Spectroscopic and computational studies of a Ni+-CO model complex: Implications for the acetyl-CoA synthase catalytic mechanism

Article abstract of DOI:10.1021/ic020441e

The four-coordinate Ni⁺ complex [PhTt^Bu]Ni^lCO, where PhTt^Bu = phenyltris((tert-buthylthio)methyl) borate (a tridentate thioether donor ligand), serves as a possible model for key Ni-CO reaction intermediates in the acetyl-CoA synthase (ACS) catalytic cycle. Resonance Raman, electronic absorption, magnetic circular dichroism (MCD), variable-temperature variable-field MCD, and electron paramagnetic resonance spectroscopies were utilized in conjunction with density functional theory and semiemperical INDO/S-Cl calculations to investigate the ground and excited states of [PhTt^Bu]Ni^lCO. These studies reveal extensive Ni⁺ → CO π-back-bonding interactions, as evidenced by a low C-O stretching frequency (1995 cm^-1), a calculated C-O stretching force constant of 15.5 mdyn/A (as compared to k_CO(free CO) = 18.7 mdyn/A), and strong Ni⁺ → CO charge-transfer absorption intensities. Calculations reveal that this high degree of π-back-bonding is due to the fact that the Ni⁺ 3d orbitals are in close energetic proximity to the CO π* acceptor orbitals. In the ACS paramagnetic catalytic cycle , the high degree of π-back-bonding in the putative Ni⁺-CO intermediate (the NiFeC species) is not expected to preclude methyl transfer from CH₃-CoFeSP.

Full text of DOI:10.1021/ic020441e

Inorg. Chem. 2003, 42, 859−867
Spectroscopic and Computational Studies of a Ni+−CO Model
Complex: Implications for the Acetyl-CoA Synthase Catalytic
Mechanism
Jennifer L. Craft,† Beaven S. Mandimutsira,‡ Koyu Fujita,‡ Charles G. Riordan,‡ and
,†
Thomas C. Brunold*
Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706, and Department of
Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
Received July 8, 2002
tBu
The four-coordinate Ni+ complex [PhTt ]NiICO, where PhTttBu ) phenyltris((tert-buthylthio)methyl)borate (a tridentate
thioether donor ligand), serves as a possible model for key Ni−CO reaction intermediates in the acetyl-CoA synthase
(ACS) catalytic cycle. Resonance Raman, electronic absorption, magnetic circular dichroism (MCD), variable-
temperature variable-field MCD, and electron paramagnetic resonance spectroscopies were utilized in conjunction
with density functional theory and semiemperical INDO/S-CI calculations to investigate the ground and excited
states of [PhTttBu]NiICO. These studies reveal extensive Ni+ f CO π-back-bonding interactions, as evidenced by
a low C−O stretching frequency (1995 cm-1), a calculated C−O stretching force constant of 15.5 mdyn/Å (as
compared to kCO(free CO) ) 18.7 mdyn/Å), and strong Ni+ f CO charge-transfer absorption intensities. Calculations
reveal that this high degree of π-back-bonding is due to the fact that the Ni+ 3d orbitals are in close energetic
proximity to the CO π* acceptor orbitals. In the ACS “paramagnetic catalytic cycle”, the high degree of π-back-
bonding in the putative Ni+−CO intermediate (the NiFeC species) is not expected to preclude methyl transfer from
CH −CoFeSP.
3
Introduction
A novel Fe-[NiFe3S4] cluster in the CODH subunit of the
protein, the so-called C cluster, is hypothesized to bind CO
substrate at the Ni center to catalyze nucleophilic attack by
OH-/H2O.2-4 A putative Ni-[Fe4S4] cluster within the ACS
subunit, termed the A cluster, is also proposed to bind CO
via the Ni atom prior to CO migratory insertion into a Ni-
CH3 bond.2,5
The oxidation state of the A cluster, particularly that of
the Ni atom, during catalysis is a heavily debated issue.
While the “paramagnetic” mechanistic proposal invokes an
EPR-active Ni+-CO intermediate (termed the NiFeC spe-
cies) prior to the methyl transfer step, the competing
“diamagnetic” proposal solely employs Ni2+-CO intermedi-
ates. Recent steady-state kinetics experiments by Ragsdale
and co-workers5 strongly advocate a catalytic role for the
paramagnetic NiFeC species, demonstrating the kinetic
competence of this Ni+-CO intermediate and further sug-
Urease, hydrogenase, methyl-CoM reductase, Ni-depend-
ent superoxide dismutase, and CO dehydrogenase/acetyl-CoA
synthase (CODH/ACS) comprise the five known classes of
Ni-containing enzymes. Nature appears to utilize Ni as either
a Lewis acid, as exemplified by urease, or redox center, as
proposed for the four other enzymes.1 Ni may also play a
key role in the binding and activation of enzymatic substrates.
In the case of CODH/ACS, the bifunctional enzyme that
catalyzes both the reversible oxidation of CO to CO2 (eq 1)
and the synthesis of acetyl-CoA by condensing the methyl
group of a methylated corrin-FeS protein (CoFeSP) with
CO and CoA (eq 2), formation of Ni-CO reaction inter-
mediates appears likely.2
CO + H2O T CO2 + 2H+ + 2e-
(1)
CH3-CoFeSP + CO + CoA T acetyl-CoA + CoFeSP (2)
(1) Ragsdale, S. W.; Riordan, C. G. J. Bioinorg. Chem. 1996, 1, 489.
(2) Ragsdale, S. W.; Kumar, M. Chem. ReV. 1996, 96, 2515.
(3) Drennan, C. L.; Heo, J. Y.; Sintchak, M. D.; Schreiter, E.; Ludden, P.
W. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 11973.
(4) Dobbek, H.; Svetlitchnyi, V.; Gremer, L.; Huber, R.; Meyer, O. Science
2001, 293, 1281.
* To whom correspondence should be addressed. E-mail: brunold@
chem.wisc.edu.
† University of Wisconsin.
‡ University of Delaware.
(5) Seravalli, J.; Kumar, M.; Ragsdale, S. W. Biochemistry 2002, 41, 1807.
10.1021/ic020441e CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/14/2003
Inorganic Chemistry, Vol. 42, No. 3, 2003 859

Craft et al.
(a) [PhTttBu]NiIP(CH3)3. [PhTttBu]NiIICl (309 mg, 0.63 mmol)
was dissolved in 100 mL of Et2O and the red solution cooled to
-78 °C in a dry ice-acetone bath. P(CH3)3 (0.63 mL of a 1.0 M
THF solution, 0.63 mmol) was injected into this solution via syringe
to give a purplish solution. CH3Li (0.43 mL of a 1.6 M Et2O
solution, 0.69 mmol) was immediately added into the purple
solution resulting in a rapid color change to yellow. A white solid
precipitated as stirring continued for 6 h as the solution warmed to
25 °C. Solvent removal under reduced pressure gave a yellowish-
white solid that was extracted with pentanes. Elution through a silica
gel plug followed by solvent removal yielded, [PhTttBu]NiIP(CH3)3
as a pale yellow solid, 203 mg (61%). X-ray-quality crystals were
obtained by cooling concentrated pentanes solutions at -40 °C.
1H NMR (C6D6): δ 86 (br, BCH2), 22 (br, P(CH3)3), 18 (br, C6H5),
11 (br, C6H5), 10 (br, C6H5), -6 (br, (CH3)3). 31P NMR (C6D6): δ
264. Anal. Calcd for C24H47BNiPS3: C, 54.2; H, 8.90. Found: C,
53.9; H, 8.76.
Figure 1. [PhTttBu]NiICO model complex and local coordinate system as
defined by the principal axes of the INDO/S-CI-calculated g-matrix.
gesting that methylation of ACS occurs by attack of the Ni+
(b) [PhTttBu]NiICO. This compound was prepared similarly to
[PhTttBu]NiIP(CH3)3 with the modification that P(CH3)3 was replaced
by CO as follows. CO was bubbled through a red solution of
[PhTttBu]NiIICl (200 mg, 0.41 mmol) at -78 °C for about 3 min
followed by the addition of CH3Li (0.28 mL of a 1.6 M Et2O
solution, 0.45 mmol) via syringe. The solution turned from red to
orange, and a white solid precipitated as stirring was continued for
6 h at 25 °C under a CO atmosphere. (The reaction vessel was
vented to relieve pressure buildup once the mixture had warmed
to 25 °C.) Upon workup [PhTttBu]NiICO was produced as a yellow
solid, 100 mg (51%). X-ray-quality crystals were obtained by
cooling concentrated pentanes solutions at -40 °C. 1H NMR
(C6D6): δ 116 (br, BCH2), 14 (br, C6H5), 10 (br, C6H5), 9 (br,
C6H5), -1 (br, (CH3)3). 13C NMR (C6D6): δ 250 (br, Ni-CO).
site on the methyl group of CH3-CoFeSP. The four-
coordinate Ni+ complex [PhTttBu]NiICO, where PhTttBu
)
phenyltris((tert-buthylthio)methyl)borate (a tridentate thio-
ether donor ligand),6 provides a possible model for the CO-
bound A cluster (Figure 1).7 EPR8 and Mo¨ssbauer9,10 studies
indicate that the NiFeC species, with g values of 2.08, 2.07,
and 2.03, contains a diamagnetic [Fe4S4]2+ cluster exchange-
coupled through a low-lying S ) 1 excited spin state to a
paramagnetic Ni+ center. As ACS-CO and [PhTttBu]NiICO
exhibit identical C-O stretching frequencies (νCO ) 1995
cm-1),11 similar Ni-CO bonding in the two species appears
likely.
Detailed spectroscopic and computational studies on
[PhTttBu]NiICO provide an opportunity to probe the nature
of Ni+-CO bonding in a well-defined small model system,
potentially lending insights into the ACS catalytic mecha-
nism. Here, resonance Raman (rR), electronic absorption,
magnetic circular dichroism (MCD), variable-temperature
variable-field MCD (VTVH MCD), and electron para-
magnetic resonance (EPR) spectroscopies are employed, in
conjunction with density functional theory (DFT) and
semiemperical INDO/S-CI calculations, to investigate the
ground and excited states of [PhTttBu]NiCO. While these
studies reveal extensive Ni+ f CO π-back-bonding interac-
tions, the Ni+ site in the putative ACS-CO intermediate is
expected to remain sufficiently nucleophilic to accommodate
methyl transfer from CH3-CoFeSP in the synthesis of acetyl-
CoA.
FT-IR (KBr): νCO, 1999 cm-1; ν CO, 1951 cm-1. Anal. Calcd for
13
C22H38BNiOS3: C, 54.6; H, 7.91. Found: C, 54.7; H, 7.73.
Solid samples utilized for resonance Raman experiments were
prepared by mixing finely ground [PhTttBu]NiICO with K2SO4 in a
∼1:4 ratio. Solid samples used to obtain absorption and MCD data
were prepared by adding a small amount of poly(dimethylsiloxane)
to the ground Ni+ complex. Solution samples used to obtain
absorption, MCD, and EPR data were prepared by dissolving
[PhTttBu]NiICO in dry toluene (∼2 mM).
Electronic Absorption and MCD Spectroscopy. Variable-
temperature electronic absorption and MCD spectra were collected
on a CD spectropolarimeter (Jasco J-715) with a sample compart-
ment modified to accommodate a superconducting magnetocryostat
(Oxford Instruments SM4-8T).
EPR Spectroscopy. EPR spectra were collected with use of a
Bruker EMX spectrometer equipped with an ER4102ST cavity.
Spectra were recorded at 4.2 K using a LHe cryostat. The instrument
was previously calibrated with DPPH. Spectra were collected with
use of the following spectrometer settings: attenuation ) 25 dB;
microwave power ) 0.64 mW; frequency ) 9.31 GHz; sweep width
Experimental and Computational Procedures
Syntheses and Sample Preparation. [PhTttBu]NiIP(CH3)3 and
[PhTttBu]NiICO were synthesized by following a published proce-
dure.6
) 5000 G; modulation amplitude ) 5.02 G; gain ) 8.93 × 10-3
;
conversion time ) 81.92 ms; time constant ) 1.28 ms; resolution
) 1024 points. Samples were prepared by adding 20 mg of sample
to an EPR tube and dissolving the sample in toluene.
(6) Schebler, P. J.; Mandimutsira, B. S.; Riordan, C. G.; Liable-Sands,
L. M.; Incavito, C. D.; Rheingold, A. L. J. Am. Chem. Soc. 2001,
123, 331.
(7) Ragsdale, S. W.; Ljungdahl, L. G.; DerVartanian, D. V. Biochem.
Biophys. Res. Commun. 1982, 108, 658.
(8) Ragsdale, S. W.; Wood, H. G.; Antholine, W. E. Proc. Natl. Acad.
Sci. U.S.A. 1985, 82, 6811.
(9) Xia, J. Q.; Hu, Z. G.; Popescu, C. V.; Lindahl, P. A.; Munck, E. J.
Am. Chem. Soc. 1997, 119, 8301.
Resonance Raman Spectroscopy. A rR excitation profile for
[PhTttBu]NiICO was obtained upon excitation with Ar+ ion (Coher-
ent I-305) and dye (Coherent 599-01) lasers with incident power
in the 10-20 mW range. Scattering was collected at ∼135° from
the surface of the sample contained in a capillary tube immersed
in a liquid N2-filled EPR dewar (77 K). The scattered light was
dispersed by a triple monochromator (Acton Research, equipped
with 300, 1200, and 2400 grooves/mm gratings) and analyzed with
(10) Russell, W. K.; Stalhandske, C. M. V.; Xia, J. Q.; Scott, R. A.; Lindahl,
P. A. J. Am. Chem. Soc. 1998, 120, 7502.
(11) Kumar, M.; Ragsdale, S. W. J. Am. Chem. Soc. 1992, 114, 8713.
860 Inorganic Chemistry, Vol. 42, No. 3, 2003

A Ni+-CO Model Complex
a deep depletion, back-thinned CCD camera (Princeton Instruments
Spec X: 100BR). Raman intensities were quantified relative to the
984 cm-1 scattering peak of K2SO4.
Normal Coordinate Analysis. A normal coordinate analysis
(NCA) of the vibrational data was performed on the NiCO unit in
the crystal structure of [PhTttBu]NiICO: Ni-C ) 1.754 Å; C-O
) 1.127 Å; Ni-C-O angle γ ) 171.0°. The analysis was based
on the Wilson FG matrix method using a Urey-Bradley force field
as implemented in a modified version of the Schachtschneider
program.12,13
DFT Calculations. DFT computations were performed using
the Amsterdam density functional (ADF) 2000.02 software package
and the ORCA 2001 program developed by Dr. Frank Neese (MPI
Mu¨lheim, Mu¨lheim, Germany).14 All calculations were of the spin-
unrestricted type, employing the local density approximation of
Vosko, Wilk, and Nusair together with the gradient corrections of
Becke15 and Perdew.16 For ADF calculations a triple-ú Slater-type
orbital basis set (ADF basis set IV) with a single polarization
function was used. Core orbitals were frozen through 1s (S, C, B,
H) and 2p (Ni). The accuracy parameter for the numerical
integration grid was set to 4.0. For all ORCA calculations, a
Gaussian-polarized double-ú valence orbital basis set was used. The
size of the integration grid was set to 3 (Lebedew 194 points). While
ADF and ORCA DFT calculations yielded essentially identical
electronic descriptions, the latter offers the advantage of plotting
orbitals using the gOpenMol software developed by Laaksonen.17,18
Semiempirical INDO/S-CI Calculations. Semiempirical cal-
culations employing the INDO/S model developed by Zerner and
co-workers19-22 were also performed using the ORCA program.
Restricted open-shell Hartree-Fock (ROHF) SCF calculations were
converged on the spin doublet ground state, which served as the
reference state for configuration interaction (CI) calculations. Stable
results were obtained by including all possible single excitations
within 46 MOs (which include 31 doubly occupied MOs (DOMOs),
1 singly occupied MO (SOMO), and 14 virtual MOs), together with
double excitations from the highest 16 DOMOs into the SOMO
and the lowest 10 virtual MOs. Only single excitations into singly
occupied and unoccupied molecular orbitals (MOs) were used to
calculate electronic transition energies and oscillator strengths.
Separate calculations to investigate possible spin-forbidden transi-
tions, i.e., doublet-to-quartet excitations, included single and double
excitations within an identical MO space. For additional details on
INDO/S-CI calculations using the ORCA program, see ref 15 and
literature cited therein.
Figure 2. Variable-field MCD spectra (1, 2, 3.5, 5, and 7 T) (top) and
electronic absorption spectrum (bottom) of solid-state [PhTttBu]NiICO
obtained at 3 K. Inset: VTVH MCD data obtained at 25 900 cm-1 (solid
lines). Brillouin curves for S ) 1/2, 1, 3/2, 2, and 5/2 obtained using g
values of 2.0 and omitting zero-field splittings are also shown for comparison
(dotted lines).
experimentally using a combination of absorption, MCD, rR,
and EPR spectroscopies.
A.1. Electronic Absorption, MCD, and VTVH MCD.
Figure 2 displays low-temperature, solid-state mull absorption
and variable-field MCD spectra. MCD signals arising from
paramagnetic species are strongly temperature dependent as
a consequence of the Boltzmann population distribution
among Zeeman-split sublevels of the electronic ground state
(MCD C-term behavior).14,23-25 Accordingly, the increase in
MCD intensity with decreasing temperature for [PhTttBu]-
NiICO (Figure S1, Supporting Information) indicates that the
observed transitions are associated with a paramagnetic Ni
species. The same transitions, albeit with slightly blue-shifted
energies, are also observed in 190 K solution MCD spectra
(Figure S2, Supporting Information), implying that the
[PhTttBu]NiICO structure is preserved in solution.26
Results and Analysis
Ground-state spin information can be readily inferred from
a VTVH MCD experiment, in which the MCD signal
intensity at a specific wavelength is monitored as a function
of temperature and field. VTVH MCD data for [PhTttBu]-
NiICO are presented in the inset of Figure 2 (solid lines).
Theoretical magnetization curves for S ) 1/2, 1, 3/2, 2, and
5/2 species obtained using g values of 2.0 and omitting zero-
field splittings (ZFS), so-called Brillouin curves, are shown
A. Spectroscopic Studies. Ground-state and excited-state
electronic properties of [PhTttBu]NiICO were investigated
(12) Schachtschneider, J. H. Technical Report No. 57-65; Shell Develop-
ment Co.: Emeryville, CA, 1966.
(13) Fuhrer, H.; Kartha, V. B.; Kidd, K. G.; Krueger, P. J.; Mantsch, H.
H. Computer Programs for Infrared Spectroscopy, Bulletin No. 15,
National Research Council of Canada: Ottawa, Canada, 1976.
(14) Neese, F.; Solomon, E. I. Inorg. Chem. 1999, 38, 1847.
(15) Becke, A. D. J. Chem. Phys. 1986, 84, 4524.
(16) Perdew, J. P. Phys. ReV. B 1986, 33, 8822.
(17) Laaksonen, L. J. Mol. Graphics 1992, 10, 33.
(18) Bergman, D. L.; Laaksonen, L.; Laaksonen, A. J. Mol. Graphics
Modell. 1997, 15, 301.
(19) Ridley, J.; Zerner, M. C. Theor. Chim. Acta 1973, 32, 111.
(20) Bacon, A. D.; Zerner, M. C. Theor. Chim. Acta 1979, 53, 21.
(21) Zerner, M. C.; Loew, G. H.; Kirchner, R. F.; Mueller-Westerhoff, U.
T. J. Am. Chem. Soc. 1980, 102, 9.
(23) Johnson, M. K.; Robinson, A. E.; Thomson, A. J. In Iron-Sulfur
Proteins; Spiro, T. G., Ed.; Wiley-Interscience: New York, 1982.
(24) Piepho, S. B.; Schatz, P. N. Group Theory in Spectroscopy with
Applications to Magnetic Circular Dichroism; Wiley: New York,
1983.
(25) Oganesyan, V. S.; George, S. J.; Cheeseman, M. R.; Thomson, A. J.
J. Chem. Phys. 1999, 110, 762.
(22) Anderson, W. P.; Edwards, W. D.; Zerner, M. C. Inorg. Chem. 1986,
25, 2728.
(26) Attempts to obtain 3 K MCD data on frozen glasses were unsuccessful
due to low light transmission.
Inorganic Chemistry, Vol. 42, No. 3, 2003 861

Craft et al.
Table 1. Fit Parameters from Gaussian Deconvolution in Figure 3
peak energy (cm-1
)
fwhm (cm-1
)
ꢀ (M-1 cm-1
)
∆ꢀ (M-1 cm-1
)
1
2
3
4
5
6
9 380
19 000
21 200
24 000
26 400
29 400
1665
3000
3000
3000
3000
3000
5
175
1210
815
1405
2100
-19
-9
18
-33
-156
113
for comparison (dotted lines). The model complex exhibits
magnetization behavior characteristic of an S ) 1/2 para-
magnetic species, as expected for a d9 system. The same
magnetization behavior is observed for all bands, indicating
that contributions from possible Ni2+ impurities are negligible
and all observed transitions are associated with the S ) 1/2
Ni+ center.
Absorption and MCD spectra were subjected to Gaussian
deconvolution to determine the number and intensities of
electronic transitions contributing to the observed features.
The minimum number of Gaussian bands were first fit to
the MCD spectrum, keeping the widths of all bands beyond
10 000 cm-1 identical (as the corresponding transitions are
similar in nature (vide infra)). Simulated absorption and
MCD spectra were then iteratively refined until the energies
for all bands in the two spectra matched. Band deconvolution
reveals that at least six electronic transitions occur in the
region between 9000 and 33 000 cm-1. Simulated spectra,
obtained as a sum of the individual Gaussian bands generated
using the fit parameters in Table 1, are plotted over the
experimental absorption and MCD spectra in Figure 3.
On the basis of its low intensity in the absorption spectrum
and its large |∆ꢀ|/ꢀ ratio, transition 1, corresponding to the
negative MCD C-term at ∼9400 cm-1 (|∆ꢀ| ≈ 20 M-1cm-1),
is assigned to a Ni-centered d f d transition. The high energy
of this d f d transition is atypical for a four-coordinate
distorted tetrahedral Ni+ species, reflecting unusual metal-
ligand bonding interactions due to the presence of the CO
ligand in [PhTttBu]NiICO (vide infra). Transitions 2-6 likely
arise from metal-to-ligand charge-transfer (CT) excitations,
the origins of which will be discussed below.
To aid in the assignment of [PhTttBu]NiICO electronic
transitions, MCD data were also taken on [PhTttBu]NiIP-
(CH3)3, an analogous Ni+ complex with P(CH3)3 replacing
the CO ligand (Figure S3, Supporting Information). Unlike
[PhTttBu]NiICO, [PhTttBu]NiIP(CH3)3 exhibits no d f d
transition in the region between 9000 and 33 000 cm-1,
reflecting a smaller overall d-orbital splitting in the latter
complex. This is consistent with the different π-acidities of
the P(CH3)3 and CO ligands. The P(CH3)3 derivative also
shows no intensity in the spectral region near 22 400 cm-1,
leading to the tentative assignment of transitions 3 and 4 for
[PhTttBu]NiICO to Ni+ f CO CT excitations. Conversely,
features resembling the derivative-shaped MCD C-term
centered at ∼28 200 cm-1 in the [PhTttBu]NiICO spectrum
are observed at slightly higher energies for [PhTttBu]NiIP-
(CH3)3 (centered at ∼30 300 cm-1). Calculations intended
to elucidate the origins of transitions 5 and 6 in [PhTttBu]-
NiICO, as well as the corresponding excitations in [PhTttBu]-
NiIP(CH3)3, are presented in section B.2.
Figure 3. Gaussian deconvolution of low-temperature absorption and MCD
spectra of [PhTttBu]NiICO.
Figure 4. 77 K rR spectrum of solid-state [PhTttBu]NiICO obtained using
514-nm laser excitation (10 mW laser power at the sample, 5 min averaging
time).
A.2. Resonance Raman and Normal Coordinate Analy-
sis. The rR spectrum of [PhTttBu]NiICO exhibits a prominent
band at 1995 cm-1 upon 514-nm (19 455-cm-1) excitation
(Figure 4), unusually low relative to the C-O stretching
frequencies of other four- and five-coordinate Ni+/Ni2+-
CO complexes (e.g., 2024 cm-1 for [NiI(DAPA)(SePh)2(CO)]-
{DAPA ) 2,6-bis[1-(phenylimino)ethyl]pyradine},27 2028
cm-1 for [NiII(CO)(SPh)(SePh)2]-,28 and 2029 cm-1 for
[NiII(PS3*)(CO)]- {PS3* ) tris(3-phenyl-2-thiophenyl)-
phosphine}29. On substitution of 12C with 13C at the carbonyl
position, the 1995-cm-1 band undergoes an isotope shift of
35 cm-1, confirming that this vibrational feature indeed arises
from the C-O stretching motion. Raman bands at lower
energies do not undergo corresponding shifts and are,
therefore, attributed to other vibrational modes within the
molecule, i.e., stretching/bending motions associated with
the tridentate thioether donor ligand.
(27) Marganian, C. A.; Vazir, H.; Baidya, N.; Olmstead, M. M.; Mascharak,
P. K. J. Am. Chem. Soc. 1995, 117, 1584.
(28) Liaw, W.-F.; Horng, Y.-C.; Ou, D.-S.; Ching, C.-Y.; Lee, G.-H.; Peng,
S.-M. J. Am. Chem. Soc. 1997, 119, 9299.
(29) Nguyen, D. H.; Hsu, H.-F.; Millar, M.; Koch, S. A.; Achim, C.;
Bominaar, E. L.; Munck, E. J. Am. Chem. Soc. 1996, 118, 8963.
862 Inorganic Chemistry, Vol. 42, No. 3, 2003

A Ni+-CO Model Complex
Figure 5. rR excitation profile (s) for theC-O stretching peak at 1999
cm-1 and corresponding low-temperature absorption spectrum.
Figure 5 shows the rR excitation profile obtained for the
C-O stretching peak (νCO ) 1995 cm-1) of [PhTttBu]-
NiICO. Strong enhancement of the C-O vibration upon
excitation in resonance with the electronic absorption feature
at ∼21 500 cm-1 confirms the assignment of transitions 3
and 4 to Ni+ f CO CT excitations.
Figure 6. EPR spectra of [PhTttBu]NiICO (bottom) and carbonyl 13C-
labeled complex (top).
the z direction requires admixture of excited-state character
into the ground state via Lz. Conversely, the CO-bound A
cluster displays g values much closer to 2.00, requiring that
spin-orbit mixing of excited-state character into the ground
state be trivial. Larger separation of ground and excited states,
possibly a result of higher Ni coordination in the protein
species, may account for these electronic differences (vide
infra).
To determine the C-O force constant kCO for [PhTttBu]-
NiICO, which reflects the degree of Ni+ f CO π-back-
bonding, we performed a NCA on the NiCO unit using the
experimental C-O stretching frequencies. Force constants
for the Ni-C and C-O stretching modes and the Ni-C-O
bending mode, as well as the Ni‚‚‚O nonbonded interaction
constant required by the Urey-Bradley force field, were fit
to best reproduce the experimental frequencies, νCO(Ni-
12CO) ) 1995 cm-1 and νCO(Ni-13CO) ) 1960 cm-1. The
C-O force constant kCO was optimized to a value of 15.5
mdyn/Å, generating frequencies of νCO(Ni-12CO) ) 2000
cm-1 and νCO(Ni-13CO) ) 1955 cm-1. This relatively small
force constant, as compared to kCO(free CO) ) 18.7 mdyn/
Å, reveals that extensive Ni+ f CO π-back-bonding in
[PhTttBu]NiCO greatly weakens the C-O bond.
As Ni-C stretching and Ni-C-O bending mode frequen-
cies could not be identified experimentally, a series of
calculations was performed to determine the sensitivity of
kCO to changes in other force constants. Variation of kNiC
and the Ni-C-O bending force constant within reasonable
limits did not significantly affect the value of kCO nor the
calculated C-O stretching frequencies, demonstrating that
the C-O stretching mode is essentially uncoupled. This
relative independence of kCO is consistent with the potential
energy distribution, in which ∼96% of the highest-energy
vibrational mode is solely attributable to the C-O stretching
motion.
A.3. EPR. The [PhTttBu]NiICO complex displays a rhom-
bic EPR signal with significant axial character and g values
of 2.64, 2.02, and 1.95 (Figure 6). Comparison of the
[PhTttBu]NiICO EPR signal to that exhibited by the CO-bound
S ) 1/2 state of the A cluster in ACS (the so-called NiFeC
signal with g values of 2.08, 2.07 and 2.03)8 indicates that
considerable electronic differences exist between the model
and the putative reaction intermediate. In the case of the
model, gz differs significantly from 2.00 due to substantial
orbital angular momentum along the z axis of the g-matrix
in the ground state. As [PhTttBu]NiICO possesses an orbitally
nondegenerate ground state, orbital angular momentum in
13C hyperfine broadening is not observed for carbonyl 13
C
(I ) 1/2) enriched [PhTttBu]NiICO (Figure 6), suggesting that
the spin of the unpaired electron does not efficiently couple
with that of the 13C nucleus. The perpendicular arrangement
2
2
of the carbonyl ligand and the singly occupied Ni 3dx -y
orbital (vide infra) likely accounts for this inefficient
hyperfine coupling mechanism.
B. Computational Studies. Spectroscopic studies were
complemented by DFT and semiempirical INDO/S-CI
calculations to assist in spectral assignments and to probe
the nature of the Ni+-CO bond. Atomic coordinates derived
from the [PhTttBu]NiICO crystal structure, with the H atom
positions obtained using standard bond angles and distances,
the phenyl ring substituted by a methyl group at B, and the
tert-butyl groups substituted by methyl groups at the S
positions, were used to model relevant structural and
electronic features of the solid-state complex (Figure 1 and
Table S1, Supporting Information).
B.1. Molecular Orbitals. Virtually identical MO energies
and compositions were calculated using both the ADF and
ORCA programs. Table 2 lists energies and compositions
of the relevant highest-energy occupied MOs and lowest-
energy unoccupied MOs obtained from the ORCA DFT
calculation. MOs are labeled according to the atomic orbitals
that produce the dominant contributions. The spin-up (R)
MOs are stabilized relative to their spin-down (â) counter-
parts due to spin polarization in the spin-unrestricted formal-
ism used. Refinement of this provisional DFT-generated
bonding description using experimental data and semi-
empirical calculations is described below.
B.2. Calculated Absorption Spectra and Spectral Band
Assignments. Table 3 gives the energies, intensities, polar-
Inorganic Chemistry, Vol. 42, No. 3, 2003 863

Craft et al.
Table 2. Energies and Compositions (%) of Relevant MOs Obtained
from Ni 3dyz f CO π*x,y excitations. Computations further
indicate that the similar MCD pseudo-A-term in [PhTttBu]-
NiIP(CH3)3 spectra originates primarily from two perpen-
dicularly polarized Ni 3dyz f P 3px,y transitions, which can
be viewed as the Ni+-P(CH3)3 counterpart to Ni 3dyz f CO
π*x,y transitions. As the majority of experimental and com-
putational data support the assignment of transitions 5 and
6 to Ni+ f CO CT excitations, the fact that Raman enhance-
ment of the C-O stretch is not detected in this region may
be ascribed to partial photodecomposition of [PhTttBu]-
NiICO at higher excitation energies.
from ORCA DFT Calculations on the [PhTttBu]NiICO Model
Ni 3d orbitals
S3
MO
E (eV) occ xy xz yz z2 x2 - y2 πx*(CO) πy*(CO) tot.
Spin-Up (R) MOs
πy*(CO) -1.1490
πx*(CO) -1.3297
x2 - y2 -5.0120
0
0
1
1
1
1
1
0
0
1
1
9
6
10
1
12
4
0
3
0
1
0
4
9
42
3
0
5
41
19
0
0
1
5
7
39
9
11
12
4
26
1
z2
-5.5129
-6.2861
-6.3857
-6.5414
1
10 48
xz
yz
xy
49 12
0
1
0
1
12 13 16
20
14
10
1
0
2
0
0
1
Spin-Down (â) MOs
πy*(CO) -1.0947
πx*(CO) -1.1659
x2 - y2 -4.1442
0
0
0
1
1
1
1
0
0
3
7
2
3
18
7
11
0
2
1
13
5
0
9
15 42
6
17
4
0
0
18
0
6
50
0
2
19
41
4
2
0
40
19
2
7
0
6
7
23
29
10
9
Weakly observed both in absorption and MCD spectra,
transition 2 appears to arise from a spin-forbidden excitation
into the lowest-energy quartet CT excited state. Semi-
empirical INDO/S-CI computations support this assignment,
z2
-4.6493
-5.2159
-5.9236
-6.0669
z2/yz
xz
xy
53
0
0
0
8
15
6
0
1
2
2
estimating quartet Ni 3dz f CO π*x,y excited states 2000-
20
4000 cm-1 lower in energy than the corresponding doublet
excited states. The formally spin-forbidden transition at ∼19
000 cm-1 can acquire intensity through spin-orbit mixing
of the corresponding excited state with the nearby doublet
Table 3. Energies, Oscillator Strengths (f), Assignments, and
Polarizations of Selected Transitions Obtained from Semiempirical
INDO/S-CI Calculationsa
energy (cm-1
transn exptl calcd
)
2
Ni dz f CO π*x CT excited state (transitions 3), consistent
f
assgnt
polarizationb
z
with the opposite signs of transitions 2 and 3 in the MCD
spectrum (Figure 3).
2
2
1
9 380
3 550 0.000 09 Ni 3dxz f 3dx -y
2c
3
19 000 30 671d
Ni 3dz f CO πx,y
*
2
To validate the [PhTttBu]NiICO spectral band assignments,
transition energies were also estimated from ADF excited-
state calculations using the Slater transition state method.30
In this approach, transition energies are obtained by transfer-
ring 0.5 electron from the occupied donor orbital to the
unoccupied acceptor orbital and then calculating the energy
difference between those two orbitals. The calculated Ni 3dxz
2
21 200 34 348 0.020 09 Ni 3dz f CO πx*
24 000 36 767 0.022 77 Ni 3dz f CO πy*
26 400 38 217 0.019 95 Ni 3dyz f CO πx*
29 400 40 768 0.010 76 Ni 3dyz f CO πy*
x, y
z
z
2
4
5
6
y, z
a Experimental transition energies are also indicated for comparison.
b Polarizations are given with respect to the local coordinate system indicated
in Figure 1. c INDO/S-CI calculations neglect spin-orbit coupling between
excited states. Transition 2, therefore, acquires no intensity in the calculated
d
f 3dx -y transition energy of 12 260 cm-1 compares
reasonably well with the experimental value of ∼9400 cm-1,
affording improvement over semiempirical results. Calculated
CT transition energies of 32 641/33 222 cm-1, corresponding
2
2
2
spectrum. Average energy of all Ni 3dz f CO πx,y* spin-forbidden
transitions.
izations, and donor/acceptor MOs for the relevant spin-
allowed d f d and Ni+ f CO CT transitions obtained from
semiempirical INDO/S-CI calculations. The only calculated
d f d transition of significant intensity involves excitation
of an electron from the Ni 3dxz orbital into the singly
2
to Ni 3dz f CO π*x/π*y transitions, and 38 198/39 086
cm-1, corresponding to Ni 3dyz f CO π*x/π*y transitions,
are somewhat higher than those observed experimentally (i.e.,
MCD pseudo-A-term maxima at ∼21 200/24 000 cm-1 and
26 500 /29 300 cm-1, respectively). Overestimation of these
CT energies can be explained in terms of the known tendency
of DFT to overestimate metal-ligand covalency; i.e.,
calculated metal d orbital energies are typically too low
relative to ligand orbital energies.31 In the present case, this
preferential stabilization of metal d orbitals leads to a greater
energy difference between Ni 3d orbitals and the unoccupied
CO π* orbitals.
2
2
occupied Ni 3dx -y orbital. Thus, transition 1, corresponding
to the negative MCD feature at ∼9400 cm-1 (Figure 2), is
2
2
assigned to a Ni 3dxz f 3dx -y excitation. Calculations
indicate further that the lowest-energy CT transitions involve
2
2
Ni 3dz f CO π*x and Ni 3dz f CO π*y electronic
excitations, consistent with our rR excitation profile data
(Figure 4). Transitions 3 and 4 can accordingly be described
as perpendicularly polarized Ni+ f CO CT transitions
involving nearly-degenerate excited states, which accounts
for observation of the MCD pseudo-A-term feature centered
at ∼22 400 cm-1 (Figure 2).
Both DFT and semiempirical INDO/S-CI calculations are
internally consistent and reproduce the essential features of
the spectroscopic data, specifically a relatively high-energy
Ni-based d f d transition and dominant Ni+ f CO CT
transitions. Figure 7 depicts DFT-generated boundary surface
plots of the MOs producing the dominant contributions to
Ni+-CO bonding. The calculations show the Ni+ 3d orbitals
Assignment of transitions 5 and 6 is not obvious on the
basis of experimental data alone. While the intense MCD
pseudo A-term near ∼28 200 cm-1 appears to arise from
additional Ni+ f CO CT transitions, this assignment
accounts for neither the relatively weak C-O stretching
enhancement in rR spectra obtained from excitation energies
exceeding 27 000 cm-1 (Figure 5) nor the observation of
similar MCD features in [PhTttBu]NiIP(CH3)3 spectra (Figure
S3). However, below 40 000 cm-1 the only other calculated
transitions of considerable intensity for [PhTttBu]NiICO arise
(30) Slater, J. C. The Self-Consistent Field for Molecules and Solids:
Quantum Theory of Molecules and Solids; McGraw-Hill: New York,
1974; Vol. 4.
(31) Szilagyi, R. K.; Metz, M.; Solomon, E. I. J. Phys. Chem. A 2002,
106, 2994.
864 Inorganic Chemistry, Vol. 42, No. 3, 2003

A Ni+-CO Model Complex
Figure 7. Experimentally calibrated bonding description of [PhTttBu]-
NiICO: (A) plots of relevant MOs from DFT calculations, where the Ni
d-orbital character in the unoccupied CO π*-based MO (right) reflects a
high degree of Ni+ f CO π-back-bonding; (B) schematic MO diagram.
to be in close energetic proximity to CO π* acceptor orbitals,
suggesting substantial π-back-bonding. In actuality, π-back-
bonding interactions may be even stronger than indicated
because the DFT calculations likely underestimate Ni+
f
CO charge donation. Both the low C-O stretching frequency
of 1995 cm-1, as compared to νCO(free CO) ) 2149 cm-1,
and the strong Ni+ f CO CT absorption intensities confirm
the importance of such bonding interactions. Further, the high
energy of the Ni 3dxz f 3dx -y transition (∼9400 cm-1)
suggests significant stabilization of the Ni 3dxz donor orbital
through π-back-bonding interactions with the CO π*x,y
acceptor orbitals.
2
2
B.3. MCD Pseudo-A-Term Analysis. Two pairs of
roughly perpendicularly polarized electronic transitions, 3/4
and 5/6, give rise to two pseudo-A-term MCD features, each
a pair of oppositely signed C-terms separated by ∼3000 cm-1
(Figure 3). In both instances, the main contribution to C-term
intensity derives from spin-orbit coupling between excited
states arising from a pair of nearly-degenerate Ni 3d f CO
π*x/π*y CT transitions. Within each pair of transitions,
coupling between excited states is sizable due to their close
energetic proximities (Table 2).
Figure 8. Graphical prediction of the MCD C-term sign associated with
the Ni 3dz f CO π*x transition observed at 21 200 cm-1: (A) structure of
2
the Ni+-CO chromophore and choice of coordinate axes; (B) donor MO
2
(Ni 3dz ) and the two proximal acceptor MOs (CO π*x and CO π*y); (C)
2
2
transition dipole moments, Mx and My for Ni 3dz f CO π*x and Ni 3dz
f
CO π*y transitions, respectively; (D) graphical determination of the spin-
orbit coupling vector Lz; (E) coordinate system illustrating that the transition
dipole moment and spin-orbit coupling vectors define a right-handed
coordinate system.
To validate further the assignments of bands 3-6, the
graphical approach developed by Neese and Solomon14 was
used to predict the C-term signs of the corresponding
transitions. Figure 7 illustrates the graphical prediction of
coupling vector Lz that points along the C-O bond axis.
Together, vectors Mx, My, and Lz define a right-handed
coordinate system (Figure 8E), which leads to the absorption
of left-handed photons and thus a positive MCD signal. This
2
the C-term sign associated with the Ni 3dz f CO π*x
transition (transition 3). Figure 8A depicts the core of the
[PhTttBu]NiICO complex from which Ni+ f CO CT transi-
result supports our assignment of transition 3 to a Ni 3dz f
2
2
tions arise. The donor MO (Ni 3dz ) and the two nearly-
CO π*x excitation. As the spin-orbit coupling vector
generated by rotation of CO π*x into CO π*y (the reverse of
degenerate acceptor MOs (CO π*x and CO π*y) are sketched
in Figure 8B. The transition dipole moments, Mx and My for
CO π*y f CO π*x rotation) leads to a vector -Lz opposite
2
2
2
the Ni 3dz f CO π*x and Ni 3dz f CO π*y transitions,
respectively, obtained from semiempirical INDO/S-CI cal-
culations (Table S2) are indicated in Figure 8C. The spin-
orbit rotation of the CO π*y orbital into the CO π*x orbital
(the acceptor orbitals for transitions 4 and 3, respectively),
demonstrated in Figure 8D, leads to a positive spin-orbit
Lz, a negative MCD signal is predicted for the Ni 3dz
f
CO π*y transition, consistent with our assignment of transi-
tion 4.
Likewise, this graphical method can be applied to the Ni
3dyz f CO π*x/π*y transitions by utilizing calculated
transition dipole vectors (Table S2) and identical spin-orbit
Inorganic Chemistry, Vol. 42, No. 3, 2003 865

Craft et al.
coupling vectors for the two transitions. In this instance a
negative MCD signal is predicted for the Ni 3dyz f CO π*x
transition, whereas a positive signal is expected for the Ni
3dyz f CO π*y transition. These MCD C-term sign predic-
tions are consistent with the assignments of bands 5 and 6
presented in Table 3.
cation is transferred to ACS from CH3-CoFeSP via an SN2
mechanism. Comparison of the Ni+ and Ni2+ model calcula-
tions suggests that a Ni+-CO intermediate, despite signifi-
cant Ni+ f CO charge donation (Figure 6), is favored for
nucleophilic attack of the CoFeSP-derived methyl group, i.e.,
the greater negative charge on the Ni atom in the Ni+-CO
species, as compared to a Ni2+-CO intermediate, translates
to a better nucleophile. While the S-donor ligands in the
Ni+-CO model and ACS differ in nature (thioether ligation
as compared to thiolate ligation), a Ni+ site in the protein
would be activated further for methyl transfer due to the
greater donor strength of thiolates as compared to thioethers.
Synthetic chemistry affords additional support for such a
mechanism, as evidenced by Ni+ nucleophilic attack of alkyl
halides and the reaction of Ni+ complexes with CH3-Co3+
complexes.34-36
B.4. g-value Calculations. Depending on the number of
excited states included in the INDO/S-CI computation,
calculated g values vary considerably for the Ni+-CO
complex. This result reflects the presence of low-lying
excited states (i.e., those involving Ni-centered d f d
transitions) that preclude the use of perturbation theory to
calculate accurate g values. However, the approximately axial
EPR signal (gz ) 2.64 and gx,y ≈ 2.00) can be rationalized
on the basis of simple perturbative methods in combination
with calculated d f d transition energies. Applying the
formalism developed by Neese and Solomon,32 we utilized
the ground-state wavefunctions and d f d exited-state
energies obtained from ADF calculations were utilized to
estimate orbital angular momentum contributions to the
ground state through spin-orbit coupling with d f d excited
states. This approach gives gz ) 2.18 and gx,y ≈ 2.00,
consistent with an axial EPR signal. Deviations of gz from
2.00 result primarily from spin-orbit coupling of the Ni 3dxz
Geometry of NiFeC Species. Unlike [PhTttBu]NiICO, the
CO-bound A cluster displays gx, gy > gz. A gz value of 2.03
indicates that spin-orbit mixing of excited states into the
2
ground state via Lz is minimal, likely indicative of a Ni 3dz
2
ground state in the protein (note that Lz|Ni 3dz ) 0).
Calculated ligand-field splittings of metal d orbitals reveal
that the only coordination geometries likely to give rise to a
2
Ni 3dz ground state are a 5-coordinate trigonal bipyramid
or a 6-coordinate, tetragonally elongated octahedron.37 If
indeed the NiFeC species is formed prior to methyl transfer,
as Ragsdale and co-workers have suggested,5 a five-
coordinate trigonal bipyramidal geometry for this ACS-CO
intermediate appears reasonable; i.e., a five-coordinate
2
2
2
2
f 3 dx -y excited state into the Ni 3dx -y ground state (hole
formalism) involving the orbital angular momentum operator
2
2
Lz; i.e., the dominant matrix element is Ni 3dxy|Lz|Ni 3dx -y
) 2i.
+
Discussion
intermediate leaves an open coordination site for “CH3 ”.
2
Due to the large destabilization of the Ni 3dz orbital relative
Ni+-CO Ws Ni2+-CO Intermediates. To address the
feasibility of an Ni2+-CO enzymatic reaction intermediate,
bonding interactions in two similar Ni-CO models were
compared: the Ni+ complex [PhTttBu]NiICO and an analo-
gous hypothetical Ni2+ complex obtained using partial ADF
geometry optimization. Optimization of the Ni-CO and
C-O bond distances in both complexes shows that increasing
the oxidation state of Ni from +1 to +2 leads to lengthening
of the Ni-CO bond by 0.079 Å and shortening of the C-O
bond by 0.019 Å. Semiempirical INDO/S-CI calculations
show that altering the metal oxidation state shifts Ni f CO
CT energies significantly, with the lowest-energy transitions
occurring at ∼34 000 and ∼40 000 cm-1 for Ni+ and Ni2+
complexes, respectively. From these computations the an-
ticipated result that Ni f CO π-back-bonding is significantly
weaker in the case of the Ni2+ complex was confirmed.
Increasing the oxidation state from +1 to +2 yields
significant changes in the calculated charge on both the
carbonyl C atom (from -0.073 to +0.026) and the Ni atom
(from -0.207 to -0.031),33 thus having important mecha-
nistic implications for the methyl transfer step requisite for
ACS catalysis.
to the other d orbitals in a trigonal bipyramid,37 d f d excited
states are expected to be sufficiently high in energy such
that spin-orbit mixing into the ground state is insignificant.
In this scenario all g values deviate little from their spin-
only values, as observed experimentally.
2
Significant destabilization of the Ni 3dz orbital, which
translates spectroscopically to a gz value near 2.00, could
result from CO binding trans to the putative X-[Fe4S4]
moiety coordinated to the Ni+ site of the A cluster. Further
support for this proposed binding scheme is provided by 13
C
(I ) 1/2) and 57Fe (I ) 1/2) enriched samples: (i) The EPR
signals of the carbonyl 13C- and 57Fe-enriched NiFeC species
demonstrate hyperfine broadening,7,8 indicative of delocal-
ization of unpaired electron density onto the CO and
X-[Fe4S4] moieties. (ii) Corresponding 13C hyperfine interac-
tions are not observed for [PhTttBu]NiICO (Figure 6), where
2
2
the unpaired electron resides in the Ni 3dx -y orbital
perpendicular to the carbonyl 13C nucleus. Thus, while no
conclusive evidence for substrate binding in ACS exists, we
advocate a trigonal bipyramidal geometry at the Ni+ site,
2
with the half-occupied Ni 3dz orbital directed along the
Methyl Transfer Step. Ragsdale and co-workers5 propose
that, following formation of the NiFeC species, a methyl
(34) Goubeaud, M.; Schreiner, G.; Thauer, R. K. Eur. J. Biochem. 1997,
243, 110.
(35) Lahiri, G. K.; Stolzenberg, A. M. Inorg. Chem. 1993, 32, 4409.
(36) Hevelston, M.; Castro, C. J. Am. Chem. Soc. 1992, 114, 8490.
(37) Solomon, E. I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.; Lee,
S.-K.; Lehnert, N.; Neese, F.; Skulan, A. T.; Yang, Y.-S.; Zhou, J.
Chem. ReV. 2000, 100, 235.
(32) Neese, F.; Solomon, E. I. Inorg. Chem. 1998, 37, 6589.
(33) The calculated charges are based on Loewdin population analysis of
DFT-generated MOs.
866 Inorganic Chemistry, Vol. 42, No. 3, 2003

A Ni+-CO Model Complex
found ACS to be a relatively poor scatterer.38 In contrast,
these [PhTttBu]NiICO studies reveal that the Ni+-CO moiety
gives rise to intense MCD pseudo-A-term features, affording
an unambiguous handle on Ni-CO interactions and thus
establishing a solid basis to experimentally evaluate CO
binding in ACS in future investigations.
Scheme 1
Clearly, these Ni-CO model studies neglect an imperative
component of the A clustersthe FeS cluster component.
Thus, a more complete understanding of the ACS catalytic
mechanism will necessarily address the role of the FeS
cluster, specifically its effects on Ni site electronics and its
interactions with enzymatic substrates. Such experimental
and theoretical studies are currently underway in our
laboratories.
unique OC-Ni-X-[Fe4S4] bond axis. This geometry could
then readily accommodate methyl transfer from CH3-
CoFeSP at a position cis to the axial CO ligand (see Scheme
1; note that the only open coordination is located within the
equatorial plane), giving way to a facile migratory insertion
step.
Conclusions
Acknowledgment. Work described here was supported
by the University of Wisconsin and the Petroleum Research
Fund Grant ACS-PRF 35685-G3, administered by the
American Chemical Society (T.C.B.), the NSF Grant CHE-
9974628 and the NIH Grant R01-GM59191 (C.G.R.), and
the NSF Graduate Research Fellowship Program (J.L.C.).
We thank Kaho Kwok for collecting a portion of the MCD
data and Dr. Frank Neese (MPI Mu¨lheim) for a free copy of
his ORCA 2001 software package.
Spectroscopic and computational studies to probe the Ni-
CO bond in the [PhTttBu]NiICO model complex reveal
extensive Ni+ f CO π-back-bonding interactions. The
similarly low C-O stretching frequency observed for ACS-
CO (the NiFeC species), νCO ) 1995 cm-1, suggests
comparable Ni-CO bonding interactions occur in the puta-
tive reaction intermediate. While our studies neither confirm
nor discount the involvement of the Ni+-CO intermediate
in the ACS catalytic cycle, methyl transfer from CH3-
CoFeSP to the Ni+ site appears feasible. Further, comparison
of [PhTttBu]NiICO and ACS-CO EPR spectra has led to the
proposal of a trigonal bipyramidal geometry at the Ni+ site
in the NiFeC species, with the CO ligand trans to the
coordinated X-[Fe4S4].
Supporting Information Available: Atomic coordinates of the
[PhTttBu]NiICO computational model (Table S1), transition dipole
vectors for relevant transitions obtained from INDO/S-CI calcula-
tions (Table S2), variable-temperature MCD spectra of solid-state
[PhTttBu]NiICO (Figure S1), MCD spectra of solid-state and solution
[PhTttBu]NiICO (Figure S2), and MCD spectra of solid-state
[PhTttBu]NiICO and [PhTttBu]NiIP(CH3)3 (Figure S3). This material
is available free of charge via the Internet at http://pubs.acs.org.
Although these conclusions rest on the widely accepted
assumption that CO binds at the Ni site of the A cluster, no
experiment has definitively shown this to be the case.
Whereas rR spectroscopy might be expected to provide an
ideal probe of Ni+-CO bonding, Spiro and co-workers have
IC020441E
(38) Spiro, T. G., personal communication.
Inorganic Chemistry, Vol. 42, No. 3, 2003 867