Chromium(V) complexes of hydroxamic acids: Formation, structures, and reactivities

Article abstract of DOI:10.1021/ic048317d

A new family of relatively stable Cr(V) complexes, [Cr^VO(L) ₂]^- (LH₂ = RC(O)NHOH, R = Me, Ph, 2-HO-Ph, or HONHC(O)(CH₂)₆), has been obtained by the reactions of hydroxamic acids with Cr(VI) in polar aprotic solvents. Similar reactions in aqueous solutions led to the formation of transient Cr(V) species. All complexes have been characterized by electron paramagnetic resonance spectroscopy and electrospray mass spectrometry. A Cr(V) complex of benzohydroxamic acid (1, R = Ph) was isolated in a pure form (as a K⁺ salt) and was characterized by X-ray absorption spectroscopy and analytical techniques. Multiple-scattering analysis of X-ray absorption fine structure spectroscopic data for 1 (solid, 10 K) point to a distorted trigonal-bipyramidal structure with trans-oriented Ph groups and Cr-ligand bond lengths of 1.58 A (Cr-O), 1.88 A (Cr-O(C)), and 1.98 A (Cr-O(N)). Under ambient conditions, 1 is stable for days in aprotic solvents but decomposes within minutes in aqueous solutions (maximal stability at pH ～ 7), which leads predominantly to the formation of Cr(III) complexes. Complex 1 readily undergoes ligand-exchange reactions with biological 1,2-diols, including D-glucose and mucin, in neutral aqueous solutions. It differs from most other types of Cr(V) complexes in its biological activity, since no oxidative cleavage of plasmid DNA in vitro and no significant bacterial mutagenicity (in the TA 102 strain of Salmonella typhimurium) was observed for 1. In natural systems, stabilization of Cr(V) by hydroxamato ligands from bacterial-derived siderophores (followed by ligand-exchange reactions with more abundant carbohydrate ligands) may occur during the biological reduction of Cr(VI) in contaminated soils.

Full text of DOI:10.1021/ic048317d

Inorg. Chem. 2005, 44, 2934−2943
Chromium(V) Complexes of Hydroxamic Acids: Formation, Structures,
and Reactivities
Swetlana Gez, Robert Luxenhofer,† Aviva Levina, Rachel Codd,* and Peter A. Lay*
Centre for HeaVy Metals Research and Centre for Structural Biology and Structural Chemistry,
School of Chemistry, UniVersity of Sydney, NSW, 2006, Australia
Received November 29, 2004
A new family of relatively stable Cr(V) complexes, [CrVO(L)2]- (LH2
) RC(O)NHOH, R ) Me, Ph, 2-HO−Ph, or
HONHC(O)(CH2)6), has been obtained by the reactions of hydroxamic acids with Cr(VI) in polar aprotic solvents.
Similar reactions in aqueous solutions led to the formation of transient Cr(V) species. All complexes have been
characterized by electron paramagnetic resonance spectroscopy and electrospray mass spectrometry. A Cr(V)
complex of benzohydroxamic acid (1, R
)
Ph) was isolated in a pure form (as a K+ salt) and was characterized
by X-ray absorption spectroscopy and analytical techniques. Multiple-scattering analysis of X-ray absorption fine
structure spectroscopic data for 1 (solid, 10 K) point to a distorted trigonal-bipyramidal structure with trans-oriented
Ph groups and Cr
conditions, 1 is stable for days in aprotic solvents but decomposes within minutes in aqueous solutions (maximal
stability at pH 7), which leads predominantly to the formation of Cr(III) complexes. Complex 1 readily undergoes
ligand-exchange reactions with biological 1,2-diols, including -glucose and mucin, in neutral aqueous solutions. It
−ligand bond lengths of 1.58 Å (Cr−O), 1.88 Å (Cr−O(C)), and 1.98 Å (Cr−O(N)). Under ambient
∼
D
differs from most other types of Cr(V) complexes in its biological activity, since no oxidative cleavage of plasmid
DNA in vitro and no significant bacterial mutagenicity (in the TA 102 strain of Salmonella typhimurium) was observed
for 1. In natural systems, stabilization of Cr(V) by hydroxamato ligands from bacterial-derived siderophores (followed
by ligand-exchange reactions with more abundant carbohydrate ligands) may occur during the biological reduction
of Cr(VI) in contaminated soils.
Scheme 1. Common Modes of Binding of Hydroxamic Acids to
Introduction
a
Transition Metal Ions (Mn+
)
Hydroxamic acid (RC(O)NHOH) moieties are widespread
in biological systems (particularly in the microbial and plant
kingdoms) as key functional groups of siderophores, which
are relatively small biomolecules that are responsible for
high-affinity scavenging of Fe(III) (and possibly of other
transition metal ions) from the environment.1 Strong binding
of metal ions provides the basis for numerous applications
of hydroxamic acids as pharmacological agents (inhibitors
of metalloenzymes).2 Metal binding to hydroxamic acids
(1:1) usually occurs in a bidentate fashion (Scheme 1) with
the formation of singly or doubly deprotonated (hydroxamato
or hydroximato) ligands.3,4 While coordination chemistry of
a References 3 and 4.
hydroxamic acids has been studied extensively (particularly
with Cu(II), Zn(II), Fe(III), Co(III), and Cr(III) ions),3,4 there
have been no reports in the literature on the formation of
Cr(V) complexes of these ligands. The current interest in
coordination chemistry of Cr(V) arises mainly from the
proposed crucial role of reactive Cr(V) intermediates in
Cr(VI)-induced genotoxicity and carcinogenicity.5 Accumu-
* To whom correspondence should be addressed. E-mail: r.codd@
chem.usyd.edu.au, p.lay@chem.usyd.edu.au.
† Present address: Technische Universita¨t Mu¨nchen, Department Chemie,
Lehrstuhl fu¨r Makromolekulare Stoffe, Lichtenbergstrasse 4, 85747
Garching, Germany.
(3) Kurzak, B.; Kozlowski, H.; Farkas, E. Coord. Chem. ReV. 1992, 114,
169-200.
(4) (a) Abu-Dari, K.; Ekstrand, J. D.; Freyberg, D. P.; Raymond, K. N.
Inorg. Chem. 1979, 18, 108-112. (b) Abu-Dari, K.; Raymond, K. N.
Inorg. Chem. 1980, 19, 2034-2040.
(1) Miller, M. J. Chem. ReV. 1989, 89, 1563-1579.
(2) Muri, E. M. F.; Nieto, M. J.; Sindelar, R. D.; Williamson, J. S. Curr.
Med. Chem. 2002, 9, 1631-1653.
2934 Inorganic Chemistry, Vol. 44, No. 8, 2005
10.1021/ic048317d CCC: $30.25
© 2005 American Chemical Society
Published on Web 03/19/2005

Chromium(V) Complexes of Hydroxamic Acids
cm) of Sephadex LH-20 (Amersham Biosciences). Elution of the
column with acetone resulted in separation of a fast-moving green
fraction (containing Cr(III)-BHA complexes, as shown by ESMS,
see Results), which was discarded, followed by a brown fraction
of 1. A yellow fraction, probably containing unreacted Cr(VI),
remained on the column. The brown fraction was collected, acetone
was removed under reduced pressure (∼30 °C), and the residue
was dried under vacuum to a constant mass (3 days at 22 °C),
leading to a brown-black microcrystalline powder (135 mg; yield
of 1, 38%). Anal. Calcd for C17H16N2O6CrK: C, 46.89; H, 3.70;
N, 6.43; Cr, 11.94; K, 8.98. Found (average values and standard
deviations for three different preparations of 1): C, 46.5 ( 0.6; H,
3.8 ( 0.2; N, 4.8 ( 0.8; Cr, 11.8 ( 0.2; K, 9.5 ( 0.5.
Thermogravimetric analysis: mass loss 14% at 25-140 °C. ESMS
(acetone, -ve ion): 338.1 m/z. UV-vis (DMF): 740 nm (2.0 ×
102 M-1 cm-1). EPR (DMF): giso ) 1.9820; Aiso(53Cr) ) 13.9 ×
10-4 cm-1; aN ) 0.89 × 10-4 cm-1 (quintet). Magnetic moment
(solid, 295 K): µeff ) 2.07 µB. IR (KBr matrix): 3600-2500 m,
br, 1710 m, 1600 m, 1543 s, 1494 m, 1445 m, 1420 m, 1373 sh,
1344 s, 1176 w, 1144 w, 1110 w, 1071 w, 1024 w, 977 m, 918 m,
775 w, 721 m, 696 s, 647 m, 550 m, br. The product is soluble (up
to ∼0.2 M) in polar aprotic solvents, such as acetonitrile (MeCN),
DMF, or dimethyl sulfoxide (DMSO), and slightly soluble (∼0.5
mM) in water at 22 °C. For preparation of aqueous solutions of 1
(0.20-1.0 mM), freshly prepared solutions in MeCN, DMF, or
DMSO (20-100 mM) were diluted with H2O or aqueous buffer
solutions.
lation and reduction of Cr(VI) by microorganisms and plants,
used for bioremediation of contaminated soils,6 may well
involve the interactions of Cr(VI) with hydroxamato moieties
of siderophores.
This work presents the first description of Cr(V) inter-
mediates formed during the reactions of Cr(VI) with hy-
droxamic acids. In a similar manner as described in previous
studies on Cr(V) complexes,7-10 a combination of electron
paramagnetic resonance (EPR) spectroscopy, electrospray
mass spectrometry (ESMS), and X-ray absorption spectros-
copy (XAS) was used for the structural characterization of
Cr(V) complexes, since they have not yet been crystallized
due to their low stabilities in solutions.
Experimental Section
Caution! Cr(VI) compounds are human carcinogens,11 and
Cr(V) complexes are mutagenic and potentially carcinogenic;5
appropriate precautions should be taken to avoid skin contact and
inhalation of their solutions and dusts.
Reagents and Solutions. Commercial reagents of analytical or
higher purity (purchased from Aldrich, Sigma, or Merck) were used
without further purification. Hydroxamic acids, including aceto-
hydroxamic acid (AHA), benzohydroxamic acid (BHA), salicyl-
hydroxamic acid (SaHA), and suberohydroxamic acid (SuHA), were
from Aldrich. Acetone and n-hexane (BP Chemicals) were distilled
after boiling for 2-3 h with activated 4 Å molecular sieves
(Aldrich). Water was purified by the Milli-Q technique. Model
Cr(V) complexes, Na[CrO(ehba)2]‚1.5H2O (ehbaH2 ) 2-ethyl-2-
hydroxybutanoic acid) and [Cr(O)2(phen)2](BF4)‚0.3Et2O (phen )
1,10-phenanthroline), were synthesized and characterized as re-
ported previously.10,12,13 Stock solutions of the buffers were treated
by Chelex 100 chelating resin (BioRad) and stored at 4 °C. Working
solutions of the buffers were prepared daily by adjusting the pH
values of the stock solutions with high purity NaOH (99.99%,
Aldrich); the pH values were measured by an Activon 210
ionometer with an AEP 321 glass/calomel electrode. Concentrations
of the catalytic metals (Fe(III) and Cu(II)) in the buffers were <0.5
µM, as determined by Buettner’s ascorbate method.14
Synthesis of K[CrVO(L)2]‚Me2CO (1, LH2 ) BHA). Finely
ground anhydrous K2Cr2O7 (120 mg, 0.816 mmol Cr) and BHA
(330 mg, 2.40 mmol) were dissolved with stirring (∼30 min at 22
°C) in N,N-dimethylformamide (DMF) (10 mL), and the solution
was kept at 22 °C for a further 5 h. To the resultant dark-brown
solution, acetone (10 mL) was added, followed by the slow addition
(∼10 min, with stirring) of n-hexane (30 mL), which resulted in
separation of a black oil. This oil was dissolved in a minimal volume
of acetone, and the solution was applied to a column (1.5 × 20
Analytical Techniques. The Cr content of 1 was determined
(after digestion of the samples with 69% HNO3) by C2H2/air flame
atomic absorption spectroscopy, using a Varian SpecAA-800
spectrometer, calibrated with standard Cr(III) solutions (Aldrich).
Determination of K+ was performed with a Sherwood 410 flame
photometer, using KCl as a standard. Elemental analyses (C, H,
N) were performed by the Australian National University Mi-
croanalytical Unit or the Microanalysis Laboratory of the Chemistry
Department, University of Otago (Dunedin, New Zealand). Ther-
mogravimetric analyses (TGA) were performed on a TGA 2950
analyzer (TA Instruments) by heating the samples in an atmosphere
of high purity N2 (BOC gases) from 25 to 500 °C at 1 °C/min.
Magnetic susceptibility was measured on a Sherwood Scientific
magnetic balance, calibrated with (NH4)2Fe(SO4)2‚6H2O, and
diamagnetic corrections for the constituent atoms were calculated
from the literature.15 Solid-state IR spectra were recorded using
the diffuse reflectance technique (for the mixtures with KBr) on a
BioRad FTS-40 spectrometer. Cyclic voltammetry experiments were
performed using a Bioanalytical Systems BAS 100B Electrochemi-
cal Analyzer (scan rate, 100 mV s-1; 100% iR compensation) with
a glassy carbon working electrode (3 mm diameter), an Ag/AgCl
reference electrode, filled with 3.0 M aqueous NaCl, Pt wire as an
auxiliary electrode, and ferrocene (Fc) as an internal standard.
Electrochemical measurements were carried out in DMF solutions,
containing (nBu4N)BF4 (0.10 M) as the supporting electrolyte, and
saturated with O2-free Ar (BOC gases). Electronic absorption
spectra were acquired on a Hewlett-Packard HP 8452 A diode-
array spectrophotometer (300-800 nm) or on a Varian Cary 5E
spectrophotometer (700-1500 nm). Concentrations of Cr(VI) in
the decomposition products of 1 were determined spectrophoto-
metrically with diphenylcarbazide (DPC, ꢀ ) 4.2 × 104 M-1 cm-1
at 540 nm),16 as described previously,17 and calibrations were
performed by a standard additions method.
(5) For a recent review, see: Levina, A.; Codd, R.; Dillon, C. T.; Lay, P.
A. Prog. Inorg. Chem. 2003, 51, 145-250 and references therein.
(6) Zayed, A. M.; Terry, N. Plant Soil 2003, 249, 139-156.
(7) Codd, R.; Levina, A.; Zhang, L.; Hambley, T. W.; Lay, P. A. Inorg.
Chem. 2000, 39, 990-997.
(8) Headlam, H. A.; Weeks, C. L.; Turner, P.; Hambley, T. W.; Lay, P.
A. Inorg. Chem. 2001, 40, 5097-5105.
(9) Levina, A.; Zhang, L.; Lay, P. A. Inorg. Chem., 2003, 42, 767-784.
(10) Weeks, C. L.; Levina, A.; Dillon, C. T.; Turner, P.; Fenton, R. R.;
Lay, P. A. Inorg. Chem. 2004, 43, 7844-7856.
(11) IARC Monographs on the EValuation of the Carcinogenic Risk of
Chemicals to Humans. Vol. 49. Chromium, Nickel and Welding;
International Agency on the Research of Cancer: Lyon, France, 1990.
(12) Krumpolc, M.; Rocˇek, J. J. Am. Chem. Soc. 1979, 101, 3206-3209.
(13) Judd, R. J.; Hambley, T. W.; Lay, P. A. J. Chem. Soc., Dalton Trans.
1989, 2205-2210.
(15) Mabbs, F. E.; Machin, D. J. Magnetism and Transition Metal
Complexes; Chapman and Hall: London, 1973.
(14) Buettner, G. R. Methods Enzymol. 1990, 186, 125-127.
Inorganic Chemistry, Vol. 44, No. 8, 2005 2935

Gez et al.
Solid-state and solution EPR spectra (X-band) were acquired on
a Bruker EMX spectrometer, using quartz capillaries (solid state)
and a Wilmad quartz flat cell (solutions); calibrations of the
magnetic field and the microwave frequency were performed with
an EMX 035 NMR gaussmeter and an EMX 048T microwave
bridge controller, respectively. Low-temperature (77 K) EPR spectra
were recorded using a “coldfinger” liquid N2 dewar. Typical
instrumental settings were as follows: center field, 3480 G; sweep
width, 6900 G (solids) or 200 G (solutions); resolution, 1024 points;
microwave power, 2.0 mW; microwave frequency, ∼9.27 GHz
(solids at 77 K), ∼9.78 GHz (solids at 295 K), or ∼9.67 GHz
(solutions); modulation frequency, 100 kHz; modulation amplitude,
5.0 G (solids) or 0.40-1.0 G (solutions); time constant, 20.48 ms;
receiver gain, 102-105; and number of scans, 5. The EPR spectra
were processed with WinEPR and simulated with WinSim software,
respectively;18 second-order corrections were applied in the deter-
minations of the giso, Aiso, and aN values.
The ESMS analyses were performed using a Finnigan LCQ mass
spectrometer; typical experimental settings were as follows: sheath
gas (N2) pressure, 60 psi; spray voltage, 4.0 kV; capillary
temperature, 200 °C; cone voltage, 25 V; tube lens offset, 20 V;
m/z range, 100-2000 (both in positive- and negative-ion modes;
no significant signals were detected at m/z > 1200). Variations in
the capillary temperature (150-220 °C) and cone voltage (3-40
V) did not cause significant changes in the spectra. Analyzed
solutions (5 µL, ∼1 mM Cr) were injected into a flow of H2O/
MeOH (1:1 v/v; flow rate, 0.20 mL min-1). Acquired spectra were
the averages of 10 scans (scan time, 10 ms). Simulations of the
mass spectra were performed using IsoPro software.19
calibrated using a Cr foil as an internal standard (calibration energy,
5989.0 eV, which corresponded to the first peak of the first
derivative of Cr(0) edge).21 Averaging, background subtraction, and
the calculations of theoretical XAFS spectra were performed using
the XFIT software package,21 including FEFF 4.0622 and FEFF
6.0123 algorithms, as described previously.20,24 The determinacies
(Ni/p, where Ni is the number of independent observations and p is
the number of varied parameters) of the models used in the single-
and multiple-scattering (SS and MS) fits to the XAFS data were
estimated by the method of Binsted et al.,25 taking into account
the applied restraints and constraints. Overdetermined models (Ni/p
> 1) were used in XAFS calculations, which enabled meaningful
fits to be obtained.25 The goodness-of-fit parameters for XAFS
fittings using different models were compared using a statistical
F-test.26,27 The random errors in the estimated XAFS parameters
that arise from the noise in the data were determined by Monte
Carlo analysis within the XFIT software.21 Starting coordinate sets
for MS XAFS calculations were obtained from molecular models
of 1 built using the HyperChem software,28 and geometric
parameters of the ligands were taken from the crystal structures of
Cr(III)-BHA complexes.4
Biological Activity Assays. The abilities of 1 and model Cr(V)
complexes to induce oxidative DNA cleavage were determined from
the decrease in the relative concentrations of form I (covalently
closed supercoiled) DNA and increases in those of form II (nicked
circular) and form III (linear) DNA. Methods for preparation of
supercoiled plasmid pUC9 DNA (supplied by the laboratory of Dr.
N. E. Dixon, Australian National University) and DNA cleavage
assays were described previously (for the reactions of [CrV-
O(ehba)2]-).29 The assays were performed in phosphate buffers
(0.10 M, pH 7.4), since the use of organic buffers leads to inhibition
of DNA cleavage by Cr(V) complexes.29b Determination of Cr-
DNA binding caused by fresh and decomposed solutions of 1 was
performed by a method developed for the reactions of DNA with
[CrVO(ehba)2],30 using Bio-Gel P-30 spin columns (Bio-Rad) for
removal of unbound Cr complexes.30 The Cr-DNA binding studies
were performed in N-(2-hydroxyethyl)-piperazine-N′-(2-ethane-
sulfonate) (HEPES) buffer (0.10 M, pH 7.4), since the use of
phosphate buffer leads to a competition for Cr binding with the
phosphate groups of DNA.30 The bacterial mutagenicity of 1 was
tested using the TA102 strain of Salmonella typhimurium, as
described previously for [CrVO(ehba)2]- and other Cr(V) com-
plexes.31,32
All the analytical results were reproduced in at least two
independent experimental series, using different sets of stock
solutions and different preparations of 1; relative deviations in the
results of parallel experiments did not exceed 10%.
X-ray Absorption Spectroscopy and Data Processing. The
chromium K-edge spectrum of 1 was recorded on the Australian
National Beamline Facility (beamline 20B) at the Photon Factory,
Tsukuba, Japan. The beam energy was 2.5 GeV, and the beam
current was 300-400 mA. A Si[111] double-crystal monochromator
was detuned by 50%. The spectrum was recorded in transmission
mode, using N2/He-filled ionization chambers. A solid sample was
mixed with boron nitride (BN, 1:1 w/w) and pressed into a 0.5-
mm pellet supported in an Al spacer between two 63.5-µm Kapton
tape windows. The sample temperature was maintained at 10 ( 1
K using a closed-cycle He CryoIndustries REF-1577-D22 cryostat.
Collection of XAS data at low temperature minimized sample
photodamage, improved the signal-to-noise ratio, and maximized
the multiple-scattering contributions to the X-ray absorption fine
structure (XAFS) spectrum.20 The spectrum was averaged from
three scans taken at different positions on the sample; no color
changes at irradiated spots were observed, and the edge energies
differed by <0.1 eV between the scans. The energy scale was
(21) (a) Ellis, P. J.; Freeman, H. C. J. Synchrotron Radiat. 1995, 2, 190-
195. (b) XFIT for Windows 95; Australian Synchrotron Research
Program: Sydney, Australia, 1996.
(22) Mustre de Leon, J.; Rehr, J. J.; Zabinsky, S. I.; Albers, R. C. Phys.
ReV. B 1991, 44, 4146-4156.
(23) Rehr, J. J.; Albers, R. C.; Zabinsky, S. I. Phys. ReV. Lett. 1992, 69,
3397-3400.
(24) Levina, A.; Armstrong, R. S.; Lay, P. A. Coord. Chem. ReV. 2005,
249, 141-160.
(25) Binsted, N.; Strange, R. W.; Hasnain, S. S. Biochemistry 1992, 31,
12117-12125.
(16) Standard Methods for the Examination of Water and Wastewater, 19th
ed.; American Public Health Association: Washington, DC, 1995; pp
3.59-3.60.
(17) Levina, A.; Bailey, A. M.; Champion, G.; Lay, P. A. J. Am. Chem.
Soc. 2000, 122, 6208-6216.
(18) (a) WinEPR, Version 960801; Bruker-Franzen Analytic: Bremen,
Germany, 1996. (b) Duling, D. R. J. Magn. Reson. 1994, B104, 105-
110. The WinSim software is available via the Internet at
http://epr.niehs.nih.gov/.
(19) IsoPro 3.0; Senko, M.: Sunnyvale, CA, 1998.
(20) (a) Rich, A. M.; Armstrong, R. S.; Ellis, P. J.; Freeman, H. C.; Lay,
P. A. Inorg. Chem. 1998, 37, 5743-5753. (b) Rich, A. M.; Armstrong,
R. S.; Ellis, P. J.; Lay, P. A. J. Am. Chem. Soc. 1998, 120, 10827-
10836.
(26) Neville, A. M.; Kennedy, J. B. Basic Statistical Methods for Engineers
and Scientists; International Textbook: Scranton, PA, 1966; pp 312-
313.
(27) Levina, A.; Lay, P. A. Inorg. Chem. 2004, 43, 324-335.
(28) HyperChem, Version 5.1; Hypercube Inc.: Gainesville, FL, 1996.
(29) (a) 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. (b) Levina,
A.; Barr-David, R.; Codd, R.; Lay, P. A.; Dixon, N. E.; Hammershøi,
A.; Hendry, P. Chem. Res. Toxicol. 1999, 12, 371-381.
(30) Levina, A.; Lay, P. A.; Dixon, N. E. Chem. Res. Toxicol. 2001, 14,
946-950.
(31) Dillon, C. T.; Lay, P. A.; Bonin, A. M.; Dixon, N. E.; Collins, T. J.;
Kostka, K. L. Carcinogenesis 1993, 14, 1875-1880.
2936 Inorganic Chemistry, Vol. 44, No. 8, 2005

Chromium(V) Complexes of Hydroxamic Acids
Chart 1. Proposed Structures of Cr(V) Hydroximato Complexes
Results
Generation of Cr(V)-Hydroximato Complexes in So-
lutions. Reactions of hydroxamic acids (0.10 M of AHA,
BHA, SaHA, or SuHA, Chart 1) with Cr(VI) (10 mM, from
K2Cr2O7) in polar aprotic solvents, such as DMF, DMSO,
or MeCN, led to color changes from yellow to green or
brown (within minutes or hours at 22 °C) and to the
appearance of strong sharp signals in the solution EPR
spectra (giso ∼ 1.982, Figure 1 and Table 1), which are
characteristic for Cr(V) complexes.5,33 The spectra (Figure
1a,c) also possessed satellite signals due to the hyperfine
splitting of the 53Cr isotope (natural abundance, 9.55%). The
Aiso values (∼14 × 10-4 cm-1, Table 1) were typical for
five-coordinate Cr(V) species.5,33 Both the main signals and
the 53Cr satellites showed well-resolved quintet patterns
(Figure 1b,c) due to the superhyperfine coupling of two
equivalent 14N nuclei (aN ) (0.8-0.9) × 10-4 cm-1, Table
1). Similar, but much weaker, EPR signals were observed
(Figure S1, Supporting Information) during the reactions of
hydroxamic acids (0.10 M of AHA or BHA, or 25 mM of
SaHA or SuHA) with Cr(VI) (Na2CrO4, 10 mM) in neutral
aqueous solutions (0.10 M HEPES buffer, pH 7.4) in the
presence of the most abundant biological reductant, glu-
tathione (GSH, 10 mM),34 or in weakly acidic aqueous
solutions (0.10 M acetate buffer, pH 5.0) in the absence of
added reductants. Maximal intensities of the EPR signals for
aqueous solutions were achieved at ∼3 min after mixing of
the reagents, and the signals completely disappeared within
∼30 min at 22 °C.
Figure 1. X-band EPR spectra (22 °C) of the solutions formed during the
reactions of hydroxamic acids (0.10 M of AHA, BHA, SaHA, or SuHA)
with Cr(VI) (10 mM, from K2Cr2O7) in DMF for 3.5 h at 22 °C: (a) first-
derivative spectra, modulation amplitude ) 1.0 G (satellite signals due to
53Cr hyperfine coupling are shown by asterisks); (b) second-derivative
spectra (central signals only, modulation amplitude ) 0.40 G); and (c) an
expanded view of the satellite signals for the Cr(V)-BHA complex (second-
derivative spectrum, modulation amplitude ) 1.0 G; the central signal is
removed for clarity).
concentrations of Cr(V) were achieved at 2, 4, 14, or >15 h
(22 °C) for the reactions of SaHA, SuHA, BHA, or AHA,
respectively; typical kinetic curves at 742 nm are shown in
Figure S2b, Supporting Information. All the formed Cr(V)
complexes, except for that of SaHA, were stable in DMF
solutions for at least 15 h at 22 °C in the presence of excess
Cr(VI) and the ligand; the signal due to the SaHA complex
decayed significantly after ∼2 h (Figure S2b). The kinetic
data (Figure S2b) also indicated that the reactions of Cr(VI)
with aromatic hydroxamic acids (BHA and SaHA) led to
∼10-fold higher yields of Cr(V) complexes than the reactions
with aliphatic ligands, AHA or SuHA (given that the
absorbances at 742 nm were similar for all of the complexes).
Studies of the reaction mixtures by ESMS (Figure S3 and
Table S1 in the Supporting Information) revealed the
formation of Cr(V) hydroximato complexes of the general
formula [CrVO(L)2]- (where LH2 is a hydroxamic acid, m/z
) -214.1, -338.1, -370.1, and -472.2 for AHA, BHA,
SaHA, and SuHA, respectively), as well as of Cr(III)
Formation of Cr(V) species during the reactions of
Cr(VI) with hydroxamic acids could also be followed by
electronic absorption (UV-vis) spectroscopy due to the
appearance of a band at 700-800 nm that resulted from a
dfd electronic transition (Figure S2a in Supporting Informa-
tion), which is known to be specific for Cr(V) complexes.35
Kinetic studies by UV-vis spectroscopy (reaction conditions
corresponded to those in Figure 1) showed that the maximal
(32) Dillon, C. T.; Lay, P. A.; Bonin, A. M.; Dixon, N. E.; Sulfab, Y.
Aust. J. Chem. 2000, 53, 411-424.
(33) Barr-David, G.; Charara, M.; Codd, R.; Farrell, R. P.; Irwin, J. A.;
Lay, P. A.; Bramley, R.; Brumby, S.; Ji, J.-Y.; Hanson, G. R. J. Chem.
Soc., Faraday Trans. 1995, 91, 1207-1216.
(34) Kosower, E. M. In Glutathione: Metabolism and Function; Arias, I.
M., Jacoby, W. B., Eds.; Raven Press: New York, 1976; pp 1-15.
(35) Farrell, R. P.; Lay, P. A. Comments Inorg. Chem. 1992, 13, 133-
175.
Inorganic Chemistry, Vol. 44, No. 8, 2005 2937

Gez et al.
Table 1. EPR Spectroscopic Parameters of Cr(V) Hydroximato
Complexes in Solutions (22 °C)
a
Aiso
,
aN,b
LW,c
complex
solvent
giso
10-4 cm-1 10-4 cm-1 10-4 cm-1
Cr(V)-AHAd DMF
Cr(V)-BHAd DMF
Cr(V)-SaHAd DMF
Cr(V)-SuHAd DMF
1.9822
1.9820
1.9820
1.9822
13.9
13.9
14.4
13.8
13.8
13.9
14.0
14.0
14.4
0.86
0.90
0.93
0.83
0.88
0.89
0.85
0.90
0.82
1.0
0.92
0.87
1.0
0.88
0.95
1.3
1e
1e
1e
1e
1f
DMSO 1.9820
DMF 1.9820
MeCN 1.9821
acetone 1.9821
H2O
1.3
1.0
1.9819
a Hyperfine splitting due to the coupling of 53Cr nuclei (relative
abundance, 9.55%). b Superhyperfine splitting due to the coupling of 14N
nuclei. c Line width of the EPR signal. d Complexes were generated in situ
by the reactions of Cr(VI) (10 mM, from K2Cr2O7) with the corresponding
hydroxamic acid (0.10 M) in DMF for 3.5 h at 22 °C. e The isolated complex
1 (1.0 mM) was dissolved in the corresponding solvent; the spectra were
taken at 3-5 min after the dissolution (22 °C). Typical spectra are shown
in Figure S6, Supporting Information. f The isolated complex 1 (0.10 M)
was dissolved in DMSO and then diluted 20-fold with H2O; the spectrum
was taken after 3-5 min (22 °C). Identical spectral parameters were
observed for 1 in an aqueous buffer solution (1.0 mM Cr, 0.10 M HEPES,
pH 7.4, Figure 9a).
Figure 2. Typical ESMS data (negative-ion mode) for the isolated complex
1 (freshly prepared ∼1.0 mM solution in acetone). No significant signals
were observed in the positive-ion mode. Experimental (lines) and simulated
(columns) isotopic distributions for the main signal are shown in the inset.
Assignment of the signals (based on the m/z values and isotopic distribution
patterns, LH2 ) BHA): [CrVO(L)2]- (m/z ) -338.1); {2[CrVO(L)2]-‚Na+}
(m/z ) -699.2); and {2[CrVO(L)2]-‚K+} (m/z ) -715.2).
hydroxamato complexes (a typical formula, [CrIII(LH)3]0, in
agreement with the literature data).4a
No evidence was obtained (from UV-vis spectroscopic
or ESMS measurements) for the formation of Cr(IV)
complexes of hydroxamic acids under the conditions that
are known to stabilize Cr(IV) complexes with other types
of ligands, including (i) the reactions of Cr(VI) with As(III)
in weakly acidic aqueous media in the presence of excess
ligand,36 (ii) the reactions of Cr(VI) with hydroxamic acids
in MeOH solutions,7 (iii) the ligand-exchange reactions of
hydroxamic acids with preformed Cr(IV) 2-hydroxycarboxy-
lato complexes,37 and (iv) the reactions of Cr(V) hydroximato
complexes with one-electron reductants (thiols) in aqueous
or DMF solutions.38
Figure 3. Solid-state X-band EPR spectra of 1 at 77 and 295 K.
Synthesis and Characterization of a Cr(V)-BHA
Complex, 1. A combination of EPR and UV-vis spec-
troscopies and ESMS was used to search for the reaction
conditions that led to maximal yields of Cr(V) hydroximato
complexes. Of all the studied ligands, the only complex to
date that could be isolated in the solid state and purified by
column chromatography was the Cr(V)-BHA complex (see
the Experimental Section). Other complexes could be
obtained in the solid state, but not in a pure form, and
attempts to purify these complexes by column chromatog-
raphy led to further decomposition of the complexes.
The purity of the isolated complex, 1, was confirmed by
the following methods: (i) ESMS (Figure 2, showing the
absence of detectable amounts of Cr(VI) and Cr(III) species,
by contrast with the crude reaction mixtures, Figure S3 and
Table S1); (ii) solid-state EPR spectroscopy (Figure 3),
showing the absence of Cr(III) species, which would lead
to broad signals with g ∼ 2.0 and g ∼ 4-5;9,10 (iii) IR
spectroscopy (a comparison of spectra for 1 and the free
ligand, BHA, is shown in Figure S4, Supporting Informa-
tion); and (iv) magnetic susceptibility measurements (µeff
)
2.07 µB at 295 K is within the range of values for well-
characterized Cr(V) 2-hydroxycarboxylato complexes, 2.05-
2.10 µB).7,12 The presence of a CrdO bond in 1 is supported
by the observation of a band at 977 cm-1 in the IR spectrum
(Figure S4).35 The results of TGA (Experimental Section and
Figure S5 in the Supporting Information) were consistent
with the presence of one mole of acetone (which was used
in the chromatographic purification of 1) per mole of Cr. A
characteristic signal due to acetone (ν(CdO), 1710 cm-1)39
was evident in the IR spectrum of 1 (Figure S4), while
several weaker peaks (e.g., 1360 and 1220 cm-1)39 were
obscured by those of the Cr(V) complex. The results of
elemental analyses (Experimental Section) corresponded to
the formulation of 1 as K[CrVO(L)2]‚Me2CO (where LH2 )
BHA, Chart 1), except for the consistently lower than
expected nitrogen contents (measured for three independent
preparations of 1 in two laboratories). This discrepancy is
likely to be due to the formation of CrN during the thermal
(36) Ghosh, M. C.; Gould, E. S. Inorg. Chem. 1991, 30, 491-494.
(37) Codd, R.; Lay, P. A.; Levina, A. Inorg. Chem. 1997, 36, 5440-5448.
(38) Gould, E. S. Coord. Chem. ReV. 1994, 135, 651-684.
(39) Gordon, A. J.; Ford, R. A. The Chemist’s Companion. A Handbook
of Practical Data, Techniques, and References; John Wiley & Sons:
New York, 1972; p 170.
2938 Inorganic Chemistry, Vol. 44, No. 8, 2005

Chromium(V) Complexes of Hydroxamic Acids
Figure 4. Electronic absorption spectra of 1 in DMF solutions (freshly
prepared, 22 °C).
Figure 6. XANES spectrum of 1 in comparison with those of two
previously characterized Cr(V) complexes: Na[CrVO(ehba)2] (Cr(V)-
ehba)42 and Na3[CrVO(LH2)2] (Cr(V)-GSH, LH5 ) GSH).9 All the spectra
were taken at 10 K from mixtures of the Cr(V) solids with BN.
cathodic current for the reduction of 1 was ∼2 times higher
than the corresponding value for [CrVO(ehba)2]- (at the same
Cr concentration and scan rate, Figure 5), which points to a
direct two-electron reduction of 1 to a Cr(III) complex. The
reduction potential of 1 is comparable with those for Cr(V)
peptide complexes (E1/2 ) -1.23 to -1.34 V vs Fc+/Fc0 in
DMF solutions),41 but these complexes are reduced quasi-
reversibly to Cr(IV) species.41
X-ray Absorption Spectroscopy of 1. The X-ray absorp-
tion near-edge structure (XANES) spectrum of 1 (Figure 6,
solid mixture with BN, 10 K) shows a pre-edge absorbance
peak typical for Cr(V) oxo complexes (due to a symmetry-
forbidden 1s f 3d transition that becomes allowed as a result
of d-p orbital mixing).7-10,42 The edge position is shifted
to lower energies by ∼1 eV compared with that of a typical
Cr(V) complex with an O-donor ligand ([CrVO(ehba)2]-,
Figure 6)42 but is higher in energy by ∼2 eV compared with
that for a Cr(V) complex with a S-/N-donor ligand
([CrVO(LH2)2]3-, where LH5 ) GSH, Figure 6).9
SS modeling of the XAFS spectrum of 1 (typical results
are listed in Table S2, Supporting Information) suggested
the presence of one short (1.57 Å) and four longer (average
length, 1.93 Å) Cr-O bonds in the first coordination shell
of 1, which is consistent with the proposed structure,
[CrVO(L)2]- (LH2 ) BHA, Chart 1). Models for MS XAFS
analyses (Tables S3 and S4 in the Supporting Information)
were built based on typical structures of five-coordinate
Cr(V) oxo complexes42 and on the geometric parameters of
doubly deprotonated BHA ligands, determined by X-ray
crystallography of Cr(III)-BHA complexes.4 The models
included all the non-H atoms that are likely to be found
within 5 Å of the Cr atoms20,24 (so that the three most distant
C atoms in both Ph groups were truncated). The results of
MS XAFS fittings for 1 and the optimized structure of the
Figure 5. Typical cyclic voltammograms of 1 and [CrVO(ehba)2]- ([Cr]
) 5.0 mM) in DMF solutions (supporting electrolyte, 0.10 M (nBu4N)BF4)
on a glassy carbon electrode (diameter, 3 mm; scan rate, 100 mV s-1).
Directions of the scans are shown by arrows.
decomposition of 1, by a process that is similar to the reaction
described for Cr(III)-urea complexes.40
The electronic absorption spectra of 1 in polar organic
solvents (e.g., DMF, Figure 4) possessed two poorly defined
shoulders at 300-400 nm and 500-600 nm and a weak
maximum at ∼740 nm (ꢀ ) 2.0 × 102 M-1 cm-1), which is
typical for Cr(V) complexes with O-donor ligands.35
Solution EPR spectra of 1 at 22 °C (Table 1 and Figure
S6 in the Supporting Information) showed a strong solvent
dependency of the line widths. The EPR signals of 1 in
MeCN or acetone solutions were much broader than those
in DMF or DMSO solutions, which led to poor resolution
of the superhyperfine splitting patterns in the former two
solvents (Figure S6). The other EPR spectral parameters (giso,
Aiso, and aN, Table 1) were not significantly solvent-
dependent. Cyclic voltammetry of 1 in DMF solutions (using
(nBu4N)BF4 as a supporting electrolyte) showed irreversible
reduction at E ) -1.33 V (vs Fc+/Fc0, solid line in Figure
5), in contrast to a quasi-reversible Cr(V/IV) couple observed
for [CrVO(ehba)2]- under the same conditions (E1/2 ) -0.89
V vs Fc+/Fc0, dashed line in Figure 5).13 The maximal
(41) Barnard, P. J.; Levina, A.; Lay, P. A. Inorg. Chem. 2005, 44, 1044-
1053.
(42) Levina, A.; Codd, R.; Foran, G. J.; Hambley, T. W.; Maschmeyer,
T.; Masters, A. F.; Lay, P. A. Inorg. Chem. 2004, 43, 1046-1055.
(40) Qui, Y.; Gao, L. J. Am. Ceram. Soc. 2004, 87, 352-357.
Inorganic Chemistry, Vol. 44, No. 8, 2005 2939

Gez et al.
factors for the atoms in the ligands (for the cis models, the
latter values reached the upper limit of 0.02 Å2, Table S4).
The initial MS XAFS models had starting Cr-O bond
lengths of 1.93 Å, which were chosen from the results of
SS XAFS calculations (Table S3). However, all the unre-
strained fits led to unequal Cr-O bond lengths, with longer
bonds adjacent to the N atoms of the ligands (1.98(2) Å for
Cr-O(2) and Cr-O(4) vs 1.88(2) Å for Cr-O(3) and
Cr-O(5), Figure 8 and Table S4). Restrained models with
the reversed order of Cr-O bond lengths or those with equal
(within 0.02 Å) Cr-O bonds led to deterioration in the fits
(the R values increased by 2-3%, Table S4), and calculations
returned to the optimized model (Figure 8) when the
restraints were released. The values of CrdO bond lengths
in 1, optimized by MS XAFS fittings, were 1.57-1.58 Å
(Table S4), in agreement with the SS XAFS results (Table
S3). The optimized geometry of 1 (Figure 8 and Table S4)
was close to those of typical distorted trigonal-bipyramidal
structures of five-coordinate Cr(V) oxo complexes,35,43
although MS XAFS calculations are unable, in most cases,
to produce unique values of bond angles in metal complexes,
due to relatively small MS contributions of the functional
groups with bond angles at <150° (Table S5 in the
Supporting Information).24
Reactivity of 1. Prolonged storage of 1 as a purified solid
(∼6 months at -20 °C, protected from light and moisture)
led to a significant reduction of 1 to Cr(III) complexes, as
determined by ESMS (Table S1) and column chromatogra-
phy (Sephadex LH-20, see the Experimental Section).
Solutions of 1 in aprotic solvents (1.0 mM in DMF, DMSO,
MeCN, or acetone) were stable for at least 3 days at 22 °C
(no significant decomposition was detected by EPR and
UV-vis spectroscopies and ESMS), but repeated attempts
to grow X-ray quality crystals of 1 from these solvents led
to precipitation of Cr(III) complexes within several weeks
or months, even at -20 °C. Solutions of 1 (1.0 mM) in
MeOH or EtOH were completely reduced to Cr(III) com-
plexes within ∼12 h at 22 °C, as determined by EPR
spectroscopy and ESMS.
Figure 7. Experimental and calculated XAFS and Fourier transform (FT)
XAFS spectra of 1 (solid, 10 K). The structure of the complex, optimized
by MS XAFS calculations, is shown in Figure 8, and details of the MS
XAFS calculations are given in Tables S3-S5, Supporting Information.
Figure 8. Structure of the complex 1, optimized by MS XAFS calculations
(the Ph groups of the ligands are truncated; calculation details are given in
Tables S3-S5, Supporting Information). Selected parameters of the
optimized model: |Cr-O(1)| ) 1.58(2) Å; |Cr-O(2)| ) |Cr-O(4)| )
1.98(2) Å; |Cr-O(3)| ) |Cr-O(5)| ) 1.88(2) Å; O(1)-Cr-O(2) )
O(1)-Cr-O(4) ) 82(2)°; O(1)-Cr-O(3) ) O(1)-Cr-O(5) ) 91-
(2)°; and O(2)-Cr-O(3) ) O(4)-Cr-O(5) ) 83(2)° (the values for
equivalent bonds in both ligands were restrained to be equal within 0.02 Å
or 2° during the calculations, Table S3). A full list of optimized parameters
is given in Table S4 (model trans-1).
Decomposition of 1 (0.20 mM) in aqueous solutions
occurred over a time scale of seconds or minutes at 22 °C
(as determined by EPR and UV-vis spectroscopies). The
maximal lifetimes of the Cr(V) complexes were observed at
pH ∼ 7 (Table 2, the t1/2 values apply only to [Cr] ) 0.20
mM and are likely to be concentration-dependent). Formation
of [CrO4]2- (λmax ) 372 nm, ꢀmax ) 4.81 × 103 M-1 cm-1)44
during the decomposition of 1 in basic or neutral media (in
the absence of added reductants) was evident from UV-vis
spectra of the decomposition products (typical examples are
shown in Figure S7, Supporting Information). The UV-vis
spectra of the decomposition products of 1 in acidic media
or in the presence of reductants, such as GSH, were
consistent with the predominant formation of Cr(III) (a
complex are shown in Figures 7 and 8, and details of the
MS XAFS calculations are given in Tables S4 and S5,
Supporting Information.
The initial models used for MS XAFS fitting to the
structure of 1 included those with cis and trans positions of
Ph groups relative to the Cr atom (similar to those in the
X-ray crystal structures of Cr(III)-BHA complexes),4 but
the results of the best fits showed a clear preference for the
trans models. This was based on better goodness-of-fit
parameters (R ) 14.4 vs 16.6% for the optimized trans and
cis models, respectively, Table S4) and lower Debye-Waller
(43) Lay, P. A.; Levina, A. In ComprehensiVe Coordination Chemistry II;
McCleverty, J. A., Meyer, T. J., Eds.; Elsevier: Oxford, U.K., 2004;
Vol. 4, pp 313-413.
(44) Brasch, N. E.; Buckingham, D. A.; Evans, A. B.; Clark, C. R. J. Am.
Chem. Soc. 1996, 118, 7969-7980.
2940 Inorganic Chemistry, Vol. 44, No. 8, 2005

Chromium(V) Complexes of Hydroxamic Acids
Table 2. Decomposition of 1 in Aqueous Solutionsa
solution
H2O (pH 6.6)
phosphate buffer (0.10 M, pH 7.4)
HEPES buffer (0.10 M, pH 7.4)
GSH (5.0 mM) in HEPES buffer
(0.10 M, pH 7.4)
t
1/2, sb
disproportionation, %c
44
65
88
25
31
27
0
<5
acetate buffer (0.10 M, pH 4.0)
NaOH (0.10 M)
HCl (0.10 M)
15
<5
<5
12
41
5
a All the experiments were performed at [Cr] ) 0.20 mM and 22.0 (
0.1 °C. b The half-life time of 1 (0.20 mM) was measured by EPR and/or
UV-vis spectroscopies. c The amount of 1 (mol %) decomposed by
disproportionation (eq 2) was calculated from the concentration of Cr(VI)
in the decomposition products (determined at t ∼ 10t1/2 by the DPC
method),16,17 using the known stoichiometry of Cr(V) decomposition (eqs
1 and 2).46
decrease in absorbance at 400-800 nm compared with the
fresh solution of 1, Figure S7).9,10,45 Determination of
Cr(VI) in the decomposition products (by the DPC method)16,17
showed that 1 decomposes predominantly by ligand or
substrate oxidation (eq 1) under all the conditions listed in
Table 2. Typically for Cr(V) oxo complexes,7,9,10,46 the role
of disproportionation (eq 2) becomes more significant with
an increase in the pH values of aqueous solutions (Table 2).
Figure 9. Typical X-band EPR spectra (solutions, 22 °C) for the ligand-
exchange reactions of 1 (1.0 mM) with D-glucose or bovine mucin in HEPES
buffers (0.10 M, pH 7.4); reaction time 3-4 min (a-c) or 8-10 min (d).
The reaction solutions contained DMF (1% v/v) that was used for complete
dissolution of 1. Assignment of the signals: signal 1 (giso ) 1.9820) is due
to 1, signal 2 (giso ) 1.9811) is due to the mixed-ligand species (Cr(V)-
BHA-carbohydrate), signal 3 (giso ) 1.9784-1.9805) is due to Cr(V)-
carbohydrate complexes (a mixture of linkage isomers),33,47,48 and signal 4
(giso ) 1.9778) is possibly due to a hydrolysis product of 1.
Cr(V)dO + S f Cr(III) + SdO
(S is the ligand or an added reductant) (1)
3Cr(V) f 2Cr(VI) + Cr(III)
(2)
Studies by ESMS of the decomposition products of 1 in
unbuffered H2O (pH 6.6, reaction time 1 h at 22 °C) revealed
the formation of PhCOO- as an oxidation product of BHA,
as well as Cr(III)-BHA complexes and [CrO4]2- (Figure
S8 in the Supporting Information).
genotoxicity and carcinogenicity.5,49 In contrast to the activity
of previously studied Cr(V) oxo complexes,29,31,32,50-52 no
significant cleavage of plasmid (pUC9) DNA was detected
in aqueous solutions (phosphate buffers, pH 4.0-7.4) of 1
(0.20-1.0 mM) in the presence or absence of GSH (2.0 mM).
Typical results of plasmid DNA cleavage assays for 1, in
comparison with those for the positive controls, [CrVO-
(ehba)2]- 29,50 and [CrV(O)2(phen)2]+,10,32 are shown in Figure
S10 (Supporting Information).
In a similar way as other Cr(V) oxo complexes,9,10,33
1
readily underwent ligand-exchange reactions with 1,2-diolato
ligands, including D-glucose (a typical carbohydrate) or
bovine mucin (a typical glycoprotein)47,48 in neutral aqueous
solutions, as shown by EPR spectroscopy (Figure 9).
Dissolution of 1 (1.0 mM) in a weakly acidic solution (pH
4.0) containing a typical 2-hydroxycarboxylato ligand,
ehbaH2 (10 mM), led to the formation of [CrVO(ehba)2]-
(detected by EPR spectroscopy, giso ) 1.9783, Figure S9 in
the Supporting Information).33 Estimation of equilibrium
constants for these ligand-exchange reactions is complicated
by the decomposition of 1, which occurs over a similar time
scale (Figures 9 and S9 and Table 2).
Biological Activity of 1. The data on the stability of 1 in
aqueous solutions (Table 2) were used to design the
conditions for testing of the biological activity of 1. The
ability of several types of Cr(V) complexes to induce
oxidative DNA cleavage in vitro has led to the postulate that
Cr(V) intermediates have an important role in Cr(VI)-induced
An additional nonoxidative pathway of Cr(V)-induced
DNA damage, recently proposed for [CrVO(ehba)2]-,30
involves the formation of Cr(III)-DNA adducts from reac-
tive positively charged Cr(III) species (which readily bind
to the negatively charged DNA backbone) that are formed
during the decomposition of Cr(V) in neutral aqueous
solutions. Incubation of 1 (0.50 mM) with fish sperm DNA
(0.10 mg mL-1) in HEPES buffer (0.10 M, pH 7.4, reaction
time 1 h at 37 °C) caused relatively low levels of Cr-DNA
binding (∼2% of total Cr vs ∼25% for [CrVO(ehba)2]- under
the same conditions).30 Furthermore, no significant changes
(49) Sugden, K. D.; Stearns, D. M. J. EnViron. Pathol., Toxicol. Oncol.
2000, 19, 215-230.
(45) Lay, P. A.; Levina, A. J. Am. Chem. Soc. 1998, 120, 6704-6714.
(46) Krumpolc, M.; Rocˇek, J. Inorg. Chem. 1985, 24, 617-621.
(47) (a) Codd, R.; Lay, P. A. J. Am. Chem. Soc. 2001, 123, 11799-11800.
(b) Codd, R.; Lay, P. A. Chem. Res. Toxicol. 2003, 16, 881-892.
(48) Codd, R.; Irwin, J. A.; Lay, P. A. Curr. Opin. Chem. Biol. 2003, 7,
213-219.
(50) Sugden, K. D.; Wetterhahn, K. E. J. Am. Chem. Soc. 1996, 118,
10811-10818.
(51) Sugden, K. D.; Campo, C. K.; Martin, B. D. Chem. Res. Toxicol. 2001,
14, 1315-1322.
(52) Levina, A.; Ludwig, C.; Lay, P. A. J. Inorg. Biochem. 2003, 96, 177.
Inorganic Chemistry, Vol. 44, No. 8, 2005 2941

Gez et al.
in the levels of Cr-DNA binding were observed when 1
was allowed to decompose in the buffer solution (1 h at 37
°C, complete disappearance of Cr(V) was verified by EPR
spectroscopy) prior to the addition of DNA (after which the
reaction was continued for another 1 h at 37 °C). By contrast,
decomposition of [CrVO(ehba)2]- prior to the addition of
DNA leads to a dramatic decrease in the levels of Cr-DNA
binding, since the initially formed reactive Cr(III) species
undergo base-catalyzed hydrolysis with the formation of
nonreactive polynuclear complexes.30 In summary, it appears
that the Cr(III)-BHA complexes, formed during the de-
composition of 1 in aqueous media (Figures S7 and S8),
possess some DNA-binding ability, but this binding does not
require the presence of Cr(V)-BHA species.
Bacterial mutagenicity assays with S. typhimurium (Ames
tests)53 are among the best-validated methods for the detec-
tion of chemical mutagens that include Cr(VI)54 and several
types of Cr(V) complexes.29a,31,32 Preliminary results of these
assays for fresh and decomposed (for 1 h at pH 7.4 and 37
°C) solutions of 1 (Table S6 in the Supporting Information)
show that the observed weak effect of 1 is likely to be due
to the formation of [CrO4]2- (a known mutagen)54 during
the decomposition of 1 in neutral aqueous media, rather than
due to a specific activity of the Cr(V) complex.
explained by the inability of Cr(IV) (a weaker electron
acceptor than Cr(V))37,38 to deprotonate the relatively weak
acidic donor groups of the hydroxamic acids.
The proposed structures of Cr(V) hydroximato complexes,
in which the O-donor atoms are adjacent to N atoms (Chart
1), are consistent with the observation of 14N superhyperfine
splitting in solution EPR spectra (Figure 1). The magnitude
of the 14N splitting (∼0.9 × 10-4 cm-1, Table 1) was lower
than that usually observed for N atoms bound directly to
1
Cr(V) ((2-3) × 10-4 cm-1)35,55 but similar to that for H
superhyperfine splitting due to the C-H bonds adjacent to
the O-donor atoms in Cr(V) complexes with cis-diolato
ligands.33,35,48,55 The strength of the 14N superhyperfine
coupling can be explained by the spatial overlap of the dxy
orbital containing the unpaired electron with orbitals of the
N atom in the chelate ring, in a similar manner as described
for rationalizing the 1H superhyperfine coupling of equatorial
protons in the chelate rings of the interaction of the equatorial
ring of cis-diolato ligands.55
Analyses of the XAFS spectrum of 1 (solid, 10 K), using
MS models (Figure 7 and Tables S3-S5), suggested the
dominance of a trans-isomer (Figure 8) in the solid state,
which can be explained by the steric hindrance of the Ph
residues in the ligands. Equilibrium mixtures of geometric
isomers can be formed in solutions of 1 and other Cr(V)-
hydroximato complexes, as such mixtures occur in the
solutions of Cr(III)-hydroximato complexes4 and of Cr(V)
2-hydroxycarboxylato complexes.7,42 In a similar manner as
that described for the well-studied Cr(V) 2-hydroxycarboxy-
lato complexes (such as [CrVO(ehba)2]-), interconversions
between the geometric isomers of 1 in solutions are likely
to occur by intramolecular Berry twists.35 The observed
solvent-dependence of the line widths in the EPR signals
(Table 1 and Figure S6) probably reflects the differences in
the rates of Berry twists in various solvents. In the more
polar solvents (DMF, DMSO, or H2O), interconversions
between the geometric isomers are fast on the EPR time
scale, leading to narrow averaged signals (Table 1 and Figure
S6). In the less polar solvents (MeCN or acetone), the rates
of Berry twists in 1 become comparable with the EPR time
scale, leading to the broadening of the signals (Table 1 and
Figure S6).35 The broadening of the outer 53Cr satellite signals
in the EPR spectra of Cr(V)-BHA species formed in situ
in DMF solutions (Figure 1c) also suggests the presence of
more than one geometric isomer in solution.33,35
Discussion
Reactions of Cr(VI) with various hydroxamic acids in
polar aprotic solvents led to relatively stable Cr(V)-
hydroximato intermediates (detected by EPR and UV-vis
spectroscopies and ESMS, Figures 1, S2 and S3), which
represent a new family of Cr(V) compounds.43 In these
reactions, hydroxamic acids act as both reductants and
ligands (in a similar manner as described for the syntheses
of other Cr(V) complexes by the reactions of Cr(VI) with
excess ligand).9,12 The structures of the resultant complexes,
[CrVO(L)2]- (where LH2 ) hydroxamic acid, Chart 1), were
assigned on the basis of the spectroscopic methods mentioned
above. Their structures were confirmed by isolation and
structural characterization (by MS analyses of XAFS spec-
troscopic data) of a Cr(V)-BHA complex, 1, and the
similarity of the EPR spectrum of this structurally character-
ized complex with the other complexes reported here.
Notably, a Cr(V)-hydroximato complex was formed with
high selectivity even with SaHA (as determined by EPR
spectroscopy, Figure 1), a ligand that contains another strong
donor group (phenolate). Thus, the hydroximato(2-) moiety
provides a favorable coordination environment for Cr(V).
By contrast, no stable Cr(IV)-hydroximato species were
observed to date (see the Results). Irreversible electrochemi-
cal reduction of 1, as well as the low reduction potentials
(Figure 5), are consistent with very low stability of Cr(IV)-
BHA species in solution. This difference in stabilities of
Cr(V) and Cr(IV) complexes with hydroximato(2-) ligands
parallels that for 1,2-diolato(2-) ligands37 and is probably
A significant difference in the Cr-O(C) and Cr-O(N)
bond lengths in 1 (1.88 and 1.98 Å, respectively, Figure 8),
revealed by MS analyses of the XAFS data, is in contrast
with the X-ray crystallographic data for Cr(III)-BHA
complexes (mer-[CrIII(LH)3]0 and mer- and fac-[CrIII(L)3]3-,
LH2 ) BHA), where these two types of bond lengths are
similar (1.96-1.98 Å for Cr-O(C) and 1.95-1.96 Å for
Cr-O(N)).4b These differences in bond lengths between the
Cr(V) and Cr(III) complexes of the same ligand are likely
to be due to the differences in geometry (distorted trigonal
(53) Maron, D. M.; Ames, B. N. Mutat. Res. 1983, 113, 173-215.
(54) Bennicelli, C.; Camoirano, A.; Petruzzelli, S.; Zanacchi, P.; De Flora,
S. Mutat. Res. 1983, 122, 1-5.
(55) Codd, R.; Dillon, C. T.; Levina, A.; Lay, P. A. Coord. Chem. ReV.
2001, 216-217, 537-582.
2942 Inorganic Chemistry, Vol. 44, No. 8, 2005

Chromium(V) Complexes of Hydroxamic Acids
bipyramidal vs octahedral) and the presence of a sterically
demanding oxo group in the Cr(V) complex.
Chromium(V) hydroximato complexes are formed as
unstable intermediates during the reactions of Cr(VI) with
the ligands in neutral or weakly acidic aqueous media, as
detected by EPR spectroscopy (Figure S1). Such complexes
may also be formed during the microbial reduction of
Cr(VI) in contaminated soils, due to the binding of Cr(V) to
hydroximato ligands of siderophores, produced by soil
bacteria or plants.1,6 Once formed, such complexes could
readily undergo ligand-exchange reactions with carbohydrate
ligands (Figure 8), which are ubiquitous in biological
systems, to form long-lived Cr(V) species.48 In summary, a
likely role of Cr(V) hydroximato complexes in biological
systems is to serve as precursors of more stable and DNA-
damaging species, such as Cr(V) carbohydrate complexes.
Like the Cr(V) 2-hydroxycarboxylato complexes7,13 but
unlike the Cr(V) complexes with amido or imine donor
ligands,10,31 the Cr(V) hydroximato complex, 1, is much more
stable in aprotic organic solvents (DMF, DMSO, MeCN, or
acetone) than in aqueous media (see the Results). The
reactivity patterns of 1 in aqueous solutions (Table 2),
including a stability maximum at pH ∼ 7 and a predominant
decomposition by a ligand-oxidation pathway (eq 1), are
reminiscent of those for [CrVO(LH2)2]3- (where LH5 )
GSH).9 Detailed mechanistic studies of the decomposition
of 1 were beyond the scope of this work, particularly since
our recent investigation of the decomposition of [CrV-
O(ehba)2]- in aqueous solutions56 showed a high degree of
complexity of such reactions. Any comparisons of (i) the
relative rates of reactions and (ii) the relative proportions of
decomposition of 1 by competing disproportionation and
ligand/solvent reduction reactions, compared with other
Cr(V) complexes, cannot be related simply to the Cr(V/IV)
reduction potential. Any analysis of such reactivity is
complicated by the facts that (i) the reduction is irreversible
in some cases (such as for 1) but reversible in many others
and (ii) there is a large range in the ease with which the
coordinated ligand is oxidized. The very negative reduction
potential combined with the relatively high proportion of
ligand oxidation for 1 compared with other Cr(V) complexes,
under the same conditions, suggests that it is the ease of
ligand oxidation that is the main contribution to the lower
stability of the complex compared with other bidentate
chelating ligands such as 2-hydroxycarboxylates. Such an
assertion is supported by the fact that 1 undergoes significant
reduction to Cr(III) in the solid state (protected from moisture
and light) over 6 months while most other Cr(V) solids with
a variety of ligands are stable under these conditions. The
relatively low stability of 1 in aqueous solutions (Table 2),
combined with its very negative reduction potential (Figure
5), is probably responsible for its low reactivity toward DNA
(Figure S10) and the absence of detectable mutagenicity in
the TA102 strain of S. typhimurium (Table S6). This low
biological activity of 1 is in stark contrast with the high
mutagenicity and/or genotoxicity in vitro of many types of
Cr(V) complexes, including those with 2-hydroxycarboxy-
lato,29,50 amido,31 imine,32,51 or thiolato (GSH)52 donor
ligands.
Acknowledgment. Financial support of this work was
provided by Australian Research Council (ARC) Large and
Discovery grants and an Australian Professorial Fellowship
to P.A.L., ARC RIEFP for the 10-element Ge detector,
Wellcome Trust and ARC RIEF grants for the ESMS and
EPR equipment, and Australian Synchrotron Research pro-
gram (ASRP) grants for access to the Australian National
Beamline Facility (ANBF) in Tsukuba, Japan. The X-ray
absorption spectroscopic (XAS) studies were performed at
ANBF with support from the ASRP, which is funded by the
Commonwealth of Australia under the Major National
Research Facilities program. R.C. is grateful for funding
through the Gritton Postdoctoral Fellowship. We thank Drs.
Garry Foran (ANBF) and Hugh Harris (University of
Sydney) for assistance in recording the XAS data, Professor
Nicholas Dixon and Mrs. Penelope Lilley (Australian
National University) for providing pUC9 plasmid DNA, Dr.
Antonio M. Bonin (University of Sydney) for providing the
Salmonella strain and assistance with the bacterial mutage-
nicity assays, and Dr. Colin L. Weeks (University of Sydney)
for assistance with TGA.
Supporting Information Available: Tables showing the as-
signment of major ESMS signals observed during the reactions of
Cr(VI) with hydroxamic acids, details of SS and MS XAFS
calculations, and typical results of Salmonella mutagenicity assays;
and Figures showing typical spectroscopic data (EPR, UV-vis, and
ESMS) for the reactions of Cr(VI) with hydroxamic acids, and
detailed characterization of 1, including IR spectra, TGA data, EPR
spectra in various solvents, UV-vis spectra and ESMS data for
decomposed aqueous solutions, and typical results of DNA cleavage
assays in comparison with those for other Cr(V) complexes.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(56) Levina, A.; Lay, P. A.; Dixon, N. E. Inorg. Chem. 2000, 39, 385-
395.
IC048317D
Inorganic Chemistry, Vol. 44, No. 8, 2005 2943