The Importance of Poly(ADP-Ribose) Polymerase as a Sensor of Unligated Okazaki Fragments during DNA Replication

Abstract

Poly(ADP-ribose) is synthesized by PARP enzymes during the repair of stochastic DNA breaks. Surprisingly, however, we show that most if not all endogenous poly(ADP-ribose) is detected in normal S phase cells at sites of DNA replication. This S phase poly(ADP-ribose) does not result from damaged or misincorporated nucleotides or from DNA replication stress. Rather, perturbation of the DNA replication proteins LIG1 or FEN1 increases S phase poly(ADP-ribose) more than 10-fold, implicating unligated Okazaki fragments as the source of S phase PARP activity. Indeed, S phase PARP activity is ablated by suppressing Okazaki fragment formation with emetine, a DNA replication inhibitor that selectively inhibits lagging strand synthesis. Importantly, PARP activation during DNA replication recruits the single-strand break repair protein XRCC1, and human cells lacking PARP activity and/or XRCC1 are hypersensitive to FEN1 perturbation. Collectively, our data indicate that PARP1 is a sensor of unligated Okazaki fragments during DNA replication and facilitates their repair.

Keywords: DNA repair; DNA replication; DNA strand break; PARP1; postreplication repair.

Graphical abstract

Figure 1

Endogenous Poly(ADP-Ribose) Is Detected Primarily during S Phase at Sites of DNA Replication (A) ADP-ribose and PCNA (indicative of S phase) immunostaining in detergent-pre-extracted U2OS cells after 30 min incubation with DMSO vehicle or PARG inhibitor (PARGi). Scale bars, 20 μm. (B) ADP-ribose and PCNA immunostaining in wild-type, PARP1^−/−, and PARP1^−/−/PARP2^−/− RPE-1 cells after 15 min incubation with DMSO vehicle or PARG inhibitor. Representative confocal images are shown. Scale bars, 5 μm. (C) Western blotting of the indicated proteins in wild-type (WT), PARP1^−/−, PARP2^−/−, PARP3^−/−, and PARP1^−/−/PARP2^−/− RPE-1 cell lines (left) and quantification of ADP-ribose levels in these cell lines after 15 min incubation with DMSO vehicle or PARG inhibitor in PCNA-negative (non-S phase) and PCNA-positive (S phase) cells (average of n = 4 with SEM). Representative ScanR images are shown in Figure S1B.

Figure 2

S Phase Poly(ADP-Ribose) Does Not Result from DNA Lesions or Replication Fork Stress (A) Representative ScanR images (left) and quantitation (right) of ADP-ribose in RPE-1 cells incubated for 20 min with 10 μM EdU in the absence or presence of either PARG inhibitor or MMS. Cell cycle populations were gated according to EdU positivity (S phase) and DNA content (G1 and G2) by DAPI staining (average of n = 3 with SEM). (B) Representative ScanR images and quantitation of ADP-ribose in wild-type and XRCC1^−/− RPE-1 cells as in (A) (average of n = 3 with SEM). (C) AP endonuclease protein (bottom left) and activity (top left) in cell extracts from wild-type and APE1 gene-targeted human HAP1 cells additionally transfected with APE1 siRNA (denoted APE1^KD) to further decrease APE1 levels and activity. Right: ADP-ribose immunostaining in pre-extracted untreated wild-type and APE1^KD cells and in cells incubated for 20 min with either PARG inhibitor or MMS. Scale bars, 20 μm. The numbers in the corners are the mean ADP-ribose intensity in all nuclei normalized to the wild-type sample, quantified in ImageJ. (D) Quantification of ADP-ribose in MMR-deficient (MSH3 and MLH1 mutant) HCT116 cells and their chromosome-complemented MMR-proficient counterparts HCT116+Ch3 (MLH1-complemented) and HCT116+Ch3+5 (MLH1- and MSH3-complemented) after 60 min incubation with or without PARG inhibitor. Cell populations were gated according to PCNA (S phase) intensity (average of n = 3 with SEM). Representative immunofluorescence images are shown in Figure S2B. (E) Quantification of ADP-ribose in PCNA-negative (non-S phase) and PCNA-positive (S phase) Rnaseh2b^+/+ and Rnaseh2b^−/− mouse embryonic fibroblasts (MEFs) after incubation for 60 min with or without PARG inhibitor (average of n = 3 with SEM). Representative ScanR images are shown in Figure S2C. (F) Representative confocal images of ADP-ribose and γH2AX immunostaining in untreated RPE-1 cells and in RPE-1 cells following incubation with or without hydroxyurea (HU) for 2 hr and with or without PARG inhibitor for the final 20 min, as indicated. Scale bars, 20 μm. Insets, right: a representative and magnified cell from each image.

Figure 3

Perturbation of the DNA Replication Proteins LIG1 and FEN1 Increases S Phase Poly(ADP-Ribose) (A) Representative ScanR images (left, PCNA-positive cells only) and quantitation of ADP-ribose (right) in 1BR- and LIG1-deficient 46BR primary fibroblasts following incubation with DMSO vehicle or PARG inhibitor for 20 min. For quantitation, PCNA-negative (non-S phase) and PCNA-positive (S phase) cells were gated according to nuclear PCNA intensity. Note the break and change in scale in the y axis required to display the very high ADP-ribose level in S phase 46BR cells (average of n = 3 with SEM). (B) Representative ScanR images (left, PCNA-positive cells only) and quantification of ADP-ribose (right) as above in wild-type, PARP1^−/−, and PARP1^−/−/PARP2^−/− RPE-1 cell lines. Cells were treated with DMSO vehicle or FEN1 inhibitor (FEN1i) for 30 min, with PARG inhibitor added or not during the last 15 min, as indicated. Note the break and change in scale in the y axis required to display the very high ADP-ribose level in FEN1 inhibitor/PARG inhibitor-treated RPE-1 cells (average of n = 3 with SEM). See also Figure S5. (C) Representative confocal images of S phase (PCNA-positive) cells from the experiment in (B), illustrating the localization of poly(ADP-ribose) with PCNA at DNA replication sites. Scale bars, 5 μm.

Figure 4

Suppression of Okazaki Fragment Formation with Emetine Prevents S Phase ADP-Ribosylation (A) Indirect immunofluorescence imaging of ADP-ribose and EdU in wild-type RPE-1 cells incubated or not with emetine (EME) and/or FEN1 inhibitor for 45 min as indicated, with or without PARG inhibitor added during the final 20 min. EdU was added to all samples during the last 20 min to detect DNA synthesis. Scale bar, 20 μm. (B) Representative ScanR images (left) and quantification (right) of mean ADP-ribose levels in pre-extracted U2OS interphase cells treated or not for 2 hr with hydroxyurea or for 1 hr with emetine, as indicated (average of n = 3 with SEM). PARG inhibitor was added or not, as indicated, during the final 30 min. Note that the ADP-ribose quantifications are the mean levels across all interphase cells (not just S phase cells). (C) EdU labeling in U2OS cells treated or not with hydroxyurea for 2 hr or emetine for 1 hr. Cells were incubated with 10 μM EdU for the final 20 min. Scale bar, 20 μm.

Figure 5

S Phase PARP Activity Recruits the DNA Repair Protein XRCC1 and Protects Cells from the Effect of the FEN1 Inhibitor (A) Representative confocal images of XRCC1 immunostaining in U2OS cells after incubation for 30 min in the presence of DMSO vehicle or PARG inhibitor. Cells were fixed immediately to detect total XRCC1 or were pre-extracted with Triton X-100 before fixation to detect chromatin-bound XRCC1. Magnified images of two representative cells are shown (right). Scale bars, 20 μm. (B) Representative confocal images of XRCC1 and PCNA immunostaining in PCNA-positive (S phase) wild-type, PARP1^−/−, PARP1^−/−/PARP2^−/−, and XRCC1^−/− RPE-1 cells after incubation in the presence or absence of PARG inhibitor for 15 min (top), as indicated, or in wild-type RPE-1 cells following incubation in the presence of FEN1 inhibitor for 30 min (bottom). Scale bars, 5 μm. (C) Clonogenic survival of wild-type, PARP1^−/−, PARP1^−/−/PARP2^−/−, and XRCC1^−/− RPE-1 cells following incubation in the presence of the indicated concentrations of FEN1 inhibitor. Data are the mean ± SEM of three independent experiments. The inset shows an independent set (n = 3) of experiments in which the above cell lines and, additionally, XRCC1^−/−/PARP1^−/− cells were incubated with 10 μM FEN1 inhibitor.