusters [Fe4S4(CN)4]3⫺ (17) or [Fe4S4(PPri3)4]1⫹ in the presence
of four equivalents of cyanide (and excess PPri3 to prevent
formation of the double cubane [Fe8S8(PPri3)6]) (16) with potassium benzophenone ketyl affords the product in 69–75%
yield. The key attribute is the use of electron-withdrawing
cyanide ligands, which for [Fe4S4(CN)4]3⫺ in dichloromethane
cause substantial positive potential shifts compared with values
(2⫹/1⫹)
⬇ ⫺0.44 V and
for [Fe4S4(SPh)4]2⫺. For this cluster, E1/2
(1⫹/0)
⬇ ⫺1.42 V. The second reduction potential is the same
E1/2
as that for the [Fe4S4]2⫹/1⫹ couple of [Fe4S4(SBut)4]2⫺ in N,Ndimethylformamide (3); [Fe4S4(SBut)4]3⫺, a cluster of comparable oxidative sensitivity, has been isolated (18). Consequently,
the potential shift of ⬇0.3 V facilitates isolation of
(Bu4N)4[Fe4S4(CN)4], which, however, is so sensitive that despite
numerous attempts, it cannot be recrystallized or maintained in
solution in the absence of a reducing agent without partial
oxidation of the cluster.
Structure. The desired cluster was obtained in red plate-like
crystals of apparent composition (Bu4N)4[Fe4S4(CN)4]䡠2MeCN
9742 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0504258102
in orthorhombic space group Pccn. The anion and two cations
reside on crystallographic twofold axes, and the other two
cations are in general positions; the anion:cation ratio is 1:4. The
cluster structure is depicted in Fig. 2; selected interatomic
distances are collected in Table 1. The cluster has the usual
cubane-type geometry (2, 19) consisting of two concentric
interpenetrating Fe4 and S4 tetrahedra. A crystallographically
imposed twofold axis bisects the opposite faces Fe(1,1⬘)S(2,2⬘)
and Fe(2,2⬘)S(1,1⬘) of the [Fe4S4]0 core, resulting in six Fe–S
[2.321(1)–2.364(1) Å] plus four Fe–Fe [2.627 (1)–2.696(1) Å]
distances that are independent. These distances are within the
normal intervals for Fe4S4 clusters. The FeII atoms exhibit
distorted tetrahedral coordination with essential linear cyanide
binding. The cluster approaches idealized D2d symmetry. We
note that Fe4S4 core structures, especially [Fe4S4]1⫹, are subject
Table 1. Selected interatomic distances in Å, angles in degrees,
and estimated standard deviations in parentheses
for [Fe4S4(CN)4]4ⴚ
Distance兾angle
Fe1-S1
Fe1-S2
Fe1-S2⬘
Fe1-C1
Fe1-Fe2
Fe1-Fe1⬘
Fe2-S1
Fe2-S1⬘
Fe2-S2
Fe2-C2
Fe2-Fe1⬘
Fe2-Fe2⬘
S1-Fe1-S2
S1-Fe1-S2⬘
S2-Fe1-S2⬘
C1-Fe1-S1
C1-Fe1-S2
C1-Fe1-S2⬘
Fe1-C1-N1
S1-Fe2-S2
S1-Fe2-S1⬘
S2-Fe2-S1⬘
C2-Fe2-S1
C2-Fe2-S1⬘
C2-Fe2-S2
Fe2-C2-N2
Value
2.325 Å (1)
2.364 Å (1)
2.326 Å (1)
2.087 Å (5)
2.6756 Å (8)
2.627 Å (1)
2.334 Å (1)
2.337 Å (1)
2.321 Å (1)
2.076 Å (4)
2.6825 Å (8)
2.696 Å (1)
106.34° (4)
105.91° (4)
109.24° (4)
115.6° (1)
104.9° (1)
114.5° (1)
175.2° (5)
107.48° (4)
106.79° (4)
105.68° (4)
110.3° (1)
112.2° (1)
114.0° (1)
177.7° (4)
Scott et al.
Table 2. Comparison of selected interatomic distances in Å and
volumes in Å3 for [Fe4S4(CN)4]3ⴚ兾4ⴚ and the fully reduced
Fe protein
Distance兾volume
Fe-Fe‡
Fe-S‡
Fe-C‡
V (Fe4)
V (S4)
V (Fe4S4)
[Fe4S4(CN)4]3⫺*
[Fe4S4(CN)4]4⫺
Fe protein†
2.71 (2)
2.29 (1)
2.024 (9)
2.34
5.64
9.41
2.67 (3)
2.33 (2)
2.082 (8)
2.25
6.21
9.48
2.65
2.33
—
2.17
6.17
9.23
Values in parentheses are standard deviations from the mean. —, no data.
*Ref. 17.
†Ref. 14.
‡Mean values.
Cyanide Stretching Frequencies. The value of vCN decreases appreciably upon reducing [Fe 4 S 4 (CN) 4 ] 3⫺ (2,103 cm ⫺1 ) to
[Fe4S4(CN)4]4⫺ (2,050 cm⫺1). The most closely related situation
occurs with K3[Fe(CN)6] and K4[Fe(CN)]6, for which vCN ⫽
2,116 and 2,041 cm⫺1, respectively. To the extent that the trends
in stretching frequencies are parallel, the data for the clusters
support reduction to the all-ferrous state, in which there appears
to be sufficient electron donation into the ␲* orbitals of the
ligand to reduce vCN to ⬍2,058 cm⫺1, the value for (Bu4N)(CN).
All data refer to polycrystalline solids.
Visible Absorption Spectrum. A characteristic property associated
with the reduction of the [Fe4S4]1⫹ state of the Fe protein to
[Fe4S4]0 is a color change in aqueous buffer from brown to red,
a decrease in absorbance over most of the visible spectrum, and
the development of a band at 520 nm (8). Synthetic
[Fe4S4(SR)4]3⫺ clusters resemble protein-bound [Fe4S4]1⫹ clusters in displaying diminished absorbance at ⬎⬇400 nm (6). As
may be seen in Fig. 3, [Fe4S4(CN)4]3⫺ does not exhibit any
discrete absorption bands throughout the UV-visible region.
However, when this cluster is reduced with a small excess of
K[Ph2CO] in acetonitrile兾THF [1:1 (vol兾vol)], the spectrum
intensifies, and a prominent shoulder appears at ⬇518 nm. This
Scott et al.
Fig. 3. UV-visible absorption spectra of [Fe4S4(CN)4]3⫺ (dotted line) in acetonitrile and [Fe4S4(CN)4]4⫺ (solid line) and K[Ph2CO] (dashed line) in acetonitrile兾THF [1:1 (vol兾vol)].
band is not due to the reductant, whose blue color arises from
the feature at 626 nm. A cyclic voltammogram of this solution
under an argon atmosphere at an initial potential of ⫺1.5 V
reveals oxidations at E1/2 ⫽ ⫺1.18 V and ⫺0.26 V, which
correspond to the ⫺1.42-V and ⫺0.44-V steps in pure acetonitrile. Because of the extreme ox idative sensitivity of
[Fe4S4(CN)4]4⫺, it was necessary to perform the absorption
spectral and electrochemical measurements in situ. Although the
visible absorption band in the protein and synthetic cluster is
likely a ligand-to-metal charge transfer, it has not been assigned.
Variable-temperature magnetic circular dichroism spectra of the
Fe protein indicate that the band at ⬇500 nm in aqueous buffer
is associated with a ground spin state S ⬎ 2 (25).
Mössbauer Spectra. Spectra of the two synthetic clusters are set
out in Fig. 4; isomer shifts ␦ and quadrupole splittings ⌬EQ are
summarized in Table 3. Spectral fits to a single site or to two
overlapping quadrupole doublets in an intensity ratio of 1:1 or
3:1 do not significantly alter the isomer shift. The spectrum of
[Fe4S4(CN)4]3⫺ at 4.2 K was reported as a 3:1 fit (17). At 77 K,
a very good simulation is obtained by one quadrupole doublet
with ␦ ⫽ 0.49 mm兾s. When this value is applied to the empirical
correlation ␦ ⫽ 1.43 ⫺ 0.40s between oxidation state s and
isomer shift of [Fe4S4(SR)4]z clusters at 77 K (2), s ⫽ 2.35
compared with Fe2.25⫹ for [Fe4S4]1⫹. The clusters differ in one
ligand of four, which in some cases measurably affects the isomer
shift (26, 27). As one example at 77 K, ␦ ⫽ 0.51 mm兾s for
[Fe4S4Cl4]2⫺ vs. ␦ ⫽ 0.43–0.44 mm兾s for [Fe4S4(SBut)4]2⫺ (28).
The spectrum of [Fe4S4(CN)4]4⫺ is equally well described by two
Fig. 4. Zero-field Mössbauer spectra of [Fe4S4(CN)4]3⫺ (a) and [Fe4S4(CN)4]4⫺
(b) at 77 K; the solid lines are fits to the data using the parameters of Table 3.
PNAS 兩 July 12, 2005 兩 vol. 102 兩 no. 28 兩 9743
CHEMISTRY
to a variety of static distortions from limiting Td symmetry,
arising from the exigencies of crystal packing, which are variant
with cation and substituent R or ligand L. For example,
[Fe4S4(CN)4]3⫺ exhibits a small crystallographically imposed
trigonal distortion (17).
Compared in Table 2 are bond distances and core volumes for
the synthetic clusters [Fe4S4(CN)4]3⫺,4⫺ and the Fe protein of the
A. vinelandii nitrogenase complex. Upon reduction of the cyanide cluster, Fe–Fe distances and Fe4 core volumes change little,
compared with the relatively large increase in Fe–S distances
and S4 core volumes; this behavior is common in the [Fe4S4]2⫹/1⫹
redox transitions of thiolate-ligated clusters. For [Fe4S4(SR)4]2⫺,
reported ranges are 2.41–2.49 Å3 for Fe4 and 5.28 –5.55 Å3 for
S4, and for [Fe4S4(SR)4]3⫺, the values are 2.38–2.50 Å3 for Fe4
and 5.74–5.80 Å3 for S4. The core total volumes increase slightly
upon reduction, from 9.50–9.70 Å3 for [Fe4S4]2⫹ to 9.57–9.93 Å3
for [Fe4S4]1⫹ (18–23). The volumes in Table 2 were calculated
from atomic coordinates (24) and indicate that the [Fe4S4]1⫹/0
transition results in similar changes. The Fe–Fe distances are not
distinguishable, the Fe4 volume shows a slight apparent decrease,
and Fe–S bond lengths and S4 volumes increase significantly.
The core volume enlarges by 0.7% because of the increased S4
volume, which is the case in all other volume comparisons. From
these observations, the added electron producing the all-ferrous
state is described by an antibonding orbital with substantial
sulfur character. Note the concordance between the metric
features of the [Fe4S4]0 cluster and the fully reduced Fe protein.
Table 3. 57Fe Mössbauer parameters for [Fe4S4(CN)4]3ⴚ兾4ⴚ and
the fully reduced Fe protein
Cluster
T, K
]3⫺
4.2†
[Fe4S4(CN)4]4⫺
Fe protein
77
77
4.2‡
[Fe4S4(CN)4
␦, mm兾s*
⌬EQ, mm兾s
Ratio
0.50, 0.57
0.49
0.65, 0.65
0.68, 0.68
1.47, 1.60
1.09
1.45, 2.00
1.2–1.7, 3.08
3:1
—
1:1
4-site fit
—, no data.
*Data referenced to Fe metal at room temperature; numbers are ⫾ 0.02 mm兾s.
†Ref. 17.
‡Refs. 8 and 11.
doublets in a 1:1 or 3:1 intensity ratio, and ␦ ⫽ 0.65 mm兾s for
both. The empirical correlation suggests s ⫽ 1.95 vs. Fe2⫹ for
[Fe4S4]0. Further, the isomer shift is in the range of tetrahedral
FeII species such as [Fe(SR)4]2⫺ and reduced rubredoxin, for
which ␦ ⬇ 0.70–0.76 mm兾s at 77 K (29). The Mössbauer data are
unequivocal in demonstrating the existence of an all-ferrous
cluster. Note the close values of isomer shifts of the cluster and
the Fe protein. However, unlike the Fe protein, the cluster does
not display a zero-field spectrum consisting of two resolved
quadrupole doublets in a 3:1 intensity ratio, with the less intense
doublet corresponding to a site with a large quadrupole splitting
(⌬EQ ⫽ 2.91 mm兾s at 130 K), as does the Fe protein (11). At 4.2
K, [Fe4S4(CN)4]4⫺ displays a complicated spectrum owing to the
internal magnetic field of the cluster. The Mössbauer spectrum
of the Fe protein has been interpreted in terms of an S ⫽ 4
ground state (11).
member of the electron transfer series 1. The closest approach
to this situation is the electrochemical detection of
[Fe4S4(SR)4]4⫺ and [Fe4S4(PR3)4]0 species. Multiple attempts to
obtain the latter resulted in isolation of [Fe8S8(PR3)6], which,
although all-ferrous, is an edge-bridged double cubane (16, 30).
The present cluster is to be distinguished from [Fe4S4(CO)12],
which has a cubane-type [Fe4S4]0 core, but with six-coordinate
diamagnetic metal sites and a mean Fe䡠䡠䡠Fe distance of 3.47 Å
(31). The cluster [Fe4S4(CN)4]4⫺, albeit with nonphysiological
terminal ligands, is, to our knowledge, the first synthetic analogue of the fully reduced cluster in the Fe protein of nitrogenase. Overall, the properties of the latter, including core dimensions, the visible absorption band, and isomer shift, are closely
approached. However, there are some differences in detail.
There is no indication of a core distortion that, from extended
x-ray absorption fine structure analysis, affords two short (2.52
Å) plus one long (2.77 Å) Fe–Fe distances and one short (2.29
Å) plus three long (2.35–2.37 Å) Fe–S distances (9). Further,
there is no evidence from Mössbauer spectroscopy of a single
iron site with a large quadrupole splitting that could be associated with discrete FeII. The cluster in the Fe protein is located
between two subunits and is connected by two Fe–SCys interactions to each. It is this arrangement that produces the largest
accessible surface area, which very likely contributes to the
stability of the [Fe4S4]0 state (14) by means of solvent (hydrogen
bonding) interactions that lead to fractional charge dissipation to
the medium. It is not surprising that a cluster bound in this way
may possess features that cannot be reproduced in a crystalline
or solution phase that generates a far less anisotropic
environment.
troscopic data establish [Fe4S4(CN)4]4⫺ as the first example of an
isolable all-ferrous cubane-type iron–sulfur cluster that is a
This work was supported by National Science Foundation (NSF) Grants
ECS-0403669 and CHE-0449634 at Miami University and by National
Institutes of Health Grant GM 28856 at Harvard University. The x-ray
diffractometer was purchased through NSF Grant EAR-0003201.
1. Holm, R. H. (2004) in Comprehensive Coordination Chemistry II: Biocoordination Chemistry, eds. Que, L., Jr., & Tolman, W. A. (Elsevier, London), Vol.
8, pp. 61–90.
2. Rao, P. V. & Holm, R. H. (2004) Chem. Rev. 104, 527–559.
3. DePamphilis, B. V., Averill, B. A., Herskovitz, T., Que, L., Jr., & Holm, R. H.
(1974) J. Am. Chem. Soc. 96, 4159–4167.
4. Pickett, C. J. (1985) J. Chem. Soc. Chem. Commun., 323–324.
5. O’Sullivan, T. & Millar, M. M. (1985) J. Am. Chem. Soc. 107, 4096–4097.
6. Cambray, J., Lane, R. W., Wedd, A. G., Johnson, R. W. & Holm, R. H. (1977)
Inorg. Chem. 16, 2565–2571.
7. Watt, G. D. & Reddy, K. R. N. (1994) J. Inorg. Biochem. 53, 281–294.
8. Angove, H. C., Yoo, S. J., Burgess, B. K. & Münck, E. (1997) J. Am. Chem. Soc.
1997, 8730–8731.
9. Musgrave, K. B., Angove, H. C., Burgess, B. K., Hedman, B. & Hodgson, K. O.
(1998) J. Am. Chem. Soc. 120, 5325–5326.
10. Angove, H. C., Yoo, S. J., Münck, E. & Burgess, B. K. (1998) J. Biol. Chem.
273, 26330–26337.
11. Yoo, S. J., Angove, H. C., Burgess, B. K., Hendrich, M. P. & Münck, E. (1999)
J. Am. Chem. Soc. 121, 2534–2545.
12. Burgess, B. K. & Lowe, D. J. (1996) Chem. Rev. 96, 2983–3011.
13. Guo, M., Sulc, F., Ribbe, M. W., Farmer, P. J. & Burgess, B. K. (2002) J. Am.
Chem. Soc. 124, 12100–12101.
14. Strop, P., Takahara, P. M., Chiu, H.-J., Angove, H. C., Burgess, B. K. & Rees,
D. C. (2001) Biochemistry 40, 651–656.
15. Nyborg, A. C., Johnson, J. L., Gunn, A. & Watt, G. D. (2000) J. Biol. Chem.
275, 39307–39312.
16. Zhou, H.-C. & Holm, R. H. (2003) Inorg. Chem. 42, 11–21.
17. Scott, T. A. & Zhou, H.-C. (2004) Angew. Chem. Int. Ed. 43, 5628–5631.
18. Hagen, K. S., Watson, A. D. & Holm, R. H. (1984) Inorg. Chem. 23, 2984–2990.
19. Berg, J. M. & Holm, R. H. (1982) in Iron–Sulfur Proteins, ed. Spiro, T. G. (Wiley
Interscience, New York), pp. 1–66.
20. Stephan, D. W., Papaefthymiou, G. C., Frankel, R. B. & Holm, R. H. (1983)
Inorg. Chem. 22, 1550–1557.
21. Mascharak, P. K., Hagen, K. S., Spence, J. T. & Holm, R. H. (1983) Inorg. Chim.
Acta 80, 157–170.
22. Carney, M. J., Papaefthymiou, G. C., Whitener, M. A., Spartalian, K., Frankel,
R. B. & Holm, R. H. (1988) Inorg. Chem. 27, 346–352.
23. Carney, M. J., Papaefthymiou, G. C., Frankel, R. B. & Holm, R. H. (1989)
Inorg. Chem. 28, 1497–1503.
24. Kaspar, J. S. & Lonsdale, K. (1967) in International Tables for X-Ray Crystallography (Kynoch, Birmingham, U.K.), 2nd Ed., Vol. 2, pp. 36–49.
25. Yoo, S. J., Angove, H. C., Burgess, B. K., Münck, E. & Peterson, J. (1998) J. Am.
Chem. Soc. 120, 9704–9705.
26. Hauser, C., Bill, E. & Holm, R. H. (2002) Inorg. Chem. 41, 1615–1624.
27. Fomitchev, D. V., McLauchlan, C. C. & Holm, R. H. (2002) Inorg. Chem. 41,
958–966.
28. Silver, J., Fern, G. R., Miller, J. R., McCammon, C. A., Evans, D. J. & Leigh,
G. J. (1999) Inorg. Chem. 38, 4256–4261.
29. Lane, R. W., Ibers, J. A., Frankel, R. B., Papaefthymiou, G. C. & Holm, R. H.
(1977) J. Am. Chem. Soc. 99, 84–98.
30. Goh, C., Segal, B. M., Huang, J., Long, J. R. & Holm, R. H. (1996) J. Am. Chem.
Soc. 118, 11844–11853.
31. Nelson, L. L., Lo, F. Y. K., Rae, A. D. & Dahl, L. F. (1982) J. Organomet. Chem.
225, 309–329.
Summary. To our knowledge, the foregoing structural and spec-
9744 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0504258102
Scott et al.