Artemis deficiency confers a DNA double-strand break repair defect and Artemis phosphorylation status is altered by DNA damage and cell cycle progression

Mutations in the Artemis gene are causative in a subset of human severe combined immunodeficiencies (SCIDs) and Artemis-deficient cells exhibit radiation sensitivity and defective V(D)J recombination, implicating Artemis function in non-homologous end joining (NHEJ). Here we show that Artemis-deficient cells from Athabascan-speaking Native American SCID patients (SCIDA) display significantly elevated sensitivity to ionizing radiation (IR) but only a very subtle defect in DNA double-strand (DSB) break repair in contrast to the severe DSB repair defect of NHEJ-deficient cells. Primary human SCIDA fibroblasts accumulate and exhibit persistent arrest at both the G1/S and G2/M boundaries in response to IR, consistent with the presence of persistent DNA damage. Artemis protein is phosphorylated in a PI3-like kinase-dependent manner after either IR or a number of other DNA damaging treatments including etoposide, but SCIDA cells are not hypersensitive to treatment with etoposide. Inhibitor studies with various DNA damaging agents establish multiple phosphorylation states and suggest multiple kinases function in Artemis phosphorylation. We observe that Artemis phosphorylation occurs rapidly after irradiation like that of histone H2AX. However, unlike H2AX, Artemis de-phosphorylation is uncoupled from overall DNA repair and correlates instead with cell cycle progression to or through mitosis. Our results implicate a direct and non-redundant function of Artemis in the repair of a small subset of DNA double-strand breaks, possibly those with hairpin termini, which may account for the pronounced radiation sensitivity observed in Artemis-deficient cells.     

Role of Artemis in DSB repair and guarding chromosomal stability following exposure to ionizing radiation at different stages of cell cycle

We analyzed the phenotype of cells derived from SCID patients with different mutations in the Artemis gene. Using clonogenic survival assay an increased sensitivity was found to X-rays (2-3-fold) and bleomycin (2-fold), as well as to etoposide, camptothecin and methylmethane sulphonate (up to 1.5-fold). In contrast, we did not find increased sensitivity to cross-linking agents mitomycin C and cis-platinum. The kinetics of DSB repair assessed by pulsed-field gel electrophoresis and gammaH2AX foci formation after ionizing irradiation, indicate that 15-20% of DSB are not repaired in Artemis-deficient cells. In order to get a better understanding of the repair defect in Artemis-deficient cells, we studied chromosomal damage at different stages of the cell cycle. In contrast to AT cells, Artemis-deficient cells appear to have a normal G(1)/S-block that resulted in a similar frequency of dicentrics and translocations, however, frequency of acentrics fragments was found to be 2-4-fold higher compared to normal fibroblasts. Irradiation in G(2) resulted in a higher frequency of chromatid-type aberrations (1.5-3-fold) than in normal cells, indicating that a fraction of DSB requires Artemis for proper repair. Our data are consistent with a function of Artemis protein in processing of a subset of complex DSB, without G(1) cell cycle checkpoint defects. This type of DSB can be induced in high proportion and persist through S-phase and in part might be responsible for the formation of chromatid-type exchanges in G(1)-irradiated Artemis-deficient cells. Among different human radiosensitive fibroblasts studied for endogenous (in untreated samples) as well as X-ray-induced DNA damage, the ranking order on the basis of higher incidence of spontaneously occurring chromosomal alterations and induced ones was: ligase 4> or =AT>Artemis. This observation implicates that in human fibroblasts following exposure to ionizing radiation a lower risk might be created when cells are devoid of endogenous damage.     

Harmonising the response to DSBs: a new string in the ATM bow

Ataxia telangiestasia mutated protein (ATM) is the major kinase that initiates the DNA damage signal transduction response following exposure to ionising radiation (IR) in mammalian cells. DNA non-homologous end-joining (NHEJ) is the most significant double strand break (DSB) repair pathway in mammalian cells. ATM-defective cell lines display cell cycle checkpoint defects and show pronounced radiosensitivity. ATM signalling was previously thought to be dispensable for NHEJ. This review discusses recent findings that ATM activates an end-processing mechanism dependent upon Artemis, a nuclease that also functions to cleave the hairpin intermediate generated during V(D)J recombination. ATM/Artemis-dependent end-processing is required for the repair of a sub-fraction (approximately 10%) of DSBs induced by IR and makes a significant contribution to survival following exposure to ionising radiation. This result represents a new role for ATM and demonstrates a novel cross communication between the DNA repair and signal transduction machinery.     

An Imperfect G2M Checkpoint Contributes to Chromosome Instability Following Irradiation of S and G2 Phase Cells

DNA double strand break (DSB) repair and checkpoint control represent two major mechanisms that function to reduce chromosomal instability following ionising irradiation (IR). Ataxia telangiectasia (A-T) cells have long been known to have defective checkpoint responses. Recent studies have shown that they also have a DSB repair defect following IR raising the issue of how ATM’s repair and checkpoint functions interplay to maintain chromosomal stability. A-T and Artemis cells manifest an identical and epistatic repair defect throughout the cell cycle demonstrating that ATM’s major repair defect following IR represents Artemis-dependent end-processing. Artemis cells show efficient G2/M checkpoint induction and a prolonged arrest relative to normal cells. Following irradiation of G2 cells, this checkpoint is dependent on ATM and A-T cells fail to show checkpoint arrest. In contrast, cells irradiated during S phase initiate a G2/M checkpoint which is independent of ATM and, significantly, both Artemis and A-T cells show a prolonged arrest at the G2/M checkpoint likely reflecting their repair defect. Strikingly, the G2/M checkpoint is released before the completion of repair when approximately 10-20 DSBs remain both for S phase and G2 phase irradiated cells. This defined sensitivity level of the G2/M checkpoint explains the prolonged arrest in repair-deficient relative to normal cells and provides a conceptual framework for the co-operative phenotype between checkpoint and repair functions in maintaining chromosomal stability.     

Methylator-induced, mismatch repair-dependent G2 arrest is activated through Chk1 and Chk2

SN1 DNA methylating agents such as the nitrosourea N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) elicit a G2/M checkpoint response via a mismatch repair (MMR) system-dependent mechanism; however, the exact nature of the mechanism governing MNNG-induced G2/M arrest and how MMR mechanistically participates in this process are unknown. Here, we show that MNNG exposure results in activation of the cell cycle checkpoint kinases ATM, Chk1, and Chk2, each of which has been implicated in the triggering of the G2/M checkpoint response. We document that MNNG induces a robust, dose-dependent G2 arrest in MMR and ATM-proficient cells, whereas this response is abrogated in MMR-deficient cells and attenuated in ATM-deficient cells treated with moderate doses of MNNG. Pharmacological and RNA interference approaches indicated that Chk1 and Chk2 are both required components for normal MNNG-induced G2 arrest. MNNG-induced nuclear exclusion of the cell cycle regulatory phosphatase Cdc25C occurred in an MMR-dependent manner and was compromised in cells lacking ATM. Finally, both Chk1 and Chk2 interact with the MMR protein MSH2, and this interaction is enhanced after MNNG exposure, supporting the notion that the MMR system functions as a molecular scaffold at the sites of DNA damage that facilitates activation of these kinases.     

DNA damage-induced G2-M checkpoint activation by histone H2AX and 53BP1

Activation of the ataxia telangiectasia mutated (ATM) kinase triggers diverse cellular responses to ionizing radiation (IR), including the initiation of cell cycle checkpoints. Histone H2AX, p53 binding-protein 1 (53BP1) and Chk2 are targets of ATM-mediated phosphorylation, but little is known about their roles in signalling the presence of DNA damage. Here, we show that mice lacking either H2AX or 53BP1, but not Chk2, manifest a G2-M checkpoint defect close to that observed in ATM(-/-) cells after exposure to low, but not high, doses of IR. Moreover, H2AX regulates the ability of 53BP1 to efficiently accumulate into IR-induced foci. We propose that at threshold levels of DNA damage, H2AX-mediated concentration of 53BP1 at double-strand breaks is essential for the amplification of signals that might otherwise be insufficient to prevent entry of damaged cells into mitosis.     

Role of a DNA damage checkpoint pathway in ionizing radiation-induced glioblastoma cell migration and invasion

Ionizing radiation (IR) induces a DNA damage response that includes activation of cell cycle checkpoints, leading to cell cycle arrest. In addition, IR enhances cell invasiveness of glioblastoma cells, among other tumor cell types. Using RNA interference, we found that the protein kinase MRK, previously implicated in the DNA damage response to IR, also inhibits IR-induced cell migration and invasion of glioblastoma cells. We showed that MRK activation by IR requires the checkpoint protein Nbs1 and that Nbs1 is also required for IR-stimulated migration. In addition, we show that MRK acts upstream of Chk2 and that Chk2 is also required for IR-stimulated migration and invasion. Thus, we have identified Nbs1, MRK, and Chk2 as elements of a novel signaling pathway that mediates IR-stimulated cell migration and invasion. Interestingly, we found that inhibition of cell cycle progression, either with the CDK1/2 inhibitor CGP74514A or by downregulation of the CDC25A protein phosphatase, restores IR-induced migration and invasion in cells depleted of MRK or Chk2. These data indicate that cell cycle progression, at least in the context of IR, exerts a negative control on the invasive properties of glioblastoma cells and that checkpoint proteins mediate IR-induced invasive behavior by controlling cell cycle arrest.     

Differential impact of diverse anticancer chemotherapeutics on the Cdc25A-degradation checkpoint pathway

When exposed to DNA-damaging insults such as ionizing radiation (IR) or ultraviolet light (UV), mammalian cells activate checkpoint pathways to halt cell cycle progression or induce cell death. Here we examined the ability of five commonly used anticancer drugs with different mechanisms of action to activate the Chk1/Chk2-Cdc25A-CDK2/cyclin E cell cycle checkpoint pathway, previously shown to be induced by IR or UV. Whereas exposure of human cells to topoisomerase inhibitors camptothecin, etoposide, or adriamycin resulted in rapid (within 1 h) activation of the pathway including degradation of the Cdc25A phosphatase and inhibition of cyclin E/CDK2 kinase activity, taxol failed to activate this checkpoint even after a prolonged treatment. Unexpectedly, although the alkylating agent cisplatin also induced degradation of Cdc25A (albeit delayed, after 8-12 h), cyclin E/CDK2 activity was elevated and DNA synthesis continued, a phenomena that correlated with increased E2F1 protein levels and consequently enhanced expression of cyclin E. These results reveal a differential impact of various classes of anticancer chemotherapeutics on the Cdc25A-degradation pathway, and indicate that the kinetics of checkpoint induction, and the relative balance of key components within the DNA damage response network may dictate whether the treated cells arrest their cell cycle progression.     

Leaky Scid phenotype associated with defective V(D)J coding end processing in Artemis-deficient mice

Radiosensitive severe combined immune deficiency in humans results from mutations in Artemis, a protein which, when coupled with DNA-dependent protein kinase catalytic subunit (DNA-PKcs), possesses DNA hairpin-opening activity in vitro. Here, we report that Artemis-deficient mice have an overall phenotype similar to that of DNA-PKcs-deficient mice-including severe combined immunodeficiency associated with defects in opening and joining V(D)J coding hairpin ends and increased cellular ionizing radiation sensitivity. While these findings strongly support the notion that Artemis functions with DNA-PKcs in a subset of NHEJ functions, differences between Artemis- and DNA-PKcs-deficient phenotypes, most notably decreased fidelity of V(D)J signal sequence joining in DNA-PKcs-deficient but not Artemis-deficient fibroblasts, suggest additional functions for DNA-PKcs. Finally, Artemis deficiency leads to chromosomal instability in fibroblasts, demonstrating that Artemis functions as a genomic caretaker.     

chk-1 is an Essential Gene and is Required for an S-M Checkpoint During Early Embryogenesis

The chk1 gene was first discovered in screens for radiation sensitive mutants in S. pombe [1]. Genetic analysis revealed that chk1 is involved in a DNA damage G2-M checkpoint. Chk1 becomes activated in response to DNA damage and prevents entry into mitosis by inhibiting the cell cycle machinery. This checkpoint decreases the risk of defective DNA being inherited by daughter cells, therefore reducing the risk of genetic instability. In higher eukaryotes, chk1 homologues have similar checkpoint functions. For example, an avian B-lymphoma cell line that is defective for Chk1 fails to arrest in G2-M after DNA damage. Nonetheless, these Chk1 defective cells are viable indicating that Chk1 is not essential for normal somatic cells to divide [2]. In spite of this, mouse and Drosophila homozygous Chk1 mutants die during embryogenesis suggesting that this is an essential gene for embryonic cell cycles [3, 4]. What particular role does Chk1 have in directing embryonic cell divisions? Here we used the model organism, C. elegans, to address the role of chk-1 during development. As expected, disruption of chk-1 by RNAi eliminated the DNA damage checkpoint response in C. elegans. In addition, we revealed that chk-1 was predominantly expressed during embryogenesis and in the postembryonic germline. Indeed, we found that chk-1 had an essential role in embryo and germline development. More specifically, disruption of chk-1 expression resulted in embryo lethality, which was attributed to a defect in an intrinsic S-M hence causing premature entry into M-phase.     

Chk1-deficient tumour cells are viable but exhibit multiple checkpoint and survival defects

The conserved protein kinase Chk1 is believed to play an important role in checkpoint responses to aberrant DNA structures; however, genetic analysis of Chk1 functions in metazoans is complicated by lethality of Chk1-deficient embryonic cells. We have used gene targeting to eliminate Chk1 function in somatic DT40 B-lymphoma cells. We find that Chk1-deficient DT40 cells are viable, but fail to arrest in G(2)/M in response to and are hypersensitive to killing by ionizing radiation. Chk1-deficient cells also fail to maintain viable replication forks or suppress futile origin firing when DNA polymerase is inhibited, leading to incomplete genome duplication and diminished cell survival after release from replication arrest. In contrast to embryonic cells, however, Chk1 is not required to delay mitosis when DNA synthesis is inhibited. Thus, Chk1 is dispensable for normal cell division in somatic DT40 cells but is essential for DNA damage-induced G(2)/M arrest and a subset of replication checkpoint responses. Furthermore, Chk1-dependent processes promote tumour cell survival after perturbations of DNA structure or metabolism.     

Mice lacking protein phosphatase 5 are defective in ataxia telangiectasia mutated (ATM)-mediated cell cycle arrest

Protein phosphatase 5 (Ppp5), a tetratricopeptide repeat domain protein, has been implicated in multiple cellular functions, including cellular proliferation, migration, differentiation and survival, and cell cycle checkpoint regulation via the ataxia telangiectasia mutated/ATM and Rad3-related (ATM/ATR) signal pathway. However, the physiological functions of Ppp5 have not been reported. To confirm the role of Ppp5 in cell cycle checkpoint regulation, we generated Ppp5-deficient mice and isolated mouse embryonic fibroblast (MEF) cells from Ppp5-deficient and littermate control embryos. Although Ppp5-deficient mice can survive through embryonic development and postnatal life and MEF cells from the Ppp5-deficient mice maintain normal replication checkpoint induced by hydroxyurea, Ppp5-deficient MEF cells display a significant defect in G(2)/M DNA damage checkpoint in response to ionizing radiation (IR). To determine whether this defect in IR-induced G(2)/M checkpoint is due to altered ATM-mediated signaling, we measured ATM kinase activity and ATM-mediated downstream events. Our data demonstrated that IR-induced ATM kinase activity is attenuated in Ppp5-deficient MEFs. Phosphorylation levels of two known ATM substrates, Rad17 and Chk2, were significantly reduced in Ppp5-deficient MEFs in response to IR. Furthermore, DNA damage-induced Rad17 nuclear foci were dramatically reduced in Ppp5-deficient MEFs. These results demonstrate a direct regulatory linkage between Ppp5 and activation of the ATM-mediated G(2)/M DNA damage checkpoint pathway in vivo.     

DNA Damage-Induced Accumulation of Centrosomal Chk1 Contributes to its Checkpoint Function

The checkpoint kinase Chk1 is an established transducer of ATR- and ATM-dependent signalling in response to DNA damage. In addition to its nuclear localization, Chk1 localizes to interphase centrosomes and thereby negatively regulates entry into mitosis by preventing premature activation of cyclin B-Cdk1 during unperturbed cell cycles. Here, we demonstrate that DNA damage caused by ultraviolet irradiation or hydroxyurea treatment leads to centrosomal accumulation of endogenous Chk1 in normal human BJ fibroblasts and in ATR- or ATM-deficient fibroblasts. Chemical inhibition of ATR/ATM by caffeine led to enhanced centrosomal Chk1 deposition associated with nuclear Chk1 depletion. In contrast to normal or ATM-deficient fibroblasts, genetically ATR-deficient Seckel-fibroblasts showed detectable constitutive centrosomal accumulation of Chk1 even in the absence of exogenous insults. After DNA damage, the centrosomal fraction of Chk1 was found to be phosphorylated at ATR/ATM phosphorylation sites. Forced immobilization of kinase-inactive but not wild-type Chk1 to centrosomes resulted in a G2/M checkpoint defect. Finally, both DNA damage, and forced centrosomal expression of Chk1 in the absence of genotoxic treatments, induced centrosome amplification in a subset of cells, a phenomenon which could be suppressed by inhibition of ATM/ATR-mediated signaling. Taken together, our results suggest that accumulation of phosphorylated Chk1 at centrosomes constitutes an additional element in the DNA damage response. Centrosomal Chk1 induces G2/M cell cycle arrest and may evoke centrosome amplification, the latter possibly providing a backup mechanism for elimination of cells with impaired DNA damage checkpoints operating earlier during the cell cycle.     

The stress kinase MRK contributes to regulation of DNA damage checkpoints through a p38gamma-independent pathway

DNA damage induced by ionizing radiation (IR) activates a complex cellular response that includes checkpoints leading to cell cycle arrest. The stress-activated mitogen-activated protein kinase (MAPK) p38gamma has been implicated in the G(2) phase checkpoint induced by IR. We recently discovered MRK as a member of the MAPK kinase kinase family that activates p38gamma. Here we investigated the role of MRK in the checkpoint response to IR. We identified autophosphorylation sites on MRK that are important for its kinase activity. A phosphospecific antibody that recognizes these sites showed that MRK is activated upon IR in a rapid and sustained manner. MRK depletion by RNA interference resulted in defective S and G(2) checkpoints induced by IR that were accompanied by reduced Chk2 phosphorylation and delayed Cdc25A degradation. We also showed that Chk2 is a substrate for MRK in vitro and is phosphorylated at Thr(68) by active MRK in cells. MRK depletion also increased sensitivity to the killing effects of IR. In addition, MRK depletion reduced IR-induced activation of p38gamma but had no effect on p38alpha activation, indicating that MRK is a specific activator of p38gamma after IR. Inhibition of p38gamma by RNA interference, however, did not impair IR-induced checkpoints. Thus, in response to IR MRK controls two independent pathways: the Chk2-Cdc25A pathway leading to cell cycle arrest and the p38gamma MAPK pathway.     

Cdc2 phosphorylation of Crb2 is required for reestablishing cell cycle progression after the damage checkpoint.

DNA damage induces cell cycle arrest (called the damage checkpoint), during which cells carry out actions for repair. A fission yeast protein, Crb2/Rhp9, which resembles budding yeast Rad9p and human BRCA1, promotes checkpoint by activating Chk1 kinase, which restrains Cdc2 activation. We show here that phosphorylation of the T215 Cdc2 site of Crb2 is required for reentering the cell cycle after the damage-induced checkpoint arrest. If this site is nonphosphorylatable, irradiated cells remain arrested, though damage is repaired, and maintain the phosphorylated state of Chk1 kinase. The T215 site is in vitro phosphorylated by purified Cdc2 kinase. Phosphorylation of T215 occurs intensely in response to DNA damage at a late stage, suggesting an antagonistic role of Cdc2 phosphorylation toward checkpoint.     

ATR pathway is the primary pathway for activating G2/M checkpoint induction after re-replication

DNA replication is tightly controlled to ensure accurate chromosome duplication and segregation in each cell cycle. Inactivation of Geminin, an inhibitor of origin licensing, leads to re-replication in human tumor cells within the same cell cycle and triggers a G(2)/M checkpoint. We find that the primary pathway to signal that re-replication has been detected is the ATR kinase and the Rad9-Rad1-Hus1 (9-1-1) clamp complex together with Rad17-RFC clamp loader. ATM kinase and the Mre11-Rad50-Nbs1 complex do not appear to play significant roles in the checkpoint. Chk1 activation occurs at early stages, whereas Chk2 activation occurs much later. Overall we conclude that ATR/Chk1 pathway is activated at an early time point after the loss of Geminin and contributes to checkpoint arrest essential for the accumulation of re-replicated cells, whereas activation of the ATM/Chk2 pathway is a by-product of DNA re-replication at a later period.     

Rad53 is essential for a mitochondrial DNA inheritance checkpoint regulating G1 to S progression

The Chk2-mediated deoxyribonucleic acid (DNA) damage checkpoint pathway is important for mitochondrial DNA (mtDNA) maintenance. We show in this paper that mtDNA itself affects cell cycle progression. Saccharomyces cerevisiae rho(0) cells, which lack mtDNA, were defective in G1- to S-phase progression. Deletion of subunit Va of cytochrome c oxidase, inhibition of F(1)F(0) adenosine triphosphatase, or replacement of all mtDNA-encoded genes with noncoding DNA did not affect G1- to S-phase progression. Thus, the cell cycle progression defect in rho(0) cells is caused by loss of DNA within mitochondria and not loss of respiratory activity or mtDNA-encoded genes. Rad53p, the yeast Chk2 homologue, was required for inhibition of G1- to S-phase progression in rho(0) cells. Pif1p, a DNA helicase and Rad53p target, underwent Rad53p-dependent phosphorylation in rho(0) cells. Thus, loss of mtDNA activated an established checkpoint kinase that inhibited G1- to S-phase progression. These findings support the existence of a Rad53p-regulated checkpoint that regulates G1- to S-phase progression in response to loss of mtDNA.     

Artemis links ATM to G2/M checkpoint recovery via regulation of Cdk1-cyclin B

Artemis is a phospho-protein that has been shown to have roles in V(D)J recombination, nonhomologous end-joining of double-strand breaks, and regulation of the DNA damage-induced G(2)/M cell cycle checkpoint. Here, we have identified four sites in Artemis that are phosphorylated in response to ionizing radiation (IR) and show that ATM is the major kinase responsible for these modifications. Two of the sites, S534 and S538, show rapid phosphorylation and dephosphorylation, and the other two sites, S516 and S645, exhibit rapid and prolonged phosphorylation. Mutation of both of these latter two residues results in defective recovery from the G(2)/M cell cycle checkpoint. This defective recovery is due to promotion by mutant Artemis of an enhanced interaction between unphosphorylated cyclin B and Cdk1, which in turn promotes inhibitory phosphorylation of Cdk1 by the Wee1 kinase. In addition, we show that mutant Artemis prevents Cdk1-cyclin B activation by causing its retention in the centrosome and inhibition of its nuclear import during prophase. These findings show that ATM regulates G(2)/M checkpoint recovery through inhibitory phosphorylations of Artemis that occur soon after DNA damage, thus setting a molecular switch that, hours later upon completion of DNA repair, allows activation of the Cdk1-cyclin B complex. These findings thus establish a novel function of Artemis as a regulator of the cell cycle in response to DNA damage.     

A Stronger DNA Damage-Induced G2 Checkpoint Due to Over-Activated CHK1 in the Absence of PARP-1

Poly(ADP-ribose) polymerase-1 (PARP-1) is involved in multi-pathways to respond to DNA damage. Lack of or inhibition of PARP-1 activity leads to slow progress of cell cycle and sensitization of cells to different stresses. Recently, it was reported that besides the Ku- dependent main non-homologous end joining (NHEJ) pathway, there is a PARP-1-dependent complementary NHEJ pathway to repair DNA double strand break (DSB). Here we show that compared with PARP-1+/+ cells, PARP-1-/- cells display a much stronger G2 checkpoint response following ionizing radiation (IR). Treatment with Chk1 siRNA abolishes the stronger G2 checkpoint response and sensitizes PARP-1-/- cells to IR. These data indicate that the stronger G2 checkpoint response in PARP-1-/- cells is CHK1-dependent, which protects cells from IR-induced killing. We also show that 4-Amino-1,8-naphthalimide (4-AN, inhibitor of PARP) but not methoxyamine (inhibitor of base excision repair (BER)), affects IR-induced G2 arrest and cell sensitivity in PARP-1+/+ cells, resulting in the phenotypes similar to those of PARP-1-/- cells. These results indicate that DSB repair from the complementary NHEJ pathway of PARP-1, but not single strand break (SSB) repair from the BER function of PARP-1, may play an essential role in the over-activated CHK1 regulated G2 checkpoint response and radiosensitivity in PARP-1-/- cells.