We are interested in elucidating the mechanisms used by DNA polymerases to increase the accuracy of DNA synthesis. In both prokaryotes and eukayotes, the fidelity of DNA synthesis results from at least two discrimination steps. Errors are prevented as the phosphodiester bond is formed due to the base selectivity of the polymerization reaction, and if errors are committed at the first step they can be corrected by 3' leads to 5' exonucleolytic proofreading. We are using the large (Klenow) fragment of E. coli DNA polymerase 1 as a model protein to dissect these two steps. The cloning and overproduction (by C. M. Joyce at Yale) of the enzymatically active Klenow polymerase and the recent determination of its structure by X-ray crystallography provided data which enabled engineering of the protein by site-directed mutagenesis. Two mutant derivatives of the Klenow polymerase, each lacking the proofreading exonuclease activity but containing normal protein structure, and a third protein containing only the polymerase domain, have been constructed. We have compared the fidelity of these proteins to that of wild-type Klenow polymerase to determine the contribution of base selectivity and proofreading to fidelity. These data demonstrate that the base selectivity of polymerization achieves an error rate of 1/10,000, and that proofreading improves fidelity about 5-fold overall. There are wide variations in discrimination at both steps, depending on the template site and error. Measurement of the steady state kinetic constants for correct and incorrect incorporation events suggest that base selectivity depends on both the ability of the polymerase to reject incorrect bases by changing active site conformation and on differences in the amount of time a nucleotide will occupy the active site based on hydrogen bonding potential and base stacking interactions. Preliminary results with the polymerase domain suggest that, independent of proofreading, the fidelity of polymerization is influenced by the small exonuclease domain, particularly for deletion errors. These studies are intended to describe the relationships between enzyme structure and the production of mutations.
