Table 4.

Error rates of DNA polymerases determined using pSJ2 and pSJ3.

Polymerase	Gapped plasmida	Number of coloniesb	Number of mutant (white) colonies	Mutation frequencyc	Error rated
Pfu-Pol B	pSJ2	25,700	11	3.2 × 10−4	1.6 × 10−6
	pSJ3	20,116	11	5.2 × 10−4	3.5 × 10−6
Pfu-Pol B (D215A/E143A)e	pSJ2	14,601	20	1.3 × 10−3	6.3 × 10−6
	pSJ3	24,766	25	1.0 × 10−3	6.7 × 10−6
Pfu-Pol B (D215A/E143A/D473G) f	pSJ2	78,431	296	3.7 × 10−3	1.8 × 10−5
	pSJ3	38,836	141	3.6 × 10−3	2.4 × 10−5
Taq-Pol	pSJ2	46,239	98	2.0 × 10−3	1.0 × 10−5
	pSJ3	20,756	34	1.6 × 10−3	1.1 × 10−5

^aAll of the gapped plasmids used in these experiments had the coding strand (inner strand in Fig. 1B) removed by treatment with Nt.BbvC1 (pSJ2) or Nt.Bpu10I (pSJ3).

^bSum of three independent experiments, each consisting of five repeats.

^cThe mutation frequencies given here have had the background mutation frequencies found for gapped pSJ2 and pSJ3 subtracted.

^dError rates were calculated using the formula given in the text. An expression frequency (P) of 0.444 was used. In the absence of extensive DNA sequencing, an N_i/N value of 1 was used and the number of detectable sites (D) was the sum of the values for base substitutions plus insertions/deletions, that is, 448 for pSJ2 and 329 for pSJ3 (Table 1).

^eThe mutation D215A/E143A disables the 3′–5′ proofreading exonuclease activity [16].

^fThe triple mutation D215A/E143A/D473G has even lower fidelity than the exo^– double mutant D215A/E143A [17].