esults, residual LIG3 activity rather than residual LIG3 protein is the relevant parameter. Therefore, we used different, independent methods to evaluate LIG3 protein activity. Notably, results summarized 4 DNA Ligases in DCC-2036 price Alternative NHEJ 5 DNA Ligases in Alternative NHEJ are optimally repaired only by LIG3. However, this result may also reflect other consequences of LIG3 depletion. Thus, it is possible that the higher Deq values measured after LIG3 depletion actually reflect the generation of secondary, replication-dependent DSBs produced from unrepaired SSBs, rather than reduced repair of IR generated DSBs. This is because LIG3 is implicated in the repair of SSBs, and its depletion may increase the generation of replication-dependent DSBs through the associated inhibition of SSB-processing. Finally, it is possible that the onset of apoptosis observed in this mutant 4 d after treatment with 4HT causes DNA degradation that is erroneously detected as reduced repair. To address these points we generated and tested additional DT40 mutants and the results obtained are presented in the following sections. Deletion of Nuclear LIG3 Reveals Conclusively the Function of LIG1 in Alternative NHEJ The inception of apoptosis at low levels of LIG3 in the conditional DT40 LIG3 mutants hampers conclusive analysis of LIG1 function in DSB repair and the investigation of a possible specific requirement for LIG3 for a subset of DSBs. To overcome this limitation we rescued lethality of LIG32/2 cells by complementing the associated mitochondria defect. Indeed, lethality of LIG32/2 mutants is rescued by the expression of diverse DNA ligases endowed with a mitochondria targeting sequence. The DT40 mutant, LIG32/M2I, carries one inactivated LIG3 allele and one mutated allele that lacks the second translation initiation site normally utilized to generate the nuclear 12504917 version of the enzyme. As a result, LIG32/M2I cells express exclusively the mitochondrial form of LIG3. Despite lack of nuclear LIG3, LIG32/M2I cells are viable and grow with kinetics similar to the wt. As expected, in LIG32/M2I cells only traces of LIG3 are detectable by western blotting in the nuclear fraction of a cellular extract, whereas high amounts of the enzyme are detected in the mitochondria and the cytoplasmic fractions. Furthermore, analysis of intracellular mtLIG3 localization through expression of a GFP-tagged protein shows clear compartmentalization in the mitochondria. Following exposure to IR, LIG32/M2I cells show effective repair of DSBs that is indistinguishable from that of wt or LIG32/2loxP cells. In LIG32/M2I cells, nuclear DSB repair is mainly supported by LIG4, and LIG1. Since extensive depletion of nuclear LIG3 in this DSB repair proficient mutant, fails to detectably alter the repair kinetics, we conclude that secondary, replication-dependent DSBs possibly arising from unrepaired SSBs as a consequence of LIG3 depletion, only have a marginal effect on Deq. To further analyze the contribution of LIG1 to DSB repair and its interplay 10604535 with LIG3, we generated the double mutant LIG32/ M2I LIG42/2 and tested its viability and its DSB repair capacity. In this mutant, which is also viable, LIG1 is the only remaining nuclear DNA 
ligase. LIG32/M2ILIG42/2 cells process DSBs with efficiency slightly lower than LIG42/2 cells and similar to that seen with the corresponding conditional LIG3 mutant after treatment with 4HT for 3.5 d. This result confirms the function of LIG1 in alternative NH
