Human cellular DNA sustains about 30,000 damages per day due to endogenous metabolic processes. Of these, about one third are due to oxidative damage. Most endogenous damages are removed by a repair process called base excision repair, which is what we study in our laboratory. The damaged purines and pyrimidines are recognized and removed by a class of enzymes called DNA glycosylases that scan the DNA molecule searching for the damaged base in a sea of normal purines and pyrimidines. Once the damaged base is found, the glycosylase kinks the DNA molecule and flips the damaged base into the substrate binding pocket of the enzyme where the N-glycosyl bond is cleaved and the damaged base released from the DNA molecule. The oxidized purines and pyrimidines recognized by these DNA glycosylases are quite structurally diverse and it has been a long term interest in the laboratory to understand how the oxidative DNA glycosylases can have such a broad substrate specificity yet not remove the normal pyrimidine or purine base. When sequenced genomes became available, it was clear that many organisms contained multiple paralogs of these oxidative DNA glycosylases which presumably evolved different functions over evolutionary time. We exploited this to study the substrate specificities of a number of DNA glycosylases from organisms that ranged from Mycobacterium tuberculosis to the giant Mimivirus to humans and determined many of these to have novel substrate specificities. We were particularly interested in the Fpg/Nei family where the Fpg glycosylases are present in all bacteria and had been known for many years to recognize oxidized purines in DNA, particularly 8-oxoguanine and formamidopyrimidine. It turned out that closely related proteins are present in plants, fungi, the giant Mimivirus and in metazoans that have very different substrate preferences. Using bioinformatics analysis in collaboration with Dr. Jeffrey Bond (MMG) we showed Nei family members to be very sparsely distributed across phyla in the actinobacteria, g-proteobacteria and in metazoans and have a variety of substrate preferences. The Nei protein in E. coli was originally identified in our laboratory over 15 years ago and shown to be specific for oxidized pyrimidines. Not only did we identify novel substrate specificities for many of the Fpg/Nei proteins, but together with Dr. Sylvie Doublié (MMG), we determined crystal structures for a number of these and defined several novel motifs. In another collaboration with Dr. David Pederson (MMG), we are looking at base excision repair in nucleosomes, a substrate much closer to the ones that the glycosylases will encounter in human cells. We showed that when the oxidized base, thymine glycol, is facing away from the histone octamer, it is almost as readily recognized by the human pyrimidine specific DNA glycosylases as it is in free DNA. However, when thymine glycol is facing inward toward the histone octamer, it must be captured by the DNA glycosylases during transient unwinding of the DNA. More recently, with Dr. Joann Sweasy, an adjunct faculty member in the Department, we are asking about the potential consequences of germ line and tumor associated single nucleotide polymorphisms in the oxidative DNA glycosylases. We have already found that a germline SNP present in human NTH DNA glycosylase, that is found in about 4% of the population in the database, renders the glycosylase inactive since the SNP is in a catalytic residue. Yet the enzyme is still able to bind damaged DNA. In studies with tissue culture cells, we are asking whether the presence of this SNP in hNTH leads to cellular transformaion. With colleague Dr. David Warshaw (Molecular Physiology and Biophysics) we are using single molecule approaches to ask how the DNA glycosylases scan the DNA molecule looking for damage.
Our laboratory has also had a long standing interest in what happens when an unrepaired lesion is encountered by a DNA polymerase. In this case, if the damage blocks the DNA polymerase, it is potentially lethal or if the damage is bypassed by the DNA polymerase, and the incorrect base inserted, the damage is potentially mutagenic/carcinogenic. We have done a substantial amount of both biochemical and biological studies over the years to delineate the biological consequences of individual lesions as predicted by their interactions with DNA polymerases and more recently we have been examining these interactions using x-ray crystallography with colleague Dr. Doublié. When a replicative DNA polymerase is complexed with DNA containing an abasic site, an incoming dAMP is incorporated but no translocation takes place and significant distortion of the primer terminus results, thus explaining the strong replication blocking effect of an abasic site lesion. In a similar approach we were able to show that a thymine glycol blocks DNA elongation because the 5’ methyl group on the thymine glycol protrudes 5’ to the lesion, pushing the 5’ base into a configuration where it can no longer pair with an incoming nucleotide. This solved a puzzle that was posed by our laboratory over twenty years ago. A nonblocking but mutagenic oxidative lesion, 5-hydroxycytosine, is being studied using similar approaches. In current studies using kinetics analysis, site directed variants of DNA polymerases are being examined for their ability to misincorporate bases opposite a number of DNA lesions. These results are being corroborated using crystallographic analysis.