PARP is activated at stalled forks to mediate Mre11-dependent replication restart and recombination

Abstract

If replication forks are perturbed, a multifaceted response including several DNA repair and cell cycle checkpoint pathways is activated to ensure faithful DNA replication. Here, we show that poly(ADP-ribose) polymerase 1 (PARP1) binds to and is activated by stalled replication forks that contain small gaps. PARP1 collaborates with Mre11 to promote replication fork restart after release from replication blocks, most likely by recruiting Mre11 to the replication fork to promote resection of DNA. Both PARP1 and PARP2 are required for hydroxyurea-induced homologous recombination to promote cell survival after replication blocks. Together, our data suggest that PARP1 and PARP2 detect disrupted replication forks and attract Mre11 for end processing that is required for subsequent recombination repair and restart of replication forks.

Figure 1

PARP is activated at stalled replication forks and required for survival of HU-induced replication stalling. (A) Surviving fraction of AA8 hamster cells treated for 10 days with increasing doses of HU in the presence or absence of PARP inhibitors NU1025 (100 nM), 1,5-dihydroxyisoquinoline (ISQ; 0.6 mM) or 4-amino-1,8-NAP (100 μM). (B) Immunofluorescence staining for PAR in AA8 hamster cells treated for 24 h with or without 0.5 mM HU. DNA was counterstained with TO-PRO-3 iodide. Bar 10 μm. (C) Quantification of immunofluorescence staining above. Percentage of AA8 cells containing sites of PARP activity induced by a 24-h treatment with 0.5 mM HU. Differences are statistically significant (Student's t-test, P<0.05). (D) Western blot analysis of PAR (top), PARP1 (middle) and α-tubulin (bottom) in myc-PARP-expressing U2OS cells treated with combinations of 0.5 mM HU and 100 μM NAP. (E) PARP activity measured by the decrease of free NAD(P)H over time during incubation with 0.5 mM HU or 1 mM MMS.

Figure 2

PARP1 binds to and is activated by DNA fork structures in vitro. (A) Biotin-labelled stalled fork construct and (B) ligated construct, containing sealed DNA ends. (C) Early replication intermediate of the plasmid pBROTB535, containing replication forks stalled in vitro by omission of topoisomerase from the replication reaction. (D) Electrophoretic mobility shift assay using biotin-labelled artificial stalled fork substrate and increasing concentrations of purified PARP1 protein with or without a 10-fold excess of non-labelled competitor stalled fork substrate or ligated construct. (E) Western blot analysis of PARP1 (bottom) and PAR (top) after incubation of 50 ng purified PARP protein with 50 ng of different DNA substrates. Automodification reduces the electrophoretic mobility of PARP1, accounting for the decreased amounts of unmodified PARP1 protein detectable at its expected molecular size in these samples. (F) PARP1 activation by increasing length of gap within the stalled fork structure (A). Recombinant human PARP1 (5 nM) was incubated with DNA constructs and biotinylated NAD⁺ for the times indicated, and blots were probed with anti-biotin antibody. Sonicated DNA was used as positive control and plasmid DNA as negative control. (G) Quantification of PARP1 activation as in (F).

Figure 3

HU induces RPA foci that co-localise with sites of activated PARP. (A) HU-induced RPA foci in wild-type MEFs, representing ssDNA, co-localise with PAR polymers formed in response to 0.5 mM HU treatment for 24 h. DNA was counterstained with TO-PRO-3 iodide. The nuclei borders are marked with blue. (B) Percentages of PAR foci co-localising with RPA. Cells with at least 10 PAR foci with a diameter over 1 μm in an optical section of 1.5 μm taken in the middle of the cell were analysed for co-localisation. The mean and standard deviation from 10 cells are depicted. (C) PARP1 associates with stalled replication forks independently of PARP activity. Forks were isolated by CldU co-immunoprecipitation (co-IP) after a 3-h treatment with 0.5 mM HU and/or inhibition of PARP using 100 μM NAP. The level of histone H3 was used as loading control.

Figure 4

PARP activity is required for restart of stalled replication forks. (A) The alkaline DNA unwinding technique releases ³H-thymidine-labelled DNA onto the ssDNA fraction when replication elongation is inhibited. The speed of replication fork elongation is measured as the time required for ³H-thymidine-labelled DNA not to be released into the ssDNA fraction (Johansson et al, 2004). (B) Dose-dependent replication elongation inhibition in AA8 hamster cells after addition of HU. (C) Time course of replication fork progression in AA8 cells during HU or after HU treatment with/without PARP inhibitor (NAP, 50 μM). The means and standard errors (bars) of three independent experiments are shown.

Figure 5

PARP1 is required for replication restart as determined using the DNA fibre assay. DNA fibre analysis of replication fork restart in U2OS cells treated with PARP inhibitor NAP or depleted of PARP1. (A) Labelling protocols for DNA fibre analysis of replication forks. U2OS cells were pulse labelled with CldU, treated with HU for 2 h, and released into IdU. Example images of replication forks are shown. (B) Fork restart in the presence or absence of 100 μM NAP (left). Stalled replication forks are shown as percentage of CldU-labelled tracks. (C) Speed of restarting forks in the presence or absence of 100 μM NAP (right). IdU fork speeds are shown as percentage of CldU fork speeds. (D) Protein levels of PARP1 and β-actin (control) in U2OS cells after 48 h depletion with siRNA. (E) Fork restart in PARP1-depleted cells, as above (left). (F) Speed of restarting forks in PARP1-depleted cells, as above (right). The means and s.d. (bars) of three independent experiments are shown. Values marked with asterisks are significantly different from control (^*P<0.05 or ^**P<0.01).

Figure 6

PARP is required for Mre11 localisation and resection at stalled replication forks. (A) Co-immunoprecipitation in U2OS cells showing proteins interacting with PARP1 in presence and absence of 0.5 mM HU and 100 μM PARP inhibitor NAP. (B) Immunofluorescence staining for PAR polymers and Mre11 protein in U2OS cells treated with 0.5 mM HU for 24 h. DNA was counterstained with TO-PRO-3 iodide. The nuclei borders are marked with blue. The close up panel shows Mre11 foci that co-localise with PAR (labelled C) those that do not co-localise (labelled N) and Mre11 foci with a diameter <0.5 μm. Bar is 0.5 μm. (C) Percentages of Mre11 foci co-localising with PAR. Cells with at least 10 Mre11 foci with a diameter over 0.5 μm in an optical section of 1.5 μm taken in the middle of the cell were analysed for co-localisation. The mean and standard deviation from 22 cells are depicted. (D) Quantification of Mre11 foci in U2OS cells induced by 0.5 mM HU in the presence or absence of PARP inhibitor. The means and s.d. (bars) of three experiments are shown. Values marked with asterisks are significantly different (Student's t-test, P<0.01). (E) Quantification of large RPA foci induced by 0.5 mM HU in PARP^+/+ (A19) and PARP^−/− (A11) MEFs. The means and s.d. (bars) of five experiments are shown. Values marked with asterisks are significantly different (Student's t-test, P<0.001). (F) Quantification of RPA foci in U2OS cells induced by 0.5 mM HU in the presence or absence of PARP inhibitor. The means and s.d. (bars) of three experiments are shown. Values marked with asterisks are significantly different (Student's t-test, P<0.01).

Figure 7

PARP1 exerts an effect in the same pathway as Mre11 to restart stalled replication forks. DNA fibre analysis of replication fork restart in U2OS cells depleted of PARP1, Mre11 or co-depleted of PARP1 and Mre11. Cells were labelled to analyse fork restart after 2 h HU as in Figure 5. (A) Western blot analysis of PARP1, Mre11 and β-actin in siRNA-treated U2OS cells. (B) Representative images of DNA fibre tracks. (C) Fork restart in PARP1- and/or Mre11-depleted cells. (D) Speed of restarting forks in PARP1- and/or Mre11-depleted cells. IdU fork speeds are shown as percentage of CldU fork speeds. Means and s.d. of three independent experiments are shown. Values marked with asterisks are significantly different from control (^*P<0.05 or ^**P<0.01).

Figure 8

PARP1 and PARP2 are required for RAD51 foci formation and HR induced at stalled replication forks. (A) Recombination frequency in the reporter construct SCneo in PARP1 and/or PARP2 depleted SW480SN.3 cells treated with 0.5 mM HU for 24 h. Means and s.d. of three independent experiments are shown. Values marked with asterisks are significantly different from control (^*P<0.05 or ^**P<0.01). (B) Recombination frequency measured in the hprt gene in SPD8 hamster cells after a 24-h treatment with 0.5 mM HU with/without PARP inhibitors NU1025 (100 nM), 1,5-dihydroxyisoquinoline (ISQ-0.6 mM) or 4-amino-1,8 NAP (100 μM). (C) Rad51 foci formation in AA8 hamster cells induced by a 24-h HU treatment (0.5 mM) with/without PARP inhibitors NU1025 (100 nM), ISQ (0.6 mM) or NAP (100 μM). (D) Rad51 foci formation induced by 0.5 mM HU treatment in PARP^+/+ and PARP^−/− MEFs. The means (symbols) and standard deviations (error bars) from at least three independent experiments are depicted.

Figure 9

PARP is activated and required for repair of replication forks stalled after dT treatments. (A) Percentage of AA8 Chinese hamster cells containing sites of PARP activity induced by a 24-h dT (2 mM) treatment. (B) PARP activity measured by the decrease of free NAD(P)H over time during incubation with 10 mM dT or 1 mM MMS. (C) Survival fraction of SW480SN.3 cells depleted of various PARP proteins after treatment for 24 h with increasing doses of the replication inhibitor dT. P<0.001 for PARP1, PARP2 and PARP1+2 as compared with scramble control. (D) Survival fraction of A11 (PARP^−/−) and A19 (PARP^+/+) after continuous exposure to increasing doses of dT. (E) Recombination frequency measured in the hprt gene in SPD8 hamster cells induced by a 24-h treatment with 10 mM dT with/without PARP inhibitor 1,5-dihydroxyisoquinoline (ISQ-0.6 mM). (F) Recombination frequency in the reporter construct SCneo within SW480SN.3 cells after 24 h treatment with 0.5 mM HU, depleted with PARP1 and/or PARP2 siRNAs. The means (symbol) and standard deviations (error bar) from at least three experiments are depicted.

Figure 10

Model for PARP-mediated repair of DNA SSBs and stalled replication forks. PARP1 rapidly binds to nicked, gapped or broken DNA regions, for example DNA SSB or stalled or collapsed replication forks, which activates the enzyme. The detection of stalled or collapsed replication forks is likely to involve activation of both PARP1 and PARP2. The PNKP protein and Mre11–RAD50–Nbs1 (MRN) complex are recruited to SSBs and stalled replication forks, respectively, to initiate end processing. Subsequent repair of SSBs includes XRCC1-ligase3, polymerase β, aprataxin (APTX) and the additional factors PCNA, polymerases δ/ɛ, FEN1 and ligase 1 in the case of long patch repair (Caldecott, 2008). The machinery involved in repair and replication restart in mammalian cells is poorly investigated, but likely to include the RAD51 protein and HR to reactivate normal PCNA-mediated replication by polymerases δ/ɛ.