Once a DSB is generated, the information about this damage needs to be propagated to downstream effector machineries such as repair proteins, cell cycle checkpoints and cell death pathways. Accordingly, recognition of the lesion is the first and initial step in the DNA damage response. In a second step, the information about the lesion can then be transported across the damaged cell. Each specific lesion of the DNA is recognized by one or a set of individual DNA damage recognition factors that are associated with certain signaling components. Double strand breaks are generally recognized by the ataxia-telangiectasia-mutated (ATM) and by the DNA-damage-dependent protein kinase (DNA-PK).
ATM is a serine/threonine protein kinase with a multitude of activities. It signals the presence of DSBs to the cell cycle machinery where it mediates proliferation arrest at G1/S, intra-S and G2/M checkpoints (Fig.
). The kinase is furthermore required for DNA repair and contributes to the initiation of apoptosis (Fig.
) [reviewed in: 43
]. Mutations in the ATM gene lead to the genetic disorder ataxia-telangiectasia (AT), which is characterized by cerebellar degeneration, immunodeficiency and an increased risk of cancer. Cells from AT patients are hypersensitive to ionizing radiation and display an increased frequency of chromosome breakage and defects in cell cycle control [reviewed in: 43
]. ATM is primarily a nuclear protein where it is present as an inactive dimer or multimer (Fig.
). The kinase is, furthermore, associated with protein phosphatase 2A (PP2A), a serine/threonine phosphatase that associates with ATM via
its scaffolding A-subunit [44
]. PP2A dephosphorylates ATM constitutively, which keeps the kinase in an inactive state [44
]. In addition to PP2A, ATM was found associated with protein phosphatase 5 (PP5). This interaction of ATM with PP5 increases after exposure of cells to DNA damaging agents and appears to be important for ATM activity in response to DSBs [45
]. How ATM becomes activated in the presence of DSBs is not entirely solved. A previous model suggested that upon sensing of a DSB, ATM becomes partially active, resulting in trans-autophosphorylation of ATM at serine 367, serine 1893 and serine 1981, and dissociation of ATM dimers/multimers into highly mobile and enzymatically active monomers that associate with damaged chromatin [46
]. More recently, this model of ATM activation has been questioned. Jean Gautier and co-workers showed that the MRN complex and DNA are sufficient to facilitate ATM monomerisation [48
]. In addition, phosphorylation of serine 1981 appears to be optional for the dissociation of ATM or its function in general [49
]. ATM even remained functional when additional autophosphorylation sites (serine 367 and/or serine 1893) were mutated [50
]. Besides phosphorylation, ATM also becomes acetylated in response to DSBs. Tip60 (HIV-1 Tat-interacting protein 60 kDa) acetylates ATM on lysine 3016 and this modification is essential for ATM activation and dissociation into monomers [51
]. It is, though, unclear by which mechanism Tip60 becomes activated in the presence of DSBs. Also Aven, an interaction partner of the anti-apoptotic protein Bcl-XL
, has been reported to contribute to ATM activation [52
]. Aven interacts with ATM in cellular extracts and overexpression of Aven promotes ATM phosphorylation on serine 1981 as well as phosphorylation of typical ATM substrates like p53 and Chk2 [52
]. However, since Aven is less efficient than DSB in inducing ATM autophosphorylation it appears to act synergistically with DNA damage rather than being a sole activator of the kinase.
Fig. (1) ATM and its target proteins. In the active state ATM phosphorylates target proteins that signal to cell cycle checkpoints, DNA repair proteins or to the cell death machinery. Phosphorylation of Brca1 or the activation of the MRN complex results in the (more ...)
Fig. (2) ATM activation and downstream signaling. (A) ATM is present in the nucleus as an inactive dimer/multimer that is associated with PP2A, a phosphatase that controls ATM phosphorylation. Upon ionizing radiation (IR), ATM is activated by trans-autophosphorylation (more ...)
Apart from ATM is DNA-PK (DNA-dependent protein kinase) able to recognize DSBs [53
]. Like ATM, DNA-PK has a strong affinity to DNA. Its recruitment to sites of DNA damage is conducted by a heterodimer of the Ku70 and Ku80 proteins [54
]. The Ku70/80 heterodimer forms a ring-shaped structure that can be threaded onto DNA ends [54
]. Due to its high affinity for loose DNA ends and its capacity to bind DNA in a sequence-independent fashion, the Ku70/80 heterodimer is believed to serve as a DSB sensor. Once the Ku70/80 heterodimer is bound to damaged DNA, it serves as a docking site for the catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs), which results in the formation of the DNA-PK holocomplex that displays serine/threonine kinase activity towards p53, H2AX, Artemis, XLF and others [55
]. Formation of the heterodimeric holocomplex is mediated by the amino-terminal and the central region of both, the Ku70 and Ku80 protein. The Ku80 carboxy-terminus is, furthermore, required for DNA-PKcs autophosphorylation at threonine 2609 [59
]. This phosphorylated form of DNA-PK co-localizes with γ-H2AX and 53PB1 at the DSB [59
After recognition of the lesion, an operation platform for repair factors needs to be generated. Since eukaryotic DNA is organized in tightly packed nucleosomes, these structures need to be widened to allow recruitment of repair proteins and to enable their interaction with damaged DNA. For this purpose, different residues of certain histone proteins become post-translationally modified. Especially acetylation, methylation, phosphorylation and ubiquitination of histone proteins have been observed during the DNA damage response [60
]. The main purpose of acetylation and phosphorylation seems to be to neutralize the positive charge of basic histone proteins. This would reduce the strength of the interaction of basic histone proteins with negatively charged DNA and may decondensate the chromatin. The earliest post-translational modification event at the chromatin as well as for the initiation of the DNA repair process is phosphorylation of the histone H2A variant H2AX at serine 139. The modified H2AX protein, which is called γ-H2AX, appears within a few minutes after exposure of cells to ionizing radiation at the break [64
]. Beside H2AX is histone H3 post-translationally modified after a DSB. Within a few minutes after introduction of a DSB, H3 becomes methylated at lysine 79 [66
]. Both histone modifications are required for recruitment of mediator proteins such as 53BP1 (p53-binding protein1), MDC1 (mediator of DNA damage checkpoint protein 1) or the MRN (Mre11, Rad50, Nbs1) complex to the DSB [66
Fig. (3) DNA lesion recognition and recruitment of repair factors. After introduction of a DNA double strand break (I), the DNA damage response is initiated by phosphorylation of H2AX (II), which is required for the recruitment of mediator proteins such as MDC1 (more ...)
The MRN complex, a highly conserved protein complex consisting of Mre11, Rad50 and Nbs1 (Nijmegen breakage syndrome 1), is implicated in DNA repair, cell cycle checkpoint control, DNA replication and telomere maintenance [70
]. The three proteins, Mre11, Rad50 and Nbs1, form a tight complex which is homogenously distributed throughout the nuclei of mammalian cells but migrates rapidly to DSBs after ionizing radiation, independent of the phase of the cell cycle [73
]. The Nbs1 protein appears to be responsible for nuclear localization and the proper assembly of the complex at DSB sites, probably by interacting with phosphorylated H2AX [68
]. Mre11 displays endo- and exonuclease activity and is important for the processing of the broken DNA ends [74
]. Rad50 also shows DNA binding capacity that possibly plays a role in tethering sister chromatids during HR [75
]. The MRN complex is composed of a single Nbs1 molecule that is associated with two dimers of Mre11 and Rad50. The Mre11 and Rad50 proteins themselves form a heterotetramer consisting of a dimer of Mre11 and Rad50. This heterotetramer contains two DNA-binding and processing domains that can bridge free DNA ends [76
]. The Mre11 dimer binds and holds the dsDNA ends in close proximity, near the Mre11 active site and it is thought that it aligns and bridges these loose DNA ends during DNA repair [78
]. Rad50 contains two long coiled-coil arms that promote intermolecular interactions through a terminal hook domain [76
]. Loss of Nbs1 or Mre11 is embryonic lethal [80
] and mutations in NBS1 and MRE11 lead to the chromosomal instability disorders Njimegen breakage syndrome (mutation in NBS) and ataxia telangiectasia-like disorder (ATLD; mutation in Mre11). Both of these syndromes are associated with enhanced sensitivity to ionizing radiation and chromosomal instability [reviewed in: 82, 83
]. MRN binds to DSBs and leads to further activation of ATM [48
]. Accordingly, cells derived from patients with Nijmegen breakage syndrome or ataxia telangiectasia-like disorders exhibit decreased ATM kinase activity despite the presence of wild type ATM [70
]. The nuclease activity of Mre11 appears to be particularly important for this process [70
]. Despite the similar timing of the appearance of Ku70/80 and the MRN complex at DSBs, recruitment of MRN and Ku seems to be independent of each other [88
]. Nbs1 and Mre11 are both targets for ATM and at least phosphorylation of Nbs1 is required for checkpoint signaling during S-phase [89
]. Therefore the MRN complex acts on the one hand as a downstream mediator of ATM but is also important for the activation of ATM and phosphorylation of downstream substrates.
MDC1, also called NFBD1 (nuclear factor with BRCT domains protein 1), appears to be a master regulator of the microenvironment at damaged chromatin. The docking of MDC1 to DSBs allows retention of multiple checkpoint and adaptor proteins including Nbs1, 53BP1 or Brca1 (breast cancer 1) at the site of the lesion where they provide a molecular platform for efficient amplification of the DNA damage signal [68
]. For this, MDC possesses two classes of phospho-binding motifs, a FHA (forkhead-associated) and a BRCT (Brca1 carboxy-terminal repeat) domain that serve as binding partners for phosphorylated proteins. BRCT and FHA modules are conserved throughout different species and are present in many proteins that are involved in the cellular response to DNA damage [reviewed in: 91, 92
]. Two repeats of the BRCT domain are located at the carboxy-terminus of MDC1. These BRCT domains associate directly with phosphorylated H2AX and this interaction seems to be essential for the recruitment and docking of MDC1 to the site of the DNA lesion (Fig.
). Accordingly, in H2AX-/-
MEFs, MDC1 fails to accumulate at DSBs [90
]. The FHA domain is positioned at the amino-terminus and binds to ATM which results in the accumulation of ATM at DSBs and enhanced phosphorylation of H2AX and of MDC1 itself [93
]. In MDC1-deficient cells, recruitment of ATM to DSBs is impaired [93
]. The central domain of MDC1 contains 14 repeats of a sequence that mediate its interaction with DNA-PKcs and the Ku heterodimer [94
]. Once recruited to DSBs, MDC1 stabilizes the MRN complex that is bound to damaged DNA and acts as a molecular scaffold for the recruitment of 53BP1, BRCA1 and additional MRN molecules to nuclear foci [68
]. Upon binding to MDC1, ATM phosphorylates the mediator protein and this phosphorylation creates a docking site for the recently identified E3-ligase RNF8 (ring finger 8) and for Ubc13 (ubiquitin-conjugating 13; Fig.
]. These proteins associate with γ-H2AX and phosphorylated MDC1 via
their FHA domains and decorate the histone protein H2A and its variant H2AX with lysine 63-linked polyubiquitin chains [22
]. Nevertheless, although RNF8 seems to be the first E3 ligase at the DSB, it appears to be insufficient for sustained ubiquitination of the chromatin. It probably rather primes the chromatin around the DSB, which then facilitates the recruitment of another E3 ligase, the RNF168 protein (Fig.
]. RNF168 recognizes the initial ubiquitin chains generated by RNF8 via
its ubiquitin-binding domain and, in concert with Ubc13, propagates and extends the formation of lysine 63-linked ubiquitin chains of histone proteins. Eventually, a threshold of lysine 63-polyubiquitin chains may be required to recruit and hold additional repair factors at the DSB. Particularly recruitment of 53BP1 and Brca1 are assumed to depend on such an “interaction trap” made of polyubiquitinated lysine 63 chains (Fig.
]. After polyubiquitination of the chromatin, Brca1, a tumor suppressor protein that also possesses ubiquitin ligase activity, especially when it is complexed with Bard1, is recruited to the lesion together with Rap80, Abraxas, Brca1 and Brcc36, which form the BRCA1-A complex [98
]. Brca1 can be found in three different complexes (Brca1 A, B and C), which depend on different adaptor proteins. The adaptor proteins for the complex A, B and C are Abraxas, Bach1/Brip1 (BRCA1-associated C-terminal helicase) and CtIP (CtBP-interacting protein), respectively. Each complex forms in a mutually exclusive manner [98
]. Whereas complex A plays an important role in DNA damage repair, the Brca1/Bach1 and the Brca1/CtIP complex form at different stages of the cell cycle and are required for a prolonged G2 phase and for G2/M transition checkpoint control, respectively [99
]. Ionizing radiation leads to ATM-dependent phosphorylation of Rap80 (receptor-associated protein 80) at serine 140, serine 402 and serine 419 [22
]. RAP80, furthermore, possesses two ubiquitin interaction motifs (UIM) which recognize lysine 63-linked ubiquitin chains. These UIMs facilitate the association of the complexes with polyubiquitinated histone proteins [22
]. Brcc36 of the complex has ubiquitin hydrolyzing activity and plays an important role in the regulation of the E3 ligase activity of Brca1 in response to ionizing radiation [101
]. The Rap80/Brcc36 complex furthermore removes lysine 63-linked polyubiquitin chains from the chromatin and is thus also involved in the termination of RNF8-Ubc13-mediated polyubiquitination once the lesion has been repaired [103
]. Although primarily implicated in HR, there is accumulating evidence that Brca1 may also be involved in NHEJ. The N-terminal region of Brca1 specifically associates with Ku80 and rapidly accumulates at DSBs in a Ku-dependent manner. Brca1 is furthermore assumed to control the fidelity of NHEJ [104
53BP1 functions, together with MDC1 and Brca1, as an adaptor/mediator protein of the DNA damage response. As such, it contributes to the activation of downstream effector molecules that function in DNA repair and DNA damage signaling, although it has no enzymatic activity. 53BP1 was originally identified due to its ability to bind to p53 [106
]. Upon ionizing radiation, it accumulates in foci at DSBs and becomes phosphorylated at several sites in an ATM-dependent manner. Among these is phosphorylation of serine 1219 particularly important as phosphorylation of this site assists in recruitment of MDC1 and γ-H2AX [107
]. 53BP1 is furthermore required for p53 accumulation, G2/M and intra-S-phase arrest, ATM autophosphorylation at serine 1981 and for the formation of Brca1-containing foci in response to ionizing radiation [108