**2. DNA-damaging treatments induce multiple and dynamic responses**

DNA is a stable material, as required for the storage of genetic information, but it is also susceptible to spontaneous changes under normal cellular conditions. It has been estimated that each cell spontaneously loses about 5000 purine bases (depurination) every day (Friedberg, 1995). The deamination of cytosine to uracil also occurs spontaneously. In addition to this inherent instability, our genomes are exposed to numerous endogenous or environmental agents, including reactive metabolites, environmental chemicals and ultraviolet radiation, capable of inducing a wide diversity of DNA lesions (Figure 1). The large number of different lesions possible – more than 100 different oxidative modifications

Fig. 1. DNA-damaging treatments induce multiple and dynamic responses mediated by DNA damage sensors and transducers. Common DNA-damaging agents (A) induce several types of DNA damage (B) directly (solid line) or indirectly (dotted line). Single- and double-strand breaks (highlighted in gray) are the most frequent end products of unrepaired damage. DNA damage is recognized by sensor proteins (C) that recruit and/or activate transducers (D), initiating a signal transduction cascade (not shown). Abbreviations: ATM, ataxia telangiectasia mutated; ATR, ataxia and rad3-related; ATRIP, ATR-interacting protein; BER, base excision repair; CSA or CSB, Cockayne Syndrome A or B; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; hOGG1, human 8-hydroxyguanine DNA-glycosylase; HR, homologous recombination repair; hHR23B, human Rad23 homolog B; MMR, mismatch repair; MRN, Mre11-RAD50-Nbs1; MYH, MutY glycosylase homologue; NEIL1, nei endonuclease VIII-like 1; NER, nucleotide excision repair; NHEJ, non-homologous end joining; PARP, poly(ADP-ribose) polymerase; RPA, replication protein A; SSBR, single-strand break repair; UV, ultraviolet; XPC, xeroderma pigmentosum group C.

2 Will-be-set-by-IN-TECH

DNA is a stable material, as required for the storage of genetic information, but it is also susceptible to spontaneous changes under normal cellular conditions. It has been estimated that each cell spontaneously loses about 5000 purine bases (depurination) every day (Friedberg, 1995). The deamination of cytosine to uracil also occurs spontaneously. In addition to this inherent instability, our genomes are exposed to numerous endogenous or environmental agents, including reactive metabolites, environmental chemicals and ultraviolet radiation, capable of inducing a wide diversity of DNA lesions (Figure 1). The large number of different lesions possible – more than 100 different oxidative modifications

Fig. 1. DNA-damaging treatments induce multiple and dynamic responses mediated by DNA damage sensors and transducers. Common DNA-damaging agents (A) induce several

double-strand breaks (highlighted in gray) are the most frequent end products of unrepaired damage. DNA damage is recognized by sensor proteins (C) that recruit and/or activate transducers (D), initiating a signal transduction cascade (not shown). Abbreviations: ATM, ataxia telangiectasia mutated; ATR, ataxia and rad3-related; ATRIP, ATR-interacting protein;

non-homologous end joining; PARP, poly(ADP-ribose) polymerase; RPA, replication protein A; SSBR, single-strand break repair; UV, ultraviolet; XPC, xeroderma pigmentosum group C.

types of DNA damage (B) directly (solid line) or indirectly (dotted line). Single- and

BER, base excision repair; CSA or CSB, Cockayne Syndrome A or B; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; hOGG1, human 8-hydroxyguanine DNA-glycosylase; HR, homologous recombination repair; hHR23B, human Rad23 homolog B; MMR, mismatch repair; MRN, Mre11-RAD50-Nbs1; MYH, MutY glycosylase homologue;

NEIL1, nei endonuclease VIII-like 1; NER, nucleotide excision repair; NHEJ,

**2. DNA-damaging treatments induce multiple and dynamic responses**

alone have been identified in DNA (Cadet et al., 1997) – has led to the evolution of many different repair pathways for sensing and repairing the various types of damage.

The complete signaling network for each damage type and its individual contribution to the cellular damage response are not fully understood, but the essential repair mechanisms have been elucidated (reviewed for example by Fortini & Dogliotti (2007); Friedberg (1995; 2001); Helleday et al. (2008); Li (2008); Wyman & Kanaar (2006)). Figure 1 summarizes the main pathways and highlights the sensors (DNA binding proteins that recognize specific DNA lesions) and transducers (enzymes that amplify the damage signal by posttranslational modification of downstream targets) involved in repair and signaling for particular types of damage. The main DNA damage transducers are the phosphoinositide 3-kinase-like kinase (PIKK) family members ATM (ataxia telangiectasia mutated), ATR (ataxia telangiectasia and Rad3-related) and DNA-PK (DNA-dependent protein kinase). A DNA break signal can also be transduced by poly(ADP-ribose) polymerases 1 or 2 (here both designated PARP), which use NAD<sup>+</sup> to catalyze the modification of their targets with negatively charged, long and branched ADP-ribose polymers. We provide below a brief description of the DNA repair pathways, the subsets of damage they repair and the transducers that are activated.

#### **2.1 Repair processes that do not directly activate transducers**

The direct repair of certain alkylation adducts and other uncomplicated base modifications by specialized single enzymes is probably the simplest repair mechanism. O6-alkylguanine DNA alkyltransferase (AGT) is a major enzyme involved in direct repair. It is encoded by the O6-methylguanine-DNA-methyltransferase (MGMT) gene and transfers the alkyl adducts produced by alkylating agents, such as temozolomide, dacarbazine or nitrosourea compounds, from O6-methylguanine, O4-methylthymine, O6-ethylguanine or O6-chloroethylguanine to a cysteine residue within the active site of the enzyme, thereby inactivating the enzyme (Gerson, 2004). Other direct repair enzymes include the DNA dioxygenases ABH2 and ABH3, which can convert 1-methyladenine and 3-methylcytosine back into adenine and cytosine, respectively (Duncan et al., 2002). The repair of alkylated lesions is a rapid process, with most alkylated sites successfully repaired within an hour (Zhu et al., 2009). The types of damage targeted by direct repair processes do not seem to be associated with the activation of damage signaling kinases, probably due to the rapid repair kinetics and the absence of intermediate strand break generation during the repair process.

#### **2.2 Repair mechanisms that activate mainly PARP as a transducer**

The base excision repair (BER) pathway recognizes and removes bases carrying non-bulky modifications that have been damaged by nonenzymatic alkylation, oxidation, ring saturation, or IR (Chan et al., 2006). BER also eliminates deaminated bases and DNA single-strand breaks (SSBs). As a first step in BER, a damage-specific DNA glycosylase (e.g. hOOG1, NEIL1 or NEIL2) recognizes and excises the damaged base, leading to the formation of a potentially cytotoxic intermediate apurinic or apyrimidinic site (AP site) (Bandaru et al., 2002; Boiteux & Radicella, 2000). The abasic sugar is cleaved by an AP endonuclease (APE1), which generates a strand break that is further processed by PARP, DNA polymerase *β* and ligase III in either short-patch or long-patch pathways (Fortini & Dogliotti, 2007). PARP not only recognizes the intermediate SSB but also acts as a damage transducer amplifying the damage signal by linking poly(ADP-ribose) (PAR) chains to its substrates, including itself. These polymers bind specific proteins, including XRCC1, DNA ligase III, p53 and DNA-PK,

**2.3.3 Double-strand break repair pathways**

Moshous et al., 2001).

It is generally accepted that the DNA double-strand break (DSB) is one of the most toxic and mutagenic DNA lesions occurring in human cells. A single DSB can, if left unrepaired, lead to the loss of a chromosome fragment and, thus, the death of the cell. However, despite the potential danger posed by DSBs, eukaryotic cells have evolved ways of improving biological processes based on the controlled induction of a DSB. Examples of this include the generation of variation during meiosis (Inagaki et al., 2010) and in the immune system (Fugmann et al., 2000), and the relaxation of supercoiled DNA by topoisomerases. Another endogenous source of DSBs are reactive oxygen species (ROS) produced by normal cellular processes, such as oxidative respiration, cytochrome P450 metabolism, peroxisomes and inflammatory

SiDNA and Other Tools for the Indirect Induction of DNA Damage Responses 337

DSB repair occurs via two main pathways: non homologous end-joining (NHEJ) and homologous recombination (HR) repair (Wyman & Kanaar, 2006). In mammalian cells, NHEJ is the major pathway for repairing breaks not associated with replication. This process may occur in all phases of the cell cycle, but predominantly in G1 phase. NHEJ involves the direct rejoining of two damaged DNA ends in a sequence-independent manner (Helleday et al., 2007; Weterings & van Gent, 2004). This end-joining mechanism is very precise for blunt ends and other simple end structures (van Heemst et al., 2004). However, the processing of incompatible ends may result in sequence alterations, such as deletions, occurring at "complicated" breaks. DNA double-strand breaks are first sensed by the ring-shaped Ku70/80 heterodimer. This DNA-Ku70/80 complex then attracts and activates the serine/threonine kinase activity of the DNA-PK catalytic subunit (DNA-PKcs). Following correct end alignment, DNA-PKcs is autophosphorylated (Weterings & Chen, 2007) and makes the ends available for ligation by ligase IV/XRCC4. Another essential NHEJ factor involved in the ligation of DSBs is XLF/Cernunnos (Ahnesorg et al., 2006; Buck et al., 2006). The MRN (Mre11 (Meiotic recombination 11)-Rad50-Nbs1 (Nijmegen breakage syndrome 1)) complex may facilitate the alignment of the two DNA ends, particularly when end processing is required (de Jager et al., 2001; Moreno-Herrero et al., 2005). The processing of "complex" lesions, such as hairpins, damaged backbone sugar residues, damaged bases, aberrant 5' hydroxyl groups or 3' phosphate groups, may involve polynucleotide kinase (Chappell et al., 2002; Koch et al., 2004), the RecQ helicase WRN (Perry et al., 2006), DNA polymerases *μ* and *λ* (Nick McElhinny et al., 2005) and the structure-specific nuclease Artemis (Ma et al., 2002;

It has recently been suggested that there is an alternative or "backup"-NHEJ (B-NHEJ) pathway that functions in conditions in which the NHEJ pathway is compromised (Iliakis, 2009). The B-NHEJ pathway seems to be dependent principally on histone H1 (Rosidi et al., 2008), PARP, which binds to DSBs with an even greater affinity than that with which it binds

Whereas NHEJ repairs DNA in a template-independent fashion by rejoining two broken ends, HR can accurately resynthesize damaged or missing sequence information at the break site, using homologous sequences as a template, preferably the adjacent sister chromatid in S or G2 phase. Several mechanisms of HR have been identified (reviewed for example by Helleday et al. (2007) and Hartlerode & Scully (2009)). All are initiated by 5'→3' resection at the DSB end, facilitated by the MRN complex (Paull & Gellert, 1998), which plays a critical role in the sensing of DSBs for HR. The MRN complex also recruits and helps to activate ATM (Lee & Paull, 2004; 2005). In addition to MRN, other factors, including CtIP (CTBP-interacting

SSBs (D'Silva et al., 1999), and DNA ligase III/XRCC1 (Audebert et al., 2004).

responses. Examples of exogenous sources of DSBs will be described below.

affecting the repair process as well as downstream responses to DNA damage (Malanga & Althaus, 2005).

#### **2.3 Repair pathways that lead to PIKK activation**

Most repair pathways involve the activation of PIKKs as transducers, especially if DNA breaks persist. Since PARPs can also sense DNA breaks, an implication of these enzymes in the pathways described in the following cannot be excluded.

#### **2.3.1 Nucleotide excision repair**

The nucleotide excision repair (NER) pathway senses and repairs various bulky, helix-distorting lesions that block DNA replication and transcription (Hanawalt, 2002). These lesions may arise, for example, following exposure to genotoxic compounds, such as polycyclic aromatic hydrocarbons or cisplatin. Two major repair mechanisms are known to be involved in this pathway: transcription-coupled repair, which specifically targets lesions blocking RNA polymerase II, and global genome repair, which deals with lesions in the rest of the genome (Cleaver, 2005). The damage sensors involved in transcription-coupled repair include, in addition to RNA polymerase II, Cockayne Syndrome A and B proteins. By contrast, XPA, Rpa and the XPC-hHR23B complex recognize lesions during global genome NER (Brown et al., 2010; Reardon & Sancar, 2005). NER is a complex multistep process involving the recognition of disrupted base pairing followed by unwinding of the DNA helix around the lesion and dual incision. The oligonucleotide patch carrying the lesion is excised, and the remaining gap is filled by regular DNA replication, using the intact complementary strand as a template. The main transducer kinase activated by the NER pathway is probably ATR, in response to UV-induced DNA damage in particular (Shell et al., 2009).

#### **2.3.2 Mismatch repair**

Mismatch repair (MMR) targets mispaired bases and nucleotides and insertion-deletion loops that arise through errors in DNA replication. The mechanisms by which eukaryotic cells distinguish mismatched from correctly matched bases in non replicating DNA remain unclear, but it is thought that recognition involves the contact of MMR proteins with the replication machinery. The eukaryotic mismatch sensors are the heterodimeric hMutS*α* (MSH2-MSH6) and hMutS*β* (MSH2-MSH3) complexes. Whereas hMutS*α* preferentially recognizes base-base mismatches and insertion/deletion mispairs of one or two nucleotides, hMutS*β* recognizes larger insertion/deletion mispairs (Li, 2008). The removal of mismatched bases and the restoration of strand integrity resemble the processes occurring in BER and NER. MMR proteins can interact with proteins in other repair pathways, such as BER, NER and homologous recombination, suggesting coordinated crosstalk between these processes (Kunkel & Erie, 2005). hMutS*α* and hMutS*β* may directly activate DNA damage signaling by physical interaction with ATM, ATR-ATRIP, c-Abl, and the p53-related transcription factor p73 (Kim et al., 2007; Shimodaira et al., 2003; Yoshioka et al., 2006). Consistently, hMutS*α* and hMutS*β*-deficient cells are defective in cell cycle arrest in response to multiple types of DNA damaging agents (Li, 2008). Another model proposes that a DDR could be activated by DNA breaks that are produced during "futile" DNA repair cycles. This model suggests that strand-specific MMR, which targets only newly replicated DNA, engages in repetitive repair cycles when it encounters a DNA lesion in the template strand, and this futile cycling activates ATR and/or ATM signaling leading to cell cycle arrest and apoptosis (Li, 1999; 2008).

#### **2.3.3 Double-strand break repair pathways**

4 Will-be-set-by-IN-TECH

affecting the repair process as well as downstream responses to DNA damage (Malanga &

Most repair pathways involve the activation of PIKKs as transducers, especially if DNA breaks persist. Since PARPs can also sense DNA breaks, an implication of these enzymes in the

The nucleotide excision repair (NER) pathway senses and repairs various bulky, helix-distorting lesions that block DNA replication and transcription (Hanawalt, 2002). These lesions may arise, for example, following exposure to genotoxic compounds, such as polycyclic aromatic hydrocarbons or cisplatin. Two major repair mechanisms are known to be involved in this pathway: transcription-coupled repair, which specifically targets lesions blocking RNA polymerase II, and global genome repair, which deals with lesions in the rest of the genome (Cleaver, 2005). The damage sensors involved in transcription-coupled repair include, in addition to RNA polymerase II, Cockayne Syndrome A and B proteins. By contrast, XPA, Rpa and the XPC-hHR23B complex recognize lesions during global genome NER (Brown et al., 2010; Reardon & Sancar, 2005). NER is a complex multistep process involving the recognition of disrupted base pairing followed by unwinding of the DNA helix around the lesion and dual incision. The oligonucleotide patch carrying the lesion is excised, and the remaining gap is filled by regular DNA replication, using the intact complementary strand as a template. The main transducer kinase activated by the NER pathway is probably ATR, in

Mismatch repair (MMR) targets mispaired bases and nucleotides and insertion-deletion loops that arise through errors in DNA replication. The mechanisms by which eukaryotic cells distinguish mismatched from correctly matched bases in non replicating DNA remain unclear, but it is thought that recognition involves the contact of MMR proteins with the replication machinery. The eukaryotic mismatch sensors are the heterodimeric hMutS*α* (MSH2-MSH6) and hMutS*β* (MSH2-MSH3) complexes. Whereas hMutS*α* preferentially recognizes base-base mismatches and insertion/deletion mispairs of one or two nucleotides, hMutS*β* recognizes larger insertion/deletion mispairs (Li, 2008). The removal of mismatched bases and the restoration of strand integrity resemble the processes occurring in BER and NER. MMR proteins can interact with proteins in other repair pathways, such as BER, NER and homologous recombination, suggesting coordinated crosstalk between these processes (Kunkel & Erie, 2005). hMutS*α* and hMutS*β* may directly activate DNA damage signaling by physical interaction with ATM, ATR-ATRIP, c-Abl, and the p53-related transcription factor p73 (Kim et al., 2007; Shimodaira et al., 2003; Yoshioka et al., 2006). Consistently, hMutS*α* and hMutS*β*-deficient cells are defective in cell cycle arrest in response to multiple types of DNA damaging agents (Li, 2008). Another model proposes that a DDR could be activated by DNA breaks that are produced during "futile" DNA repair cycles. This model suggests that strand-specific MMR, which targets only newly replicated DNA, engages in repetitive repair cycles when it encounters a DNA lesion in the template strand, and this futile cycling activates

ATR and/or ATM signaling leading to cell cycle arrest and apoptosis (Li, 1999; 2008).

Althaus, 2005).

**2.3 Repair pathways that lead to PIKK activation**

**2.3.1 Nucleotide excision repair**

**2.3.2 Mismatch repair**

pathways described in the following cannot be excluded.

response to UV-induced DNA damage in particular (Shell et al., 2009).

It is generally accepted that the DNA double-strand break (DSB) is one of the most toxic and mutagenic DNA lesions occurring in human cells. A single DSB can, if left unrepaired, lead to the loss of a chromosome fragment and, thus, the death of the cell. However, despite the potential danger posed by DSBs, eukaryotic cells have evolved ways of improving biological processes based on the controlled induction of a DSB. Examples of this include the generation of variation during meiosis (Inagaki et al., 2010) and in the immune system (Fugmann et al., 2000), and the relaxation of supercoiled DNA by topoisomerases. Another endogenous source of DSBs are reactive oxygen species (ROS) produced by normal cellular processes, such as oxidative respiration, cytochrome P450 metabolism, peroxisomes and inflammatory responses. Examples of exogenous sources of DSBs will be described below.

DSB repair occurs via two main pathways: non homologous end-joining (NHEJ) and homologous recombination (HR) repair (Wyman & Kanaar, 2006). In mammalian cells, NHEJ is the major pathway for repairing breaks not associated with replication. This process may occur in all phases of the cell cycle, but predominantly in G1 phase. NHEJ involves the direct rejoining of two damaged DNA ends in a sequence-independent manner (Helleday et al., 2007; Weterings & van Gent, 2004). This end-joining mechanism is very precise for blunt ends and other simple end structures (van Heemst et al., 2004). However, the processing of incompatible ends may result in sequence alterations, such as deletions, occurring at "complicated" breaks. DNA double-strand breaks are first sensed by the ring-shaped Ku70/80 heterodimer. This DNA-Ku70/80 complex then attracts and activates the serine/threonine kinase activity of the DNA-PK catalytic subunit (DNA-PKcs). Following correct end alignment, DNA-PKcs is autophosphorylated (Weterings & Chen, 2007) and makes the ends available for ligation by ligase IV/XRCC4. Another essential NHEJ factor involved in the ligation of DSBs is XLF/Cernunnos (Ahnesorg et al., 2006; Buck et al., 2006). The MRN (Mre11 (Meiotic recombination 11)-Rad50-Nbs1 (Nijmegen breakage syndrome 1)) complex may facilitate the alignment of the two DNA ends, particularly when end processing is required (de Jager et al., 2001; Moreno-Herrero et al., 2005). The processing of "complex" lesions, such as hairpins, damaged backbone sugar residues, damaged bases, aberrant 5' hydroxyl groups or 3' phosphate groups, may involve polynucleotide kinase (Chappell et al., 2002; Koch et al., 2004), the RecQ helicase WRN (Perry et al., 2006), DNA polymerases *μ* and *λ* (Nick McElhinny et al., 2005) and the structure-specific nuclease Artemis (Ma et al., 2002; Moshous et al., 2001).

It has recently been suggested that there is an alternative or "backup"-NHEJ (B-NHEJ) pathway that functions in conditions in which the NHEJ pathway is compromised (Iliakis, 2009). The B-NHEJ pathway seems to be dependent principally on histone H1 (Rosidi et al., 2008), PARP, which binds to DSBs with an even greater affinity than that with which it binds SSBs (D'Silva et al., 1999), and DNA ligase III/XRCC1 (Audebert et al., 2004).

Whereas NHEJ repairs DNA in a template-independent fashion by rejoining two broken ends, HR can accurately resynthesize damaged or missing sequence information at the break site, using homologous sequences as a template, preferably the adjacent sister chromatid in S or G2 phase. Several mechanisms of HR have been identified (reviewed for example by Helleday et al. (2007) and Hartlerode & Scully (2009)). All are initiated by 5'→3' resection at the DSB end, facilitated by the MRN complex (Paull & Gellert, 1998), which plays a critical role in the sensing of DSBs for HR. The MRN complex also recruits and helps to activate ATM (Lee & Paull, 2004; 2005). In addition to MRN, other factors, including CtIP (CTBP-interacting

Fig. 2. Transformation of DNA damage. DNA lesions are normally repaired by the

of a stalled replication fork results frequently in DSBs. DNA damage symbols and

abbreviations are as for Figure 1.

with the DNA (Hsiang et al., 1989; Kohn et al., 1987).

corresponding repair pathways. However, deficient repair may result in SSBs or DSBs. If a lesion persists during S-Phase (blue circle), stalled replication forks may arise. The collapse

SiDNA and Other Tools for the Indirect Induction of DNA Damage Responses 339

progression are adducts of DNA bases (Helleday et al., 2008). By the same mechanism, inhibitors of DNA synthesis may also indirectly cause DSBs, as they impair replication fork progression (Lundin et al., 2002). Such inhibitors include aphidicolin, which inhibits DNA polymerases (Ikegami et al., 1978) and hydroxyurea, an inhibitor of ribonucleotide reductase (Bianchi et al., 1986). Topoisomerase inhibitors induce DSBs by exploiting the natural activity of topoisomerases during DNA replication. Topoisomerases resolve the DNA torsions induced during replication, by introducing a transient break in the DNA. Inhibitors of topoisomerases prevent the resealing of the break, by trapping the enzyme in a complex

Thus, DSBs are the final outcome of unrepaired damage at the end of all these transformation processes (Figure 2). It is therefore not surprising that redundant and well regulated mechanisms have evolved for detecting, in particular, the presence of this toxic lesion and for activating DDR. DSBs can activate DNA-PK directly and they also activate ATM and ATR after end resection (Lopez-Contreras & Fernandez-Capetillo, 2010; Smith et al., 2010). Under certain conditions, PARP may also signal the presence of a DSB (Iliakis, 2009). The direct precursors of DSBs – SSBs and stalled replication forks – may themselves induce DDR, but there is less redundancy in the detection of these structures. SSBs are probably recognized and signaled to damage checkpoints mostly by PARP (Bouchard et al., 2003) and aberrant replication forks induce ATR activity through the recognition of RPA-coated stretches of ssDNA (Lopez-Contreras & Fernandez-Capetillo, 2010). The lack of redundancy in the signaling of these structures may account for their frequent transformation into DSBs.

protein), Exo1 and BLM (Bloom's syndrome protein), are required for 5'-end resection in mammalian cells (Hartlerode & Scully, 2009; Sartori et al., 2007; Yun & Hiom, 2009). After resection, single-stranded DNA (ssDNA) rapidly binds the ssDNA-binding protein RPA, which is then replaced by multimers of the Rad51 recombinase, forming a nucleoprotein filament at the end of the ssDNA. Rad51 loading involves direct interaction with BRCA2 (Pellegrini et al., 2002) and other factors (Hartlerode & Scully, 2009; Sy et al., 2009). The Rad51 nucleoprotein filament then captures double-stranded DNA (dsDNA) and scans it for homology (Bianco et al., 1998). When a homologous region is encountered, the 3'-end of the invading strand is extended by a polymerase, using the duplex DNA as a template. From this stage on, the repair pathway may diverge. The DSBR (DNA double-strand break repair pathway, also known as the double Holliday junction model) pathway mostly results in chromosomal crossover, whereas the SDSA (synthesis-dependent strand annealing) pathway ends with non crossover products (Johnson & Jasin, 2000; Liu & West, 2004; Van Dyck et al., 2001).

#### **2.4 Dynamics and heterogeneity of DNA damage**

One challenge in the study of the cellular response to DNA damage is the multitude of lesions introduced by most genotoxic agents. For instance, the exposure of cells to IR results in damage to all components of the cell, including lipids, proteins and nucleic acids. IR acts directly on the DNA, causing breaks in its phosphodiester backbone. This process accounts for about 30% of the DNA damage induced by IR (Chapman et al., 1973). The radicals produced by the indirect effects of radiation may account for as much as 70% of the DNA damage induced by IR (Chapman et al., 1973). These radicals damage DNA, resulting in a wide diversity of DNA lesions, such as damage to bases and the backbone sugar (oxidation, rearrangement, adducts), intrastrand crosslinks, the formation of abasic sites, single- and double-strand breaks and DNA-protein crosslinks (Jeggo & Lavin, 2009). Complex lesions, such as clustered DSBs and LMDS (locally multiply damaged sites) may also occur. After these complex lesions, DSBs are the most harmful lesions to the cell (Ward, 1975). It has been shown, in rodent cells, that the extent of cell death is directly correlated with the yield of DSB under various X-ray irradiation conditions (Radford, 1985). IR is therefore often used in investigations of the cellular response to DSBs. However, DSBs are not the most frequent type of lesion induced by IR. A dose of 1 Gy, for example, induces about 1000 SSBs and 150 protein-DNA crosslinks, but only 40 DSBs (Friedberg, 1995).

The reaction of various alkylating agents with DNA leads to the formation of highly heterogeneous products. Some agents may preferentially produce certain alkylation products, but the DNA damage generated is never limited to a single type (De Bont & van Larebeke, 2004). Furthermore, as for IR, other cell components, including proteins and ribonucleic acids, may be modified. Cellular responses to these modifications, such as activation of the proteasomal degradation pathway, may interfere with DDR pathways, or be involved in crosstalk with these pathways.

One type of damage can be transformed into another by inefficient repair and DNA replication or transcription (Figure 2). As described above, DNA repair pathways, such as BER, MMR and NER, generate intermediate SSBs. These SSBs may result in DSBs, if the repair is incomplete and the lesion persists (Bonner et al., 2008). The transformation of SSBs into DSBs occurs, for example, when replication forks encounter a SSB on the template and collapse (Strumberg et al., 2000) (Figure 2). Common types of DNA damage interfering with replication fork 6 Will-be-set-by-IN-TECH

protein), Exo1 and BLM (Bloom's syndrome protein), are required for 5'-end resection in mammalian cells (Hartlerode & Scully, 2009; Sartori et al., 2007; Yun & Hiom, 2009). After resection, single-stranded DNA (ssDNA) rapidly binds the ssDNA-binding protein RPA, which is then replaced by multimers of the Rad51 recombinase, forming a nucleoprotein filament at the end of the ssDNA. Rad51 loading involves direct interaction with BRCA2 (Pellegrini et al., 2002) and other factors (Hartlerode & Scully, 2009; Sy et al., 2009). The Rad51 nucleoprotein filament then captures double-stranded DNA (dsDNA) and scans it for homology (Bianco et al., 1998). When a homologous region is encountered, the 3'-end of the invading strand is extended by a polymerase, using the duplex DNA as a template. From this stage on, the repair pathway may diverge. The DSBR (DNA double-strand break repair pathway, also known as the double Holliday junction model) pathway mostly results in chromosomal crossover, whereas the SDSA (synthesis-dependent strand annealing) pathway ends with non crossover products (Johnson & Jasin, 2000; Liu & West, 2004; Van Dyck et al.,

One challenge in the study of the cellular response to DNA damage is the multitude of lesions introduced by most genotoxic agents. For instance, the exposure of cells to IR results in damage to all components of the cell, including lipids, proteins and nucleic acids. IR acts directly on the DNA, causing breaks in its phosphodiester backbone. This process accounts for about 30% of the DNA damage induced by IR (Chapman et al., 1973). The radicals produced by the indirect effects of radiation may account for as much as 70% of the DNA damage induced by IR (Chapman et al., 1973). These radicals damage DNA, resulting in a wide diversity of DNA lesions, such as damage to bases and the backbone sugar (oxidation, rearrangement, adducts), intrastrand crosslinks, the formation of abasic sites, single- and double-strand breaks and DNA-protein crosslinks (Jeggo & Lavin, 2009). Complex lesions, such as clustered DSBs and LMDS (locally multiply damaged sites) may also occur. After these complex lesions, DSBs are the most harmful lesions to the cell (Ward, 1975). It has been shown, in rodent cells, that the extent of cell death is directly correlated with the yield of DSB under various X-ray irradiation conditions (Radford, 1985). IR is therefore often used in investigations of the cellular response to DSBs. However, DSBs are not the most frequent type of lesion induced by IR. A dose of 1 Gy, for example, induces about 1000 SSBs and 150

The reaction of various alkylating agents with DNA leads to the formation of highly heterogeneous products. Some agents may preferentially produce certain alkylation products, but the DNA damage generated is never limited to a single type (De Bont & van Larebeke, 2004). Furthermore, as for IR, other cell components, including proteins and ribonucleic acids, may be modified. Cellular responses to these modifications, such as activation of the proteasomal degradation pathway, may interfere with DDR pathways, or be involved

One type of damage can be transformed into another by inefficient repair and DNA replication or transcription (Figure 2). As described above, DNA repair pathways, such as BER, MMR and NER, generate intermediate SSBs. These SSBs may result in DSBs, if the repair is incomplete and the lesion persists (Bonner et al., 2008). The transformation of SSBs into DSBs occurs, for example, when replication forks encounter a SSB on the template and collapse (Strumberg et al., 2000) (Figure 2). Common types of DNA damage interfering with replication fork

2001).

**2.4 Dynamics and heterogeneity of DNA damage**

protein-DNA crosslinks, but only 40 DSBs (Friedberg, 1995).

in crosstalk with these pathways.

Fig. 2. Transformation of DNA damage. DNA lesions are normally repaired by the corresponding repair pathways. However, deficient repair may result in SSBs or DSBs. If a lesion persists during S-Phase (blue circle), stalled replication forks may arise. The collapse of a stalled replication fork results frequently in DSBs. DNA damage symbols and abbreviations are as for Figure 1.

progression are adducts of DNA bases (Helleday et al., 2008). By the same mechanism, inhibitors of DNA synthesis may also indirectly cause DSBs, as they impair replication fork progression (Lundin et al., 2002). Such inhibitors include aphidicolin, which inhibits DNA polymerases (Ikegami et al., 1978) and hydroxyurea, an inhibitor of ribonucleotide reductase (Bianchi et al., 1986). Topoisomerase inhibitors induce DSBs by exploiting the natural activity of topoisomerases during DNA replication. Topoisomerases resolve the DNA torsions induced during replication, by introducing a transient break in the DNA. Inhibitors of topoisomerases prevent the resealing of the break, by trapping the enzyme in a complex with the DNA (Hsiang et al., 1989; Kohn et al., 1987).

Thus, DSBs are the final outcome of unrepaired damage at the end of all these transformation processes (Figure 2). It is therefore not surprising that redundant and well regulated mechanisms have evolved for detecting, in particular, the presence of this toxic lesion and for activating DDR. DSBs can activate DNA-PK directly and they also activate ATM and ATR after end resection (Lopez-Contreras & Fernandez-Capetillo, 2010; Smith et al., 2010). Under certain conditions, PARP may also signal the presence of a DSB (Iliakis, 2009). The direct precursors of DSBs – SSBs and stalled replication forks – may themselves induce DDR, but there is less redundancy in the detection of these structures. SSBs are probably recognized and signaled to damage checkpoints mostly by PARP (Bouchard et al., 2003) and aberrant replication forks induce ATR activity through the recognition of RPA-coated stretches of ssDNA (Lopez-Contreras & Fernandez-Capetillo, 2010). The lack of redundancy in the signaling of these structures may account for their frequent transformation into DSBs.

DNA break sensors include the MRN complex (Rupnik et al., 2008), Ku70/80, PARP and RPA. Break recognition by these sensors leads to the activation of transducer kinases, such as the PIKKs ATM, ATR and DNA-PK as well as PARP (Iliakis, 2009; Stiff et al., 2004; Ward & Chen, 2001). ATM, DNA-PK and ATR can phosphorylate the serine 139 residue of the histone variant H2AX (yielding *γ*-H2AX), at nucleosomes around the lesion. H2AX phosphorylation is probably the earliest posttranslational modification in DDR and may be considered the initial signal amplification step. *γ*-H2AX formation is followed by binding of the mediator protein MDC1 (mediator of DNA damage checkpoint 1) to the DSB-flanking chromatin (Jungmichel & Stucki, 2010). Mediator or adaptor proteins help to transmit, enhance and sustain the signaling between sensors and transducers, leading to the spread of the repair machinery along the chromosome. Other mediators include 53BP1 (p53-binding protein 1) and BRCA1 (Li & Zou, 2005; Misteli & Soutoglou, 2009). The recruitment of 53BP1 and BRCA1 to the DSB is indirect, requiring the activity of the E3 ubiquitin ligases RNF8 (Huen et al., 2007; Kolas et al., 2007) and RNF168 (Doil et al., 2009; Stewart et al., 2009). ssDNA compartments may be bound by RPA, which subsequently recruits ATR (Lobrich & Jeggo, 2007) and Rad51. Foci of Rad51 binding colocalize with Rad52 (Liu & Maizels, 2000), Rad54 (Essers et al., 2002), RPA (Raderschall et al., 1999), BRCA1 (Scully et al., 1997) and BRCA2 (Chen et al., 1998). Sustained activation of the transducers results in the transmission of the damage signal to effectors, which relay the signal to downstream pathways with endpoints in different cellular processes, such as

SiDNA and Other Tools for the Indirect Induction of DNA Damage Responses 341

Not all the actors in DNA damage signaling and repair form characteristic foci. Unlike ATM, MDC1, 53BP1, BRCA1/BARD1 and MRN, the central NHEJ proteins Ku70/80 and DNA-PK do not spread to the adjacent chromatin upon recruitment to the break, presumably because they are required at low copy number at sites of damage (Bekker-Jensen et al., 2006; Lukas et al., 2003). The same is true for the effector kinases Chk1 and Chk2, and for p53, which interact only transiently with damage sites, subsequently diffusing rapidly to relay the signal

One endpoint of the described signaling cascade in response to DNA damage is the activation of checkpoints to provide the cell with more time for DNA repair. DDR checkpoints have been identified at the G1/S and G2/M boundaries, and during S phase and, potentially, in mitosis (reviewed by Lukas et al. (2004)). After activation, the transducer kinases ATM, ATR or DNA-PK phosphorylate p53 either directly or via ATM-induced activation of the effector kinase Chk (checkpoint kinase) 2. Phosphorylated p53 then induces transcription of the gene encoding the Cdk inhibitor p21, which ultimately prevents transition from G1 to S-phase. Both Chk 1 and 2 activate the G2/M and intra-S checkpoints (Smith et al., 2010). It was long thought that ATM principally phosphorylated Chk1 and that ATR preferentially phosphorylated Chk2. However, this view has been modified by the discovery of various crosstalk between these kinases (Bartek & Lukas, 2003). The precise role of DNA-PK in this regulation remains unclear. PARP may contribute to checkpoint signaling by activation of p53. p53 exhibits high affinity for automodified PARP (Malanga et al., 1998) and p53 functions are impaired in PARP-deficient cells (Wang et al., 1998; Wieler et al., 2003). Furthermore, PARP activation in response to excessive DNA damage leads to extensive NAD<sup>+</sup> consumption. The cellular NAD<sup>+</sup> depletion can induce cell death through several mechanisms, depending on

checkpoint arrest or apoptosis (Kastan & Bartek, 2004) (Figure 3).

the cellular context (reviewed by Rouleau et al. (2010)).

to their soluble downstream targets.

#### **2.5 Cellular DNA damage response**

DNA breaks, including DSBs in particular, induce a highly coordinated DDR process leading to signal amplification, enhanced repair functions, cell cycle arrest or apoptosis. Many proteins are implicated in the DDR, which involves complex spatial and temporal coordination and many dynamic interactions between repair proteins and DNA.

#### **2.5.1 Spatiotemporal organization of the DNA damage response**

The components of the DDR pathway may be classified roughly as DNA-damage sensors, mediators, transducers and effectors (Figure 3A). After the sensing of a DNA break, mediator and repair proteins rapidly accumulate on the chromatin surrounding the lesion, to form subnuclear repair foci (Fernandez-Capetillo et al., 2003) (Figure 3B). Protein recruitment to DSBs normally occurs in a hierarchical manner and involves multiple posttranslational modifications, such as phosphorylation, ubiquitination, PARylation or acetylation (Essers et al., 2002; Lukas et al., 2004; Polo & Jackson, 2011). The massive accumulation of DNA repair and signaling factors may lead to structural stabilization of the break. The amplification and maintenance of the DNA-damage signal through the recruitment of multiple copies of transducer kinases to sites of damage is probably an even more important function (Misteli & Soutoglou, 2009).

Fig. 3. The DDR signal transduction cascade. (A) DNA damage is first physically recognized by sensor proteins (gray). Mediator proteins (blue) facilitate the recruitment and activation of transducer kinases (red). A positive feedback loop between mediators and transducers leads to the maintenance and amplification of the signal. The transducer kinases then phosphorylate various effector proteins (green), including kinases, transcription factors and repair proteins. Depending on the severity of the damage, this can lead to various cellular responses (purple). (B) Formation of multiprotein complexes at the sites of DSBs (Ward & Chen, 2004) and microscopic visualization of the formation of *γ*-H2AX foci in response to IR. The exposure of cells to IR results in the rapid recruitment of numerous proteins to the sites of DNA lesions. The signal transducing kinases ATM and the related DNA-PK initiate a cascade of phosphorylation events (P), amplifying the signal to activate, if necessary, cell cycle checkpoint pathways or apoptosis, in situations in which the damage is too great to be repaired.

8 Will-be-set-by-IN-TECH

DNA breaks, including DSBs in particular, induce a highly coordinated DDR process leading to signal amplification, enhanced repair functions, cell cycle arrest or apoptosis. Many proteins are implicated in the DDR, which involves complex spatial and temporal

The components of the DDR pathway may be classified roughly as DNA-damage sensors, mediators, transducers and effectors (Figure 3A). After the sensing of a DNA break, mediator and repair proteins rapidly accumulate on the chromatin surrounding the lesion, to form subnuclear repair foci (Fernandez-Capetillo et al., 2003) (Figure 3B). Protein recruitment to DSBs normally occurs in a hierarchical manner and involves multiple posttranslational modifications, such as phosphorylation, ubiquitination, PARylation or acetylation (Essers et al., 2002; Lukas et al., 2004; Polo & Jackson, 2011). The massive accumulation of DNA repair and signaling factors may lead to structural stabilization of the break. The amplification and maintenance of the DNA-damage signal through the recruitment of multiple copies of transducer kinases to sites of damage is probably an even more important function (Misteli &

Fig. 3. The DDR signal transduction cascade. (A) DNA damage is first physically recognized by sensor proteins (gray). Mediator proteins (blue) facilitate the recruitment and activation of transducer kinases (red). A positive feedback loop between mediators and transducers leads

phosphorylate various effector proteins (green), including kinases, transcription factors and repair proteins. Depending on the severity of the damage, this can lead to various cellular responses (purple). (B) Formation of multiprotein complexes at the sites of DSBs (Ward & Chen, 2004) and microscopic visualization of the formation of *γ*-H2AX foci in response to IR. The exposure of cells to IR results in the rapid recruitment of numerous proteins to the sites of DNA lesions. The signal transducing kinases ATM and the related DNA-PK initiate a cascade of phosphorylation events (P), amplifying the signal to activate, if necessary, cell cycle checkpoint pathways or apoptosis, in situations in which the damage is too great to be

to the maintenance and amplification of the signal. The transducer kinases then

coordination and many dynamic interactions between repair proteins and DNA.

**2.5.1 Spatiotemporal organization of the DNA damage response**

**2.5 Cellular DNA damage response**

Soutoglou, 2009).

repaired.

DNA break sensors include the MRN complex (Rupnik et al., 2008), Ku70/80, PARP and RPA. Break recognition by these sensors leads to the activation of transducer kinases, such as the PIKKs ATM, ATR and DNA-PK as well as PARP (Iliakis, 2009; Stiff et al., 2004; Ward & Chen, 2001). ATM, DNA-PK and ATR can phosphorylate the serine 139 residue of the histone variant H2AX (yielding *γ*-H2AX), at nucleosomes around the lesion. H2AX phosphorylation is probably the earliest posttranslational modification in DDR and may be considered the initial signal amplification step. *γ*-H2AX formation is followed by binding of the mediator protein MDC1 (mediator of DNA damage checkpoint 1) to the DSB-flanking chromatin (Jungmichel & Stucki, 2010). Mediator or adaptor proteins help to transmit, enhance and sustain the signaling between sensors and transducers, leading to the spread of the repair machinery along the chromosome. Other mediators include 53BP1 (p53-binding protein 1) and BRCA1 (Li & Zou, 2005; Misteli & Soutoglou, 2009). The recruitment of 53BP1 and BRCA1 to the DSB is indirect, requiring the activity of the E3 ubiquitin ligases RNF8 (Huen et al., 2007; Kolas et al., 2007) and RNF168 (Doil et al., 2009; Stewart et al., 2009). ssDNA compartments may be bound by RPA, which subsequently recruits ATR (Lobrich & Jeggo, 2007) and Rad51. Foci of Rad51 binding colocalize with Rad52 (Liu & Maizels, 2000), Rad54 (Essers et al., 2002), RPA (Raderschall et al., 1999), BRCA1 (Scully et al., 1997) and BRCA2 (Chen et al., 1998). Sustained activation of the transducers results in the transmission of the damage signal to effectors, which relay the signal to downstream pathways with endpoints in different cellular processes, such as checkpoint arrest or apoptosis (Kastan & Bartek, 2004) (Figure 3).

Not all the actors in DNA damage signaling and repair form characteristic foci. Unlike ATM, MDC1, 53BP1, BRCA1/BARD1 and MRN, the central NHEJ proteins Ku70/80 and DNA-PK do not spread to the adjacent chromatin upon recruitment to the break, presumably because they are required at low copy number at sites of damage (Bekker-Jensen et al., 2006; Lukas et al., 2003). The same is true for the effector kinases Chk1 and Chk2, and for p53, which interact only transiently with damage sites, subsequently diffusing rapidly to relay the signal to their soluble downstream targets.

One endpoint of the described signaling cascade in response to DNA damage is the activation of checkpoints to provide the cell with more time for DNA repair. DDR checkpoints have been identified at the G1/S and G2/M boundaries, and during S phase and, potentially, in mitosis (reviewed by Lukas et al. (2004)). After activation, the transducer kinases ATM, ATR or DNA-PK phosphorylate p53 either directly or via ATM-induced activation of the effector kinase Chk (checkpoint kinase) 2. Phosphorylated p53 then induces transcription of the gene encoding the Cdk inhibitor p21, which ultimately prevents transition from G1 to S-phase. Both Chk 1 and 2 activate the G2/M and intra-S checkpoints (Smith et al., 2010). It was long thought that ATM principally phosphorylated Chk1 and that ATR preferentially phosphorylated Chk2. However, this view has been modified by the discovery of various crosstalk between these kinases (Bartek & Lukas, 2003). The precise role of DNA-PK in this regulation remains unclear. PARP may contribute to checkpoint signaling by activation of p53. p53 exhibits high affinity for automodified PARP (Malanga et al., 1998) and p53 functions are impaired in PARP-deficient cells (Wang et al., 1998; Wieler et al., 2003). Furthermore, PARP activation in response to excessive DNA damage leads to extensive NAD<sup>+</sup> consumption. The cellular NAD<sup>+</sup> depletion can induce cell death through several mechanisms, depending on the cellular context (reviewed by Rouleau et al. (2010)).

**3. Methods and mechanisms for inducing a damage response in the absence of**

The induction of DNA damage repair pathways by exogenous DNA was first reported in bacteria, into which UV-irradiated lambda bacteriophages (D'Ari & Huisman, 1982; George et al., 1974) or plasmids (Bailone et al., 1984) were introduced. This indirect response is controlled by activation at the sites of exogenous DNA damage of the RecA protein, the key enzyme of the bacterial damage signaling response known as the "SOS response" (see Schlacher & Goodman (2007)). No such mechanism was initially found in mammals, despite a few publications reporting that UV-irradiated H-1 parvovirus or SV40 simian virus induced an "SOS-like" repair pathway in infected mammalian cells (Cornelis et al., 1982). Elucidation of the mechanisms underlying the response to damaged DNA or RNA in the cell took much longer in mammals. We review here the mammalian cell response to oligonucleotides, viral and immunostimulatory DNA and the insight into DDR gained from the analysis of artificial

SiDNA and Other Tools for the Indirect Induction of DNA Damage Responses 343

In bacteria, ssDNA has been identified as the signal triggering the bacterial SOS repair response. The RecA protein (Rad51 in humans) is directly stimulated by ssDNA, inactivating the LexA repressor and triggering the repair response (Craig & Roberts, 1981). It has been suggested that ssDNA (at SSBs, stalled replication forks or resected DSBs) acts also as the major stimulatory signal for DNA damage responses in eukaryotic cells (Li & Deshaies, 1993; Nur et al., 2003). Several studies based on transfection or the microinjection of synthetic DNA oligonucleotides have analyzed the response of the cell to ssDNA. Studies by Nur et al. (2003) have shown that ssDNA acts upstream from ATM/p53 in DNA damage signaling. The transfection of cells with short (as few as 5 bases) ssDNA molecules with random sequences induced ATM activation and apoptosis, whereas very short (8 bp) dsDNA molecules did not (Nur et al., 2003). The induction of apoptosis by ssDNA is consistent with earlier studies, in which transfection with randomly fragmented DNA (Schiavone et al., 2000) or the nuclear injection of linearized plasmid DNA, circular DNA containing a gap, or single-stranded circular phagemids induced cell cycle arrest or apoptosis (Huang et al., 1996). The activation of ATM by fragmented DNA requires the MRN-assisted assembly of short, linear ssDNA fragments into high-molecular weight complexes, as shown by experiments in *Xenopus laevis*

So, if ssDNA can activate ATM directly, how is ATM activated in response to DNA double-strand breaks? When DNA double-strand breaks are sensed by the MRN complex, MRN partially unwinds the ends to expose ssDNA (Lee & Paull, 2005). It has been shown in *Xenopus laevis* egg extracts that 70 bp synthetic double-stranded molecules are rapidly resected in an MRN-dependent manner to generate ssDNA oligonucleotides, which activate ATM (Jazayeri et al., 2008). Consistent with these findings, the injection of small synthetic ssDNA oligomers into undamaged cells also induces ATM activation, and the elimination of ssDNA oligomers results in the rapid extinction of ATM activity. In summary, the results obtained from experiments with single-stranded or long double-stranded DNA fragments suggest that short ssDNA molecules are the essential signal for the induction of ATM-dependent cell cycle

**chromatin damage**

**3.1 Cellular response to DNA oligonucleotides**

extracts (Costanzo et al., 2004).

arrest and apoptosis.

repair foci.

#### **2.5.2 Complexity of DNA damage response regulation**

The outcomes of DNA damage signaling are, literally, a matter of life or death. Depending on the severity of DNA damage, the cell will either repair the damage to enable it to continue dividing or enter apoptosis. Complex, redundant signaling pathways converging on several central node proteins have emerged to ensure that this signaling remains under tight control. These node proteins must obtain input signals from several sources, in the form of protein modifications, before they can relay the signal to downstream effectors. This has a protective effect, greatly reducing the risk of important cell fate decisions, such as entry into apoptosis, occurring in response to a single erroneous input signal (Yarosh, 2001). The nodes in DDR include checkpoint proteins, such as Chk1 and Chk2, which control cell cycling and require input signals from several sources for full activation (McAdams & Arkin, 1999). Another prominent example for a node protein in DNA damage signaling is p53 (Kohn, 1999). The p53 protein has 11 sites for phosphorylation and acetylation, and can theoretically assume about 2000 modification states, if all the possible independent combinations are taken into account. Twelve different kinases can phosphorylate p53, and activated p53 can interact with at least 15 downstream proteins (Yarosh, 2001). The kinases that phosphorylate p53 in response to DNA damage include DNA-PK, ATM and ATR. Full p53 activity requires phosphorylation by both DNA-PK and ATM, at least (Wang et al., 2000). It is therefore now thought that p53 plays a key role in determining the degree of damage, through the assessment of input signals, on which the decision as to whether apoptosis is necessary is based (Kohn, 1999). There is a need to determine the specific conditions under which individual PIKKs become activated. Which (genotoxic) stresses lead to the activation of all transducer kinases? Is it possible to activate a single PIKK specifically, without affecting the others, and what are the cellular consequences of this?

There is also direct interplay between the transducer kinases (Chen et al., 2007). PIKKs can phosphorylate each other in response to DNA damage, resulting in mutual control of their activities. In addition to mediating posttranslational modifications, the kinases seem to regulate each other, either directly or indirectly (Peng et al., 2005). Studies on mutants and siRNA experiments have shown that a decrease in the amount of one of these kinases often leads to a decrease in the amounts of the PIKK sister kinases. The location of the kinases also seems to play an important role. For ATM, for example, the concentration of multiple copies in repair foci plays an important role in kinase activation, whereas DNA-PK does not need such an accumulation of multiple copies for full activation.

Another layer of complexity is added by the partially overlapping substrate specificities of the transducers ATM, DNA-PK and ATR (reviewed by Durocher & Jackson (2001) and Yang et al. (2003)). These transducers signal different types of DNA damage, but it was recently shown that the PIKK-mediated signaling network is highly extensive, with hundreds of phosphorylation events at ATR, ATM and DNA-PK consensus target sites induced by IR (Matsuoka et al., 2007). As discussed above, the plethora of types of damage induced by IR results in the activation of all three PIKKs. It remains to be determined which substrates are specific or overlapping for which transducer kinases in this long list of potential PIKK targets. Further insight into the contributions of individual repair signaling pathways has been provided by studies of the responses induced by damage signals in the absence of chromatin damage. This aspect will be discussed below.
