Higher sensitivity of denaturing gradient gel electrophoresis than sequencing in the detection of mutations in DNA from tumor samples.

A wide variety of human tumors has been shown to be associated with mutations in tumor-suppressor genes or oncogenes. Traditionally, these mutations have been studied by constant denaturing gel electrophoresis (CDGE), denaturing gradient gel electrophoresis (DGGE) or single-strand conformation polymorphism (SSCP). It has been suggested (8,11) that DGGE is the most sensitive and easy to use. In our laboratory, DGGE is routinely used to screen for mutations in exons 5 to 8 of p53 because these conserved regions tend to accumulate mutations. However, no mutation could be detected by sequencing in some of the DGGE-positive samples. Moreover, we found these false-negatives using DNA from tumor tissue but not from peripheral blood cells from members of families with reported history of breast cancer (10). These observations suggested to us that sequencing could have a lower sensitivity than DGGE to detect mutations from tumor samples. Indeed, this was consistent with the observations made by others using CDGE and DGGE (2) and with those obtained by Newmark et al. with exon G of the androgen-receptor gene (12) and Beck et al. (3) with the p53 gene. Moyret et al. discussed the relative sensitivities of DGGE and SSCP. According to these authors, mutation detection by sequencing of positive samples that were positive by either method was not always easy, and they suggest that a low proportion of mutated DNA could be a cause (11). On theoretical considerations, Bosari and Viale (6) have suggested that mutations can be detected by sequencing when the proportion of wildtype alleles does not exceed two thirds of the material to be analyzed. Loda (9) suggested that at least 20% of the cells analyzed should carry the mutation. We have determined more accurately the proportion of mutant DNA necessary to detect a single point mutation using both DGGE and direct sequencing of polymerase chain reaction (PCR) products. The approach followed has been to compare the sensitivity of DGGE and sequencing in samples with the same amount of total DNA but containing increasing proportions of a DNA carrying a previously defined mutation. As a source of mutant DNA, we used peripheral blood of a patient with a family history of breast cancer. These cells are heterozygous for a neutral polymorphism (CGA) and were verified by direct sequencing of a PCR product. Normal DNA (wild-type) was extracted from peripheral blood of a donor without the polymorphism, and its normal sequence was verified by the same methods. Samples of the mutant and the wild-type DNA were analyzed simultaneously to avoid sample-tosample variations. DNA was extracted by conventional methods. For PCR, approximately 1 μg of each DNA was amplified as previously described (8). The primers used were located on introns 4 and 6, respectively, and the resulting fragment of 506 bp encompasses exons 5 to 6. PCR products were purified using the Sephaglas BandPrep Kit (Pharmacia Biotech, Uppsala, Sweden) after excising the band from a 1.8% agarose gel. To prepare a mixture containing known relative proportions of mutant and wild-type DNA, serial dilutions of the purified PCR product from the heterozygous donor were electrophoresed and stained with ethidium bromide, and pictures were analyzed by densitometry. The areas of the bands of the mentioned dilutions were plotted, and the regression line obtained (r2 = 0.98, P <0.01) was used to calculate the relative concentration of wild-type products run in the same gel. PCR product from the heterozygous donor was used as the 50% mutant DNA. Samples containing decreasing relative proportions of mutant DNA (40%, 35%, 30%, 25%, 20%, 15% and 10%) were prepared by mixing with wild-type DNA. These samples were analyzed as described below. DGGE was performed on a 6.5% acrylamide/polyacrylamide gel (16 cm long) with a denaturing gradient of 30% to 80% urea. Each of the above samples (20 μL) was run simultaneously at 160 V for 3.5 h at a constant temperature of 60°C (2) and stained in 1 μg/mL ethidium bromide in TAE buffer (40 mM Tris, 20 mM sodium acetate, 1 mM EDTA, pH 7.4). The two PCR products (20 μL each) that were included as controls were the wild-type and another mutant control that was heterozygous for the same polymorphism. Sequencing was performed using the Thermo Sequenase Cycle Sequencing Kit following the manufacturer’s instructions (Amersham International plc, Bucks, England, UK), with the reverse primer previously used for PCR and labeling with [α-33P]dATP. Figure 1 shows the DGGE pattern of eight mutant/wild-type DNA mixtures