Mechanisms for DNA Damage Recognition by Alkyladenine DNA Glycosylase

Abstract

Alkyladenine DNA glycosylase (AAG) initiates the base excision repair of a diverse range of alkylated and deaminated DNA lesions in human cells. These sites of damage are removed from the genome through hydrolytic cleavage of the N-glycosidic bond to produce a shared abasic site repair intermediate. Lesions recognized by AAG include 1,N6-ethenoadenine, hypoxanthine, and the charged lesion 3-methyladenine. To recognize such a diverse set of lesions from among a vast excess of undamaged bases, AAG must be highly efficient in searching for targets while maintaining strong selectivity to prevent excision of undamaged bases. AAG binds nonspecifically to DNA and rapidly scans the genome in search of targets, but it remains unclear how AAG identifies compatible sites of damage for specific binding. Once a lesion is found, AAG utilizes a nucleotide flipping mechanism to rotate the lesion out of the duplex and into the active site for excision. This provides another selectivity filter for discrimination between damaged and undamaged nucleobases. To gain a better understanding of the substrates recognized by AAG, as well as the active site features important for specificity, we characterized the kinetics for the excision of the controversial target lesion 1,N2-ethenoguanine (ɛG). We found that ɛG is recognized and excised much less efficiently than the primary substrates of AAG. We also identified a key active site residue (N169) involved in discrimination against ɛG as a substrate. Collectively, these results suggest that poor recognition by AAG would prevent repair of this lesion in cells, which instead likely rely on the expanded repertoire of one of the AlkB homologs. Glycosylases such as AAG have long been speculated to use DNA bending to identify sites of damage during searching, as bending is observed in crystal structures of glycosylases bound to both lesions and undamaged DNA. We developed a stopped flow fluorescence resonance energy transfer (FRET) assay to measure the timing of DNA bending by AAG to place this step within the established AAG mechanism. We found that DNA bending and base flipping occur on the same timescale, after the lesion is already bound by AAG. This precludes DNA bending from contributing to searching by AAG. However, structural differences between glycosylases could lead to distinct strategies for DNA searching and lesion recognition. This and previous work point to the role of DNA intercalation by a conserved β-hairpin of AAG, which could serve to directly sense altered base pairing, a common feature of the best AAG substrates. Finally, we sought to explore the activity of AAG in the HAP1 cell line, an increasingly popular near-haploid cell line with great potential for use in the study of DNA repair. We determined that AAG does not contribute to the sensitivity of HAP1 cells to the alkylating agent methyl methanesulfonate. This lack of a phenotype contrasts with previously characterized cell lines, suggesting alkylation repair or tolerance is specifically perturbed in HAP1 cells. More work is needed to understand why HAP1 does not benefit from the repair activity of AAG, however, this work makes it clear that other cell models are better suited to the study of AAG function. Together, these results provide a deeper understanding of the mechanisms of searching and specificity used by AAG to recognize DNA damage, and the scope of tools available to study this process.