The function of protein phosphatase 1 nuclear‐targeting subunit (PNUTS)—one of the most abundant nuclear‐targeting subunits of protein phosphatase 1 (PP1c)—remains largely uncharacterized. We show that PNUTS depletion by small interfering RNA activates a G2 checkpoint in unperturbed cells and prolongs G2 checkpoint and Chk1 activation after ionizing‐radiation‐induced DNA damage. Overexpression of PNUTS–enhanced green fluorescent protein (EGFP)—which is rapidly and transiently recruited at DNA damage sites—inhibits G2 arrest. Finally, γH2AX, p53‐binding protein 1, replication protein A and Rad51 foci are present for a prolonged period and clonogenic survival is decreased in PNUTS‐depleted cells after ionizing radiation treatment. We identify the PP1c regulatory subunit PNUTS as a new and integral component of the DNA damage response involved in DNA repair.
Protein phosphatase 1 nuclear‐targeting subunit (PNUTS, also known as p99 and R111) is predominantly localized to the interphase nucleus of mammalian cells (Jagiello et al, 1995; Kreivi et al, 1997; Allen et al, 1998; Landsverk et al, 2005). It binds to the α‐ and γ‐isoforms of the catalytic subunit of PP1 (PP1c; Jagiello et al, 1995; Allen et al, 1998; Trinkle‐Mulcahy et al, 2006) through a docking K397SVTW401 (RVXF) motif. A W401A mutation or phosphorylation by protein kinase A within the RVXF motif disrupts binding to PP1c in vitro (Kreivi et al, 1997; Kim et al, 2003).
The function of PNUTS is largely uncharacterized. The upregulation of PNUTS occurs due to hypoxia, and overexpression or depletion of PNUTS causes apoptosis (Kim et al, 2003; Lee et al, 2007). In response to hypoxia (Udho et al, 2002) or chemotherapeutic agents (Krucher et al, 2006), PNUTS dissociates from PP1c, an event that correlates with retinoblastoma protein dephosphorylation and G1 arrest. Recently, PNUTS was shown to interact with the telomeric protein telomere repeat factor 2 (Kim et al, 2009). As several telomere repeat factor 2‐binding proteins are involved in DNA damage repair, PNUTS has been suggested to function in the DNA damage response (DDR).
The DDR coordinates DNA repair, cell cycle arrest and—depending on the severity of the insult—cell death after exposure to double‐strand breaks (DSBs). It is essential for genome integrity and has been proposed to be an early barrier against cancer progression (Jackson & Bartek, 2009). A key signalling cascade in DDR is the ataxia telangiectasia and Rad3‐related/Chk1 pathway that controls the S and G2 checkpoints and homologous recombination repair (Liu et al, 2000; Sorensen et al, 2005). A role for protein phosphatases, including PP1c, in DDR has recently emerged, with reports that PP1c is involved in DSB repair, cell survival and the DNA damage G2 checkpoint (den Elzen & O’Connell, 2004; Tang et al, 2008; Hamilton et al, 2009; Peng et al, 2010).
We have previously shown that PNUTS enhances mitotic chromosome decondensation in a cell‐free system, suggesting that it has a role in the regulation of chromosome structure (Landsverk et al, 2005). In this study, we examined the effect of PNUTS depletion on cell cycle progression through mitosis. We show a prolonged mitotic prophase caused by a checkpoint‐dependent delay at the G2–M transition. Furthermore, we show an involvement of PNUTS in DDR, with a prolonged ionizing radiation (IR)‐induced G2 checkpoint arrest, rapid and transient targeting of PNUTS to damage sites, prolonged presence of DNA damage markers and decreased cell survival after IR treatment. Our results indicate a role for PNUTS in the regulation of DNA repair.
Results And Discussion
Our earlier results suggested a role for PNUTS in the regulation of chromosome structure at mitosis (Landsverk et al, 2005). Consequently, we used time‐lapse imaging of HeLa cells stably expressing histone H2B–enhanced green fluorescent protein (EGFP) to monitor the duration of mitotic phases after small interfering RNA (siRNA)‐mediated PNUTS depletion. A marked lengthening of prophase—defined morphologically as the stage between the first visible signs of chromosome condensation and nuclear envelope breakdown—was observed, compared with controls. In PNUTS‐depleted compared with control cells, average prophase duration increased threefold (60.1±38.4 and 20.1±6.8 min, respectively, P<0.0001; Fig 1A,B; Table 1) and frequency distribution of prophase duration was much broader (median values of 48 min and 18 min, respectively; Fig 1C). To quantify chromatin condensation, we randomly selected subsets of cells for computerized image processing to measure the heterogeneity in the spatial distribution of fluorescently labelled chromatin (supplementary information online). Chromatin in PNUTS‐depleted cells showed signs of condensation as early as 2.5 h before nuclear envelope breakdown, compared with 0.5 h in control cells (Fig 1D).
Prolonged prophase is characteristic of cells arrested at the G2–M transition (Pines & Rieder, 2001). To test this, we monitored mitotic entry after 1 h incubation with the G2 checkpoint inhibitor caffeine and the Chk1‐inhibitor SB218078 (Jin et al, 2005). Remarkably, more PNUTS‐depleted than control cells entered mitosis in the presence of caffeine or SB218078 (Fig 1E,F). Similar results were observed with an independent PNUTS siRNA oligonucleotide (supplementary Fig S1A,D online). Thus, PNUTS depletion activates a caffeine‐ and SB218078‐sensitive G2 checkpoint in HeLa cells.
The above results led us to ask whether PNUTS might also be involved in the DNA‐damage‐induced G2 checkpoint. To assess this, we exposed siRNA‐treated cells to IR treatment and monitored G2 accumulation after 24 h. In agreement with a previous study (De Leon et al, 2008), PNUTS‐depleted and control cells showed similar cell cycle profiles in the absence of IR treatment (Fig 2A). However, IR‐treated PNUTS‐depleted cells showed enhanced accumulation in G2, compared with control cells (Fig 2A). This was accompanied by increased phosphorylation of Chk1 at Ser 317 and Ser 345 and of replication protein A (RPA; Fig 2B), consistent with increased activation of ataxia telangiectasia and Rad3‐related/Chk1 in PNUTS‐depleted cells. Similar effects were seen with an independent PNUTS siRNA oligonucleotide (supplementary Fig S1B,C online). Conversely, overexpression of PNUTS–EGFP inhibited accumulation in G2 phase (Fig 2C). To verify that the IR‐treated PNUTS–EGFP‐expressing cells were dividing, we added nocodazole to trap cells in mitosis (Fig 2C). The G2 arrest was quantified as the ratio of cells in G2–M after IR treatment relative to IR+noco (averages were 37.8% for PNUTS–EGFP‐expressing and 74.6% for GFP‐negative cells; Fig 2D).
The decreased G2 accumulation in PNUTS–EGFP‐expressing cells after IR treatment is reminiscent of that observed in G2‐checkpoint‐challenged HeLa cells expressing EGFP–PP1α (den Elzen et al, 2004), suggesting a role for PNUTS in the DDR through activation of PP1c. To investigate this we expressed the PP1c‐docking mutant PNUTS(W401A) (Kreivi et al, 1997; Kim et al, 2003) tagged to EGFP. First, we verified that PNUTS(W401A)–red fluorescent protein could not delocalize PP1γ–EGFP, even when it was expressed at high levels (supplementary Fig S2 online), confirming that it is defective for PP1 targeting. PNUTS(W401A)–EGFP‐expressing cells also showed a reduced G2 arrest 20 h after IR treatment (P<0.01, Fig 2C,D), indicating that inhibition of G2 arrest by PNUTS overexpression is not affected by disruption of PP1c binding and is also not due to the delocalization of PP1c by PNUTS. This suggests that the role of PNUTS in the DDR is not restricted to the activation of PP1c.
As several proteins involved in the DDR are recruited at sites of DSBs, we assessed whether PNUTS is recruited at DNA damage sites. Although PNUTS–EGFP did not show visible recruitment at DNA damage sites after IR treatment (supplementary Fig S3C online), we observed a transient recruitment of PNUTS–EGFP at DNA damage sites induced by 405‐nm microirradiation (Fig 3; supplementary Fig S3A,B online) in Hoechst 33258‐sensitized cells (Ayoub et al, 2008). The PNUTS–EGFP fusion protein was recruited within a time duration of seconds to DNA damage sites and returned to original levels after approximately 5 min (Fig 3). Maximal intensity compared with initial intensity was observed within 1 min and was 1.3‐fold for PNUTS–EGFP and 1.5‐fold for PNUTS(W401A)–EGFP (Fig 3A,C,G; Table 2). The slightly stronger and prolonged recruitment of PNUTS(W401A)–EGFP could suggest that PP1c binding negatively influences the recruitment of PNUTS and participates in the removal of PNUTS from DNA damage sites. Notably, PNUTS–EGFP intensity returned to initial levels before maximal recruitment of GFP–53BP1 (Fig 3A,F,H; supplementary Fig S3B online; supplementary Movie S1 online). Fluorescence recovery after photobleaching experiments using a 488‐nm laser for irradiation—which does not excite Hoechst 33258 or create DNA damage—did not result in PNUTS–EGFP recruitment, suggesting that recruitment was due to DNA damage (Fig 3B,D,G). Furthermore, a fluorescently tagged construct of another nuclear‐targeting subunit of PP1c, nuclear inhibitor of PP1 (NIPP1)–EGFP, was not recruited (Fig 3E,H), demonstrating specific recruitment of PNUTS. Interestingly, the nonhomologous end‐joining proteins Ku70, Ku80 and DNA‐dependent protein kinase are also visibly recruited to laser‐induced DNA damage, but do not form foci after IR treatment (Kim et al, 2005; Bekker‐Jensen et al, 2006; Mari et al, 2006). One possibility is that PNUTS, similarly to these proteins, binds to DSBs or other DNA lesions at low abundance and this recruitment is therefore only visible after high amounts of localized DNA damage, as achieved by laser irradiation.
The above results are consistent with a role for PNUTS in the regulation of DNA repair, checkpoint signalling or both. To clarify this, we examined the DSB marker γH2AX. From analysis by western blotting, PNUTS‐depleted and control cells showed similar γH2AX levels 2 h after IR treatment; however, γH2AX levels were higher in PNUTS‐depleted cells 24 h after IR treatment (Fig 4A; supplementary Fig S1C online). Similar results were obtained from flow cytometry analysis 24 h after IR treatment for two independent PNUTS siRNA oligonucleotides (Fig 4B). Immunofluorescence analysis revealed that although all PNUTS‐depleted and control cells contained more than 10 γH2AX foci 1 h after IR treatment, 74.2% of PNUTS‐depleted compared with 26.1% of control cells contained more than 10 γH2AX foci 24 h after IR treatment (P<0.05; Fig 4C,D). Thus, removal of γH2AX is delayed in PNUTS‐depleted cells, suggesting that PNUTS is involved in the regulation of DNA repair or participates in dephosphorylation of γH2AX.
To address whether PNUTS controls dephosphorylation of γH2AX, we analysed γH2AX levels in exponentially growing and IR‐treated cells after overexpression of PNUTS–EGFP. As PNUTS overexpression did not result in a decrease in γH2AX levels (Fig 4E), the PNUTS–PP1c holoenzyme probably does not dephosphorylate γH2AX in vivo, although PP1c does so in vitro (Siino et al, 2002). We also analysed 53BP1 foci, a different marker for DSBs (Schultz et al, 2000). Consistent with delayed repair, PNUTS‐depleted cells showed more 53BP1 foci 24 h after IR treatment than control cells (Fig 4F). PNUTS‐depleted cells also showed increased Rad51 and RPA foci at 5 and 24 h after IR treatment (Fig 4F; data not shown), consistent with increased homologous recombination repair (Sorensen et al, 2005) and presence of single‐strand DNA (Jackson & Bartek, 2009). These data strongly suggest that PNUTS regulates DNA repair rather than γH2AX dephosphorylation.
Finally, we reasoned that if PNUTS regulates DNA repair, PNUTS‐depleted cells should be hypersensitive to IR treatment in cell survival assays. The PNUTS siRNA oligonucleotides with PNUTS almost completely downregulated (siPNUTS 1; supplementary Fig S1D online) were unsuitable for clonogenic survival assays, as they showed reduced cloning efficiency, even in the absence of IR treatment (<10% cloning efficiency compared with control transfected cells). However, an increase in IR‐induced death could be observed by measuring the percentage of sub‐G1 cells in DNA histograms 72 h after 5‐Gy IR treatment (supplementary Fig S4 online). More conclusively, a different PNUTS siRNA oligonucleotide (siPNUTS 3)—which resulted in only partial downregulation of PNUTS—showed decreased clonogenic survival after IR treatment in HeLa cells (Fig 4G). Thus, and consistent with defective DNA repair, PNUTS‐depleted cells are hypersensitive to IR‐induced cell death.
Specificity is provided to PP1c—a protein phosphatase with a broad range of substrates—through association with regulatory subunits that target the holoenzyme to specific locations or substrates and control its activity (Moorhead et al, 2007; Bollen et al, 2009). Although PP1c has previously been implicated in the DDR (den Elzen et al, 2004; Tang et al, 2008; Hamilton et al, 2009), we identify here for the first time, to the best of our knowledge, that the regulatory subunit of PP1 PNUTS is an integral component of this pathway. The rapid and transient recruitment of PNUTS–EGFP to DSBs together with the long‐term effects of PNUTS knockdown on DNA damage markers, Chk1 activation, G2 arrest and cell survival after IR treatment suggests that PNUTS regulates early events in the DDR, which affects major downstream targets involved in DNA repair.
Cell culture, irradiation and drugs. Human epithelial HeLa, HeLaEGFP–PP1γ (Trinkle‐Mulcahy et al, 2003) and HeLaH2B–EGFP (Hirota et al, 2004) were grown in Dulbecco's Modified Eagle Medium containing 5% fetal calf serum (Invitrogen). Cells were irradiated using a 137Cs source at a dose rate of 4.3 Gy/min or with an X‐ray generator (Faxitron CP160, 160 kV, 6.3 mA) at a dose rate of 1 Gy/min. Caffeine and nocodazole were purchased from Sigma‐Aldrich and SB218078 from Calbiochem.
DNA constructs and cloning. Wild‐type PNUTS was cloned from HeLa complementary DNA into the enhanced green and monomeric red fluorescent protein vectors (Clontech) without the seven carboxy‐terminal amino acids of PNUTS. The PNUTS(W401A)–EGFP mutant was produced by site‐directed mutagenesis (Quickchange II, Stratagene). Mouse GFP–53BP1 was obtained from J. Lukas (Bekker‐Jensen et al, 2006) and NIPP1–EGFP from L. Trinkle‐Mulcahy (Trinkle‐Mulcahy et al, 1999).
Western blotting, immunofluorescence and flow cytometry. Western blotting and immunofluorescence were performed as described previously (Landsverk et al, 2005). PNUTS antibody was provided by M. Bollen (Lesage et al, 2004), phospho‐Chk1 (Ser 317 and Ser 345) and phospho‐H3 (pH3S10) were obtained from Cell Signaling Technology, phospho‐H2AX (γH2AX clone JBW301) from Millipore, RPA (p34) from AH Diagnostics and γ‐tubulin (GTU‐88) from Sigma. Chk1 antibody (DCS310) has been described previously (Sorensen et al, 2005). Flow cytometry with double labelling was performed as described previously (Naderi et al, 2005) omitting the acid/pepsin and neutralization steps, and triple labelling is described in the supplementary information online.
Live cell imaging and laser irradiation. Procedures are described in the supplementary information online.
Clonogenic survival assay. Twenty‐four hours after siRNA transfection, between 150 and 1,200 cells (depending on IR dose to yield 50–100 colonies per dish) were seeded on to 6‐cm dishes, incubated for about 20 h, and treated with IR (0, 2, 4 and 6 Gy). After 14 days, cells were stained with methylene blue and colonies containing more than 50 cells were counted as survivors. Survival fractions after each IR dose were calculated as the average cloning efficiency ratio (from three parallel dishes in each experiment) of irradiated cells to non‐irradiated cells.
Statistics. Statistical analysis was performed using unpaired Student's t‐test and all experiments were performed three times or more, unless otherwise indicated. Error bars represent standard deviation values.
Supplementary information is available at EMBO reports online (http://www.emboreports.org).
Conflict of Interest
The authors declare that they have no conflict of interest.
Supplementary Movie S1
We thank L. Trinkle‐Mulcahy for HeLaPP1γ‐EGFP cells, the NIPP1–EGFP construct and input; M. Beullens and M. Bollen for PNUTS antibodies and input; and J. Lukas for the green fluorescent protein–p53‐binding protein 1 construct. We acknowledge the use of the Norwegian Molecular Imaging Consortium (NORMIC‐UIO) live‐imaging facilities at the Department of Molecular Biosciences. This study was supported by the Norwegian Cancer Society, the Research Council of Norway and the Harald Andersens foundation.
- Copyright © 2010 European Molecular Biology Organization