Functional analogue reaction systems of the DMSO reductase isoenzyme family: Probable mechanism of S-oxide reduction in oxo transfer reactions mediated by bis(dithiolene)-tungsten(IV,VI) complexes

Article abstract of DOI:10.1021/ja012735p

The recent development of structural and functional analogues of the DMSO reductase family of isoenzymes allows mechanistic examination of the minimal oxygen atom transfer paradigm M^IV + QO →M^VIO + Q with the biological metals M = Mo

Full text of DOI:10.1021/ja012735p

Published on Web 03/29/2002
Functional Analogue Reaction Systems of the DMSO
Reductase Isoenzyme Family: Probable Mechanism of
S-Oxide Reduction in Oxo Transfer Reactions Mediated by
Bis(dithiolene)-Tungsten(IV,VI) Complexes
Kie-Moon Sung and R. H. Holm*
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received December 20, 2001
Abstract: The recent development of structural and functional analogues of the DMSO reductase family
of isoenzymes allows mechanistic examination of the minimal oxygen atom transfer paradigm MIV + QO
f MVIO + Q with the biological metals M ) Mo and W. Systematic variation of the electronic environment
at the WIV center of desoxo bis(dithiolene) complexes is enabled by introduction of para-substituted phenyl
groups in the equatorial (eq) dithiolene ligand and the axial (ax) phenolate ligand. The compounds [W(CO)2-
(S2C2(C6H4-p-X)2)2] (54-60%) have been prepared by ligand transfer from [Ni(S2C2(C6H4-p-X)2)2] to [W(CO)3-
(MeCN)3]. A series of 25 complexes [WIV(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2]1- ([X4,X′], X ) Br, F, H, Me, OMe;
X′ ) CN, Br, H, Me, NH2; 41-53%) has been obtained by ligand substitution of five dicarbonyl complexes
with five phenolate ligands. Linear free energy relationships between E1/2 and Hammett constant σp for the
electron-transfer series [Ni(S2C2(C6H4-p-X)2)2]0,1-,2- and [W(CO)2(S2C2(C6H4-p-X)2)2]0,1-,2- demonstrate a
substituent influence on electron density distribution at the metal center. The reactions [WIV(OC6H4-p-X′)-
(S2C2(C6H4-p-X)2)2]1- + (CH2)4SO f [WVIO(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2]1- + (CH2)4S with constant
substrate are second order with large negative activation entropies indicative of an associative transition
state. Rate constants at 298 K adhere to the Hammett equations log(k[X ,X′]/k[X ,H]) ) Faxσp and log(k[X ,X′]
/
4
4
4
k[H ,X′]) ) 4Feqσp. Electron-withdrawing groups (EWG) and electron-donating groups (EDG) have opposite
effects on the rate such that kEWG > kEDG. The effects of X′ on reactivity are found to be ∼5 times greater
than that of X (Fax ) 2.1, Feq ) 0.44) in the Hammett equation. Using these and other findings, a stepwise
oxo transfer reaction pathway is proposed in which an early transition state, of primary WIV-O(substrate)
bond-making character, is rate-limiting. This is followed by a six-coordinate substrate complex and a second
transition state proposed to involve atom and electron transfer leading to the development of the WVIdO
group. This work is the most detailed mechanistic investigation of oxo transfer mediated by a biological
metal.
4
Introduction
frequently) 1,2-diphenylethylene-1,2-dithiolate(2-) ligands in
complexes whose desoxo and monooxo coordination and metric
Our development of structural and reactivity analogues of
the active sites of the DMSO reductase (DMSOR) family of
molybdoenzymes1-3 has resulted in the functional oxo transfer
reaction systems summarized generally in Figure 1.4,5 Under
the Hille enzyme classification scheme,1 this family includes
DMSO reductase itself and trimethylamine N-oxide reductase.
Catalytic sites are characterized by two pterin dithiolene cofactor
ligands bound to the molybdenum atom. In functional systems,
we have adopted the minimal reaction paradigm MoIV + QO
f MoVIO + Q,6 and utilized 1,2-dimethylethylene- or (less
features are consistent with X-ray absorption and crystal-
lographic results for the enzymes.7-11 The paradigm now
appears to apply to the two most extensively studied DMSO
reductases, from Rhodobacter sphaeroides12 and Rhodobacter
capsulatus.13 These ligands have been selected in order to
provide a credible representation of the structural and electronic
features of the native metal coordination. Square pyramidal
(7) Czjzek, M.; Dos Santos, J.-P.; Pommier, J.; Me´jean, V.; Haser, R. J. Mol.
Biol. 1999, 284, 435-447.
(8) George, G. N.; Hilton, J.; Temple, C.; Prince, R. C.; Rajagopalan, K. V. J.
Am. Chem. Soc. 1999, 121, 1256-1266.
* Corresponding author. E-mail: holm@chemistry.harvard.edu.
(1) Hille, R. Chem. ReV. 1996, 96, 2757-2816.
(9) Li, H.-K.; Temple, C.; Rajagopalan, K. V.; Schindelin, H. J. Am. Chem.
Soc. 2000, 122, 7673-7680.
(2) Roma˜o, M. J.; Kna¨blein, J.; Huber, R.; Moura, J. J. G. Prog. Biophys.
Mol. Biol. 1997, 68, 121-144.
(10) Temple, C. A.; Hilton, J. C.; George, M. G.; Prince, R. C.; Barber, M. J.;
Rajagopalan, K. V. Biochemistry 2000, 39, 4046-4052.
(11) Bray, R. C.; Adams, B.; Smith, A. T.; Bennett, B.; Bailey, S. Biochemistry
2000, 39, 11258-11269.
(3) Kisker, C.; Schindelin, H.; Rees, D. C. Annu. ReV. Biochem. 1997, 66,
233-267.
(4) Lim, B. S.; Holm, R. H. J. Am. Chem. Soc. 2001, 123, 1920-1930.
(5) Sung, K.-M.; Holm, R. H. J. Am. Chem. Soc. 2001, 123, 1931-1943.
(6) Lim, B. S.; Sung, K.-M.; Holm, R. H. J. Am. Chem. Soc. 2000, 122, 7410-
7411.
(12) Garton, S. D.; Temple, C. A.; Dhawan, I. K.; Barber, M. J.; Rajagopalan,
K. V.; Johnson, M. K. J. Biol. Chem. 2000, 275, 6798-6805.
(13) Bell, A. F.; He, X.; Ridge, J. P.; Hanson, G. R.; McEwan, A. G.; Tonge,
P. J. Biochemistry 2001, 40, 440-448.
9
4312
J. AM. CHEM. SOC. 2002, 124, 4312-4320
10.1021/ja012735p CCC: $22.00 © 2002 American Chemical Society

S-Oxide Reduction in Oxo Transfer Reactions
A R T I C L E S
Figure 1. Upper: Depiction of previously studied oxygen atom transfer reactions of N-oxides and S-oxides mediated by bis(dithiolene)Mo/W complexes
1 and 2 (R ) Me, Ph; R′ ) Ph, p-C6H4X′, C6F5, Pri). The MoVIO reaction product is not stable to an internal redox reaction affording the indicated MoVO
product 3 and an aryloxide radical that abstracts a hydrogen atom, apparently from solvent. Lower: Oxygen atom transfer reactions in this work mediated
by WIV,VI complexes 4 and 5 with variable equatorial (X) and axial (X′) para substituents.
desoxo MoIV complex 1 reacts with S-oxides or N-oxides to
afford distorted octahedral MoVIO complex 2 and reduced
substrate.Atypicalreactionsystemcontains[Mo(OPh)(S2C2Me2)2]1-,
(CH2)4SO or Me3NO, [MoO(OPh)(S2C2Me2)2]1-, and (CH2)4S
or Me3N.4 In this and related systems, the MoVIO product is
not stable and decays to the MoVO complex 3 by an intramo-
lecular redox reaction forming a RO• radical that abstracts a
hydrogen atom, presumably from acetonitrile solvent. The results
of a detailed kinetics study of the reduction of S-oxide and
N-oxide substrates have been described.4
Oxo transfer reactions mediated by bis(dithiolene)molybde-
num and -tungsten complexes exhibit large negative entropies
of activation indicative of an associative transition state.4-6
A
transition state of this sort in the reaction system [Mo(OMe)-
(S2C2Me2)2]1-/Me2SO follows from density functional calcula-
tions and a structure has been presented.22 In our previous work,
we have proposed a stepwise process of oxo transfer consisting
of substrate binding through the oxygen atom, S-O bond
cleavage, and atom transfer to the metal atom with concomitant
oxidation of the latter. Within this manifold of events, the rate-
determining event itself remains an open question. To examine
this matter, we utilize an extensive series of complexes
[WIV(OR′)(S2C2R2)2]1- (4) in reactions with (CH2)4SO as shown
in Figure 1. Here para-substituents X and X′ of the phenyl rings
of the equatorial dithiolene and axial phenolate ligands can be
varied to provide an extended data set reflecting the dependence
of rate constants on electronic factors at the metal center. We
have selected the complexes 4 because of manipulatable
substitutents, convenient rates, very clean reactions, and the
stability of product complexes 5.
Given the existence of Mo/W isoenzymes in the DMSOR
family,14-16 a series of bis(dithiolene)WIV, WVO, and WVIO2
complexes containing the foregoing ligands has been pre-
pared.5,17-19 The analogous reaction paradigm WIV + QO f
WVIO + Q has been followed, with [W(OPh)(S2C2Me2)2]1-
,
[WO(OPh)(S2C2Me2)2]1-, and substrate comprising a represen-
tative reaction system (Figure 1). Tungsten and molybdenum
oxo transfer systems with S-oxide or N-oxide substrates exhibit
complementary second-order reactions described by the rate law
-d[MIV]/dt ) k[MIV][QO] with kW/kMo ≈ 2-30 at 298 K.4-6,20
The converse rate relationship for substrate oxidation has also
been demonstrated.21 An important difference between the
molybdenum and tungsten systems is that monooxo tungsten
complex 2 is substantially more stable than is its molybdenum
counterpart, thereby simplifying the kinetics analysis.5
Experimental Section
Preparation of Compounds. All operations were performed anaero-
bically under a pure dinitrogen atmosphere with use of an inert
atmosphere box or standard Schlenk techniques unless otherwise stated.
Acetonitrile, dichloromethane, diethyl ether, and THF were purified
with an Innovation Technology solvent purification system; 1,4-dioxane
and solvents used in column chromatography were of HPLC grade,
and deuterated solvents were stored over 4-Å molecular sieves for at
least 1 d. Water was purified by a Barnsted NANOpure Diamond
system. The compounds [Ni(S2C2Ph2)2],23 [W(CO)2(S2C2Ph2)2],17 and
(Et4N)[W(OPh)(S2C2Ph2)2]5 were prepared as previously described. The
compounds 4,4′-dibromobenzil, 4,4′-difluorobenzil, and anisoin (Alfa
(14) Buc, J.; Santini, C.-L.; Giordani, R.; Czjzek, M.; Wu, L.-F.; Giordano, G.
Mol. Microbiol. 1999, 32, 159-168.
(15) Stewart, L. J.; Bailey, S.; Bennett, B.; Charnock, J. M.; Garner, C. D.;
McAlpine, A. S. J. Mol. Biol. 2000, 299, 593-600.
(16) Stewart, L. J.; Bailey, S.; Collison, D.; Morris, G. A.; Garner, C. D.
ChemBioChem 2001, 2, 703-706.
(17) Goddard, C. A.; Holm, R. H. Inorg. Chem. 1999, 38, 5389-5398.
(18) Sung, K.-M.; Holm, R. H. Inorg. Chem. 2000, 39, 1275-1281.
(19) Sung, K.-M.; Holm, R. H. Inorg. Chem. 2001, 40, 4518-4525.
(20) Ueyama, N.; Oku, H.; Nakamura, A. J. Am. Chem. Soc. 1992, 114, 7310-
7311.
(22) Webster, C. E.; Hall, M. B. J. Am. Chem. Soc. 2001, 123, 5820-5821.
(23) Schrauzer, G. N.; Mayweg, V. P. J. Am. Chem. Soc. 1965, 87, 1483-
1489.
(21) Tucci, G. C.; Donahue, J. P.; Holm, R. H. Inorg. Chem. 1998, 37, 7, 1602-
1608.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4313

A R T I C L E S
Sung and Holm
Aesar) and 4,4′-dimethylbenzil (Aldrich) were commercial samples and
were used without further purification. The compounds NaOC6H4-p-
X′ (X′ ) CN, Br, H, Me, NH2) were prepared as described elsewhere.5
Tetramethylene sulfoxide was distilled over CaH2 and stored under
dinitrogen. Solvent removal and drying steps were performed in vacuo.
Because of the large number of related compounds prepared, charac-
terization information (elemental analyses and/or absorption and NMR
spectral data) are limited to representative compounds of each type.
Further spectral data are available as Supporting Information.24
[Ni(S2C2(C6H4-p-X)2)2] (X ) Br, F, Me, OMe). These complexes
were prepared by a procedure analogous to that for [Ni(S2C2Ph2)2]23
with use of 4,4′-disubstituted benzil ((p-XC6H4CO)2) or benzoin (p-
XC6H4CH(OH)COC6H4-p-X). A mixture of benzil or benzoin (3-4 g)
and 1-1.5 equiv of P2S5 in 20 mL of dioxane was heated to reflux at
110 °C for 1-4 h. When the initial yellow suspension became a clear
dark brown solution, the latter was cooled to room temperature and
filtered to remove an insoluble pale yellow solid. To the filtrate was
added a solution of 0.5 equiv of NiCl2‚6H2O per equiv of benzil or
benzoin in 1 mL of water, and the mixture was heated to reflux until
a black crystalline precipitate developed. The volume of the reaction
mixture was reduced to half and the precipitate was collected by
filtration, washed with a minimal amount of dioxane, water, methanol,
and ether, and dried in air to afford the product as a dark green-brown
solid (57-64%).25 These compounds are satisfactorily identified by
spectral data and by electrochemical properties (vide infra).
Hz). 19F{1H} NMR (acetone-d6) δ -114.81. Absorption spectrum (THF)
λmax (ꢀM) 313 (sh, 14 700), 538 (13 000), 660 (sh, 1780) nm. Anal.
Calcd for C30H16O2F4S4W: C, 45.24; H, 2.02; F, 16.10; S, 9.54.
Found: C, 46.80; H, 3.15 F, 15.70; S, 9.22.
X ) Me. Chromatography: ether/n-pentane ) 1/9 (v/v), Rf ) 0.85.
1
IR (KBr) VCO 2024 (vs), 1974 (s) cm-1. H NMR (acetone-d6) δ 2.33
(s, 12), 7.14 (d, 8, J ) 8.8 Hz), 7.20 (d, 8, J ) 8.0 Hz). Absorption
spectrum (THF) λmax (ꢀM) 317 (sh, 13 600), 556 (13 400) nm. Anal.
Calcd for C34H28O2S4W: C, 52.31; H, 3.61; S, 16.43. Found: C, 52.48;
H, 3.66; S, 16.51.
X ) OMe. Chromatography: ether/n-pentane ) 1/1 (v/v), Rf ) 0.94.
1
IR (KBr) vCO ) 2023 (vs), 1974 (s) cm-1. H NMR (acetone-d6) δ
3.83 (s, 12), 6.89 (d, 8, J ) 9.2 Hz), 7.24 (d, 8, J ) 8.8 Hz). Absorption
spectrum (THF) λmax (ꢀM) 319 (sh, 21 300), 587 (21 000) nm. Anal.
Calcd for C34H28O6S4W: C, 48.35; H, 3.34; S, 15.18. Found: C, 48.52;
H, 3.43; S, 15.29.
(Et4N)[W(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2] (X ) Br, F, H, Me,
OMe; X′ ) CN, Br, H, Me, NH2). Twenty four compounds have been
prepared by a procedure analogous to that for (Et4N)[W(OPh)(S2C2-
Ph2)2]5 from the combination of five dicarbonyl complexes [W(CO)2-
(S2C2(C6H4-p-X)2)2] and five ligands NaOC6H4-p-X′ on a 0.03-0.05
mmol scale. To a suspension of NaOC6H4-p-X′ and 1 equiv of Et4NCl
in 0.5 mL of acetonitrile was added a solution of [W(CO)2(S2C2(C6H4-
p-X)2)2] in 1.5 mL of THF. The color of the solution changed to brown
instantly and the solvents were evaporated. The resulting dark brown
solid was redissolved in a minimal volume of acetonitrile, filtered, and
recrystallized by addition of ether. Products were obtained as black or
dark brown crystals (41-53%).
1
X ) Br. H NMR (CDCl3) δ 7.25 (d, 8, J ) 8.0 Hz), 7.44 (d, 8, J
) 8.0 Hz). Absorption spectrum (dichloromethane) λmax (ꢀM) 280 (26
600), 323 (34 600), 594 (1530), 864 (21 800) nm.
X ) F. 1H NMR (CDCl3) δ 7.01 (t, 8, JH-H ) JH-F ) 9.2 Hz), 7.35
(dd, 8, JH-H ) 8.8 Hz, JH-F ) 5.2 Hz). 19F{1H} NMR (CDCl3) δ
-111.49. Absorption spectrum (dichloromethane) λmax (ꢀM) 272
(24 400), 317 (31 000), 594 (1430), 859 (20 600) nm.
1
X ) Br, X′ ) H. H NMR (CD3CN, anion) δ 6.58 (d, 2, J ) 7.2
Hz), 6.91 (t, 1, J ) 7.6 Hz), 7.10 (t, 2, J ) 8.4 Hz), 7.23 (d, 8, J ) 8.8
Hz), 7.38 (d, 8, J ) 8.4 Hz). Absorption spectrum (acetonitrile) λmax
(ꢀM) 307 (35 200) nm. Anal. Calcd for C42H41NOBr4S4W: C, 41.78;
H, 3.42; N, 1.16; Br, 26.49; S, 10.62. Found: C, 41.59; H, 3.36; N,
1.22; Br, 26.56; S, 10.52.
X ) Me. 1H NMR (CDCl3) δ 2.32 (s, 12), 7.09 (d, 8, J ) 8.0 Hz),
7.28 (d, 8, J ) 8.0 Hz). Absorption spectrum (dichloromethane) λmax
(ꢀM) 277 (34 200), 321 (38 800), 602 (1862), 877 (28 400) nm.
1
X ) F, X′ ) H. H NMR (CD3CN, anion) δ 6.58 (d, 2, J ) 7.2
1
X ) OMe. H NMR (CDCl3) δ 3.82 (s, 12), 6.82 (d, 8, J ) 8.0
Hz), 6.91 (t, 1, J ) 7.6 Hz), 6.98 (t, 8, JH-H ) JH-F ) 8.8 Hz), 7.10
(t, 2, J ) 8.4 Hz), 7.31 (dd, 8, JH-H ) 8.4 Hz, JH-F ) 5.6 Hz). 19F-
{1H} NMR (CD3CN) δ -118.27. Absorption spectrum (acetonitrile)
Hz), 7.34 (d, 8, J ) 8.0 Hz). Absorption spectrum (dichloromethane)
λmax (ꢀM) 301 (28 000), 335 (20 100), 629 (1230), 928 (20 400) nm.
[W(CO)2(S2C2(C6H4-p-X)2)2] (X ) Br, F, OMe, Me). These
compounds were synthesized by a procedure similar to that for
[W(CO)2(S2C2Ph2)2].17 A mixture of [W(CO)3(MeCN)3] (0.8-1.4
mmol) and [Ni(S2C2(C6H4-p-X)2)2] in a 1:4 mol ratio was suspended
in 50 mL of dichloromethane. A dark violet or purple solution developed
immediately. The reaction mixture was stirred for 24 h, the volume of
the reaction mixture was reduced to 2-3 mL in vacuo, and the
concentrated solution was eluted on a silica gel column in di-
chloromethane to remove an insoluble major impurity. A second column
chromatography with an ether/pentane eluant system led to the
separation of a dark violet or purple band, which was collected. The
solvent was removed to give the product as a black or dark violet solid
(54-60%). The chromatographic steps were performed in air.
X ) Br. Chromatography: ether/n-pentane ) 1/9 (v/v), Rf ) 0.90.
λmax (ꢀM) 306 (34 200) nm. Anal. Calcd for C42H41NOF4S4W: C, 52.34;
H, 4.29; N, 1.45; F, 7.88; S, 13.31. Found: C, 52.38; H, 4.32; N, 1.51;
F, 7.96; S, 13.28.
X ) Me, X′ ) H. 1H NMR (CD3CN, anion) δ 2.31 (s, 12), 6.57 (d,
2, J ) 8.4 Hz), 6.88 (t, 1, J ) 7.2 Hz), 7.05 (d, 8, J ) 8.4 Hz), 7.09
(t, 2, J ) 7.6 Hz), 7.20 (d, 12, J ) 8.0 Hz). Absorption spectrum
(acetonitrile) λmax (ꢀM) 306 (38 800) nm. Anal. Calcd for C46H53-
NOS4W: C, 58.28; H, 5.63; N, 1.48; S, 13.53. Found: C, 58.12; H,
5.72; N, 1.52. S, 13.45.
1
X ) OMe, X′ ) H. H NMR (CD3CN, anion) δ 3.76 (s, 12), 6.57
(d, 2, J ) 8.8 Hz), 6.79 (d, 8, 8.8 Hz), 6.88 (t, 1, J ) 7.6 Hz), 7.08 (t,
2, J ) 8.4 Hz), 7.25 (d, 8, J ) 8.8 Hz). Absorption spectrum
(acetonitrile) λmax (ꢀM) 310 (43 400), 492 (sh, 1410), 714 (1020) nm.
Anal. Calcd for C46H53NO5S4W: C, 54.59; H, 5.28; N, 1.38; S, 12.67.
Found: C, 54.51; H, 5.35; N, 1.32; S, 12.62.
1
IR (KBr) VCO ) 2032 (vs), 1985 (s) cm-1. H NMR (acetone-d6) δ
7.56 (d, 8, J ) 8.8 Hz), 7.28 (d, 8, J ) 8.8 Hz). Absorption spectrum
(THF) λmax (ꢀM) 311 (sh, 18 900), 540 (20 700), 660 (sh, 2330) nm.
Anal. Calcd for C30H16O2Br4S4W: C, 34.64; H, 1.55; Br, 30.73; S,
12.33. Found: C, 34.57; H, 1.62 Br, 30.57; S, 12.46.
Kinetics Measurements. All reactions were monitored under
anaerobic conditions in acetonitrile solutions with a Cary 3 spectro-
photometer operating at 250-900 nm. Reduction reactions of tetra-
methylene sulfoxide to tetramethylene sulfide by the series of W(IV)
complexes were run under pseudo-first-order conditions at 25 ( 0.5
°C. Initial W(IV) concentrations were [WIV]0 ) 0.50-1.2 mM in sys-
tems containing 150-4000 equiv of TMSO. As the reaction proceeds,
the color of the solution changes from yellow-brown to violet. The
spectral changes accompanying each reaction are similar; a sharp isos-
bestic point is observed at 340-350 nm and a distinct band of the
product W(VI) complex develops at λmax ) 517-520 nm, at which
X ) F. Chromatography: ether/n-pentane ) 1/9 (v/v), Rf ) 0.87.
1
IR (KBr) VCO 2034 (vs), 1989 (s) cm-1. H NMR (acetone-d6) δ 7.13
(t, 8, JH-H ) JH-F ) 9.2 Hz), 7.36 (dd, 8, JH-H ) 8.8 Hz, JH-F ) 5.2
(24) See the paragraph at the end of this article for Supporting Information
available.
(25) These compounds have been reported to be obtained in higher yield with
use of different stoichiometry of reactants in 1,2-diethyl-2-imidazolidinone
as solvent: Takuma, K.; Irizato, Y.; Katho, K. PCT Int. Appl. 1990; patent
no. WO9012019. However, we were unable to reproduce these results and
thus followed the method of Schrauzer and Mayweg.23
wavelength absorbances were taken for the evaluation of rate constants
kobs and k[X ,X′]
.
4
9
4314 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

S-Oxide Reduction in Oxo Transfer Reactions
A R T I C L E S
Figure 2. Synthesis of the complexes [Ni(S2C2(C6H4-p-X)2)2], [W(CO)2(S2C2(C6H4-p-X)2)2], and [W(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2] (4). The matrix
[X4,X′] of 25 WIV complexes utilized in this work is indicated.
V)27 are substantial, as are those for [WO(S2C2R2)2]1-/2- (0.29
Standard deviations for rate constants were estimated by using a
V)17 and [W(CO)2(S2C2R2)2]0/1-/2- (0.21, 0,20 V).28 In these
linear least-squares error analysis with uniform weighting of data
points.26 Other details of the experimental procedure and data analysis
and all other cases, EPh > EMe for corresponding couples,
consistent with the relative inductive effects of phenyl and
are described in a previous report.5
Other Physical Measurements. Absorption spectra were obtained
methyl groups. In metal dithiolene electron transfer series, the
on Varian Cary 50 Bio or Cary 3 spectrophotometers. 1H and 19F NMR
ligands may assume a non-innocent role and, particularly in the
spectra were taken with a Varian Mercury 400 spectrometer: chemical
shifts were referenced to Me4Si or CCl3F (19F). IR spectra were
measured in KBr pellets with a Nicolet Nexus 470 FT-IR instrument.
Electrochemical measurements were performed with a PAR Model 263
potentiostat/galvanostat with a Pt working electrode and 0.1 M Bu4-
NPF6 supporting electrolyte in dichloromethane solutions. Potentials
are referenced to the saturated calomel electrode (SCE).
more oxidized forms, metal and ligand oxidation states are not
29
well-defined. This is the case in two series, [Ni(S2C2Me2)]2-,1-,0
and [W(CO)2(S2C2Me2)2]2-,1-,0 28
, certain members of which are
pertinent to this work. However, in the M ) Mo and W
complexes 1-3, ring C-C and C-S bond lengths conform to
the values expected for double and single bonds, respective-
ly.5,17-19,30 Consequently, from a structural criterion the ligands
behave as classical ene-1,2-dithiolate dianions in the indicated
oxo transfer reactions and metal oxidation states of reactant 1
and product 2 are well-defined.
With the foregoing background in mind, we have selected
oxo transfer reactions between the desoxo WIV complexes 4
and the constant substrate (CH2)4SO in acetonitrile solution for
detailed kinetics evaluation. With reference to Figure 2, the
complexes [Ni(S2C2(C6H4-p-X)2)2] are first prepared by the
Results and Discussion
Design of Complexes. The oxo transfer reactions mediated
by the M ) Mo and W complexes in Figure 1 involve species
with chelate ring substitutents R ) Me or Ph. As has been long
recognized with metal dithiolenes, the principal electronic
function of R substituents is modulation of redox potentials27
by influencing electron distribution and frontier orbital energies.
For example, potential differences between the R ) Me/Ph
couples [Ni(S2C2R2)2]1-/2- (0.32 V) and [Ni(S2C2R2)2]0/1- (0.28
(28) Fomitchev, D. V.; Lim, B. S.; Holm, R. H. Inorg. Chem. 2001, 40, 645-
(26) Clifford, A. A. MultiVariate Error Analysis; Halsted Press: New York,
1973.
(27) McCleverty, J. A. Prog. Inorg. Chem. 1968, 10, 49-221.
654.
(29) Lim, B. S.; Fomitchev, D. V.; Holm, R. H. Inorg. Chem. 2001, 40, 4257-
4262.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4315

A R T I C L E S
Sung and Holm
indicated method.23 Several examples of this general type of
complex had been synthesized previously,23,31,32 and substitutent
X varied to tune the energy of the intense near-infrared
electronic band characteristic of neutral nickel bis(dithiolene)s.
An extensive set has been reported in the Japanese patent
literature.25 The indicated ligand transfer reaction17 affords
[W(CO)2(S2C2(C6H4-p-X)2)2] (54-60%). The carbonyl groups
are readily displaced by phenolates to give the desired complexes
[W(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2] (4, 41-53%), which un-
doubtedly have the square-pyramidal structure of the parent or
reference complex [W(OPh)(S2C2Ph2)2]1-.5 The 25 complexes
of interest here are displayed in matrix form in Figure 2.
Hereafter, individual complexes are designated as [X4,X′]; only
the parent [H4,H] has been prepared previously.
We note the facility with which the synthetic scheme
conveniently accommodates independent electronic variation of
axial and equatorial ligands within the same molecular frame-
work. Equatorial ligand substitutents are varied four at a time.
We are unaware of another confluence of ligand structure and
metal that allows a similarly extensive electronic perturbation
of a parent molecule under the constraints of effectively constant
molecular structure and steric influence on substrate binding.
Linear Free Energy Relationships. Before turning to oxo
transfer kinetics, we provide evidence that substituent variation
in the groups R ) p-C6H4X/X′ can lead to linear free energy
relationships (LFERs). Although redox data abound for one or
both steps of electron transfer series 1, there has been only one
previous analysis of the dependence of potentials on R groups.33
Series 2, whose neutral members are precursors to the set
[X4,X′], is likewise subject to an analysis of substituent
dependence of potentials. The well-defined chemically reversible
one-electron redox steps of these series are apparent from the
cyclic voltammograms in Figure 3.
Figure 3. Cyclic voltammograms (100 mV/s) of [Ni(S2C2(C6H4-p-X)2)2]
(upper) and [W(CO)2(S2C2(C6H4-p-X)2)2] (lower) with X ) Br and OMe
in dichloromethane, demonstrating the three-member electron-transfer series.
Peak potentials are indicated. Complexes with X ) F, H, and Me behave
similarly.
[Ni(S2C2R2)2]0 T [Ni(S2C2R2)2]1- T [Ni(S2C2R2)2]2- (1)
Table 1. Reversible Redox Potentials for Bis(dithiolene) Nickel
and Tungsten Complexes in Dichloromethane
[W(CO)2(S2C2R2)2]0 T [W(CO)2(S2C2R2)2]1-
T
a
complex
E
1/2, V (∆E , mV)
p
[W(CO)2(S2C2R2)2]2- (2)
Ni
W
[Ni(S2C2(C6H4-p-Br)2)2]
[Ni(S2C2(C6H4-p-F)2)2]
[Ni(S2C2Ph2)2]b
[Ni(S2C2(C6H4-p-Me)2)2]
[Ni(S2C2(C6H4-p-OMe)2)2]
0.075 (110), -0.738(100)
0.032 (110), -0.794 (110)
-0.045 (110), -0.870 (110)
-0.097 (140), -0.922 (140)
-0.113 (110), -0.932 (110)
Values of E1/2 for the two series with X ) Br, F, H, Me, and
OMe are collected in Table 1. The spread in potentials is 190
mV in series 1 and 120-130 mV in series 2. Because potentials
are reproducible to (10 mV, the potential variations are
adequate for analysis using Hammett eq 3 and relationships 4
and 5 (T ) 298 K). Plots of (EX - EH)/0.059 vs 4σp, shown
[W(CO)2(S2C2(C6H4-p-Br)2)2]
[W(CO)2(S2C2(C6H4-p-F)2)2]
[W(CO)2(S2C2Ph2)2]c
[W(CO)2(S2C2(C6H4-p-Me)2)2]
[W(CO)2(S2C2(C6H4-p-OMe)2)2]
-0.458 (80), -0.971 (90)
-0.495 (100), -1.010 (110)
-0.540 (90), -1.060 (90)
-0.568 (90), -1.089 (90)
-0.578 (80), -1.101 (90)
log(KX/KH) ) 4Fσp
∆∆Go ) -F(EX - EH) ) -2.303RT log(KX/KH) (4)
log(KX/KH) ) (EX - EH)/0.059
(3)
a E1/2 ) (Epa + Epc)/2, 100 mV/s, vs SCE, at ∼25 °C. b In DMF, 0.13
and -0.88 V, respectively, ref 33. c In acetonitrile, -0.37 and -0.93 V,
repectively, ref 17.
(5)
∆∆Go is larger in series 1 than series 2 (FNi/FW ) 1.5). In the
latter, back-bonding of the carbonyl groups may act as a leveling
effect in electronic changes produced by variation of X. Values
of VCO are J10 cm-1 lower for X ) Me, OMe than for X )
Br, F, consistent with marginally greater back-bonding in
molecules with the more electron releasing substituents. The
principal point is that electronic effects of substituents are
propagated through phenyl groups to the unsaturated chelate
in Figure 4, are linear with the slopes FNi ) 2.0 and FW ) 1.3.
Potentials of both steps are taken into account to obtain EX -
EH values. The substituent effect on the free energy difference
(30) Lim, B. S.; Donahue, J. P.; Holm, R. H. Inorg. Chem. 2000, 39, 263-273.
(31) Freyer, W. Z. Chem. 1980, 24, 32-33.
(32) Mueller-Westerhoff, U. T.; Vance, B.; Yoon, D. I. Tetrahedron 1991, 47,
909-932.
(33) Olson, D. C.; Mayweg, V. P.; Schrauzer, G. N. J. Am. Chem. Soc. 1966,
88, 4876-4882.
9
4316 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

S-Oxide Reduction in Oxo Transfer Reactions
A R T I C L E S
Figure 4. Hammett plots of redox potentials vs 4σp for [Ni(S2C2(C6H4-
p-X)2)2] and [W(CO)2(S2C2(C6H4-p-X)2)2]. The first (9 Ni, b W) and second
(0 Ni, O W) potential differences relative to the X ) H species are nearly
constant and are taken together to generate the linear free energy
relationships for the two types of complexes.
Figure 5. Absorption spectral changes in the pseudo-first-order reaction
of [WIV(OPh)(S2C2(C6H4-p-F)2)2]1- and (CH2)4SO in acetonitrile at 298 K
([F4,H]0 ) 0.69 mM, [(CH2)4SO]0 ) 0.36 M). Spectra were recorded every
4 min. The inset shows the tight isosbestic point maintained throughout
the reaction.
rings of the nickel29 and tungsten28 complexes to measurable
extents which are manifested as LFERs.
decomposition is qualitatively faster with the stronger electron-
withdrawing groups X/X′, which presumably effect a more
electron-deficient WVI center.
(b) Substituent Dependence of Rate Constants. The rate
constant matrix of Table 2 demonstrates that the rates of reaction
6 are sensitive to the substituents X/X′, with relative rate
Kinetics and Mechanism of Oxo Transfer. (a) Reaction
Systems. Previously, we have shown that the rate constants for
reaction 6 (R ) Me) in acetonitrile at 298 K adhere to the
Hammett equation log(kX′/kH) ) Fσp, with the positive slope F
) 2.1 showing that electron-withdrawing groups enhance the
rate.5 Using the same substrate, selected because it offers
minimal steric hindrance of any S-oxide and reacts ca. 20 times
faster than Me2SO under the same conditions,5 we provide a
more extensive examination of reaction 6 using the complexes
constants reaching a maximum at k[Br ,CN]/k[(OMe) ,NH ] ≈ 390.
Hammett eqs 8 and 9 are used for analysis of the influence of
para-substituents on reaction rates. Each column of the rate
matrix represents a set of five rate constants that are solely
dependent on variable substituent X′ of the axial phenolate
ligand at constant X (eq 8). Likewise, each row represents a
set of five rate constants that are solely dependent on variable
substituent X of the equatorial ligands at constant X′ (eq 9).
4
4
2
[WIV(OC6H4-p-X′)(S2C2R2)2]1- + (CH2)4SO f
[WVIO(OC6H4-p-X′)(S2C2R2)2]1- + (CH2)4S (6)
[X4,X′]. The reactions follow the second-order rate law 7 as
log(k[X ,X′]/k[X ,H]) ) Faxσp
log(k[X ,X′]/k[H ,X′]) ) 4Feqσp
(8)
(9)
4
4
-d[WIV]/dt ) k[WIV][(CH2)4SO]
(7)
4
4
previously shown. Kinetics under pseudo-first-order conditions
afford linear plots of kobs vs [(CH2)4SO], from which the second-
The 10 data sets when plotted as eqs 8 and 9 afford the linear
relationships evident in Figure 6. Values of Fax and Feq were
obtained from the averaged slopes for axial X′ and equatorial
X contributions, respectively. The slopes are positive, indicating
that relative to [H4,H] an electron-withdrawing group (EWG)
accelerates the reaction and an electron-donating group (EDG)
has the opposite effect. The value Fax ) 2.1 matches exactly
that for the system [W(OC6H4-p-X′)(S2C2Me2)2]1-/(CH2)4SO,5
supporting the consistency of the results. The much smaller
value Feq ) 0.44 with Fax/Feq ≈ 4.8 results from the less
significant influence of X, which is located farther from the
tungsten center than X′. That kEWG > kEDG is further made
evident by the three-dimensional plot of rate constants relative
to the reference complex [H4,H] in Figure 7. Complexes lying
order rate constants k[X ,X′] are evaluated. Reactions are notably
4
clean and exhibit tight isosbestic points, as is illustrated by
spectrophotometric monitoring of the reaction involving [F4,H]
in Figure 5. Final spectra are closely related to that of
[WO(OPh)(S2C2Me2)2]1- (λmax 408, 514, 637 nm),5 and provide
certain identification of the [WO(OC6H4-p-X′)(S2C2(C6H4-p-
X)2)2]1- reaction products. While not highly stable, [WO(OPh)-
(S2C2Me2)2]1- has been isolated as its Et4N+ salt and shown to
have distorted octahedral stereochemistry.5,6 In general, these
WVIO complexes are stable over the period of the kinetics
measurement, but at longer times (3-5 h after oxo transfer is
complete) slowly undergo an autoredox reaction to form blue-
green solutions of [WVO(S2C2(C6H4-p-X)2)2]1- (λmax ∼ 715 nm).
The analogous molybdenum complex [MoO(OPh)(S2C2Me2)2]1-
is too unstable to isolate and also undergoes autoreduction to
liberate the phenoxy radical that is detected as phenol.4 A similar
instability pathway is likely for the tungsten complexes. Their
above the log(k[X ,X′]/k[H ,H]) ) 0 plane carry combinations of
X ) Br, F and X′ ) Br, CN.
4
4
(c) Activation Parameters. These parameters were deter-
mined for three reactions and are summarized in Table 3 together
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4317

A R T I C L E S
Sung and Holm
Table 2. Second-Order Rate Constants for the Reduction of (CH2)4SO Mediated by Bis(dithiolene)W(IV) Complexes [X4,X′] (4)a
[X ,X′] (M-1s-1 b
)
4
k
X′
X ) Br
0.51(2)
X ) F
0.21(2)
X) H
0.14(1)
X) Me
X ) OMe
CN
Br
H
Me
NH2
6.6(1) × 10-2
1.4(1) × 10-2
2.7(3) × 10-3
1.7(4) × 10-3
1.4(2) × 10-3
5.5(1) × 10-2
1.0(2) × 10-2
2.1(1) × 10-3
1.7(1) × 10-3
1.3(3) × 10-3
5.9(1) × 10-2
1.4(1) × 10-2
7.7(1) × 10-3
4.7(3) × 10-3
2.9(1) × 10-2
6.6(1) × 10-3
5.2(1) × 10-3
3.2(1) × 10-3
1.7(2) × 10-2
3.0(1) × 10-3 c
2.0(5) × 10-3
1.8(1) × 10-3
a In acetonitrile, at 298 K. b -d[WIV]/dt ) k[X ,X′][WIV][(CH2)4SO] for the reaction series 4 + (CH2)4SO f 5 + (CH2)S. c Reference 5.
4
Figure 8. Linear free energy relationships in the reduction of (CH2)4SO
Figure 6. Hammett plots of the relative rates of reaction 6 as log(k[X ,X′]
/
4
mediated by [W(OC6H4-p-X′)(S2C2Me2)2]1- (X′ ) CN, H, NH2) (b) and
[X4,H] (X ) Br, H, OMe) (O) in acetonitrile at 298 K.
k[X ,H]) vs σp (b, variable X′ at specified X) and log(k[X ,X′]/k[H ,X′]) vs 4σp
(O, variable X at specified X′).
4
4
4
transition state. Relative rate constants at 298 K are within a
factor of 100 for the six reactions, resulting in a variation of
only 2.8 kcal/mol in ∆Gq298. However, values of ∆Hq and ∆Sq
are somewhat differently apportioned in the two sets. For the
reactions of [Br4,H], [H4,H], and [(OMe)4,H] values of |∆Hq|
are lower and those of |∆Sq| are higher, such that T∆Sq does
not vary widely and is 48-57% of ∆Gq298. As anticipated by
Hammett plots (Figure 6), there is a linear correlation between
∆Gq298 and substituent parameters of the two sets of complexes
in Table 3. These LFERs, demonstrated in Figure 8, emphasize
the influence of substituents on rate. Given the linear relation-
ships in Figures 6 and 8, the factors differentiating rates we
take to be largely enthalpic in nature. Irregular activation
entropies are unlikely to afford linear behavior. We assume the
variable enthalpic contributions to apply to all reactions of
[X4,X′] complexes.
(d) Mechanistic Considerations. We proceed from the
establishment of an associative transition state in and linear free
energy relationships for reaction 6. The data of Table 1 clearly
reveal that EWGs increase redox potentials. This property of
increasing ease of oxidation should extend to desoxo WIV,
although it cannot be directly established here because the parent
complex [H4,H] and other [X4,X′] species do not show reversible
Figure 7. Three-dimensional display of the relative rates of reaction 6 as
log(k[X ,X′]/k[H ,H]) vs X and X′. The black squares lie on the log 1 ) 0
plane; symbols extending above and below this plane correspond to
complexes [X4,X′] whose rate constants are larger and smaller, respectively,
than for [H4,H].
4
4
with those for three reactions of [W(OC6H4-p-X′)(S2C2Me2)2]1-
included for comparison. The slowest and fastest reactions with
the constant axial substituent X′ ) H were investigated in the
[X4,H] series. As in previous reactions of [M(OPh)(S2C2Me2)2]1-
(M ) Mo, W) with N-oxides and S-oxides,4-6 activation
entropies are large and negative, consistent with an associative
redox reactions. Rate constants exhibit the order kEWG > kEDG
,
which is opposite to the expected ease of oxidation of WIV,
i.e., EEWG > EEDG. There being no other suitable measure
available, we take this relationship of ground-state potentials
to reflect the intrinsic relative oxidizability of a WIV center in
a transition state, whether by partial charge transfer or discrete
9
4318 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

S-Oxide Reduction in Oxo Transfer Reactions
A R T I C L E S
Table 3. Activation Parameters for the Reduction of (CH2)4SO Mediated by Bis(dithiolene)W(IV) Complexes
q
q
W(IV) complex
ka (M-1 s-1
)
∆H (kcal/mol)
∆S (eu)
∆Gq a (kcal/mol)
[W(OC6H4-p-CN)(S2C2Me2)2]1- b
[W(OPh)(S2C2Me2)2]1- b
3.5(1) × 10-2
9.0(3) × 10-4
3.8(7) × 10-4
1.4(1) × 10-2
3.0(1) × 10-3 b
2.1(1) × 10-3
10.8(1)
11.6(4)
15(1)
8.6(7)
9.0(4)
10.9(1)
-29(1)
-33(1)
-24(4)
-38(5)
-40(8)
-34(4)
19.4
21.4
22.2
20.0
20.8
21.0
[W(OC6H4-p-NH2)(S2C2Me2)2]1- b
[W(OPh)(S2C2(C6H4-p-Br)2)2]1-
[W(OPh)(S2C2Ph2)2]1-
[W(OPh)(S2C2(C6H4-p-OMe)2)2]1-
a In acetonitrile, at 298 K, -d[WIV]/dt ) k[WIV][(CH2)4SO]. b Data from ref 5.
interpretation, the rate-determining transition state is early. Its
exact geometry cannot be described, but from the Hammond
postulate it presumably requires reduction of the chelate ring
dihedral angle from the original 128° in [H4,H]. The extent to
which it approaches ca. 100°, as in [WO(OPh)(S2C2Me2)2]1- 5
,
is unknown. In support of a transition state with dominant
WIV‚‚‚OQ bond-making in related systems is our previous
observation of parallel trends in rate constants for the reaction
of [W(OPh)(S2C2Me2)2]1- and QO and pKa values of the
substrate.5 Webster and Hall22 have proposed a transition state
model for the simplified reaction [M(OMe)(S2C2Me2)2]1-
+
Me2SO f [MO(OMe)(S2C2Me2)2]1- + Me2S on the basis of
density functional calculations. From theoretical M-OSMe2 and
O-S bond lengths, the transition state is substantially biased
toward the T2 description. The reaction [MoO2(SCH2CHNH)2]
+ Me3P f [MoO(SCH2CHNH)2] + Me3PO, a model for an
actual system,34 has also been examined by DFT calculations.35
In both cases a single transition state is considered for the oxo
transfer event. Our proposal of two such states follows from
the lack of correlation of WIV oxidizability and rate constants.
It is well established that for reduction of the same substrate
mediated by strictly analogous bis(dithiolene) molybdenum and
tungsten complexes, kW > kMo, although the effect is not large.
Thus for reaction 10 in acetonitrile at 298 K, kW/kMo ) 6 for R
) Me and 8.8 for R ) Ph.4-6 For single-turnover reaction 11
of R. capsulatus DMSOR isoenzymes, kW/kMo ) 17.15,16 While
Figure 9. Proposed qualitative reaction coordinate for reaction 6 with
schematic representations of transition states (T1, T2) and intermediate (I).
The thick line refers to a system initially containing reference complex
[W(OPh)(S2C2Ph2)2]1-, and other lines to complexes [W(OR′)(S2C2R2)2]1-
containing electron-donating groups (EDG) or electron-withdrawing groups
(EWD) X/X′ in the substituents R′ ) p-C6H4X′ and R ) p-C6H4X.
[MIV(OPh)(S2C2R2)2]1- + (CH2)4SO f
[MoVIO(OPh)(S2C2R2)2]1- + (CH2)4SO (10)
MIV-DMSOR + Me2SO f MVIO-DMSOR + Me2S (11)
oxidation. The lack of correlation between oxidizability and rate
indicates that the event of electron transfer, wherein an oxygen
atom is transferred to the metal center with concomitant
oxidation to WVI and reduction to oxide, is not, or is not
simultaneous with, the rate-determining step. We conclude that
we have not carried out a similar study of the molybdenum
analogues of reaction 6, a reasonable inference is that the
pathway of Figure 9 applies to the reactions of [MoIV(OC6H4-
p-X′)(S2C2R2)2]1- 4
. In this event, the binding of constant
substrate to MoIV is weaker than that for WIV. Structural and
mechanistic studies of molybdoenzymes of the DMSOR family
have concentrated on catalytic cycles.12,13,36-39 Little information
is available on rate constants for individual steps in the cycle.
The intrinsic property kW > kMo at parity of substrate and
complex structure does not, however, necessarily translate into
faster substrate reduction reactions for tungstoenzymes in vivo.
Garner and co-workers16 have provided the instructive demon-
the most probable description of the reaction coordinate involves
two transition states and one intermediate in the stepwise
pathway set out in Figure 9. The first transition state T1 primarily
involves WIV‚‚‚OQ bond-making.5 The intermediate I with a
discrete W-OQ bond and distorted octahedral stereochemistry
is next formed, followed by transition state T2 in which the O-Q
bond is largely weakened and the multiply bonded WdO
interaction develops. The reaction is completed by separation
of reduced substrate Q from the complex, which adopts the
stereochemistry of [WO(OPh)(S2C2Me2)2]1-
.
(34) Schultz, B. E.; Holm, R. H. Inorg. Chem. 1993, 32, 4244-4248.
(35) Thomson, L. M.; Hall, M. B. J. Am. Chem. Soc. 2001, 123, 3995-4002.
(36) Adams, B.; Smith, A. T.; Bailey, S.; McEwan, A. G.; Bray, R. C.
Biochemistry 1999, 38, 8501-8511.
(37) Canne, C.; Lowe, D. J.; Fetzner, S.; Adams, B.; Smith, A. T.; Kappl, R.;
Bray, R. C.; Hu¨ttermann, J. Biochemistry 1999, 38, 14077-14087.
(38) Pollock, V. V.; Barber, M. J. Biochemistry 2001, 40, 1430-1440.
(39) Bray, R. C.; Adams, B.; Smith, A. T.; Richards, R. L.; Lowe, D. J.; Bailey,
S. Biochemistry 2001, 40, 9810-9820.
The substituent dependence is expressed in the activation free
energy to form transition state T1. EWGs would render the WIV
center relatively more electrophilic than EDGs, thereby promot-
ing substrate binding and lowering ∆Gq1 (Figure 9). In the limit
of an infinitely strong EWG, I and T2 are suppressed, and the
reaction coordinate reduces to a single transition state. In this
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4319

A R T I C L E S
Sung and Holm
stration that turnover of R. capsulatus W-DMSOR is much
slower than that of the native Mo-DMSOR. In this case the
MVIO f MIV step is postulated as rate-determining because of
the lower redox potentials of the tungstoenzyme.
Summary. The following are the principal results and
conclusions of this investigation.
1. A series of 25 square pyramidal desoxo bis(dithiolene)
complexes of the type [WIV(OC6H4-p-X′)(S2C2(C6H4-p-X)2)2]1-
have been prepared in which equatorial (X ) Br, F, H, Me,
OMe) and axial (X′ ) CN, Br, H, Me NH2) substituents are
independently varied. The complexes provide dissimilarity in
electron density at the metal center at essentially constant
structure.
2. The complexes [Ni(S2C2(C6H4-p-X)2)2]0,1-,2- and [W(CO)2-
(S2C2(C6H4-p-X)2)2]0,1-,2- constitute reversible electron-transfer
series, neutral members of which are synthetic precursors of
the complexes in (1). Substituent dependence of redox potentials
is demonstrated by a linear free energy relationship between
E1/2 and the Hammett constant σp, with potentials decreasing
in the order X ) OMe < Me < H < F < Br.
transition state is early and has substantial WIV-OQ bond
making character (Figure 9). A six-coordinate intermediate
containing bound QO is formed thereafter followed by a second
transition state in which most or all of the atom and concomitant
electron transfer occurs. From the LFER in (3), ∆GqEWG
∆GqEDG and ∆Gq ≈ ∆Gq . ∆Gq .
<
1
2
5. Reaction system 6 and others preceding it4-6 involving
structurally and electronically realistic representations of the
active sites in the DMSOR enzyme family, inclusive of
isoenzymes, are functional models of enzyme action.
This work is directed toward a description of the rate-limiting
transition state in reaction 6. It is the most detailed investigation
yet of the mechanism of oxo transfer mediated by a biological
metal. We consider the reaction coordinate in Figure 9 to be
intrinsic to the complexes and substrate. The extent to which it
might be modified by protein structure and environment remains
to be learned. Equally important is the influence of substrate
on reaction mechanism. A beginning has been made in our
previous investigation, where we note that rate constants do not
uniformly correlate with the bond dissociation energy order P-O
> As-O > S-O > N-O.5 This behavior also suggests that
the rate-determining step does not involve significant Q-O bond
breaking.
3. Oxo transfer reaction 6 has been examined for the 25
complexes in (1) in acetonitrile. Reactions are first order in WIV
and in the constant substrate (CH2)4SO, and show large negative
activation entropies (-34 to -40 eu) consistent with an
associative transition state. Rate constants k vary linearly under
the Hammett equation log(kX/kH) ) Fσp with a larger influence
of the axial substituent X′ than the more distant equatorial
substituent X (Fax/Feq ) 4.8). Equivalently, a LFER exists
Acknowledgment. This research was supported by NSF Grant
CHE 98-76457.
Supporting Information Available: Spectroscopic charac-
terization data for 20 complexes of the type [X4,X′] (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
between ∆Gq and σp(X′) or 4σp(X). The rate constant order
298
kEWG > kEDG is general and synergistic effects are observed, as
indicated by, e.g., k[Br ,CN]/k[(OMe) ,CN] ≈ 390.
4
4
4. The results in (2) and (3) lead to the proposal of a stepwise
oxo transfer reaction pathway in which the rate-limiting
JA012735P
9
4320 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002