DNA replication fidelity has been assayed by using a modified DNA sequencing reaction. In one experimental approach, dideoxycytidine 5'-triphosphate (ddCTP) was used as a chain terminator during replication of M13 phage DNA by the large fragment of DNA polymerase I. The deoxyribonucleotide analogue BrdUTP was used to compete against ddCTP-induced chain terminations as an assay for B X G base mispairing (B represents bromodeoxyuridine when the analogue is present as a base pair or base mispair). By comparing BrdUTP to dCTP for competition against ddCTP, an average misincorporation frequency for BrdUMP of 0.2% was found. A similar average misincorporation frequency has been measured previously for the incorporation of radioactively labeled BrdUMP and dCMP into the synthetic template-primer poly-[d(G,T)] X oligo(dA). The advantage of the sequencing method is that an error frequency is determined for each template guanine in a defined DNA sequence, thus providing information on the effect of neighboring base sequences on fidelity. Misincorporation frequencies varied no more than 5-fold among 50 template guanines tested. The approach used here is not limited for use with nucleotide analogues but is generally applicable in determining misincorporation frequencies and sequence specificities for any deoxynucleoside triphosphate substrate. In a second experimental approach, base mispairing between bromouracil and guanine was demonstrated directly by using 5-bromodideoxyuridine 5'-triphosphate (BrddUTP). A comparison of chain terminations attributable to BrddUTP and to dideoxythymidine 5'-triphosphate (ddTTP) revealed that B X A and T X A base pairs formed at about the same rate, whereas B X G mispairs occurred 4-10 times more frequently than T X G. The elevation in the frequency of B X G over T X G mispairs is consistent with the mutagenic behavior of the base analogue.