The radiosensitive mutant xrs-5, a derivative of the Chinese hamster ovary (CHO) K1 cell line, is defective in DNA double-strand break repair and V(D)J recombination. The defective phenotypes of xrs-5 cells are complemented by the 86 kDa subunit of Ku antigen. OBA is a protein,previously purified from HeLa cells, that binds in a sequence-specific manner to mammalian origins of DNA replication. The DNA-binding subunit of OBA has been identified as Ku86. We tested the xrs-5 cell line for its ability to replicate a mammalian origin-containing plasmid, p186, in vivo and in vitro. In vivo, the p186 episomal DNA replication in transfected xrs-5 cells was reduced by 45% when compared with the CHO K1 cells transfected with p186. In vitro, although total and cytoplasmic cell extracts from xrs-5 cells replicated the p186 with the same efficiency as the parental CHO K1 cell extracts, xrs-5 nuclear extracts did not possess any detectable replication activity. Addition of affinity-purified OBA/Ku restored replication in the xrs-5 nuclear extract reaction. Western blot analyses showed that the levels of other replication proteins (Orc2,PCNA, DNA polymerase ϵ and δ, Primase and Topoisomerase IIα)were comparable in both the xrs-5 mutant and CHO K1 wild-type cell lines. In addition, the in vivo association of Ku with the DHFR origin-containing sequence (oriβ) was examined in both the CHO K1 and xrs-5 cell lines by a chromatin immunoprecipitation (ChIP) assay. Anti-Ku antibodies did not immunoprecipitate a detectable amount of Ku from the xrs-5 cells in the origin-containing sequence, in contrast to the CHO K1 cells, wherein Ku was found to be associated with the oriβ origin. The data implicate Ku antigen in in vivo and in vitro DNA replication and suggest the existence of another protein with Ku-like functions in the xrs-5 cells.
The radiosensitive mutant cell line, xrs-5, a derivative of the Chinese hamster ovary (CHO) K1, is defective in V(D)J recombination and DNA double-strand break (DSB) rejoining (Kemp et al., 1984; Pergola et al.,1993; Taccioli et al.,1994; Rathmell and Chu,1994). The xrs mutants were initially described as the first mammalian cell mutants with a defect in DSB rejoining(Kemp et al., 1984). They were isolated by a simple technique involving transfer of heavily mutagenized cells from single colonies, followed by irradiation with a dose determined not to affect survival of wild-type CHO K1 cells(Jeggo and Kemp, 1983). The xrs-5 cells, along with other members of the X-ray complementation group 5, also lack DNA end-binding activity(Rathmell and Chu, 1994; Getts and Stamato, 1994; Singleton et al., 1997). The defective repair and recombination phenotypes of the xrs-5 cells have been traced to the 86 kDa subunit of the Ku protein, since Ku86 cDNA restored DNA-end binding, X-ray sensitivity and V(D)J recombination(Getts and Stamato, 1994; Rathmell and Chu, 1994; Smider et al., 1994; Taccioli et al., 1994).
Ku (reviewed in Tuteja and Tuteja,2000) is a heterodimeric DNA-binding protein of 70 kDa and 86 kDa subunits (Mimori et al.,1986). Ku binds avidly to DNA ends, whether blunt or with 5′or 3′ overhangs, as well as to other discontinuities in the DNA structure (Mimori and Hardin,1986; Taghva et al.,2002; Arosio et al.,2002). Sequence-specific binding of Ku has also been demonstrated,including binding to mammalian origins of DNA replication(Giffin et al., 1996; Ruiz et al., 1999; Araujo et al., 1999; Novac et al., 2001). Ku is necessary for the processing of DNA DSBs that are induced by damaging agents such as ionizing radiation or oxidative reactions, by endogenous recombination processes and by certain chemotherapeutic drugs (reviewed in Featherstone and Jackson,1999; Tuteja and Tuteja,2000; Doherty and Jackson,2001). An involvement of Ku has also been suggested in transcription (Giffin et al.,1996; Finnie et al.,1993; Kuhn et al.,1995; Generesch et al., 1995; Camara-Clayette et al., 1999),telomeric maintenance (Boulton and Jackson,1996; Porter et al.,1996; Polotnianka et al.,1998; Haber, 1999; Baumann and Cech, 2000), replicative senescence(Woo et al., 1998; Lim et al., 2000), suppression of chromosomal aberrations, malignant transformation and tumour progression(Difilippantonio et al., 2000; Pucci et al., 2001), aging(Vogel et al., 1999; Cooper et al., 2000; Li and Comai, 2001) cell cycle regulation (Munoz et al.,2001), multidrug resistance(Um et al., 2001) and DNA replication (Ruiz et al.,1999; Novac et al.,2001).
A role for Ku in DNA replication and replication fork movement was proposed a decade ago, on the basis of the observation that Ku bound to DNA ends during S phase (Paillard and Strauss,1991). Since then, evidence for the involvement of Ku in DNA replication has been accumulating. The HeLa cell DNA-dependent ATPase,co-fractionating with a 21S multiprotein complex capable of supporting SV40 in vitro DNA replication (Vishwanatha and Baril, 1990), is identical to Ku antigen(Cao et al., 1994). Ku binds to viral and eukaryotic origins of DNA replication(de Vries et al., 1989; Toth et al., 1993; Araujo et al., 1999; Ruiz et al., 1999; Novac et al., 2001), to matrix attachment regions (MARs), which serve as initiation sites for DNA replication(Galande and Kohwi-Shigematsu,1999) and to Alu DNA, which provides cis-elements affecting chromatin structure during DNA replication(Tsuchiya et al., 1998). Ku also binds in vivo to genomic sequences comprising origins of DNA replication in a cell-cycle-dependent manner, with fivefold higher binding at the G1/S interphase by comparison to G0 (Novac et al., 2001). Recently, Ku was shown to interact with WRN(Cooper et al., 2000), a DNA helicase with 3′-5′ exonuclease activity that is part of the replication complex (Lebel et al.,1999) and interacts with replication protein A (RPA)(Brosh et al., 1999). Furthermore, OBF2, a Saccharomyces cerevisiae Ku homologue, supports the formation of a stable multiprotein replication complex by binding to essential replication sequences (Shakibai et al., 1996), whereas a Ku70 yeast mutant strain exhibits a high DNA content during mitotic growth (Barnes and Rio, 1997). Finally, the phenotypes of the Ku knockout mice,such as their small size, the failure of the cells to proliferate in culture,their prolonged doubling time and their premature senescence may also suggest a role of Ku in DNA replication(Featherstone and Jackson,1999). Extracts prepared from Ku86-/- cells,deficient in Ku protein, have approximately 70% lower in vitro DNA replication activity compared with Ku86+/+ extracts(Novac et al., 2001).
We have previously shown that OBA (Ruiz et al., 1995), a HeLa cell activity whose DNA binding subunit has been identified as Ku86, is involved in mammalian DNA replication(Ruiz et al., 1999). The affinity-purified OBA (Ku) binds to mammalian origins of replication,including A3/4, a 36 bp sequence that is common to mammalian origins of DNA replication including the Chinese hamster DHFR origin-containing sequence(oriβ), the human c-myc, dmnt-1 and lamin B2 origins and the monkey ors sequences, among others(Araujo et al., 1999; Ruiz et al., 1999). Depletion of Ku, either by inclusion of an oligonucleotide comprising its binding site(A3/4) or by antibodies directed against the Ku protein, inhibited DNA replication to as low as 10-20% in a mammalian in vitro replication system(Ruiz et al., 1999).
The mammalian in vitro replication cell-free system used in this study mimics nuclear, semi-conservative DNA replication in vivo(Pearson et al., 1991; Pearson et al., 1994; Zannis-Hadjopoulos et al.,1994; Todd et al.,1995; Pelletier et al.,1997; Pelletier et al.,1999) and is comparable to other mammalian in vitro cell-free systems (Krude, 2000). It uses a plasmid containing a mammalian origin of DNA replication and extracts from HeLa cells (Pearson et al.,1991) and is based on the SV40 in vitro replication system, but it does not require the viral T antigen. This system has been used to study the effects on DNA replication of cancer chemotherapeutic drugs(Diaz-Perez et al., 1996; Diaz-Perez et al., 1998), of various proteins including the Oct-1 transcription factor(Matheos et al., 1998), the GATA-1 factor, Ku antigen and DNA polymerase δ(Ruiz et al., 1999; Matheos et al., 2002), a mammalian polynucleotide kinase (Jilani et al., 1999), the cruciform binding 14-3-3 proteins(Novac et al., 2002) and the replication competence of Ku86+/+ and Ku86-/- mouse embryonic fibroblasts(Novac et al., 2001).
Since the xrs-5 mutant cells have severely reduced levels of Ku86 and Ku70 proteins, here we tested their potential for replication in vivo and in an in vitro replication system. We found that the xrs-5 cells had a reduced ability to support in vivo DNA replication upon transfection of p186, a mammalian origin-containing plasmid. However, total and cytoplasmic extracts from xrs-5 cells replicated the origin-containing plasmid in vitro with the same efficiency as the wild-type CHO K1 cell extracts. By contrast, xrs-5 nuclear cell extracts did not possess any detectable in vitro replication activity, but addition of affinity-purified OBA/Ku restored in vitro replication activity to wild-type levels; this is similar to the way that Ku86 cDNA complements the defective repair and recombination phenotypes (Smider et al.,1994). Also, the levels of other replication proteins such as Orc2, PCNA, DNA polymerase ϵ and δ, Primase and Topoisomerase IIα were comparable in both the xrs-5 mutant and CHO KI wild-type cell lines. Using a ChIP assay, we also found that in xrs-5 cells no Ku could be detected bound to the hamster DHFR oriβ in vivo,unlike the result in CHO K1 cells, in which Ku was bound to this origin. Moreover, we identified a factor in the xrs-5 cytoplasmic cell extracts that bound in a sequence-specific manner to the A3/4 origin sequence,possibly accounting for the efficient replication of the xrs-5 total and cytoplasmic cell extracts in vitro. The data suggest that Ku participates in mammalian DNA replication in vivo and in vitro and that the factor present in the cytoplasmic extracts can compensate for the lack of Ku in the nuclear extracts.
Materials and Methods
The xrs-5 cell line was derived from the CHO K1 cell line on the basis of its sensitivity to ionizing radiation(Jeggo and Kemp, 1983). CHO K1 and xrs-5 cells were cultured in RPMI medium (Gibco-BRL, Burlington,Ontario, Canada) supplemented with penicillin (100 μg/ml), streptomycin(0.1 mg/ml), 1 mM glutamine and 10% fetal bovine serum (Gibco-BRL). Cells were maintained at 37°C in 5% CO2.
Cell extracts from log phase CHO K1 and xrs-5 cell monolayers were prepared as previously described by Pearson et al.(Pearson et al., 1991). Briefly, monolayers were washed twice with isotonic buffer (20 mM Tris-HCl pH 7.4, 137 mM NaCl, 5 mM KCl, 1 mM CaCl2, 0.5 mM MgCl2 and 250 mM glucose). The cells were collected and lysed in a Type B Dounce homogenizer. Nuclei were removed by 5 minutes of centrifugation at 1200 g and were subsequently used to prepare the nuclear extract. The supernatant was centrifuged for 1 hour at 100,000 g in a Beckman type 50 Ti rotor, and the supernatant was used as the cytoplasmic extract. The nuclear pellet was suspended in hypertonic buffer (hypotonic and 500 mM potassium acetate) in half the volume used for the cytoplasmic extract. The mixture was incubated for 90 minutes on ice, spun in a Beckman SW50.1 rotor at 300,000 g for 1 hour and used as the nuclear extract.
Preparation of template DNA for replication assay
Plasmid p186 consists of the minimal origin of ors8 (GenBank accession number M26221). It contains the NdeI/RsaI sub-fragment of ors8 cloned in the NruI site of pBR322(Todd et al., 1995). Plasmid DNA was isolated using the QIAGEN-tip 500 column (QIAGEN, Mississauga,Ontario, Canada).
In vivo DNA replication
In vivo DNA replication was performed as previously described(Landry and Zannis-Hadjopoulos,1991; Pearson et al.,1991; Wu et al.,1993; Nielsen et al.,1994; Zannis-Hadjopoulos et al., 1994; Todd et al.,1995) with the following modification. HeLa cells, at a density of 3×105, were seeded and transfected with 10 μg plasmid DNA(p186 or equimolar amounts of pBR322) using the calcium phosphate transfection kit (Invitrogen). Cells were also transfected with the pcDNA6.0 HislacZ to control for transfection efficiency. 72 hours post-transfection, cells were lysed as previously described (Todd et al., 1995) and low molecular weight DNA was isolated and purified using the QIAquick columns (QIAGEN). The DNA was then digested with 1 unit of DpnI for 1 hour to digest unreplicated (bacterially methylated)plasmid. The DpnI-digested and undigested DNAs were used to transform TOP 10 E. coli cells (Invitrogen), as previously described(Landry and Zannis-Hadjopoulos,1991). The in vivo relative DNA replication was determined by counting the number of colonies obtained with the transformation using the DpnI-digested plasmid and correcting it for the amount of total plasmid DNA isolated as determined by counting the number of bacterial colonies obtained with undigested plasmid DNA isolated by Hirt lysis. The replication level of the xrs-5 cells is expressed as a percentage of the wild-type CHO K1 cells.
Mammalian in vitro DNA replication
Replication assays were performed as previously described(Pearson et al., 1991; Matheos et al., 1998; Novac et al., 2002), with slight modifications. CHO K1 or xrs-5 nuclear cell extracts (16μg), cytoplasmic cell extracts (45 μg) or nuclear and cytoplasmic cell extracts together (8:15 ratio) were added to the replication mixture with 100 ng of input p186 plasmid DNA. Unmethylated pBluescript KS+ was also typically included in each in vitro replication reaction as a control for DNA recovery and DpnI resistance. A reaction of pBR322, a methylated,non-replicating plasmid, was also performed for each in vitro replication experiment to show that the observed DNA replication was origin dependent.
Experiments involving the addition of A3/4-affinity-purified Ku(Ruiz et al., 1999) were performed by pre-incubating the nuclear cell extracts with 160 ng or 600 ng of affinity-purified Ku on ice for 20 minutes, prior to the replication reaction. Replication reactions were performed at 30°C for 1 hour. The in vitro replication reaction products were purified using the QIAquick PCR purification kit (QIAGEN). Reaction samples were digested with DpnI(New England Biolabs, Mississauga, Ontario, Canada), as previously described(Matheos et al., 1998), and resolved by electrophoresis on 1% agarose gel. Quantification was performed on DpnI-digested products by densitometric measurements using a phoshorimager analyzer (Fuji BAS2000, Stamford, CT), as described previously(Diaz-Perez et al., 1996; Matheos et al., 1998; Novac et al., 2001). The total amount of DNA recovered from the in vitro replication reaction was determined by quantitative analysis of the picture of the ethidium-bromide-stained gel.
Nuclear or cytoplasmic extract proteins were resolved by 8% SDS-PAGE,transferred to Immobilon-P (Millipore, Mississauga, Ontario, Canada), probed with human anti-Ku70 (C-19), anti-Ku86 (C-20), anti-PCNA (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), anti-ORC2 (Medical and Biological Laboratories CO, Ltd, Watertown, MA), anti-DNA Polymerase δ (BD Transduction Laboratories, Inc., Mississauga, Ontario), anti-DNA Polymeraseϵ (Clone 93G1A), anti-DNA Primase (p49) (Neomarkers, Union City, CA) or anti-Topoisimerase IIα (Ki-S1) (Chemicon International, Inc., Temecula,CA) followed by treatment with horseradish-peroxidase-conjugated donkey anti-goat, anti-mouse or anti-rabbit IgG (Santa Cruz Biotechnology, Inc). The blots were developed using the ECL western blotting detection kit(Amersham-Pharmacia, Baie d'Urfé, Québec, Canada).
In vivo crosslinking
ChIP assays were performed as described previously(Novac et al., 2001). Briefly,CHO K1 and xrs-5 cells were crosslinked with 1% formaldehyde (Sigma)in serum-free medium at room temperature for 5 minutes. Cells were resuspended in lysis buffer, and the DNA was fragmented by sonication to an average size of 500-1000 bp in length. Sheared chromatin extracts were precleared with protein G-agarose beads (Roche Molecular Biochemicals) at 4°C for 1 hour and were then incubated with either 50 μl of preimmune goat serum (Santa Cruz Biotechnology) or 10 μg of anti-Ku70 (M-19) or anti-Ku86 (C-20) goat polyclonal antibodies (Santa Cruz Biotechnology) or anti-clone162 (against the Ku70/86 heterodimer) mouse monoclonal antibody (NeoMarker). The protein-DNA immune complexes were then precipitated with 50 μl of protein G-agarose beads, washed thoroughly and eluted in 200 μl of extraction buffer (1% SDS in TE). Crosslinks were reversed overnight at 65°C, then incubated at 37°C for 2 hours with proteinase K (Roche Molecular Biochemicals). The extracted DNA was purified by QIAquick PCR purification columns (QIAGEN).
Real-time PCR quantification analysis of immunoprecipitated DNA
PCR reactions were carried out as described previously(Novac et al., 2002), using the LightCycler Instrument (Roche Molecular Biochemicals), with some modifications. The PCR reaction contained 1/20th of the immunoprecipitated material, 3 mM Mg2+ and 1 μM of each primer(oriβ F: 5′-ATCCTCCTAGCTCGGAGTCA-3′, 1534-1553 accession no. X94372; oriβ R: 5′-GGCTTATCTGCATCCTATTC-3′, 1677-1696 accession no. X94372, AF028017F: 5′-GAGCAGGTATAAGGGCCTTGG-3′,138-158 accession no. AF028017; AF028017R:5′-CGGTCTGGGTATGTTTAGCAAGAC-3′, 362-392 accession no. AF028017)using the LightCycler capillaries (Roche Molecular Biochemicals) and the QuantiTect SYBR Green PCR reaction mix (QIAGEN). These two primer sets amplify a 152 bp fragment from the DHFR oriβ origin of DNA replication(Kobayashi et al., 1998)(Fig. 5A) or a 255 bp fragment 17 kb downstream from the DHFR gene(Kobayashi et al., 1998).
Quantification of the PCR products was assessed by the LightCycler instrument (Roche Molecular Diagnostics) using SYBR Green I dye as detection format and LightCycler software version 3.5. An initial denaturation of 15 minutes at 95°C was followed by 40 cycles with denaturation for 15 seconds at 95°C, annealing at 55°C for 10 seconds and polymerization for 20 seconds at 72°C. Genomic CHO K1 DNA (24 ng, 28 ng, 72 ng and 96 ng) was used to build the standard curve necessary for the quantification of the PCR products (Fig. 4A). A melting curve analysis was also performed at the end of the PCR amplification to asses the specificity of the amplified product(Fig. 4B). The melting curve analysis cycle was composed of three segments: 95°C for 0 seconds(temperature transition rate of 20°C/second), 55°C for 30 seconds(temperature transition rate of 20°C/second) and a final segment at 95°C for 0 seconds (temperature transition rate of 0.2°C/second). The specificity of the 152 bp and 255 bp PCR products was also assessed by agarose gel electrophoresis and visualized with ethidium bromide(Fig. 4C,D).
Electrophoretic mobility-shift assay (EMSA)
Nuclear or cytoplasmic cell extracts (10 μg) were incubated with 0.5 ng of 32P-end-labeled A3/4 probe (5′CCTCAAATGGTCTCCAATTTTCCTTTGGCAAATTCC 3′) for 30 minutes on ice in the presence of 1 μg poly dI-dC (Amersham-Pharmacia), used as non-specific competitor, in a final volume of 20 μl including binding buffer (10 mM Tris-HCl pH 7.5, 80 mM NaCl, 1 mM EDTA, 10 mM 2-mercaptoethanol, 0.1% Triton X-100, 4% glycerol). The mixtures were electrophoresed on a 6% PAGE gel at 180 Volts in TBE (45 mM Tris-HCl pH 8.0, 45 mM Boric Acid, 1 mM EDTA), and the gel was dried and subjected to autoradiography.
For electrophoretic mobility-shift competition assays, 0.5 ng of 32P-labeled A3/4 was mixed with increasing molar excess amounts of A3/4 or non-specific (5′ TTCCGAATACCGCAAG 3′) cold competitor oligonucleotides. A 10 μg extract protein was then added, and the reaction was left to proceed as described above.
Electrophoretic mobility-supershift assay
Electrophoretic mobility shift mixtures were prepared as described above,and after the standard 30 minutes incubation, 1.5 μg of clone 162 antibody(Neomarkers, Union City, CA) or control goat IgG (Sigma, St Louis, MO) was added and further incubated for an additional 3 hours on ice. The samples were then applied to native 6% PAGE and electrophoresed, as described above.
In vivo replication activity of CHO K1 and xrs-5 cells
Previous studies have shown that xrs-5 mutants are defective in Ku86 (Taccioli et al., 1994; Getts and Stamato, 1994; Rathmell and Chu, 1994; Smider et al., 1994), a subunit of the Ku heterodimeric protein. This renders the mutants severely defective in the repair and recombination processes. Since Ku86 binds to origins of DNA replication and has been implicated in mammalian DNA replication (Ruiz et al.,1999; Novac et al.,2001), we tested the ability of xrs-5 cells to support the in vivo replication of p186, a pBR322-based plasmid that contains the minimal monkey cell origin of ors8(Todd et al., 1995). The p186 plasmid replicates autonomously in vivo, when transfected to mammalian cells and in a cell-free in vitro system that uses HeLa cell extracts. The replication of this plasmid initiates within the ors sequence, is semi-conservative, bi-directional and dependent on the replicative DNA polymerases α and δ (Pearson et al., 1991; Pearson et al.,1994; Todd et al.,1995). Furthermore, in vivo binding of Ku to this origin sequence was recently demonstrated, with a five-fold higher binding at the G1/S interphase than at G0 (Novac et al.,2001).
Supercoiled plasmid DNA from either p186 or pBR322 was transfected into either xrs-5 or CHO K1 cells, and the ability of the plasmids to undergo autonomous replication was assayed by the DpnI-resistance assay, to distinguish between input plasmid and plasmid replicated in the eukaryotic cells (Frappier and Zannis-Hadjopoulos, 1987; Landry and Zannis-Hadjopoulos,1991). p186 DNA was isolated at 72 hours post-transfection from both CHO K1 and xrs-5 cells and was digested with DpnI; the plasmids recovered from the xrs-5 cells yielded DpnI-resistant DNA, which transformed E. coli with an efficiency that was 45% lower than that DNA recovered from the CHO K1 cells(Fig. 1). No bacterial colonies were obtained with DpnI-digested DNA recovered from either the xrs-5 or CHO K1 cells that had been transfected with pBR322 (data not shown) as expected, since this plasmid does not contain a mammalian origin of DNA replication and thus is fully digested by DpnI. To verify that the decreased recovery of replicated plasmids in the xrs-5 cells was not due to the introduction of breaks during transfection and the requirement of NHEJ to repair those breaks prior to replication, undigested DNA from p186 and pBR322 recovered from xrs-5 and CHO K1 cells was also used to transform bacteria. No difference in total DNA recovered was found (data not shown), indicating that the observed effect with DpnI digestion(Fig. 1) is due to a replication effect.
Absence of Ku protein from xrs-5 mutant cells
To examine whether any residual Ku protein could be detected in xrs-5 cell extracts that might account for their observed in vitro DNA replication activity when using the total or cytoplasmic extracts, western immunoblot analyses were performed using anti-Ku antibodies(Fig. 2). An anti-human Ku86 antibody (C-20; Santa Cruz) was used that was able to recognize the hamster Ku86 protein, reacting with both the CHO K1 nuclear (K1 N) and cytoplasmic (K1 C) cell extracts (Fig. 2A,lanes 2 and 3). No material crossreacting with this antibody was detectable either in the cytoplasmic (xrs-5 C) or nuclear (xrs-5 N)cell extracts derived from the xrs-5 cells(Fig. 2A, lanes 4 and 5), which is in agreement with previous reports stating that xrs-5 cells lack Ku86 protein, shown by western blotting, and Ku86 RNA transcript, shown by northern blotting (Rathmell and Chu,1994; Singleton et al.,1997). However, the Ku86 transcript was detected in the xrs-5 mutants when the more sensitive technique of RT-PCR was used(Singleton et al., 1997). Each subunit of the Ku heterodimer is required to stabilize the other, and the absence of Ku86 in xrs-5 has been shown to result also in the loss of the Ku70 subunit (Taccioli et al.,1994; Smider et al.,1994; Singleton et al.,1997). To examine the levels of Ku70 in the xrs-5 cell extracts, the same membrane was probed with an anti-human Ku70 antibody (C-19;Santa Cruz) recognizing the Ku70 protein in the CHO K1 nuclear and cytoplasmic cell extracts (Fig. 2B, lanes 2 and 3). By contrast, xrs-5 showed dramatically decreased levels of Ku70 (Fig. 2B, lanes 4 and 5). Upon longer exposures, a faint band crossreacting with the anti-Ku70 antibody was detected in the xrs-5 cytoplasmic cell extracts.
Examination of the levels of replication protein in xrs-5 vs CHO K1
Western blot analyses were also performed to verify the level of other replication proteins in the mutant cell line and to compare it to the wild-type cells. Nuclear extracts from either CHO K1 or xrs-5 cells showed equivalent amounts of ORC2, PCNA, DNA polymerase ϵ, DNA polymeraseδ, Primase or Topoisomerase IIα proteins in both the mutant and wild-type cell lines (Fig. 3A-F).
Ku associates in vivo with the DHFR oriβ in CHO K1 cells but not in xrs-5 cells
To analyze whether the Ku protein associates with replication origins in Chinese hamster cells, as it was found to do in monkey (CV-1) cells(Novac et al., 2001), its interaction with the DHFR replication origin, oriβ(Fig. 5A), was analyzed in wild-type CHO K1 cells and compared to that in the Ku86 mutant cells(xrs-5) using a ChIP assay, as previously described(Novac et al., 2001). Genomic CHO K1 DNA was used to build the standard curve necessary for the quantification of the immunoprecipitated DNA using the LightCycler instrument(Fig. 4A). The LightCycler allows the quantification of the PCR products (considered background level)that are undetectable by agarose gel electrophoresis stained with ethidium bromide (Fig. 4C,D). To verify the specificity of the PCR products, a melting curve analysis was performed at the end of the amplification cycle, and no primer-dimers interfered with the quantification (Fig. 4B). In addition, PCR products were also separated on a 2% agarose gel to verify their size (Fig. 4C,D).
Anti-Ku70, anti-Ku86 and anti-clone 162 (directed against the Ku70/86 heterodimer) antibodies were used to immunoprecipitate in vivo protein-DNA complexes that had been crosslinked by treatment of the cells with formaldehyde (Novac et al.,2001). The isolated DNA was analyzed for an origin-containing-sequence using the quantitative real-time PCR approach(Novac et al., 2001; Novac et al., 2002; Ladenburger et al., 2002). In agreement with other ChIP analyses(Alexandrow et al., 2002; Ladenburger et al., 2002; Novac et al., 2001; Novac et al., 2002), the formaldehyde crosslinking approach used to study protein-DNA interactions in vivo largely prevents covalent binding of non-specific proteins to chromatin. With the CHO K1 cells, when preimmune goat serum (NGS) was used, the quantified DNA abundance in the origin-containing-sequence corresponded to background levels (3×103 molecules/2×108cells) (Fig. 5B). This is considered to be the non-specific DNA pulled down when performing the ChIP assay. Similarly, the DNA immunoprecipitated non-specifically with anti-clone162 antibody in non-crosslinked CHO K1 and xrs-5 cells was also quantified and corresponded to background levels(Fig. 5B). In addition, a primer set amplifying a fragment outside of the DHFR initiation zone (17 kb downstream from the DHFR gene) was also used as an additional control for non-specific DNA binding (Fig. 4A,B,D; Fig. 5B). The DNA immunoprecipitated from that region with anti-clone162, anti-Ku86 or anti-Ku70 antibodies from crosslinked CHO K1 cells was comparable to background levels (Fig. 5B). When the immunoprecipitation was performed with anti-Ku70, anti-Ku86 or anti-Ku70/86 antibodies in CHO K1 formaldehyde-crosslinked cells, the DNA abundance in the oriβ was approximately 13-fold, seven-fold or sixfold higher than background level, respectively(Fig. 5B). By contrast,anti-Ku70, anti-Ku86 or anti-Ku70/86 antibodies precipitated the oriβ DNA in the xrs-5 cells at the background level(Fig. 5B), as expected. Thus,Ku associates with the origin of DNA replication in vivo in CHO K1 cells but not in the xrs-5 mutant cells, where it is mutated.
A3/4 binding activity in CHO K1 and xrs-5 cell extracts
Electrophoretic mobility shift and super-shift assays were performed to examine whether there was an A3/4-specific binding activity present in the wild-type and mutant CHO cell extracts(Fig. 6A,B) that might account for the in vitro replication results. First, titration of the amount of protein and oligonucleotide allowed us to determine the optimum amounts to use in order to avoid multiple bindings of Ku molecules, producing a ladder effect. When HeLa cell extracts were reacted with radiolabeled A3/4, two main complexes arose (Fig. 6A, lane B). The slower migrating complex (*) results from the interaction of A3/4 with the Ku heterodimer, Ku70/Ku86, whereas the faster migrating complex (**) arises from its interaction with the truncated Ku,Ku70/Ku69. Ku69 is a truncated form of Ku86 that results from site-specific proteolytic cleavage by a leupeptin-sensitive protease(Quinn et al., 1993; Han et al., 1996; Jeng et al., 1999). The levels of the truncated complex vary depending on the leupeptin concentration; at concentrations above 5 μM the truncated complex is not formed(Jeng et al., 1999) (D.M.,O.N., G.B.P. and M.Z.-H., unpublished). This complex is not obtained with the CHO K1 or xrs-5 extracts owing to the higher leupeptin concentrations. However, a band immediately below the ** band was obtained, which is probably due to non-specific binding since this complex is also present with the control antibody. The migration of these complexes was further retarded by clone 162 antibody (Fig. 6A, lane C, ***), which recognizes the Ku70-Ku86 heterodimer, but not by the control antibody(Fig. 6A, lane D). The reaction of radiolabeled A3/4 oligonucleotide with the CHO K1 nuclear cell extracts yielded a complex of similar migration(Fig. 6A, lane E) to that obtained from its reaction with the HeLa cell extracts (compare with Fig. 6A, lane B). The complexes obtained with the CHO extracts were consistently fainter. This might be due to the CHO extracts having lower levels of Ku(Fig. 2) or because total HeLa extracts were used as opposed to either nuclear or cytoplasmic extracts used for the CHO extracts. This complex was also supershifted by the clone 162 antibody (Fig. 6A, lane F) but not by the control antibody (Fig. 6A, lane G). The EMSA reaction of radiolabeled A3/4 oligonucleotide with the xrs-5 nuclear cell extracts did not result in a Ku-A3/4 complex (Fig. 6A,lanes H, J), consistent with the results obtained by the western blot analysis, in which neither Ku70 nor Ku86 were detected (see Fig. 3). The EMSA reaction with xrs-5 cytoplasmic cell extracts(Fig. 6B) also resulted in a complex with similar migration to the HeLa(Fig. 6A, lane B) and CHO K1 nuclear (Fig. 6A, lane E) and CHO K1 cytoplasmic (Fig. 6B,lane A) cell extracts. This complex, however, was not supershifted by the clone 162 antibody (Fig. 6B,lane E), unlike the complex generated with the CHO K1 cytoplasmic cell extracts (Fig. 6B, lane B; not visible), suggesting that the epitope that is normally recognized by this antibody is either not present in the xrs-5 cytoplasmic extracts or it is not accessible. Faster migrating complexes, probably due to degradation,were also detected.
An A3/4-specific binding protein is present in CHO K1 and xrs-5 cytoplasmic cell extracts
The binding specificity for the A3/4 oligonucleotide in the complexes formed with the xrs-5 cytoplasmic cell extracts was tested by competition bandshift assays, using increasing molar excess of cold A3/4(Fig. 7, lanes 3, 4, 8, 9) as a specific competitor and cold pBR322-derived oligonucleotide(Fig. 7, lanes 5, 6, 10, 11),which was used as non-specific competitor. The A3/4-Ku complex formed with both the CHO K1 (Fig. 7, lane 7) and xrs-5 (Fig. 7,lane 2) cytoplasmic cell extracts was specifically competed with increasing concentrations of cold A3/4 but not with cold non-specific competitor,indicating that the complex formed with these extracts represents a specific protein interaction with A3/4. Again some faster migrating complexes were also detected, which were probably attributable to degradation products.
In vitro replication activity of CHO K1 and xrs-5 cell extracts
Since the in vivo autonomous replication assay showed that the xrs-5 cells were impaired in DNA replication, we tested their replication activity in the mammalian in vitro replication system, which allows the dissection and study of the proteins required for DNA replication(Diaz-Perez et al., 1996; Diaz-Perez et al., 1998; Matheos et al., 1998; Ruiz et al., 1999; Jilani et al., 1999; Novac et al., 2001; Novac et al., 2002; Matheos et al., 2002). The CHO cell extracts, both wild-type (K1) and mutant (xrs-5), yielded similar in vitro replication products (Fig. 8A) to those routinely obtained with the HeLa cell extracts(Pearson et al., 1991; Zannis-Hadjopoulos et al.,1994; Matheos et al.,1998), namely, relaxed circular (form II), linear (form III) and supercoiled (form I). However, in vitro replication of p186 DNA using the CHO cell extracts was consistently less efficient (approximately nine times) than with the HeLa cell extracts (data not shown). This is consistent with our previous findings and those of other laboratories, suggesting that HeLa cell extracts may be producing higher concentrations of initiator proteins,resulting in more efficient replication than that observed with CV-1 or COS-7 cell extracts (Pearson et al.,1991; Stillman and Gluzman,1985; Wobbe et al.,1985; Li and Kelly,1984; Guo et al.,1989). Other laboratories have also reported differences in in vitro replication activities of cell extracts, depending on their source(Krude, 2000; Stoeber et al., 1998).
We next performed in vitro DNA replication assays, using either cytoplasmic or nuclear extracts separately, or the two together, from either CHO K1 or xrs-5 cells (Fig. 8A,B). The CHO K1 cytoplasmic cell extracts(Fig. 8A, lanes 3 and 4)replicated the p186 DNA as efficiently as the CHO K1 total cell extracts(Fig. 8A, lanes 1 and 2),whereas the nuclear cell extracts alone(Fig. 8A, lanes 5 and 6) were approximately sixfold less efficient. The profiles obtained for both the total incorporation of radioactive precursor nucleotide (data not shown), indicating total incorporation owing to both replication and repair synthesis, and DpnI-resistance (Fig. 8A), indicating incorporation owing to replication alone, were similar for the two types of cell extracts. Furthermore, the quantification profiles of the DpnI-resistant bands, generated by the two reactions and corresponding to DNA forms II and III, were virtually the same(Fig. 8A). The xrs-5 cytoplasmic cell extracts (Fig. 8A, lanes 9 and 10) also showed comparable replication activity to the xrs-5 total (Fig. 8A, lanes 7 and 8) and the CHO K1 total(Fig. 8A, lanes 1 and 2) and cytoplasmic (Fig. 8A, lanes 3 and 4) cell extracts. By contrast, the xrs-5 nuclear cell extracts did not support the in vitro replication of the p186 DNA, as indicated by the failure to incorporate any radioactive precursor(Fig. 8A, lanes 11 and 12). When pBR322 was used as template DNA for the in vitro replication reaction, no DpnI-resistant products were obtained, although some incorporation of radionucleotides into the DNA with both the CHO K1 extracts (nuclear and cytoplamic) and the xrs-5 cytoplasmic extracts was seen (data not shown), owing to DNA repair, as also observed previously(Pearson et al., 1991).
OBA/Ku restores replication activity of xrs-5 nuclear cell extracts
To determine whether Ku could restore replication activity in the xrs-5 nuclear cell extracts, additional assays were performed where affinity-purified OBA/Ku was added to the in vitro replication reaction(Fig. 8C,D). The exogenous addition of affinity-purified OBA/Ku increased replication of the xrs-5 nuclear cell extracts from 0% to approximately 60%, relative to the replication activity of the CHO K1 nuclear extract(Fig. 8D). Hence, Ku was able to complement the lack of replication activity in the xrs-5 nuclear extract, which is similar to the way that Ku86 cDNA complements the defective repair and recombination phenotypes(Smider et al., 1994). No significant effect was observed when affinity-purified OBA/Ku was added to the CHO K1 nuclear cell extracts in the in vitro replication reaction(Fig. 2D, lanes 2 and 3; Fig. 2E).
In the present study, we have examined the replication activity of the xrs-5 cells, defective in the 86 kDa subunit of Ku antigen, a protein known to be involved in numerous cellular metabolic processes, including DNA replication. The in vivo replication results(Fig. 1, Fig. 5B) clearly demonstrate an involvement of Ku in DNA replication, whereby in the absence of the protein the replication level was reduced by almost half when compared to the control cells. Furthermore, Ku was found associated with the DHFR oriβorigin of DNA replication in CHO K1, but not in xrs-5 cells, in agreement with data previously obtained in vitro and in vivo with Ku86-/- cells (Novac et al., 2001) and with HeLa cells depleted of Ku antigen(Ruiz et al., 1999). The in vitro replication results with the xrs-5 nuclear extracts of the present study, showing a complete loss of replication activity(Fig. 8) that is restored upon addition of affinity-purified OBA/Ku protein(Fig. 8C,D), further supports a role for Ku in DNA replication. Surprisingly, xrs-5 cytoplasmic cell extracts (Fig. 8A,B) and xrs-5 total cell extracts (Fig. 8A,B) replicated the DNA as efficiently as the wild-type CHO K1 extracts. The discrepancy between the in vivo and the in vitro replication results with total extracts may be related to the absence of nuclear structure in vitro. An intact nuclear structure has been suggested to be required for the initiation of DNA replication in human(Krude et al., 1997), yeast(Pasero et al., 1997) and Xenopus (Wu et al.,1997; Dimitrova and Gilbert,2000) cells. Furthermore, nuclei from mimosine-arrested HeLa cells do not require the nuclear membrane for initiation of DNA replication in soluble cell extracts from proliferating human cells(Krude et al., 1997; Krude, 2000). One role of the nuclear structure may be in establishing initiation sites. When sperm chromatin or naked DNA is added to Xenopus egg extracts, DNA replication is initiated randomly and at many sites along the DNA, whereas, if intact nuclei are used, site-specific initiation occurs(Gilbert et al., 1995). It is possible that some Ku protein sequestered in the nuclear membrane of the xrs-5 cells (Yasui et al.,1999) is released into the cytoplasm during the extract preparation, thereby allowing the total and cytoplasmic xrs-5 cell extracts to replicate the DNA as efficiently as the wild-type cells. Alternatively, another protein, or a modified form of Ku antigen, may exist in the Ku-deficient cells, which would allow the cells or extracts to initiate replication. Such a protein with sequence-specific binding to the A3/4 sequence was apparently present in the xrs-5 cytoplasmic extracts(Fig. 6A,B, Fig. 7A). Ku is an origin-binding protein involved in DNA replication, and in its absence, other protein(s) with a similar function may be present, since (1) knockout mice are viable (Nussenzweig et al.,1996), (2) extracts from Ku86-/- cells or HeLa cells with depleted Ku antigen have a basal level of replication in vitro(Ruiz et al., 1999; Novac et al., 2001) and (3)the in vivo replication activity of xrs-5 cells is reduced by half(Fig. 1) but is not altogether abolished.
Upon examination of the Ku protein levels in the xrs-5 cell extracts by western blot analyses, it was found that xrs-5 cytoplasmic or nuclear cell extracts did not have any detectable levels of Ku70 or Ku86 proteins (Fig. 2A,B). This is in agreement with previous studies in which neither Ku86 protein nor Ku86 transcript were detected in xrs-5 cells or other cell lines of the XRCC5 group, by western and northern analyses,respectively (Rathmell and Chu,1994; Singleton et al.,1997; Errami et al.,1998). Since both Ku subunits are required for protein stability(Errami et al., 1996; Singleton et al., 1997) and since Ku86 probably regulates Ku70 levels post-transcriptionally(Chen et al., 1996), it is not surprising to find reduced Ku70 levels as a result of reduced Ku86 levels. However, studies with Ku knockout mice have demonstrated that inactivation of Ku70 resulted in a phenotype that was distinct from that obtained with Ku86-knockout mice, suggesting that Ku70 and Ku86 have functions that are independent of each other (Nussenzweig et al., 1996; Gu et al.,1997; Li et al.,1998) and that the remaining protein functions as a monomer or a homodimer (reviewed in Featherstone and Jackson, 1999). Low levels of wild-type Ku86 transcripts, however,have been detected in the xrs-5 cells by the more sensitive technique of RT-PCR (Singleton et al.,1997). The low level of wild-type transcript could be either due to the presence of some revertants in the xrs-5 cell population or to a low level of transcription in each cell. By contrast, Yasui et al. have recently reported the presence of both Ku70 and Ku86 proteins in the xrs-5 cells, albeit at low concentrations, using western blot and 2D gel electrophoresis analyses (Yasui et al., 1999). Furthermore, by indirect immunofluorescence, Ku70 and Ku86 were found distributed over both the cytoplasm and nuclei of CHO K1 and xrs-5 cells, with enhanced perinuclear localization of Ku86 in the xrs-5 cells (Yasui et al.,1999). The subcellular localization of Ku has been controversial,as discussed in Koike et al. (Koike et al., 1999). Reports of purely nuclear(Koike et al., 1999; Bakalkin et al., 1998),membrane (Dalziel et al.,1992), cytoplasmic (Bakalkin et al., 1998) and both nuclear and cytoplasmic(Fewell and Kuff, 1996)localization of Ku have been published (reviewed in Koike et al., 1999). The discrepancy is probably due to differences in the detection methods used or to a change in Ku's subcellular localization during the cell cycle. Ku has been shown to be present in the cytoplasm of mitotic cells and at the periphery of condensed chromosomes (Koike et al.,1999). The same investigators found Ku70 and Ku86 associated as a heterodimer throughout the cell cycle. However, recent evidence suggests that Ku70 and Ku86 may not always be dimerized, since the nuclear translocation of Ku70 precedes that of Ku86 at the late telophase/early G1 phase during the cell cycle (Koike et al.,1999). Although an enhanced perinuclear localization of only Ku86 in xrs-5 cells (Yasui et al.,1999) is surprising, the localization observed may be that of the truncated form of Ku86, which is stable when not complexed to Ku70(Singleton et al., 1997). Unlike the wild-type protein, truncated stable forms of Ku86 have been more frequently detected without Ku70, perhaps because of protein conformational changes or loss of degradation signal sequences(Singleton et al., 1997). Yasui et al. postulate that the concentration of Ku86 at the nuclear periphery of xrs-5 cells keeps the Ku complex sequestered, thereby preventing it from accessing the DNA (Yasui et al.,1999). In terms of DNA repair, this might prevent the Ku protein complex from accessing DNA DSBs. Recently, it was shown that each Ku subunit can translocate to the nucleus independently of its own NLS and that this translocation is dependent on its interaction with the other subunit(Koike et al., 2001). Furthermore, irradiation of cells resulted in an upregulation of the cellular level of Ku70 but not Ku86 (Brown et al.,2000). The inability of the xrs-5 cells to replicate the DNA efficiently in vivo and of the xrs-5 nuclear cell extracts to replicate DNA in vitro may be due to either a lack of or the presence of very low levels of Ku86. Alternatively, Ku86 might be localized in the nuclear periphery and thus be unavailable for interaction with the DNA and the replication complex. The replication of the xrs-5 cytoplasmic extract activity might be due to the presence of another protein with some affinity for the replication origin. Such a protein may also account for the observed in vivo replication activity of xrs-5 cells of 55%(Fig. 1). It is surprising that this protein is found in the cytoplasm, but this may reflect a more dynamic situation in the cells, as seen in the in vitro replication reaction. That is,this protein may translocate to the nucleus and bind to the origin at G1/S only and then dissociate and exit from the nucleus. Since the extracts used for the in vitro replication reaction are prepared from log phase cells, the majority of this protein would be present in the cytoplasm, with undetectable levels in the nucleus (Figs 6, 7 and 8). This would also explain why the in vivo association of Ku with the DHFR oriβ origin of DNA replication, in logarithmically growing xrs-5 cells, was comparable to background levels, whereas its association with the same origin of DNA replication was approximately ninefold higher in CHO K1 cells(Fig. 5B).
The lack of in vitro replication activity of the xrs-5 nuclear extracts was restored upon addition of affinity-purified OBA/Ku(Fig. 8C,D). The ability of affinity-purified OBA/Ku to rescue replication activity in the xrs-5 nuclear cell extracts further demonstrates that it plays a role in mammalian DNA replication. Addition of OBA/Ku to the CHO K1 nuclear cell extracts had no effect on replication (Fig. 8D,E), suggesting that the level of Ku in CHO K1 extracts was sufficient for optimum replication.
Reaction of the xrs-5 cytoplasmic cell extracts with radiolabeled A3/4 produced a protein-DNA complex with a similar migration to the Ku-A3/4 complex produced by the HeLa and CHO K1 cell extract(Fig. 6). Furthermore, this complex was specific, as determined by competition with cold A3/4 DNA(Fig. 7). However, the xrs-5 cytoplasmic cell extracts A3/4 complex was not recognized by the anti-Ku (clone 162) antibody, which recognizes a conformational epitope of the Ku70-Ku86 heterodimer (Fig. 6B) or by the individual anti-Ku70 or anti-Ku86 antibodies (data not shown), suggesting that this epitope is not present. This implies that the cytoplasmic protein that recognizes A3/4 is another origin-specific binding protein with similar affinities for A3/4, or a modified form of Ku86 that is able to complex with Ku70, albeit in a manner that is not recognized by the clone 162 antibody. Further studies are underway to address the nature of this protein.
Absence of the Ku86 protein, either in Ku86-/- mice or cells, results in hypersensitivity to ionizing radiation, defective DSB-repair pathways and lymphocyte development and early onset of an age-related phenotype including osteopenia, hepatocellular degeneration and shortened life span (Vogel et al., 1999; Nussenzweig et al., 1996; Gu et al., 1997). Furthermore,Ku86 knockout mice are half the size of their heterozygous littermates, their cells have prolonged doubling times in culture, owing to rapid loss of proliferating cells, and they exhibit replicative senescence(Nussenzweig et al., 1996; Gu et al., 1997). In vitro DNA replication using extracts prepared from Ku86-/- cells showed a 70% decrease in the replication level, compared to the control(Novac et al., 2001). Depletion of Ku from HeLa cells also resulted in inhibition of DNA replication to a level that was 10-20% of normal in vitro replication(Ruiz et al., 1999). The fact that the knockout mice are viable and DNA replication is not completely abolished when Ku is absent suggests that other mechanisms may take over in the absence of Ku. We are presently investigating these mechanisms in the xrs-5 and the Ku86-/- cells with the aim of further understanding the involvement of Ku antigen in DNA replication.
We thank Terry Chow (Montréal General Hospital, Montréal,Québec, Canada) for generously providing the xrs-5 cell line and Marcia T. Ruiz (McGill University Cancer Centre, Montréal,Québec, Canada) for providing A3/4 affinity-purified Ku. This work was supported by funds from the Canadian Institute of Health Research (M.Z.-H.)and the Cancer Research Society (G.B.P.). D.M. is a recipient of studentships from the Cancer Research Society, the McGill University Faculty of Medicine and the Défi Corporatif Canderel. O.N. is a recipient of studentship from the Fonds de Recherche en Sante du Quebec (FRSQ).