Research Article: A plasmid-based lacZα gene assay for DNA polymerase fidelity measurement

Date Published: February 15, 2013

Publisher: Academic Press

Author(s): Brian J. Keith, Stanislaw K. Jozwiakowski, Bernard A. Connolly.


A significantly improved DNA polymerase fidelity assay, based on a gapped plasmid containing the lacZα reporter gene in a single-stranded region, is described. Nicking at two sites flanking lacZα, and removing the excised strand by thermocycling in the presence of complementary competitor DNA, is used to generate the gap. Simple methods are presented for preparing the single-stranded competitor. The gapped plasmid can be purified, in high amounts and in a very pure state, using benzoylated–naphthoylated DEAE–cellulose, resulting in a low background mutation frequency (∼1 × 10−4). Two key parameters, the number of detectable sites and the expression frequency, necessary for measuring polymerase error rates have been determined. DNA polymerase fidelity is measured by gap filling in vitro, followed by transformation into Escherichia coli and scoring of blue/white colonies and converting the ratio to error rate. Several DNA polymerases have been used to fully validate this straightforward and highly sensitive system.

Partial Text

A considerably improved plasmid-based DNA polymerase fidelity assay has been described. The new plasmids, pSJ2 and pSJ3, are elaborations of an earlier version, pSJ1 [14], and contain a lacZα gene flanked by two nicking endonuclease sites. With pSJ2 and pSJ3, both the number of detectable sites and the expression frequency have been determined, factors necessary for determining polymerase error rates and not previously available for pSJ1. The expression frequency of 44.4% found for pSJ2 is near the theoretical maximum of 50% expected for random repair of mismatches, suggesting high plasmid integrity with minimal base damage. The critical step in applying pSJ2 and pSJ3 is gapping. As found with pSJ1, gapping cannot simply be carried out using heat to remove the strand excised in the nicking step. Rather, a complementary single-stranded competitor needs to be added to sequester the excised strand. The assay described in this article improves the preparation of relatively long competitors, using PCR followed by λ-exonuclease digestion, by exploiting the exonuclease resistance of phosphorothioate-protected primers. Although the phosphorothioate improvement consistently resulted in higher yields of competitor, PCR-based methods tend to give relatively small quantities of amplified material. Higher amounts of single-stranded DNA are available by direct chemical synthesis, but here the lengths that can be prepared become limiting. Thus, two competitors were required to successfully gap pSJ3; unfortunately, the use of four or five competitors to gap the longer pSJ2 failed. Gapping remains the most demanding part of the protocol, and research is still required to enable the production of long competitors in high quantity or to develop an alternative protocol. The use of BND–cellulose to isolate gapped plasmids represents a major improvement over gel-based purification techniques. Greater amounts can be prepared in a very straightforward manner and (importantly) the resulting gapped plasmids appear to be of much higher quality with little damage to the bases. This is evidenced by the very low background mutation rates seen for both pSJ2 and pSJ3 prepared using BND–cellulose, reduced approximately 5-fold as compared with gel-purified material. The low backgrounds seen with gapped pSJ2 (1 × 10−4) and pSJ3 (3 × 10−5) are extremely advantageous when studying high-fidelity polymerases. The suitability of the new plasmids was confirmed by determining the error rates of Pfu-Pol and Taq-Pol, which gave values similar to those obtained by a number of other methods. So far, pSJ2 and pSJ3 have been used only to test the fidelity of DNA polymerases in vitro. However, due to the compatibility of plasmids with many bacteria and eukaryotes, the in vivo study of DNA replication and repair should be possible, as has already been described for a plasmid-based mismatch repair activity assay [15].




Leave a Reply

Your email address will not be published.