Identification of the oxidized products formed upon reaction of Chromium(V) with thymidine nucleotides

Article abstract of DOI:10.1021/ja962428x

Two pathways have been observed for Cr(V)-mediated nucleotide oxidation in reactions of bis(2-ethyl-2-hydroxybutyrato)oxochromate(V) [CrO(ehba)₂]^- with thymidine nucleotides dTMP, dTDP, and dTTP. The extent of Cr(V)-induced nucleotide oxidation was greatest for thymidine diphosphate (dTDP), as measured by the production of thiobarbituric acid reactive species (TBARS; indicative of pathway 1) and thymine release (indicative of pathway 2). The nucleoside thymidine showed no reaction, suggesting a phosphate-dependent oxidation. Amounts of TBARS and thymine were maximal at a pH range of 6.0-6.5, and both TBARS formation and thymine release correlated with decay of the Cr(V) EPR signal. Formation of TBARS was maximal in 100% O₂ but decreased markedly under argon, whereas thymine release was maximal under argon, but remained the major product observed under aerobic conditions. Pathway 1 for the reaction of Cr(V) with dTDP led to formation of glycolic acid and trans-thymine propenal at approximately equimolar amounts, consistent with a mechanism involving oxygen-dependent sugar oxidation following hydrogen atom abstraction at the C-4' carbon of the deoxyribose sugar. Pathway 2 led to release of free thymine, but much less (barely detectable) 2-deoxy-D-pentitol was formed from postreduction of the reactive aldehydic sugar fragment. Thus, the oxygen-independent release of thymine does not appear to result from reaction at the C-4' hydrogen unless decomposition of the aldehydic intermediate occurred. Determination of the oxidation state of chromium responsible for the observed oxidative damage was carried out using Mn(II), a Cr(IV)-specific reductant. Mn(II) essentially abolished all activity for both TBARS formation (pathway 1) and thymine release (pathway 2). These results suggest that Cr(IV), formed upon disproportionation of Cr(V), oxidizes the nucleotide deoxyribose sugar moiety via a phosphate-bound intermediate. pathway 1 involves oxygen-dependent oxidation at the C-4' position; however, the mechanism for oxygen-independent thymine release (pathway 2) is still unclear.

Full text of DOI:10.1021/ja962428x

J. Am. Chem. Soc. 1996, 118, 10811-10818
10811
Identification of the Oxidized Products Formed upon Reaction
of Chromium(V) with Thymidine Nucleotides
Kent D. Sugden and Karen E. Wetterhahn*
Contribution from 6128 Burke Laboratory, Department of Chemistry, Dartmouth College,
HanoVer, New Hampshire 03755-3564
ReceiVed July 15, 1996X
Abstract: Two pathways have been observed for Cr(V)-mediated nucleotide oxidation in reactions of bis(2-ethyl-
2-hydroxybutyrato)oxochromate(V) [CrO(ehba)2]- with thymidine nucleotides dTMP, dTDP, and dTTP. The extent
of Cr(V)-induced nucleotide oxidation was greatest for thymidine diphosphate (dTDP), as measured by the production
of thiobarbituric acid reactive species (TBARS; indicative of pathway 1) and thymine release (indicative of pathway
2). The nucleoside thymidine showed no reaction, suggesting a phosphate-dependent oxidation. Amounts of TBARS
and thymine were maximal at a pH range of 6.0-6.5, and both TBARS formation and thymine release correlated
with decay of the Cr(V) EPR signal. Formation of TBARS was maximal in 100% O2 but decreased markedly under
argon, whereas thymine release was maximal under argon, but remained the major product observed under aerobic
conditions. Pathway 1 for the reaction of Cr(V) with dTDP led to formation of glycolic acid and trans-thymine
propenal at approximately equimolar amounts, consistent with a mechanism involving oxygen-dependent sugar
oxidation following hydrogen atom abstraction at the C-4′ carbon of the deoxyribose sugar. Pathway 2 led to release
of free thymine, but much less (barely detectable) 2-deoxy-D-pentitol was formed from postreduction of the reactive
aldehydic sugar fragment. Thus, the oxygen-independent release of thymine does not appear to result from reaction
at the C-4′ hydrogen unless decomposition of the aldehydic intermediate occurred. Determination of the oxidation
state of chromium responsible for the observed oxidative damage was carried out using Mn(II), a Cr(IV)-specific
reductant. Mn(II) essentially abolished all activity for both TBARS formation (pathway 1) and thymine release
(pathway 2). These results suggest that Cr(IV), formed upon disproportionation of Cr(V), oxidizes the nucleotide
deoxyribose sugar moiety Via a phosphate-bound intermediate. Pathway 1 involves oxygen-dependent oxidation at
the C-4′ position; however, the mechanism for oxygen-independent thymine release (pathway 2) is still unclear.
Introduction
lesions induced by oxidative stress (DNA strand breaks, 8-oxo-
2′-deoxyguanosine, abasic sites) and chromium directly meta-
lated to the DNA strands (Cr-DNA adducts, both inter- and
intrastrand cross-links, and Cr-DNA-protein cross-links).9-12
At the genetic level, chromium has been shown to affect the
expression of some inducible genes and to interact with oxidative
stress inducible transcription factors in the signal transduction
pathway.13,14 In general, the chromium oxidation state which
is responsible for producing a particular DNA lesion and the
mechanism by which the lesion is produced are currently
unknown. However, there is a general belief that the higher
valent states of chromium, acting either directly or through
reactive oxygen species produced from molecular oxygen or
hydrogen peroxide, are responsible for much of the oxidative
stress and DNA damage observed in biological systems.
Bis(2-ethyl-2-hydroxybutyrato)oxochromate(V) [CrO(ehba)2]-
has properties that make it a favorable Cr(V) model for substrate
oxidation studies. The pentacoordinate complex with oxygen
ligands resembles the postulated ligand environment, and has a
The carcinogenicity of chromium(VI), is well established but
this oxidation state is not considered to be responsible for the
DNA lesions associated with the observed carcinogenic activity.
Instead, the role of the Cr(VI) is in membrane transport where
its tetrahedral conformation is isostructural with phosphate and
sulfate and allows active transport into the cell Via nonspecific
anion transport channels. Once inside the cell, intracellular
reduction occurs and generates the “ultimate” carcinogenic
chromium species.1 Chromium(V) is considered to be a
candidate for one of these ultimate carcinogenic species since
it has been observed both in ViVo and in Vitro during the
biological reduction of Cr(VI) to the final intracellular stable
oxidation state Cr(III).2-4 This biological reduction of Cr(VI)
to lower oxidation states has been observed with a wide variety
of naturally occurring cellular reductants including ascorbate,
glutathione, cysteine, NADH, and cytochrome P450 reductase.5-8
A consequence of this intracellular reduction is the production
of a wide array of adverse biological effects including DNA
(8) O’Brien, P.; Barrett, J.; Swanson, F. Inorg. Chim. Acta 1985, 108,
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, October 15, 1996.
(1) DeFlora, S.; Wetterhahn, K. E. Life Chem. Rep. 1989, 7, 169-244.
(2) Liebross, R. H.; Wetterhahn, K. E. Chem. Res. Toxicol. 1990, 3, 401-
403.
L19-L20.
(9) Stearns, D. M.; Kennedy, L. J.; Courtney, K. D.; Giangrande, P. H.;
Phieffer, L. S.; Wetterhahn, K. E. Biochemistry 1995, 34, 910-919.
(10) Borges, K. M.; Wetterhahn, K. E. Carcinogenesis 1989, 10, 2165-
2168.
(3) Sugiyama, M.; Tsuzuki, K.; Hidaka, T.; Ogura, R.; Yamamoto, M.
Biol. Trace Elem. Res. 1991, 30, 1-8.
(11) Casadevall, M.; Kortenkamp, A. Carcinogenesis 1994, 15, 407-
409.
(4) Jiang, J.; Liu, K. J.; Shi, X.; Swartz, H. M. Arch. Biochem. Biophys.
1995, 319, 570-573.
(12) Kortenkamp, A.; Oetken, G.; Beyersman, D. Mutat. Res. 1990, 232,
155-161.
(5) Stearns, D. M.; Wetterhahn, K. E. Chem. Res. Toxicol. 1994, 7, 219-
230.
(13) Hamilton, J. W.; Wetterhahn, K. E. Mol. Carcinog. 1989, 2, 274-
286.
(6) Jennette, K. W. J. Am. Chem. Soc. 1982, 104, 874-875.
(7) Goodgame, D. M. L.; Hayman, P. B.; Hathaway, D. E. Polyhedron
1982, 1, 497-499.
(14) Ye, J.; Zhang, X.; Young, H. A.; Shi, X. Carcinogenesis 1995, 16,
2401-2405.
(15) Krumpolc, M.; Rocek, J. J. Am. Chem. Soc. 1979, 101, 3206-3209.
S0002-7863(96)02428-6 CCC: $12.00 © 1996 American Chemical Society

10812 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Sugden and Wetterhahn
similar EPR g value, as the Cr(V)-ascorbate complex.5,15 The
R-hydroxy ligands of the model Cr(V) complex are not as prone
to intramolecular oxidation as the corresponding ascorbate
ligands, thus limiting the reactivity observed to the metal center.
Also, the decay of this Cr(V) complex has been well studied
and is known to proceed by disproportionation through a Cr(IV)
intermediate at neutral pH.16 At progressively more acidic pH
values, disproportionation is slower, with the Cr(V) complex
being quite stable at pH 3-4. Importantly, the Cr(V) complex
does not show formation of carbon-based or oxygen radicals;17
thus, any oxidative damage should be due to a direct metal-
substrate oxidative mechanism.
The mode of interaction of Cr(V) with DNA is currently
unknown, although Farrell et al. initially postulated that direct
binding of Cr(V) to DNA led to oxidation and DNA strand
breakage.18 Recently, the ability for [CrO(ehba)2]- to undergo
ligand exchange with free phosphate and pyrophosphate has
been shown,19 suggesting phosphate as a tethering point with
potential relevance to binding of Cr(V) to DNA and deoxy-
ribonucleotides. Upon reduction of the high valent chromi-
um(V) species to the more stable +3 state, the ensuing oxidation
of nucleotides can be envisioned. Previously, in situ generated
Cr(V)-glutathione and Cr(V)-ascorbate species had been
suggested to undergo an oxidative mechanism with DNA at the
sugar moiety.11,20,21 However, under the reaction conditions
employed, carbon-based radicals5 and thiyl radicals22 are formed
which could account for the observed reactivity and no product
analysis was performed to show formation of the specific sugar
and base degradation products expected.
In this paper, we show two oxidative pathways for the
reaction of a high valent chromium(V) species with thymidine
nucleotides by the identification and quantitation of specific
degradation products. We have found that Cr(V)-induced
oxidative damage of the thymidine series of nucleot(s)ides is
wholly dependent on the presence of a phosphate moiety within
the substrate. Of the thymidine nucleotides, dTDP was sig-
nificantly more prone to oxidation than either dTMP or dTTP,
while dT alone showed no reactivity. Identification of thiobar-
bituric acid reactive species (TBARS) arising from trans-
thymine propenal, and formation of the remaining 3-carbon sugar
unit, glycolic acid, showed that pathway 1 involved an oxygen-
dependent oxidation at the C-4′ hydrogen atom of the deoxy-
ribose sugar. In addition, the release of the free base thymine
in this reaction showed the occurrence of an oxygen-independent
mechanism (pathway 2). Finally, we demonstrated using Mn(II)
as a Cr(IV)-specific reductant that these two pathways involve
Cr(IV) as the reactive high valent species in these sugar
oxidation mechanisms and not Cr(V) itself.
N,O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimeth-
ylchlorosilane (TMCS) and dry acetonitrile as the silylation cosolvent
was purchased from the Pierce Chemical Co. Propargyl alcohol,
NaBH4, 2-ethyl-2-hydroxybutyric acid, 1,1,3,3-tetramethoxypropane,
and 2-thiobarbituric acid were obtained from Aldrich Chemical Co.
Dowex AG 50W-X8 cation exchange resin (H+ form), Dowex AG
50W-X2 cation exchange resin (H+ form), Dowex AG 1-X2 anion
exchange resin (Cl- form), and Chelex 100 were obtained from Bio-
Rad Co. Whatman silica gel TLC plates (both analytical and
preparative), trichloroacetic acid, MnCl2‚4H2O, ZnCl2, and HPLC grade
methanol were purchased from Fisher Scientific.
HPLC Conditions. HPLC separations and quantitations were
carried out on a HP-1090 HPLC system with diode array detection at
wavelengths of 254 nm for thymine, 275 nm for cis-thymine propenal,
and 305 nm for trans-thymine propenal. Analyses were performed on
a Rainen Microsorb-MV 100 Å C18 column, 3 µm particle size (4.6
mm i.d. × 10 cm length), as described23 and afforded the following
retention times and detection limits (in parentheses where applicable)
for the compounds of interest: thymine (1.65-1.70 min, 0.1 µM),
thymidine (1.75-1.80 min, 0.1 µM), dTMP (1.10-1.20 min), dTDP
(0.90-1.00 min), trans-thymine propenal (3.05-3.15 min, 0.1 µM),
cis-thyminepropenal (2.10-2.20 min).
EPR Conditions. EPR spectra were recorded at RT using a Bruker
ESP-300 spectrometer with spectral parameters of 100 kHz field
modulation, 1.0 G modulation amplitude, 5.12 ms time constant, 9.769-
9.773 GHz microwave frequency, 1 × 105 receiver gain, 2 mW
microwave power attenuated at 20 dB, 3380-3580 G sweep width,
and a 21 s scan time. All signals were averaged over nine scans.
Measurements were done on ca. 100 µL volume samples, drawn into
a capillary tube, and sealed on one end with Dow-Corning high vacuum
grease. The g values were determined with respect to 2,2-diphenyl-
1-picrylhydrazyl radical (DPPH), g ) 2.0036. Concentrations were
determined from a standard curve of [CrO(ehba)2]- in 100 mM aqueous
2-ethyl-2-hydroxybutyric acid.
GC/MS Conditions. GC/MS analyses were carried out on the
trimethylsilylated derivatives of the compounds by using a HP-5890
GC with a HP-5971 mass selective detector. Silylations were ac-
complished by addition of a 1:1 solution of 1% TMCS in BSTFA/
acetonitrile followed by heating at 60 °C for 1 h. GC conditions were
100 mL/min He flow rate with 2.5 mL/min split flow, injector
temperature of 250 °C, detector temperature of 312 °C, and a column
temperature gradient of 100-300 °C at 16 °C/min with a Supelco SPB-5
column (30 m × 0.2 mm).
Synthesis of Cr(V). The sodium salt of bis(2-ethyl-2-hydroxy-
butyrato)oxochromate(V) was prepared in a crystalline form using the
method of Krumpolc and Rocek.15 IR and UV-vis analysis yielded
identical results as that shown previously for this complex. Purity was
determined to be >98% by oxidation to Cr(VI) using alkaline H2O2
and analysis by UV-vis using ꢀ ) 4830 cm-1 M-1 24
Caution! Cr(VI)
.
is a known human carcinogen, and Cr(V) complexes are potentially
carcinogenic. Appropriate precautions should be taken in handling
these materials.
Synthesis of cis- and trans-Thymine Propenals. Preparation of
cis- and trans-thymine propenals was carried out by the method of
Ajmera et al.23 by reacting bis(trimethylsilyl)thymine25 in a neat solution
of propargylaldehyde.26 The product of interest, trans-thymine pro-
penal, was purified by preparative TLC,27 and was determined by HPLC
to be ∼95% pure. The contaminants were a small amount of free
thymine and the corresponding cis isomer. UV-vis spectra of trans-
thymine propenal demonstrated a wavelength maximum of 306 nm and
Experimental Section
Materials. Thymine, thymidine, dTMP, dTDP, dTTP, MgCl2‚6H2O,
glycolic acid, 2-deoxy-D-ribose, Na2Cr2O7‚2H2O, alkaline phosphatase
(type VII-L from bovine intestinal mucosa), and Sephadex DEAE A-25
anion exchange resin were purchased from the Sigma Chemical Co.
The deoxyribonucleotides, as the sodium salts, were used as supplied.
a calculated extinction coefficient of 11 270 M-1 cm-1 1H-NMR (300
.
MHz, D2O, δ (HDO) ) 4.67 ppm) δ 9.63 (1H, d, J ) 7.8 Hz), δ 8.09
(1H, d, J ) 14.4 Hz), δ 7.67 (1H, s), δ 6.25 (1H, q, J ) 6.6 and 14.7
Hz), δ 1.82 (3H, s).
(16) Krumpolc, M.; Rocek, J. Inorg. Chem. 1985, 24, 617-621.
(17) Sugden, K. D.; Wetterhahn, K. E. Inorg. Chem. 1996 35, 651-
657.
(18) Farrell, R. P.; Judd, R. J.; Lay, P. A.; Dixon, N. E.; Baker, R. S.
U.; Bonin, A. M. Chem. Res. Toxicol. 1989, 2, 227-229.
(19) Sugden, K. D.; Wetterhahn, K. E. Inorg. Chem. 1996, 35, 3727-
3728.
(20) Casadevall, M.; Kortenkamp, A. Carcinogenesis 1995, 16, 805-
809.
(21) da Cruz Fresco, P.; Shacker, F.; Kortenkamp, A. Chem. Res. Toxicol.
1995, 8, 884-890.
(22) Aiyar, J.; Berkovits, H. J.; Floyd, R. A.; Wetterhahn, K. E. Chem.
Res. Toxicol. 1990, 3, 595-603.
(23) Ajmera, S.; Wu, J. C.; Worth, Jr, L.; Rabow, L. E.; Stubbe, J.;
Kozarich, J. W. Biochemistry 1986, 25, 6586-6592.
(24) Haupt, G. W. J. Res. Natl. Bur. Stand. 1952, 48, 414-423.
(25) Kotick, M. P.; Szantay, C.; Bardos, T. J. J. Org. Chem. 1969, 34,
3806-3813.
(26) Sauer, J. C. In Organic Syntheses; Rabjohn, N., Ed.; John Wiley &
Sons: New York, 1963; Collect. Vol. IV, pp 813-815.
(27) Giloni, L.; Takeshita, M.; Johnson, F.; Iden, C.; Grollman, A. P. J.
Biol. Chem. 1981, 256, 8608-8615.

Reaction of Cr(V) with Thymidine Nucleotides
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10813
Synthesis of 2-Deoxy-D-erythro-pentitol. The synthesis of 2-deoxy-
D-erythro-pentitol was accomplished using the method of Tymiak and
Rinehart,28 as modified by Rabow et al.29 Derivatization of the oil
with 1% TMCS in BSTFA and subsequent GC/MS afforded a mass
spectrum identical with that reported by Rabow et al.,29 m/z 321, 307,
231, 219, 205, and the 103 base peak. 1H-NMR (300 MHz, D2O,
3-(trimethylsilyl)propionic-2,2,3,3-d4 acid internal standard) δ 3.74 (4
H, m), δ 3.60 (2 H, m), δ 1.86 (1 H, dtd, J ) 14.4, 7.5, 3 Hz), δ 1.65
(1 H, tdd, J ) 12.3, 5.3, 6.6 Hz). 13C NMR (75.4 MHz, D2O,
3-(trimethylsilyl)propionic-2,2,3,3-d4 acid internal standard) δ 74.9 (C-
4), 69.1 (C-3), 62.6 (C-5), 58.7 (C-1), 34.3 (C-2). The carbon
numbering scheme is in reference to the sugar group on a nucleotide
and does not conform to IUPAC conventions for this complex.
Conditions for Reaction of Cr(V) with Thymidine and Thymidine
Deoxyribonucleotides. (A) Identification and Quantification of
TBARS, Base Release, and trans-Thymine Propenals. Typical reac-
tions for determination of TBARS, base release and trans-thymine
propenals were carried out simultaneously using 0.8 mL reaction
volumes containing 10 mM nucleotide or nucleoside and 0.618 mM
Cr(V) (from a 50 mM stock solution of Na[CrO(ehba)2] in H2O). The
pH was maintained at 6.0 unless otherwise specified with 100 mM
NaOAc buffer rigorously demetalated with Chelex-100 ion exchange
resin. The reactions were carried out under ambient oxygen pressure
at RT for 1 h, except for reactions aimed at determination of oxygen
dependence which were carried out under an atmosphere of 100%
oxygen or after rigorously degassing with argon (note: only the reaction
itself was conducted under the argon or oxygen atmosphere but not
the subsequent TBARS, thymine release, and trans-thymine propenal
analyses which were carried out in air).
Half of the reaction solution was assayed for TBARS using the
method of Greenwald et al.30 and quantified from the absorbance at
532 nm using a standard of 1,1,3,3-tetramethoxypropane (ꢀ ) 1.6 ×
105 M-1 cm-1).23 Thymine release and trans-thymine propenal forma-
tion were determined by passing the remaining 0.4 mL reaction volume
through a column (0.5 × 5 cm) of DEAE Sephadex A-25 anion
exchange resin followed by elution with 0.6 mL of H2O. Thymine
was quantified by HPLC at 254 nm Versus a standard curve of authentic
thymine, and trans-thymine propenal was quantified by HPLC at 305
nm using synthetic trans-thymine propenal as a standard.
Figure 1. (A) Comparison of the reactivity of the thymine-containing
nucleoside and nucleotides (10 mM) with Cr(V) (0.618 mM) at pH
6.0. Solid bars represent thymine release, and cross-hatched bars
represent TBARS formation. (B) pH dependence for the formation of
TBARS (open circles) and thymine release (closed circles) in the
reaction of 0.618 mM Cr(V) with 10 mM dTDP. Reactions were carried
out in 100 mM NaOAc for 1 h at RT in air. Data represent the mean
( SD (n ) 3).
(B) Identification and Quantification of Glycolic Acid and
2-Deoxy-D-erythro-pentitol. Determination of glycolic acid and
2-deoxy-D-erythro-pentitol were hampered by the lack of a chromophore
thus entailing the use of larger volume reactions and subsequent analysis
by GC/MS. Glycolic acid formed in the reaction between dTDP and
Cr(V) was determined essentially by the method of McGall et al.31
and 2-deoxy-D-erythro-pentitol determined by the method of Rabow
et al.29 The glycolic acid and 2-deoxy-D-erythro-pentitol were deriva-
tized with TMCS in BSTFA as described above and analyzed by GC/
MS. Quantitations were achieved using authentic glycolic acid and
synthetic 2-deoxy-D-erythro-pentitol as standards. Essentially quantita-
tive recovery was achieved using these methods as demonstrated with
reactions using these standards. The detection limit was ca. 1.0 µM
for both compounds.
(C) Trapping of Cr(IV) by Mn(II). The role of Cr(IV) in the
oxidative mechanism was determined using Mn(II) as a Cr(IV) trap
and Mg(II) as a cationic control under the conditions described above
for the formation of TBARS, thymine release, and trans-thymine-
propenal. Mg(II) or Mn(II) (final concentration of 1.25 mM) was added
to the deoxyribonucleotide solution prior to addition of Cr(V).
of the pyrimidine base and the known oxidation products derived
from these substrates’ reaction with ionizing radiation and
bleomycin.32,33 These studies have shown oxygen-dependent
formation of base propenal (a TBA reactive substance) as well
as oxygen-independent base release as a result of two separate
pathways for sugar degradation. The role of the phosphate
moiety in Cr(V)-mediated nucleotide oxidation was probed by
reacting Cr(V) with the dTMP, dTDP, and dTTP nucleotide
series, and the dT nucleoside at pH 6.0 under ambient oxygen
pressure for 1 h. Formation of thiobarbituric acid reactive
species (TBARS) as well as release of free thymine was
observed for all three nucleotides, while thymidine showed no
reaction (Figure 1A). TBARS formation was greatest for dTDP
(2.15 ( 0.08 µM) followed by dTMP (0.71 ( 0.05 µM) and
dTTP (0.70 ( 0.01 µM). An identical trend was observed for
thymine release: dTDP (20.95 ( 0.62 µM) > dTMP (7.5 (
0.10 µM) > dTTP (6.79 ( 0.76 µM).
Disproportionation of Cr(V) in aqueous neutral solutions,
3Cr(V) f 2Cr(VI) + Cr(III), yields a stoichiometry of 67%
Cr(VI) and 33% Cr(III).16 The presence of an oxidizable
substrate, such as a nucleotide, should shift the redox percent-
ages toward the more reduced form Cr(III), resulting in a
decrease in the final Cr(VI) concentration. Measurement of the
final Cr(VI) concentration formed during the reactions described
Results
Role of Phosphate in Cr(V) Reactions with Nucleotides.
Thymidine nucleotides were chosen as a model for Cr(V)-
mediated nucleotide oxidation because of the inherent stability
(28) Tymiak, A. A.; Rinehart, Jr, K. L. J. Am. Chem. Soc. 1983, 105,
7396-7401.
(29) Rabow, L. E.; Stubbe, J.; Kozarich, J. W. J. Am. Chem. Soc. 1990,
112, 3196-3203.
(32) Janicek, M. F.; Haseltine, W. A.; Henner, W. D. Nucleic Acids Res.
1985, 13, 9011-9029.
(33) (a) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107-1136.
(b) Kane, S. A.; Hecht, S. M. Progress in Nucleic Acid Research and
Molecular Biology; Academic Press Inc.: New York, 1994; Vol. 49, pp
313-352.
(30) Greenwald, R. A.; Rush, S. W.; Moak, S. A.; Weitz, Z. Free Rad.
Biol. Med. 1989, 6, 385-392.
(31) McGall, G. H.; Rabow, L. E.; Stubbe, J. J. Am. Chem. Soc. 1987,
109, 2836-2837.

10814 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Sugden and Wetterhahn
Table 1. Oxidation Products and Cr(VI) Loss Observed in the
Reaction between Cr(V) and dTDPa
oxidized product
glycolic acid, III
trans-thymine propenal, I
thymine
2-deoxy-D-pentitol, V
loss of Cr(VI)
concn (µM)
6.5 ( 0.7b
1.1 ( 0.1,b 5.6 ( 0.5c
20.95 ( 0.62b
<1.0
19.8 ( 4.6b
a All reactions were carried out at 100 mM NaOAc, pH 6.0, in air
at RT for 1 h with 10 mM dTDP and 0.618 mM Cr(V). b Mean ( SD
(n ) 3). c Corrected concentration for anion exchange column loss.
above showed that disproportionation was stoichiometric, yield-
ing 67% Cr(VI), for the control Cr(V) reaction as well as the
reaction with dT. A small yet significant decrease in Cr(VI)
concentration, between 3 and 4%, was observed in reactions of
Cr(V) in the presence of dTDP (Table 1) and dTTP (23.5 (
4.5 µM), indicating that substrate oxidation had occurred. This
number was considered to correspond roughly to the “oxidative
reactivity range” for Cr(V) with these nucleotides. However,
the low overall reaction with dTMP could not be detected using
this method. This “oxidative reactivity range” was consistent
with the total oxidized products measured by TBARS and
thymine release for the Cr(V) reaction with dTDP (Table 1).
The fact that dT showed no reaction in any of these assay
systems demonstrates the absolute requirement for phosphate
in the redox reaction between Cr(V) and nucleotides.
pH and Time Dependence of TBARS Formation and
Thymine Release. Maximum formation of both TBARS and
free thymine for the reaction of Cr(V) with dTDP was observed
at pH 6.0-6.5 when carried out at RT for 1 h under ambient
oxygen pressure (Figure 1B). However, some reaction could
be detected in the full pH range measured, 5.0-7.5. The range
of greatest reactivity was considered to be a consequence of
balancing the ligand exchange lability of Cr(V) leading to rapid
disproportionation at higher pH values with ligand exchange
stability and thus lower reactivity at more acidic pH values.
The 1 h reaction time used throughout this study corresponded
to complete decay of the Cr(V) EPR signal in the pH ranges of
7.5-6.5 at RT, and approximately 95% Cr(V) decay at pH 6.0
(Figure 2A). The Cr(V) EPR signal showed significantly greater
stability at pH < 6.0, which would account for the lower
reactivity observed under these conditions. The presence of
dTDP in the reaction solution with Cr(V) did not significantly
alter the normal decay pathway for Cr(V) as measured by EPR
(Figure 2A). The appearance of oxidative damage, as measured
by thymine release and TBARS formation (Figure 2B), was
inversely related to the disappearance of Cr(V) (Figure 2A).
Neither Cr(VI) itself nor an aged solution of Cr(V) (which had
been allowed to completely disproportionate to Cr(VI) and
Cr(III)) showed significant TBARS formation or thymine release
(data not shown). Reaction of Cr(V) with dTMP and dTTP at
pH 6.0 demonstrated similar conditions for maximal formation
of TBARS and thymine release (data not shown). These data
support the hypothesis that oxidative damage arises from either
Cr(V) itself or an in situ generated Cr(IV) species under
conditions which are within a physiologically relevant range.
Oxygen Dependence for TBARS Formation and Thymine
Release. The requirement for oxygen was determined for the
reaction of Cr(V) with dTDP by measuring the formation of
TBARS and thymine release under an oxygen, air, or argon
atmosphere. Under a pure oxygen atmosphere, TBARS forma-
tion was enhanced only slightly over air but was decreased by
42% in the absence of oxygen (Figure 3A). The opposite was
true for thymine release where production of free thymine was
maximal under argon, but decreased by 35% under oxygen with
Figure 2. (A) Loss of Cr(V) EPR signal over time for the dispropor-
tionation reaction of Cr(V) in the presence (open triangles) or absence
(closed triangles) of dTDP and time-dependent formation of trans-
thymine propenal (closed squares) in the reaction of 0.618 mM Cr(V)
with 10 mM dTDP in 100 mM NaOAc, pH 6.0. Data represent the
mean ( SD (n ) 3). (B) Time-dependent formation of TBARS (open
circles) and thymine release (closed circles) in the reaction of 0.618
mM Cr(V) with 10 mM dTDP in 100 mM NaOAc, pH 6.0, at RT.
air being intermediate in reactivity (Figure 3A). Since only the
direct 1 h reaction of Cr(V) with dTDP was carried out under
a pure argon or oxygen atmosphere, it is possible that the
subsequent addition of thiobarbituric acid to form the TBARS
chromophore (carried out in air) may have been measuring a
long-lived O2-dependent species. This long-lived O2-dependent
species may account for the residual TBARS which were
observed under anoxic conditions. The oxygen-dependent
formation of TBARS is a classical marker of C-4′ H-atom
abstraction and suggests an oxygen-dependent mechanism for
Cr(V)-induced oxidation of dTDP (pathway 1) outlined in
Scheme 1.34 However, the formation of free thymine in an
oxygen-independent fashion suggests a separate pathway of
sugar oxidation (pathway 2).
Pathway 1: Identification of Products from the Oxygen-
Dependent Pathway in the Cr(V)-dTDP Reaction. (A)
trans-Thymine Propenal. The specific species being measured
in the TBARS assay is considered to be trans-thymine propenal
(for this substrate), I, formed following abstraction of the C-4′
hydrogen of deoxyribose, addition of molecular oxygen to the
reactive C-4′ radical intermediate, and anti-elimination of the
2′-pro-R hydrogen of the oxidized sugar (Scheme 1).34 Con-
current with the formation of I would be a phosphor-
ylated glycolic acid species, II, in a 1:1 ratio. Specifically, this
(34) (a) Burger, R. M.; Projan, S. J.; Horwitz, S. B.; Peisach, J. J. Biol.
Chem. 1986, 261, 15955-15959. (b) McGall, G. H.; Rabow, L. E.; Ashley,
G. W.; Wu, S. H.; Kozarich, J. W.; Stubbe, J. J. Am. Chem. Soc. 1992,
114, 4958-4967.

Reaction of Cr(V) with Thymidine Nucleotides
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10815
incomplete release of I from phosphate was demonstrated in
the reaction of bleomycin with DNA.34a In control reactions,
authentic I was inert to decomposition by either chromium(V)
or chromium(VI) under the normal reaction conditions. How-
ever, a control reaction for TBARS production either with or
without passing through an anion exchange resin demonstrated
that only 19% of the total TBARS formed in the one hour
reaction had undergone hydrolysis resulting in release of the
phosphate. This demonstrated that the actual concentration of
I formed in the reaction of Cr(V) with dTDP was ∼5-fold higher
(5.6 ( 0.5 µM). Formation of I was also observed by HPLC
in reactions between Cr(V) and dTMP or dTTP at pH 6.0, but
at much lower levels as suggested by the ∼3-fold lower level
of TBARS observed.
(B) Glycolic Acid. The second sugar fragment formed in
pathway 1 (see Scheme 1) is expected to be the phosphorylated
glycolic acid species II. Treatment with alkaline phosphatase
releases glycolic acid, III, which is readily measured by GC/
MS upon derivatization to the bis(trimethylsilyl)glycolate spe-
cies. Formation of III in the reaction of Cr(V) with dTDP at
pH 6.0 was confirmed by GC/MS analysis (Table 1). The
glycolic acid formed in the reaction had an identical retention
time on the GC as authentic III, 4.6 min, as well as identical
molecular ion, M - CH3 of m/z ) 205, and identical fragmenta-
tion pattern. The concentration of III formed was consistent
with that obtained for fragment I (Table 1), as expected from
the 1:1 ratio predicted in Scheme 1. The fact that the amounts
of I and III are consistent with each other but higher than the
TBARS suggests that a reaction between chromium and TBARS
may occur and lead to an underestimation of the actual reaction.
Formation of III in the reaction between Cr(V) and dTMP and
dTTP was not attempted due to the low level of reaction for
these two substrates and the fact that levels of III were near
the detection limits for the much higher yielding Cr(V)-dTDP
reaction.
Figure 3. Oxygen dependence of (A) the formation of TBARS (cross-
hatched bars) and thymine release (solid bars) and (B) the formation
of trans-thymine propenal in the reaction of 0.618 mM Cr(V) with 10
mM dTDP in 100 mM NaOAc, pH 6.0, for 1 h at RT. Data represent
the mean ( SD (n ) 3).
would be a diphosphorylated species from the dTDP substrate,
but a monophosphorylated species if this reaction occurred in
DNA.
Formation of I was demonstrated in the reaction of Cr(V)
with dTDP at pH 6.0 using HPLC (Figure 4). The HPLC trace
showed that, in this reaction, a peak eluting at ∼3.1 min had
an identical retention time as synthetic trans-thymine propenal
(Figure 4). Diode array detection of the HPLC eluent showed
that the 3.1 min peak had an identical absorption spectra as
authentic I with a wavelength maximum at 306 nm. Collection
of the 3.1 min peak and subsequent thin layer chromatography
Versus authentic I demonstrated an identical Rf value (0.73),
and when sprayed with a 1% solution of 2-thiobarbituric acid
in 0.1 N NaOH developed the red-pink color associated with
TBARS. No cis isomer of I was detected in any of the reactions,
which is consistent with the mechanism of anti-elimination of
the 2′-pro-R hydrogen.34
Production of TBARS should directly correlate with formation
of I, since the malondialdehyde-like species measured at 532
nm is derived from acid hydrolysis of I.27 Formation of I in
the reaction of Cr(V) with dTDP (Figure 2A) followed the same
time dependence as the formation of TBARS (Figure 2B),
suggesting that I is the ultimate species measured in this assay.
Formation of I was also observed to follow the same trend as
TBARS formation for oxygen dependence (Figure 3B). How-
ever, when quantified versus a standard curve of authentic trans-
thymine propenal, the concentration of I was determined to be
1.1 ( 0.1 µM, much less than expected on the basis of the
concentration of TBARS. This result suggested either decom-
position of I was occurring during the reaction, or there was an
incomplete release of I from the phosphate moiety. Incomplete
release would result in binding to the anion exchange resin used
during workup of the product and thus an underestimation of
total I produced during the reaction. Precedence for this
Pathway 2: Identification of Products from an Oxygen-
Independent Pathway in the Cr(V)-dTDP Reaction. (A)
2-Deoxy-D-erythro-pentitol. A two-electron, O2-independent
pathway (Scheme 2)29 or a one-electron pathway where OH-
addition is favored over O2 addition could account for the release
of the free base thymine. The mechanism invoked involves
formation of a carbocation intermediate at the C-4′ position of
deoxyribose by a two electron Cr(IV)/Cr(II) redox mechanism,
followed by addition of OH- from water and concomitant
thymine release. In tandem with thymine release is the
production of the reactive 4′-keto-1′-aldehyde species IV
(Scheme 2).29 This reactive aldehydic species is not isolable
itself29 but instead is reduced by NaBH4 to the polyalcoholic
species 2-deoxy-D-pentitol, V, with both erythro and threo
isomers expected.29 With dTDP as the substrate, any V formed
would be the diphosphorylated species that upon dephos-
phorylation with alkaline phosphatase would give free V for
GC/MS analysis. Reaction of Cr(V) with dTDP at pH 6 resulted
in production of V that was below the detection limits afforded
by this assay (data not shown). A small, unquantifiable peak
eluting at a retention time of 10.6 min, identical to that for
authentic V, was observed but the concentration was less than
lowest standard (<1.0 µM, Table 1). It cannot be ruled out
that the reactive aldehydic intermediate IV was formed but
decomposed in the presence of oxidizing chromium species in
solution, although the reduced species V was observed to be
stable to the presence of both Cr(V) and Cr(VI). Repeating
the experiment under anaerobic conditions to maximize forma-
tion of V (suggested by the anaerobic results for thymine release,
Figure 3A) resulted in no detectable V. This result strongly

10816 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Sugden and Wetterhahn
Scheme 1
2Cr(V) f Cr(IV) + Cr(VI)
Cr(IV) + Mn(II) f Cr(III) + Mn(III)
(1)
(2)
as a competitor for Cr(IV) can thus determine what role this
oxidation state of chromium may play in nucleotide oxidations.
At a 2-fold molar excess of Mn(II) to Cr(V) in the reaction
with dTDP at pH 6.0, thymine release was completely inhibited
and TBARS formation was reduced to essentially background
levels (Figure 5). A control reaction using a nonredox-active
divalent cation, Mg(II), under identical conditions showed
essentially no inhibition of either thymine release or TBARS
formation (Figure 5). Formation of trans-thymine propenal, I,
was detected in both control and Mg(II) reactions but not in
the Mn(II) reactions (data not shown). The data clearly
demonstrate that Cr(IV) is the species responsible for oxidizing
the nucleotides in both pathways 1 and 2 (Schemes 1 and 2).
Figure 4. HPLC chromatogram of the products formed from the
reaction of Cr(V) (0.618 mM) and dTDP (10 mM) in 100 mM NaOAc,
pH 6.0, at ambient oxygen pressure and RT for 1 h. Trace A: the
reaction monitored at 305 nm for detection of trans-thymine propenal
at 3.1 min, also showing thymine, T, at 1.6 min. Trace B shows the
elution profile of authentic trans-thymine propenal, synthesized as
described in the Experimental Section.
Discussion
Role of Phosphate in Cr(V)-Mediated Nucleotide Oxida-
tions. Previously, we have shown that Cr(V) can bind to both
phosphate and pyrophosphate, with pyrophosphate being the
more kinetically favorable.19 Since dTDP can be considered a
pyrophosphate analog and dTMP a phosphate analog, the fact
that dTDP shows a greater oxidation reaction with Cr(V) than
dTMP is consistent with a kinetic basis for oxidation of
nucleotides. The reaction of Cr(V) with dTTP would be
considered to be the most reactive if the trend of increasing
phosphates giving increasing oxidation held true. This trend
was observed in the loss of Cr(VI) concentration but not in either
of the two biomarkers used in this study, thymine release and
TBARS. This can be explained on a steric basis where a
bidentate binding mode to the γ- and â-phosphates of dTTP
would give a greater degree of flexibility to the putative Cr(IV)-
nucleotide complex in contrast to binding at the R- and
suggests that production of V, and the mechanism that it
represents, was a minor pathway and reaction at different sugar
sites such as C-3′, C-5′, or C-1′ must be considered to explain
the excess production of free thymine.
Role of the +4 Chromium Oxidation State in Nucleotide
Oxidations. The role that the +4 oxidation state of chromium
may play in the oxidation of nucleotides was investigated using
a Cr(IV)-specific reductant, Mn(II).35 Cr(IV) produced from
the initial disproportionation of Cr(V), eq 1, can react with
Mn(II) to produce Cr(III) and Mn(III), eq 2. The use of Mn(II)
(35) Beattie, J. K.; Haight, G. P., Jr. Prog. Inorg. Chem. 1972,17, 93-
145.

Reaction of Cr(V) with Thymidine Nucleotides
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10817
Scheme 2
TBARS and thymine release were within physiologically
significant pH ranges, ionic strengths, and concentrations. The
greatest reactivity appeared to be in a range which best balances
stability with disproportionation. Disproportionation was neces-
sary for reactivity, as is shown by the Mn(II) studies and by
the low reactivity observed at acidic pH values, although at high
pH values disproportionation appears to be favored over
substrate oxidation, leading to TBARS and thymine release.
The reaction of Cr(V) with nucleotides demonstrates an
oxygen-dependent mechanism (pathway 1, Scheme 1) similar
to that observed with other DNA damaging agents such as
bleomycin, neocarzinostatin, and ionizing radiation.33,39 The
requirement of oxygen for production of both single strand
breaks (SSBs) in DNA and base release has been demonstrated
for in situ generated Cr(V)-glutathione and Cr(V)-ascorbate
complexes.20,21 This oxygen dependence has been ascribed to
a Cr(IV)- or Cr(V)-peroxo species formed from the reaction
with molecular oxygen during the reduction process. In this
study there was an oxygen dependence in TBARS formation,
which would be manifested as SSBs in DNA, but increased
oxidation as measured by thymine release was observed under
anoxic conditions between Cr(V) and dTDP which is not in
agreement with that seen previously.20,21 Instead, this data is
in agreement with our previous results showing that Cr(V) can
directly oxidize EPR spin traps without the presence of oxygen17
and that invoking a Cr(IV)- or Cr(V)-peroxo species with
molecular oxygen is not necessary for the [CrO(ehba)2]- species.
Identification of Oxidized Products from the Cr(V)-
dTDP Reaction. Formation of base propenals in reactions with
DNA or nucleotides has been shown to be oxygen-dependent,
occurring through a one-electron, hydrogen atom abstraction
mechanism at the C-4′ of deoxyribose.34 In the reaction of
Cr(V) with dTDP we have demonstrated the unequivocal
formation of trans-thymine propenals. As well, we have identi-
fied the remaining sugar fragment from the C-4′ H-atom
abstraction, glycolic acid. To our knowledge, this is the first
time that products from a known oxidative mechanism have
been isolated and characterized from the reaction of high valent
chromium complexes with nucleotides. Extrapolation of this
mechanism to DNA suggests that strand breakage should be
Figure 5. Effect of the Cr(IV)-specific reductant Mn(II) or the divalent
cation control Mg(II) (1.2 mM) on the release of thymine (solid bars)
and TBARS formation (cross-hatched bars) in the reaction of 0.618
mM Cr(V) with 10 mM dTDP in 100 mM NaOAc, pH 6.0, for 1 h at
RT. Data represent the mean ( SD (n ) 3). N.D. ) not detectable.
â-phosphates as in dTDP. The end result of this greater
flexibility would be increased oxidation at other sugar sites such
as C-3′ or C-5′, yielding a higher overall oxidation as measured
in Cr(VI) concentration loss but through the formation of
oxidized products not measured in this study. The binding of
the +3 oxidation state of chromium to nucleotides through the
phosphate is well known,36 but the ability of Cr(IV) to bind to
phosphate has not been demonstrated, due to its transient nature.
The importance of phosphate in chromium reactions as a
potential binding site has been recently shown for the formation
of ternary Cr(III)-amino acid complexes with DNA.37 As well,
reduction reactions of Cr(VI) and ascorbate leading to single
strand breaks in DNA have been observed to occur primarily
in phosphate buffer.38 These results suggest a mechanism where
a reactive Cr(IV) species transiently binds to a terminal or
internucleotide phosphate allowing favorable steric interactions
for sugar oxidation to occur.
Conditions for Maximum Formation of TBARS and
Thymine Release. The conditions for maximal formation of
(36) DePamphilis, M. L.; Cleland, W. W. Biochemistry 1973, 12, 3714-
3719.
(37) Zhitkovich, A.; Voitkun, V.; Costa, M. Biochemistry 1996, 35,
7275-7282.
(38) da Cruz Fresco, P.; Kortenkamp, A. Carcinogenesis 1994, 15, 1773-
(39) Pratviel, G.; Bernadou, J.; Meunier, B. Angew. Chem., Int. Ed. Engl.
1995, 34, 746-769.
1778.

10818 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Sugden and Wetterhahn
observed. Indeed, strand breaks have been shown for this Cr(V)
complex by supercoiled plasmid DNA relaxation assays.18
in the reaction showed that oxidation of thymidine nucleotides
is dependent on the formation of Cr(IV). Neither Cr(VI) nor
Cr(III) showed significant activity in any of the reactions.
Previously, the use of V(IV) as a Cr(IV)-specific reductant has
been observed to inhibit strand breakage by the [CrVO(ehba)2]-
complex, as measured by a supercoiled plasmid DNA relaxation
assay.43 However, this oxidation appears to be ligand specific
since certain macrocyclic tetraamide Cr(V) complexes appear
to relax supercoiled plasmid DNA without undergoing dispro-
portionation/reduction to Cr(IV).44 While it is still too early to
make general conclusions on the ultimate oxidation state of
chromium responsible for intracellular DNA oxidation, the +4
oxidation state must be considered a leading candidate.
Conclusions. The results reported here have established that
oxidation of nucleotides by high valent chromium complexes
can occur at the sugar moiety. This oxidation is dependent upon
the presence of a phosphate presumably as a tethering point
for the chromium complex. An oxygen-dependent mechanism
(pathway 1) which involves the abstraction of the C-4′ hydrogen
atom has been demonstrated by isolating and identifying the
two major products which are known to be associated with this
The release of the free base thymine was also observed and
initially ascribed as the product of a two-electron oxidation at
the C-4′ of deoxyribose (pathway 2, Scheme 2). However, this
putative two-electron oxidation pathway was not verified since
the sugar fragment expected in this oxidation, 2-deoxy-D-
pentitol, was not observed in an equimolar amount with free
thymine. A wide variety of other possible oxidation mecha-
nisms could explain the thymine release, including C-1′, C-3′,
and C-5′ oxidations. An alternative mechanism which would
lead to thymine propenals, glycolic acid, and malondialdehyde
formation but not to the 4′-keto-1′-aldehyde sugar fragment by
hydrogen atom abstraction at the C-4′ of deoxyribose has been
observed for 3′-O-acylated nucleosides.40 This mechanism
appears to be unlikely for the reaction of chromium with
nucleotides since the nucleotides lack the 3′-acyl protecting
group necessary for migration and polarization of the hydro-
peroxide intermediate formed when molecular oxygen reacts
with the C-4′ radical of deoxyribose.40 As well, equal amounts
of malondialdehyde and free thymine should be detected, which
was not observed.
mechanism, trans-thymine propenal and glycolic acid.
A
second, as yet unidentified, mechanism is also occurring to
account for the release of free thymine. As well, the oxidation
state of chromium which appears to be responsible for this
oxidation has been determined to be the Cr(IV) species which
is formed during disproportionation of the parent Cr(V) complex.
These results are the first in which a specific oxidation state of
chromium has been assigned a definitive oxidative mechanism
for a biologically relevant substrate unambiguously identified
by product analysis. The mechanism, when coupled with
phosphate dependence, has implications for oxidative DNA
damage as well as disruption of intracellular nucleotide pools.
This study shows that caution in assigning mechanisms must
be used until both expected fragments from the oxidation
reaction are identified. Due to the possibility of sugar oxidation
intermediates reacting with high valent chromium species in
solution, further studies with specifically tritiated nucleotides
are needed to unequivocally establish the point of C-H bond
cleavage leading to release of the free thymine.
Role of the +4 Oxidation State in Nucleotide Oxidations.
During disproportionation of Cr(V), Cr(IV) is known to be
formed and has been shown to be an inherently more oxidative
species.41 As well, Cr(IV) in perchloric acid solutions has been
shown to readily oxidize a variety of organic complexes.42
However, unlike Cr(V), Cr(IV) is refractory to spectroscopic
observations in aqueous neutral solutions. The use of Mn(II)
as a Cr(IV) trap can be used to infer the involvement of Cr(IV)
in a mechanism since it does not appear to interact with the
other potential oxidizing complex Cr(V).5 The use of Mn(II)
Acknowledgment. This research was funded by PHS Grant
ES07167 from the National Institute of Environmental Health
Sciences, DHHS (K.E.W.). The EPR spectrometer was pur-
chased with NSF Grant CHE-8701406. K.D.S. was supported
by a NRSA postdoctoral fellowship (Grant CA60434) from the
National Cancer Institute, DHHS.
JA962428X
(40) McGall, G. H.; Stubbe, J.; Kozarich, J. W. J. Org. Chem. 1991, 56,
48-55.
(41) Gould, E. S. Coord. Chem. ReV. 1994, 135/136, 651-684.
(42) (a) Scott, S. L.; Bakac, A.; Espenson, J. H. J. Am. Chem. Soc. 1991,
113, 7787-7788. (b) Al-Ajlouni, A.; Bakac, A.; Espenson, J. H. Inorg.
Chem. 1994, 33, 1011-1014.
(43) Barr-David, G.; Hambley, T. W.; Irwin, J. A.; Judd, R. J.; Lay, P.
A.; Martin, B. D.; Bramley, R.; Dixon, N. E.; Hendry, P.; Ji, J-Y.; Baker,
R. S. U.; Bonin, A. M. Inorg. Chem. 1992, 31, 4906-4908.
(44) Dillon, C. T.; Lay, P. A.; Bonin, A. M.; Dixon, N. E.; Collins, T.
J.; Kostka, K. L. Carcinogenesis 1993, 14, 1875-1880.