Mispair-, site-, and strand-specific error rates during simian virus 40 origin-dependent replication in vitro with excess deoxythymidine triphosphate.

We have measured the fidelity of leading and lagging strand DNA replication in HeLa cell extracts. Providing an excess of one dNTP in reactions induces replication errors consistent with misincorporation of that dNTP. With excess dTTP, both substitutions and single-nucleotide frameshifts are induced. Error distribution is nonrandom; reproducible hot spots for a substitution and a frameshift error are observed. Measurements with two vectors having the origin of replication on opposite sides of the mutational target demonstrate that error rates for G.dTTP and C.dTTP mispairs depend on whether the strand is replicated as the leading or lagging strand. Also, the two hot spots are only observed in one origin-target orientation. Replication reactions reconstituted from two fractions derived from extracts are 3-fold less accurate, but the error specificity with excess dTTP is similar to that with extracts. This suggests that the processes responsible for the nonrandom error rates are not lost as a result of fractionation. Furthermore, the reconstituted system is devoid of mismatch repair activity. Thus, mismatch repair is not responsible for the mispair-, site-, and strand-specific differences observed.