Generation of hydroxyl radical by chromate in biologically relevant systems: role of Cr(V) complexes versus tetraperoxochromate(V).

While Cr(V) species and .OH radicals have been suggested to play significant roles in the mechanism of chromate-related carcinogenesis, controversy still exists regarding the identity of the Cr(V) species and their role in the generation of .OH radicals. Some recent studies have suggested that the primary Cr(V) species involved is the tetraperoxochromate(V) (CrO8(3-)) ion, which produces .OH radical either on decomposition or by reaction with H2O2. The present study utilized ESR and spin trapping techniques to probe this mechanism. The results obtained show that (i) CrO8(3-) is not formed in any significant quantity in the reaction of chromate with biologically relevant reductants such as glutathione, glutathione reductase, NAD(P)H, ascorbate, vitamin B2, etc. (ii) Decomposition of CrO8(3-), or its reaction with H2O2 does not generate any significant amount of .OH radicals. (iii) The major Cr(V) species formed are complexes of Cr(V) with reductant moieties as ligands. (iv) These Cr(V) complexes generate .OH radicals from H2O2 via Fenton-like reaction. The present study thus disagrees with the recently proposed "tetraperoxochromate(V) theory of carcinogenesis from chromate." Instead, it suggests an alternative mechanism, which might be labeled as "the Cr(V)-complexation-Fenton reaction model of carcinogenesis from chromate.


Introduction
Chromate and Cr(VI)-containing compounds have been found to exert serious toxic and carcinogenic effects on humans and animals and to cause mutations in bacteria and transformation of mammalian cells (1)(2)(3)(4)(5)(6)(7). Since it has been reported that such Cr(VI) compounds do not react with isolated DNA (8), the reduction of Cr(VI) by cellular reductants has been thought to be an important step in the mechanism of Cr(VI)-induced DNA damage (6,8). Using ESR spectroscopy, Jennette has shown that a long-lived Cr(V) species is formed in the course of microsomal reduction of Cr(VI) in the presence of NADPH (9). Since Cr(V) complexes are generally characterized as being labile and reactive, whereas Cr(III) complexes are substitutionally inert, the detection of Cr(V) formation led Jennette to suggest that the Cr(V) interme-diates are the likely candidates for the "ultimate" carcinogenic forms of chromium compounds (9). While several earlier studies have shown that the Cr(VI)-induced DNA damage is strongly dependent on the Cr(V) intermediates (8,(10)(11)(12)(13), recent studies have suggested that hydroxyl (-OH) radicals may also play an important role (10)(11)(12)(14)(15)(16)(17)(18)(19)(20)(21). It has been reported, for example, that treatment of Chinese hamster V-79 cells with FAD and Cr(VI) resulted in an increase in Cr(VI)-induced DNA strand breaks over that observed upon treatment of cells with Cr(VI) alone (21). This increase in DNA strand breakage was attributed to OH radical generation in the presence of FAD. In contrast, incubation of Chinese hamster V-79 cells with vitamin E, an OH radical scavenger, prior to treatment with Cr(VI) led to a decrease in Cr(VI)-induced DNA strand breaks (12,16).
As to the mechanism of Cr(VI) related 'OH radical generation, our studies (22)(23)(24) suggest that the reaction of a Cr(V) complex with hydrogen peroxide via a Fenton-like reaction is the important source of chromium-mediated -OH radical generation in biological systems (Equation [1]). Cr(V) + H202 -4Cr(VI) + OH + OH- [1] Cr(V) species has been reported to be generated in the reduction of Cr(VI) by vari-ous biological systems (6), in particular, microsomes (9,25), mitochondria (26), superoxide radical (27), certain flavoenzymes (28)(29)(30), mitochondrial electron transfer chain complexes (31), ascorbate (32,33), and thiol-and diol-containing molecules (6,23,34,35). In contrast to the mechanism outlined in Equation [1], it has been frequently suggested (10,11,15,18,36) that Cr(VI) reacts with cellular H202 to form tetraperoxochromate(V) (CrO83) ions, which decompose to produce -OH radicals via Equation [2], as given below: Cr(VI) + H202 -CrO83--Cr(VI) + -OH [2] The Cr(VI)-mediated 'OH radical generation via the CrO8 3intermediate was recently cited (37) as the basis of "the tetraperoxochromate(V) theory of carcinogenesis from chromate." The first evidence for the role of CrO8 3in the Cr(VI)-mediated -OH radical generation was reported by Kawanishi et al. (15). They observed CrO8 3formation by ESR from a mixture containing 40 mM Na2CrO4 and 400 mM H202 at pH 8.0. However, the concentrations used for both Cr(VI) and H202 were orders of magnitude higher than any in vivo estimate. Later Aiyar et al. (10) reaffirmed the Kawanishi model of 'OH radical generation from a mixture of Cr(VI) and H202, i.e., that CrO8 3was the species responsible Environmental Health Perspectives for -OH generation. In a subsequent report (11), these authors reported the detection of CrO83ions from mixtures containing Cr(VI), GSH and H202, although the detected ESR signal appears to be too weak for any reliable assignment. In a separate study, Sugiyama et al. (36) reported that a mixture of Cr(VI) and vitamin B2 generates CrO8 3 and vitamin B2 enhances the OH radical generation from a mixture of Cr(VI) and H202' They attributed the enhancement of-OH radical generation by vitamin B2 to CrO83formation. This conclusion disagrees with our reports (22)(23)(24) that any diol-containing molecule will generate a Cr(V)-diol complex upon reaction with Cr(VI) (23), and that this Cr(V)-diol complex will react with H202 to generate OH radical via a Fenton-like mechanism. The present investigation was undertaken with the view of resolving this and related controversies.

ESR Measurements
All ESR measurements were made using a Varian E3 ESR spectrometer and a flat cell assembly. Hyperfine couplings were measured (to 0.1 G) directly from magnetic field separations using K3CrO8 and DPPH as reference standards. Reactants were mixed in test tubes in a final volume of 250 pl. The reaction mixture was then transferred to a flat cell for ESR measurement. The concentrations given in the figure leg-  (15) observed via ESR formation of CrO83from a mixture containing 40 mM Na2CrO4 and 400 mM H202 at pH 8.0. However, the concentrations used for both Cr(VI) and H202 were orders of magnitude higher than any in vivo estimate. We tried to detect the CrO83formation at concentrations of K2Cr207 and H202 as low as possible. As shown in Figure (Figure lb), which was assigned to Cr(V)-NADPH complex formation (28)(29)(30). The spectra in Figure 1 demonstrate that in an enzymatic reduction of Cr(VI), Cr(V)-complex formation is much more efficient than CrO8 3 formation from direct reaction of Cr(VI) with H202 Does a Mixture ofCr(VI) and Vitamin B2 Generate CrO837? As shown in Figure  2a generates an ESR signal at g = 1.9795. The measured peak-to-peak linewidth is about 5 G. While in an earlier report (36) this signal had been assigned to the CrO8 3ions, this g value and the linewidth are typical of those of Cr(V)-diol complexes (23).
To ascertain whether this signal was due to CrO8 , a synthesized K3CrO8 crystal was added to the Cr(VI)/vitamin B2 mixture. The g value of CrO83signal was measured to be 1.9720 (Figure la), in agreement with an earlier report (38). Since vitamin B2 contains diol functionalities, a key group for Cr(V) formation, spectra were recorded from reaction of Cr(VI) with other diol-containing molecules ( Figures  2b and 2c). We also recorded the ESR spectrum from a mixture of Cr(VI) and vitamin B2 in DMSO because of the higher solubility of vitamin B2 in that medium. As can be noted in Figure 2d, this spectrum was strong enough that a superhyperfine splitting of about 0.85 G could be resolved. Such superhyperfine splitting is considered to be characteristic of the superhyperfine interaction of the methylene-type of hydrogens in Cr(V)-diol complexes (23). At higher spectral gain and wider scan width, we observed four weak satellite signals due to the 53Cr isotope (9.55% abundance, I = 3/2) (Figure 2e).
The measured 17.7 G 53Cr hyperfine splitting (indicated in Figure 2e) is very similar to those reported earlier (23) for Cr(V) complexes, confirming that the Cr(V) species detected in the mixture of Cr(VI) and vitamin B2 is indeed a Cr(V)-diol type complex, and not CrO8 3 ions.
Does Reaction of Cr(VI) with GSH Generate CrOf ? We examined whether the reaction of Cr(VI) with GSH can generate CrO83. GSH was chosen because it is considered to be one of the major Cr(VI) reductants in cellular systems (18). Figure  3a shows the ESR spectrum obtained from an aqueous solution of GSH and K2Cr207 in a 5:1 molecular ratio at pH = 4.0. This spectrum was recorded 2 min after solution preparation. The spectrum is dominated by signals at g = 1.995 (5.8 G peak-to-peak derivative width) and at g = 1.985. These two peaks have different decay rates and have been already assigned by earlier workers to two different Cr(V) species, coordination not defined (34,35). The spectra in Figures 3a-3g show the time dependence of the decay of the Cr(V) complexes. It may be noted that the g = 1.985 species decays at a slower rate than the g = 1.995 species; the g = 1.985 signal was still observable after at least 15 min of mixing. Variation of the solution pH within the range pH 3 to pH 8 yield spectra generally similar to those outlined above for pH 4.0, with an enhanced rate of decay occuring at higher pH values. This observation essentially rules out the identification of these species as CrO83-, because CrO 3becomes more stable at the higher pH. Figure 4 shows the effect of changing the GSH:K2Cr207 ratio. A decrease in the GSH:K2Cr207 ratio decreases the g = 1.995 signal while increasing the g = 1.985 signal (Figures 4d-4f). However, for a GSH:K2Cr207 ratio of less than one, not even the g = 1.995 signal was observed (Figure 4g). We reiterate that the chemical structures of these Cr(V) species are not known at present but their dependence on the GSH:Cr(VI) ratio suggests that the g = 1.985 species might be a 1:1 and the g = 1.995 species a 2:1 GSH complex of Cr(V). The results show that the major species generated in the reaction of Cr(VI)  with GSH is Cr(VI)-GSH complexes, with no evidence for CrO8 3 formation.
It was recently reported (11) that a mixture of K2Cr207, GSH and H202 in a Tris-HCl medium generates CrO8 -, although the ESR peak was too weak to make any reliable assignment. We were unable to detect any CrO83formation from a mixture of K2Cr707, GSH and H202 in Tris-HCI. To obtain further clues, we measured the stability of K3CrO8 in phosphate buffer as well in Tris-HCl buffer. The ESR spectrum of CrO83in phosphate buffer is shown in Figure 5a. This spectrum can be detected even 10 minutes after solution preparation. In Tris-HCl medium, however, CrO8 3rapidly decayed and converted to another Cr(V) species at g = 1.9721 (Figure 5b). This Cr(V) species is likely to be a Cr(V)-Tris-HCI complex, but the detailed structure has not yet been elucidated. The intensity of the newly formed Cr(V) species reduced to about 20% within 1.5 min after reaction initiation while that of CrO8 3 became  (Figure 5c). When GSH was added to a K3CrO8/Tris-HCl solution, no Cr(V) species can be detected (Figure 5d), indicating that GSH also reacts with K3CrO8. These results show that CrO8 3is unstable in the Tris-HCl (pH 8) medium and that CrO83formation in the mixture of K2Cr2O7, GSH and H202 in Tries-HCl buffer (pH 8.0) is unlikely. We also studied the effect of NADPH on the stability of K CrO8 As shown in Figure 5e, addition o?NADPH to the K3CrO8 solution caused a sharp decrease in CrO83 signal and led to the formation of the Cr(V)-NADPH complex, suggesting that CrO8 3 undergoes fast ligand exchange with NADPH to form a Cr(V)-NADPH complex. Thus in biological systems in which thiols like GSH and diols such as NADPH are plentiful, it is unlikely that CrO8 3could exist in any significant amount. Does Reaction ofCr(VI) with Ascorbate Generate CrO83 ? We used ascorbate as another Cr(VI) reductant since ascorbate is also thought to be a major Cr(VI) reductant in cell systems (6,32). As shown in Figure 6a, a mixture of Cr(VI) and ascorbate in phosphate buffer (pH 7.2) generates an ESR signal at g = 1.9794, which can be assigned to a Cr(V)-ascorbate complex according to an earlier report (33). No CrO 3-ESR signal was detected. When H202 and DMPO (as a spin trap) were added, a 1:2:2:1 quartet with hyperfine splittings of aN = aH = 14.9 G was observed ( Figure 6b). Based on these splittings (39), the 1:2:2:1 signal was assigned to the DMPO/'OH adduct, as evidence of -OH radical generation. Upon addition of H202, however, Cr(V) became nondetectable (Figure 6b), indicating that -OH radicals were generated in the reaction between Cr(V) and H202 via a Fenton-like mechanism.
Does K3CrO8 Generate -OH Radical? Figures 7a and 7b present spin trapping studies of possible -OH generation from K3CrO8. Decomposition of CrO83in phosphate buffer does not generate any detectable amount of -OH radicals 3Figure 7a). Addition of H202 to the CrO8 -solution did not cause any significant enhancement in OH generation, nor any significant decrease in the CrO83peak (Figure 7b). Figure 7c shows the Cr(V)-NADPH complex formation from a mixture of Cr(VI) and NADPH. Addition of H202 reduced the intensity of the Cr(V)-NADPH signal to a barely detectable level with a concomitant appearance of the DMPO/-OH spin adduct signal ( Figure  7d), indicating that reaction of Cr(V)-NADPH with H202 does generate -OH radicals. Similarly, reaction of Cr(VI) and GSH generated Cr(V)-GSH complexes and the DMPO/GS-spin adduct ( Figure  7e) (23). Addition of H202 reduced the intensity of Cr(V)-GSH signals and generated DMPO/'CHOHCH3 spin adduct signal (as indicated by asterisks) (Figure 70. Since the hyperfine splitting of DMPO/GS-and DMPO/-OH is very similar, ethanol was added for the identification of -OH radical. It is known that -OH radical efficiently reacts with ethanol to form 'OCHOHCH3 radical, which, in turn, is trapped by DMPO to produce DMPO/ICHOHCH3. These results show that -OH radicals were generated via reaction of the Cr(V)-GSH complex with H202, without any significant contribution from CrO8.
Earlier studies have shown that reaction of Cr(VI) with H202 at high concentrations generates -OH radical (15). The present study confirms that while OH radicals can be generated in the reaction of Cr(VI) with H202, the yield is very low. As can be noted from Figure 8, a mixture of 2 mM K2Cr207, 2 mM NADPH and 3 mM H202 in phosphate buffer (pH 7.2) generates -OH radicals in a yield 15 times higher than from a mixture of 5 mM K2Cr2O7 and 10 mM H202. Addition of GSSG-R increases this yield by a factor of 90. These results demonstrate that in cellular systems, the majority of Cr(VI)-mediated 'OH radicals would be expected to be generated via a Fenton-like reaction of Cr(V)-complexes rather than decomposition of CrO83.
Environmental Health Perspectives 234

Discussion
The results obtained in the present study demonstrate that under biologically relevant conditions, Cr(VI) does not generate CrO8 3in significant amounts. Moreover, 83 CrO8 3 decomposition or its reaction with H202 does not generate any significant amount of -OH radicals. Instead, reduction of Cr(VI) by NADPH, GSH, NADPH/GSSG-R, and ascorbate generates significant amount of Cr(V) complexes.
No CrO83species could be detected in the reaction of Cr(VI) with these materials. The Cr(V) complexes thus produced readily react with H202 to generate -OH radicals via a Fenton-like mechanism [Cr(V) + H202 -* Cr(VI) + -OH + OH-]. This mechanism is the likely major pathway for Cr(VI)-mediated -OH radical generation because Cr(V) complexes can be generated in various biological systems, including cytochrome P-450 (6), cytochrome b5 (40), the electron transport complexes of the inner mitochondrial membrane (31), microsomal systems (9,25), certain flavoenzymes such as glutathione reductase (28)(29)(30), superoxide radicals (27), and a number of thiol and diol-containing molecules (6,23,34,35). In particular, GSH is considered by far the most important among the possible thiol reductants of Cr(VI), mainly because of its ubiquitous occurrence in mammalian cells in millimolar concentrations (18,41), and because GSH levels can be easily modulated with various agents (41). In addition to GSH, ascorbate is also considered the likely Cr(VI) non-enzymatic reductant in vivo on the basis of its high reactivity with Cr(VI) and its abundance within the cell (6,32,33). As for the molecular mechanism of OH radical generation by Cr(V) complexes with H202, it may be noted that CrO83has a tetrahedral structure with all four covalent bonds fully occupied by 022 moieties as shown below. Cr 00 Thus the CrO83/H202 complex would not easily split H202 to -OH radical. On the other hand, a Cr(V) complex such as Cr(V)-NADPH, is expected to be octahedral, with one vacant site as shown below: Thus the H202 can attach to the vacant coordination site (indicated by the arrow) 0 o 0 r Cr(V) 01 H202 and form a long-lived complex to generate OH radical. This mechanism is similar to the oxidation of Fe(II) with H202 in the Fenton reaction as the production of OH radical from Fe(II) via Fenton reaction is facilitated greatly by the formation of Fe(II) complexes that have vacant sites for H202 coordination (42).
In conclusion, the present study demonstrates that the reduction of Cr(VI) by cellular reductants does not generate any significant amount of CrO8 3ions. ESR spin trapping studies using laboratory synthesized K3CrO8 showed that CrO8 3decomposition by itself, or the reaction of CrO8 3with H202 does not generate any significant amount of -OH radicals. Reactions of Cr(VI) with several major Cr(VI) reductants generate Cr(V) complexes, which produce OH radical via a Fenton-like mechanism. The present investigation does not support the "tetraperoxochromate(V) theory of carcinogenesis from chromate." Instead, we propose that the mechanism involves two species, Cr(V) complexes and -OH radicals generated via a Fenton-like mechanism. Hence, this mechanism might be labeled as "Cr(V)-complexation-Fenton reaction model of carcinogenesis from chromate."