Randomization of the Damaged Base-Capping Loop Rewires Substrate Specificity in Formamidopyrimidine–DNA Glycosylase

Abstract

Abstract Formamidopyrimidine–DNA glycosylase (Fpg) is a bacterial base excision repair enzyme that removes 8-oxoguanine (8-oxoG) and some other oxidatively damaged nucleobases from DNA. Fpg hydrolyzes the N-glycosidic bond of the damaged nucleotide and nicks DNA at the resulting apurinic/apyrimidinic (AP) site by β-elimination. Together with bacterial endonuclease VIII and eukaryotic homologues, more specific for damaged pyrimidines, Fpg belongs to a helix–two-turn–helix (H2TH) superfamily of DNA glycosylases. Crystal structures of Fpg reveal that the enzyme recognizes 8-oxoG by flipping it out of the DNA duplex and capping it with a flexible tetrapeptide loop. However, this loop is poorly conserved and contacts 8-oxoG only through the main-chain amides, raising questions about the enzyme specificity determinants. To clarify the mechanism of 8-oxoG recognition, we have produced and characterized five variants of Fpg from Lactococcus lactis with a randomized capping loop. All of these mutants were deficient in 8-oxoG repair, both in vitro and in the E. coli complementation assay, yet retained the ability to bind AP site-containing DNA, form a Schiff base covalent reaction intermediate, and catalyze β-elimination. The mutants were still active in the removal of 5-hydroxyuracil (OHU), an alternative Fpg substrate. All loop mutants showed relaxed opposite-base specificity, processing AP sites and OHU opposite either C or A. Thus, whereas the general architecture of the active site in H2TH DNA glycosylases is permissive for the excision of a wide range of oxidized bases, the capping loop endows Fpg with the specificity for its biologically relevant substrate, 8-oxoguanine.