Choosing the right path: does DNA-PK help make the decision?

Abstract

DNA double-strand breaks are extremely harmful lesions that can lead to genomic instability and cell death if not properly repaired. There are at least three pathways that are responsible for repairing DNA double-strand breaks in mammalian cells: non-homologous end joining, homologous recombination and alternative non-homologous end joining. Here we review each of these three pathways with an emphasis on the role of the DNA-dependent protein kinase, a critical component of the non-homologous end joining pathway, in influencing which pathway is ultimately utilized for repair.

Fig. 1. Pathways for repairing DNA double-strand breaks

DSBs are repaired by one of three pathways, c-NHEJ (left panel) HR (middle panel) or a-NHEJ (right panel). c-NHEJ occurs with rapid kinetics throughout the cell cycle, and is initiated when DSBs are recognized and bound by the Ku 70/86 heterodimer. The MRN complex may co-localize to the same break (potentially through interaction with Ku70). Ku recruits DNA-PKcs forming the DNA-PK holoenzyme. Artemis, which is less abundant than DNA-PKcs, is complexed with DNA-PKcs. Upon recruitment to a DSB by Ku, the nuclease activity of Artemis is activated by trans autophosphorylation at the ABCDE cluster within DNA-PKcs. ABCDE phosphorylations promote end processing whereas PQR phosphorylations (also in trans) limit end processing. This provides tight regulation of end access to various processing factors, including Artemis, PNKP, and X family polymerases. If c-NHEJ fails, the HR and a-NHEJ pathways can gain access to the break; this would be facilitated by the presence of the MRN complex. In this case, DNA-PK might be inactivated by phosphorylation at the J, K or T sites as well as other, not yet identified, sites. J, K and T phosphorylations have been shown to inhibit c-NHEJ and promote HR. It is not known whether J, K or T phosphorylations occur in cis or trans, but clearly DNA-PKcs autophosphorylates some sites in cis (K.M. unpublished data). Ligation is carried out by the XLF/XRCC4/LigIV complex. Filament formation by XRCC4 and XLF may also contribute to synapsis. Further autophosphorylation of DNA-PKcs induces kinase inactivation (N, J, K, and T sites) and finally dissociation (sites responsible for dissociation are not known, although phosphorylation of ABCDE may contribute to complex dissociation). Although assembly and action or the c-NHEJ pathway is presented sequentially, it is also possible that c-NHEJ factors are simultaneously recruited to the site of a DSB through interaction with Ku and/or DNA-PKcs. Such a supercomplex would likely require the presence of Ku to be maintained, and as a result may cause Ku to be “stuck” on the DNA following repair. Proteolysis could provide a mechanism for subsequent removal of Ku. If c-NHEJ fails or if the DSB is first recognized by PARP, MRN is recruited to the DSB. During the S and G₂ phases of the cell cycle, MRN along with CtIP and BRCA1 perform initial, short range DNA end resection which promotes HR (adapted from [179]). HR proceeds with slower kinetics than c-NHEJ. More extensive end resection is carried out by EXO1 in collaboration with the BLM helicase. The exposed 3’ overhangs are then rapidly coated with RPA, which is ultimately replaced by Rad51 through the assistance of “mediator” proteins such as BRCA2 and/or the Rad51 paralogs. The Rad51 filament invades a homologous segment of DNA while simultaneously displacing the complementary strand forming a D-loop. Following D-loop formation several sub-pathways of HR can occur including formation of Holiday junctions or SDSA. Limited DNA end resection by MRN and CtIP can result in a-NHEJ (adapted from [179]), which can occur throughout the cell cycle and may involve ligation by the XRCC1/DNA ligase III complex; a-NHEJ also proceeds with slower kinetics than c-NHEJ.

Fig. 2. DNA-PK’s autophosphorylation is functionally complex

DNA-PKcs is phosphorylated on numerous sites (likely more than 40) in vitro and in vivo [22]. Our laboratory has studied (by mutagenesis) ~20. The cartoon depicts only sites shown to affect repair in living cells. Phosphorylation at T in the activation loop of the kinase domain inactivates the kinase; thus blocking NHEJ and promoting HR. JK phosphorylation also impedes NHEJ while promoting HR; however, JK phosphorylation does not affect enzymatic activity. Phosphorylation of N impedes (but does not block) kinase activation, thus inhibiting NHEJ. However, (unlike JK, and T phosphorylations) N phosphorylation does not promote HR. Phosphorylation of sites within either of the two major clusters (ABCDE and PQR) enhances NHEJ by reciprocal regulation of end processing. None of the ABCDE or PQR sites (alone or in combination) alter enzymatic activity or mediate autophosphorylation induced kinase dissociation (from Ku bound DNA), although phosphorylation of ABCDE sites may contribute to kinase dissociation. Studies are ongoing to define autophosphorylations responsible for complete complex dissociation.