Activation of PARP-1 in response to bleomycin depends on the

Transcription

1 ORIGINAL ARTICLE Activation of PARP-1 in response to bleomycin depends on the Ku antigen and protein phosphatase 5 FDong 1, S Soubeyrand 1 and RJG Hache 1,2,3 (2010) 29, & 2010 Macmillan Publishers Limited All rights reserved /10 $ Ottawa Health Research Institute, Ottawa, Ontario, Canada; 2 Department of Medicine, University of Ottawa, Ottawa, Ontario, Canada and 3 Department of Biochemistry, Microbiology, and Immunology, University of Ottawa, Ottawa, Ontario, Canada Poly (ADP-ribose) polymerase-1 (PARP-1) has an important role in the cellular response to a broad spectrum of DNA lesions. PARP-1 is strongly activated in response to double-stranded DNA breaks (DSBs), yet its contribution to the DSB response is poorly understood. Here we used bleomycin, a radiomimetic that generates DSBs with high specificity to focus on the response of PARP-1 to DSBs. We report that the induction of PARP-1 activity by bleomycin depends on the Ku antigen, a nonhomologous-dna-end-joining factor and protein phosphatase 5 (PP5). PARP-1 activation in response to bleomycin was reduced over 10-fold in Ku-deficient cells, whereas its activation in response to U.V. was unaffected. PARP-1 activation was rescued by reexpression of Ku, but was refractory to manipulation of DNA-dependent protein kinase or ATM. Similarly, PARP-1 activation subsequent to bleomycin was reduced 2-fold on ablation of PP5 and was increased 5-fold when PP5 was overexpressed. PP5 seemed to act directly on PARP-1, as its basal phosphorylation was reduced on overexpression of PP5, and PP5 dephosphorylated PARP-1 in vitro. These results highlight the functional importance of Ku antigen and PP5 for PARP-1 activity subsequent to DSBs. (2010) 29, ; doi: /onc ; published online 25 January 2010 Keywords: Ku antigen; PARP-1; protein phosphatase 5; DNA repair; bleomycin Introduction Poly (ADP-ribose) polymerase-1 (PARP-1) is a pleiotropic molecular sensor of DNA damage that has an important role in maintaining genomic integrity, with PARP-1 deficient cells exhibiting a heightened sensitivity to a broad variety of genotoxic agents (Meyer-Ficca et al., 2005). PARP-1, a 113-kDa phosphoprotein, is comprised of an N-terminal zinc-fingered DNA binding domain separated from a C-terminal catalytic domain Correspondence: Dr RJG Haché, Department of Biological Sciences, University of Calgary, 2500 University Drive, Calgary, Alberta, Canada T2N 1N4. rjhache@ucalgary.ca Received 19 May 2009; revised 23 November 2009; accepted 7 December 2009; published online 25 January 2010 by a central region containing an automodification domain and a BRCT domain that mediates heteromolecular interactions (Ame et al., 2004). The binding of PARP-1 to damaged DNA activates its poly (ADPribose) (PAR) polymerase activity that adds extended branched PAR chains onto target proteins. As automodification of PARP-1 accumulates, the affinity of PARP-1 for DNA decreases, eventually leading to displacement of the PARP-1 from the DNA and inactivation of its enzymatic activity (Ame et al., 2004). A poly (ADP-ribose) glycohydrolase (PARG) reverses the PAR modification of proteins through cleavage of the PAR chains (Meyer-Ficca et al., 2005). PARP-1 is a phosphoprotein and several studies have suggested that modulation of PARP-1 phosphorylation can regulate its activity (Tanaka et al., 1987; Bauer et al., 1992; Ju et al., 2004; Kauppinen et al., 2006; Midorikawa et al., 2006). For example, PARP-1 can be phosphorylated by protein kinase C, and phosphorylation of PARP-1 and PARP-1 interacting factors have been shown to regulate PARP-1 activity (Tanaka et al., 1987; Bauer et al., 1992). Recently, it has been proposed that PARP-1 can be activated in the absence of DNA damage through association with the active form of the mitogenactivated kinase ERK2 (Cohen-Armon et al., 2007). PARP-1 has important regulatory roles in base excision repair and single-stranded break repair, and is also activated during nucleotide excision repair (Huber et al., 2004). PARP-1 mediated ADP ribosylation promotes a relaxation of chromatin structure over regions of DNA damage (Poirier et al., 1982) and directly promotes the recruitment and function of DNA repair factors, such as the architectural base excision repair/single-stranded break repair factors, XRCC1 and DNA ligase III to sites of DNA damage (Masson et al., 1998; Galande and Kohwi-Shigematsu, 1999; Leppard et al., 2003; Lis and Kraus, 2006). Hyperactivation of PARP-1 in response to elevated levels of DNA damage results in energy depletion and a particular form of apoptosis triggered by NAD þ deficiency (Bouchard et al., 2003; Andrabi et al., 2006). Evidence for the involvement of PARP-1 in the repair of double-stranded DNA breaks (DSBs) has also recently emerged, pointing to roles in both homologous recombination and nonhomologous-dna-end joining (NHEJ). Genetic studies have established a role for PARP-1 in HR (Hochegger et al., 2006; Idogawa et al., 2007), whereas biochemical and genetic studies have

2 2094 linked PARP-1 to NHEJ (Schultz et al., 2003; Perrault et al., 2004; Audebert et al., 2006; Wang et al., 2006). PARP-1 has been shown to interact with the NHEJ proteins, Ku antigen and DNA-dependent protein kinase (DNA-PK) with reciprocal modifications and mutual alterations in activity suggested (Ruscetti et al., 1998; Ariumi et al., 1999; Galande and Kohwi- Shigematsu, 1999). Another study indicated that PARP-1 may specifically facilitate HR by decreasing HR-inhibitory effects of Ku antigen (Hochegger et al., 2006). This correlated with a report that the interaction of Ku with PARP-1 modulated the affinity of Ku antigen for DNA ends (Li et al., 2004). However, the exact interplay between PARP-1 and Ku remains to be established, as another study reported that assembly of the HR machinery at DSBs occurs independently from the NHEJ machinery and that Ku binding to DSBs both precedes the recruitment of PARP-1 and occurs with higher affinity (Kim et al., 2005). Agents that generate DSBs generally damage DNA in a complex manner, with DSBs comprising only a small component of the resultant DNA damage. Thus, approximately only 0.5% of the DNA damage caused by ionizing radiation, the classical agent used for DSB studies, are actually DSBs, with damage being mostly a combination of single-stranded DNA breaks (SSBs) and DNA base damage (Ward, 1988). Moreover, the structures of the DSBs generated by ionizing radiation are varied and complex (Ward, 1988). By contrast, radiomimetic glycopeptides of the bleomycin family are free radical-based DNA damaging agents that produce much higher levels of DSBs with relatively uniform and simple DNA ends (Povirk, 1996). For example, about 10% the DNA damage generated by bleomcyin are DSBs with 5 0 phosphate and 3 0 phosphoglycolate termini with either blunt DNA ends or DNA ends offset by one basepair 5 0 overhang (Povirk et al., 1989). Here, we have determined that PARP-1 is strongly activated in response to bleomycin and that this activation is dependent on Ku antigen and is enhanced by protein phosphatase 5 (PP5). Bleomycin treatment induced activation of PARP-1 in a manner that was strongly dependent on Ku and resulted in a dramatic reduction of PARP-1 response in cells lacking Ku. Similarly, silencing of PP5 reduced the activation of PARP-1, whereas overexpression of PP5 enhanced the PARP-1 response and increased its isoelectric point. By contrast, the response of PARP-1 to UV was unaffected by the absence of Ku. These results highlight a dependence of PARP-1 on Ku in the cellular response to this DSB-inducing agent and provide a first indication of a positive role of specific dephosphorylation in the activation of PARP-1. Results Bleomycin is a potent inducer of PARP-1 poly-adp-ribose activity Treatment of HeLa cells with bleomycin (30 mg/ml, 1 h) led to the rapid induction of DNA damage as reflected by the prominent accumulation of phosphorylated H2AX (gh2ax) in western blots of cell extracts (Figure 1a). Indirect immunofluorescence analysis of gh2ax foci suggested the presence of damage sites per cell (data not shown). gh2ax induction was accompanied by a strong increase in the overall level of protein poly-adp ribosylation (PAR) in the cells as visualized by a PAR antibody in western blots (Figure 1a). This was accompanied by a shift in the migration of PARP-1 through SDS PAGE that is indicative of the auto-poly-adp ribosylation that accompanies PARP-1 activation. Indirect immunofluorescence analysis of whole cells confirmed that bleomycin treatment led to a strong induction of PAR and showed that the PAR accumulated primarily in the nucleus (Figure 1b). PARP-1 activity can also be detected in cell extracts incubated in the presence of [ 32 P] NAD, where it is primarily reflected by the automodification of PARP-1 (Ogata et al., 1981). Nonetheless, when we compared 32 P incorporation from [ 32 P] NAD into proteins in HeLa cell extracts, we observed a dramatic bleomycin-dependent induction of 32 P incorporation, with the spontaneous activity of PARP-1 in extracts from mock-treated cells not visible on same exposures (Figure 1c). This increase in PARP-1 activity could not be ascribed to a difference in cellular DNA released into the extracts, which we measured to be in the pg range and did not differ quantitatively or qualitatively between extracts from control and bleomycin-treated cells (Supplementary Figure 1, data not shown). Predictably, the 32 P signal coimmunopreciptiated with PARP-1 and exhibited an upwards smearing in the PARP-1 signal on western blots that is typical of PARP-1 automodification (Figure 1d). A comparison of [ 32 P] NAD incorporation in wild type and PARP-1 / murine embryonic fibroblasts (MEFs) (Figure 1e), verified that the 32 P incorporation was PARP-1-dependent. PARP-1 activation in response to bleomycin was also observed in several other cell types and occurred rapidly, with maximal activity being detectable within 15 min (data not shown). The activation of PARP-1 activity as reported by [ 32 P] NAD automodification of PARP-1 in vitro after bleomycin treatment was very strong, exceeding the level of PARP-1 activation after a 10 min treatment with 0.5 mm H 2 O 2 (Figure 2a), a DNA damaging agent that elicits a strong activation of PARP-1 through the generation of single-stranded DNA breaks (SSBs) and oxidative damage (Cantoni et al., 1986). g-irradiation, which introduces predominantly single-stranded DNA breaks that become processed over time to DSBs, seemed to be less effective than bleomycin in inducing PARP-1 activity, even when administered at very high doses (Figure 2b) that induce thousands of DNA breaks per cell (Cole et al., 1975). PARP-1 activity in response to bleomycin also exceeded that induced by a dose of 0.25 J/m 2 of UV radiation (Figure 2c). Activation of PARP-1 in response to bleomycin depends on Ku antigen PARP-1 has previously been suggested to have a role in regulating the response of Ku antigen to DSBs

3 2095 Figure 2 Bleomycin induction of poly (ADP-ribose) polymerase-1 (PARP-1) activity exceeds the response to H 2 O 2,, ionizing radiation and U.V. (a) Comparison of 32 P NAD incorporation into PARP-1 in HeLa cell extracts after bleomycin (Bleo, 30 mg/ml, 1 h) and H 2 O 2 (0.5 mm, 10 min) treatment. (Ctl, vehicle-treated cells) Top panel shows 32 P incorporation, lower panel a western blot of PARP-1. (b) Comparison PARP-1 activation subsequent to bleomycin treatment and 30 min after g-irradiation (8 100 Gy) as in (a). (c) Comparison of activation of PARP-1 automodification by bleomycin and 1 h after UV irradiation (25 J/m 2 )asin(a). Figure 1 Bleomycin induces poly (ADP-ribose) polymerase-1 (PARP-1) auto-adp-ribosylation activity and accumulation of PAR. (a) Western analysis of HeLa cell extracts for gh2ax, poly ADP-ribose (PAR), PARP-1 and nucleolin subsequent to 1 h treatment with bleomycin (Bleo, 30 mg/ml) or vehicle alone. (b) Indirect immunofluorescence detection of PAR in mock- and bleomycin-treated HeLa cells compared with DAPI nuclear staining of the same cells. (c) Comigration of 32 P NAD labeling with PARP-1 in cell extracts separated by SDS PAGE (8%) after bleomycin treatment. Left, 32 P signal; right, PARP-1 western blot. (d) 32 P NAD labeling of cell extracts after bleomycin treatment and immunoprecipitation with PARP-1. SDS PAGE of PARP-1 immunoprecipitate, control immunoprecipitate with a nonspecific antibody (Ctrl Ab), compared with the signal in 10% of the original extract±bleomycin treatment. 32 P NAD incorporation is at the top, PARP-1 western below. (e) 32 P NAD PARP-1 labeling is lost in PARP-1 / cells. Comparison of extracts prepared from 3T3 fibroblasts derived from WT and PARP / (KO) mice for 32 P NAD incorporation and the presence of PARP-1 as in (c). (Hochegger et al., 2006; Wang et al., 2006). To determine whether Ku might also be important for the activation of PARP-1 subsequent to the DSBs induced by bleomycin, we compared the activity of PARP-1 after treatment in extracts prepared from V79 cells, a hamster embryonic lung fibroblast cell line and V15B cells, a line clonally derived from V79 containing a mutation in the Ku80 gene that leads to the absence of Ku70 and Ku80 (Errami et al., 1996). Typical strong activation of PARP-1 activity was observed in extracts prepared from V79 cells and the response included a shifting and spreading of the PARP-1 band on immunoblots (Figure 3a). Strikingly, although V15B cells contained the same level of PARP-1 as V79 cells, bleomycin treatment failed to activate PARP-1 activity, with 32 P NAD incorporation at the same low level as in extracts prepared from vehicle-treated V79 cells that reflects the relatively low level of spontaneous activation of PARP-1 that occurs on cell lysis. Further, PARP-1 activation was rescued in the V15B cells on reexpression of Ku80 from a viral vector, which reconstitutes hamster/human Ku70/Ku80 (Schild-Poulter et al., 2003), reinforcing the dependence of PARP-1 activation on Ku rather than other changes that might have arisen in the cells. By contrast, UV treatment of V15B cells resulted in an approximately 2-fold enhancement in PARP-1 activation (Figure 3b), highlighting the independence of PARP-1 activation on UV DNA damage from Ku. Similar results were obtained with MEFs lacking Ku through abrogation of Ku70 expression (Figure 3c). Here, Ku70 / MEFs failed to accumulate significant

4 2096 Figure 3 Bleomcyin induction of poly (ADP-ribose) polymerase-1 (PARP-1) is dependent on Ku antigen. (a) Comparison of 32 P NAD incorporation into PARP-1 and PARP-1 electrophoretic mobility on SDS PAGE in V79 cells (Chinese hamster ovary cells) expressing WT Ku (V79), clone V15B lacking Ku80 (V15B) and V15B cells supplemented with a cdna expressing human Ku80 (V15B/Ku80) that reconstitutes Ku levels and function (Giffin & Haché, 1995; Schild-Poulter et al., 2003). Bleomycin (bleo) treatment (1 h, 30 mg/ml) was as indicated. 32 P NAD incorporation into PARP-1 immediately after the 1 h bleo treatment was quantified relative to bleomycin-treated V79 cells, with error representing standard deviation over three independent experiments performed in duplicate. (b) PARP-1 automodification as in (a) subsequent to bleomycin and UV irradiation (25 J/m 2 ) of V79 and V15B cells. Quantification (±s.d.) represents the average of duplicate determinations, normalized to the value of bleomycintreated samples, and is representative of three independent experiments. (c) Indirect immunofluorescence detection of PAR in mock- and bleomycin-treated (1 h, 30 mg/ml) murine embryonic fibroblasts (MEFs) lacking Ku through deletion of Ku70 / (Ku70 / MEFs) and MEFs reconstituted to express WT Ku through stable reexpression of a human Ku70 cdna (Ku70 / MEF þ hku70) (Crawley et al.), compared with DAPI nuclear staining of the same cells. PAR in response to bleomycin treatment, whereas reconstitution of Ku through reexpression of Ku70 resulted in a prominent induction of nuclear PAR subsequent to bleomycin as detected by indirect immunefluorescence analysis of the cells. Together, these results highlight that loss of Ku70 or Ku80 separately, which led to similar Ku phenotypes, each abrogated the induction of PARP-1 activity in response to bleomycin. We next investigated whether the Ku-dependent regulatory control point for PARP-1 activation in response to bleomycin included dependence on factors upstream or downstream from Ku in the NHEJ response. In the first instance, Ku recognition of DSBs leads to the recruitment and activation of DNA-PK at the break (Gottlieb and Jackson, 1993). Inhibition of Figure 4 Bleomcyin induction of poly (ADP-ribose) polymerase-1 (PARP-1) activity is independent of DNA-dependent protein kinase (DNA-PK) and ATM. (a) PARP-1 automodification in HeLa cell extracts subsequent to bleomycin treatment in the presence and absence of DNA-PK inhibitor Nu7026 (10 mm), with Nu7026 treatment initiated 2 h before bleomycin treatment and continued through the incubation of the extract with 32 P NAD. (b) Western analysis of PARP-1 mobility on SDS PAGE subsequent to bleomycin treatment of DNA-PKcs positive MO59J/Fus1 cells and DNA-PKcs-deficient MO59J/Fus9 cells. (c) PARP-1 automodification in extracts prepared from bleomycintreated ATM-deficient AT-fibroblasts and AT-fibroblasts (AT) with ATM restored (AT þ ATM). (d) Western analysis of PARP-1 mobility on SDS PAGE subsequent to bleomycin treatment of HeLa cells in the presence or absence of 10 mm (2 h pre-treatment) of the global large-pi3k-like kinase inhibitor wortmannin. DNA-PK with the specific inhibitor Nu7026 had no visible effect on PARP-1 activation subsequent to bleomycin treatment (Figure 4a). Similarly, the shift in PARP-1 mobility in SDS PAGE that marks its automodification was the same in MO59J/Fus1 cells that express DNA-PK and MO59J/Fus9 cells that are DNA- PK deficient (Figure 4b). Thus DNA-PK signaling

5 downstream from Ku had no discernable role in PARP- 1 activation in response to bleomycin. Second, ATM is a kinase that acts upstream of Ku to induce the phosphorylation of H2AX at DSBs and to initiate the DSB response signaling cascade (Kinner et al., 2008). PARP-1 activation in response to bleomycin, was also unaffected by ATM status (Figure 4c). Both ATMdeficient (AT) cells and cells with WT ATM restored exhibited the same activation of PARP-1 in response to bleomycin. Finally, inhibition of ATM and DNA-PK at the same time by Wortmanin also failed to alter the activation of PARP-1 after the bleomycin treatment (Figure 4d). These results suggested a specific dependence of PARP-1 on Ku for its activation at DSBs Enhanced interaction between PARP-1 and Ku in response to bleomycin Both Ku and PARP-1 are recruited to DNA ends and previous reports have suggested that Ku can interact physically with PARP-1, directly or in combination with other factors such as WRN (Galande and Kohwi- Shigematsu, 1999; Li et al., 2004). Attempts to coimmunoprecipitate PARP-1 and Ku from HeLa cell extracts prepared from control cells were unsuccessful under our stringent conditions (Figure 5a). However, after bleomycin treatment, antibodies to the Ku heterodimer and PARP-1 reciprocally coprecipitated the other factor. Further, the PARP-1 that was coimmunoprecipitated by Ku showed strong PARP activity (Figure 5b), supporting the activation of PARP-1 being linked to its interaction with Ku. Extensive PARP-1 automodification did not seem to be required for Ku binding, as addition of the PARP-1 enzymatic inhibitor 4ANI, which effectively blocked the shift in PARP-1 mobility on SDS PAGE (Figure 5c) and inhibited 32 P NAD incorporation by over 95% (data not shown), enhanced the coimmunoprecipitation of PARP-1 with Ku, suggesting a preferential interaction between Ku and PARP-1 before its auto-adp ribosylation, but subsequent to DNA damage. Finally, addition of ethidium bromide to the extracts, which disrupts protein-dna interactions (Lai and Herr, 1992), prevented coimmunoprecipitation of PARP-1 and Ku (Figure 5d). Although this result indicates that DNA contact by PARP-1 and/or Ku is important to detect their physical association, it leaves open whether the two factors are preferentially recruited to the same DNA ends or whether the DNA binding of one or both factors is required for their physical interaction. Activation of PARP-1 after bleomycin treatment depends on PP5 The interaction of PARP-1 with Ku was a specific response to bleomycin suggested that signaling events acting on PARP-1 preceded its activation in response to bleomycin. It has been suggested previously that phosphorylation may regulate PARP-1 activity (Kauppinen et al., 2006; Midorikawa et al., 2006). As it seemed that neither ATM-initiated signaling nor DNA-PK influenced activation of PARP-1 in response Figure 5 Bleomycin induces strong, DNA-dependent, association between poly (ADP-ribose) polymerase-1 (PARP-1) and Ku. (a) HeLa cell lysates prepared from cells treated with bleomycin ( þ Bleo, 30 mg/ml) or vehicle ( Bleo) for 1 h were immunoprecipitated with antibodies against the Ku heterodimer (left) or PARP- 1 (right). After SDS PAGE separation, the immunoprecipitates were compared with 10% of the original extract (input) for the presence of PARP-1 and Ku70, as indicated to the right of the panel. Ctl indicates samples incubated with protein-g sepharose alone (b) 32 P NAD labeling of PARP1-1 after immunoprecipitation from cell extracts with Ku antibody (Ku) or protein-g sepharose alone (Ctl) compared with labeling in 10% of the extracts used for immunoprecipitation (top), with western blots of Ku (middle) and PARP-1 (bottom) also shown. (c) Immunoprecipitation of PARP-1 by Ku subsequent to bleomycin treatment is enhanced by inhibition of PARP-1 activity. Analysis of PARP-1 coimmunoprecipitation with Ku as in (a) with the addition of 4ANI (30 mm) beginning 1 h before bleo treatment. The panels are assembled from one exposure of a single membrane, with intervening lanes deleted. (d) Addition of 0.2 mg/ml ethidium bromide disrupts the coimmunoprecipitation of Ku with PARP-1 subsequent to bleomycin treatment. Note for (a) and (c), exposures of the left and right hand panels have been varied to allow for side by side display, with exposure of the inputs being shortened.

6 2098 to bleomycin, we sought to determine whether a protein phosphatase activity might be important for PARP-1 activation. In the first test, we examined the effect of the broad protein phosphatase inhibitor okadaic acid (OA) on the response of PARP-1 to bleomycin treatment of HeLa cells (Figure 6a). Preincubation of the cells with 250 nm OA reduced the subsequent activation of PARP-1 in response to bleomycin to 27±5% of cells incubated in parallel without OA, suggesting an important role for a protein phophatase in the PARP-1 response. Further, this result suggested PP1, PP2A, PP4 and PP5 as potential effectors, but excluded PP2B and PP7, which are insensitive to OA at this concentration. In an effort to reduce the number of candidate phosphatases, we pretreated the cells with two additional inhibitors, fostriecin and tautomycetin (Figure 6b). Neither 10mM fostriecin, which effectively inhibits PP2A and PP4, nor 0.3 mm tautomycetin, which inhibits PP1, had a significant effect on the response of PARP-1 to bleomycin. Thus, by the process of elimination, these results identified PP5 as the prime candidate to modulate the response of PARP-1 to bleomycin. In the first test of the importance of PP5 for the activation of PARP-1, we examined the effect of sirna-mediated ablation of PP5 on the response of PARP-1 to bleomycin in HCT 116 cells (Figure 7a). Transient transfection of a Smart pool of anti PP5 RNA sirna reduced PP5 levels by approximately 80%, and this led to a 40% reduction in 32 P NAD incorporation into PARP-1. By contrast, a nonspecific sirna had no effect on PP5 levels. Conversely, increasing cellular PP5 levels through overexpression of PP5 with a Flag tag (fpp5) resulted in a 5-fold increase in PARP-1 activity subsequent to bleomycin treatment without affecting PARP-1 activity in control cells (Figure 7b). Together, these data provided strong evidence for a role for PP5 in the activation of PARP-1 poly-adp-ribose polymerase activity in response to bleomycin. To test whether PP5 could overcome the dependence of PARP-1 on Ku in responding to bleomycin, we overexpressed fpp5 in V79/V15B cells (Figure 7c). In V79 cells, fpp5 overexpression enhanced the activation of PARP-1 in response to bleomycin 2-fold. In V15B cells, the fold effect of fpp5 expression on the much lower level of PARP-1 activity seemed to be similar, but fpp5 expression was clearly unable to compensate for the absence of Ku. Although these results provide strong evidence that PP5 cannot substitute for Ku in the activation of PARP-1, it is difficult to comment definitely on whether the effect of PP5 is dependent on Ku for two reasons. First, while the ratio of fpp5 to endogenous PP5 expression in V15B cells was comparable to that in HCT 116 cells, it was several folds lower than the level of expression in V79 cells. Second, the very low levels of PARP-1 activation in V15B cells precluded determination of whether a 2- fold enhancement in PARP-1 activation was significant. Finally, to explore whether PP5 was also implicated in the activation of PARP-1 in response to other DNA damaging agents, we tested the effect of PP5 Figure 6 Phosphatase inhibitor sensitivity suggests a role for protein phosphatase 5 (PP5) in the activation of poly (ADP-ribose) polymerase-1 (PARP-1) in response to bleomycin in HeLa cells. (a) Effect of 250 nm okadaic acid (OA) on 32 P NAD incorporation into PARP-1 in extracts from bleomycin ( þ ) and vehicle-treated ( ) cells. OA treatment was initiated 1 h before bleomycin. The upper panel shows SDS PAGE separation of 32 P incorporation, the lower panel a western of the same gel for PARP-1. OA sensitivity of the activation eliminates the kinases over scored with an X as candidates to participate in PARP-1 activation (Mitsuhashi et al., 2001; Honkanen and Golden, 2002; Golden and Honkanen, 2003). (b) Comparison of 32 P incorporation into PARP-1 as in (a), except that cells were incubated with 1.0 mm fostriecin or 0.3 mm tautomycetin as indicated. overexpression on the activation of PARP-1 in response to UV (Figure 7d). Expression of fpp5 in HCT116 cells enhanced the response of PARP-1 to UV 2.3-folds (P ¼ 0.02), suggesting that the stimulatory effect of PP5 will be broadly based and likely to be a precursor to the response rather than a consequence of DNA damage. Phosphorylation of PARP-1 is reduced by PP5 To determine the phosphorylation status of PARP-1 in naı ve cells, we incubated HeLa, HCT 116 and 293T cells with a pulse of 32 P orthophosphate. In all three cell lines, this resulted in the prominent incorportation of 32 P into PARP-1 (Figure 8a). Further, western analysis of PARP-1 immunoprecipitated from cell extracts from asynchronously growing 293T cells and separated on two-dimension gel revealed a number of PARP-1 isoforms (Figure 8b). Treatment of the extract with calf intestinal phosphatase before immunoprecipitation of PARP-1 resulted in a strong basic shift in PARP-1 on isoelectric focusing. These results establish that PARP-1 is basally phosphorylated on many sites.

7 2099 Figure 7 Poly (ADP-ribose) polymerase-1 (PARP-1) activation in reponse to bleomycin is dependent on PP5. (a) PP5 ablation reduces bleomycin-induced 32 P NAD incorporation into PARP-1 in HCT 116 cell extracts. Cells were harvested 72 h after transfection with protein phosphatase 5 (PP5) specific sirna ( þ ) or control sirna ( ) and were treated with bleomycin ( þ ) or vehicle ( ) for the last hour as indicated. After SDS PAGE, 32 P incorporation is compared with the levels of PARP-1, PP5 and nucleolin. Quantification of 32 P incorporation into PARP-1 is relative to bleomycin-treated controls and is corrected for nucleolin levels, with the values compiled from three independent experiments performed in duplicate (±s.d.). (b) PP5 overexpression enhances bleomycin-induced 32 P NAD incorporation into PARP-1 in HCT 116 cell extracts. Cells were harvested 72 h after transfection with a Flag-tagged PP5 (fpp5) expression vector or control vector and 1 h after 30 mg/ml bleomycin treatment. 32 P NAD incorporation into PARP-1 in cell extracts is compared with the levels of PARP-1, fpp5/endogenous PP5 and Ku80 determined by western blot. The mobility of fpp5 and endogenous PP5 are indicated by the arrowheads. (c) Effect of fpp5 overexpression on 32 P NAD incorporation into PARP-1 into V79 and V15B cell extracts. Experiment was performed and quantified as in (b). fpp5 expression in V15B cells was approximately 30% of the level of expression in V79 cells whereas endogenous PP5 levels were indistinguishable. (d) Effect of over expression of PP5 on activation of PARP-1 in response to UV performed in HCT 116 cells as in (b), except that cells were harvested 1 h after exposure to 25 J/m 2 of 32 UV. P NAD incorporation into PARP-1 is compared with western blot determination of the levels of expression of PARP-1 and PP5. To investigate whether PP5 directly affected the phosphorylation state of PARP-1 we performed a twodimensional separation of PARP-1 immunoprecipitated from 293T in which PP5 was overexpressed (Figure 9a). Overexpression of PP5 resulted in a basic shift in the major PARP-1 isoforms on isoelectric focusing that was Figure 8 Poly (ADP-ribose) polymerase-1 (PARP-1) is phosphorylated before exposure of the cell to DNA damaging agents. (a) HeLa, HCT116 and 293T cells were incubated with 32 P orthophosphate for 4 h. PARP-1 was immunoprecipitated from cellular extracts, subjected to SDS PAGE. 32 P incorporation [ 32 P] and western analysis of PARP-1 are compared with the levels of PARP-1 in the crude extracts. (b) Migration of PARP-1 in two dimensions is altered by calf intestinal phosphatase. Western analysis of two-dimensional separation (isoelectric focusing ph 3-10/SDS PAGE) of PARP-1 immunoprecipitated from 293T cell extracts and either mock treated (Ctl) or treated with calf intestinal phosphatase before separation. Arrows highlight the relative migration of the major forms of control and calf intestinal phosphatase-treated PARP-1. very similar to the shift observed on calf intestinal phosphatase treatment. Further, when we immunoprecipitated fpp5 and combined it with immunoprecipitated PARP-1 under conditions permissive for phosphatase activity, the isoelectric point of PARP-1 in isoelectric focusing again underwent a strong basic shift (Figure 9b). Thus these results show that PP5 induces the dephosphorylation of PARP-1 in vivo and that PP5 can directly dephosphorylate PARP-1 in vitro. Together, these results suggest basal phospohorylation of PARP-1 reduces its response to at least DSBs and that PP5 contributes to alleviating the effect of this phosphorylation. Discussion Activation of PARP-1 poly (ADP-ribose) polymerase activity is a broadly based cellular response to DNA damage that is important to several DNA repair processes. We report here that activation of PARP-1 in response to bleomycin-induced DNA damage is strongly enhanced by Ku antigen and by PP5-mediated dephosphorylation of PARP-1. Although the role of Ku seemed to be specific to DSB response, the role of PP5 in activating PARP-1 may be more broadly based. In the first instance, we determined that Ku antigen has a major and specific role in the activation of PARP-1

8 2100 Figure 9 Protein phosphatase 5(PP5) alters the isoelectric point of poly (ADP-ribose) polymerase-1 (PARP-1). (a) Western analysis of PARP-1 after two-dimensional separation of extracts prepared from 293T cells transfected with control vector (top) or fpp5 expressing vector (bottom). Verification that ERK and MEK phosphorylation in the extracts was unaffected by fpp5 was used as a control against potentially relaxed specificity of overexpressed fpp5 (data not shown). (b) Western analysis of PARP-1 separation in two-dimensions after incubation of PARP-1 immunoprecipitated from 293T cells with flag immunoprecipate from control cells (top) or fpp5 expressing cells (bottom). The right of the panels shows PARP-1 levels in 5% of the extracts used for the IP (input). subsequent to bleomycin treatment. The absence of Ku resulted in a strong decrease in PARP-1 activation and this activity could be reconstituted by reexpression of the Ku in Ku-deficient cells. However, the presence of Ku was not sufficient to promote the activation of PARP-1 in cell extracts. Rather, PARP-1 activation depended on previous treatment of the cells with bleomycin, which rendered the subsequent extract competent for Ku to enhance PARP-1 activity. Thus, our results suggest that Ku-dependent events are required for the activation of PARP-1 activity subsequent to the induction of DNA damage. The introduction of DSBs per cell by bleomycin is a relatively modest amount of DNA damage and it is not surprising that it did not have a discernable effect on the trace amounts of DNA released as a consequence of the mechanics of extract preparation. Similarly, it is also unlikely that the 5 0 phosphate and 3 0 phosphoglycolate termini generated by bleomycin would be substantially more efficient in attracting or activating PARP-1 than the bulk of the DNA ends in the extracts. Certainly, sheared or nuclease treated DNA are very potent inducers of PARP-1 activity in vitro (Okayama et al., 1977; Pion et al., 2005). Thus, our results indicate that a Ku-dependent signal, likely originating from the bleomycin-generated DSBs within the cell before extraction served either to directly modulate the enzymatic potential of PARP-1 or acted on Ku antigen in a manner that encouraged recruitment of PARP-1 to DNA damage sites from which it could be activated. The generation of DSBs initiates a cellular signaling response that activates cell cycle checkpoints and induces the accumulation of gh2ax and DNA repair factors at the sites of the DSBs. Most of these signaling events occur downstream from the activation of ATM, yet the events that promoted the activation of PARP-1 seemed to be ATM-independent. Similarly, PARP-1 activation was also independent of DNA-PK signaling at the DNA break sites. Thus, an important question for future experimentation will be to identify the nature of the signal that provides for the stimulation of PARP-1 activity by Ku. Analysis of protein phosphatase activities suggested a specific role for PP5 in promoting the activation of PARP-1. PP5 overexpression promoted the dephosphorylation of PARP-1, as visualized by a shift in PARP-1 isoelectric point and PP5 was able to directly dephosphorylate PARP-1 in vitro. Thus, it seems that dephosphorylation of PARP-1 by PP5 facilitates PARP- 1 activity or PARP-1 activation. The work of other groups and our preliminary analysis indicates that PARP-1 is phoshosphorylated on many sites (Kauppinen et al., 2006, data not shown). It will be interesting to identify the specific sites that are dephosphorylated by PP5 and to determine how they affect PARP-1 activity. PP5 is a ubiquitously expressed Ser/Thr protein phosphatase distributed between nucleus and cytoplasm (Chen et al., 1994). However, PP5 differs from other phosphatases in having low intrinsic activity and thus a presumed requirement for activation (Chen et al., 1994; Skinner et al., 1997). Earlier studies have shown a role for PP5 in the activation of cell cycle checkpoints in response to DSBs. Thus, MEFs from PP5 deficient mice exhibit a reduction in ATM activation subsequent to DSBs and an alteration in the G2/M checkpoint (Ali et al., 2004). In this context a role for PP5 in the activation of PARP-1 is aligned with a role for PP5 in the activation of ATM. By contrast, antisense and overexpression experiments also have provided evidence of a role for PP5 in suppressing p53 and 53BP1 activity through repression of their phosphorylation in response to DSBs (Zuo et al., 1998; Kang et al., 2009). Additional studies will be required to reconcile these seemingly contradictory roles of PP5 in the cellular response to DSBs. Although PP5 has been found to interact readily with ATM and DNA-PK (Ali et al., 2004; Wechsler et al., 2004), PARP-1 activation occurred independently from both kinases argues against a bridging model for the recruitment of PP5 to PARP-1 at DNA ends. We have been unable to show an interaction between Ku and PP5, despite extensive attempts at coimmunoprecipation. Indeed, our results indicating that PP5 also has the ability to facilitate the activation of PARP-1 in response to UV-initiated DNA damage suggests that PP5 may be involved in a broadly based regulation of PARP-1. The effect of Ku on PARP-1 activity was very potent, leading to a 10-fold activation of PARP. Indeed, in the absence of Ku, PARP-1 activation subsequent to bleomycin treatment was comparatively reduced to low level that was obtained as a result of the DNA present in the extracts from control cells. Kinetic

9 experiments have shown that Ku is recruited to sites of DNA damage in advance of PARP-1 and in vitro DNA binding experiments have determined the DNA binding affinity of Ku to be higher than that of PARP-1 (Ariumi et al., 1999; Kim et al., 2005). At the same time, it has been suggested that the presence of Ku at DNA ends can inhibit DNA binding of PARP-1. Many studies have linked PARP-1 and Ku either directly or as components of multiprotein complexes (Ruscetti et al., 1998; Ariumi et al., 1999; Galande and Kohwi-Shigematsu, 1999; Grote et al., 2006; Yin and Glass, 2006; Hossain et al., 2009). Although cell lines extraction procedures and coimmunoprecipitation protocols vary between studies, there have been consistent observations of Ku/PARP-1 interactions and a frequent observation that PARP activity or DNA-PK mediated phosphorylation modulates the complexes. Of note, these earlier experiments were performed with naı ve factors or extracts that were prepared from cells that had not been exposed to a DNA damaging agent such as the bleomycin. In addition, in at least one study (Ruscetti et al., 1998), Ku/PARP-1 binding was determined to be resistant to DNase I digestion, which would disrupt colocalization of two factors to the same DNA molecule or DNA-dependent bridging interactions between the two factors. However, unlike the ethidium bromide used in our study, protein-dna interactions can be resistant to DNAse I. Thus our results contribute two additional brush strokes to the still blurred picture of Ku/PARP-1 binding. First, we show that the interaction is enhanced subsequent to DNA damage. Second, our data suggest that the conformational change induced in Ku on DNA binding (Rivera-Calzada et al., 2007), or similar effect on PARP-1 may increase the affinity of the PARP-1/Ku interaction. In addition, Ku also has been shown to be posttranslationally modified in a regulated manner (Douglas et al., 2005; Postow et al., 2008). Thus, it is possible that the interaction between Ku and PARP-1 is enhanced by the post-translational modification of one or both factors that occurs subsequent to bleomycin. Clarification of what seems to be a complex interaction awaits a definitive molecular analysis of Ku/PARP-1 binding under consistent conditions. Finally, while we have shown in this study the effects of Ku on PARP-1 activity, the work of others have suggested that PARP-1 may also directly modify Ku activity (Li et al., 2004). In a follow-up study that is presently being prepared for publication, we show evidence that PARP-1 is important for NHEJ and have identified specific PAR modification sites on Ku through which the effects of PARP-1 on NHEJ are mediated (Crawley et al., in preparation). Thus the activities of PARP-1 and Ku are interrelated and mediate complementary roles in repairing DSBs. Materials and methods Reagents and procedures are described in the Supplementary Information Conflict of interest The authors declare no conflict of interest. Acknowledgements We are very grateful to Dr V Schreiber and Dr G de Murcia for providing the PARP-deficient and control mouse fibroblasts. We would also like to thank Dr Y Shiloh for providing us with the ATM-restored and AT fibroblast cell lines (PEBS- YZ5 and PEBS, respectively). Finally, we are grateful to L Pope for her logistical support. This work was funded by an operating grant from the Canadian Institutes of Health Research (CIHR MT 13412) to RJGH. RJGH was a University of Ottawa Health Research Chair References Ali A, Zhang J, Bao S, Liu I, Otterness D, Dean NM et al. (2004). Requirement of protein phosphatase 5 in DNA-damage-induced ATM activation. Genes Dev 18: Ame JC, Spenlehauer C, de Murcia G. (2004). The PARP superfamily. Bioessays 26: Andrabi SA, Kim NS, Yu SW, Wang H, Koh DW, Sasaki M et al. (2006). Poly(ADP-ribose) (PAR) polymer is a death signal. Proc Natl Acad Sci USA 103: Ariumi Y, Masutani M, Copeland TD, Mimori T, Sugimura T, Shimotohno K et al. (1999). Suppression of the poly(adp-ribose) polymerase activity by DNA-dependent protein kinase in vitro. 18: Audebert M, Salles B, Weinfeld M, Calsou P. (2006). Involvement of polynucleotide kinase in a poly(adp-ribose) polymerase-1-dependent DNA double-strand breaks rejoining pathway. J Mol Biol 356: Bauer PI, Farkas G, Buday L, Mikala G, Meszaros G, Kun E et al. (1992). Inhibition of DNA binding by the phosphorylation of poly ADP-ribose polymerase protein catalysed by protein kinase C. Biochem Biophys Res Commun 187: Bouchard VJ, Rouleau M, Poirier GG. (2003). PARP-1, a determinant of cell survival in response to DNA damage. Exp Hematol 31: Cantoni O, Murray D, Meyn RE. (1986). Effect of 3-aminobenzamide on DNA strand-break rejoining and cytotoxicity in CHO cells treated with hydrogen peroxide. Biochim Biophys Acta 867: Chen MX, McPartlin AE, Brown L, Chen YH, Barker HM, Cohen PT. (1994). A novel human protein serine/ threonine phosphatase, which possesses four tetratricopeptide repeat motifs and localizes to the nucleus. EMBO J 13: Cohen-Armon M, Visochek L, Rozensal D, Kalal A, Geistrikh I, Klein R et al. (2007). DNA-independent PARP-1 activation by phosphorylated ERK2 increases Elk1 activity: a link to histone acetylation. Mol Cell 25:

10 2102 Cole A, Shonka F, Corry P, Cooper WG. (1975). CHO cell repair of single-strand and double-strand DNA breaks induced by gammaand alpha-radiations. Basic Life Sci 5B: Douglas P, Gupta S, Morrice N, Meek K, Lees-Miller SP. (2005). DNA-PK-dependent phosphorylation of Ku70/80 is not required for non-homologous end joining. DNA Repair (Amst) 4: Errami A, Smider V, Rathmell WK, He DM, Hendrickson EA, Zdzienicka MZ et al. (1996). Ku86 defines the genetic defect and restores X-ray resistance and V(D)J recombination to complementation group 5 hamster cell mutants. Mol Cell Biol 16: Galande S, Kohwi-Shigematsu T. (1999). Poly(ADP-ribose) polymerase and Ku autoantigen form a complex and synergistically bind to matrix attachment sequences. J Biol Chem 274: Giffin W, Haché RJ. (1995). Nuclear factor binding to a DNA sequence element that represses MMTV transcription induces a structural transition and leads to the contact of single-stranded binding proteins with DNA. DNA Cell Biol 14: Golden TA, Honkanen RE. (2003). Regulating the expression of protein phosphatase type 5. Methods Enzymol 366: Gottlieb TM, Jackson SP. (1993). The DNA-dependent protein kinase: requirement for DNA ends and association with Ku antigen. Cell 72: Grote J, Koenig S, Ackermann D, Sopalla C, Benedyk M, Los M et al. (2006). Identification of poly(adp-ribose)polymerase-1 and Ku70/ Ku80 as transcriptional regulators of S100A9 gene expression. BMC Mol Biol 7: 48. Hochegger H, Dejsuphong D, Fukushima T, Morrison C, Sonoda E, Schreiber V et al. (2006). Parp-1 protects homologous recombination from interference by Ku and Ligase IV in vertebrate cells. EMBO J 25: Honkanen RE, Golden T. (2002). Regulators of serine/threonine protein phosphatases at the dawn of a clinical era? Curr Med Chem 9: Hossain MB, Ji P, Anish R, Jacobson RH, Takada S. (2009). Poly(ADP-ribose) polymerase 1 interacts with nuclear respiratory factor 1 (NRF-1) and plays a role in NRF-1 transcriptional regulation. J Biol Chem 284: Huber A, Bai P, de Murcia JM, de Murcia G. (2004). PARP-1, PARP- 2 and ATM in the DNA damage response: functional synergy in mouse development. DNA Repair (Amst) 3: Idogawa M, Masutani M, Shitashige M, Honda K, Tokino T, Shinomura Y et al. (2007). Ku70 and poly(adp-ribose) polymerase- 1 competitively regulate beta-catenin and T-cell factor-4-mediated gene transactivation: possible linkage of DNA damage recognition and Wnt signaling. Cancer Res 67: Ju BG, Solum D, Song EJ, Lee KJ, Rose DW, Glass CK et al. (2004). Activating the PARP-1 sensor component of the groucho/tle1 corepressor complex mediates a CaMKinase IIdelta-dependent neurogenic gene activation pathway. Cell 119: Kang Y, Lee JH, Hoan NN, Sohn HM, Chang IY, You HJ. (2009). Protein phosphatase 5 regulates the function of 53BP1 after neocarzinostatin-induced DNA damage. J Biol Chem 284: Kauppinen TM, Chan WY, Suh SW, Wiggins AK, Huang EJ, Swanson RA. (2006). Direct phosphorylation and regulation of poly(adp-ribose) polymerase-1 by extracellular signal-regulated kinases 1/2. Proc Natl Acad Sci USA 103: Kim JS, Krasieva TB, Kurumizaka H, Chen DJ, Taylor AM, Yokomori K. (2005). Independent and sequential recruitment of NHEJ and HR factors to DNA damage sites in mammalian cells. J Cell Biol 170: Kinner A, Wu W, Staudt C, Iliakis G. (2008). Gamma-H2AX in recognition and signaling of DNA double-strand breaks in the context of chromatin. Nucleic Acids Res 36: Lai JS, Herr W. (1992). Ethidium bromide provides a simple tool for identifying genuine DNA-independent protein associations. Proc Natl Acad Sci USA 89: Leppard JB, Dong Z, Mackey ZB, Tomkinson AE. (2003). Physical and functional interaction between DNA ligase IIIalpha and poly(adp-ribose) polymerase 1 in DNA single-strand break repair. Mol Cell Biol 23: Li B, Navarro S, Kasahara N, Comai L. (2004). Identification and biochemical characterization of a Werner s syndrome protein complex with Ku70/80 and poly(adp-ribose) polymerase-1. J Biol Chem 279: Lis JT, Kraus WL. (2006). Promoter cleavage: a topoiibeta and PARP-1 collaboration. Cell 125: Masson M, Niedergang C, Schreiber V, Muller S, Menissier-de Murcia J, de Murcia G. (1998). XRCC1 is specifically associated with poly(adp-ribose) polymerase and negatively regulates its activity following DNA damage. Mol Cell Biol 18: Meyer-Ficca ML, Meyer RG, Jacobson EL, Jacobson MK. (2005). Poly(ADP-ribose) polymerases: managing genome stability. Int J Biochem Cell Biol 37: Midorikawa R, Takei Y, Hirokawa N. (2006). KIF4 motor regulates activity-dependent neuronal survival by suppressing PARP-1 enzymatic activity. Cell 125: Mitsuhashi S, Matsuura N, Ubukata M, Oikawa H, Shima H, Kikuchi K. (2001). Tautomycetin is a novel and specific inhibitor of serine/ threonine protein phosphatase type 1, PP1. Biochem Biophys Res Commun 287: Ogata N, Ueda K, Kawaichi M, Hayaishi O. (1981). Poly(ADP-ribose) synthetase, a main acceptor of poly(adp-ribose) in isolated nuclei. J Biol Chem 256: Okayama H, Edson CM, Fukushima M, Ueda K, Hayaishi O. (1977). Purification and properties of poly(adenosine diphosphate ribose) synthetase. J Biol Chem 252: Perrault R, Wang H, Wang M, Rosidi B, Iliakis G. (2004). Backup pathways of NHEJ are suppressed by DNA-PK. J Cell Biochem 92: Pion E, Ullmann GM, Ame JC, Gerard D, de Murcia G, Bombarda E. (2005). DNA-induced dimerization of poly(adp-ribose) polymerase-1 triggers its activation. Biochemistry 44: Poirier GG, de Murcia G, Jongstra-Bilen J, Niedergang C, Mandel P. (1982). Poly(ADP-ribosyl)ation of polynucleosomes causes relaxation of chromatin structure. Proc Natl Acad Sci USA 79: Postow L, Ghenoiu C, Woo EM, Krutchinsky AN, Chait BT, Funabiki H. (2008). Ku80 removal from DNA through double strand break-induced ubiquitylation. J Cell Biol 182: Povirk LF. (1996). DNA damage and mutagenesis by radiomimetic DNA-cleaving agents: bleomycin, neocarzinostatin and other enediynes. Mutat Res 355: Povirk LF, Han YH, Steighner RJ. (1989). Structure of bleomycininduced DNA double-strand breaks: predominance of blunt ends and single-base 5 0 extensions. Biochemistry 28: Rivera-Calzada A, Spagnolo L, Pearl LH, Llorca O. (2007). Structural model of full-length human Ku70-Ku80 heterodimer and its recognition of DNA and DNA-PKcs. EMBO Rep 8: Ruscetti T, Lehnert BE, Halbrook J, Le Trong H, Hoekstra MF, Chen DJ et al. (1998). Stimulation of the DNA-dependent protein kinase by poly(adp-ribose) polymerase. J Biol Chem 273: Schild-Poulter C, Shih A, Yarymowich NC, Hache RJ. (2003). Downregulation of histone H2B by DNA-dependent protein kinase in response to DNA damage through modulation of octamer transcription factor 1. Cancer Res 63: Schultz N, Lopez E, Saleh-Gohari N, Helleday T. (2003). Poly(ADPribose) polymerase (PARP-1) has a controlling role in homologous recombination. Nucleic Acids Res 31: Skinner J, Sinclair C, Romeo C, Armstrong D, Charbonneau H, Rossie S. (1997). Purification of a fatty acid-stimulated proteinserine/threonine phosphatase from bovine brain and its identification as a homolog of protein phosphatase 5. J Biol Chem 272: Tanaka Y, Koide SS, Yoshihara K, Kamiya T. (1987). Poly (ADPribose) synthetase is phosphorylated by protein kinase C in vitro. Biochem Biophys Res Commun 148: Wang M, Wu W, Rosidi B, Zhang L, Wang H, Iliakis G. (2006). PARP-1 and Ku compete for repair of DNA double strand breaks by distinct NHEJ pathways. Nucleic Acids Res 34: