E. coli Lhr is an uracil DNA glycosylase requiring an active site aspartate
We investigated whether E. coli Lhr is capable of DNA glycosylase activity, as suggested from structural similarities between glycosylases and the uncharacterized C-terminal region of E. coli Lhr (Lhr-CTD, LHR amino acids 876-1538) [11, 12] (Figure 1A ).E. coli Lhr-CTD protein fragment and full-length Lhr protein (1538 amino acid) were purified (Figure 1B )  when Lhr-CTD (50 – 800 nM) mixed with a Cy5-end labelled 37-nt ssDNA molecule modified to contain a single uracil nucleotide located 18 nucleotides from the 5’ ssDNA end (uracil-ssDNA) a single product was observed on alkaline treatment of reactions (Figure 1C , compare lanes 1-6 with 7-12), indicating DNA strand breakage at a DNA abasic site consistent with glycosylase activity. No product was observed from Lhr-CTD mixed with the same ssDNA lacking chemical modification (Figure 1C lanes 13-24).
To validate the uracil-DNA glycosylase activity of E. coliLhr-CTD we sought to identify single amino acid substitutions that would inactivate it. Sequence alignment of E. coli Lhr-CTD and AlkZ, with which it has structural similarity [11, 12], were unproductive at identifying highly conserved residues because Lhr-CTD lacks the ‘QxQ’ motif characteristic of AlkZ protein active sites [16], therefore suggesting an alternate catalytic mechanism in LHR. Alternatively, potential active site amino acids were identified through visual scrutiny of the Phyre2 [17] predicted model of the E. coliLhr-CTD, and in particular the positioning of side chains proximal to a proposed glycosylase active site (Figure 1D). Purified Lhr-CTDD1536A (50 – 800 nM), gave no product when titrated into the uracil modified ssDNA after alkaline treatment, compared with Lhr-CTD (Figure 1E ). We then tested whether substitution of the Lhr Asp-1536 residue inactivated uracil-DNA glycosylase active site chemistry or had some other effect on the protein that perturbed DNA binding. Unmutated Lhr-CTD was unable to form stable complex with DNA in EMSAs, when compared with full length Lhr (Figure 2A ), therefore we purified and tested full LhrD1536A. Lhr was also active as an uracil-DNA glycosylase compared with Lhr-CTD (Figure 2B ), but the LhrD1536A mutation inactivated glycosylase activity in agreement with inactive Lhr-CTDD1536A fragment (Figure 2C ). In EMSAs LhrD1536A formed stable complex with DNA similarly to Lhr (Figure 2D ), therefore we conclude that Lhr is a uracil DNA glycosylase that requires an active site aspartic acid residue.