Biochemical evolution of DNA polymerase eta: properties of plant, human, and yeast proteins

To assess how evolution might have biochemically shaped DNA polymerase eta (Poleta) in plants, we expressed in Escherichia coli proteins from Arabidopsis thaliana (At), humans (Hs), and the yeast Saccharomyces cerevisiae (Sc), purified them to near homogeneity, and compared their properties. Consistent with the multiple divergent amino acids within mostly conserved polymerase domains, the polymerases showed modest, appreciable, and marked differences, respectively, in salt and temperature optima for activity and thermostability. We compared abilities to extend synthetic primers past template cyclobutane thymine dimers (T[CPD]T) or undamaged T-T under physiological conditions (80-110 mM salt). Specific activities for "standing-start" extension of synthetic primers ending opposite the second template nucleotide 3' to T-T were roughly similar. During subsequent "running-start" insertions past T-T and the next 5' ( N + 1) nucleotide, AtPoleta and HsPoleta appeared more processive, but DNA sequence contexts strongly affected termination probabilities. Lesion-bypass studies employed four different templates containing T[CPD]Ts, and two containing pyrimidine (6-4')-pyrimidinone photoproducts ([6-4]s). AtPoleta made the three successive insertions [opposite the T[CPD]T and (N + 1) nucleotides] that define bypass nearly as well as HsPoleta and somewhat better than ScPoleta. Again, sequence context effects were profound. Interestingly, the level of insertion opposite the ( N - 1) nucleotide 3' to T[CPD]T by HsPoleta and especially AtPoleta, but not ScPoleta, was reduced (up to 4-fold) relative to the level of insertion opposite the ( N - 1) nucleotide 3' to T-T. Evolutionary conservation of efficient T[CPD]T bypass by HsPoleta and AtPoleta may reflect a high degree of exposure of human skin and plants to solar UV-B radiation. The depressed ( N - 1) insertion upstream of T[CPD]T (but not T-T) may reduce the extent of gratuitous error-prone insertion.