Clinical Chemistry 48:8 1321–1328 (2002)
Cancer Diagnostics: Discovery and Clinical Applications
Homogeneous Amplification and Mutation Scanning of the p53 Gene Using Fluorescent Melting Curves Haleigh Millward,1 Wade Samowitz,2 Carl T. Wittwer,2 and Philip S. Bernard2* Background: In malignancy, gene mutations frequently occur in tumor suppressor genes such as p53 and are sporadically located. We describe a homogeneous method for amplification and mutation scanning, and apply the method to the p53 gene. Methods: Using a series of overlapping fluoresceinlabeled oligonucleotides complementary to a wild-type p53 sequence, we detected somatic mutations in colorectal cancers by aberrant probe:target melting temperatures (Tm). The probes were designed so that fluorescence decreased on target annealing as a result of deoxyguanosine quenching. Probes were walked along the sequence to be scanned, using two to three probes per cuvette and placing overlapping probes in separate reactions. After amplification, the reaction was cooled to anneal probes and then slowly heated (0.1 °C/s) while fluorescence was continuously monitored. Somatic mutations in tumor tissue were detected by changes from a characteristic wild-type melting curve profile using leukocyte DNA. Results: A complete scanning of the DNA binding domain (exons 5– 8) of the p53 gene was completed in a single run (!30 min) starting from genomic leukocyte DNA. To show proof-of-principle, p53 exons 6 – 8 from 63 colon cancers were probe-scanned and showed 100% agreement with direct sequencing for detecting alterations from wild-type DNA. Conclusions: p53 mutation scanning by single-labeled hybridization probes is a homogeneous, rapid, and sensitive method with application in both research and clinical diagnostics. © 2002 American Association for Clinical Chemistry
1
Idaho Technology Inc., Salt Lake City, UT 84105. University of Utah School of Medicine, Department of Pathology, Salt Lake City, UT 84132. *Address correspondence to this author at: University of Utah School of Medicine, Department of Pathology, 30 North 1900 East, Salt Lake City, UT 84132. Fax 801-581-4517; e-mail
[email protected]. Received in revised form March 21, 2002; accepted May 2, 2002. 2
The p53 tumor suppressor gene (17p13.1) encodes a multifunctional transcription factor that orchestrates cell cycle events, including transcription, DNA replication and repair, differentiation and development, and programmed cell death (1–3 ). The 393-amino acid nuclear phosphoprotein can be divided into five structural and functional domains that affect and are affected by various modulators (4 ). From the 11 exons, the first being noncoding, exons 5– 8 (codons 102–292) encode the DNAbinding domain (5, 6 ). Among the 10 000 documented somatic mutations in p53, 93% are found in this region (7 ). Mutations within the DNA-binding domain disrupt the proper function of p53, particularly when they occur within critical hotspots necessary for protein:DNA contact (codons 248 and 273) or stability (codons 175, 249, and 282). These sites account for !25% of the mutations in cancers overall (4, 7 ), but many cancers acquire their own unique mutation patterns based on exogeneous and endogenous factors (4, 8 ). Finally, mutations within the DNA-binding domain show prognostic and predictive significance in common cancers, such as head and neck (9 ), colorectal (10 –12 ), and breast cancer (13–18 ). Despite the demonstrated value of p53 to cancer medicine, routine clinical testing of p53 continues to be debated (19 ). Reasons for this include conflicting results between testing methodologies and finding a method that is amenable to use in the clinical laboratory. Immunohistochemistry is a convenient method to score for molecular markers in anatomic pathology, but p53 protein overexpression and p53 mutation status have poor concordance in most tumors (20, 21 ). Furthermore, immunohistochemistry is less specific than nucleic acid analysis and is subject to variability because of differences in antibody specificity, scoring criteria, and storage (22, 23 ). Additional advantages of using nucleic acid analysis are that it allows specific mutations to be correlated with treatment response and it provides a marker for monitoring residual disease (24, 25 ). Various scanning methods and direct sequencing are used to establish associations between genotype and
1321
1322
Millward et al.: Mutation Scanning by Fluorescent Melting Curves
disease (26 ). Conventional scanning methodologies use slab gels to resolve mobility shifts attributable to mutations that cause single-stranded conformational changes or changes in heteroduplex melting. Gel methods that differentiate homo- and heteroduplexes by exploiting differences in duplex stability include denaturing gradient gel electrophoresis (DGGE)3 (27 ), constant denaturant gel electrophoresis (CDGE) (28 ), and temporal temperature gradient gel electrophoresis (29 ). Alternatively, heteroduplexes may be detected by chemical or enzymatic cleavage followed by gel fragment sizing (30 ). Advances in scanning methodologies include the incorporation of fluorescence with single-strand conformational polymorphism (SSCP) analysis and the use of denaturing HPLC (dHPLC) (31–33 ), both of which offer more potential for speed, resolution, and automation than slab gels. In contrast to scanning methods, many methods for the detection of known mutations use real-time homogeneous genotyping (34 ). These technologies include genotyping of known mutations/polymorphisms by use of hybridization probes (35, 36 ), exonuclease probes (TaqMan) (37 ), hairpin probes (Beacons) (38 ), and self-probing amplicons (Scorpions) (39 ). In this study, we extended the application of homogeneous genotyping by fluorescent hybridization probes beyond the identification of known mutations to scanning for unknown mutations. Single-labeled fluorescein probes were designed complementary to the wild-type p53 sequence. Overlapping probes covering exons 5– 8 were used to scan for mutations by comparing the melting curve profile for wild-type DNA to that of tumor DNA. The method has widespread application in research and clinical laboratories for the repetitive scanning of genes known to acquire germline or somatic mutations.
(http://www.iarc.fr/p53/Index.html). Information about gender, grade, and other gene mutations in these samples is available elsewhere (40, 42 ). All identifiers were removed from the samples in this study, and the investigators were blinded to the position and type of mutation found by sequencing until after probe scanning. Tumor samples were chosen so that there would be approximately the same number of mutations in each exon. An amplicon not used for sequencing was diluted 1:1000 in 10 mmol/L Tris-HCl (pH 8.0) for reamplification and scanning with the LightCyclerTM (Roche). Genomic leukocyte DNA was used as a control for comparison to tumor profiles (tumors scanned for exons 6 – 8) and to illustrate feasibility of complete scanning of the DNA-binding domain (exons 5– 8). Genomic leukocyte DNA was prepared from whole blood with the PUREGENETM DNA Isolation Kit (Gentra Systems) and according to the manufacturer’s instructions. The concentration was adjusted to 50 ng/!L genomic DNA (A260 " 1).
Materials and Methods
Primers and probes were selected from the genomic wild-type p53 sequence (GenBank accession no. U94788). Primers were positioned to allow complete scanning of each exon and intron-exon boundary. Primer Designer© 4 (Scientific and Educational Software) was used to minimize primer-dimers and hairpins. Primer uniqueness was verified using BLAST (http://www.ncbi.nlm.gov). All primers were designed with melting points (Tm) of 61 $ 2 °C, and probes used in Tm multiplexing were designed for an !10 °C separation. Primer and probe Tms were calculated using Tm Utility (http://www.idahotech.com). Each exon was asymmetrically amplified with probes designed to hybridize to the excess strand. Each amplicon scanned contained the entire exon, except for exon 5, which was scanned from two overlapping amplicons. Each amplicon was between 200 and 250 bp in length. Each probe was designed so that a fluorescein molecule, conjugated to a terminal deoxycytosine, would approximate two to three G residues on annealing to its complementary target. There is maximum fluorescence quenching when the fluorescein anneals opposite at least two G residues: one residue in direct contraposition to the fluo-
samples
Hybridization probe scanning for p53 exons 6 – 8 was performed on diluted amplicon DNA from 63 colon cancers that were previously analyzed for p53 mutations by SSCP analysis and direct sequencing (40 ). Each sample had been microdissected and the DNA extracted from formalin-fixed, paraffin-embedded tissue blocks, as described previously (41 ). Eligible participants in the initial study were men and women with any stage of primary adenocarcinoma or carcinoma of the colon diagnosed between 30 and 79 years of age and without a family history of colon disease (e.g., adenomatous polyposis, Crohn disease, or ulcerative colitis) (40 ). The specific mutations and polymorphisms reported have been submitted to the International Agency for Research on Cancer
3 Nonstandard abbreviations: DGGE, denaturing gradient gel electrophoresis; CDGE, constant denaturant gel electrophoresis; SSCP, single-strand conformational polymorphism; dHPLC, denaturing HPLC; and Tm, melting temperature.
primer and probe synthesis Primers and probes were obtained from ITBiochem (Idaho Technology Inc., Salt Lake City, UT) and synthesized as described previously (43 ). Briefly, the 5#-fluoresceinlabeled oligonucleotides were synthesized using a phosphate-controlled pore glass cassette (Glen Research) to block extension from the 3# end. Fluorescein was directly incorporated to the 5# end by use of an amidite (BioGenex). The 3#-fluorescein-labeled oligonucleotides were synthesized using a fluorescein-controlled pore glass cassette (BioGenex). All probes used had calculated concentration fluorescein-to-oligonucleotide ratios of 0.8 –1.2.
primer and probe selection
Clinical Chemistry 48, No. 8, 2002
rescein, and the other in the first base overhang position (44 ). Overlapping hybridization probes were placed in separate reaction vessels so to not interfere with hybridization. Probes were designed to overlap by approximately 5 bp on either end. Primer and probe sequences are provided in an online supplement at LightCycler University (http://www.idahotech.com/lightcycler_u/ index.html).
homogeneous pcr and scanning PCR and scanning were performed in 10-!L reactions that included 50 ng of genomic leukocyte DNA or 1 !L of colon cancer amplicon (!1 % 105 copies); 200 !M each of dATP, dCTP, dGTP, and dTTP; 50 mM Tris, pH 8.3 (25 °C); 3 mM MgCl2, 500 ng/!L bovine serum albumin, 0.1 !M site-specific fluorescein-labeled probe; 2 U of KlenTaq1TM (AB Peptides); and 0.09 !g of TaqStart antibody (CLONTECH Laboratories). In addition, the primer extending the strand to which the probes annealed was included at 0.5 !M, and the opposite primer was present at 0.2 !M. Each reaction was placed in a plastic/glass cuvette and amplified in the LightCycler through 40 cycles of PCR consisting of 3 steps: 95 °C for 0 s; 57 °C for 10 s; and 72 °C for 0 s. A single fluorescence acquisition was taken each cycle at the end of the 57 °C annealing step. The temperature transition rate between annealing (57 °C) and extension (72 °C) was 2 °C/s. Otherwise, the temperature ramp rates were programmed at 20 °C/s. After PCR, the LightCycler immediately executed a melting curve program consisting of brief denaturation (94 °C for 0s; 20 °C/s), rapid cooling (20 °C/s) down to a hold for probe annealing (40 °C for 90 s), and a melting ramp at 0.1 °C/s to 94 °C while continuously monitoring fluorescence.
detection of sequence alterations Colon tumor tissue DNA and DNA from healthy leukocytes were compared using melting curve profiles generated by plotting temperature vs dF/dT [derivative of fluorescence (F) with respect to temperature (T)]. Because quenching probes have an increase in fluorescence on melting, the derivative of fluorescence rather than the negative derivative was plotted to show “right-side-up” melting peaks. Fluorescence values were normalized for a baseline of zero. A tumor sample melting profile was considered aberrant if, when compared with the wildtype profile, the Tm (or &Tm) exceeded the expected Tm by 2 SD, there was an obvious change in peak areas, or there was an additional peak that could not be accounted for by a polymorphism. The sensitivity of single-labeled probes was determined from mixing studies using a true heterozygous sample (A639G) (45 ).
p53 scanning
Results
A total of 29 single-labeled probes divided among 13 reaction vessels provided complete scanning of the p53
1323
DNA-binding domain (exons 5– 8). Fig. 1 provides an overview of the scanning technique and specifically illustrates scanning exon 7 of p53 by use of six walking probes. PCR and scanning were done homogeneously, and overlapping probes were separated into two reaction vessels so to not interfere with probe hybridization. The detection of a point mutation in exon 7 of p53 is shown in Fig. 2. All probe:target duplexes were completely complementary in the wild-type sample (Fig. 2A), but the colon cancer sample contained an allele with a T-to-C base change at codon 248 that created a destabilizing C:A mismatch under the first (lowest Tm) probe (Fig. 2B). Melting curve analysis showed a single melting curve transition for each site in the wild-type control (Fig. 2C). By comparison, the tumor sample had an extra transition during the first probe melt attributable to the mutation (Fig. 2D). Thus, the derivative melting curve profile gave the expected three peaks for the wild-type control (Fig. 2E), but four peaks for the colon cancer sample (Fig. 2F). Note that in Fig. 2, quenching probes give an increase in fluorescence on melting and are plotted as dF/dT (rather than 'dF/dT) to show peaks with positive values. Mutation detection sensitivity for single-labeled probes was tested using a sample with the A639G polymorphism and a homozygous wild-type sample. Fig. 3 shows the resulting melting curves after amplification of different ratios of alleles. The true (50:50) heterozygous sample gave nearly symmetric derivative melting curves. The allele with the polymorphism could be detected even when diluted 1:20 (95% background) in the homozygous wild-type sample.
comparing scanning with direct sequencing Sixty-three colon cancers were both scanned and sequenced for exons 6 – 8 of the p53 gene. We detected p53 mutations in 57 of 63 colon cancer samples by probe scanning. Six of the 63 samples had no alterations de-
Fig. 1. Schematic for scanning the p53 DNA-binding domain. Primers and single-labeled fluorescein probes are included from the beginning of the reaction for homogeneous genotyping. Primers are positioned within introns to scan the entire exon and the intron/exon boundaries. Overlapping probes are walked along the region(s) to be scanned. For example, exon 7 was spanned using six probes. Probes 1, 3, and 5 (' ' ') were placed in a separate reaction from probes 2, 4, and 6 (——–) to prevent interference with hybridization. Multiple sites are interrogated within a single reaction by designing probes with different Tms.
1324
Millward et al.: Mutation Scanning by Fluorescent Melting Curves
Fig. 2. Mutation detection in exon 7 of p53. Shown are hybridization probe scanning profiles comparing a wild-type sample with a colon cancer sample containing a point mutation in exon 7. After PCR, the reactions were cooled, and the probes were hybridized to single-stranded product. The wild-type sample had both alleles completely complementary to the scanning probes (A), whereas the tumor sample had an allele with a T-to-C base change at codon 248, creating a C:A mismatch (B). A plot of temperature vs fluorescence shows a single transition for each site in the wild-type sample (C). The same plot of the tumor sample shows an additional melting transition for the first probe site representing the presence of both a wild-type and a mutant allele (D). Thus, a plot of the derivative melting curves (temperature vs dF/dT) gives three peaks for the wild-type sample (E), but four peaks for the tumor sample (F).
tected by probe scanning. There was 100% concordance between probe scanning and sequencing for the detection of an alteration from wild type. Within the 57 samples
Fig. 3. Detecting p53 alterations from a background of wild type. The sensitivity for detecting an alteration was tested by mixing a wild-type sample with a germline heterozygous sample containing the A639G polymorphism. Derivative melting curves (temperature vs dF/dT) are shown for a homozygous sample (' ' '), a 50:50 heterozygous sample (——–), and a 95:5 mixture of the wild-type allele to polymorphic allele (' ! ! ').
with mutations/polymorphisms, there were 62 variants identified by direct sequencing and 63 detectable Tm shifts under the scanning probes. Five samples had two mutations/polymorphisms detected by sequencing. In three cases, both mutations were present under one probe, producing a single Tm shift, and in two cases, the mutations produced Tm shifts for two probes. There were four samples in which one mutation occurred in an overlapping probe region, creating a detectable Tm shift under two probes. Single-base changes accounted for 58 of 62 (93%) of the alterations. The other four mutations were small deletions (C deletion and GC deletion) and insertions (ACT insertion and T insertion). Shown in Table 1 are the numbers of resolvable and nonresolvable Tm shifts produced by different mutations under each probe used to scan exons 6 – 8. A resolvable profile was counted when a sequence was different from other samples and produced a unique &Tm. Uniqueness was defined by having a &Tm for a given reaction (2 SD away from any other aberrant profile. Overall, there were 48 mutations that gave unique melting curve profiles. Shown in Table 2 are data analyses from 20 runs scanning p53 exons 5– 8 in wild-type leukocyte DNA. The absolute Tm, &Tm, and SD were determined for probes in each reaction. We found that using the &Tm between two probes at different sites in the same reaction is more precise than using a shift in a single Tm (F-test, P )0.1).
Table 1. Mutations producing resolvable and nonresolvable Tm shifts. Resolvable
Exon 6 Probe Probe Probe Probe Probe Probe Exon 7 Probe Probe Probe Probe Probe Probe Exon 8 Probe Probe Probe Probe Probe Probe Total a
1 2 3 4 5 6
0 4 4 0 2 4
0 4a 0 0 0 0
1 2 3 4 5b 6
0 3 2 5 2 5
0 0 0 0 2c 0
1 2 3 4 5 6
1 6 2 3 4 1 48
0 0 0 0 0 0 6
Four different mutations produced the same Tm shift. Hybridizes over hotspot codon 248. c Two different mutations produced the same Tm shift. b
Nonresolvable
1325
Clinical Chemistry 48, No. 8, 2002
Table 2. Comparison of absolute Tm and !Tm for precision of mutation detection. Tm 1, °C
Exon 5aa 1, 3 3, 5 2, 4 Exon 5b 1, 5 2, 4 3, 6 Exon 6 1, 5 2, 4 3, 6 Exon 7a,b 1, 3 3, 5 2, 4 4, 6 Exon 8 1, 4 2, 5 3, 6 a b
Tm 2, °C
!Tm, °C
Mean
SD
Mean
SD
Mean
SD
58.4 69.1 66.4
0.26 0.19 0.24
69.1 74.9 73.6
0.19 0.20 0.19
10.7 5.8 7.2
0.21 0.14 0.13
66.1 63.3 70.3
0.22 0.18 0.23
73.1 71.3 75.0
0.29 0.21 0.25
7.0 8.0 4.8
0.26 0.18 0.07
54.9 61.2 70.1
0.28 0.29 0.29
68.7 68.8 76.4
0.26 0.36 0.26
13.7 7.6 6.3
0.14 0.24 0.13
54.2 60.7 58.3 69.1
0.24 0.20 0.25 0.21
60.7 70.8 69.1 75.2
0.20 0.22 0.21 0.29
6.4 10.1 10.7 6.2
0.20 0.12 0.10 0.16
65.9 68.0 66.3
0.30 0.38 0.28
71.5 73.7 79.1
0.29 0.26 0.23
5.7 5.7 12.8
0.20 0.27 0.28
Probes 1, 3, and 5 were multiplexed in the same reaction. Probes 2, 4, and 6 were multiplexed in the same reaction.
Discussion Homogeneous genotyping methods commonly use fluorescence resonance energy transfer or reporter-quencher systems for real-time allele identification (34 ). For example, genotyping on the LightCycler uses oligonucleotides labeled with either a donor or acceptor fluorophore. During probe/target hybridization, these fluorophores are brought into close proximity and fluorescence resonance energy transfer occurs. Genotyping is performed by slowly heating after amplification and monitoring changes in fluorescence as the probe:target duplex thermally denatures and the dyes physically separate. Other homogeneous genotyping methods, such as those using exonuclease (37 ) or conformational probes (38, 39 ), score alleles during PCR and require two dyes to identify each allele. Color multiplexing allows these methods to genotype multiple alleles in a single reaction (46 ). In contrast, hybridization probes can identify multiple alleles with a single probe (43 ). Additional sites can be identified by color multiplexing (47– 49 ), Tm multiplexing (50 ), or both. The use of single-fluorescein-labeled probes designed for quenching on target hybridization allows melting curve analysis without the need for a dye pair (44 ). This simplifies the design, synthesis, and purification of hybridization probes. It also reduces cost and makes scanning feasible. Multiple sites can be interrogated with a single color by use of Tm multiplexing, but there is a potential problem in that a mutation may not be detected if the Tm for a variant
allele is within the melting range of another probe. For this reason, slight changes in &Tm and changes in peak areas were important issues to consider when comparing a tumor profile to a wild-type control. Proper Tm spacing for each probe site can reduce this potential problem. For example, a 10 °C separation between wild-type probe Tms usually provides a large enough temperature range that even a very destabilizing single-base change will not cause a shift into the lower melting probe. The extent of destabilization also depends on probe length and sequence. Although a shift into the melting domain of another probe site is certainly possible, changes from a wild-type profile can be easily detected, as shown by the 100% agreement between probe scanning and SSCP analysis/sequencing in the samples investigated. The sensitivity for detecting a mutation in a background of wildtype allele can be as high as 95%, depending on the experimental conditions. Single-labeled probes are currently limited to Tm multiplexing, but other dyes that change fluorescence with hybridization could be used with fluorescein to provide color multiplexing and increase the power of the method. The estimated Tm from the derivative melting curve is highly reproducible under constant reaction conditions of heating rate, salt concentration, and probe:target concentrations (51 ). We have shown that single-labeled probes are also highly accurate. When we analyzed the scanning results from 20 wild-type samples run over different days and on different instruments, the between-run SD for &Tm
1326
Millward et al.: Mutation Scanning by Fluorescent Melting Curves
ranged from 0.07 °C (0.18 °C for absolute Tm shift) to 0.28 °C (0.38 °C for absolute Tm shift). Depending on the particular probe set used and the method of analysis (&Tm or absolute Tm), Tm shifts as small as 0.2 °C from that empirically tested can be suspicious for a sequence alteration. Using a theoretical nearest-neighbor model and generating random sequence by a computer, von Ahsen and coworkers (52, 53 ) found that only 0.055% of mismatches should cause a melting shift lower than 1.25 °C in a 19mer probe. This excludes mismatches in the ultimate and penultimate positions, which may produce smaller Tm shifts. For this reason, we scanned with probes that overlapped on the ends to provide increased sensitivity. This technique should allow all possible small alterations to be detected. The p53 gene is mutated in most human cancers by the time of invasive growth (54 ). The types of mutations and their frequencies and distributions vary depending on environmental exposures and endogenous factors (4 ). For example, benzo[a]pyrene exposure from tobacco smoke causes an unusual predominance of G-to-T base substitutions that cluster in hotspots unique to lung cancer (8 ). Other cancers frequently acquire CpG transitions (23% of all mutations) in p53 (4, 55 ). These alterations occur by methylation of cytosine (C to T) or deamination of 5-methycytosine (G to A) and are particularly common in colorectal cancer (45%) and adenocarcinoma of the esophagus (48%). Databases of somatic p53 mutations help characterize the spectrum of alterations that occur in p53 in different cancers and can be used to develop more tailored scanning assays (7 ). For example, an assay for colorectal cancer may use additional probes that are specific (i.e., complementary) for mutations in CpG hotspot codons 175, 213, 245, 248, 273, and 282 (4 ). Such an assay design would show a Tm shift for any alteration other than the mutation of interest. This type of scheme has been used for genotyping the factor V Leiden (G1691A) mutation, in which polymorphisms (G1689A and A1692C) under the genotyping probe can decrease specificity (51 ). Of the 57 colon cancer samples with detectable p53 mutations in exons 6 – 8, 17 (30%) of those tumors had mutations in CpG hotspots. We did not perform mutation detection on the colon cancers for exon 5 (containing codon 175) because the exon 5 amplicon that was initially generated for the colon cancers in a previous study did not contain the primer sites that were used for our genomic DNA amplification. Nevertheless, we found a high preponderance of CpG mutations in the five hotspots scanned in exons 6 – 8. Although both the nearest-neighbor model and our p53 scanning results show that properly designed probes can detect any mismatch, some mutations may not be resolvable from others by their Tm shift. For example, probe 5 in exon 6 of p53 spanned the A639G polymorphism in codon 213 (45 ). This polymorphism, which was present in two samples, caused the same apparent Tm shift as a mutation
in codon 216. The polymorphism created an A:C mismatch 11 bp from the 5# end of the 27mer probe, whereas the mutation created a G:T mismatch 18 bp from the 5# end. An A:C mismatch is more destabilizing than a G:T mismatch (56, 57 ), but the former was flanked by stabilizing G:C neighbors and the latter was flanked by less stabilizing A:T neighbors (58 ), producing the same Tm shift. Shifts in Tm created by polymorphisms can be easily clarified by including a patient’s genomic leukocyte DNA as a proper control to be compared with their tumor DNA. For example, if tumor and germline DNA produce the same Tm profile, then the variant is either a polymorphism or, less likely, a germline mutation. In assessing the effectiveness of a mutation detection scheme, several factors must be considered, including sensitivity, specificity, and cost. The sensitivity of a scanning method is highly dependent on the assay design, the specific target, and the type of mutation being detected. For example, DGGE and CDGE are sensitive to singlebase changes but require a GC clamp for robust melting profiles and have decreased sensitivity in GC-rich regions (DGGE) (27 ) or regions containing multiple melting domains (CDGE) (28 ). More recent methods of scanning include fluorescent SSCP analysis and dHPLC; however, both of these methods require extensive optimization of temperature and buffers to achieve the proper running conditions for mutation detection (31, 33 ). Although probe scanning has high sensitivity, it is generally less specific than sequencing. In all cases in which SSCP or dHPLC detects a mutation, sequencing is needed for confirmation. In only some cases will sequencing be needed after probe scanning because the use of complementary probes for hot spots (most mutations) will decrease the need for a second step. We estimate that the reagent expense for probe scanning of p53 exons 5– 8 is between the cost of SSCP (least expensive) and sequencing (most expensive). The major advantages that probe scanning has over fluorescent SSCP, dHPLC, sequencing, and microarrays are that it is homogeneous and faster, which are important for labor costs, sample tracking, contamination, and turnaround time. In conclusion, hybridization probe scanning is well suited for the detection of single-base substitutions and small alterations such as those encountered in the p53 gene. It is a promising method that is sensitive, homogeneous, and faster (15-min postamplification analysis) than existing methods. In addition, hybridization probe schemes directed at particular hotspots may in many cases allow commonly encountered mutations to be identified without sequencing. Because of the up-front cost of probe synthesis, probe scanning is best suited for the repeated analysis of genes known to have sporadic or germline mutations. This work was supported by the NIH STTR program (Grant GM60063-03). The LightCycler and related tech-
Clinical Chemistry 48, No. 8, 2002
nologies are licensed from the University of Utah to Idaho Technology and Roche Applied Systems. C.T.W. holds equity interest in Idaho Technology; P.S.B. is a consultant for Idaho Technology. A table listing the sequences of the primers and probes used in this study as well as the locations of the probe targets is available through the online version of this article, available at the Clinical Chemistry Online web site (http://www.clinchem.org/ content/vol48/issue8/).
References 1. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000;408:307–10. 2. Agarwal ML, Taylor WR, Chernov MV, Chernova OB, Stark GR. The p53 network. J Biol Chem 1998;273:1– 4. 3. Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997;88:323–31. 4. Hainaut P, Hollstein M. p53 and human cancer: the first ten thousand mutations. Adv Cancer Res 2000;77:81–137. 5. Pavletich NP, Chambers KA, Pabo CO. The DNA-binding domain of p53 contains the four conserved regions and the major mutation hot spots. Genes Dev 1993;7:2556 – 64. 6. Cho Y, Gorina S, Jeffrey PD, Pavletich NP. Crystal structure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations. Science 1994;265:346 –55. 7. Hernandez-Boussard T, Rodriguez-Tome P, Montesano R, Hainaut P. IARC p53 mutation database: a relational database to compile and analyze p53 mutations in human tumors and cell lines. International Agency for Research on Cancer. Hum Mutat 1999; 14:1– 8. 8. Hernandez-Boussard TM, Hainaut P. A specific spectrum of p53 mutations in lung cancer from smokers: review of mutations compiled in the IARC p53 database. Environ Health Perspect 1998;106:385–91. 9. Erber R, Conradt C, Homann N, Enders C, Finckh M, Dietz A, et al. TP53 DNA contact mutations are selectively associated with allelic loss and have a strong clinical impact in head and neck cancer. Oncogene 1998;16:1671–9. 10. Goh HS, Yao J, Smith DR. p53 point mutation and survival in colorectal cancer patients. Cancer Res 1995;55:5217–21. 11. Goh HS, Elnatan J, Low CH, Smith DR. p53 point mutation and survival in colorectal cancer patients: effect of disease dissemination and tumour location. Int J Oncol 1999;15:491– 8. 12. Pricolo VE, Finkelstein SD, Hansen K, Cole BF, Bland KI. Mutated p53 gene is an independent adverse predictor of survival in colon carcinoma. Arch Surg 1997;132:371– 4, discussion 374 –5. 13. Borresen AL, Andersen TI, Eyfjord JE, Cornelis RS, Thorlacius S, Borg A, et al. TP53 mutations and breast cancer prognosis: particularly poor survival rates for cases with mutations in the zinc-binding domains. Genes Chromosomes Cancer 1995;14: 71–5. 14. Bergh J, Norberg T, Sjogren S, Lindgren A, Holmberg L. Complete sequencing of the p53 gene provides prognostic information in breast cancer patients, particularly in relation to adjuvant systemic therapy and radiotherapy. Nat Med 1995;1:1029 –34. 15. Aas T, Borresen AL, Geisler S, Smith-Sorensen B, Johnsen H, Varhaug JE, et al. Specific P53 mutations are associated with de novo resistance to doxorubicin in breast cancer patients. Nat Med 1996;2:811– 4. 16. Berns EM, Foekens JA, Vossen R, Look MP, Devilee P, HenzenLogmans SC, et al. Complete sequencing of TP53 predicts poor response to systemic therapy of advanced breast cancer. Cancer Res 2000;60:2155– 62.
1327
17. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 2001;98:10869 –74. 18. Geisler S, Lonning PE, Aas T, Johnsen H, Fluge O, Haugen DF, et al. Influence of TP53 gene alterations and c-erbB-2 expression on the response to treatment with doxorubicin in locally advanced breast cancer. Cancer Res 2001;61:2505–12. 19. Yarbro JW, Page DL, Fielding LP, Partridge EE, Murphy GP. American Joint Committee on Cancer prognostic factors consensus conference. Cancer 1999;86:2436 – 46. 20. Soong R, Robbins PD, Dix BR, Grieu F, Lim B, Knowles S, et al. Concordance between p53 protein overexpression and gene mutation in a large series of common human carcinomas. Hum Pathol 1996;27:1050 –5. 21. Sjogren S, Inganas M, Norberg T, Lindgren A, Nordgren H, Holmberg L, et al. The p53 gene in breast cancer: prognostic value of complementary DNA sequencing versus immunohistochemistry. J Natl Cancer Inst 1996;88:173– 82. 22. Baas IO, Mulder JW, Offerhaus GJ, Vogelstein B, Hamilton SR. An evaluation of six antibodies for immunohistochemistry of mutant p53 gene product in archival colorectal neoplasms. J Pathol 1994;172:5–12. 23. Jacobs TW, Prioleau JE, Stillman IE, Schnitt SJ. Loss of tumor marker-immunostaining intensity on stored paraffin slides of breast cancer. J Natl Cancer Inst 1996;88:1054 –9. 24. Shao ZM, Wu J, Shen ZZ, Nguyen M. p53 mutation in plasma DNA and its prognostic value in breast cancer patients. Clin Cancer Res 2001;7:2222–7. 25. Gonzalez R, Silva JM, Sanchez A, Dominguez G, Garcia JM, Chen XQ, et al. Microsatellite alterations and TP53 mutations in plasma DNA of small-cell lung cancer patients: follow-up study and prognostic significance. Ann Oncol 2000;11:1097–104. 26. Shi MM. Enabling large-scale pharmacogenetic studies by highthroughput mutation detection and genotyping technologies. Clin Chem 2001;47:164 –72. 27. Ferretti G, Curigliano G, Pastorino U, Cittadini A, Flamini G, Calabro MG, et al. Detection by denaturant gradient gel electrophoresis of tumor-specific mutations in biopsies and relative bronchoalveolar lavage fluid from resectable non-small cell lung cancer. Clin Cancer Res 2000;6:2393– 400. 28. Borresen AL, Hovig E, Smith-Sorensen B, Malkin D, Lystad S, Andersen TI, et al. Constant denaturant gel electrophoresis as a rapid screening technique for p53 mutations. Proc Natl Acad Sci U S A 1991;88:8405–9. 29. Chen TJ, Boles RG, Wong LJ. Detection of mitochondrial DNA mutations by temporal temperature gradient gel electrophoresis. Clin Chem 1999;45:1162–7. 30. Taylor GR. Enzymatic and chemical cleavage methods. Electrophoresis 1999;20:1125–30. 31. Moore L, Godfrey T, Eng C, Smith A, Ho R, Waldman FM. Validation of fluorescent SSCP analysis for sensitive detection of p53 mutations. Biotechniques 2000;28:986 –92. 32. Marsh DJ, Theodosopoulos G, Howell V, Richardson AL, Benn DE, Proos AL, et al. Rapid mutation scanning of genes associated with familial cancer syndromes using denaturing high-performance liquid chromatography. Neoplasia 2001;3:236 – 44. 33. Ellis LA, Taylor CF, Taylor GR. A comparison of fluorescent SSCP and denaturing HPLC for high throughput mutation scanning. Hum Mutat 2000;15:556 – 64. 34. Foy CA, Parkes HC. Emerging homogeneous DNA-based technologies in the clinical laboratory. Clin Chem 2001;47:990 –1000. 35. Bernard PS, Reiser A, Pritham GH. Mutation detection by fluorescent hybridization probe melting curves. In: Wittwer C, Meuer S,
1328
36.
37.
38. 39.
40.
41.
42.
43. 44.
45. 46.
47.
48.
Millward et al.: Mutation Scanning by Fluorescent Melting Curves
Nakagawara K-I, eds. Rapid cycle real-time PCR methods and applications. Heidelberg: Springer-Verlag, 2001:11–9. Bernard PS, Wittwer CT. Homogeneous amplification and variant detection by fluorescent hybridization probes. Clin Chem 2000; 46:147– 8. Lee LG, Connell CR, Bloch W. Allelic discrimination by nicktranslation PCR with fluorogenic probes. Nucleic Acids Res 1993; 21:3761– 6. Tyagi S, Kramer FR. Molecular beacons: probes that fluoresce upon hybridization. Nat Biotechnol 1996;14:303– 8. Whitcombe D, Theaker J, Guy SP, Brown T, Little S. Detection of PCR products using self-probing amplicons and fluorescence. Nat Biotechnol 1999;17:804 –7. Samowitz WS, Curtin K, Ma KN, Edwards S, Schaffer D, Leppert MF, Slattery ML. Prognostic significance of p53 mutations in colon cancer at the population level. Int J Cancer 2002;99:597– 602. Spirio LN, Samowitz W, Robertson J, Robertson M, Burt RW, Leppert M, et al. Alleles of APC modulate the frequency and classes of mutations that lead to colon polyps. Nat Genet 1998;20:385– 8. Samowitz WS, Curtin K, Ma KN, Schaffer D, Coleman LW, Leppert M, et al. Microsatellite instability in sporadic colon cancer is associated with an improved prognosis at the population level. Cancer Epidemiol Biomarkers Prev 2001;10:917–23. Lay MJ, Wittwer CT. Real-time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262–7. Crockett AO, Wittwer CT. Fluorescein-labeled oligonucleotides for real-time PCR: using the inherent quenching of deoxyguanosine nucleotides. Anal Biochem 2001;290:89 –97. Carbone D, Chiba I, Mitsudomi T. Polymorphism at codon 213 within the p53 gene. Oncogene 1991;6:1691–2. Lee LG, Livak KJ, Mullah B, Graham RJ, Vinayak RS, Woudenberg TM. Seven-color, homogeneous detection of six PCR products. Biotechniques 1999;27:342–9. Bernard PS, Pritham GH, Wittwer CT. Color multiplexing hybridization probes using the apolipoprotein E locus as a model system for genotyping. Anal Biochem 1999;273:221– 8. von Ahsen N, Oellerich M, Schutz E. Use of two reporter dyes
49.
50.
51.
52.
53.
54.
55. 56. 57.
58.
without interference in a single-tube rapid-cycle PCR: "1-antitrypsin genotyping by multiplex real-time fluorescence PCR with the LightCycler. Clin Chem 2000;46:156 – 61. Herrmann MG, Dobrowolski SF, Wittwer CT. Rapid #-globin genotyping by multiplexing probe melting temperature and color. Clin Chem 2000;46:425– 8. Bernard PS, Ajioka RS, Kushner JP, Wittwer CT. Homogeneous multiplex genotyping of hemochromatosis mutations with fluorescent hybridization probes. Am J Pathol 1998;153:1055– 61. Lyon E, Millson A, Phan T, Wittwer CT. Detection and identification of base alterations within the region of factor V Leiden by fluorescent melting curves. Mol Diagn 1998;3:203–10. von Ahsen N, Schutz E. Using the nearest neighbor model for the estimation of matched and mismatched hybridization probe melting points and selection of optimal probes on the LightCycler. In: Meuer S, Wittwer CT, Nakagawara K, eds. Rapid cycle real-time PCR: methods and applications. Heidelberg: Springer-Verlag, 2001:43–56. von Ahsen N, Oellerich M, Armstrong VW, Schutz E. Application of a thermodynamic nearest-neighbor model to estimate nucleic acid stability and optimize probe design: prediction of melting points of multiple mutations of apolipoprotein B-3500 and factor V with a hybridization probe genotyping assay on the LightCycler. Clin Chem 1999;45:2094 –101. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54:4855–78. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759 – 67. Allawi HT, SantaLucia J Jr. Thermodynamics and NMR of internal G.T mismatches in DNA. Biochemistry 1997;36:10581–94. Allawi HT, SantaLucia J Jr. Nearest-neighbor thermodynamics of internal A.C mismatches in DNA: sequence dependence and pH effects. Biochemistry 1998;37:9435– 44. Ke SH, Wartell RM. Influence of neighboring base pairs on the stability of single base bulges and base pairs in a DNA fragment. Biochemistry 1995;34:4593– 600.
