The role of replication blockage by the exocyclic DNA adducts propanodeoxyguanosine (PdG) and pyrimido[1,2-alpha]purin-10(3H)-one (M1G) was determined through the use of site-specifically adducted M13MB102 genomes containing a C:C-mismatch approximately 3000 base-pairs from the site of adduct incorporation. Genomes containing either dG, PdG, or M1G positioned at site 6256 of the (-)-strand were transformed into repair-proficient and repair-deficient Escherichia coli strains and the percent template utilization was determined by hybridization analysis. Unmodified genomes containing a C:C-mismatch resulted in a percent template utilization of approximately 60 and 40% for the (-)- and (+)-strands, respectively. Transformation of PdG- or M(1)G-adducted genomes resulted in approximately a 60-40% and 50-50% (-)-strand to (+)-strand ratio, respectively, indicating that PdG and M(1)G are negligible blocks to replication in repair-proficient E. coli. This is in contrast to previous studies using (PdG:T)- and (M1G:T)-mismatched M13MB102 genomes, which resulted in a majority of the replication events using the unadducted (+)-strand and suggested that both adducts were significant blocks to replication [J. Biol. Chem. 272 (1997) 11434; Proc. Natl. Acad. Sci. U.S.A. 94 (1997) 8652]. The C:C-mismatch results, though, indicate that the large strand bias detected in the earlier studies is due to repair of the adducts and resynthesis of the (-)-strand using the (+)-strand as a template for repair synthesis. Transformation of adducted C:C-mismatched genomes into E. coli strains deficient in nucleotide excision repair did result in an increased strand bias with only approximately 20 and 34% of the replication events using the (-)-strand for PdG- and M1G-adducted genomes, respectively. The increased strand bias indicates the importance of nucleotide excision repair in the removal of PdG and M1G.