Many DNA-carcinogen adducts not only compromise polymerase fidelity but also inhibit replication. This polymerase stalling or "idling" may then contribute to misincorporations if the polymerase is not completely blocked, such as the G:C to A:T mutations caused by O(6)-substituted guanines. Kinetic experiments were conducted to address the mechanism of polymerase stalling of T7 DNA polymerase exo(-) (T7(-)) and HIV-1 reverse transcriptase (RT) during replication of primer/template DNA containing guanine (G), O(6)-methylguanine (O(6)-MeG), or O(6)-benzylguanine (O(6)-BzG), thus, extending work presented in the preceding paper in this issue [Woodside, A. M., and Guengerich, F. P. (2002) Biochemistry 41, 1027-1038]. Substitution of a thio-substituted dNTP did not appear to strongly affect the chemistry of phosphodiester bond formation because the rate decreased <3-fold. Although the for "productive" binding increased for both T7(-) and RT as a function of the O6 substituent, fluorescence titrations indicate that the ground-state DNA binding was not affected for O(6)-alkylG substrates. DNA dissociation rates (k(off)) did not differ between unmodified and adduct-containing substrates. The presence of the correct nucleotide stabilized E*DNA interactions, resulting in a 10-fold slower k(off). Pre-steady-state experiments done in the presence of trap DNA revealed two rates of incorporation at the adduct, differing approximately 100-fold. Kinetic modeling fit the experimentally determined data (i.e., low burst amplitude at the adduct) only if the mechanism included an inactive E*DNA*dNTP complex. In summary, several lines of evidence indicate that the existence of a nonproductive polymerase complex best explains polymerase kinetics at DNA-carcinogen adducts, specifically O(6)-alkylguanine.