Use of denaturing-gradient gel electrophoresis to study chromium-induced point mutations in human cells.

A large number of hprt-mutants were obtained by treating human lymphoblast cells (TK6) with 5 microM K2Cr2O7 for 5 hr and selecting by growth in 6-thioguanine. A combination of high fidelity polymerase chain reaction (PCR) and denaturing gradient gel electrophoresis (DGGE) allowed us to measure mutant frequencies as a function of DNA sequence. Chromium(VI) induced four hotspots in a 104 bp domain of hprt exon 3. Substitutions at G:C base pairs were the predominant mutations. One of the chromium-induced hotspots was located at the same position as previously determined hydrogen peroxide and benzo(a)pyrene diol epoxide hotspots.


Introduction
Chromium(VI) is a human and animal carcinogen (1). Exposure to chromium(VI) compounds is common in industries like steel manufacturing and leather tanning, and has been linked to excess risk for lung cancer among the workers exposed to various forms of chromium(VI) (2,3). It is believed that chromium(VI) exerts its mutagenic effect by being rapidly transported into cells and reduced intracellularly. However, it is not clear whether the ultimate mutagen is a chromium intermediate or an oxygen radical generated during chromium(VI) reduction.
Mutational spectra, the unique patterns of DNA alterations associated with a mutagen, provide useful information for the study of mechanisms of mutations (4-7). Such cell-culture spectra are usually obtained by treating cells with a mutagen and analyzing a large set of mutant colonies to avoid the bias of sib selection. This "clone by clone" approach demands an inordinate amount of labor to generate a statistically valid spectrum. We have previ- ously described an approach to obtain a mutational spectrum of great statistical precision by treating several large cell cultures with a mutagen to induce large number of independent mutations, amplifying the target DNA sequence with high fidelity polymerase chain reaction (PCR) and then separating mutant sequences from nonmutant sequences by denaturing gradient gel electrophoresis (DGGE) (8). The positions, types, frequencies and sequence specificity are studied by sequencing the mutant DNA. Such an approach reduces the time and labor necessary to obtain the mutational spectrum.
Obtaining Chromium-induced Mutants To achieve a precision of ±20% for the induced frequency of a particular mutation, one must analyze at least 100 independent mutants for that particular mutant type. If we select the hprt-mutants by 6thioguanine (6TG), and consider 1% of 6TG-resistant (6TGr) mutations as a hotspot, then we need to induce 10,000 independent 6TGr mutants to confirm the existence of such a hotspot. In addition, assuming 50% of the 6TG' mutants are due to large DNA alteration (e.g., large deletion), one needs to analyze 20,000 independent mutants to obtain a point mutation spectrum. In this study, 6 liters of human lymphoblast cells (TK6) at a concentration of 1 x 106 cells/ml were treated with 5 pM K2Cr207 for 5 hr. Survival and mutant fraction were deter-mined (9) to be 33% and 1.2 x 10-5, respectively. The treatment resulted in: The experiment was performed twice. The number of chromium induced mutants met our requirement to obtain a mutational spectrum with good statistical precision.

Application of DGGE to Examine hprt-Mutations
DGGE is a technique developed by Fischer and Lerman (10) to separate DNA sequences that differ by a single base pair.
The technique is based on the fact that the electrophoretic mobility of DNA in polyacrylamide gel is sensitive to the secondary structure of the molecule. When a DNA sequence containing both a high-melting and a low-melting domain is run on a polyacrylamide gel with increasing concentrations of denaturant, the low-melting domain will make the transition from helix to random coil (melted), resulting in greatly reduced mobility of the molecule. The melting property of a domain is extremely sequence-dependent. We extended this approach by boiling and reannealing mutant sequences together with wild-type sequences to create mutant/wild-type heteroduplexes. By this means, all point mutations in the low-Environmental Health Perspectives  melting domain are separated from the wildtype homoduplexes (11). In this study, the low melting region of hprt exon 3, which contains 104 base pairs (bp), was screened. The high-melting domain of e'xon 3 serves as a "clamp," allowing DNA molecules to obtain a partially melted structure and focus in the gel. The target sequence was PCR amplified with high-fidelity Vent polymerase (New England Biolabs, Beverly, MA) under optimal conditions (12). The wildtype probe was hybridized with mutant sequences to form mismatched heteroduplexes that decreased the stability of DNA molecules. As a result, mutant/wild-type heteroduplexes focused at lower denaturant concentrations, and thus were separated from the wild-type sequences. The mutational spectrum was obtained by analyzing the intensity and sequence of the mutant bands.

Results and Discussion Mutational Spectrum of K2Cr2O7
Results of DGGE analysis of mutations in the 104-bp low-melting domain of hprt exon 3 are shown in Figure 1.
Results of sequence analysis of the  mutant bands are displayed in Table 1. Four hotspot mutations induced by K2Cr2O7 treatment were found. The frequencies of these mutations ranged from 2 to 4.5% of 6TGr mutants and contained 475 to 1069 independent mutants in each hotspot. Three of the 4 mutations were G:C base substitutions.

DGGE as a Tool to Study Chromium Mutagenisis
This study demonstrated a clear reproducible spectrum for chromium(VI). The number of mutants studied was large enough to permit statistical analysis of the mutant frequency as a function of DNA sequence. The approach of high-fidelity PCR and DGGE can now be further applied to study the mechanism of chromium mutagenisis. The mutational spectra of chromium(VI) and the suspected ultimate mutagen could be compared. The comparison could reveal either no common hotspot, suggesting that the suspected ultimate mutagen does not contribute significantly to chromium mutagenisis; or some common hotspots, suggesting a shared pathway related to the suspected ultimate mutagen. It is interesting to see that chromium(VI), as well as hydrogen peroxide (13) and benzo(a)pyrene diol epoxide (8), induced mutations at position 243. Chromium(VI) and benzo(a)pyrene diol epoxide hotspots are both G:C -* A:T; but the hydrogen-peroxide hotspot is G:C -> C:G. These results suggest either that all three mutagens may share some but not all mutagenic pathways or that position 243 is particularly prone to DNA damage leading to hprt-mutations.