Reaction of Chromium (VI) with Ascorbate Produces Chromium (V

Chem. Res. Toxicol. 1994, 7, 219-230

219

Reaction of Chromium(V1) with Ascorbate Produces Chromium(V), Chromium(IV), and Carbon-Based Radicals Diane M. Stearns and Karen E. Wetterhahn’ Department of Chemistry, 612%Burke Laboratory, Dartmouth College, Hanover, New Hampshire 03755-3564 Received May 12,1999

Reaction of potassium dichromate with sodium ascorbate was studied by EPR spectroscopy a t room temperature, in 0.10 M N-[2-hydroxyethyl]piperazine-N’-[2-ethanesulfonicacid] (HEPES),phosphate, cacodylate, and tris(hydroxynethy1)aminomethane hydrochloride (Tris-HC1) buffers a t p H 7.0, in the presence of 0.10 M spin trap [5,5-dimethyl-l-pyrrolinel-oxide or 2-methyl-N-(4-pyridinylmethylene)-2-propanamine NJV’-dioxide]. Chromium(V), ascorbate radical, C02-, and other carbon-based spin trap-radical adducts were observed. Chromium(V), COz-, and the carbon-based radicals were observed a t low ratios of ascorbate to chromium, and ascorbate radical was observed at high ratios of ascorbate t o chromium. The presence of Cr(1V) was detected indirectly by reaction with Mn(I1) and a subsequent decrease in the Mn(I1) E P R signal. More Cr(1V) was found for the higher reaction ratios of ascorbate t o Cr(V1). The only buffer effect observed was arelative decrease of the Cr(V) signal in Tris.HC1us HEPES, phosphate, and cacodylate buffers, no change in the radical adducts was observed. There was no evidence for reactive oxygen species as intermediates in this reaction. Addition of the singlet oxygen trap 2,2,6,6-tetramethyl-4-piperidone hydrochloride showed no 2,2,6,6-tetramethyl-l-piperidinyloxy radical formation. The Cr(V) species did not react with dioxygen, and dioxygen did not affect the formation of carbon-based radicals. A mechanism consistent with these observations is discussed.

Introduction Chromium(V1) is a known carcinogen in humans and animals (1,2);however, Cr(V1) does not react with DNA in vitro unless reducing agents are added. The uptakereduction model (3) proposes that Cr(V1) is reduced intracellularly, producing reactive intermediates that target DNA. These potentially genotoxic species include Cr(V), Cr(IV), Cr(III), free radicals, and reactive oxygen species. Although glutathione has been the focus of major study ( 4 ) ,there has been renewed interest in the reaction between Cr(V1) and ascorbate (vitamin C) since anumber of reports have suggested that ascorbate is also involved in the reductive pathway of Cr(V1) in vivo. Ascorbate was found to be more reactive than glutathione for reduction of Cr(V1) in rat lung (5). Ascorbate has been shown to be the major reductant of Cr(V1) in rat lung, kidney, and liver ultrafiltrates (6,7).Cr-DNA binding resulting from the reaction of Cr(V1) with DNA in the presence of rat lung ultrafiltrates was correlated to ascorbate-dependent metabolismof Cr(VI) (7). Increases in the ascorbate levels of V-79 cells affected relative levels of Cr(V1)-inducedDNA damage, specifically decreasing alkali-labile sites and increasing DNA-protein cross-links and cytotoxicity (8). Ascorbate was found to block extracellular dissolution of lead chromate particles and to decrease genotoxicity of Cr(V1) in Chinese hamster ovary cells (9). One basic question that remains to be fully answered is the role of cellular reductants such as ascorbate and glutathione toward activation of Cr(V1) to reactive Cr(V), Cr(IV),or radicals us detoxification of Cr(V1)byreductioil to the stable end product Cr(II1) and/or scavanging of

* Author to whom correspondence should be addressed. Telephone: (603) 646-3413;Fax: (603) 646-3946. 0 Abstract published in Advance ACS Abstracts, March 1, 1994. 0893-228%J94/2707-0219$04.50/0

radicals. It is a major hypothesis in this laboratory that Cr(V1)-induced DNA damage will be dependent on the relative concentrationsof these small molecule reductants, and that the different types of DNA damage observed in different animal tissues and cell lines will be related to different intracellular stoichiometries of Cr(V1) and reducing agents. One manifestation of the different reactivity of Cr(V1) with cellular reductants may be two pathways of Cr(V1)-induced DNA damage, namely a chromium-mediatedpathway and a free radical pathway. Although the understanding of Cr(V1) metabolism in vivo will likely involve a combined effect of ascorbate and glutathione, one of the first steps is to understand the reaction of Cr(V1) with ascorbate alone. In light of these goals the previous literature on Cr(VI)/ascorbate chemistry is incomplete. A previous EPR study (10) in 1 M Tris.HC1 and 0.10 M N-12hydroxyethyllpiperazine-N’-[2-ethanesulfonicacid] (HEPEW buffers at pH 7-8.5, at room temperature (RT) had shown that the in vitro reduction of Cr(V1) by ascorbate produces Cr(V) and ascorbate radical as unstable intermediates; however, spin traps were not used to look for more reactive freeradicals. The interaction of TriseHCl buffer with Cr(V) was reported (10)but the effect of buffer on the stability of Cr(V) was not determined, which is crucial for evaluating the potential reactivity of Cr(V) toward DNA in vitro (111,nor was an attempt made to quantitate the Cr(V) or ascorbate radical formed. While 1 Abbreviations: ascH-, ascorbate; asc-, ascorbate radical, DHA, dehydroaecorbate; DKG, 2,a-diketogularic acid; DMPO,5,5-dimethyl1-pyrroline1-oxide;DPPH, 2,2-diphenyl-l-pi&ylhyd?a.zylradical,EHBA, 2-ethyl-2-hydroxybutyric acid; HEPES, N-[2-hydroxyethyllpiper~~~ N’-[2-ethanesulfonicacid]; POBN,a-(4-pyridyll-oxide)-N-tert-butylnitrone, 2-methyl-N-(4-pyridinyhnethylene)-2-propanamine N,”-dioxide; R, carban-based radical; RT, mom temperature; TEMPO, 2,2,8,6tetramethyl-l-piperidinyloxy.

6 1994 American Chemical Society

220 Chem. Res. Toxicol., Vol. 7, No. 2, 1994

studies were underway in this laboratory two reports were published (12, 13) which presented evidence for involvement of dioxygen in the reaction of Cr(V1) with ascorbate. A recent kinetic study in phosphate-citrate and Tris.HC1 buffers (12) determined a 10-fold increase in the reaction rate for reactions in the absence of oxygen. A second report (13) suggested the formation of activated oxygen species upon reaction of Cr(VI) with ascorbate in phosphate buffer at pH 7.4, and 37 “C. This conclusion was based on the observations that the COO- adduct of 5,5-dimethyl-1pyrroline 1-oxide(DMPOICOi-) was generated during the reaction of Cr(V1)with ascorbate in the presence of formate and DMPO and that this DMPO/C02-signal was oxygen dependent. The authors concluded that the Cr(V) was reacting with dioxygen to form reactive oxygen species which were trapped by formate. There were some discrepancies between their results (13)and ours (vide infra) which needed to be clarified in order to understand the reaction of Cr(V1) and ascorbate in vitro and in vivo. We have used EPR spectroscopy t o reexamine the reaction of Cr(V1) with ascorbate under a variety of conditions and the work described here shows that the reaction is much more complicated than previously believed (10, 12-1 7). Under our conditions more than one carbon-basedradical is produced from the fragmentation of ascorbate upon reaction with Cr(V1) and there is a buffer dependence for the stability of the Cr(V) signal. We see no reactivity of Cr(V) with dioxygen; nor any oxygen dependence for the formation of the free radicals in the absence of trace iron. The presence of Cr(1V) as an intermediate was detected indirectly by reaction with Mn(I1) and subsequent loss of the Mn(I1) EPR signal. Although the presence of Cr(1V) was postulated in previous kinetic studies (12-13, this is the first report of the detection of Cr(1V) produced by reaction of Cr(V1) and ascorbate. A mechanism consistent with our EPR observations is discussed.

Experimental Section Chemicals. The Cr(V1)sourcewas KzCrzO,, the Cr(II1)source was Cr(NOs)s.SH20, and the Mn(I1) source was MnClz.4HzO (Fisher ScientificCo.,Pittsburg, PA). Chromium(VI)is a human carcinogenand should be handled with care. Ascorbatewas used as the sodium salt (Sigma Chemical Co., St. Louis, MO). Trace iron wasremoved from all buffer and ascorbatesolutionsby %fold treatment with Chelex-100 resin, sodium form, 100-200 mesh (Bio-RadLaboratories, Richmond, CA) unless otherwise noted. The purity of the ascorbate was verified by 1H and lSCNMR spectroscopyin DzO (18). The actual concentration of ascorbate stock solutions was determined spectrophotometrically at 265 nm (c = 14 500 M-1 cm-1) (19) before and after the EPR experiments. The Cr(V) complexes Na[CrO(EHBA)z] and KsCrOs were prepared by literature methods (20, 21). DMPO (Fluka,Ronkonoma,NY) and 2-methyl-N-(pyridinylmethylene)2-propanamine Nfl-dioxide (POBN) (Aldrich,Milwaukee, WI) were purified by solution filtration through activated charcoal (22). Dehydroascorbate (DHA) and 2-ethyl-2-hydroxybutyric acid (EHBA) were obtained from Aldrich, sodium formate and potassium phosphate were from Fisher Scientific Co., HEPES and dimethylarsinicacid (cacodylate)were fromSigma Chemical Co., and Tris.HC1 was from GIBCO BRL, Gaithersburg, MD. For reactions at pH 7.0 the pH of all stock solutions was checked and adjusted before final dilution if necessary. All solutionswere prepared immediately before use. EPR Spectroscopy. EPR spectra were recorded on a Bruker ESP-300 spectrometer with 100-kHz field modulation, 1.0 G

Stearns and Wetterhahn

modulation amplitude, 1 X lo6 receiver gain, 5.12-ms time constant, 9.772-9.769 microwave frequency, 20 dB (2-mW) microwave power, a 3430 to 3530 G sweep width, and 21-s scan time. Spectra were recorded at RT after 80 s or 18 h reaction time, as noted, and were signal averaged over nine scans for a 3.15 min total scan time. Spectra recorded in a flat cell were acquired with the above conditions,exceptfor a microwavepower of 27 dB (390 pW). Spectra in the presence of Mn(I1) were acquired under the standard conditions except the sweep width was expanded to 3480 f 100 G and only one scan was measured. A portion of each reaction solution (-60 pL) was drawn into an open-ended capillary tube (1.2 f 0.1 i.d.) and the bottom was sealedwith Dow-Corninghigh vacuum grease. Thegvalueswere calculated from calibration against 2,2-diphenyl-l-picrylhydrazyl radical (DPPH,g = 2.0036 f 0.0003) (23);however,since solution spectra of the 2,2,6,6-tetramethyl-l-piperidinyloxy (TEMPO) radical gave a g value of 2.007 compared to the literature value of 2.006 (24) the g values reported in this work were corrected for this difference. Reaction Solutions for EPR Spectroscopy. Reactionswere run with 9.0-9.2 mM Cr(V1) and 0.5-3 equivalents of ascorbate in 0.10M HEPES, potassium phosphate, cacodylate,or Tris.HC1 buffer (pH 7.0, RT). Spin traps were present at 0.10 M concentration. The concentrations of ascorbate radical and spin trapradical adducts were estimated by comparisonof the signal intensity (I@ where I = peak to trough intensity and AH = line width between maximum and minimum of the first derivative spectral peak) to that of solutions of known concentration of TEMPO radical (Sigma) (25). The concentration of Cr(V) was estimated by comparisonof I M to that of known concentrations of Cr(V) as KSCrOs (21) in 0.2 M NaOH, 0.5% HzOz. Spectra of standard samples were acquired under identical conditions as spectra of reaction samples. For reactions in the presence of manganese(II), Mn(I1) was used at the same concentration as Cr(V1). The Mn(1I) EPR signal completely masked the Cr(V) and radical signals. Control solutions are discussed in the text. +l/2 The six line Mn(I1) signal corresponding to the -l/2 transition (26)had a g value of 1.996and AM^ = 95.60 G. Signal intensity (I@) was compared for reaction solutions and the same concentration of Mn(I1) in buffer alone to calculate the concentration of Mn(I1) lost, which was then equated to the concentration of Cr(1V)produced as an intermediate. Reactions were initiated by addition of Cr(V1). The capillary was placed in a quartz tube before placement in the microwave cavity. Reproducibility was verified for all reactions. Reactions under Argon or Dioxygen Atmosphere. For experimenta carried out under argon or 0 2 , stock solutions were sparged for 15-30 min, and reactant solutions transferred by gas-tight syringe. Reactions were carried out under constant argonor Ozflow. For someexperimentssolutionswere transferred to a deoxygenated flat cell by syringe. For others, reaction solutions were drawn into sparged capillaries under positive pressure. For experimentstesting carbon-basedradical formation under argon, reactions were carried out under argon flow for 17 h, sample solutions were drawn, and spectra were acquired in air in capillary tubes. Reactions under argon were compared to reactions in air after the same reaction time. Consistent results were obtained using either the flat cell or the capillary tubes; however, EPR signalreproducibility was found to be better using capillaries than the flat cell.

-

Results Reaction of Cr(V1) with Ascorbate Produces Cr(V), Ascorbate Radical, COO-, and Other CarbonBased Radicals. In order to detect possible L(reactive intermediates” such as Cr(V) and radicals, the reaction of Cr(V1) (9.2 mM) with ascorbate (4.65-27.5 mM, 0.5-3 equiv) was monitored by EPR spectroscopy in 0.10 M HEPES buffer (pH 7.0, RT) in the presence of 0.10 M spin trap (DMPO or POBN). The EPR spectra obtained

Chem. Res. Toxicol., Vol. 7, No. 2, 1994 221

Cr(Vr) Reduction by Ascorbate

Table 1. Hyperfine Coupling Constants and g Values for DMPO/R’ and POBN/R’ Formed from Reaction of Cr(V1) and Ascorbate adduct AH (G) AN (G) g value relative amount

A

asc:Cr 3.1:l

~~~~~

DMPO/R’0

1.6:l

1.1:l

0.53:l

25.7

1 2 3

18.8

22.8

15.5 2.007 15.5 2.007 15.8 2.007 POBN/R’ b 15.82 2.005 14.70 2.005

1 0.40 0.35

2.69 1 0.315 6.26 0 9.0 mM Cr(V1) + 9.1 mM sodium ascorbate with 0.10M DMPO in 0.10 M HEPES (pH 7.0, RT), 80-6 reaction. Simulation shown in Figure 2. b 9.0 mM Cr(V1) + 9.0 mM sodium ascorbate with 0.10M POBN in 0.10 mM HEPES (pH 7.0,RT), 64-min reaction.

2

3430

3480 Gauss

major

minor

3530

B

I 3430

3440 3450 3460 3470 3480

I

3480 Gauss

3530

3490

3500 3510

Gauss

Figure 1. (A) EPR spectrafor reaction of potassium dichromate (9.2 mM Cr(V1)) with varying sodium ascorbate concentration (4.8-28 mM) in 0.10 M HEPES buffer (pH 7.0, RT) with 0.10M DMPO. Spectra were acquired after 80 s of reaction time at RT. Part B shows EPR spectra for time course of the reaction of dichromate [10.0 mM Cr(VI)]and sodium ascorbate (10.2mM) in 0.10 M HEPES (pH 7.0, RT) with 0.10M DMPO. Procedural details are given in the Experimental Section.

Figure 2. (Top)EPR spectrum of DMPO/R from the reaction of potassium dichromate (9.2mM Cr(V1))and sodium ascorbate (9.6 mM) in 0.10 M HEPES (pH 7.0, RT) with 0.10 M DMPO. Spectrawere acquired after 80sof reactiontime at RT. Procedural details are given in the experimental section. The bottom line is a simulated spectrum showing assignment of three carbon-based DMPO radical adducts. Hyperfine splitting constanta are listed in Table 1.

upon reaction of chromium(V1) with ascorbate in the presence of DMPO in HEPES showed the presence of ascorbate radical, chromium(V), and DMPO radical adducts (Figure 1A). The ascorbate radical anion (g = 2.006, AH = 1.82(8) G; lit. (27) g = 2.00518, AH = 1.76 G) was observed at ascorbate to chromium ratios of 3:l to 1.5:l and was not trapped by DMPO or POBN. The concentration of ascorbate radical estimated by comparison of IAW to that of a solution of the TEMPO free ’ and0.007radical corresponded to 1.0(2)pM, or 0.004(1)% (1)96 of the initial ascorbate concentration for the 3:l and 151ratios, respectively. Solutions of ascorbate alone in buffer showed insignificant amounts of ascorbate radical. Chromium(V) and the DMPO radical adducts were only observed at lower ratios of ascorbate to Cr(V1). The amount of Cr(V) (g = 1.980, AH = 1.17 G)increased as the ascorbate to Cr(V1) ratio decreased. The concentration of Cr(V) was estimated to be 0.014(6) mM, or 0.15(4) 96 of total Cr for the 1:l ratio and 0.017(6) mM, or 0.19(4) 96 for the 0.5:l ratio. The intensity of the Cr(V) signal was unaffected by the presence of spin trap (data not shown). Under these conditions the Cr(V) signal was too weak to detect 53Cr hyperfine coupling; however, at higher concentrations of 0.10 M Cr(V1) and 0.05 M ascorbate 53Cr hyperfine coupling was observed at 17.8(1) G which is in agreement with previous work (10). The 53Cr hyperfine coupling andg value are in the range expected for a square

pyramidal bis-diol-oxochromium(V)(CrV=O) species (28, 29). The Cr(V) species had a half-life on the order of 10-15 min at RT, with little or no Cr(V) detectable after 45 min (Figure 1B). The highest amount of DMPO radical adducts was observed at a 1:l ratio of ascorbate to &(VI). Three different DMPO radical adducts were distinguished. The g values and hyperfine coupling constants were calculated from simulations and are listed in Table 1, and the simulation is shown in Figure 2. The relative magnitudes of AH > ANare expected for carbon-based DMPO-radical adducts (30). The DMPO adduct listed as 2 was assigned as the C02- adduct based on the reported spectral data, g = 2.0058, AH = 18.7 G, AN = 15.6 G (31). The DMPO radical adducts were not stable, having half-lives of 15 min at RT, and little or no adducts were detectable after 45 min (Figure 1B). The total concentration of radicals trapped by DMPO at the 1:l ratio corresponds to 0.2(1) % of the initial ascorbate concentration,or 0.02(1) mM, based on comparison of IA@ with that of the TEMPO free radical. It should be noted that the amounts of carbonbased radical adducts are affected by the trapping efficiency of DMPO and POBN and are not necessarily a measure of the actual amounts of radicals formed. The EPR spectra obtained for the reaction of Cr(V1) with ascorbate in the presence of POBN in HEPES showed the same types of intermediates as were observed in the

-

222 Chem. Res. Toxicol., Vol. 7, No. 2, 1994

Stearns and Wetterhahn A

asc:Cr

10.0

1.5:l

1.O:l 4.0 3.0 3430

3480

3530

Gauss

B 6 h,.

l h

I

3430

3480

3530

Gauss Figure 3. (A) EPR spectrafor reaction of potassium dichromate [9.2 mM Cr(VI)] with varying sodium ascorbate concentration (4.6-28 mM) in 0.10 M HEPES buffer (pH 7.0, RT) with 0.10 M POBN. Spectra were acquired after 80 s of reaction time at RT. Part B showsthe EPR spectraover time for reaction of potassium dichromate [9.2 mM Cr(VI)]and sodium ascorbate (9.3 mM) in 0.10M HEPES (pH 7.0, RT) with 0.10 M POBN spin trap. Bars at 1 h mark the minor POBN radical adduct that was formed more slowly and was less stable than the major radical adductb). Hyperfine splitting constants are listed in Table 1. Procedural details are given in the Experimental Section. presence of DMPO, namely ascorbate radical, Cr(V), and POBN radical adducts (Figure 3A). As was observed in the presence of DMPO, the optimal ascorbate to Cr(V1) ratios for POBN free-radical adducts and Cr(V) were 1:l and 0.5:1, respectively. However,at the 1:land 0 5 1ratios of ascorbate to Cr(V1) apparently one major POBN radical adduct signal, or more than one adduct with overlapping signals, was observed initially at 0.02 mM, or 0.2(1) % of initial ascorbate concentration. The POBN radical adduct(@ were quite stable, being detectable even after 23 h. The asymmetry of the POBN adduct signals suggested the presence of multiple adducts, and this assumption is also supported by observation of three radical adducts in the presence of DMPO. A second minor POBN radical adduct signal was distinguished at 12 min and had disappeared by 6 h (Figure 3B). Control reactions of Cr(V1) with spin trap and ascorbate with spin trap over this time range showed no POBN radical adducts. The spectrum of the POBN adduct signals at 64 min was simulated with the g values and hyperfine coupling constants listed in Table 1. The major adduct was assigned

I

I

0.0

5.0

I

I

10.0

15.0

I

I

20.0

25.0

30.0

Concentration of Ligand (mM)

Figure 4. Intensity of the Mn(I1) EPR signal [9.0 mM Mn(1I)I in the presence of (0)0-27 mM ascorbate, (A)0-27 mM DKG, (m) 0-27 mM ascorbate + 9.0 mM Cr(V1).The loss of the Mn(1I) signal was converted to concentration by comparison with Mn(I1) standards and was equated with the presence of Cr(IV) intermediate, n = 3-6. The concentration of Cr(IV) as percent total chromium concentration is listed in Table 2. Spectra were acquired after 80 s of reaction time at RT. Procedural details are given in the Experimental Section.

as including the POBN/C02- adduct from comparison to literature values of AH = 3.0, AN = 15.5 (30),and because DMPO/COz-was observedin reaction solutions containing DMPO (see above). Reaction of Cr(V1) with Ascorbate Produces Cr(1V). Cr(1V) has been implied to be formed as a “reactive intermediate” in the reaction of Cr(V1) with ascorbate, and therefore, the possible production of Cr(V1) in this reaction was examined. Cr(1V) was detected indirectly by reaction with Mn(I1) (eq 1) (32): Cr(1V) + Mn(I1)

-

Cr(II1)

+ Mn(1II)

(1)

Reaction of 9.0-9.2 mM Cr(V1) with 4.5-27 mM ascorbate in the presence of 9.0-9.2 mM Mn(I1) in 0.10 M HEPES buffer (pH 7.0, RT) in the presence or absence of 0.10 M DMPO resulted in a measurable loss of the Mn(I1) EPR signal relative to control solutions. The Mn(I1) signal intensity decreased with increasing ascorbate concentration in reaction samples but was not affected by the aame concentrations of ascorbate in the absence of Cr(V1) (Figure 4). In order to determine whether the decrease in Mn(I1) was specific for Cr(IV), a variety of control reactions of Mn(I1) with the starting materials [Cr(VI) and ascorbate], and reaction products [Cr(III),DHA, and 2,3-diketogularicacid (DKG,the open-ringform of DHA)], were examined. The Mn(I1) EPR signal was not affected by 0.10 M DMPO or 27 mM DHA (data not shown) but the signal decreased slightly with increasing concentration of DKG (33)(Figure 4). Control reactions were measured on solutions of Mn(1I) with 0-27 mM DKG even though the maximum amount of DKG that could be formed in reaction solutions is 1.5 equiv, or 13.5 mM. Chromium(V1) alone showed no reactivity with Mn(I1) under these conditions, nor did 9.0 mM Cr(II1) incubated with 27 mM ascorbate, DHA or DKG (Table 2). The “reactive intermediates”besides Cr(IV) which could react with Mn(I1) are Cr(V), carbon-based radicals, and ascorbate radical. The Cr(V) complex Na[CrO(EHBA)d (20) was included as a test for Cr(V) reactivity with Mn(I1) (Table 2). The Cr(V)EHBA complex is known to

Chem. Res. Toxicol., Vol. 7, No. 2, 1994 223

Cr( VI)Reduction by Ascorbate Table 2. Relative Amount of Cr(1V) Formed in Reaction of Cr(V1) w i t h Ascorbate or Controls Estimated from Reaction w i t h Mn(II)*

% Cr(1V)

sample Cr(V1)b [Cr(V)O(EHBA)$ Cr(II1) + 3ascHCr(II1) + 3DHAf Cr(II1) 3DKGe Cr(V1) + 0.5ascHCr(V1) + 1.OascHCr(V1) + 1.5ascHCr(V1) + 3.0ascH-

+

5.6 f 0.4 11.4 0.1 nde 0.74 f 0.03 nd 20f 1

*

*

24f3 36 f 5 49 f 4

a Calculated from decrease in EPR signal intensity (ZIAZP)of 9.0 mM MnClp4HzO in the presence of potassium dichromate and sodium ascorbate, or controls, 80-s reaction time. Reported as % of total chromiuminsample(n = 3-10). Cr(VI)at9.0mMinO.lOMHEPES (pH 7.0,RT). Stock solution Cr(V) in HzO added to Mn(I1) solution in 0.20M HEPES for final solution concentrations of 9.0 mM Cr(V) and 9.0 mM Mn(I1) in 0.10 M HEPES (reaction pH 6.6, RT). Signal intensity corrected for loss due to 18 mM EHBA (see text). Cr(NO3)rSHzOa t 9.0 mM, ascorbate a t 27 mM in 0.10 M HEPES (pH 7.0,RT), 1h reaction before Mn(I1) addition. e Decrease in signal not detected. f Cr(NO&SHZO a t 9.0 mM, DHA a t 27 mM in 0.10 M HEPES (reaction pH 6.0, RT), 4-minreaction before Mn(I1)addition. No estimation or correction was made for conversion of DHA to DKG. 8 Cr(N03)~9HzOa t 9.0mM, DKGat 27 mM inO.10 M HEPES (pH 7.0,RT), l-h reaction before Mn(I1) addition, signal corrected for 3 DKG. Reaction ratio of 9.0 mM Cr(V1) + 0-27 mM ascorbate in 0.10 M HEPES (pH 7.0,RT). Signal intensity was corrected for stoichiometric equivalent of DKG.

be stable only in the presence of excessligand under acidic conditions and undergoes disproportionationat pHs above 3-4 (20). A UV/vis study of the stability of 10 mM Na[ CrO(EHBA)z] complex showed an immediate decay in 0.10 M HEPES at pH 7.0, a 6% decay over 35 min for a solution in 0.10 M EHBA adjusted to pH 7.0 and a 7% decay over 35 min for a solution in HzO. For the current EPR study the Cr(V)EHBA complex was tested in the absence of excess ligand for two reasons. First, reaction of Mn(I1) with EHBA alone in solutions adjusted to pH 7.0 showed an EPR signal decrease proportional to EHBA concentration (discussed below). Second, testing the reaction of the Cr(V)EHBA complex with Mn(I1) under conditions which deviate from the Cr(VI)/ascorbate reaction and are known to stabilize Cr(V) defeat the purpose of using it as a control for reactivity. A freshly prepared 9.0 mM solution of Na[CrO(EHBA)zI in HzO showed a decrease in the Mn(I1) EPR signal of 19.896, whereas Mn(I1) with 2-fold EHBA at pH 7.0 showed a decrease of 8.4 % . The decrease of the Mn(I1) signal in the presence of EHBA ligand could be due to three possibilities. Coordination of Mn(I1)by EHBA could produce a lowering of symmetry about Mn(I1) resulting in zero-field splitting (34) and a quenching of the EPR signal, or coordination by EHBA could favor formation of EPR silent antiferromagnetically coupled di-p-hydroxo-Mn(I1) dimers (35, 36). Coordination of Mn(I1) by EHBA may be enhancing Mn(I1) reactivity with 02 to yield EPR silent Mn(1V) or Mn(II1)complexes (35,37,38). The later explanation was discarded for the following reasons: solutions of Mn(I1) with 2-fold EHBA showed the same decrease in the Mn(I1) EPR signal under argon as in air. Mn(II), a d5 ion, has essentially no visible absorbance, whereas Mn(II1)with polyhydroxo ligands shows d-d transitions with E = 100300 M-l cm-1 (35,37,38). The UV/vis spectra of 9.0 mM Mn(I1) in the presence of either 2-fold or 20-fold EHBA (adjusted to pH 7.0 in 0.10 M HEPES) showed no visible absorbance, confirming the absence of detectable

Mn(II1). Air oxidation of Mn(I1) was observed by UV/vis in the presence and absence of EHBA ligand only when the pH was raised to 13. The decrease of the Mn(I1) EPR signal (11%when corrected for the effect of the EHBA ligand) in the presence of [Cr(V)O(EHBA)zI- could be due to at least two possibilities: reactivity with Cr(V), or more likely the known disproportionation of two Cr(V) to Cr(V1) and Cr(1V) (39) and subsequent reaction with Cr (IV). On the basis of the Cr(V)EHBA control experiment it does not appear that reaction of Mn(I1) with Cr(V) could be the major pathway for the loss of the Mn(I1)EPR signal. The maximum amount of “Cr(V) ascorbate” observed in reaction solutions was 0.01-0.02 mM. Reaction of 9 mM Mn(I1) with 9 mM Cr(V)EHBA in the absence of competing reductant for Cr (V)produced only an 11% decrease in the Mn(I1) signal. It seems unlikely that reaction of 9 mM Mn(I1) with “Cr(V) ascorbate” in the presence of excess ascorbate which would compete for reduction of Cr(V)could account for the -50% loss of the Mn(I1) signal. If “Cr(V) ascorbate” were reactive with Mn(I1) then the loss of the Mn(I1) EPR signal would be expected to be greater at lower ascorbate concentrations; however, the loss of the signal was greater at higher ascorbate concentrations. Another observation that fails to support Cr(V) reactivity was that the same loss of the Mn(I1) EPR signal was found when Cr(V1) was reacted with ascorbate in Tris.HC1 buffer, even though little or no “Cr(V) ascorbate” was observed in Tris.HC1 buffer in the absence of Mn(I1) (vide infra). It also seems unlikely that the carbon-based radicals were involved in the decrease in the Mn(I1) EPR signal. If the radicals were reacting with Mn(I1) then addition of DMPO should have resulted in less of a decrease in the Mn(I1) signal since DMPO would be competing with Mn(1I)for the radicals. The presence of DMPO in reaction solutions had no effect on the Mn(I1) signal. By the same argument presented for Cr(V) above, if Mn(I1) were reacting with the carbon-based radicals then there should be less of a decrease of Mn(I1) at high ascorbate concentrations where ascorbate and Mn(I1) would be competing for the carbon-based radicals. The trend for greater loss of the Mn(I1) signal a t higher ascorbate concentration does fit the trend observed for the ascorbate radical signal. The reactivity of Mn(I1)with ascorbate radical cannot be directly measured; however, an argument against oxidation of Mn(I1) by ascorbate radical can be made from the reduction potentials of about -1.6 V for Mn(III)/Mn(II) (32)and 0.320 V for asc-/ascH(40). Coordination of ascorbate radical to Mn(I1) to produce zero-field splitting is not likely because ascorbate did not affect the Mn(I1) signal (Figure 4), and reaction of 9 mM Mn(I1) with 9 mM TEMPO radical showed only a 5.0 f 5.0%loss in signal. This was not surprising since the coordination of Mn(I1) with nitroxyl radical ligands has been reported to affect the line width of the radical EPR signal, but not that of Mn(I1) (41). On the basis of these observations it is not unreasonable to surmise that ascorbate radical is not responsible for the loss of the Mn(I1) signal. Assuming that the loss of the Mn(I1) EPR signal was due to reaction with Cr(IV), the amounts of Cr(1V) measured in reactions of Cr(V1) with varying ascorbate are listed in Table 2. The results clearly show that the level of Cr(1V) formed in the reaction of Cr(V1) with

224

Chem. Res. Toxicol., Vol. 7,No.2, 1994 0 Argon

e H

2,

5000

0 Oxygen

-

I

.-x

4000

CI

v1

3 3000

-1

I

M

2000

in' 1000

i 0

6

0

U

I

I

1

I

I

I

10

20

30

40

50

60

70

Time (min) 3420

3450

3480 Gauss

3510

3540

Figure 6. Effect of buffer on the EPR-detectableintermediates for the reaction of potassium dichromate and sodium ascorbate: (a) 0.10 M HEPES buffer, 9.2 mM Cr(VI),9.6 mM ascorbate; (b) 0.10 M phosphate buffer, 8.6 mM Cr(VI),8.6 mM ascorbate; (c) 0.10 M cacodylatebuffer, 9.6 mM Cr(VI),9.4 mM ascorbate;and (d) 0.10 M Tris.HC1 buffer, 9.0 mM Cr(VI),9.4 mM ascorbate. All buffers were pH 7.0, RT, and all reaction solutionscontained 0.10 M DMPO. Spectra were acquired after 80 s of reaction time at RT. Procedural details are listed in the Experimental Section.

ascorbate increases with increasing ascorbate concentration, suggesting a predominate two-electron pathway of Cr(V1) reduction by ascorbate. Buffer-Dependenceof the Reaction of Cr(V1)with Ascorbate. In order to show that the carbon-centered radicals detected in the reaction of Cr(V1) with ascorbate were derived from ascorbate and not the buffer, the reaction was studied using a variety of organic and inorganic buffers. When 0.10 M Tris-HC1buffer (pH 7.0, RT) was substituted foro. 10M HEPES buffer in the above reaction, ascorbate radical and carbon-based radicaladducts were detected but the amount of Cr(V) was significantly decreased (Figure 5a,d). The 1:l ascorbate to Cr(V1) reaction in 0.10 M Tris.HC1 showed the same three DMPO adducts as were observed in 0.10 M HEPES (Figure 5a,d). The same amounts of DMPO radical adducts were observedin both buffers, within experimental error (0.02 mM, or 0.2% of ascorbate concentration). We (11) have previously shown that reaction of Cr(V1) and ascorbate in Tris.HC1buffer gave an EPR spectrum with lower levels of Cr(V) relative to reaction in HEPES buffer and a shift in the g value for Cr(V) from 1.980 (HEPES) to 1.978 (Tris.HC1). The reaction carried out in either potassium phosphate or cacodylate buffer (0.10 M, pH 7.0, RT) produced the same Cr(V) and DMPO radical adducts signals in the same relative amounts as were observed in HEPES buffer (Figure 5b,c) except for the addition of a minor Cr(V) signal in cacodylate buffer with a g value of 1.983, AH = 1.17 G (Figure 5c). This suggests that the Cr(V) species in HEPES or phosphate buffer is not coordinated by the buffer, whereas the Cr(V) species in Tris.HC1 is associated with buffer, and cacodylate buffer shows a minor interaction. The presence of the same DMPO radical adducts in different buffers is evidence that the carbon-based radicals were produced by fragmentation of ascorbate by chromium(V1) and not a reaction involving buffer. Effect of Dioxygen on the Reaction of Cr(V1) with Ascorbate. The involvementof dioxygenin the reduction

Figure 6. EPR signal intensity of Cr(V) species from reaction of potassium dichromate [9.0 mM Cr(V1)I and sodium ascorbate (9.0 mM) in 0.10 M HEPES (pH 7.0, RT) over time. Reactions were carried out under argon (0)or dioxygen (0).Procedural

details are listed in the Experimental Section.

of Cr(V1) by ascorbate would have important implications for the metabolism of Cr(V1) in vivo. For this reason several experiments were carried out to explorethe possible reactivity of 02 with intermediates produced in this reaction. The first step was to determine if reactive oxygen species were being scavanged by the spin traps. Under no conditions was the *OHadduct of DMPO (42)or POBN (43)observed, and addition of the singlet oxygen trap 2,2,6,6-tetramethyI-4-piperidone hydrochloride (44)showed no formation of the TEMPO radical; therefore, no direct evidence for reactive oxygen species was found in this system. Unfortunately, a negative result is inconclusive. The reduction of TEMPO radical by ascorbate is not thermodynamicallyfavored (401,and therefore, occurs very slowly. The reported reaction rate was 7.9 p M min-l(45). Thus, reactivity of TEMPO with ascorbate cannot be used to explain the lack of a TEMPO radical signal. One other possible explanation for this result is that if the carbonbased radicals and '02 were formed by independent or competing pathways, the TEMPO radical EPR signal could be quenched by reaction of TEMPO with carbonbased or ascorbate radicals (eqs 2 and 3) (46,47): (TEMP0)N-0' (TEMPOIN-0'

-

+ R'

+ asc*--

(TEMP0)N-0-R

(TEMP0)N-OH

+ DHA

(2)

(3)

Therefore the involvement of reactive oxygen species was explored further. Because the previous study (13) suggested that reactive oxygen species were produced by reaction of Cr(V) with dioxygen we attempted to test this hypothesis. Stock solutions of Cr(V1) and ascorbate were sparged for 15-30 min with either argon or pure dioxygen. Chromium(V1) was added to ascorbate by gas-tight syringe and reactions were continuously sparged over 60 min while aliquota were taken for EPR measurement. EPR samples were drawn into sparged capillary tubes under positive pressure, and the decay of the Cr(V) signal was monitored over time. The Cr(V) signal under 02 had a AH(pp) of 1.37 G us 1.17 G for that under argon. The signal intensity calculated as I A P showed no difference in the rate of decay of Cr(V) under 0 2 or argon (Figure 6), and thus, it was concluded that Cr(V) from the reaction of Cr(V1) and ascorbate does not react with 02,

Chem. Res. Toxicol., Vol. 7, No. 2, 1994 226

Cr(Vr) Reduction by Ascorbate I

3440

I

I

I

3460

3480 Gauss

3500

I

1-

3520

3430

Figure 7. Effect of dioxygenon the carbon-basedradicalsformed in the reaction of potassium dichromate 19.0 mM Cr(V1)I and sodium ascorbate (9.0 mM) in 0.10 M HEPES buffer (pH 7.0, RT) with 0.10 M POBN,after 18 h (a) reaction in air and (b) reaction under argon. Procedural details are given in the Experimental Section. The possible 0 2 dependence for the formation of the carbon-based radicals was also tested. This reactivity was explored by taking advantage of the stability of the POBN radical adducts. The reaction of Cr(V1) (9.0 mM) with ascorbate (9.0-9.2 mM) was carried out in 0.10 M HEPES buffer with 0.10 M POBN, in air and under argon for 18-h at RT. During the 18 h reaction under argon the Cr(V) and Cr(1V) species were completely reduced to Cr(II1) without exposure to dioxygen. There was no difference in the amounts of POBN adducts observed between reactions in air or under argon (Figure 7). The concentrations of POBN adducts ranged from 12 to 13 pM. The reaction of Cr(V1) with ascorbate under argon showed that the carbon-based radicals were formed by an oxygenindependent pathway. Role of Trace Iron in the Cr(VI)/Ascorbate Reaction. If Cr(V) does not react with 0 2 and the carbonbased radicals were formed in the absence of 0 2 , then it was necessary to explain how the previous study (13) observed an increase in DMPO/C02*- for reactions of Cr(V1)and ascorbate in the presence of formate. Formate is known to trap *OHgenerated by activated dioxygen, producing CO2- by hydrogen atom abstraction (48).On the basis of the above experiments evidence pointed to trace iron in the previous study (13) as a way to explain an increase in the DMPO/C02*-signal in the presence of formate and 0 2 . The reactivity of trace iron was explored under conditions identical to the prevous study (13), namely 10 mM Cr(VI), 10 mM ascorbate, 50 mM DMPO, fl.O M formate in 0.10 M phosphate buffer, pH 7.4 at 37 "C. Stock solutions were treated or untreated with Chelex resin. Reactions were carried out in air at 37 "C for 10 min and then transferred to a flat cell for acquisition. Reactions in the absence of formate showed the same Cr(V) and DMPO/R spectra as were observed in capillary tubes whether or not solutions had been treated for iron removal, although the signal intensities were lower here than reported above due to further extent of reaction. The addition of formate showed an increase in DMPO/ COf- as was reported (13) with a slightly greater signal observed for untreated solutions. The most important spectrum, however, was that of the control solution of 10 mM Cr(V1) + 1.0 M formate + 50 mM DMPO which showed a strong DMPO/*OHsignal for untreated solutions

I

I

I

I

I

I

3445

3460

3475 Gauss

3490

3505

3520

Figure 8. Effect of 1.0 M formate on carbon-baeed DMPO adducts formed in the reaction of potassium dichromate L9.0 mM Cr(VI)] and sodium ascorbate (9.0 mM) in 0.10 M HEPES buffer (pH 7.0, RT) with 0.10 M DMPO. Reactions were carried out at RT under argon and acquired in a flat cell, total reaction time of 2 min: (a) reaction in the absence of formate and (b) reaction in the presence of 1.0 M formate. Formate and ascorbate solutions were treated for iron removal. Procedural details are given in the Experimental Section.

us no signal for chelex-treated solutions. This EPR signal in control solutions was also reported in the previous paper (13). On the basis of these and the above experiments it was concluded that the 0 2 dependence reported in the prevous paper (I3 )was due to trace iron, most likely coming from formate. In light of these results it is difficult to evaluate the recent kinetic study (12) in which removal of 02 resulted in a 10-fold increase in the rate of reduction of Cr(V1) by ascorbate. Because no effort was made to remove iron from the buffers it is not clear whether the increase in the rate of Cr(V1) reduction was aided by reaction of Fe(I1) with Cr(V1) (49). The standard reaction of 9.0 mM Cr(V1) with 9.0 mM ascorbate in the presence of 0.10 M DMPO f 1.OM formate in 0.10 M HEPES at pH 7.0, with a reaction time of 2 min under argon at R T gave a flat-cell EPR spectrum showing DMPO/R (Figure 8a), in the absence of formate, and DMPO/R and DMPO/C02- (Figure 8b), in the presence of formate. The formation of DMPO/C02-in the absence of 02 suggests that the carbon-based radicals react with formate to produce COa-. Discussion Reaction of Cr(V1) with ascorbate resulted in different EPR-detectable intermediates depending on the reaction ratio. Chromium(V) and carbon-based radicals were observed at low ratios of ascorbate to Cr(VI),and ascorbate radical was observed under conditions of excessascorbate. Chromium(1V) was detected indirectly by reaction with Mn(II), and amounts measured were proportional to ascorbate. These chromium and carbon-based radical intermediates may be responsible for the genotoxicactivity of carcinogenicchromium(V1)compounds which has been observed in vitro and/or in vivo. Chromium(V) Complexes. The g value of 1.980 for the Cr(V) complex formed upon reaction of Cr(V1) and ascorbate in HEPES and potassium phosphate buffers suggests a square pyramidal, axial oxo coordination mode (28). The 63Cr hyperfine coupling of 17.8(1) G suggests bis-diol coordination (29). The signal was too weak to

Stearns and Wetterhahn

226 Chem. Res. Toxicol., Vol. 7, No. 2, 1994

Scheme 1

observesuperhyperfine coupling (29). The shifting value to 1.978 for the Cr(V) species in TriwHC1(10, this study) and the additional Cr(V) signal at g = 1.983 in cacodylate buffer suggest buffer interaction, although in cacodylate buffer this interaction is minor. This shift ing value from 1.980 to 1.978 upon buffer coordination in TriwHC1 is consistent with that previously observed for increases in coordination number of Cr(V) complexes of oxalic acid (a shift from 1.9766 to g = 1.9714) (50) and EHBA (a shift from g = 1.9785 t o g = 1.9714) (51). Likely coordination environments for the Cr(V) species in the noninteracting buffers are shown in Scheme lA, and that in Tris.HC1 in Scheme 1B. The coordination of Cr(V) by Tris buffer to give a 6-coordinate oxo species is consistent with the lower amount of Cr(V) observed in this buffer, presumably because a 6-coordinate Cr(V) species would be less stable toward reduction to 6-coordinate Cr(II1) (32). It is becoming well accepted that Cr(V) is produced during Cr(V1) metabolism and likely plays a role in some types of DNA damage. Sugiyama and co-workers have observed a Cr(V) species with a gl value of 1.989 after treatment of Chinese hamster V-79 cells with Cr(V1) (52). Formation of Cr(V) in V-79 cells has been correlated with Cr(V1)-induced DNA alkali-labile sites (8, 53, 54) and chromosomal aberrations (55,56). Chromium(V) with a g value of 1.987 has been observed in 14-day chick embryo liver after in vivo treatment with Cr(V1) at nontoxic doses (57). The formation of Cr(V) (57) has been linked with observation of DNA-protein cross-links (58) and Cr-DNA adducts2 in chick embryo liver. Thus, the stability of Cr(V) in vivo will likely be a function of the relative intracellular concentration of reducing agents and will affect the types of DNA damage observed. Indirect Detection of Cr(1V). Chromium(1V) has been proposed to be an intermediate in the reduction of Cr(V1) by inorganic and organic substrates (32,59). The classic test consists of observation of the effect of Mn(I1) on the rate of Cr(V1) reduction (32)where removal of Cr(1V) by reaction with Mn(I1) depletes substrate for Cr(VI), thus inhibiting the rate of Cr(V1) reduction. Recently, Cr(1V) species have been detected in a number of experiments. The most thoroughly studied is the relatively stable Cr(1V) complexof EHBA, which has been observed by UV/vis spectroscopy (60-63). Stability is conferred by the presence of excess ligand at low pH. This Cr(1V) complex has recently been proposed to be more reactive than Cr(V) toward nicking of plasmid DNA (64). 2

Misra, M. and Wetterhahn, K. E. Unpublished results.

Chromium(1V) has also been shown to be a long-lived intermediate in the reduction of Cr(V1) with glutathione by solution magnetic susceptibility measurements and HPLC detection (65). The aqua Cr(1V) oxo dication has recently been prepared and its oxidation of organic substrates has been studied (66). We report here the detection of Cr(1V) formed during the reduction of Cr(V1) by ascorbate in the presence of Mn(I1). This method allows for an estimation of the relative amounts of Cr(1V)formed under different reaction ratios of ascorbate and Cr(V1);however, it does not allow quantitation of the total Cr(1V) formed. The effect of Mn(I1) on the rate of Cr(V1) reduction has not been measured here. The amount of Cr(1V) was determined by measurement of the decrease in the Mn(II) EPRsignal. The Mn(I1) EPR signal was not significantly decreased by Cr(V1) at a rate that would compete with that of the Cr(V1) and ascorbate reaction, nor was the signal affected by Cr(III), ascorbate or DHA (Tables 1and 2). There was a slight loss of the Mn(1I) signal in the presence of ligands DKG and EHBA alone, which has been interpreted to be due to coordination of Mn(I1) to form EPR silent species rather than oxidation to Mn(II1). It was assumed that the slight reactivity of Cr(II1) with 3-fold DHA was due to some conversion of DHA to DKG. The Cr(V) complex [CrO(EHBA)& was included as a control; however, this was complicated by the known instability of this complex in the absence of excessligand (20)and reactivity of EHBA with Mn(I1). There are other reactions that may occur with rate constants on the order of that of Cr(1V) with Mn(I1) such as (eqs 4-6): Cr(V1) + Cr(1V) 2Cr(IV) Cr(1V)

-

-

2Cr(V)

(4)

Cr(II1) + Cr(V)

(5)

+ ascH--

Cr(II1) + asc*-

(6)

These would cause an underestimation of the total amount of Cr(1V) formed but would still allow for determination of the relative amounts. There has been a report (67) of the reduction of an EHBA-coordinated complex of Cr(1V) by Mn(I1) in which the Mn(I1) is catalytic; however, this result is not inconsistent with our method because the previous study stated that catalysis was only important at high concentrations of Mn(I1) relative to Cr(1V) a t pH 3.7, whereas this study was conducted at 1:l Mn(I1) to chromium ratio at pH 7.0. The Mn(II1) in this study may be reduced back to Mn(I1) by any of the following reactions (eqs 7-9): Mn(II1)

+ ascH--

Cr(1V) + Mn(II1)

-

2Mn(III) -MnO,

Mn(I1) + asc*Cr(V)

+ Mn(I1)

+ Mn(I1)

(7) (8)

(9)

However, if the re-formed Mn(I1) was EPR silent, as for example, a monomeric species with zero-field splitting (34) or a hydroxy-bridged dimer (35,36)then the decrease in EPR signal would still be a measure of reaction of Cr(1V) with Mn(I1) only. If the reformed Mn(I1) was not EPR silent, then the result would again be an underestimate of levels of Cr(1V). This is the first example of indirect detection of Cr(1V) using EPR spectroscopy, and the first observation of

Chem. Res. Toxicol., Val. 7,No.2, 1994 227

Cr(Vl) Reduction by Ascorbate Cr(1V) as the major intermediate formed during the reduction of Cr(V1) by ascorbate. This technique may prove useful for estimation of the relative amount of Cr(1V) intermediates under different reaction conditions. Detection of Carbon-Based Radicals. Reaction of Cr(V1) with ascorbate a t equimolar concentrations or excess Cr(V1) resulted in formation of carboe-based radicals as reactive intermediates. The observation of carbon-based radicals was independent of the buffer. From comparison to literature values (30)the adducts DMPOI *OH(AH = 14.9 G , AN = 14.9 G ) ,DMPO/H (AH = 22.5 G, A N = 16.6 G ) ,POBN/*OH,(AH = 1.7 G,AN 15.0 G ) , and POBN/H (AH = 10.2 G, AN = 16.2 G ) can be ruled out. The spectrum of the adduct DMPO/H would include a signal at -3440 G which was not observed (Figure 1A and 2). The assignment of the DMPO adduct 2 (AH = 18.8 G, AN = 15.5 G ) as the COz- adduct is also based on literature values (AH = 18.7 G, AN = 15.6 G ) (30).The COz- is probably a product of the cleavage of ascorbate at the C1 position. The DMPO adduct 3 (AH = 22.8 G , AN = 15.8G )has splitting constants equal to those observed for the DMPO adducts of hydroxyalkyl fragments .C(OH)zCH(OH)CH&H and *C(OH)2CHaOH[AH= 22.6 G ,AN = 15.8G (68)lwhich would be consistent with possible types of ascorbate cleavage products. The carbon-basedradicals forming DMPO adducts 2 and 3 would presumably be trapped by POBN to give superimposed spectra represented by the major POBN adduct (Figure 3,parts A and B) since POBN adducts of hydroxyalkyls and COZ- give similar coupling constants (30)and since POBN has been observed to show a higher trapping efficiency for COzthan DMP0.3 The DMPO radical adduct 1 (AH = 25.7 G, A N = 15.5 G)showed a larger 8-H hyperfine coupling constant than is precedented for known classes of DMPO radical adducts (301,and thus, its identity is not clear. The minor POBN adduct that grows in after 1 h of reaction time could be a radical species produced in a decay step with a chromium intermediate, or it could result from a rearrangement of an ascorbate fragment bound to POBN. This minor POBN adduct also has a 8-H hyperfine coupling constant greater than any previously reported (30).The rate of decay of the signal intensity of both the major POBN adduct and Cr(V) could account for the increase in the minor adduct's signal after 1 h of reaction. Attempts to isolate and/or distinguish the POBN adducts by the chromatographic methods of Ortiz de Montellano and co-workers (69) were unsuccessful, due to the small amounts of radical adduct present. Carbon-based radicals have been implicated in many mechanisms of DNA damage, for example, those of the carcinogenic hydrazine Phenelzine (701, the antitumor and certain agents neocarzinostatin and calicheamicin (71), Fe(I1) chelates used as antimalarial drugs (72).A carbonbased radical formed from 2,2'-azobis(2-amidinopropane) hydrochloride has recently been shown to cause single strand breaks in plasmid DNA (73).While the observation of carbon-based radicals formed by Cr(V1) metabolism in vivo would be hindered by the instability of the spin traps and/or spin trap adducts, the present Cr(VI)/ascorbate model system allows for evaluation of in vitro carbonradical reactivity with DNA. Preliminary results suggest that these carbon radicals do indeed cause single strand breaks in pBR322 DNA (11). 8

Yuann, J-.M. and Wetterhehn, K. E. Unpublished results.

Scheme 2 Excess Ascorbate

Excess Chromium(VI)

Role of 0 2 in the Cr(VI)/Ascorbate Reaction. Observation of carbon-based radicals in the reaction of Cr(V1)with ascorbate under our conditions contrasts with the results of Lefebvre and PBzerat (13)who detected Cr(V) and ascorbate radical as the only EPR active intermediates for the 1:l reaction of 10 mM Cr(V1) and ascorbate in 0.10 M phosphate buffer (pH 7.4)at 37 "C, and observed the DMPO/COz-adduct only in the presence of 1M formate. We believe these discrepancies are due to differences in reaction time and temperature and trace iron. The EPR spectra of Lefebvre and PBzerat (13)were obtained after 10 min of reaction at 37 "C, whereas the spectra in this study were acquired after 80 s a t RT. The half-lives for the DMPO radical adducts and Cr(V) at RT were 10and 15min respectively (Figure 1B)and the signals were barely detectable after 1 h. Our results suggest that the observationof DMPO/COz-in the presence of formate is due to carbon-based radicals rather than *OHreacting with formate. We have found that Cr(V) did not react with 0 2 , that the carbon-based radicals were produced by an oxygen-independent pathway, and that solutions of Cr(V1) and DMPO without removal of trace iron showed DMPO/'OH in the presence of formate. In solutions treated for trace iron no DMPO/.OH was detected in reactions and/or controls, and the TEMPO radical precursor did not trap singlet oxygen; therefore, we cannot invoke involvement of reactive oxygen species for any major pathway of this reaction. Mechanism for the Reduction of Cr(V1) by Ascorbate. Previous kinetic studies have shown that the reaction of Cr0d2-Withascorbate was first order in Cr(V1) and ascorbate (13-17) The kinetic study run at physiological pH (1M TriwHCl, pH 7.4)determined a secondorder rate constant of 36 f 1 M-1 min-' (16). Scheme 2 outlines a mechanism that is consistent with the previous studies (12-17)and the EPR results presented above. It is proposed that the reduction of Cr(V1) by ascorbate is a two electron process (eq 10) because the amount of Cr(1V)was much higher than the amount of Cr(V)detected in these studies, and a two-electron transfer would be expected for a cis-coordinating diol (74, 75). If the reduction does occur by a one electron pathway (eqn 17-

.

19):

Stearns and Wetterhahn

228 Chem. Res. Toxicol., Vol. 7, No. 2, 1994

+ ascHCr(V) + ascHCr(1V) + ascHCr(V1)

-

-

+ asc'Cr(1V) + ax*Cr(II1) + Cr(V)

-

(17) (18) (19)

then the reaction of Cr(V) with ascorbate must be faster than the reaction of Cr(1V) with ascorbate to account for the predominance of Cr(1V) trapped by Mn(I1) and absence of a Cr(V) signal at high ascorbate to Cr(V1) ratios. This relative reactivity was observed by Ghosh et al. (76) who showed that the Cr(V) complex of EHBA reacted with ascorbate faster than the Cr(1V) complex by a factor of 65 between pH 3.3 and 4.3 resulting in a buildup of Cr(1V) prior to complete reduction of Cr(V). For conditions with excess ascorbate the reaction steps are as would be expected for the overall reaction of the three-electron reduction of Cr(V1) by the two-electron reductant ascorbate (eq 13). Ascorbate radical would be consumed by disproportionation to ascorbate and DHA (eq 12) (77). If any carbon-based radicals are formed under these conditions, the nitroxyl radical adduct would presumably be quenched by the excess ascorbate (eq 20) (47) (data not shown): R-NO'

+ ascH-

-

R-NOH

+ awe-

(20)

Under conditions of excess Cr(V1) the reactivity is more complicated. Since Cr(V) and spin trap radical adducts were observed only in the presence of excess Cr(VI), this suggested that the carbon-based radicals were products of a Cr(V1) reaction with a species other than reduced ascorbate. By this scheme, the carbon-based radicals would be produced by reaction of Cr(V1) with DHA (eq 15), essentially the three-electron oxidation of ascorbic acid. This hypothesis is currently being tested. Chromium(V) may then be formed by reaction of Cr(V1) with Cr(1V) (eq 14) (50, 51, 78) or by reaction of Cr(V1) with DHA (eq 15). The disproportionation of two molecules of Cr(1V) to give Cr(V) and Cr(II1) has been invoked in other Cr(V1)oxidation mechanisms (63, but since Cr(V) is only observed in the presence of excess Cr(V1) reaction of Cr(1V)withCr(V1) (eqn 14) is proposed here. However, these EPR experiments do not rule out the possible formation of Cr(V) by reaction of Cr(V1) with Cr(II1) (79).

Conclusions EPR spectroscopy experiments have shown that the reaction between Cr(V1) and ascorbate produces Cr(V), Cr(IV), and carbon-based radicals as reactive intermediates. No oxygen dependence for formation of either Cr(V) or the radical adducts was observed. The Cr(V) species was unstable in Tris.HC1buffer. The carbon-based radicals were presumed to come from further oxidation of dehydroascorbate by Cr(VI), making ascorbate a threeelectron reductant with Cr(V1). We are interested in the types of intermediates that are formed in the reaction of Cr(V1) with ascorbate (vitamin C) since it is one of the major cellular reductants of carcinogenic Cr(V1) (6,7)and it is still unclear as to whether this reaction is a detoxification or toxification process (BO). The reactions described here should not be interpreted to suggest that the Cr(V) and free-radical intermediates are only relevant under conditions of excess Cr(VI), which would not be biologically relevant; rather this is a model

system which provides a means to generate these reactive species to explore their role in DNA damage in vitro. Preliminary results (11) suggest that the presence of Cr(V) correlates with Cr-DNA binding and that formation of free radicals results in DNA strand breakage. Therefore, this system is a useful model to study the mechanism of Cr(V1)-induced DNA damage.

Acknowledgment. This investigation was supported by PHS grant CA34869 awarded by the National Cancer Institute, DHHS (K.E.W.). The EPR spectrometer was purchased with funding from NSF grant CHE-8701406. D.M.S. was supported by a postdoctoral fellowship from the Norris Cotton Cancer Center, Dartmouth/Hitchcock Medical Center and an NRSA fellowship (CA59292) from the National Cancer Institute, DHHS.

References Alcedo, J. A., and Wetterhahn, K. E. (1990)Chromium toxicity and carcinogenesis.Int. Reu. Exp. Pathol. 31,85-108. De Flora, S.,and Wetterhahn, K. E. (1989)Mechanismsof chromium metabolism and genotoxicity. Life Chem. Rep. 7, 169-244. Connett, P. H.,and Wetterhahn, K. E. (1983)Metabolism of the carcinogen chromate by cellular constituents. Struct. Bonding 54, 93-124. Standeven, A. M., and Wetterhahn, K. E. (1991)Possible role of glutathione in chromium(V1) metabolism and toxicity in rats. Pharmacol. Toxicol. 68,469-476 and references therein. Suzuki,Y .,and Fukuda,K. (1990)Reduction of hexavalent chromium by ascorbic acid and glutathione with special reference to the rat lung. Arch. Toxicol. 64,169-176. Standeven, A. M., and Wetterhahn, K. E. (1991)Ascorbate is the principal reductant of chromium(V1) in rat liver and kidney ultrdiltrates. Carcinogenesis 12,1733-1737. Standeven, A. M., and Wetterhahn, K. E. (1992)Ascorbate is the principal reductant of chromium(V1)in rat lung ultrafiltrates and cytosols and mediates Cr-DNA binding in uitro. Carcinogenesis 13, 1319-1324. Sugiyama, M., Tsuzuki, K., and Ogura, R. (1991)Effect of ascorbic acid on DNA damage, cytotoxicity, glutathione reductase, and formation of paramagnetic chromium in Chinese hamster V-79cells treated with sodium chromate(V1). J. Biol. Chem. 266, 33833386. Wise, J. P., Orenstein, J. M., and Patierno, S. R. (1993)Inhibition of lead chromate clastogenesisby ascorbate relationship to particle dissolution and uptake. Carcinogenesis 14, 429-434. Goodgame, D. M. L., and Joy, A. M. (1987)EPR study of the Cr(V) and radical speciesproduced in the reduction of Cr(V1)by ascorbate. Inorg. Chim. Acta 135,115-118. Stearns, D. M., Courtney, K.D., Giangrande, P. H., Phieffer, L. S., and Wetterhahn, K. E. (1994)Reduction of chromium(V1) by ascorbate: role of reactive intermediates in DNA damage in uitro. Enuiron. Health Perspect. (in press). Dixon, D. A., Sadler, N. P., and Dasgupta, T. P. (1993)Oxidation of biologicalsubstrates by chromium(VI). Part 1. Mechanism of the oxidation of L-ascorbic acid in aqueous solution. J. Chem. Soc., Dalton Trans. 3489-3495. Lefebvre, Y., and PBzerat, H. (1992)Production of activated species of oxygen during the chromate(V1)-ascorbatereaction: implication in carcinogenesis. Chem. Res. Toxicol. 5,461-463. BgUdea, I., and Munteanu, L. (1980)The kinetics of the oxidation of ascorbic acid by chromate. Studia Uniu. Babes-Bolyai. Chem. 25,24-31. Banat?, B. (1981)A stopped-flow study of the oxidation reaction of ascorbic acid with chromic acid. Inorg. Chim. Acta 53,L13-Ll5. Connett, P. H., and Wetterhahn, K. E. (1985)In vitro reaction of the carcinogen chromate with cellular thiols and carboxylic acids. J. Am. Chem. SOC. 107,4282-4288. Agrawal, A,, Rao, I., and Sharma, P. D. (1993)Trace metal-ion catalysis: kinetics and mechanism of oxidation of I-ascorbic acid by chromium(VI)in phosphate buffers. Transition Met. Chem. 18,191196. Paukstelis, J. V.,Mueller, D. D., Seib, P. A., and Lillard, D. W., Jr. (1982)NMR spectroscopy of ascorbic acid and ita derivatives. In Ascorbic Acid: Chemistry, Metabolism, and Uses.(Seib, P. A. and Tolbert, B. M., Eds.) Advances in Chemistry Series 200,American Chemical Society, pp 125-151,Washington, DC.

Chem. Res. Toxicol., Vol. 7, No. 2, 1994 229

Cr(VI)Reduction by Ascorbate Buettner, G. R. (1988) In the abaence of catalytic metals ascorbate does not autoxidize at pH 7: ascorbate as a test for catalytic metals. J. Biochem. Biophys. Methods 16,27-40. Krumploc, M., and RoEek, J. (1979) Synthesis of stable chromium(V) complexes of tertiary hydroxy acids. J. Am. Chem. SOC.101, 3206-3209.

Dalal, N. S., Millar, J. M., Jagadeesh, M. S., and Seehra, M. S. (1981) Paramagnetic resonance, magnetic susceptibility, and antiferromagnetic exchange in a CP+ paramagnet: potassium perchromate (KaCrOs).J. Chem. Phys. 74, 1916-1923. Buettner, G. R., and Oberley, L. W. (1978) Considerations in the spin trapping of superoxide and hydroxyl radical in aqueous systems using 5,5-dimethyl-l-pyrroline-l-oxide. Biochem. Biophys. Res. Commun. 83,69-74. Wertz, J. E., and Bolton, J. R. (1986) Experimental methods; spectrometer performance. In Electron Spin Resonance: Elementary Theory andPractica1 Applications pp 464-465. Chapman and Hall, New York. Zang, L. Y., Zhang, Z. Y., and Misra, H. P. (1990) EPR studies of generated during photoirradiation of trapped singlet oxygen (102) hypocrellin A. Photochem. Photobiol. 52, 677-683. Tan, J. D., Hudson, S. E., Brown, S. J., Olmstead, M. M., and Mascharak, P. K. (1992) Syntheses, structures, and reactivities of synthetic analogues of the three forms of Co(II1)-bleomycin: proposed mode of light-induced DNA damage by the Co(II1) chelate of the drug. J.Am. Chem. SOC. 114,3841-3853. Chan, S. I.,Fung, B. M., and Lcltje,H. (1967) Electron paramagnetic resonance of Mn(I1) complexes in acetonitrile. J.Chem. Phys. 47, 2121-2130.

Laroff,G. P., Fessenden,R. W.,and Schuler,R. H. (1972) The electron spin resonance spectra of radical intermediates in the oxidation of ascorbic acid and related substances. J.Am. Chem. SOC.94,90629073.

Russev, P., Mitewa, M., and Bontchev, P. R. (1981) EPR study on Cr(V)-oxalicacid complexesin nonaqueous medium. J. Znorg. Nucl. Chem. 43, 35-40. Aiyar, J., De Flora, S., and Wetterhahn K. E. (1992) Reduction of chromium(V1)to chromium(V)by rat liver cytosolic and microsomal fractions: is DT-diaphorase involved? Carcinogenesis 13, 11591166 and references therein. Buettner, G. R. (1987) Spin trapping: ESR parameters of spin adducts. Free Radicals Biol. Med. 3, 259-303. Makino, K., Mossoba, M. M., and Riesz, P. (1983) Chemical effects of ultrasound on aqueous solutions. Formation of hydroxyl radicals and hydrogen atoms. J. Phys. Chem. 87, 1369-1377. Beattie, J. K.,and Haight, G. P., Jr. (1972) Chromium(V1)oxidations of inorganic substrates. Prog. Znorg. Chem. 17,93-145. Tolbert, B. M., and Ward, J. B. (1982) Dehydroascorbic Acid. In Ascorbic Acid: Chemistry, Metabolism and Uses (Seib, P. A,, and Tolbert,B. M., Eds.) Advances in Chemistry Series 200, pp 116-118, American Chemical Society, Washington, DC. BorrBs-Almenar,J. J., Burriel,R., Coronado,E., Gatteschi, D., G6mezGarcia, C. J., and Zanchini, C. (1991) Magnetic interactions and single-ionzero-field-splittingeffectsin the two-sublatticemanganese chain MnMn(EDTA).SH20: Magnetism and single-crystal EPR spectra. Znorg. Chem. 30,947-950. Bodini, M. E., Willis, L. A., Riechel, T. L., and Sawyer, D. T. (1976) Electrochemical and spectroscopic studies of manganese(II), -(III), and -(IV)gluconate complexes. 1. Formulas and oxidation-reduction stoichiometry. Znorg. Chem. 15, 1538-1543. Kitajima, N., Singh, U. P., Amagai, H., Osawa, M., and Moro-oka, Y. (1991) Oxidative conversion of a Mn(p-OH)zMnto a Mn(p-O)zMn moiety. Synthesis and molecular structures of a (p-hydroxo)dimanganese(I1,II) and (p-0xo)dimanganese(111,111) complex with a hindered Na ligand. J. Am. Chem. SOC.113, 7757-7758. Magers, K. D., Smith, C. G., and Sawyer,D. T. (1978) Polarographic and spectroscopic studies of the manganese(II), -(III), and -(IV) complexes formed by polyhydroxy ligands. Znorg. Chem. 17, 515523.

Richens, D. T., Smith, C. G., and Sawyer, D. T. (1979) Sorbitol and related poly01 complexes of manganese(II), -(III), and -(IV): redox and oxygenation equilibria. Znorg. Chem. 18, 706-712. Krumploc, M., and RoEek, J. (1985) Chromium(V) oxidations of organic compounds. Znorg. Chem. 24,617-621. Buettner G. R. (1993) The pecking order of free radicals and antioxidants: lipid peroxidation, a-tocopherol, and ascorbate. Arch. Biochem. Biophys. 300,535-543. Eaton, S. S., and Eaton, G. R. (1978) Interaction of spin labels with transition metals. Coord. Chem. Rev. 26. 207-262. Finkelstein, E., Rosen, G. M., Raukman, E. J., and Paxton, J. (1979) Spin trapping of superoxide. Mol. Pharmacol. 16,676-685. . . Janzen, E. G., Wang, Y. Y., and Shetty, R. V. (1978) Spin trapping (43) with a-pyridyl 1-oxideN-tert-butyl nitrones in aqueous solutions. A unique electron spin resonance spectrum for the hydroxyl radical adduct. J. Am. Chem. SOC.100, 2923-2925.

Kawanishi, S., Inoue, EL, and Sano, S. (1986) Mechanism of DNA cleavage induced by sodium chromate(V1) in the presence of hydrogen peroxide. J.Biol. Chem. 261, 5952-5958. Belkin, S., Mehlhorn, R. J., Hideg, K., Hankovsky, O., and Packer, L. (1987) Reduction and destruction rates of nitroxide spin probes. Arch. Biochem. Biophys. 266,232-243. Brownlie, I. T., and Ingold, K. U. (1967) The inhibited autoxidation of styrene. Part VII. Inhibition by nitroxides and hydroxylamines. Can. J. Chem. 46,2427-2432. Nigam, S., Asmus, K-. D., and Willson, R. L. (1976) Electron transfer and addition reactions of free nitroxyl radicals with radiation induced radicals. J. Chem. SOC.,Faraday Trans. 1 72,2324-2340. Morehouse, K. M., and Mason, R. P. (1988) The transition metalmediated formation of the hydroxyl free radical during the reduction of molecular oxygen by ferredoxin-ferridoxin:NADP+oxidoreductase. J.Biol. Chem. 263,1204-1211. Espenson, J. H. (1970) Oxidation of transition metal complexes by chromium(V1).Acc. Chem. Res. 3,347-353. Bramley, R., Farrell, R. P., Ji, J-.Y., and Lay, P. A. (1990) Ligand exchange and redox reactions of chromium(V): isolation of the chromium(V) oxolato complexesfrom the ligand exchange reaction of bis[2-ethyl-2-hydroxybutanoato(2-)loxochromate(V)with oxalate in 50% aqueous acetic acid. A u t . J. Chem. 43,263-279. Farrell, R. P., Judd, R. J., Lay, P. A., Bramley, R., and Ji, J-. Y. (1989) Ligand exchange and reduction reactions of oxochromate(V) complexes: characterization of the common chromium(V) intermediates in the reductions of chromium(V1) and of trans-bis(2ethyl-2-hydroxybutanoato(2-))oxochromate(V)by oxalic acid.Znorg. Chem. 28, 3401-3403. Sugiyama, M., Tsuzuki, K., Hidaka, T., Ogura, R., and Yamamoto, M. (1991) Reduction of chromium(V1)in Chineae hamster V-79cells. Biol. Trace Elem. Res. 30, 1-8. Sugiyama, M. (1991) Effects of vitamins on chromium(V1)-induced damage. Enuiron. Health Perspect. 92, 63-70. Sugiyama, M., Tsuzuki, K., and Haramaki, N. (1993) Influence of o-phenanthroline on DNA single-strand breaks, alkali-labile sites, glutathione reductase, and formation of chromium(V) in Chinese hamster V-79 cells treated with sodium chromate(V1). Arch. Biochem. Biophys. 305, 261-266. Sugiyama, M., Lin, X., and Costa, M. (1991) Protective effect of vitamin E against chromosomal aberrations and mutation induced by sodium chromate in Chinese hamster V79 cella. Mutat. Res. 260, 19-23.

Sugiyama,M.,Tsuzuki, K., Lin, X.,and Costa, M. (1992) Potentiation of sodium chromate(V1)-induced chromosomal aberrations and mutation by vitamin B2 in Chinese hamster V79 cells. Mutat. Res. 283,211-214.

Liebross, R. H., and Wetterhahn, K. E. (1992) In uiuo formation of chromium(V) in chick embryo liver and red blood cells. Carcinogenesis 13, 2113-2120. Hamilton, J. W., and Wetterhahn, K. E. (1986) Chromium(V1)induced DNA damage in chick embryo liver and blood cells in uiuo. Carcinogenesis 7 , 2085-2088. Westheimer,F. H. (1949) The mechanismsof chromicacid oxidations. Chem. Rev. 45, 419-451. Ghosh, M. C., and Gould, E. S. (1990) Electron transfer. 103. Reactions of aqueous chromium(1V).Znorg. Chem. 29,4258-4262. Ghosh, M. C., and Gould, E. S. (1991) Electron transfer. 106. Stabilized aqueous chromium(IV), as prepared from the chromium(VI)-arsenic(II1) reaction. Znorg. Chem. 30, 491-494. Ghosh, M. C., and Gould, E. S. (1992) Carboxy-bound chromium(IV); a formal potential and estimated self-exchangerate for CrmJV pertaining to 2-ethyl-2-hydroxybutanoatechelates. J. Chem. SOC., Chem. Commun. 195-196. Ghosh, M. C., and Gould, E. S. (1993) Electron transfer. 114. Structure-reactivity relationships in the redox series chromium(V,IV,III) and vanadium(V,IV,III). J. Am. Chem. SOC.115, 31673173.

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-.Ye, Baker, R. S. U., and Bonin, A. M. (1992) Supression by vanadium(IV) of chromum(V)-mediated DNA cleavage and chromium(VI/ V)-induced mutagenesis. Synthesis and crystal structure of the vanadium(1V)complex (Nh)~V(O~(HOC(Et&OO)(OC(Et&OO]l. Znorg. Chem. 31,4906-4908. Bose, R. N., Moghaddas, S., and Gelerinter, E. (1992) Long-lived chromium(1V)and chromium(V) metabolites in the chromium(V1)glutathione reaction: NMR, ESR, HPLC and kinetic characterization. Znorg. Chem. 31, 1987-1994. Scott, S. L., Bakac, A., and Espenson, J. H. (1992) Oxidation of alcohols,aldehydes, and carboxylatesby the aquachromium(1V)ion. J. Am. Chem. SOC.114,4205-4213. Ghosh, M.C.,Gelerinter,E.,andGould,E. S. (1992) Electrontransfer. 111. Disproportionation of carboxylato-bound chromium(IV). Catalysis by manganese(I1). Znorg. Chem. 31, 702-705.

230 Chem. Res. Toxicol., Vol. 7, No. 2, 1994 (68)Thornalley, P. J., and Stern,A. (1984)The production of free radicals during the autooxidation of monosaccharides by buffer ions. Carbohydrate Res. 134,191-204. (69) Ortiz de Montellano,P. R., Augusto, O.,Viola, F.,and Kimze. K. L. (1983)Carbon radicals in the metabolism of alkyl hydrazines. J. Biol. Chem. 268,8623-8629. (70) k i t e , L.C. C., and Augusta, 0. (1989)DNA alterations induced by the carbon-centered radical derived from the oxidation of 2-phenylethybydrazine. Arch. Biochem. Bibphys. 270,560-572. (71) Dedon, P. C., and Goldberg, I. H. (1992)Free-radical mechanisms involved in the formation of sequence-dependent bistranded DNA lesions by the antitumor antibiotics bleomycin, neocarzinostatin, and calicheamicin. Chem. Res. Toxicol. 6,311-332. (72) Iheanacho, E. N., Sarel, S., Samuni, A., Avramovici-Griaaru,S., and Spira, D. T. (1991)Fe(11)-chelates based on redox-active pyridoxal-betaines as C-centered radicals causing single- and doublestrand scissions to DNA. Free Radicals Res. Commun. 11, 307315. (73) Hiramoto, K., Johkoh, H., Sako, K-.I., and Kikugawa, K. (1993) DNA breaking activity of the carbon-centered radical generated from 2,2-azobie(2-amidinopropane)hydrochloride (APPH). Free Radicals Res. Commun. 19,323-332.

Stearns and Wetterhahn (74) RoEek, J., and Weatheimer, F. H.(1962)The cleavage of cis- and tra~-1,2-dimethyl-1,2-cyclopentanediolby chromic acid. J. Am. 84,2241-Q246. Chem. SOC. (75) Rameeh,S., Uahapatro,S. N.,Liu, J. H.,andRoEek, J. (1981)Threeelectron oxidations. 18.Carbon-13and deuterium isotope effects in the cooxidation of 2-hydroxy-2-methylbutyric acid and 2-propanol. Evidence for a two-step mechaniim. J. Am. Chem. SOC. 103,61725175. (76) Ghosh, S.K., Base, R. N., and Gould, E. S. (1987)Electron transfer. 85.Reduction of chromium(V) with aacorbic acid. Znorg. Chem. 26, 2684-2687. (77)Bieleki, B. H.J., Augustine, 0. A., and S c h w a , H. A. (1981) Mechanismof disproportionation of ascorbateradicals. J.Am. Chem. SOC.103,3616-3518. (78) Sen Gupta, K. K., and Chakladar, J. K. (1973)Kinetics of the metalion oxidation of phenylphosphonoua acid. Part II. Oxidation by chromium(VI). J. Chem. Sec., Perkin Trans.2 929-932. (79) Hasan, F., and RoEek, J. (1972)Three-electron oxidations. II. The 94,9073chromiw(VI) oxidation of oxalic acid. J. Am. Chem. SOC. 9081. (80)Sugiyama,M.(1992)Role of physiologicalantioxidanta in chromium(VI)-induced cellular injury. Free Radicale Biol. Med. 12,397-407.