Implications of alternative substrate binding modes for catalysis by uracil-DNA glycosylase: an apparent discrepancy...

A theoretical study showed that the base excision repair enzyme uracil-DNA glycosylase (UDG) exploits electrostatic interactions with backbone phosphate groups in the substrate for catalysis. Although experiments performed to test the calculated results confirmed the predicted importance of the -2, -1, and +1 phosphate groups, there was an apparent disagreement with regard to the +2 phosphate group. The calculations indicated that it made an important contribution, while experimentally, the effect of its deletion or neutralization was small. The +2 phosphate group interacts directly with an active site histidine (H148 in humans) in the crystal structure of UDG in complex with double-stranded (ds) DNA. We previously calculated that H148 has a strong anticatalytic effect due to its protonation, and here we use alchemical free energy simulations to estimate its site-specific pKa. We find that it is positively charged over the entire experimental pH range (4-10), so its deprotonation cannot compensate for deletion or neutralization of the +2 phosphate group. The free energy simulations are facilitated by an efficient charge-scaling procedure that allows quantitative correction for the implicit treatment of solvent far from the active site; improvements are made to that method to account carefully for differences in the truncation of electrostatic interactions in the contributing molecular-mechanical and continuum-electrostatic approaches. Additional simulations are used to demonstrate that the +2 phosphate group is fully solvent exposed in complexes with single-stranded DNA substrates like those used in the experiments. In contrast, it is well-structured and buried in the dsDNA complex used in the original simulations. Differences in solvent shielding thus account for the apparent lack of an effect observed experimentally upon neutralization or deletion of this group.