Point mutations in the ras proto-oncogenes are amongst the most frequent changes found in human malignancies (1) and may have prognostic importance. A variety of methods have been published for the detection of ras mutations (2–4). However, the most frequently used assays are limited either by an unsatisfactory sensitivity (2,3), the spectrum of detectable mutations (4) or require radioactive labeling (2,3). Here, we present a novel approach for the detection of ras point mutations based on the recently described method of peptide nucleic acid (PNA) mediated PCR clamping (5). PNAs are DNA-mimics where the phosphoribose backbone is replaced by a peptide-like repeat of (2-aminoethyl)-glycine units. Due to this chemical difference PNAs differ from DNA molecules in several aspects: (i) PNA/DNA-hybrids have a higher thermal stability compared with the corresponding DNA/DNA hybrids (∼1°C/base for mixed sequences); (ii) PNA/DNA hybrids are more destabilized by single base pair mismatches than the corresponding DNA/DNA hybrids (5); and (iii) PNAs could not serve as primer molecules in PCR.
The basic idea to use these features for the detection of ras gene mutations was to extend the original assay described for mutations at a single position (5) to several mutations clustered in a 4–5 bp span in one PCR. The principle was to hybridize chromosomal human DNA to a 15mer PNA complementary to the wild-type (wt) Ki-ras sequence surrounding codons 12 and 13 (schematically illustrated in Fig. 1). We reasoned that, in the case of wt Ki-ras, formation of PNA/DNA hybrids would be favoured. The bound PNA should sterically hinder annealing of a partially overlapping generic oligonucleotide, thus excluding the normal Ki-ras sequence from sufficient PCR amplification. In the case of mutant alleles, the melting temperature of the PNA/DNA hybrid was reduced, thereby allowing the 23mer oligonucleotide to outcompete PNA annealing and preferential amplification of mutant sequences.
This model was tested on six of 12 possible Ki-ras mutations in codons 12 and 13 derived from several tumor entities, which were available to us and had been characterized previously (6,7). A concentration of 2.84 µM PNA-1 was found sufficient to inhibit any detectable Ki-ras amplification starting from wt DNA (data not shown). In contrast, even at 14.2 µM PNA-1 no reduction in amplification was seen with primers for the human γ-interferon gene (8) or the human growth hormone gene, indicating the specificity of this inhibition. Several reaction parameters were evaluated for their influence on discrimination between mutant and wt Ki-ras alleles, e.g. the Tm of the competing generic primer, PCR temperature profiles, the total number of cycles, buffer composition and different polymerases. Among these, the total number of PCR cycles, the oligonucleotide annealing temperature and the amount of template DNA added were most important. A significant improvement was achieved by addition of glycerol (7.5% v/v) to the reaction.
After optimization the assay was able to detect all tested Ki-ras mutations irrespective of the site or type of the mutation in a single PCR (Fig. 2). A 429 bp fragment of the human growth hormone gene was coamplified in each reaction to exclude unspecific PCR inhibition in cases with no Ki-ras amplification. We also checked the sensitivity of our method by diluting a sample carrying a heterozygous GGT→GAT mutation in codon 12 in wt DNA. Under the conditions described, the mutation was detected down to one mutant allele in a background of 200 wt alleles. Direct sequencing confirmed that the mutant allele was amplified predominantly (Fig. 3D), whereas sequence analysis of the same sample reacted without PNA failed to detect the alteration (Fig. 3C).
In our eyes the major advantage of PNA mediated PCR clamping over published assays seems to be the higher flexibility which allows detection of mutations stretched over 4–6 bp in a single reaction. In addition, this method is not restricted to specific base exchanges, a major drawback of procedures using allele specific amplification (9). In conclusion, PNA-mediated PCR clamping is an attractive tool for the detection of ras gene point mutations. The simplicity and versatility make it especially helpful in large scale screening programs. Due to the special situation, ras gene mutations are ideally suited targets, but the assay could also be a rapid and sensitive prescreening method for common clustering mutations in other genes, such as hot-spot mutations in codons 175, 248 or 273 of the p53 tumor suppressor gene.
This study was supported in part by the Deutsche Forschungsgemeinschaft through grants to A.N. (Ne 310/6–2) and B.W. (SFB 366), and by the Wilhelm-Sander Stiftung (A.N.).