Analysis of repair and mutagenesis of chromium-induced DNA damage in yeast, mammalian cells, and transgenic mice.

Chromium (Cr) is a widespread environmental contaminant and a known human carcinogen. We have used shuttle vector systems in yeast, mammalian cells, and transgenic mice to characterize the mutational specificity and premutational DNA damage induced by Cr(VI) and its reduction intermediates in order to elucidate the mechanism by which Cr induces mutations. In the yeast system, treatment of vector-containing cells with Cr(VI) results in a dose-dependent increase in mutations in the SUP4-o target gene of the vector; mutagenesis is enhanced in an apn-1 yeast mutant, deficient in the capacity to repair oxidative-type DNA damage. In vector-containing mammalian cells, treatment with Cr(VI) also results in a dose-dependent increase in mutations in the vector target gene supF. The Cr-induced mutations in supF occurred mostly at G:C base pairs and were widely distributed across the gene, a pattern similar to those observed with ionizing radiation or hydrogen peroxide. These results support the hypothesis that Cr(VI)-induced oxidative-type DNA damage is responsible for Cr mutagenesis in the cell. Recently these studies were extended into the Big Blue transgenic mouse system in which Cr-induced mutagenesis was observed in the lung, the target organ for Cr carcinogenesis in humans. Analysis of the spectrum of these mutations will test whether Cr mutagenesis occurs by similar mechanisms in the intact animal as in cell culture systems and yeast.


Introduction
Chromium (Cr) is an essential nutrient and a human carcinogen (1,2). Occupational exposure to hexavalent chromium [Cr(VI)] compounds through inhalation is associated with an increased incidence of bronchogenic carcinoma (2,3). In occupationally exposed workers, lung tissue levels of 100 ig/g wet weight have been reported (4). Chromium continues to be in widespread use in industry in paints, metal plating, and as a corrosion inhibitor. It has been identified as a primary contaminant at approximately one-third of the Superfund Hazardous Wastesites (5). Since Cr is an element, it cannot be removed by biodegradation or incineration methods, and furthermore it may be mobilized into groundwater sources (6). The risk to human health from ingestion is unknown, although a drinking water standard of 0.05 mg!l has been established (7).
In the environment, Cr exists primarily in the trivalent [Cr(III)] and hexavalent forms; interconversion of the two forms can occur under certain conditions (8). Generally, Cr(VI) and not Cr(III) This paper is based on a presentation at the Symposium on the Superfund Basic Research Program: A Decade of Improving Health through Multi-Disciplinary Research held [23][24][25][26] February 1997 in Chapel Hill, North Carolina. Manuscript received at EHP 1  compounds have been found to be mutagenic and carcinogenic in a variety of test systems (2,9). This appears to be due to the selective penetration of cells by chromate (CrO42-) anions (9,10). Once in the cell, Cr(VI) is subject to reduction by a wide variety of cellular reductants (11)(12)(13). Glutathione (GSH) is thought to be one of the most important intracellular reducing agents for Cr(VI); GSH is present intracellularly at mM levels and alterations in intracellular concentrations has been shown to affect Cr(VI)-induced toxicity (14,15). GSH appears to have a dual role: reactive intermediates are generated during reduction of Cr by GSH (13,16) and GSH serves as a radical scavenger to protect cells agains Cr-induced oxidative damage (17). Cr induces both oxidative-type DNA damage (e.g., DNA strand breaks and oxidized bases) and cross-links (ternary complexes of GSH or amino acids with Cr and DNA) (15,(18)(19)(20). Cr(VI) has been shown to be mutagenic in a variety of bacterial and mammalian assay systems (21)(22)(23), but the nature of the premutagenic DNA damage responsible for mutation induction has not been clearly defined. Furthermore, the link between the mutagenic and carcinogenic activities of Cr compounds has not been established.
We have undertaken a systematic analysis of Cr mutagenesis to clarify the role of GSH in Cr activation and to better define the Cr-induced DNA damage responsible for the mutagenic activity of Cr compounds. We have used shuttle vectors in yeast (24) and mammalian systems (25,26) and the Big Blue transgenic mouse system (27) to investigate the mechanism of Cr mutagenesis and to characterize Cr-induced mutations. The ultimate goal of this research is to elucidate the mechanism of Cr carcinogenesis, thereby clarifying the risk to human health from Cr exposure.

mmalian SupFSystem
The pZ189 shuttle vector (31) carries the bacterial supFgene as the mutational target, pBR322 sequences for replication and selection in bacteria, and SV40 virus sequences for replication in primate cells. The E. coli tester strain MBM7070 (31) carries the lacZ(Am)CA7020, which can be suppressed by the supF tRNA suppressor. When MBM7070 carries pZ189, it forms blue colonies in the presence of isopropyl-fP-D-thiogalactoside (IPTG) and 5bromo-4-chloro-3-indoyl--D-galactoside (X-gal); bacteria with supF mutant plasmids form white or light blue colonies. Almost any mutational change in the supF gene causes loss of suppressor function.
For mutagenesis testing, the vector was transfected into African green monkey kidney CV-1 cells grown in monolayer (24,25). For in vitro treatment of the vector, DNA was first washed in Tris-EDTA buffer (10 mM Tris, 1 mM EDTA, pH 7.6) and then treated with potassium dichromate in the presence of a 5-fold excess of GSH in 0.1 M phosphate buffer, pH 7.0, for 90 min at 37°C; DNA was washed with 100 mM EDTA before transfection. The EDTA treatment removes bound Cr from the DNA. Vector DNA was isolated 48 hr after transfection. For treatment of the transfected cells, potassium dichromate was added to the growth medium for 12 hr, starting 6 hr after transfection; vector DNA was harvested from the cells 72 hr after transfection. In all cases, vector DNA harvested from transfected cells was treated with DpnI (to degrade input DNA) and transformed into E. coli MBM7070 by electroporation.

Big Blue Transgenic Mouse Systnm
The C57BL/6 Big Blue transgenic mouse carries the bacterial lad gene in a bacteriophage X-based vector integrated into the mouse chromosome (32). For chromium treatment, mice (4to 6-weekold females) were anesthetized, and a small surgical cut was made in the ventral neck to expose the trachea; a chromium solution was injected with a 30.5-gauge needle through the wall of the trachea into the lungs. High molecular weight lung DNA was isolated as described by Stratagene (33) and the X-based vector DNA was recovered by X packaging using the Stratagene Transpack (Stratagene, La Jolla, CA). Vector lacI mutants were identified as blue plaques on agar plates with SCS-8 E. coli in the presence of X-gal; wild-type vector produced colorless plaques. Mutant sequence analysis was carried out as described by Erfle et al. (34).

Results
To investigate Cr mutagenesis, we used three mutagenesis assay systems. Each system allowed the examination of a particular aspect of Cr mutagenesis. In the yeast SUP4-o system (28), we took advantage of the availability of particular DNA repairdeficient strains to aid in the identification of the types of premutagenic DNA damage responsible for Cr mutagenesis. In the mammalian cell pZ189 system (31), we compared the mutation spectrum that resulted from treatment of the plasmid in vitro under controlled conditions with that observed when the DNA was damaged intracellularly. In the Big Blue transgenic mouse system (32), we examined the mutagenic potential of Cr in the lung, the target organ for Cr carcinogenesis in humans. By comparing results from all three systems, we have determined mutational mechanisms that appear to be common to all systems and that likely reflect the mechanisms that contribute to human disease.

Yeast SUP4-o System
In the yeast system (28), the mutational target is the yeast SUP4-o suppressor tRNA gene (25) carried on the single-copy YCpMP2 plasmid. To characterize premutagenic DNA damage in this system, we introduced this plasmid into two repairdeficient yeast strains. These were the DRY373 strain, which carries the apnl mutation and lacks apurinic endonucleasel 3'-diesterase activity (30), and the KAM-1 strain, which carries the radl mutation and is deficient in the incision step of nudeotide excision repair (NER) (29). Thus, the DRY373 strain is sensitive to agents that cause oxidative DNA damage (e.g., hydrogen peroxide) and the KAM-1 strain is sensitive to agents that cause bulky DNA adducts (e.g., UV radiation). We used these strains to determine the sensitivity to Cr toxicity and Cr mutagenicity.
For toxicity measurements, these strains were treated with Cr compounds during exponential growth, and survival was measured by a colony assay. For the mutagenesis studies, the plasmid-containing yeasts were treated with Cr and then screened for mutants. SUP4-o mutant plasmids were further analyzed by DNA sequencing. To investigate the role of GSH in Cr toxicity and mutagenesis, we used BSO to lower the intracellular level of GSH (35) and the precursor amino acids (L-glutamate, L-cysteine, and glycine) to raise the intracellular level of GSH (36).
We found that Cr toxicity was the greatest in the apnl DRY strain and under conditions of lowered GSH (Table 1). Cr mutagenesis was the highest in the DRY strain ( Figure 1) and was increased in all strains when GSH levels were raised by treatment of cells with precursor amino acids (26). These results suggest that Cr causes premutagenic DNA damage that is normally repaired by the product of the APNJ gene, but not by NER. Furthermore, GSH appears to play a role in protecting cells against Cr toxicity; but at the same time, GSH appears to enhance Cr mutagenesis. In the DRY strain, Cr induced base substitution mutations, deletions, and insertions (   one-third of the mutations analyzed were insertions of the yeast Ty transposable element into one of two sites within the SUP4-o gene. In previous studies, mobilization of the Ty element was found to be induced primarily by agents that cause DNA strand breaks (37). These results suggest that DNA strand breaks are another important type of Cr-induced premutagenic DNA damage. Mammalian pZ189 Sysm In the mammalian system, the mutational target was the bacterial supFgene carried by an SV40-based shuttle vector plasmid, pZ189 (31) (Figure 2). In this system, the plasmid DNA was treated with Cr and GSH under controlled in vitro conditions and then introduced into the cells for replication and mutation fixation (24); or the untreated plasmid was introduced into the cells, and the cells were treated with Cr (25). Under the latter conditions, the Cr will be activated by normal cellular constituents and gcGSW'.  Replicated plasmid is recovered and tested for mutations as in section A. interact with the chromatized replicating plasmid in the cell nucleus. If Cr-induced mutations that were generated after in vitro treatment of the plasmid are similar to those generated when plasmid-containing cells are treated with Cr, this would suggest that similar mechanisms are involved.
When the pZ189 plasmid was treated in vitro with Cr(VI) in the presence of GSH (1:5 molar ratio) there was a Cr dose-dependent increase in single-strand and doublestrand breaks (assayed on agarose gels) and a concomitant loss in biologic activity (assayed as transformation efficiency in E. coli) (24).
In contrast, there was no measurable effect when the DNA was treated with Cr(III) in the presence of GSH, or with either Cr(VI) or Cr(III) in the absence of GSH. Replication of this vector DNA in CV-1 monkey cells led to a Cr dose-dependent increase in supFmutants with Cr(VI) plus GSH-treated DNA, but not Cr(III)-treated DNA (24) (Table 3). These results demonstrate that reactive intermediates which cause premutational DNA damage are generated in the reduction of Cr(VI) by GSH.
We next asked whether Cr(VI) was mutagenic in the supF system when vectorcontaining cells were treated with Cr(VI). Here again, mutagenesis increased in a dosedependent manner with Cr(VI) [(25); Table  4). To determine whether premutational  Big Blue mouse/lac/ DNA damage induced inside the cell by Cr was similar to that generated in vitro by Cr in the presence of GSH, we compared the spectra of supFmutations induced under the two conditions (25). We argued that similar spectra would suggest similar mutational mechanisms and different spectra would argue for different mechanisms. Mutant plasmids were characterized by agarose gel electrophoresis to screen for large insertions and deletions and by DNA sequencing to identify smaller changes. A total of 141 mutants were analyzed from the in vitro Cr treatment (24) and 89 mutants were analyzed from the treatment ofvector-containing cells (Table 4) (25). Comparison of the mutants induced following in vitro treatment of the vector with Cr plus GSH and the mutants induced when vector-containing cells were treated with Cr in vivo revealed many similaritites (Table 5) (24,25). A large percentage of the mutations were deletions, similar to that observed with H202-treated plasmid (38) and unlike UWC-treated plasmid (39). Most of the mutations occurred at GC base pairs with a large fraction of GC --TA as well as GC ->AT base substitutions; this distribution is also similar to H202-treated plasmid and not UVCtreated plasmid. The fraction of multiple mutations appeared to be somewhat higher when the cells were treated with Cr versus treatment of plasmid.
To compare the mutation spectra with respect to the distribution of mutations within the supFgene, we carried out an analysis (25) using the method described by Adams and Skopek (40). A comparason of the spectra of mutations within the tRNA coding sequence generated by Cr(VI) in vivo and in vitro revealed no significant difference (p>0.05). A similar comparison of the two H202 spectra revealed no significant difference. In contrast, comparison of the Cr(VI) and H202 in vivo spectra revealed significant differences (p<0.001); in particular, several hot spots were observed in the Cr(VI) spectrum that were not observed in the H202 spectrum. Thus, despite overall similarities-i.e., mutations occurred primarily at GC base pairs and were widely distributed along the target sequence-differences at mutational hot spots were observed in vivo. These in vivo differences may reflect differences in the way Cr(VI)and H202-generated reactive intermediates interact with DNA in the context of other cellular proteins.

Big Blue Transgenic Mouse System
To relate mutagenesis data from cell culture to Cr mutagenesis in the target organ relevant to carcinogenesis, we next used the Big Blue transgenic mouse system (Figure 3). In this system, the lad mutagenesis target is carried in a bacteriophage X-based vector tandemly integrated into the mouse chromosome. These mice were treated with Cr(VI) by intratracheal instillation into the lung, and 1 to 4 weeks later the mice were sacrificed, the lungs were removed, and high molecular weight DNA was isolated. The vector DNA was recovered by X packaging and lad mutants were identified by plaque color on indicator plates. We observed a time-dependent (1-4 weeks; data not shown) and Cr(VI) dosedependent (1.7-6.8 mg/kg) increase in the frequency of lad mutants in vector DNA from the lungs of treated mice (Table 6). This is the first demonstration of Cr(VI)induced mutations in an experimental animal. Further study will be required to relate this mutagenesis to the carcinogenic activity of Cr compounds. Nevertheless, these results suggest that Cr has the mutagenic potential to be a genotoxic carcinogen.
Plaques Figure 3. The Big Blue mouse system. The transgenic mice carrying the lacl target gene in an integrated X-based vector, are treated with Cr by instillation into the lung. Later, animals are sacrificed, the lungs are removed, and lung DNA is isolated. Target gene sequences are recovered by x packaging and lacl mutants are identified by plaque color on indicator plates.

Discussion
The carcinogenic potential of certain chromium compounds has been recognized for several decades. This has led to the implementation of restrictions on human exposures and concern about the high levels of chromium contamination at many waste sites. Despite these concerns, chromium compounds are still used in large quantities in industry and the potential for human exposure remains. Although chromium compounds have been shown to be mutagenic in a variety of test systems, the mechanism of mutagenesis has not been well understood. Furthermore, it is not clear what level of danger is posed by environmental sources of chromium. The work described here focuses on an analysis of mechanisms by which chromium induces DNA damage and causes mutations. The long-range goal of this work is to enhance our understanding of the links between chromium exposure, DNA damage, mutagenesis, and carcinogenesis to provide a rationale for assessing the risks associated with human environmental exposures.
By using both yeast and mammalian cells we have demonstrated that Cr(VI)induced DNA damage leads to deletions, DNA rearrangements, and base substitution mutations. The pattern of base substitution mutations resembles that induced by agents, such as hydrogen peroxide and ionizing radiation, that cause oxidative DNA damage. In particular, mutations were observed primarily at GC base pairs and were distributed widely within the target gene. This is the first comparison of the mutational specificity of chromium to other oxidative agents in a system that allows the detection of a broad spectrum of mutational events. We did observe differences among these spectra at a few mutational hot spot sites (25). The mutational specificity of chromium in human cells was also examined using the denaturing gradient gel electrophoresis method developed by Chen and colleagues (23). Of the four sites at which chromiuminduced base substitutions were detected in that system, three were GC base pairs; these hot spots differed from those detected in that system after H202 treatment (23). This difference in mutational hot spots is consistent with our observations in the shuttle vector systems. Yang et al. (22) also examined the mutational specificity of chromium in the hprt gene of Chinese hamster ovary-KI cells. They reported dominant base changes at AT sites with mainly T--A and T-G transversions. These data may indicate that the base substitution specificity of chromium mutagensis is species specific.
Our studies with transgenic mice demonstrate that Cr(VI) is mutagenic in the target organ (lung) relevant to human chromium carcinogenesis. Despite the relative insensitivity of this system to mutagenesis, we detected a large increase in mutagenesis after exposure to chromium. While the levels of chromium used in this study are well above those associated with human exposure, these studies confirm that chromium has the potential to act as a genotoxic carcinogen in the intact organism. Further analysis of the spectrum of mutations in this system should allow a better understanding of the mutational mechanisms operative in the intact organism. In addition, this system can be utilized to determine the potential for chromium genotoxicity in other organs and by other routes of exposure. In particular, it will be of interest to determine whether exposures that are more relevant to environmental contamination (e.g., drinking water) also cause mutagensis in this system.