A Positive Role for the Ku Complex in DNA Replication Following Strand Break Damage in Mammals*

Ku70-Ku80 complex is the regulatory subunit of DNA-dependent protein kinase (DNA-PK) and plays an essential role in double-strand break repair following ionizing radiation (IR). It preferentially interacts with chromosomal breaks and protects DNA ends from nuclease attack. Here we show evidence that cells defective in Ku80 exhibit a significantly slow S phase progression following DNA damage. IR-induced retardation in S phase progression in Ku80-/- cells was not due to the lack of DNA-PK kinase activity because both wild-type cells and DNA-PKcs-deficient cells showed no such symptom. Instead, proliferating cell nuclear antigen (PCNA) dissociated from chromosomes following IR in Ku80-deficient cells but not in wild-type or DNA-PKcs-deficient cells. Treatment of HeLa cells with IR induced colocalization of the Ku complex with PCNA on chromosomes. Together, these results suggest that binding of the Ku complex at chromosomal breaks may be necessary to maintain the sliding clamps (PCNA) on chromatin, which would allow cells to resume DNA replication without a major delay following IR.

Upon DNA damage, eukaryotic cells invoke several mechanisms that control DNA metabolism as well as cell cycle progression (1)(2)(3). Damage checkpoint mechanism is the first line of damage defense in eukaryotic cells, which not only blocks initiation of replication but also acts as formidable replication/ transcription barriers and causes slowing/stalling of replication/transcription machinery at damaged DNA sites (4). Stalled replication or transcription machinery requires checkpoint genes such as rad53 and mec1 for stabilization (5)(6)(7), although it triggers physical/functional interaction with repair factors to initiate the repair of damaged DNA (8,9).
Unlike other DNA damages, strand break damage does not appear to act itself as a replication barrier. Instead, it induces the damage checkpoint pathways that activate a series of S phase checkpoint genes to prevent initiation of replication (4). Damage-induced S phase checkpoint pathways are initiated by activation of phosphatidylinositol (PI) 1 3-kinase family members (ataxia telangiectasia mutated (ATM), ATM-related, and DNA-dependent protein kinase (DNA-PK)) resulting in phosphorylation of their targets such as Chk2, Nbs1, and MDC1 (KIAA0170/NFBD1) (10 -12). Both Chk2 and Nbs1 are direct targets of ATM and contribute to inhibition of replicative DNA synthesis, but are involved in two separate intra-S phase checkpoint pathways (11). Chk2 phosphorylates CDC25A phosphatase, which results in proteolytic degradation. Lack of CDC25A leads to an accumulation of inactive Cdk2 (phosphorylated at Thr-14 and Tyr-15) (10), which likely prohibits loading of a replication initiation factor, CDC45, to the origin of replication (13). As a result, wild-type cells exhibit an IRsensitive inhibition in DNA synthesis, whereas cells lacking a checkpoint gene exhibit radio-resistant DNA synthesis (RDS) following DNA damage (10). On the other hand, Nbs1 is a part of the multisubunit nuclease, called Mre11 complex, consisting of Mre11, Rad50, and Nbs1 (also known as xrs2) (14,15), and mutations in these genes have been related to DNA repair deficiencies (16,17). A number of studies have uncovered functions for the Mre11 complex in checkpoint signaling and in DNA replication (18,19). Nonetheless, a detailed understanding of how this complex functions in the intra-S phase checkpoint pathway and inhibition of DNA replication following DNA damage are not clear.
DNA-dependent protein kinase (DNA-PK) is a member of the PI 3-kinase family that is regulated by various stresses and shares amino acid sequence homology in its C-terminal kinase domain with other family members (ATM, ATM-related, and p110 PI 3-kinase) (20,21). It is a nuclear serine/threonine protein kinase that is not only involved in the damage checkpoint pathway but is also an essential factor for double-strand break (dsb) repair (also known as non-homologous end-joining (NHEJ) repair). It is a three-protein complex consisting of a 460-kDa catalytic subunit (DNA-PKcs) and two regulatory subunits (Ku70/Ku80 heterodimer). The latter regulates kinase activity upon binding to DNA termini (22). Studies with mouse and human cells defective in DNA-PKcs indicated that DNA-PK is not only necessary for exit from the DNA damageinduced G 2 arrest (23) but is also involved in S phase arrest following DNA damage (24,25). Both DNA-PK and ATM phosphorylate human replication protein A, a key replication/repair factor involved in replication and repair, suggesting a possible role for ATM/DNA-PK in controlling DNA replication following DNA damage, although the functional implication for replication protein A phosphorylation in DNA replication and repair remains unclear (26 -28).
Ku70/Ku80 are the regulatory subunits for DNA-PKcs that preferentially interact with chromosomal breaks following IR, which not only protects DNA ends from nuclease attack but is also essential for double-strand break repair (29 -31). In addition, human Ku complex has been shown to associate physically with telomerase and plays a unique role in telomere maintenance (32,33). In this study, we investigated a role for the Ku complex in S phase arrest following IR. We found that mammalian cells lacking Ku80 exhibit a prolonged retardation in S phase progression following DNA damage. This was not directly related to the lack of DNA repair activity in Ku80Ϫ/Ϫ cells. Instead, the Ku complex binds to chromosomal breaks following IR and prevents a key replication factor, proliferating cell nuclear antigen (PCNA), from being unloaded from chromatin. In this way, intact replication machinery can be maintained on the chromosome without falling apart, which eventually facilitates cells to resume DNA replication after repair.

MATERIALS AND METHODS
Cell Cultures-Mouse embryonic fibroblast cells (Ku80Ϫ/Ϫ, AtmϪ/Ϫ, and DNA-PKcsϪ/Ϫ) (34) and HeLa cells were grown in minimum essential medium and Dulbecco's modified Eagle's media F-12 (Invitrogen), respectively, which were supplemented with 10% fetal calf serum (Invitrogen), penicillin (10 unit/ml, Sigma), and streptomycin (0.1 mg/ml, Sigma). For irradiation, cells were grown in 150-mm dishes with 70 -80% confluence at the time of irradiation. Tissue culture medium was renewed 1 day prior to irradiation to remove non-viable and unattached cells. A Gammacell-40 exactor (Nordion) containing a 137 Cs source was used as a source for ionizing radiation.
BrdUrd Tracking Experiment-To analyze S phase progression of wild-type (WT) and mutant cells following IR treatment, cells were pulse-labeled with 10 M bromodeoxyuridine (BrdUrd) for 1 h, washed twice with PBS, treated with 0 -40 Gy of IR, and cultured for the indicated times before harvest. BrdUrd-labeled cells were detected us-ing FITC-conjugated anti-BrdUrd monoclonal antibody (Pharmingen) essentially according to the manufacturer's specifications. Briefly, cells labeled with BrdUrd were trypsinized, fixed in 70% ethanol, treated with 2 M HCl for denaturation, and washed in PBS. After neutralization with 0.1 M sodium borate, cells were resuspended in PBS, 0.5% Tween 20 containing anti-BrdUrd monoclonal antibody and incubated for 20 min at room temperature. After PBS wash, cells were resuspended in propidium iodide (50 g/ml, Sigma) and analyzed on a FACScan flow cytometer (BD Biosciences). Watson Pragmatic analysis was used to quantify cell cycle distribution of BrdUrd-incorporated cells (35).
Preparation of Intact Nuclei and Cytosolic Extracts-Cytosolic extracts and intact nuclei were prepared from asynchronously grown WT or mutant murine cells essentially as described previously (36). Cells were washed twice with PBS and allowed to swell in 20 ml of ice-cold hypotonic buffer (20 mM Hepes-KOH, pH 7.8, 5 mM potassium acetate, 0.5 mM MgCl 2 , and 0.5 mM DTT) per 150-mm dish for 10 min. All subsequent steps were carried out at 4°C. After removal of mitotic cells that detached from the dish under hypotonic conditions, interphase cells were scraped off the plates and disrupted with 25 strokes in a Dounce homogenizer using a loose-fitting pestle. Nuclei were pelleted at 1,500 ϫ g for 3 min, and the supernatant (cytosolic extracts) was collected and recentrifuged at 14,000 ϫ g for 20 min at 4°C. Cytosolic extracts were aliquoted and frozen in liquid N 2 . Pelleted nuclei, after a brief wash with hypotonic buffer, were resuspended at 1.5 ϫ 10 8 nuclei/ml in hypotonic buffer. Aliquots were frozen in drops in liquid nitrogen and stored at Ϫ80°C until use.
In Vitro Replication with Intact Nuclei-Replication reactions (30 l) containing cytosolic extracts (300 g of protein), a buffered mixture of NTPs, dNTPs, an ATP-regenerating system, and 1.0 ϫ 10 5 nuclei were performed as described previously (36). In vitro DNA replication reactions were incubated at 37°C in the presence of [␣-32 P]dATP for 2 h. Two volumes of a stop mixture (1% SDS, 5 mM DTT, 1 mM EDTA, 50 mM Tris, pH 6.8) were added after the incubation, and aliquots were pipetted onto GFC filters in quadruplicate. Two filters were dried and used to measure total counts, although the other two were trichloroacetic acid-precipitated (10% trichloroacetic acid containing 2% sodium pyrophosphate). By using the ratio of the incorporated radiolabel to the total radiolabel available in the reaction and relating this to the dATP pool, the mass of DNA synthesized was calculated using the formula ((trichloroacetic acid cpm/total cpm) ϫ 0.101 ϭ ng/l DNA synthesized).
In Vivo DNA Synthesis-Rate of DNA synthesis was measured by Relative DNA synthesis (%) was determined by measuring the incorporation of [ 3 H]thymidine into chromosomal DNA as described previously (24). Cells (1 ϫ 10 5 /16-mm diameter well) were incubated with [ 3 H]thymidine (5 Ci/ml; 67 Ci/mmol from ICN) and cold thymidine (0.2 g/ml) for the indicated times, and metabolism was stopped by the addition of 1/10 volume 2.3 M citric acid. Cells were washed once with phosphate-buffered saline (PBS), and DNA was precipitated with 10% trichloroacetic acid at 4°C for at least 2 h. Trichloroacetic acidprecipitated materials were solubilized in 250 l of 0.2 M NaOH and transferred to a vial containing 5 ml of scintillation fluid and 10 l of concentrated acetic acid for radioactivity measurement.
Double-strand Break Repair Assay-Kinetics of the rejoining of radiation-induced damaged DNA in cells following exposure to ␥ irradiation ( 137 Cs) were measured by pulsed field gel electrophoresis. Cells were grown in the presence of 2.5 M [ 14 C]thymidine (0.1 Ci/ml) (37,38) and treated with ionizing radiation. Following IR, cells were incubated at 37°C with prewarmed (42°C) fresh medium to allow dsb repair and then harvested at various times and resuspended in serumfree medium at a concentration of 2-5 ϫ 10 6 cells/ml. Cells were mixed with an equal volume of 1% agarose, and the solidified cell-agarose suspensions were lysed with buffer containing 10 mM Tris, pH 8.0, 50 mM NaCl, 0.5 M EDTA, 2% N-laurylsarcosyl, and proteinase E and O (0.1 mg/ml) for 16 -18 h at 50°C (37). DNA double-strand breaks were analyzed by asymmetric field inversion gel electrophoresis using 0.5% agarose gel in 0.5ϫ TBE at 10°C for 40 h. After electrophoresis, gels were analyzed by fluorography. For quantification of damaged DNA repair, intact chromosome and damaged DNA were separately removed from the gel and measured for 14 C using liquid scintillation counter.
Assessment of Chromatin-bound Proteins-Immunoblot analyses of chromatin-bound Ku complex and PCNA from WT and Ku80Ϫ/Ϫ cells were carried out following various doses of IR treatment. Cells were treated with cell lysis buffer I (25 mM Hepes, pH 7.5, 0.3 M NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5% Triton X-100, 0.5 mM DTT, and 1 mM phenylmethylsulfonyl fluoride) and incubated for 90 min on ice. After centrifugation at 14,000 ϫ g for 30 min, pellets were extensively washed four times with cell lysis buffer I, resuspended in SDS-PAGE loading buffer, and analyzed by a 10% SDS-PAGE. Gels were transferred to polyvinylidene difluoride membrane, probed with either Ku70/Ku80 antibody (goat polyclonal IgG, Santa Cruz Biotechnology) or PCNA antibody (monoclonal mouse IgG, Oncogene Sciences) followed by horseradish peroxidase-conjugated secondary antibody. Proteins were visualized by using the ECL system (Amersham Biosciences).
Laser-scanning Confocal Microscopy-Cells were grown up to 70% confluence on a sterilized Labtek II coverslip, irradiated with a 137 Cs source, and further incubated at 37°C until harvest. Immunofluorescence microscopy was carried out as described previously (8).
Treatment of Cells with Antisense Oligonucleotides-Cells were seeded in a 100-mm tissue culture plate (5 ϫ 10 6 ) 24 h prior to an antisense oligonucleotide treatment. Indicated amounts of either control-or Ku80-antisense oligonucleotides (morpholino, Gene Tools, LLC.) were mixed with 200 l of special delivery solution in a serum-free RPMI for 15 min at room temperature. After further incubation at room temperature for 20 min, antisense oligonucleotide solutions were added to cells in a 100-mm dish, briefly mixed, and returned to a 37°C incubator for 3 h. Solutions were removed from dishes and replaced with fresh serum-containing media for further incubation at 37°C 72 h before cell harvest.
We further examined whether IR-induced inhibition of replication activity in Ku80Ϫ/Ϫ cells is due to a blockade of G 1 /S phase transition or an alteration of the intra-S phase checkpoint pathway. WT and mutants (DNA-PKcsϪ/Ϫ and Ku80Ϫ/Ϫ) were initially pulse-labeled with BrdUrd to mark a defined population of cells in S phase and monitor their progression through G 2 /M phase of the cell cycle. Following 1-h pulse with BrdUrd, cells were treated with IR (5 Gy), cultured for the indicated times, and analyzed for cell cycle progression using fluorescence cytometry. In both WT and DNA-PKcsϪ/Ϫ, BrdUrd-labeled cells progressed through S phase but showed cell cycle arrest at G 2 /M phase following IR treatment ( Fig. 2B and Table I). In contrast, treatment of Ku80-deficient cells with IR not only induced G 2 /M phase arrest but also exhibited a prolonged retardation in S phase of the cell cycle that continued for at least 16 h ( Fig. 2B and Table I). A prolonged delay of Ku80Ϫ/Ϫ cells in S phase progression following IR was different from the ATM-mediated intra-S phase checkpoint pathway (Fig. 2B) that exhibits a transient replication arrest (10,11). Together, these results suggest the following: 1) IR-induced inhibition of S phase progression in Ku80Ϫ/Ϫ cells is not related to G 1 /S phase or the intra-S phase checkpoint pathway; 2) the Ku complex, unlike DNA-PKcs or ATM, may have a unique role in DNA replication following DNA damage.

IR-induced Retardation of S Phase Progression in Ku80Ϫ/Ϫ Cells Is Independent of DNA-PK Activity or dsb Repair-To
investigate further whether IR-induced lower replication activity in Ku80Ϫ/Ϫ cells is due to the lack of the Ku complex and not associated with DNA-PK kinase activity, WT murine fibroblast cells were treated with a PI 3-kinase inhibitor, wortmannin, and examined for replication activity following IR treatment. Regardless of IR, wortmannin treatment had no effect on DNA synthesis in WT cells (data not shown), suggesting that the role for Ku complex in replication is unique and not related to DNA-PK kinase activity. Because the Ku complex also plays a crucial role in dsb repair, we examined repair kinetics of WT and mutants following a high dose of IR (Fig. 3). Both WT and AtmϪ/Ϫ cells showed efficient repair activity immediately following IR treatment, whereas Ku80Ϫ/Ϫ cells exhibited very little or no repair (Fig. 3). In the meanwhile, DNA-PKcsϪ/Ϫ After washing with TBST (20 mM Tris-HCl, pH 7.4, 137 mM NaCl, and 0.2% Tween 20), slides were treated with TBST containing 5% non-fat dried milk for 30 min and incubated with an anti-PCNA monoclonal antibody for 1 h at 37°C. After washing with TBST three times, the slides were incubated for 1 h with donkey anti-mouse IgG antibody conjugated with FITC (Molecular Probes). The slides were washed in TBST, counterstained for DNA with 4,6-diamidino-2-phenylindole (0.1 g/ml, Vector Laboratories), and covered with a cover glass. Images were collected using Zeiss LSM-510 microscope. cells showed effective but a much slower dsb repair compared with WT or AtmϪ/Ϫ cells (Fig. 3). This result suggests that IR-induced prolonged inhibition of replication activity in Ku80Ϫ/Ϫ cells may not be due to the lack of dsb repair because DNA-PKcsϪ/Ϫ cells, despite of its lower dsb repair activity, showed very little or no change in replication activity following IR.

IR-induced Lower Replication Activity in Ku80Ϫ/Ϫ Cells Cannot Be Overcome by the Addition of Ku Complex in Vitro-
IR-induced inhibition of replication activity in Ku80Ϫ/Ϫ may be overcome by addition of Ku70/Ku80 if the Ku complex is directly responsible for it. To test this, we carried out an in vitro chromatin replication reactions with intact nuclei and cytosolic extracts because most replication factors are present in cytosolic fractions of the cell extracts (39). Intact nuclei from WT or Ku80Ϫ/Ϫ cells efficiently supported DNA synthesis in vitro (Fig. 4), whereas a considerably lower DNA synthesis was observed in reactions containing nuclei from irradiated Ku80Ϫ/Ϫ cells regardless of the source of cytosolic extracts. This suggests that IR-induced lower replication activity was dependent on the nuclear fraction and not the cytosolic extracts (protein fractions). It should be pointed out that the addition of cytosolic extracts (from WT) to the nuclei (from irradiated Ku80-/) was not able to rescue lower replication activity (Fig.  4), although more than 80% of the Ku complex was in cytosolic extracts (data not shown). This suggests that IR treatment may modulate nuclei in the absence of functional Ku complex.

IR Induces Dissociation of PCNA from Chromatin in Ku80Ϫ/Ϫ Cells but Not WT and DNA-PKcsϪ/Ϫ Cells-The
observation that the addition of Ku complex was not able to overcome damage-induced replication inhibition in Ku80Ϫ/Ϫ cells (Fig. 4) raises the possibility that a lack of functionally active Ku complex may cause a change in chromatin structure or distribution of replication factor(s) following DNA damage, leading to a lower replication activity in vivo. In keeping with this notion, an increased interaction between the Ku70-Ku80 complex and PCNA was observed following DNA damage (8). This prompted us to examine whether the presence of Ku complex affects the association of PCNA with chromatin following DNA damage. For this, nuclei were isolated from WT or Ku80Ϫ/Ϫ cells following IR and examined for the presence of PCNA. As reported previously (22,40,41), treatment of cells with IR induced the interaction of Ku complex with chromatin in a dose-dependent manner (Fig. 5A). Interestingly, the amount of PCNA was markedly reduced following DNA damage in chromatin from Ku80Ϫ/Ϫ cells but not from WT cells (Fig. 5A, middle panel). Kinetic studies showed that PCNA dissociated from chromatin within an hour following DNA damage and remained dissociated for at least 6 h (Fig. 5B). This result is in keeping with the immunofluorescence studies that dissociation of PCNA from chromatin was observed within 4 h following IR in Ku80Ϫ/Ϫ cells but not in WT or DNA-PKcsϪ/Ϫ cells (Fig. 6), suggesting that binding of the Ku complex at DNA breaks protects PCNA and perhaps other replication factors from dissociating from chromatin following IR, which may be necessary for S phase cells to immediately resume DNA replication upon completion of DNA repair.
Suppression of Ku80 Expression Exhibits a Prolonged Retardation in Replication and Dissociation of PCNA from Chromatin Following DNA Damage-We next examined whether IRinduced inhibition of replication in Ku80Ϫ/Ϫ murine cells also occurred in human cells. HeLa cells were treated with either a control (scrambled) or Ku80-targeted antisense RNA (morpholino) and analyzed for in vivo replication activity following IR treatment. A significantly reduced expression of Ku80 was observed in HeLa cells treated with Ku80 antisense RNA but not with a control RNA (Fig. 7A). Treatment of cells with Ku80 antisense RNA not only lowered replication activity in vivo (Fig. 7B) but also decreased the amount of PCNA associated with chromatin following IR (Fig. 7C). These observations were similar to those observed with Ku80-deficient murine cells (Figs. 1 and 5), suggesting that the Ku complex plays a unique and direct role in replication following DNA damage.
Colocalization of PCNA and Ku80 (Ku Complex) at DNA Breaks Following IR-The Ku complex, once bound to DNA ends, may interact with PCNA to keep the sliding clamp and other replication factors from dissociating from chromatin. If so, we may be able to see colocalization of the Ku complex and PCNA at DNA damage sites. To examine this, HeLa cells were treated with IR and prepared for chromatin following extensive washing with buffer containing 0.3 M NaCl. Chromatin-associated proteins were examined by immunofluorescence for the presence of Ku complex and PCNA. PCNA was largely associated with chromatin regardless of DNA damage (Fig. 8, A  versus D). Without IR treatment the Ku complex was not associated with chromatin but became associated with chromatin following DNA damage (Fig. 8, B versus E). In merged images, PCNA and Ku80 were separately localized in HeLa cells without IR treatment (Fig. 8C) but colocalized on chromatin following DNA damage (Fig. 8F).
To see whether the interaction between PCNA and Ku complex is mediated by DNA damage, cytosolic extracts from Ku80ϩ/ϩ or Ku80Ϫ/Ϫ cells were mixed with streptavidin-Sepharose in the presence or absence of biotin-labeled duplex DNA (Fig. 9). Association of PCNA with Ku complex was observed only when cell extracts were incubated with duplex DNA, suggesting that the interaction between two proteins occurs at DNA ends. However, we cannot rule out a possibility that PCNA may simply be trapped at DNA breaks in the presence of Ku complex without any direct interaction between the two proteins.

DISCUSSION
In this study, we describe a novel role for dsb repair factor, the Ku complex, in replication following IR-induced DNA damage. Unlike cells deficient in S phase checkpoint genes such as atm, Ku80Ϫ/Ϫ cells triggered the RDS checkpoint following IR but failed to recover from RDS (Fig. 3), suggesting that IRinduced replication inhibition observed in Ku80Ϫ/Ϫ cells is independent of the damage-induced intra-S phase checkpoint. A potential role for the Ku complex in initiation of replication has been suggested as it associates with a mammalian origin of DNA replication as well as with replication proteins (42). Although we did not see a notable difference between WT and Ku80Ϫ/Ϫ cells in S phase progression/DNA replication in the absence of DNA damage, the Ku complex seems to have a novel role in replication following DNA damage, which allows cells to resume S phase progression after repair of DNA damage.
The Ku complex, due to its high affinity for DNA ends, immediately interacts with DNA following strand break damage. Binding of the Ku complex elicits conformational changes that likely allow it to bind DNA-PKcs. Thus, the presumed role for the Ku complex is to bind DNA breaks and recruit DNA-PKcs to the damaged sites for NHEJ repair. The Ku complex may also serve as an alignment factor that not only increases NHEJ efficiency but also enhances the accuracy (30,43). Upon the assembly of DNA-PK holoenzyme on DNA breaks, the repair complex activates its serine/threonine protein kinase activity and phosphorylates target substrates that colocalize with it at the ends of broken DNA. Binding of the Ku complex to ends of DNA is not only necessary to promote NHEJ repair but also to protect chromosomal DNA from nuclease attack (30,44). IR-induced inhibition of replication in Ku80-deficient cells was not likely due to the lack of DNA repair activity because DNA-PKcsϪ/Ϫ showed no defect in S phase progression following IR treatment although exhibited much slower DNA repair activity than WT cells (Fig. 3). It should be pointed out, however, that delayed S phase progression in Ku80Ϫ/Ϫ was directly related to the presence of unrepaired DNA damage; accordingly, lack of dsb repair in Ku80Ϫ/Ϫ is one of the major factors in inducing a delayed S phase progression. In fact, Ku complex physically interacts with XRCC4-ligase IV complex, which increases the rate of DNA ligation by 20-fold (45,46). This reaction is independent of DNA-PKcs, suggesting a unique role for Ku complex outside the context of DNA-PK, which may be essential for S phase cells to carry out DNA replication without a delay following IR.
The Ku complex was shown to interact physically with PCNA, which can be significantly enhanced by DNA damage (8). Interaction between the Ku complex and PCNA may be necessary to keep PCNA sliding clamps on chromatin following IR and to avoid any delay in resuming DNA synthesis. However, neither Ku70 nor Ku80 possesses the PCNA interaction motif (47). Moreover, we were not able to detect a physical interaction between the Ku complex and PCNA in the absence of DNA (data not shown). It is possible that the interaction between PCNA and the Ku complex only occurs at DNA ends ( Fig. 9), or the presence of Ku complex at DNA breaks may force PCNA to sit close to DNA ends without falling off. Alternatively, the independent function of Ku in binding at DNA ends and stimulation of DNA ligation (45,46) may play a crucial role in keeping PCNA at damaged DNA foci, whereas Ku80Ϫ/Ϫ cells were unable to join dsbs. In any event, more detailed study is necessary to elucidate the role for the Ku complex in DNA replication.
PCNA forms a stable homotrimer through the interdomain loops located at the N and C termini, resulting in forming a ring structure around duplex DNA (48,49). Although PCNA is not a DNA-binding protein, it is introduced to the DNA through an interaction with the replication factor C and acts as a sliding clamp that tethers DNA polymerase ␦ to DNA, resulting in a highly processive DNA synthesis (50). Synthesis of lagging strand in eukaryotes involves the polymerase switch from pol ␣ to pol ␦, which requires reloading of PCNA (51,52). It is not clear, however, what exactly happens to the PCNA trimer once it gets off DNA during replication. In our in vitro experiments, the addition of cytosolic extracts from WT cells did not stimulate poor replication activity associated with intact chromatin from Ku80Ϫ/Ϫ (Fig. 4), although excess amount of both PCNA and Ku complex were present in cytosolic extracts (data not shown). This may simply be due to the presence of unrepaired DNA damage that prohibits DNA replication even after the addition of the Ku complex and PCNA in vitro. There may not be an adequate mechanism in vivo to reload PCNA once it slips through DNA ends because replication was significantly delayed for up to 24 h in Ku80Ϫ/Ϫ cells following IR (Fig. 1). Alternatively, the presence of unprotected DNA breaks in the absence of the Ku complex may induce an alteration of chromatin structure and/or modification of chromatin-associated factors. We have also observed that Ku80-deficient cells not only had significantly lowered PCNA in chromatin but also showed decreased amount of replication factor C associated with chromatin following IR (data not shown), suggesting that binding of the Ku complex at DNA ends may be necessary to keep other replication proteins from slipping away from the chromosome through DNA breaks.
PCNA is tightly associated with chromatin in S phase cells and functions as a sliding clamp for elongation of the DNA chain by DNA pol ␦. Our immunofluorescence studies showed that PCNA in asynchronously grown cells were also tightly associated with chromosomal DNA (Figs. 6 and 8). Most of the PCNA may form the ring structure throughout the cell cycle as it interacts with proteins involved in excision repair, mismatch repair, cellular regulation, and DNA processing in addition to replication factors (53)(54)(55)(56)(57). The role for PCNA in DNA metabolism is beyond DNA replication; therefore, the functional implication for the Ku complex in relation to PCNA following DNA damage would certainly be extended to other DNA metabolism.