Finding a needle in a haystack: detection and quantification of rare mutant alleles are coming of age.

Detecting rare mutant alleles in the background of wildtype DNA sequences and quantifying a specific DNA sequence in a clinical sample are two of the most challenging problems in DNA diagnostics. Successful approaches can expect important applications. For example, prenatal diagnosis of genetic disorders can be accomplished by invasive procedures such as amniocentesis, chorionic villus sampling, or fetal blood sampling, or by labor-intensive methods such as fetal cell enrichment from maternal circulation (1 ). If one can detect the mutant allele in the small number of fetal cells found in the maternal circulation, prenatal diagnosis of many devastating genetic disorders can be done with a simple blood test that poses minimal risk to the mother or the fetus. Other applications of the ability to find “a needle in the haystack” include cancer marker detection for early diagnosis or detection of relapse (2 ), and detection of specific alleles in pooled population samples (3 ). Similarly, the ability to quantify the amount of a specific DNA species present in clinical samples can help define the viral load in HIV-positive patients (4 ), determine minimal residual disease in patients with leukemia (5, 6), and monitor specific gene expression (7 ). In this issue of the journal, a group of researchers in Australia describe two promising methods that address these two problems. In the first report, Fuery et al. (8 ) describe the restriction endonuclease-mediated selectivePCR (REMS-PCR) by which a mutant allele can be detected in the presence of as much as 1000-fold excess of wild-type alleles. The idea is quite simple and is an extension of their previously published work (9, 10). If the wild-type allele is part of a restriction site for a thermostable restriction endonuclease (such as BstNI), conditions can be found under which the restriction endonuclease retains activity despite the high annealing temperatures during PCR. Under these conditions, any amplification product derived from the wild-type sequence will be cut by the restriction endonuclease and can no longer serve as template for the next round of amplification. This suppresses the amplification of the wild-type allele and allows only the mutant allele to continue in the PCR and yield a product detectable by gel electrophoresis. In the second report by the same group, Todd et al. (11 ) describe a novel general strategy for the detection and quantification of specific DNA sequences. The approach they have developed is based on the ability of a DNA enzyme to cleave an RNA-containing reporter probe (12 ). Specifically, the antisense sequence of a 10-23 DNAzyme is added to one of the PCR primers for the assay such that the active DNAzyme is formed only if PCR amplification occurs. A DNA/RNA chimeric reporter substrate containing fluorescent and quencher dye molecules on opposite sides of the cleavage site is added to the reaction mixture and is cleaved as the DNAzyme forms during PCR amplification. The accumulation of PCR products is monitored in real time by changes in the fluorescence released by the separation of the fluoro/quencher dye molecules as the reporter substrate is cleaved by the newly formed DNAzyme. In a serial dilution experiment, Todd et al. (11) showed that the DzyNA-PCR method was able to resolve 10-fold dilutions of K-ras from 10–10 copy number with high precision. The REMS-PCR method is simple to perform and appears to have great sensitivity and reliability. Because it is a “closed-tube” reaction with all reagents assembled together initially, the chance for contamination is minimized and automation is possible. The obvious drawback of the REMS-PCR method is the need for a thermostable restriction enzyme that cleaves the wild-type allele. This may not always be possible, and this requirement limits the general utility of the REMS-PCR approach. The authors describe techniques to identify suitable enzymes. Moreover, with protein engineering of restriction enzymes, one may foresee the production of thermostable variants. The DzyNA-PCR DNA detection and quantification method is novel and attractive because the only specialty reagent, the energy-transfer DNA/RNA hybrid reporter substrate of the DNAzyme, can be used in any assay. The only target-specific reagents are the two PCR primers, one of which is modified with the antisense sequence of the DNAzyme. This feature compares well with the other methods capable of DNA quantification in a closed-tube format. For example, the 59-nuclease reaction (TaqMan assay) (13 ) and Molecular Beacons (14 ) both detect specific DNA sequences and can quantify DNA in clinical samples, but they use target-specific TaqMan probes and molecular beacons that add to the expense of the reaction. However, there is no apparent reason that a universal probe or beacon cannot be constructed to act in a way similar to the DzyNA-PCR method such that the fluoro/ quencher dye pair are separated only when amplification occurs. In other words, if one adds a synthetic sequence to one of the PCR primers that is complementary to a universal TaqMan probe, cleavage of the TaqMan probe will occur only when amplification occurs. Similarly, one can add a synthetic sequence to one of the PCR primers that is the same as the recognition sequence of a universal molecular beacon; the complementary sequence will form for molecular beacon annealing only when amplification occurs. The DzyNA-PCR also compares well against the Invader assay (15 ), where flap endonucleases (FENs) isolated from archaea are used to recognize and cleave a branched structure formed when two overlapping oligonucleotides hybridize to a target DNA strand. The isothermal Invader assay can be used to detect specific DNA sequences directly from genomic DNA without amplifying the template. The linear amplification of the cleavage Editorial