Sensitivity of Electrospray Ionization Mass Spectrometry Detection of Codon 249 Mutations in the p53 Gene Compared with RFLP¹

HCC³ is a major cause of cancer death in certain parts of the world, including Asia and sub-Saharan Africa, accounting for over 200,000 deaths annually in the People’s Republic of China. The major etiological factors that have been associated with development of the disease in these regions are infection with hepatitis B or hepatitis C virus and exposure to high levels of AFB₁ in the diet (1, 2).

Mutations in the p53 tumor suppressor gene are found in a majority of human cancers, and distinct mutational spectra have been observed for different cancer types (3). One of the most striking examples of a “molecular fingerprint” in the p53 gene is a guanine to thymine (G to T) transversion at the third base of codon 249, resulting in an amino acid change of arginine to serine, that is found in up to 70% of HCCs from regions with high exposure to AFB₁ (4, 5, 6). In contrast, this mutation is absent from HCCs in regions with negligible levels of AFB₁ exposure (7). In vitro evidence also indicates that exposure to AFB₁ results induces a guanine to thymine transversion at codon 249 of the p53 gene (8, 9, 10, 11).

Several studies have now demonstrated that DNA isolated from the serum and plasma of cancer patients contains the same genetic aberrations as DNA isolated from the individual’s tumor (12, 13, 14, 15). The process by which tumor DNA is released into circulating blood is unclear but may be due to necrosis, apoptosis, or other processes (16). Recently, p53 mutations have been detected in DNA isolated from the plasma of individuals with HCC (17, 18). Because the specific G to T mutation at codon 249 results in the loss of a restriction enzyme site present in the wild-type sequence, Kirk et al. (17) were able to use RFLP to detect the mutations. In our report (18), we used an ESI-MS-based method called SOMA (19) to detect the p53 mutations. Here we describe a study in which the relative sensitivity of both methods was evaluated.

A DNA sample extracted from a human HCC that was known to have the G to T mutation at codon 249 of the p53 gene was serially diluted with DNA from a cell line with a wild-type p53 sequence. The mutant HCC sample has a mutant allele:wild-type allele ratio of 50:50 (as determined by SOMA analysis), and dilutions were therefore made from 50% mutant allele down to 0.1% mutant allele. In total, a series of 20 samples was prepared, and each sample was analyzed in triplicate by each method.

Liver tumor and normal liver samples were obtained as part of an ongoing prospective cohort investigation of liver cancer in high-risk areas of China. This collaboration between the Shanghai Cancer Institute and Johns Hopkins School of Public Health has been approved by each institution’s respective Institutional Review Board for Human Research.

RFLP was performed essentially as described by Kirk et al. (17). Using the primers p1 and p2 described by Kirk et al. (17), 100 ng of DNA for each of the samples in the series described above were amplified by PCR in a total volume of 50 μl. The reaction mixture contained 16.6 mm NH₄SO₄, 67 mm Tris (pH 8.8), 6.7 mm MgCl₂, 10 mm β-mercaptoethanol, 8% DMSO, 0.6 mm each of four deoxynucleoside triphosphates, 0.5 μm each primer, and 2.5 units of platinum Taq DNA polymerase (Life Technologies, Inc., Gaithersburg, MD). The thermocycling conditions were 94°C for 2 min; 35 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s; and a final extension of 72°C for 5 min. Negative controls (no DNA added) were included for each set of PCR reactions. The PCR product was purified by ethanol precipitation, and 5 μl were digested with 20 units of HaeIII (New England Biolabs) at 37°C for 4 h in a final volume of 50 μl. The digest reaction was purified by ethanol precipitation, the pellet was resuspended in 20 μl of Tris-EDTA buffer, and 4 μl (equivalent to 1 μl of original PCR product) were run on a 3% Metaphor agarose gel (BioWhittaker Molecular Applications, Rockland, ME). A control sample that did not have the mutation was included in each experiment to determine whether complete digestion was achieved.

SOMA was performed as described previously (18). Primers p53–8F1 and p53–8R1 were used to amplify 100 ng of DNA using the same PCR reaction mixture as for RFLP analysis. The thermocycling conditions were 95°C for 2 min; 40 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s; and a final extension of 72°C for 2 min. Negative controls (no DNA added) were included for each set of PCR reactions. The PCR product was purified by ethanol precipitation and digested with 8 units of BpmI (New England Biolabs) for 2 h at 37°C in a volume of 50 μl to release 8-bp internal fragments. A phenol-chloroform extraction, followed by an ethanol precipitation in the presence of SeeDNA (Amersham Pharmacia, Piscataway, NJ), was performed to purify samples for analysis by ESI-MS.

The digested fragments were resuspended in 10 μl of the HPLC mobile phase [70:30 (v:v) solvent A:solvent B, where solvent A was 0.4 m 1,1,1,3,3,3-hexofluoro-2-propanol (pH 6.9), and solvent B was 50:50 (v:v) 0.8 m 1,1,1,3,3,3-hexafluoro-2-propanol:methanol], and 8 μl were introduced into the HPLC coupled to the ESI-MS. HPLC was carried out at 30 μl/min using a 1 × 150-mm Luna C18, 5μ reversed phase column (Phenomenex, Torrance, CA) and Surveyor pumps (ThermoFinnigan Corp., San Jose, CA). The gradient conditions were 70% A:30% B programmed to 100% B in 5 min, where it was held for 10 min.

Mass spectra were obtained with a LCQ Deca ion-trap mass spectrometer (ThermoFinnigan Corp.) equipped with an electrospray ionization source operated in the negative ionization mode. The spray voltage was set at – 4.0 kV, and the heated capillary was held at 240°C. Each of the oligonucleotide ions was isolated in turn and subjected to collision-induced dissociation at 30% collision energy. Full scan spectra of the resultant fragment ions from m/z 600 to m/z 2000 were acquired, and signals from up to three specific fragment ions were summed as a function of time for each of the oligonucleotides. The mass spectrometer was programmed to acquire data in the centroid mode (1 μscan; 200 ms; isolation width 3 Da) using four scan events monitoring each [M − 2H]²⁻ oligonucleotide individually [scan event 1, AGG-s (5′-CGGAGCCC-3′), m/z 1256.3→600–2000; scan event 2, AGG-as (5′-CCTCCGGT-3′), m/z 1219.8→600–2000; scan event 3, AGT-s (5′-CGGAGTCC-3′), m/z 1244.3→600–2000; scan event 4, AGT-as (5′-ACTCCGGT-3′), m/z 1231.8→600–2000]. Reconstructed ion chromatograms were generated and smoothed from this raw data using an isolation width of 1.0 Da. The fragment ions used for each oligonucleotide were AGG-s (m/z 1047.3 + 1180.7), AGG-as (m/z 1268.6 + 1347.8 + 1637.2), AGT-s, (m/z 1437.4 + 1542.4), and AGT-as (m/z 1075.0). A sample was considered positive when fragments were observed in both sense and antisense channels for the mutant allele in at least three scans across the peak.

For samples that were predigested with HaeIII before SOMA analysis, 100 ng of DNA were incubated with 5 units of HaeIII in a volume of 10 μl at 37°C for 2 h. PCR was performed on 4 μl of this reaction mix using the conditions described above.

During the initial analyses of samples by RFLP, incomplete digestion was frequently seen in the control normal samples, and optimization of the method was required. The PCR buffer system, the amount of PCR product digested, and the amount of the digest reaction run on a gel were all examined. We found that using the buffer system described by Kirk et al. (17) did not always result in amplification product, whereas the buffer used for SOMA analysis was more robust. When large amounts of PCR product were digested, undigested product was routinely observed on the gel. We found that digesting 5 μl of PCR product gave the most sensitive and consistent results (data not shown). If the entire digested product was loaded onto the agarose gel, undigested bands were often seen in the normal controls. We therefore only loaded 4 of 20 μl of the purified product on the gel.

Having optimized the method, the serially diluted set of samples was analyzed by RFLP. Fig. 1 shows that complete digestion of the normal control (Lane 14) was achieved, and the expected bands for the wild-type alleles (66 and 92 bp) could be seen. Undigested product from the mutant allele (159 bp) could be detected in Lanes 2–8, corresponding to samples containing 50% to 6.25% mutant allele in the presence of wild-type allele. Mutant allele could not be detected at lower levels.

The serially diluted samples were analyzed by SOMA, and representative chromatograms are shown in Fig. 2. AGGs and AGGas represent the wild-type sense and antisense allele, and AGTs and AGTas are the mutant sense and antisense fragments, respectively. Thus, Fig. 2, A and B, depicts the 50:50 mix of alleles and 100% wild-type, respectively. Fig. 2,C is the chromatogram for the sample containing 6.25% mutant alleles that corresponds to the limit of sensitivity of the RFLP method. Fig. 2, D and E, illustrates analysis of the sample with 0.39% mutant alleles without and with HaeIII digestion before SOMA analysis, respectively. Mass chromatographic peaks indicating the presence of a mutation were consistently detectable in both sense and antisense alleles in samples containing 2.4% or more mutant alleles in the presence of wild-type alleles when there was no predigestion of the samples with HaeIII.

Digestion of DNA samples with HaeIII cleaves wild-type sequences, which are then unable to be amplified during PCR. There is thus enrichment for mutant alleles, leading to an increase in sensitivity of mutation detection by SOMA. When the serially diluted samples were analyzed by SOMA after a predigestion with HaeIII, peaks were detected for both the sense and antisense mutant alleles in the 0.15% mutant sample but were detected only in the sense channel for the 0.3% mutant sample. The limit of detection of mutant allele was therefore considered to be 0.39% because both sense and antisense alleles were consistently detected in these samples after predigestion with HaeIII (Fig. 2 E).

A set of DNA samples extracted from HCC and normal liver samples was analyzed for the p53 mutation by RFLP and SOMA. The same 5 of 26 samples were positive for the p53 mutation by RFLP and SOMA during the initial analysis (Table 1). Of these positive samples, four were detected in tumor tissue and one was detected in normal tissue from different individuals. SOMA was repeated using HaeIII predigestion, and an additional four samples were found to be positive for the mutation (Table 1). The p53 mutation was detected in the paired normal and tumor DNA from two individuals, in tumor DNA only from three individuals, and in normal DNA only from two individuals.

The presence of a highly specific p53 mutation in the sera of HCC patients may provide a biomarker that can be used for early detection of cancer or as an end point in intervention trials. To determine which of the two reported methods for detection of this specific mutation was more sensitive, we compared both methods using an identical set of serially diluted samples. The sensitivity of each method was determined by the sample with the lowest percentage of mutant allele in the serially diluted series in which a p53 mutation was consistently detected. The findings indicated that the electrospray mass spectrometry method, SOMA, is about 2.5 times more sensitive than RFLP, and this is increased to 15-fold more sensitive if the wild-type alleles are predigested with HaeIII before PCR amplification during SOMA.

When a set of HCC and normal liver samples was analyzed by RFLP and SOMA, a greater number of DNA samples from HCC patients were found to be positive for the p53 mutations by SOMA after predigestion with HaeIII compared with RFLP or SOMA with no predigestion. The increase in sensitivity provided by predigestion of the wild-type alleles allowed detection of a mutation in the normal tissue of patient 11, whose tumor tissue was seen to contain a mutation by all techniques. The presence of a mutation in the normal tissue of patient 6 may indicate an early lesion because mutations at codon 249 of the p53 gene have previously been detected in nonmalignant tissue, suggesting that this is an early event in tumorigenesis (20).

RFLP has the advantage of being a relatively simple technique that is easy to perform in any molecular biology laboratory. However, it relies on the fortuitous presence of a restriction endonuclease site that is either created or destroyed by the mutation of interest. For the aflatoxin-specific p53 mutation, a HaeIII site is present in the wild-type sequence that is lost when there is a mutation at the third base of codon 249. To avoid a false positive result, there must be complete digestion of the wild-type sequence, and optimization of the digest conditions was required to achieve reliable digestion. For samples with a high percentage of mutant allele relative to wild-type allele, this method provides a simple and effective mutation detection methodology. However, for low levels of mutant allele in the presence of a large amount of wild-type allele, great care has to be taken to avoid false positive results, and subsequent confirmation with another technique, such as DNA sequencing, is required.

The presence of mutations at the second base of codon 249 or the first and second bases of codon 250 of the p53 gene will also result in the loss of the HaeIII restriction endonuclease site, and the RFLP method is therefore not specific to the G to T transversion mutation at the third base of codon 249. In contrast, SOMA will only detect the specific mutation of interest because the mass spectrometer is programmed to monitor only the mass to charge ratios of the oligonucleotides produced by digestion. An additional level of specificity is achieved by fragmenting isolated parent ions and monitoring the resultant daughter ions. Mutation analysis by SOMA is very specific, and the incidence of false positive results is minimized by the fact that a gain of signal, rather than loss of signal, is being monitored.

In addition to the better sensitivity of SOMA compared with RFLP, the use of mass spectrometry as the detection method provides the potential to develop SOMA as a quantitative method. In order for this to be achieved after predigestion of the wild-type alleles, an internal standard will need to be developed against which the mutant alleles can be quantitated. A quantitative approach would have important applications in using the p53 codon 249 mutation as a biomarker for aflatoxin exposure and HCC development.