Comparison of structurally-related alkoxide, amine, and thiolate-ligated MII (M = Fe, Co) complexes: The influence of thiolates on the properties of biologically relevant metal complexes

Article abstract of DOI:10.1016/j.ica.2007.07.038

Mechanistic pathways of metalloenzymes are controlled by the metal ion's electronic and magnetic properties, which are tuned by the coordinated ligands. The functional advantage gained by incorporating cysteinates into the active site of non-heme iron enzymes such as superoxide reductase (SOR) is not entirely understood. Herein, we compare the structural and redox properties of a series of structurally-related thiolate, alkoxide, and amine-ligated Fe(II) complexes in order to determine how the thiolate influences properties critical to function. Thiolates are shown to reduce metal ion Lewis acidity relative to alkoxides and amines, and have a strong trans influence thereby helping to maintain an open coordination site. Comparison of the redox potentials of the structurally analogous compounds described herein shows that alkoxide ligands favor the higher-valent Fe³⁺ oxidation state, amine ligands favor the reduced Fe²⁺ oxidation state, and thiolates fall somewhere in between. These properties provide a functional advantage for substrate reducing enzymes in that they provide a site at the metal ion for substrate to bind, and a moderate potential that facilitates both substrate reduction and regeneration of the catalytically active reduced state. Redox potentials for structurally-related Co(II) complexes are shown to be cathodically-shifted relative to their Fe(II) analogues, making them ineffective reducing agents for substrates such as superoxide.

Full text of DOI:10.1016/j.ica.2007.07.038

Available online at www.sciencedirect.com
Inorganica Chimica Acta 361 (2008) 1070–1078
www.elsevier.com/locate/ica
Comparison of structurally-related alkoxide, amine, and
thiolate-ligated MII (M = Fe, Co) complexes: The influence of
thiolates on the properties of biologically relevant metal complexes
Lisa M. Brines, Gloria Villar-Acevedo, Terutaka Kitagawa, Rodney D. Swartz,
1
1
*
Priscilla Lugo-Mas, Werner Kaminsky , Jason B. Benedict , Julie A. Kovacs
Department of Chemistry, University of Washington, P.O. Box 351700, Seattle, WA 98195-1700, United States
Received 3 July 2007; accepted 31 July 2007
Available online 15 August 2007
Dedicated to Edward Solomon
Abstract
Mechanistic pathways of metalloenzymes are controlled by the metal ion’s electronic and magnetic properties, which are tuned by the
coordinated ligands. The functional advantage gained by incorporating cysteinates into the active site of non-heme iron enzymes such as
superoxide reductase (SOR) is not entirely understood. Herein, we compare the structural and redox properties of a series of structurally-
related thiolate, alkoxide, and amine-ligated Fe(II) complexes in order to determine how the thiolate influences properties critical to func-
tion. Thiolates are shown to reduce metal ion Lewis acidity relative to alkoxides and amines, and have a strong trans influence thereby
helping to maintain an open coordination site. Comparison of the redox potentials of the structurally analogous compounds described
herein shows that alkoxide ligands favor the higher-valent Fe3+ oxidation state, amine ligands favor the reduced Fe2+ oxidation state,
and thiolates fall somewhere in between. These properties provide a functional advantage for substrate reducing enzymes in that they
provide a site at the metal ion for substrate to bind, and a moderate potential that facilitates both substrate reduction and regeneration
of the catalytically active reduced state. Redox potentials for structurally-related Co(II) complexes are shown to be cathodically-shifted
relative to their Fe(II) analogues, making them ineffective reducing agents for substrates such as superoxide.
ꢀ 2007 Elsevier B.V. All rights reserved.
Keywords: Synthetic thiolate-ligated iron complexes; Non-heme iron; X-ray crystal structures; Redox properties; Bioinorganic; Metalloenzyme function
1. Introduction
bound substrates. Our group has put considerable effort
into determining how coordinated cysteinate residues
might control the chemistry taking place at the active-site
of a metalloenzymes. Cysteinates are known to form highly
covalent bonds to transition metals [17,18] and facilitate
electron transfer reactions [19,20]. Our work has shown
that thiolate ligands make low-spin iron accessible even
in a non-heme environment [21–23], stabilize iron in the
+3 oxidation state [21–23], and labilize sites trans to the
thiolate thereby promoting reactivity by releasing product
[24], even with low-spin Co(III) which is typically inert
[25,26]. Recently, we showed that the thiolate ligand of a
biomimetic superoxide reductase (SOR) model, [FeII-
(cyclam-PrS)]+, and its trans positioning relative to the
Cysteinate-ligated metalloenzymes and metalloproteins
promote a number of critical biological processes [1–16].
Mechanistic pathways are controlled by the metal ion’s
electronic and magnetic properties, which are tuned by
the coordinated ligands, and subtly altered by H-bonding
interactions within the protein. Supporting ligands play
an important role in determining function by controlling
a metal ion’s redox potential, as well as the basicity of
*
Corresponding author.
E-mail address: kovacs@chem.washington.edu (J.A. Kovacs).
UW staff crystallographers.
1
0020-1693/$ - see front matter ꢀ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2007.07.038

L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
1071
substrate binding site, contribute significantly to the cata-
lyst’s function by dramatically lowering the Fe–O stretch-
ing frequency (mFe–O = 419 cmꢀ1) and force constant (k =
1.20 mdynes/cm) well-below that of any other reported
iron-peroxo, thereby favoring Fe–O bond cleavage and
H2O2 release [24]. Superoxide reductase, peptide deformy-
lase, and nitrile hydratase occupy an unique niche amongst
known non-heme enzymes in that they contain cysteinate
ligands, as opposed to the more frequently encountered
nitrogen and oxygen based ligands. The functional advan-
tage gained by incorporating cysteinates is not entirely
understood. By systematically replacing a single atom or
functional group within identical ligand frameworks and
comparing the structural and redox properties of the result-
ing iron complexes, we hope to address this question. For
this purpose, we synthesized alkoxide and amine analogues
of our functional SOR models [FeII(SMe2N4(tren))]+ (1;
tren = tris (2-aminoethyl)amine) and [FeII(cyclam-PrS)]+
(5). Herein, we examine and compare the structures, and
electronic, magnetic, and redox properties of this series of
structurally related complexes. We also compare primary
versus secondary versus tertiary amines, and structurally
analogous Co(II) and Fe(II) complexes in order to deter-
mine how the metal ion, amine basicity, and steric bulk
influences key properties critical to metalloenzyme
function.
spectrometer. Magnetic moments were determined using
a Quantum Design SQUID Magnetometer and by Evans’
method as modified for a superconducting solenoid
[31,32]. Temperatures were obtained using Van Geet’s
method [33]. Elemental analyses were performed by
Galbraith Laboratories (Knoxville, TN) and Atlantic
Microlab Inc. (Norcross, GA).
2.2. Synthesis of [FeII(NEt2N4(tren))Cl](PF6) (3)
A solution of tris(2-aminoethyl) amine (tren) (0.584 g,
4 mmol) was dissolved in 10 ml of dry methanol, followed
by the addition of diethylaminoacetone (0.532 g,
4.00 mmol). This solution was allowed to stir for 10 min.
Fe(II)Cl2 (0.504 g, 4.00 mmol) was dissolved in 5 ml of
dry methanol and added to the ligand mixture dropwise,
immediately producing
a yellow precipitate. NaPF6
(1.34 g, 8.00 mmol) was added and the reaction mixture
was allowed to stir overnight. The precipitate was filtered
and washed with methanol to yield a pale yellow solid.
Yield: 680 mg (53%). Yellow needle-like crystals were
obtained by slow diffusion of diethyl ether into an acetoni-
trile solution. E1/2 (MeCN, 298 K) = 303 mV versus SCE.
Solution magnetic moment (303 K, MeCN) leff
=
5.37 BM. Magnetic moment in solid state: leff = 5.23 BM.
Elemental Anal. Calc. for FeN5C13H31ClPF6: C, 31.63;
N, 14.19; H, 6.33. Found C, 31.68; N, 14.17; H, 6.40%.
2. Experimental
2.3. Synthesis of [CoII(SMe2N4(tren))](Cl) (4)
2.1. General methods
Anhydrous cobaltous chloride (519 mg, 4.00 mmol) in
methanol (20 ml) was added dropwise to a solution
containing NaOCH3 (216 mg, 4.00 mmol), 3-methyl-3-mer-
capto-2-butanone (472 mg, 4 mmol), tren (585 mg,
4.00 mmol) and left to stir overnight. The resulting green
solution was dried under reduced pressure, re-dissolved
with a minimal amount of acetonitrile, and filtered through
a fritted funnel. The solution was layered with diethyl ether
and left to re-crystallize overnight giving pure 4 (700 mg,
52% yield) as a crystalline solid. Electronic absorption
(CH3CN): kmax (e) = 362 (1275) nm, 456 (243) nm, 504
(204) nm, 595 (179) nm. Redox potential (MeCN versus
SCE): Epa = +270 mV, Epc = ꢀ729 mV. Elemental Anal.
Calc. for C11H25ClCoN4S: C, 38.88; H, 7.42; N, 16.49.
Found: C, 38.64; H, 7.34; N, 16.69%.
All reactions were performed using standard Schlenk
techniques under an atmosphere of dinitrogen. The iron
complexes [FeII(SMe2N4(tren))](PF6) (1) [27], [FeII(OMe2N4-
(tren))(H2O)](PF6) (2) [28], [FeII(cyclam-PrS)](BPh4) (5)
[24], and [FeII(SMe2N4(tren-Et4))]+ (8) [29], as well as 3-
Boc-cyclam and cyclam-PrSAc Æ 4HCl [24], were synthe-
sized as previously described. All other reagents were
obtained from Aldrich and used without further purifica-
tion. Solvents were purified through standard procedures
[30]. 3-Methyl-3-mercapto-2-butanone was prepared
according to a published procedure [23]. NMR spectra
were recorded using Bruker AV-300, AV-301, or DPX
500 spectrometers and referenced to the residual protio sol-
vent. 1H NMR chemical shifts (d) are reported in parts per
million (ppm) and coupling constants (J) are reported in
Hz. IR spectra were recorded on a Perkin-Elmer 1720
FT-IR. Electrospray mass spectra were recorded on a Bru-
ker Esquire LC-Ion Trap Mass Spectrometer. EPR spectra
were obtained using a Varian CW-EPR spectrometer at
4 K equipped with an Oxford helium cryostat. Cyclic vol-
tammograms were recorded on a PAR 263A potentiostat
with a glassy carbon working electrode, a platinum counter
electrode, and Ag/Ag+ (0.1 M) in acetonitrile as the
reference electrode. Potentials were then converted to val-
ues referenced versus SCE. Electronic absorption spectra
were recorded using a Hewlett–Packard 8453 diode array
2.4. Synthesis of CyclamPrOH Æ 4HCl
2.4.1. Synthesis of 3-iodo-1-propanol
To a 50 ml round bottom flask was added 3-chloro-1-
propanol (2.93 g, 316 mmol) and sodium iodide (9.47 g,
632 mmol) in 20 ml acetone. The solution was refluxed
overnight. After cooling to room temperature, the solution
was filtered and the solvent was removed by rotary evapo-
ration. The residue was redissolved in methylene chloride,
filtered again, and the solvent was removed. The resulting
red-orange oil was redissolved in 5 ml methylene chloride

1072
L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
and passed through a small silica plug (100% CH2Cl2). The
first several fractions were collected and evaporated to give
a brown oil (3.06 g, 52% yield). 1H NMR (CDCl3): 3.73 (q,
2H, J = 5.70 Hz), 3.30 (t, 2H, J = 6.60 Hz), 2.13 (s, 1H),
2.06 (quintet, 2H, J = 6.00 Hz).
2.6. Synthesis of [CoII(cyclam-PrS)][BPh4] Æ MeCN (7)
Sodium methoxide (170 mg, 5.00 mmol) and cyclam-
PrSAc Æ 4HCl (200 mg, 1.00 mmol) were dissolved in
15 ml of methanol, allowed to stir for 30 min and cooled
to ꢀ30 ꢁC. CoCl2 Æ 6H2O (82 mg, 1.00 mmol) was dissolved
in 5 ml of methanol, cooled to ꢀ30 ꢁC and added dropwise
to the cold ligand while stirring. A pink solution instantly
formed. Sodium tetraphenylborate (216 mg, 1.00 mmol)
dissolved in 10 ml of methanol, was then added and the
reaction mixture was stirred overnight. The solution was
filtered, the filtrate was concentrated to 5 ml, layered with
ether (25 ml) and cooled to ꢀ35 ꢁC to yield pink crystals.
Yield: 65 mg (46.5%). Electronic absorption (CH3CN):
kmax (e) = 324 (2029) nm, 366 (975) nm, 549 (116) nm;
(MeOH) kmax (e) = 317 (2348) nm, 359 (1128) nm, 557
(134) nm. Epa (MeCN) = +0.900 V versus SCE. Solution
magnetic moment (310.2 K; MeOH) leff = 3.97 BM. EPR
DCM/toluene glass (1:1), 10.3 K: g1 = 4.99, g2 = 3.93,
g3 = 2.61. ESI-MS cald. for [CoN4SC13H29]+: 332.8; found
332.2. Elemental Anal. Calc. for CoC37H49N4SB: C, 66.30;
N, 8.60; H, 7.62. Found C, 66.07, N, 8.43, H, 7.72%.
2.4.2. Synthesis of 3Boc-cyclamPrOH
In a 50 ml round bottom flask was added 3-iodo-1-propa-
nol (1.79 g, 9.62 mmol), 3Boc-cyclam (2.41 g, 4.81 mmol),
and K2CO3 (1.33 g, 9.62 mmol) in 20 ml DMF. The solution
was stirred under nitrogen at room temperature for 9 days.
100 ml of DI water was added to the solution and the prod-
uct was extracted with ethyl acetate (3 · 30 ml). The organic
layer was then washed with brine (2 · 30 ml) and dried over
MgSO4. Evaporation led to a sticky white residue of the
product and excess 3-iodo-1-propanol. TLC revealed that
the products would be difficult to separate, so the Boc
deprotection step was carried out on the mixture. ESI-MS
(M+1) calc. for M = [N4C28H54O7]: 558.7, found: 559.9.
2.4.3. Synthesis of CyclamPrOH Æ 4HCl
To the residue formed in the above step was added 15 ml
of 4 M HCl/dioxane. The solution was allowed to stir at
room temperature overnight. The solvent was removed
by rotary evaporation and the residue was taken up in a
MeOH/Et2O solution (1:3). The solution was filtered and
the precipitate was washed with additional ether
(2 · 15 ml). The solid was further purified by triturating
for 2 days in an Et2O/CH2Cl2 mixture (1:1). The residue
was then filtered, washed with additional ether, and
pumped dry under vacuum. Yield: 1.92 g (ꢁ100%). 1H
NMR (D2O): 3.65 (t, 2H, J = 5.70 Hz), 3.58–3.15 (m,
20H), 2.03 (m, 4H), 1.87 (m, 2H). ESI-MS (M+1) calc.
for M = N4C13H30O: 258.4; found: 259.3.
2.7. X-ray crystallography
A bi-refringent yellow crystal prism of 3 was cut down
to 0.24 · 0.12 · 0.12 mm and mounted on a glass capillary
with oil. Data was collected at 23 ꢁC on a Nonius Kappa
CCD diffractometer. The crystal-to-detector distance was
set to 30 mm and exposure time was 15 seconds per degree
for all data sets with a scan width of 1.8ꢁ. The data collec-
tion was 98.6% complete to 24.71ꢁ in #. A total of 20518
partial and complete reflections were collected covering
the indices, h = ꢀ15 to 16, k = ꢀ21 to 21, l = ꢀ10 to 10.
Reflections (3472) were symmetry independent and the
Rint = 0.116 indicated that the data was of less than aver-
age quality (average quality = 0.07). Indexing and unit cell
refinements indicated an orthorhombic P lattice in the
space group Pna21 (no. 33). Data was collected at room
temperature because a slow splitting of peaks was observed
upon cooling the crystal. The splitting of peaks appeared to
indicate that a low temperature crystallographic phase
transition was occurring, though no structure of this phase
could be obtained. This does, however, explain the large
thermal parameters of several atoms in the structure. Fur-
thermore, the BASF parameter indicated the crystal was
comprised of approximately 46% of the inversion twin of
this structure.
A red prism, size 0.44 · 0.38 · 0.27 mm of 4 was
mounted on a glass capillary with oil. Data was collected
at ꢀ143 ꢁC. on a Nonius Kappa CCD diffractometer.
The crystal-to-detector distance was set to 30 mm and
exposure time was 15 seconds per degree for all data sets
with a scan width of 2ꢁ. The data collection was 99.8%
complete to 25ꢁ in #. A total of 20162 partial and complete
reflections were collected covering the indices, h = ꢀ13 to
13, k = ꢀ15 to 17, l = ꢀ15 to 15. Reflections (3700) were
2.5. Synthesis of [FeII(cyclam-PrO)](BPh4) Æ MeOH (6)
Sodium methoxide (270 mg, 5.00 mmol) and cyclam-
PrOH Æ 4HCl (404 mg, 1.00 mmol) were dissolved in
15 ml methanol and allowed to stir for 30 min. A solution
of FeCl2 (127 mg, 1.00 mmol) in methanol (10 ml) was
added dropwise to the ligand solution at room tempera-
ture. Sodium tetraphenylborate (342 mg, 1.00 mmol) was
added and the solution was allowed to stir overnight.
The solution was then filtered through Celite and the meth-
anol was evaporated. The residue was redissolved in 20 ml
acetonitrile and filtered again. The filtrate was concen-
trated to 10 ml, layered with ether (35 ml), and cooled to
ꢀ40 ꢁC to yield white solid 6. Yield: 368 mg (58.2%). Single
crystals suitable for X-ray diffraction were grown at room
temperature over several days by layering ether (20 ml)
onto a methanolic solution (10 ml) of 6. ESI-MS calc. for
[FeN4OC13H29]+: 313.2; found: 312.9. Magnetic moment
(solid state): 4.79 BM. Redox potential (MeCN versus
SCE): E1/2 = +35 mV. Elemental Anal. Calc. for FeC38H53-
N4O2B: C, 68.68; H, 8.04; N, 8.43. Found: C, 68.63; H,
7.93; N, 8.40%.

L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
1073
Table 1
Crystal data, intensity collectionsa and structure refinement parameters for [FeII(NEt2N4(tren))(Cl)](PF6) (3), [CoII(SMe2N4(tren))](Cl) (4), [FeII(cyclam-
PrO)](BPh4)MeOH (6), and [CoII(cyclam-PrS)](BPh4)MeCN (7)
3
4
6
7
Formula
MW
T (K)
FeC13H31ClF6N5P
493.70
296
CoC11H25ClNS
339.79
130(2)
FeC38H53BN4O2
664.50
130(2)
CoC39H52BN5S
692.66
130(2)
Unit cell
Orthorhombic
13.6500(4)
18.6480(6)
8.575(1)
90
90
90
2182.7(3)
4
1.502
Monoclinic
10.3700(3)
12.9290(4)
11.5130(4)
90
94.654(1)
90
1538.50(8)
4
Monoclinic
8.1770(6)
12.7180(6)
17.977(1)
90
132.2251(8)
90
3535.3(1)
4
Monoclinic
9.2860(2)
19.6960(5)
21.3530(7)
90
111.338(1)
90
1863.4(2)
4
˚
a (A)
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
V (A )
Z
d(calc) (g/cm3)
1.467
1.248
1.265
Space group
R
Pna21
P21/c
P21/c
P21/c
0.0500b
0.1284c
0.998
0.0443b
0.1180c
1.094
0.0548b
0.1549c
0.982
0.0553b
0.1238d
0.997
Rw
GOF
a
˚
Mo Ko(a, ꢁ) (k = 0.71070 A) radiation; graphite monochromator; ꢀ90 ꢁC.
R = RiFoj ꢀ jFci/RjFoj.
b
c
2
1=2
2
Rw ¼ ½RwðjF oj ꢀ jF cjÞ =RwF 2oꢂ , where wꢀ1 ¼ ½rc2ount þ ð0:05 F 2Þ ꢂ=4F 2.
2
2
1=2
2
d
Rw ¼ fR½wðF 2o ꢀ F 2c Þ ꢂ=R½wðF o2Þ ꢂg ; w ¼ 1=½r2ðF o2Þ þ ð0:1395PÞ þ 1:3703Pꢂ, where P ¼ ½F o2 þ 2F 2cꢂ=3.
Table 2
Selected geometrical parameters for the cations of [FeII(SMe2N4(tren))](PF6) (1), [FeII(OMe2N4(tren))(H2O)](PF6) (2), [FeII(NEt2N4(tren)(Cl))](PF6) (3),
[CoII(SMe2N4(tren))](Cl) (4), [FeII(cyclam-PrS)](BPh4) (5), [FeII(cyclam-PrO)](BPh4)MeOH (6), [CoII(cyclam-PrS)](BPh4) Æ MeCN (7), and [FeII(SMe2N4-
(tren-Et4))]+ (8)
1
2
3
4
5
6
7
8
M–X(S,O,N)
M–N(1)
M–N(2)
M–N(3)
M–N(4)
M–L
2.3281(9)
2.091(3)
2.268(3)
2.131(3)
2.117(3)
N/A
2.001(1)
2.129(2)
2.257(2)
2.249(2)
2.222(2)
2.149(1)
2.293(5)
2.131(5)
2.267(5)
2.215(15)
2.232(14)
2.424 (2)
2.3006(7)
2.050(2)
2.223(2)
2.086(2)
2.107(2)
N/A
2.286(1)
2.181(4)
2.138(4)
2.174(4)
2.166(4)
N/A
1.906(2)
2.225(2)
2.206(2)
2.151(2)
2.142(2)
N/A
2.263(1)
2.221(2)
2.099(3)
2.175(2)
2.086(3)
N/A
2.3171(9)
2.104(3)
2.227(3)
2.219(3)
2.220(3)
N/A
X–M–N(1)
X–M–N(2)
X–M–N(3)
X–M–N(4)
N(1)–M–N(2)
N(1)–M–N(3)
N(1)–M–N(4)
N(2)–M–N(3)
N(3)–M–N(4)
s-value
84.02(8)
163.02(7)
108.8(8)
107.9(9)
79.9(1)
124.9(1)
110.3(1)
79.9(1)
115.2(1)
0.64
76.49(6)
153.14(6)
105.45(6)
108.00(6)
76.71(7)
94.77(6)
103.51(7)
78.57(7)
144.73(7)
N/A
75.1(2)
152.3(2)
107.0(6)
104.7(6)
77.2(2)
97.8(5)
92.5(5)
76.7(6)
148.2(2)
N/A
84.81(6)
163.96(6)
106.90(6)
112.4(2)
79.49(8)
125.14(8)
122.54(8)
79.93(8)
102.54(8)
0.65
101.2(1)
112.0(1)
104.9(1)
101.9(1)
90.4(2)
153.7(2)
82.5(2)
82.8(2)
146.1(2)
0.13
98.77(8)
95.70(8)
110.54(8)
121.16(8)
165.45(8)
92.37(8)
82.21(8)
81.09(9)
128.26(8)
0.62
99.07(7)
141.68(8)
91.29(8)
112.61(8)
91.9(1)
169.6(1)
84.3(1)
80.1(1)
91.3(1)
0.47
83.41(7)
161.53(7)
105.63(8)
109.64(7)
78.6(1)
121.4(1)
111.4(1)
81.3(1)
118.49(9)
0.67
˚
Bond lengths and bond angles are given in A and degrees, respectively.
X = S, N, or O.
L = sixth ligand (O or Cl).
symmetry independent and the Rint = 0.0366 indicated that
the data was excellent (average quality = 0.07). Indexing
and unit cell refinements indicated a monoclinic lattice in
the space group P21/c (no. 14).
A yellow crystal prism size 0.54 · 0.34 · 0.19 mm of 6
was mounted on a glass capillary with oil. Data were col-
lected at ꢀ143 ꢁC on a Nonius Kappa CCD diffractometer.
The crystal-to-detector distance was set to 30 mm and
exposure time was 20 seconds per degree for all data sets
with a scan width of 1.1ꢁ. The data collection was 99.0%
complete to 25ꢁ in #. A total of 78675 partial and complete
reflections were collected covering the indices, h = ꢀ 19 to
18, k = ꢀ12 to 12, l = ꢀ36 to 43. Reflections (8278) were
symmetry independent and the Rint = 0.0829 indicated that
the data was of average quality (average quality = 0.07).
Indexing and unit cell refinements indicated a primitive
monoclinic lattice in the space group P21/c (no. 14).
A
purple crystal plate of 7 was cut down to
0.29 · 0.12 · 0.12 mm and mounted on a glass capillary
with oil. Data was collected at ꢀ143 ꢁC on a Nonius

1074
L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
Kappa CCD diffractometer. The crystal-to-detector dis-
tance was set to 30 mm and exposure time was 40 seconds
per degree for all data sets with a scan width of 1.6ꢁ. The
data collection was 99.5% complete to 25ꢁ in #. A total
of 47402 partial and complete reflections were collected
covering the indices, h = ꢀ12 to 12, k = ꢀ23 to 26,
l = ꢀ28 to 28. Reflections (8679) were symmetry indepen-
dent and the Rint = 0.0738 indicated that the data was of
average quality (average quality = 0.07). Indexing and unit
cell refinements indicated a primitive monoclinic lattice in
the space group P21/c (no. 14).
The data for 3, 4, 6, and 7 were all integrated and scaled
using HKL-SCALEPACK, and an absorption correction was
performed using SORTAV. Solution by direct methods (SIR
97; default 4) produced complete heavy atom phasing mod-
els consistent with the proposed structures. All non-hydro-
gen atoms were refined anisotropically by full-matrix least-
squares methods, while all hydrogen atoms were initially
located in a difference Fourier map and were refined using
a riding model. Crystal data for 3–4, 6–7 is presented in
Table 1. Selected bond distances and angles are assembled
in Table 2.
PrS)]+ (7) were synthesized via a similar route using CoCl2
or [Co(H2O)6]lCl2 in place of FeCl2. Thiolate-ligated 8,
which is a tertiary amine derivative of our functional
SOR model 1, was synthesized in order to determine if
the amine N–H protons or H-atoms were involved in the
proton-dependent mechanism of superoxide reduction.
Alkoxide-ligated [FeII(OMe2N4(tren))(H2O)]+ (2) was syn-
thesized as described elsewhere.28 The cyclamPrOH Æ HCl
ligand of alkoxide-ligated [FeII(cyclam-PrO)]+ (6) was
pre-assembled via a four step synthesis (outlined in Scheme
2), analogous to that previously reported for cyclam-PrSA-
c Æ 4HCl [24]. For the sulfur-containing ligand, the thiolate
must be protected (as a thioacetate group) throughout the
synthesis to prevent oxidation of the thiol. For the alcohol-
substituted ligand, no protecting group was needed for the
alcohol. The alkoxide complex [FeII(cyclam-PrO)]+ (6) was
prepared via in situ deprotonation of the alcohol and
amines using five equivalents of sodium methoxide, fol-
lowed by the addition of FeCl2 and NaBPh4 in MeOH.
Recrystallization from MeOH/Et2O afforded analytically
pure 6 as a white solid. The ESI-MS spectrum of this solid
redissolved in methanol shows the major peak at m/
z = 312.9, consistent with a 5-coordinate geometry for
the iron-containing cation. This structure was confirmed
by X-ray crystallography. Amine-ligated [FeII(NEt2N4-
(tren)(Cl)]+ (3) was synthesized via the FeCl2-templated
Schiff-base condensation between tris (2-aminoethyl)-
amine) (tren) and diethylaminoacetone in MeOH.
3. Results and discussion
3.1. Synthesis of thiolate-, alkoxide-, and amine-ligated 1–8
In order to correlate properties with metal ion structure
and ligated atoms, one needs to compare structurally anal-
ogous systems in which only the heteroatoms or metal ion
are varied and the ligand backbone remains the same. Mul-
tidentate ligands are generally required to maintain a some-
what rigid structure, since the biologically important first-
row transition-metal ions tend to be substitution labile.
Schematic drawings of the series of thiolate, alkoxide,
and amine complexes described in this study are shown
in Scheme 1.
3.2. Comparison of structures 1–8
Single crystals of 3, the Clꢀ salt of 4, 6, and 7 were
obtained via layer diffusion of Et2O into an MeCN (3, 4,
7) or MeOH (6) solution. The structures of 1, 2, the PF6ꢀ
salt of 4, 5, and 8 were described elsewhere [24,27–29]. Met-
rical parameters for 1–8 are compared in Table 2. ORTEP
diagrams of 3, 4, 6, and 7 are shown in Figs. 1 and 2. In
contrast to the other structures, X-ray data for 3 was col-
lected at ambient temperature because a phase change
occurred at lower temperatures, thus the significantly lar-
ger thermal ellipsoids (Fig. 1). Alkoxide and amine-ligated
Thiolate-ligated [FeII(SMe2N4(tren))]+ (1) [27], [FeII-
(cyclam-PrS)]+ (5) [24], and [FeII(SMe2N4(tren-Et4))]+ (8)
[29] were synthesized as previously described. Thiolate-
ligated [CoII(SMe2N4(tren))](Cl) (4) and [CoII (cyclam-
Scheme 1.
Scheme 2.

L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
1075
acetonitrile nitrogen (N(5)) points directly towards the
Co ion of 7, is H-bonded to two of the cyclam secondary
˚
amine protons (N(5)Hꢃ ꢃ ꢃN(3) = 2.245(9) A; N(5)Hꢃ ꢃ ꢃ
˚
N(4) = 2.223(9) A), and trans to the thiolate. These are
not the first examples of coordinatively unsaturated thio-
late complexes observed in our lab: our bis-thiolate-ligated
iron complex ½FeIIIðS2Me2N3ðPr; PrÞÞꢂþ remains five-coordi-
nate even though it contains a more Lewis acidic Fe3+
metal ion, and the first-row series [MII(tren)N4SMe2 +
]
Fig. 1. ORTEP of [CoII(SMe2N4(tren))](Cl) (4) and the cation of
[FeII(NEt2N4(tren))(Cl)](PF6) (3) showing that Clꢀ binds to the amine-,
but not the thiolate derivative. All hydrogen atoms, with the exception of
the amine hydrogens, have been removed for clarity.
(M = Mn, Fe (1), Co, Ni, Cu, Zn) [16] all maintain five-
coordinate structures even in coordinating solvents such
as MeCN. These findings are unusual in that the synthesis
of complexes containing an open-coordination site usually
requires the use of steric bulk, non-coordinating solvents,
and strict avoidance of potentially coordinating counteri-
ons [35–38].
Why alkoxide-ligated 6 does not bind a sixth ligand,
whereas 2 does, is unclear. The secondary versus primary
amines may be responsible (by lowering the metal ion
Lewis acidity), although the redox potentials (vide infra)
imply that this is not the case. Ligand constraints favoring
a distorted trigonal bipyramidal geometry may also be
involved. This was the explanation given for the consis-
tently observed five-coordinate structure of the first-row
Fig. 2. ORTEP of the cations of [FeII(cyclam-PrO)](BPh4) (6) and
[CoII(cyclam-PrS)](BPh4) Æ MeCN (7) showing the MeCN H-bonding to
the N–H protons, but not coordinated to the metal ion of 7. All hydrogen
atoms, with the exception of the amine hydrogens, have been removed for
clarity.
tren-ligated [MII(tren)N4SMe2 +
] (M = Mn, Fe (1), Co, Ni,
Cu, Zn) series recently reported by our group [16]. The fact
that alkoxide and amine-ligated tren complexes 2 and 3
afford six-coordinate structures would appear to rule this
out. On the other hand, it may be that the cyclam ligand
of 6 constrains the geometry to be five-coordinate when a
metal ion that is too large to fit inside the macrocyclic cav-
ity is involved. The high spin Fe2+ ion of 6 (vide infra)
would be larger than the preferred low-spin Fe2+, or oxi-
2 and 3 consistently crystallize as six-coordinate complexes
with a water (2) or Clꢀ (3) occupying the sixth site (Fig. 1,
Scheme 1). In contrast, all of the thiolate-ligated complexes
1, 4, 5, 7, and 8 remain five-coordinate, even in the presence
of coordinating solvents (MeCN, MeOH) or a Clꢀ counter-
ion. This suggests that the metal ion Lewis acidity is lower
for thiolate-ligated 1, 4, 5, 7, and 8, and greatest for amine
and alkoxide-ligated 2 and 3. The noticeable trans influence
of the thiolate [25,34], as well as reduced molecular charge,
could also be a factor partially responsible for the low
coordination number. Amine-ligated 3 (Fig. 1) would be
a dication if the Clꢀ anion were not incorporated into its
primary coordination sphere, explaining its higher affinity
for Clꢀ. This is in contrast to all of the other molecules
included in this study in which charge is conserved to afford
monocationic metal complexes. The fact that the Clꢀ coun-
terion is not incorporated into the primary coordination
sphere of 4 (Fig. 1) can be attributed to a decreased Lewis
acidity (relative to 3) caused by the replacement of a neu-
tral amine ligand with a monoanionic thiolate. Charge is
conserved however, by replacing the thiolate of 1 with an
alkoxide in 2, and yet the alkoxide complex reproducibly
binds a sixth ligand, whereas the thiolate complex 1 does
not. Again this implies that the thiolate decreases the metal
ion Lewis acidity helping to maintain an open coordination
site. Thiolate-ligated 7 (Fig. 2) remains five-coordinate
even though it cocrystallizes with a MeCN solvent. The
dized Fe3+
.
With the exception of 5, the five-coordinate structures
are all tetragonally distorted trigonal bipyramidal
(s = 0.47–0.67) approximately half-way between square
pyramidal (s = 0.0) and trigonal bipyramidal (s = 1.0).
Thiolate-appended cyclam-ligated 5 is much closer to
square pyramidal (s = 0.13). The X = S, O, N heteroatom
occupies the apical position of the tren-derived complexes
(1, 4, and 8), and the equatorial plane of the cyclam-
derived complexes 6 and 7. The CoII-cyclam complex 7
crystallizes as the less frequently encountered trans-V
(R,R,R,R) isomer, whereas 5 and 6 are both isolated as
the more common trans-I (R,S,R,S) isomer. The trans-I
configurational isomer is the most stable isomer for five-
coordinate cyclam complexes.
For the five-coordinate structures, the alkoxide Fe–O
˚
bond (1.906(2) A in 6) is significantly shorter than the thio-
˚
˚
late Fe–S bonds (range: 2.286(1) A (in 5) – 2.3281(9) A (in
1)), as one would expect based on differences between the
covalent radii of O versus S. The alkoxide Fe–O bond
˚
(2.001(1) A) for the six-coordinate structure 2 is longer
than that of the five-coordinate structure 6, as expected
based on the size difference of 5- versus 6-coordinate metal
ion radii [39]. The tertiary amine Fe–N(5) bond of 3

1076
L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
˚
(2.293(5) A) is longer than the tertiary amines of five-coor-
The iron complexes 1 [40], 2 [28], 3 (leff = 5.23 BM;
Fig. 4), 5, [24] 6 (leff = 4.79 BM; Figure S-2), and 8 [29]
are all high-spin S = 2 both in the solid state and in
solution. This is irrespective of whether the fifth ligand
X (Table 2) is an alkoxide, thiolate, or amine (Fig. 4).
The cobalt complexes 4 (leff = 4.05 BM), [16] and 7
(leff = 3.97 BM; supplemental) are both high-spin as well
with an S = 3/2 ground state over the temperature range
4–300 K.
˚
˚
dinate 5 (2.181(4) A), 6 (2.225(2) A), and 8 (mean:
˚
2.222(5) A), most likely because it is six-coordinate. Within
the tren-ligated systems, there is less variability of these ter-
tiary amine distances (range: 2.223(2)–2.268(3) A) most
likely due to ligand constraints. The secondary amine Fe–
N distances (2.138A in 5 to 2.206(2) in 6) are noticeably
shorter than the tertiary amine Fe–N bonds, but longer
than the imine Fe–N bonds, and the primary amine Fe–
N bonds (2.117 A and 2.131(3) A in 1). The Co–S, Co–O,
and Co–N distances in 4 and 7 are noticeably shorter than
those of their iron analogues. This is expected based on
periodic trends [16].
˚
˚
˚
˚
3.4. Comparison of redox potentials – correlation with
apical ligand X (S, O, N) and metal ion (Co2+, Fe2+
)
Upon oxidation, the iron and cobalt complexes 1 and 4
both have been shown to convert to low-spin (S = 1/2 or 0)
six-coordinate M3+ structures [27,41]. The iron complex,
[FeII(SMe2N4(tren))]+ (1), reacts with HO2 to afford six-
coordinate [FeIII(SMe2N4(tren))(MeCN)]2+ via a hydroper-
oxo intermediate [FeIII(SMe2N4(tren))(OOH)]+ [27,42].
Cyclic voltammograms of 3, 6, and 7, are shown in
3.3. Comparison of the electronic and magnetic
properties of 1–8
In contrast to the intensely colored, low-spin Fe(III)-
thiolates, the Fe(II)-thiolates 1, 5, and 8 described herein
are pale in color, and high-spin, due to the absence of
charge transfer bands in the visible region, and the
absence of pentet excited states. There is a poorer energy
match between the reduced metal ion d- and sulfur-orbi-
tals, so these charge transfer bands fall outside the visible
region. The alkoxide and amine complexes 2, 3, and 6 are
pale yellow and high-spin. The pale color of these solu-
tions is due to the weak intensity of Laporte-forbidden
d–d transitions for the six-coordinate complexes, or poor
energy match between the metal ion d- and alkoxide-
orbitals. The thiolate-ligated cobalt complexes 4 and 7
are slightly more intensely colored due to the presence
of lower energy charge transfer bands that tail into the
visible (Fig. 3 and supplemental Figure S-4). The charge
transfer bands are lower in energy for Co2+ because the
more electronegative metal ion’s orbitals lie lower than
those of Fe2+. The electronic absorption spectra of 1
and 4–8 are all solvent-independent implying that solvents
do not bind to the metal ion’s open coordination site,
even in coordinating solvents.
Fig. 4. Inverse molar magnetic susceptibility (1/vm) vs. temperature (T)
plot for [FeII(NEt2N4(tren))(Cl)](PF6) (3) fit to an S = 2 spin-state.
Fig. 5. Cyclic voltammogram of [FeII(NEt2N4(tren))(Cl)](PF6) (3) in
MeCN at 298 K (0.1 M (Bu4N)PF6, glassy carbon electrode, 150 mV/s
scan rate). Peak potentials versus SCE indicated.
Fig. 3. Electronic absorption spectrum of [CoII(cyclam-PrS)](BPh4) Æ
MeCN (7) in MeCN at ambient temperature.

L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
1077
Table 3
coordination number, implying that 3 would have an even
more cathodically shifted potential in the absence of the
sixth (Clꢀ) ligand. Alkoxide-ligated 2 would have a less
anodically shifted potential in the absence of its sixth
(H2O) ligand. Another property influencing redox poten-
tials is overall molecular charge. Molecular charge is con-
served for 1, 2, and 3 – all are monocations. This would
not be the case for 2 at higher pH where deprotonation
of the water would afford a neutral hydroxide-bound spe-
cies, or for 3 in the absence of the Clꢀ ligand where instead
a dicationic (2+) species would form.
In summary, comparison of redox potentials indicates
that the tertiary amine tren-ligands of 3 and 8 stabilize iron
in the +2 oxidation state, whereas the alkoxide tren-ligand
stabilizes iron in the +3 oxidation state. The primary amine
thiolate tren-ligand falls somewhere in between, favoring
the +3 oxidation state more than the amine, but less than
an alkoxide. Thiolate-ligated [FeII(SMe2N4(tren))]+ (1) is a
better reducing agent than thiolate-ligated [CoII(SMe2N4(t-
ren))]+ (4), and the ligand stabilizes iron in the +3 oxida-
tion state, and cobalt in the +2 oxidation state. Within
the cyclam-ligated series 5–7 similar trends are observed.
Replacement of the apical thiolate of [FeII(cyclam-PrS)]+
(5) with an apical alkoxide in [FeII(cyclam-PrO)]+ (6)
causes the redox potential to shift anodically by
ꢀ100 mV, and replacement of the Fe2+ ion of 5 with a
Co2+ ion (in 7) causes the redox potential to shift cathod-
ically by +765 mV.
Comparison of redox potentials for thiolate, alkoxide, and amine-ligated
M(II) complexes
E1/2 (vs SCE) (mV)
[FeII(SMe2N4(tren))]+ (1)
[FeII(SMe2N4(tren-Et4))]+ (8)
[FeII(OMe2)N4(tren)(H2O)]+ (2)
ꢀ200
+410 (Epa, irrev.)
ꢀ285 (pH = 6.8, Epa, irrev.)
ꢀ420 (pH = 10.6, Epa, irrev.)
[FeII(NEt2N4(tren)(Cl))]+ (3)
[CoII(SMe2N4(tren))]+ (4)
[FeII(cyclam-PrS)]+ (5)
[FeII(cyclam-PrO)]+ (6)
[FeII(cyclam)(MeCN)2]2+
[CoII(cyclam-PrS)]+ (7)
+303
+270 (Epa, irrev.)
+135
+35
+700
+900 (Epa, irrev.)
Fig. 5, and supplemental Figs. S-6 and S-7, respectively.
Redox potentials are assembled in Table 3. Comparison
of the redox potential (E1/2 and Epa; Table 3) of tren-
ligated 1 versus that of tren-ligated 4 [16] shows that
replacement of the Fe2+ with Co2+ within an identical
ligand environment causes the potential to shift cathodi-
cally by +470 mV. Similarly, comparison of E1/2 for
cyclam-ligated 5 versus E1/2 for cyclam-ligated 7 shows that
replacement of Fe2+ with Co2+, within an identical ligand
environment, causes the potential to shift cathodically by
+765 mV. These differences reflect differences in ionization
potentials for the two metals, and follow expected periodic
trends. As demonstrated by the lower energy charge trans-
fer bands in their electronic absorption spectra (vide
supra), the Co2+ d-orbitals lie lower in energy than the
Fe2+ orbitals making Co2+ more difficult to oxidize. Given
this, Co2+ would not be the metal ion of choice for metal-
loenzymes, such as SOR, that function as reducing agents.
Comparison of the redox potentials (E1/2; Table 3) of 8
versus those of 5 and 1, shows that a cathodic shift in the
redox potential by +275 mV upon replacement of primary
amines with secondary amines, and +610 mV upon
replacement of primary amines with tertiary amines. Ter-
tiary amines stabilize iron in the +2 oxidation state,
whereas primary amines stabilize iron in the +3 oxidation
state, and secondary amines fall somewhere in between.
One would expect the opposite trend on the basis of their
relative bond lengths and Lewis basicities.
Comparison of the redox potential (E1/2; Table 3) of 1
versus that of 3, shows that the replacement of a thiolate
sulfur (X in Table 2) with an amine nitrogen causes the
potential to shift cathodically, by +503 mV. In contrast,
replacement of the thiolate sulfur in 1 with an alkoxide
oxygen in 2 causes the potential to shift in the opposite
direction. The redox potential of the water-ligated alkoxide
complex 2 (Scheme 1) is pH-dependent, shifting anodically
by ꢁ60 mV/pH unit, most likely because the coordinated
water readily deprotonates.28 Although the ligands are con-
served with only minor changes to the apical X-atom, the
coordination numbers differ for 1 (5-coordinate) versus 2
(6-coordinate) and 3 (6-coordinate), making these compar-
isons somewhat less valid. One would expect the redox
potential to shift anodically in response to an increase in
3.5. Summary and conclusions
Comparison of the structurally-related derivatives
described herein, and elsewhere [23,26,43], shows that thio-
late ligands reduce metal ion Lewis acidity relative to alk-
oxides and amines, and have a strong trans influence
thereby helping to maintain an open coordination site.
With an open coordination site, the metal ion is accessible
so that substrates can react via an inner sphere mechanism.
Comparison of the redox potentials of the structurally
analogous compounds described herein indicates that the
higher-valent Fe3+ oxidation state is favored most by alk-
oxide ligands, and least by amine ligands. This fits with
what one would expect based on the hard–soft acid–base
classification of metal ions and ligands (Lewis acids and
bases) [44]. Thiolates favor Fe3+ more than amines but less
than alkoxides. This is somewhat surprising given that sul-
fur is usually considered to be a soft Lewis base, and Fe3+
a
hard Lewis acid. As shown previously however [45], exten-
sive p-interaction and delocalization stabilizes the Fe(III)–
S bonds more than one would anticipate. For a redox
active iron enzyme to function as a catalyst capable of
reducing substrates, both the Fe3+ and the Fe2+ states must
be accessible. Although a catalyst with a more anodic
potential would be a more potent reducing agent, an exces-
sively anodic potential would result in an inaccessible cat-
alytically active reduced Fe2+ state [46]. Since an open
coordination site, and a moderately low redox-potential

1078
L.M. Brines et al. / Inorganica Chimica Acta 361 (2008) 1070–1078
[17] D.W. Randall, S.D. George, B. Hedman, K.O. Hodgson,
K. Fujisawa, E.I. Solomon, J. Am. Chem. Soc. 122 (2000) 11620.
[18] T. Glaser, B. Hedman, K.O. Hodgson, E.I. Solomon, Acc. Chem.
Res. 33 (2000) 859.
[19] S.D. George, M. Metz, R.K. Szilagyi, H. Wang, S.P. Cramer, Y. Lu,
W.B. Tolman, B. Hedman, K.O. Hodgson, E.I. Solomon, J. Am.
Chem. Soc. 123 (2001) 5757.
are both necessary for efficient, inner sphere, substrate (e.g
superoxide) reduction to occur, the combination of a thio-
late (as opposed to an alkoxide or amine) coordinated to
iron (as opposed to cobalt), with primary (as opposed to
secondary or tertiary) amines in the synthetic models cre-
ates both of these optimum conditions needed for this
chemistry to occur. This implies that the cysteinate ligand
of the non-heme iron enzyme SOR plays an important role
in promoting function.
[20] M.D. Lowery, J.A. Guckert, M.S. Gebhard, E.I. Solomon, J. Am.
Chem. Soc. 115 (1993) 3012.
[21] S. Shoner, D. Barnhart, J.A. Kovacs, Inorg. Chem. 34 (1995) 4517.
[22] H.L. Jackson, S.C. Shoner, D. Rittenberg, J.A. Cowen, S. Lovell, D.
Barnhart, J.A. Kovacs, Inorg. Chem. 40 (2001) 1646.
[23] J.J. Ellison, A. Nienstedt, S.C. Shoner, D. Barnhart, J.A. Cowen, J.A.
Kovacs, J. Am. Chem. Soc. 120 (1998) 5691.
Acknowledgement
[24] T. Kitagawa, A. Dey, P. Lugo-Mas, J. Benedict, W. Kaminsky,
E. Solomon, J.A. Kovacs, J. Am. Chem. Soc. 128 (2006) 14448.
[25] J. Shearer, I.Y. Kung, S. Lovell, W. Kaminsky, J.A. Kovacs, J. Am.
Chem. Soc. 123 (2001) 463.
We thank Ed Solomon and his group for a very fruitful
and synergistic ongoing collaboration. This work was sup-
ported by the NIH (Grant Number: GM 45811-17).
[26] J. Shearer, H.L. Jackson, D. Schweitzer, D.K. Rittenberg, T.M.
Leavy, W. Kaminsky, R.C. Scarrow, J.A. Kovacs, J. Am. Chem. Soc.
124 (2002) 11417.
Appendix A. Supplementary material
[27] J. Shearer, J. Nehring, W. Kaminsky, J.A. Kovacs, Inorg. Chem. 40
(2001) 5483.
[28] L.M. Brines, W. Kaminsky, M.L. Kirk, J.A. Kovacs, 2007, manu-
script in preparation.
[29] R.M. Theisen, J. Shearer, W. Kaminsky, J.A. Kovacs, Inorg. Chem.
43 (2004) 7682.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.ica.
2007.07.038.
References
[30] D.D. Perrin, W.L.F. Armarego, D.R. Perrin, Purification of Labo-
ratory Chemicals, 2nd ed., Pergamon Press, Elmsford, NY, 1980.
[31] D.H. Live, S.I. Chan, Anal. Chem. 42 (1970) 791.
[32] D.A. Evans, J. Chem. Soc. (1959) 2005.
[33] A.L. Van Geet, Anal. Chem. 40 (1968) 2227.
[34] J. Shearer, I. Kung, S. Lovell, J.A. Kovacs, Inorg. Chem. 39 (2000)
4998.
[1] E.I. Solomon, Inorg. Chem. 45 (2006) 8012.
[2] J.A. Kovacs, Chem. Rev. 104 (2004) 825.
[3] P.V. Rao, R.H. Holm, Chem. Rev. 104 (2004) 527.
[4] H.B. Gray, B. Malstro¨m, R.J.P. Williams, J. Biol. Inorg. Chem. 5
(2000) 551.
[35] D.V. Yandulov, R.R. Schrock, J. Am. Chem. Soc. 124 (2002)
6252.
[36] J.M. Smith, A.R. Sadique, T.R. Cundari, K.R. Rodgers, G. Lukat-
Rodgers, R.J. Lachicotte, C.J. Flaschenriem, J. Vela, P.L. Holland,
J. Am. Chem. Soc. 128 (2006) 756.
[5] P.A. Lindahl, Biochemistry 41 (2002) 2097.
[6] P.V. Rao, S. Bhaduri, J. Jiang, D. Hong, R.H. Holm, J. Am. Chem.
Soc. 127 (2005) 1933.
[7] J.V. Rodrigues, I.A. Abreu, D. Cabelli, M. Teixeira, Biochemistry 45
(2006) 9266.
[37] B.J. O’Keefe, L.E. Breyfogle, M.A. Hillmyer, W.B. Tolman, J. Am.
Chem. Soc. 124 (2002) 4384.
[38] S.D. Brown, T.A. Betley, J.C. Peters, J. Am. Chem. Soc. 125 (2003)
322.
[39] R.D. Shannon, Acta. Crystallogr., Sect. A 32 (1976) 751.
[40] R.M. Theisen, Ph.D. thesis, University of Washington, 2005.
[41] R.D. Swartz, J.A. Kovacs, manuscript in preparation.
[42] J. Shearer, R.C. Scarrow, J.A. Kovacs, J. Am. Chem. Soc. 124 (2002)
11709.
[43] S.C. Shoner, A. Nienstedt, J.J. Ellison, I. Kung, D. Barnhart, J.A.
Kovacs, Inorg. Chem. 37 (1998) 5721.
[44] S.J. Lippard, J.M. Berg, Principles of Bioinorganic Chemistry,
University Science, Mill Valley, 1994.
[45] P. Kennepohl, F. Neese, D. Schweitzer, H.L. Jackson, J.A. Kovacs,
E.I. Solomon, Inorg. Chem. 44 (2005) 1826.
[46] J. Shearer, S.B. Fitch, W. Kaminsky, J. Benedict, R.C. Scarrow, J.A.
Kovacs, Proc. Natl. Acad. Sci., USA 100 (2003) 3671.
[8] D.M. Kurtz Jr., Acc. Chem. Res. 37 (2004) 902.
[9] A.T. Fiedler, P.A. Bryngelson, M.J. Maroney, T.C. Brunold, J. Am.
Chem. Soc. 127 (2005) 5449.
[10] J. Wuerges, J.-W. Lee, Y.-I. Yim, H.-S. Yim, S.-O. Kang, K.D.
Carugo, Proc. Natl. Acad. Sci. USA 101 (2004) 8569.
[11] J. Shearer, L.M. Long, Inorg. Chem. 45 (2006) 2358.
[12] I.G. Denisov, T.M. Makris, S.G. Sligar, I. Schlichting, Chem. Rev.
105 (2005) 2253.
[13] H. Ogata, S. Hirota, A. Nakahara, H. Komori, N. Shibata, T. Kato,
K. Kano, Y. Higuchi, Structure 13 (2005) 1635.
[14] P.T.R. Rajagopalan, X.C. Yu, D. Pei, J. Am. Chem. Soc. 119 (1997)
12418.
[15] S. Chang, V.V. Karambelkar, R.C. diTargiani, D.P. Goldberg, Inorg.
Chem. 40 (2001) 194.
[16] L.M. Brines, J. Shearer, J.K. Fender, D. Schweitzer, S.C. Shoner, D.
Barnhart, W. Kaminsky, S. Lovell, J.A. Kovacs, Inorg. Chem 46, in
press.