The first all-cyanide Fe4S4 cluster: [Fe 4S4(CN)4]3-

Article abstract of DOI:10.1002/anie.200460879

A compound with potential: The title compound (see structure) not only resembles the relevant protein active sites geometrically and spectroscopically, but also possesses the least negative [Fe₄S₄(L) ₄]^2-/3- and [Fe₄S₄(L) ₄]_3-/4- (L is a monoanionic ligand) redox potentials of all protein Fe₄S₄ analogues. The value of the Fe ₄S₄^+/0 redox potential implies a new route to the isolation of the elusive Fe₄S₄⁰ cluster.

Full text of DOI:10.1002/anie.200460879

Communications
Bioinorganic Chemistry
The First All-Cyanide Fe4S4 Cluster:
[Fe4S4(CN)4]3ꢀ**
Thomas A. Scott and Hong-Cai Zhou*
In pervasiveness of occurrence and multiplicity of function,
iron–sulfur clusters rival the biological prosthetic groups such
as hemes and flavins.[1] Starting from the first spontaneous
self-assembly of the [Fe4S4(SR)4]2ꢀ cluster in 1972,[2] and the
identification of the protein-bound Fe4S4 in the same year, the
study of iron–sulfur clusters has evolved into a mature field in
which the synthetic inorganic chemistry now resembles the
total synthesis of natural products in organic chemistry.[3]
0
However, the identification of an all-ferrous Fe4S4 state in
the Fe protein of nitrogenase,[4,5] which may have important
implications in the mechanism of nitrogen fixation,[6,7] has
posed a great challenge to inorganic synthetic chemists
because when isolated from the protein environment, the
0
Fe4S4 state is difficult to access chemically. For example, the
midpoint potential of [Fe4S4(SPh)4]3ꢀ/4ꢀ is ꢀ1.72 V versus
saturated calomel electrode (SCE).[8] Holm et al.[9] initiated
the stabilization of the low oxidation states of the Fe4S4 cluster
by using sterically demanding trialkyl phosphine ligands, and
synthesized a number of iron sulfur clusters in the all-ferrous
state. However, the all-ferrous [Fe4S4(PR3)4]0 cluster was
identified only in solution. Attempts to isolate the cluster in
pure form resulted in ligand loss and the formation of higher
nuclearity clusters.[10] To prevent ligand loss, a p-acid ligand
which can bind the Fe4S4 core through s donation and
stabilize low oxidation states by p backbonding is needed.
0
The stabilization of the all-ferrous Fe4S4 core using CO
resulted in an [Fe4S4(CO)12] cluster[11] containing six-coordi-
nate iron atoms and an Fe···Fe separation of 3.47 Š (com-
pared to 2.73 Š in reduced ferredoxin); this all-ferrous cluster
is thus biologically irrelevant. Herein we report an
+
[Fe4S4(CN)4]3ꢀ cluster that resembles the protein Fe4S4
active sites geometrically and spectroscopically, and possesses
[Fe4S4L4]2ꢀ/3ꢀ (L is a mono-anionic ligand) and [Fe4S4L4]3ꢀ/4ꢀ
midpoint potentials of ꢀ0.4 and ꢀ1.38 V versus Ag/AgCl, the
least negative such potentials among all synthetic analogues
of the Fe4S4 protein active sites (For the conversion among
different reference electrodes, see Table 2). These redox
[*] T. A. Scott, Prof. Dr. H.-C. Zhou
Department of chemistry and Biochemistry
Miami University
Oxford, OH 45056 (USA)
Fax: (+01)513-529-8091
E-mail: zhouh@muohio.edu
[**] We thank Dr. Catalina Achim for Mössbauer studies, Dr. Brian
Bennett for collecting and simulating the EPR spectra, and Tracy
Petersen, David Collins for discussion. We also thank Miami
University, the donors of the American Chemical Society Petroleum
Research Fund (PRF No. 39794-G3), and the National Science
Foundation (ECS-0403669) for financial support. H.C.Z. also thanks
the Research Corporation for an Award (RI1188). The X-ray
diffractometer is supported by the NSF (EAR-0003201).
5628
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200460879
Angew. Chem. Int. Ed. 2004, 43, 5628 –5631

Angewandte
Chemie
Table 1: Selected interatomic distances [Š] and angles [8] for 1.
potentials are less negative than expected and imply a
possible new path to a biologically relevant all-ferrous
[Fe4S4L4]4ꢀ cluster, which is a goal in iron–sulfur protein
analogue chemistry.
ꢀ
Fe S
ꢀ
Fe C
ꢀ
C N
Fe···Fe
Fe-C-N
2.6906(6)
2.2791(8)
2.2846(10)
2.2863(7)
2.3003(7)
2.031(3)
1.132(4)
174.6(3)
The reduction of [Fe4S4Cl4]2ꢀ in the presence of CNꢀ
followed by the addition of Na(BPh4) leads to the formation
of [Fe4S4(CN)4]3ꢀ (1; Figure 1). Cluster 1 can also be made in
2.7240(7)
2.016(7)
1.21(1)
180
Fe···Fe 2.726 Š) in C. acidi-urici Fdred (reduced ferredoxin),[13]
and corresponding synthetic analogue [Fe4S4(SPh)4]3ꢀ (Fe S
ꢀ
2.29, Fe···Fe 2.74)[14] One of the Fe-CN groups is perfectly
linear (crystallographically imposed), and the other three
deviate slightly from strict linearity (Fe-C-N 174.6(3)8). An
IR spectrum of 1 shows an n(C/N) stretch at 2103 cmꢀ1, which
is almost the same as the n(C/N) stretch in [FeII(CN)6]4ꢀ
(2098 cmꢀ1) but quite different from that in [FeIII(CN)6]3ꢀ
(2135 cmꢀ1), which indicates a strong p backbonding and the
predominantly ferrous character of 1.[15]
It is informative to compare the geometric parameters of 1
with those of the recently determined crystal structure of the
0
all-ferrous Fe4S4 form of the nitrogenase Fe protein from
Azotobacter vinelandii (Av Fe protein).[16] The average Fe···Fe
0
distance of 2.699 Š is slightly longer than that in the Fe4S4
cluster in Av Fe protein (2.65 Š), implying stronger Fe···Fe
interaction in the all-ferrous state. In contrast, the average
Figure 1. The preparation of the [Fe4S4(CN)4]3ꢀ cluster 1.
ꢀ
Fe S separation of 2.288 Š in 1 is slightly shorter than that in
the all-ferrous Av Fe protein cluster (2.33 Š), which reflects
higher yield by using [Fe4S4(PiPr3)4]+ as the starting cluster.[10]
The reaction between [Fe4S4Cl4]2ꢀ and [Bu4N][CN] gave an
unidentified black solid, which showed none of the electro-
chemical characteristics of 1 or of the one-electron oxidized
form of 1. Thus, the substitution of the Clꢀ ion for the CNꢀ ion
the elongation of the Fe S bonds in a Fe4S4 cluster upon
ꢀ
+
0
ꢀ
reduction from Fe4S4 to Fe4S4 . This elongation of Fe S
bonds has also been observed upon the reduction of the
2+
+
Fe4S4 state to the Fe4S4 state in other synthetic ana-
logues.[14] The exact iron oxidation states in 1 have also been
confirmed by electrochemical, EPR, Mössbauer, and UV/Vis
studies.
2+
needs to occur after the one-electron reduction of the Fe4S4
core. This result is consistent with the observation that the
reduction of the Fe4S4 core facilitates ligand substitution.[9,10]
Compound 1 crystallizes in space group R3c with the N2,
As shown in Figure 3, the cyclic voltammogram of 1 using
glassy carbon as the working electrode shows two reversible
electrochemical pairs. For clarity, the redox potentials have all
been converted into those versus normalized hydrogen
electrode (NHE) and are listed in Table 2. In the following
C2, Fe2, and S2 atoms residing on
a threefold axis
(Figure 2).[12] Every unit cell contains six formula units of 1.
The cluster anion has approximate Td symmetry, but is slightly
elongated along the crystallographic threefold axis. Selected
interatomic distances and angles are listed in Table 1.
ꢀ
The Fe S and Fe···Fe separations of 1 (Table 1) closely
+
ꢀ
resemble those of the protein Fe4S4 core (Fe S 2.281, 2.285;
Figure 3. Cyclic voltammogram (100 mVsꢀ1) of [Fe4S4(CN)4]3ꢀ in 0.1m
solution of [Bu4N][PF6] in CH2Cl2. Peak potentials are those versus Ag/
AgCl (E1/2 versus Fc/Fc+ 0.69 V; Fc=[(C5H5)2Fe]) using a glassy carbon
working electrode. Potentials versus NHE can be found in Table 2.
Figure 2. The structure of [Fe4S4(CN)4]3ꢀ thermal ellipsoids set at 50%
probability (see Table 1).
Angew. Chem. Int. Ed. 2004, 43, 5628 –5631
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5629

Communications
cally, and has the least negative [Fe4S4L4]2ꢀ/3ꢀ and [Fe4S4L4]3ꢀ/4ꢀ
redox potentials among all known synthetic analogues.
Table 2: Comparison of redox potentials (V) of 1 with those of the
relevant Fe4S4 cores.[a]
L
E1/2 [Fe4S4L4]2ꢀ/3ꢀ
E1/2 [Fe4S4L4]3ꢀ/4ꢀ
Ref.
CNꢀ
ꢀ0.18
ꢀ0.41
ꢀ0.76
ꢀ1.16
this work
[19]
[8,19]
Experimental Section
ꢀSC6H4-p-NO2
ꢀSPh
1 Method A: A black crystalline solid of (Bu4N)2[Fe4S4Cl4] (0.293 g,
0.3 mmol) was suspended in THF (2.5 mL) and mixed with a [Bu4N]
[CN] (0.322 g, 1.2 mmol) solution in THF (2.5 mL). Freshly made
potassium benzophenoketyl (0.33 mmol) in THF (2 mL) was added
drop-wise to the mixture, and the mixture was stirred for 15 min. The
addition of Na(BPh4) (0.308 g, 0.3 mmol) in THF (2 mL) caused
precipitation of a white solid, presumably NaCl. The resulting
suspension was stirred for 1.5 h and dried under vacuum to yield a
black solid. The solid was extracted with MeCN (5 mL), and the
extraction was filtered through Celite. Diethyl ether vapor diffusion
into the filtrate caused the precipitation of a black solid and colorless
crystals within three days. The compound was purified through re-
crystallization from MeCN/diethyl ether with minimal amounts of
MeCN to dissolve only the dark solid. Black block crystals of 1
formed after three crystallizations. Yield: 0.145 g (41%).
ꢀ1.48
ꢀ0.8
Enzyme:
Ferredoxin
Av Nitrogenase
Fe protein
ꢀ0.4
[17]
[18]
[a] All redox potentials are those versus NHE. Standard potentials (V) in
aqueous solutions at 258C versus NHE: SCE+0.2415; Ag/AgCl+
0.2223.[24]
discussion, potentials are those versus NHE only. Cluster 1
possesses midpoint potentials of ꢀ0.18 and ꢀ1.16 V. The
former corresponds to that of the [Fe4S4L4]2ꢀ/3ꢀ redox pair,
which is less negative than that of Fdox/Fdred (ꢀ0.4 V) in native
proteins;[17] the latter represents E1/2 of the [Fe4S4L4]3ꢀ/4ꢀ
redox pair, the closest to that of iron protein of nitrogenase
(ꢀ0.8 V).[18] Previously, the least negative [Fe4S4L4]2ꢀ/3ꢀ
potential in synthetic analogues was ꢀ0.41 V in
[Fe4S4(SC6H4-p-NO2)4]2ꢀ,[19] and that for [Fe4S4L4]3ꢀ/4ꢀ was
ꢀ1.48 V in [Fe4S4(SPh)4]2ꢀ.[8,19]
Method B: A black crystalline solid of (BPh4)[Fe4S4(PiPr3)4]
(0.394 g, 0.3 mmol) was dissolved in THF (2.5 mL), and mixed with a
[Bu4N][CN] (0.322 g, 1.2 mmol) solution in THF (2.5 mL). The
mixture was stirred for 2 h to form a suspension, and then allowed
to settle. The supernatant solution was decanted and the solid was
washed with aliquots of THF and dried under vacuum. The black
residue was dissolved in MeCN (8 mL) and crystallized from MeCN/
diethyl ether yielding black block crystals of 1. Yield: 0.240 g (67%).
Elemental analysis calcd (%) for C52H108N4Fe4S4: C 52.79, H 9.20, N
8.28, S 10.84; found: C 52.57, H 9.22, N 7.98, S 10.87.
Using platinum instead of glassy carbon as the working
electrode, the peak currents of 1 diminish progressively when
a multiple scan is performed. Whether this is due to cyanide
adsorption and polymerization on the platinum surface[20] or
interaction between the clusters and the platinum surface
needs further investigation.
Preliminary EPR (axial, g2 = 2.114, gz = 1.897 at 9 K) and
Mössbauer (d at 4.2 K relative to Fe metal: 0.503, 0.566; DEq:
1.329, 1.869) studies have confirmed the oxidation state of 1.
Received: June 4, 2004
Revised: July 15, 2004
Keywords: bioinorganic chemistry · cyanides · iron ·
.
nitrogenases · sulfur
[1] P. V. Rao, R. H. Holm, Chem. Rev. 2004, 104, 527 – 559.
[2] T. Herskovitz, B. A. Averill, R. H. Holm, J. A. Ibers, W. D.
Phillips, J. F. Weiher, Proc. Natl. Acad. Sci. USA 1972, 69, 2437 –
2441.
[3] R. H. Holm in Comprehensive Coordination Chemistry II, Vol. 8
(Eds.: J. A. McCleverty, T. J. Meyer), New York, 2004,
61 – 90.
[4] G. D. Watt, K. R. N. Reddy, J. Inorg. Biochem. 1994, 53, 281 –
294.
[5] H. C. Angove, S. J. Yoo, B. K. Burgess, E. Münck, J. Am. Chem.
Soc. 1997, 119, 8730 – 8731.
[6] B. K. Burgess, D. J. Lowe, Chem. Rev. 1996, 96, 2983 – 3011.
[7] A. C. Nyborg, J. L. Johnson, A. Gunn, G. D. Watt, J. Biol. Chem.
2000, 275, 39307 – 39312.
[8] J. Cambray, R. W. Lane, A. G. Wedd, R. W. Johnson, R. H.
Holm, Inorg. Chem. 1977, 16, 2565 – 2571.
[9] C. Goh, B. M. Segal, J. Huang, J. R. Long, R. H. Holm, J. Am.
Chem. Soc. 1996, 118, 11844 – 11853.
[10] H.-C. Zhou, R. H. Holm, Inorg. Chem. 2003, 42, 11 – 21.
[11] L. L. Nelson, F. Y. K. Lo, A. D. Rae, L. F. Dahl, J. Organomet.
Chem. 1982, 225, 309 – 329.
+
The spectroscopic features are very similar to those of Fe4S4
proteins and their corresponding synthetic analogues,[14] but
the final answer for the ground state needs high-field
Mössbauer and variable-temperature EPR studies (works
currently under way and will be presented elsewhere). In
addition, the absorption spectrum of 1 in MeCN is featureless,
in agreement with those of all other proteins and synthetic
analogues in the Fe4S4+ core oxidation state.[1]
The closest examples in the literature to 1 are the
octanuclear complexes [Fe4S4(NC-MLn)4]2ꢀ (M = W and
Mn),[21,22] where the cyanide bridges were reversed and act
as linkers between an Fe4S4 cluster and four other metal
cluster units. The core oxidation state was Fe4S42+. The key for
1 remaining monomeric is the reduction of the cluster from
2+
+
Fe4S4 to Fe4S4 . It is now possible to use 1 as a secondary
building unit to clusters of higher nuclearities or coordination
polymers by simply oxidizing 1 in the presence of other metal
species with open coordination sites.
A monosubstituted cyanide cluster, [Fe4S4(LS3)(CN)]2ꢀ,
was also detected spectroscopically,[23] but 1 represents the
first all-cyanide Fe4S4 cluster.
[12] Crystal structure determination of 1: C52H108Fe4N7S4, Mr =
1183.09; a black crystal (0.30 ” 0.27 ” 0.15 mm) mounted on a
Pyrex fiber under a cold stream of N2, T= 173(2) K, l(MoKa) =
0.71073 Š, rhombohedral, space group R3c, a = 16.9244(3), c =
38.4541 (14) Š, a = b = 908, g = 1208, V= 9538.9(4) Š3, Z = 6,
In summary, we have isolated and characterized the first
all-cyanide Fe4S4 cluster, which resembles the protein Fe4S4
clusters geometrically, electrochemically, and spectroscopi-
, , empirical
1calcd = 1.236 gcmꢀ3 2qmax = 56.68, m = 1.062 mmꢀ1
absorption correction using SADABS, F(000) = 3822, 22584
5630
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org Angew. Chem. Int. Ed. 2004, 43, 5628 –5631

Angewandte
Chemie
reflections, 3859 unique reflections, 314 parameters, 1 restraint,
GOF = 1.151, R1 = 0.0339 [I > 2s(I)], wR2 = 0.0825, min./max.
residual electron density ꢀ0.265/0.584 eŠꢀ3, refined on j F2 j .
CCDC-240511 (1) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cam-
bridge Crystallographic Data Centre, 12 Union Road, Cam-
bridge CB21EZ, UK; fax: (+ 44)1223-336-033; or deposit@
ccdc.cam.ac.uk).
[13] Z. Dauter, K. S. Wilson, L. C. Sieker, J. Meyer, J. M. Moulis,
Biochemistry 1997, 36, 16065 – 16073.
[14] E. J. Laskowski, R. B. Frankel, W. O. Gillum, G. C. Papaefthy-
miou, J. Renaud, J. A. Ibers, R. H. Holm, J. Am. Chem. Soc.
1978, 100, 5322 – 5337.
[15] K. R. Dunbar, R. A. Heintz in Progress in Inorganic Chemistry,
Vol. 45 (Ed.: K. D. Karlin), Wiley, New York, 1997, pp. 283 –
391.
[16] P. Strop, P. M. Takahara, H.-J. Chiu, H. C. Angove, B. K.
Burgess, D. C. Rees, Biochemistry 2001, 40, 651 – 656.
[17] J. M. Berg, R. H. Holm in Iron-Sulfur Proteins (Ed.: T. G. Spiro),
Wiley, New York, 1982, pp. 1 – 66.
[18] M. Guo, F. Sulc, M. W. Ribbe, P. J. Farmer, B. K. Burgess, J. Am.
Chem. Soc. 2002, 124, 12100 – 12101.
[19] C. Zhou, J. W. Raebiger, B. M. Segal, R. H. Holm, Inorg. Chim.
Acta 2000, 300–302, 892 – 902.
[20] G. Baltrunas, E. Morkyavichius, T. Yankauskas, Russ. J. Electro-
chem. 1998, 33, 572 – 575.
[21] N. Zhu, J. Pebler, H. Vahrenkamp, Angew. Chem. 1996,
108, 984 – 985; Angew. Chem. Int. Ed. Engl. 1996, 35, 894 –
895.
[22] N. Zhu, R. Appelt, H. Vahrenkamp, J. Organomet. Chem. 1998,
565, 187 – 192.
[23] R. H. Holm, S. Ciurli, J. A. Weigel, in Progress in Inorganic
Chemistry, Vol. 38 (Ed.: S. J. Lippard), Wiley, New York, 1990,
pp. 1 – 74.
[24] A. J. Bard, L. R. Faulkner, Electrochemical Methods-Fundamen-
tals and Applications, 2nd ed., Wiley, New York, 2001, pp. 808 –
809.
Angew. Chem. Int. Ed. 2004, 43, 5628 –5631
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5631