DNA-dependent protein kinase (DNA-PK) was discovered in early 1990s, when the catalytic activity of a protein kinase triggered by double-stranded DNA was reported for the first time. Interestingly, it was noted that the kinase functions efficiently only in the presence of linear but not supercoiled DNA (Carter et al. 1990). Independently, DNA-PK was reported to phosphorylate transcription factor Sp1 and a number of other DNA-binding proteins. Furthermore, protein Ku was identified as an essential component of DNA-PK and confirmed as a binding subunit that served to recruit DNA-PK catalytic subunit (DNA-PKcs) (Gottlieb and Jackson 1993). Another important breakthrough came from Stamato and his colleagues in 1994, who reported lack of double-strand break (DSB) DNA binding activity in radiosensitive rodent cell line and attributed it to the Ku heterodimer absence (Getts and Stamato 1994). Since then, DNA-PK has been studied extensively as a major player in DNA DSB repair.
Phosphatidyinositol 3-Kinases Superfamily
DNA-PK is a member of large superfamily of phosphatidyinositol 3-kinases (PI3K), which is generally divided into three classes based on primary structure with an additional group of distantly related and structurally heterogeneous enzymes involved in monitoring of genomic integrity and cell signaling in order to regulate cell growth. These kinases are referred as PI3K functional subgroup – PI3K-related kinases (PIKKs). This class comprises of ataxia telangiectasia mutated kinase (ATM), ataxia telangiectasia and Rad3-related kinase (ATR), mammalian target-of-rapamycin (mTOR), suppressor with a morphological effect on genitalia family member (SMG-1), transactivation/transformation-domain-associated protein (TRRAP), and DNA-PK.
The crystal structure of DNA-PKcs revealed that many alpha-helical HEAT repeats facilitate bending of the polypeptide chain and allow it to fold into a hollow circular structure. The FAT-C is located on top of this structure; inside is a small HEAT domain, which is likely to bind DNA. The structure was described as a flexible open-ring cradle promoting DSB repair (Sibanda et al. 2010).
C-terminal region of Ku80 contains globular helical domain, which is involved in protein-protein interactions and is required for the interaction of Ku heterodimer with DNA-PKcs. On the other hand, C-terminal region of Ku70 contains a scaffold-associated protein domain, which is required for interaction with chromatin. The shape of Ku70 allows binding in the major groove of DNA (proximal to DSB), while Ku80 binds in the minor groove (distal to DSB site) and thus both subunits form a circular structure mounting the molecule of DNA. This setting allows the C-terminal region of Ku80 to recruit and establish DNA-PKcs at DSB and facilitates stabilization of another molecule of DNA-PKcs at the opposite site of the break. In addition, this particular interaction stimulates DNA-PK complex activity (Novotna et al. 2013).
Heterodimer of Ku70/80 ensures initiation of nonhomologous end joining (NHEJ) and it binds to DSB within seconds of its creation and does so in all cell cycle phases before DNA-PKcs to allow its access to the DNA end. This results in activation of the catalytic activity of the enzyme. Once localized in the DNA damage site, it performs its primary function, i.e. being a scaffold protein to recruit the NHEJ machinery to the DNA lesion. The secondary function is maintenance and stabilization of DSB ends, which abrogates their nonspecific processing. In this respect, Ku has been shown to block DNA end-processing enzymes such as exonuclease-1 in vitro. This function is crucial for diminishing chromosomal aberrations and keeping genomic integrity (Davis et al. 2014).
DNA-PKcs has a limited activity in the absence of Ku70/80 and DNA, which makes it truly DNA-dependent. Although the mechanisms regulating DNA-PKcs activity are quite well-described, the exact role of DNA-PKcs enzymatic activity in NHEJ is still unclear. Activated DNA-PKcs phosphorylates a number of proteins in vitro, including p53, transcription factors, RNA polymerase, Ku70/Ku80, X-ray cross complementing protein 4 (XRCC4)-like factor (XLF), Artemis, DNA ligase IV, and others; however, the in vivo significance of these phosphorylations is unclear too. The implicated NHEJ factor Werner (WRN) has been recently reported as a DNA-PK substrate required for efficient NHEJ. Additionally, some new proteins that are phosphorylated by DNA-PKcs after DNA damage have been identified, but likewise, their importance in NHEJ or DNA damage response (DDR) has not been established yet. Those substrates are for instance protein kinase B (PKB/Akt), heat shock protein HSP90α, nuclear receptor 4A, and scaffold attachment factor SAF-A.
One of the best-characterized DNA-PKcs substrates is actually DNA-PKcs itself. It is phosphorylated on more than 40 sites after DSB induction, thus the impact of each phosphorylation on DNA-PKcs function seems to be very complex. Many sites are localized in two well-described clusters: ABCDE (Thr2609, Ser2612, Thr2620, Ser2624, Thr2638, and Thr2647) and PQR (Ser2023, Ser2029, Ser2041, Ser2051, Ser2053, and Ser2056). Thr2609 was originally thought to be an autophosphorylation site only, but late studies revealed that it is phosphorylated also by ATM or ATR, depending on the stimulus. Anyway, it is required for rejoining of DSB and studies on mice demonstrated its critical function in vivo. The ABCDE cluster phosphorylation promotes end-processing via DNA-PKcs dissociation from Ku and DSB, making the damaged site accessible to downstream repair factors. On the other hand, PQR cluster autophosphorylation retains DNA-PKcs at damaged DNA ends (Summers et al. 2011). JK cluster contains phosphorylations at Thr943 and Ser1003, which were shown to specifically promote homologous recombination (HR) while inhibiting NHEJ, suggesting a possible mechanism of protection of certain DNA ends from NHEJ repair (Neal et al. 2011).
One particular DNA-PKcs substrate deserves special attention in respect to DDR. In mammalian cells, DSB in chromatin immediately triggers phosphorylation of histone H2AX at Ser139 (γH2AX), which creates foci in the megabase regions localized around the damage site. H2AX is a key component of repair machinery and ensures recognition of DSB, recruitment of other repair molecules, and coordination of signaling cascades required for efficient repair.
DNA-PKcs Role in DNA Repair
NHEJ plays the largest role in DSB repair in humans. It does not require a homologous template and thus is not restricted to a certain phase of the cell cycle, unlike the other well-characterized DSB repair mechanism, homologous recombination (HR), which is believed to be active only during S and G2 phases when the sister chromatid is available.
The process of NHEJ starts with DNA end recognition and assembly and stabilization of the NHEJ complex. Ku heterodimer is bound to the DSB ends; as a scaffold protein, it recruits directly or indirectly the main NHEJ factors, including DNA-PKcs, XRCC4, DNA Ligase IV, XLF, and Aprataxin-and-PNK-like factor (APLF) (Goodwin and Knudsen 2014). One of the helping proteins, XRCC4, was discovered in highly radiosensitive cell lines with defective DSB reparation and it was reported as a specifically phosphorylation target of DNA-PK in gamma-irradiated cells. XRCC4 has been shown to bind to an important part of the system – DNA ligase IV. In mice lacking XRCC4 or LIG4 gene, massive apoptosis occurred in embryonic neural cells and mutations in human fibroblast cell line 180BR (derived from patient with lymphatic leukaemia) leading to higher radiosensitivity linked to DNA ligase IV and the inability to repair the radiation damage (Riballo et al. 1999). XRCC4-DNA Ligase IV complex then ligates the compatible ends. This reaction is stimulated by XLF, which interacts with XRCC4. Frequently, DNA ends are not compatible. They contain damaged bases and/or DNA backbone sugars and preprocessing or excision before ligation is required. DNA polymerases μ and λ, polynucleotide kinase, and the Artemis nuclease are responsible for this end processing and occasionally required DNA polymerization (Iliakis et al. 2015).
Although Ku is the first NHEJ factor to bind to the DSB end, DNA-PKcs is not necessarily the next one. The order of the factors recruitment is believed to be flexible and may depend on the actual complexity of the DNA damage. Simple DSB may be repaired as described rapidly – involving only the Ku heterodimer and XRCC4-Ligase IV complex together with XLF – whereas more complicated DSB require DNA-PKcs and possibly contribution of ATM. Recently identified paralogue of XRCC4 and XLF DNA repair factors (PAXX) was identified and some functional redundancy with XLF was revealed.
In addition to its role in NHEJ, DNA-PK is also involved in HR, the second major DSB repair mechanism. During cell cycle, HR pathway is predominantly performed within S and G2 phases, and it is particularly crucial for genomic stability during DNA replication. In contrast to NHEJ, which is error-prone, HR is characterized by high fidelity. Generally, it can be divided into four phases: (i) resection of DNA ends; (ii) formation of RAD51 filament; (iii) strand invasion and Holliday junction (HJ) formation; and (iv) HJ dissolution.
DNA-PKcs blocks exonuclease 1 (Exo1) binding required for resection. Autophosphorylation of DNA-PKcs promotes DNA-PKcs dissociation and consequently Exo1 binding. ATM can compensate for DNA-PKcs autophosphorylation and promote resection under conditions where DNA-PKcs catalytic activity is inhibited. The Mre11-Rad50-Nbs1 (MRN) complex, which is crucial for initial detection of a lesion, further stimulates resection in the presence of Ku and DNA-PKcs by recruiting Exo1 and enhancing DNA-PKcs autophosphorylation, and it also inhibits DNA ligase IV/XRCC4-mediated end rejoining. Hence, in addition to its pivotal role in NHEJ end joining, DNA-PKcs also cooperates with ATM and MRN to regulate resection and thus influences the choice of DNA repair pathway (Fig. 2).
DNA-PK Functions in Cell
In addition to its well-established role in DSB repair, DNA-PKcs was found to be activated by hypoxia, independent of DNA damage. Consistently, DNA-PKcs positively regulates hypoxia-inducible factor α-1 and its activation by histone acetylation during hypoxia was reported to affect cellular survival. Furthermore, new roles are still emerging such as roles in basal transcription machinery, viral infection and immune response networks, maintenance of telomeres, and other processes involved in hormone-dependent cancers and tumor suppression.
Nonetheless, a key cellular function of DNA-PK in relation to consequence of DDR is genomic stability. Interaction of DNA-PKcs and Snail1 (an E-box binding transcription factor known to promote epithelial-mesenchymal transition during development and progression of cancer) has been attributed to promotion of genomic instability and aggressive cancer characteristics. Interestingly, DNA-PKcs-dependent phosphorylation of Snail1 was shown to stabilize Snail1 and to inhibit DNA-PKcs activity (Pyun et al. 2013). DNA-PKcs also maintains chromosomal integrity with important role in mitosis. siRNA-mediated depletion of DNA-PKcs or chemical inhibition led to various mitotic defects including increased number of misaligned mitotic chromosomes, abnormal nuclear morphologies, and lagging chromosomes (Jette and Lees-Miller 2015).
DNA-PK has been also implicated in crucial mechanism affecting cell fate, transcription. It interacts with transcriptional factors and critically impacts cell behavior and disease progression, especially in context of cancer. The best-known example of direct cancer-associated transcriptional regulation is interaction (phosphorylation) with p53 tumor suppressor. Moreover, DNA-PKcs overexpression or increased activity in numerous tumor tissues is closely associated with metastases, poor prognosis, and radioresistance. However, the expression of DNA-PK shows large intra-tumor heterogeneity. In regard to the high-energy needs of tumor environment, DNA-PKcs inhibition or depletion resulted in reduced AMP-activated protein kinase phosphorylation and activation in response to glucose deprivation, confirming a role for DNA-PKcs in metabolic regulation (Amatya et al. 1823). Interestingly, DNA-PK negatively regulates phosphorylation of cryptochrome 1 at Ser 588 and thus contributes to the determination of the circadian period (Gao et al. 2013).
Given the critical roles of DNA-PK, the mechanisms that control its activity are of a crucial importance. In addition to autoregulation, several prosurvival signaling pathways have been shown to regulate DNA-PKcs kinase activity and may serve as proximal effectors of DNA damage and impact cell behavior. Serine/threonine kinase Akt was reported to form a complex with DNA-PKcs, promoting autophosphorylation in the initial phase of radiation-induced DNA damage repair. Moreover, a cell-type dependent positive feedback loop exists as DNA-PKcs phosphorylates Akt at Ser 473 (Toulany et al. 2011). Additionally, epidermal growth factor receptor can bind to DNA-PKcs and enhance its activity in response to DNA damage. Casein kinase II has been shown to colocalize with γH2AX at DNA damage sites and presumably facilitates stabilization of DNA-PKcs and Ku80 interaction at broken DNA ends (Olsen et al. 2010). Similarly, RNA-binding protein nuclear factor 90 that was also shown to regulate DNA-PKcs activity as well as protein phosphatase 6, which contributes to DNA-PKcs activation via direct interaction (Reichman et al 2002, Hosing et al. 2012). A functional interaction between DNA-PK and a tyrosine kinase c-Abl in response to DNA damage has been also hypothesized. Clearly, all the mentioned studies indicate that DNA-PKcs activation is a complex process and additional studies are required for a complete understanding of its regulation.
DNA-PK is a dynamic serine/threonine protein kinase that requires binding to DNA in order to express its catalytic properties. It serves as a versatile molecular sensor of DNA damage and is required for DSB repair and V(D)J recombination. Recently, a plethora of studies described multiple crucial functions besides that in repair machinery. These include role in genomic instability, hypoxia, metabolism, inflammatory response, and other. Alterations in DNA-PK expression or activity suggest that the consequences of its inhibition should be useful in antitumor therapy. Therefore, DNA-PK became an attractive target in oncology and a variety of different strategies has been applied in cancer treatment (Durisova et al. 2016).
In spite of the fact that DNA-PK has been studied extensively over three decades, some key questions remain opened for future research. For instance, DNA-PK has been defined as a transcriptional modulator, but what factors control this specific function in cancer cells? And how are the other functions altered during cancer initiation and progression? Although autophosphorylations of various sites have been established as activation, processes that precede these events are not fully characterized as well as factors and specific mechanisms that are required for the dissolution of NHEJ complex once the repair procedure is completed.
DNA-PK inhibitors entering clinical trials are well defined in regard to inhibition of kinase activity but little is known about general impact of kinase-independent functions that may influence therapeutic efficacy of malignancies. Hence, additional studies are needed because DNA-PK is indisputably a critical component of cellular response to DNA damage and a major player in human cancer.
- Iliakis G, Murmann T, Soni A. Alternative end-joining repair pathways are the ultimate backup for abrogated classical non-homologous end-joining and homologous recombination repair: Implications for the formation of chromosome translocations. Mutat Res Genet Toxicol Environ Mutagen. 2015;793:166–75.PubMedCrossRefGoogle Scholar