Detection of Cellular DNA Cleavage Using Non-Proofreading Thermostable DNA Polymerases

Few techniques are available to detect DNA lesions in cultured cells at the nucleotide level (8). One such method is primer extension of genomic DNA (5) that may be improved using linear amplification by repeated PCR cycles (1,11). This method has been used successfully to map the genomic sites of topoisomerase II (2,3,7) and topoisomerase I activity (10). DNA topoisomerases modulate DNA structure through a concerted DNA breakage-rejoining action (12). Antitumor drugs, such as doxorubicin and etoposide, poison the enzyme by stabilizing a reaction intermediate in which DNA strands are broken and covalently linked to enzyme tyrosine residues (4,9). Topoisomerase poisons may thus help in detecting genomic sites of enzyme activity. In the PCR-based primer extension method, a thermostable DNA polymerase is used to amplify a fragment corresponding to a topoisomerase IImediated strand cut. This is necessary because the cleaved strand is relatively rare and is present in a mixture of largely intact strands. It has been reported that the thermostable Taq DNA polymerase adds non-template nucleotides at the 3′ end of extended strands (3,6). Thus, the 5′ termini of the DNA template cannot be mapped exactly, and consequently, DNA breaks can be assigned to an accuracy of 1–2 bp in such experiments. To verify whether polymerases with proofreading activity that do not add extra nucleotides to the template may allow a better localization of strand breaks, we tested several thermostable polymerases for their ability to detect DNA breaks by topoisomerase II. Table 1 shows that DNA breaks can be detected only when DNA polymerase has no proofreading activity (see also Figure 1, left panel), in agreement with previous data indicating that Taq DNA polymerase (Promega, Milan, Italy) is more efficient than Ultma enzyme from PE Biosystems Italia (3). Indeed, some cleavage could be observed with Ultma only when the amount of cleaved template was high, such as for the highly repetitive satellite III DNA (3). This suggests that the Taq-added bases at the 3′ ends are needed to detect genomic DNA breaks with this method. Thus, we performed PCR-based primer extension studies to determine the effects of mixing Ultma and Taq DNA polymerases in the same reaction tube. Cellular DNA breaks were stimulated by dh-EPI, a potent anthracycline analog (3), at the histone H2A-H2B intergenic region of the Drosophila Kc cell genome (Figure 1), where cleavage sites were known to be present (3) (unpublished data). Using Taq DNA polymerase, topoisomerase II-mediated DNA breaks were readily detected after 20–30 thermal cycles, and no cleavage was observed using the Ultma enzyme (Figure 1). In a second set of experiments, after 20 cycles with Taq DNA polymerase, the reaction products were analyzed in sequencing gels or subjected to 15 additional cycles with both Taq and Ultma DNA polymerases (Figure 1, right panel). The results showed that the proofreading Ultma enzyme extended the previously present bands, corresponding to topoisomerase IIcleaved template strands (Figure 1, sites 246, 270 and 284), thus resulting in a loss of the signal in the gel. Taq-extended cleavage products bearing 3′ base overhangs could thus be Benchmarks