Nonhomologous end joining (NHEJ) is important for the repair of ionizing radiation and radiomimetic drug-generated DSBs, which are often associated with ligation-obstructing nucleotide damage. To facilitate ligation at such breaks, NHEJ employs a host of processing factors (i.e. nucleases, polymerases, etc.) that prepare DNA ends for joining. While this mechanism is efficient at joining broken chromosomes, it can frequently be inaccurate (i.e. loss of sequence at the DSB), because repair is mediated without the assistance of a template. My dissertation demonstrates how NHEJ-mediated repair of DSBs with associated abasic sites is an exception to this phenomenon. I show that abasic sites at DSB termini severely block NHEJ's ligation step and must be excised for joining to proceed. Despite the many processing enzymes associated with NHEJ, none are capable of excising this damage. Instead we found that the NHEJ core factor, Ku, has intrinsic lyase activity that removes these abasic sites. Analysis of Ku's substrate specificity reveals that lyase activity is restricted to abasic sites near a 5' terminus that directly block ligation. Furthermore, sequence 5' of abasic sites embedded in double stranded DNA (+4 bps) is mostly preserved due to Ku's limited activity in this context. By characterizing Ku's active site I identified eight lysine residues that contribute to lyase activity and determined that the primary nucleophile is within the N-terminus of Ku70 (K31). These amino acids reside on the outer surface of the Ku heterodimer nearest the DNA end - an optimal position for interacting with abasic sites closest to the break terminus. My results provide mechanistic insight into how NHEJ deals with one type of damage induced by ionizing radiation and may explain why loss of Ku leads to severe radiation sensitivity. Additionally, my results suggest NHEJ is more than a simple ligation machine but rather it is a sophisticated pathway suited to repair and join DSBs with associated nucleotide lesions.