DNMT3b protects centromere integrity by restricting R-loop-mediated DNA damage

Abstract

This study used DNA methyltransferase 3b (DNMT3b) knockout cells and the functional loss of DNMT3b mutation in immunodeficiency-centromeric instability-facial anomalies syndrome (ICF) cells to understand how DNMT3b dysfunction causes genome instability. We demonstrated that R-loops contribute to DNA damages in DNMT3b knockout and ICF cells. More prominent DNA damage signal in DNMT3b knockout cells was due to the loss of DNMT3b expression and the acquirement of p53 mutation. Genome-wide ChIP-sequencing mapped DNA damage sites at satellite repetitive DNA sequences including (peri-)centromere regions. However, the steady-state levels of (peri-)centromeric R-loops were reduced in DNMT3b knockout and ICF cells. Our analysis indicates that XPG and XPF endonucleases-mediated cleavages remove (peri-)centromeric R-loops to generate DNA beaks, causing chromosome instability. DNMT3b dysfunctions clearly increase R-loops susceptibility to the cleavage process. Finally, we showed that DNA double-strand breaks (DSBs) in centromere are probably repaired by error-prone end-joining pathway in ICF cells. Thus, DNMT3 dysfunctions undermine the integrity of centromere by R-loop-mediated DNA damages and repair.

Fig. 1. R-loops are the sources of DNA damage in BKO and ICF cells.

a, b HCT116 and BKO cells were fixed for a γH2AX IF staining (scale bar, 10 μm). The relative intensity of γH2AX in cells (n > 150) from three independent experiments is expressed, ***P < 0.001 by the Mann–Whitney test. b Western blot analysis of DNA damage response markers. c γH2AX IF staining of HCT116 and BKO cells that were infected with retrovirus of empty HA-vector, HA-RNaseH1-WT, and HA-RNaseH1-D209N (scale bar, 10 μm, left). Relative fluorescent intensity of γH2AX in cells (n > 150) from three independent experiments (upper right). Western blot of the expression of HA-RNase H1 (bottom right). d The comparison of DNA damage signal in wild-type versus ICF LCL. Cells were fixed for IF staining by γH2AX antibody (scale bar, 10 μm). The relative intensity of γH2AX in cells (n = 100) from three independent experiments. e ICF LCLs were infected with retrovirus of empty HA-vector, HA-RNaseH1-WT, and HA-RNaseH1-D209N for γH2AX IF staining (scale bar, 10 μm, Left). Relative fluorescent intensity of γH2AX in cells (n = 100) from three independent experiments (upper right). Western blot analysis of HA-RNaseH1-WT and HA-RNaseH1-D209N (bottom right).

Fig. 2. The loss of DNMT3b and the acquired p53 mutation cause prominent DNA damage in HCT116 cells.

a NGS targeting sequencing of 54 genes in HCT116 and BKO cells. HCT116 and BKO shared common mutations in four genes. Additional mutations at p53 (TP53) were found in BKO cells. b BKO cells were infected with Tet-on Flag-p53-IRES-GFP virus followed by treatment with or without 2 μM doxycycline for 3 days for western blot (left) and viability assay (right, means ± SEM, n = 3, *** indicated P < 0.001 by two-tailed unpaired Student’s t-test. ns. P > 0.05). c BKO cells stably expressing Flag-DNMT3b were selected and infected with Flag-p53-IRES-GFP virus. Afterward, cells were treated with or without 1 μM doxycycline for 48 h and fixed for γH2AX IF staining (scale bar, 10 μm, left). Relative intensity of γH2AX in cells (n > 80), *** P < 0.001 by Mann–Whitney test (upper right). Western blots of Flag-DNMT3b, p53, p21, and Actin (bottom right). d ICF LCLs were transfection with 2 μg control siRNA or p53 siRNA using Amaxa® Cell Line Nucleofector® Kit V. After post-transfection at 48 h, cells were fixed for γH2AX IF staining (scale bar, 10 μm, left). Fluorescent intensity of γH2AX in cells (n > 100) was quantitated by Image J and relative intensity is expressed, ***P < 0.001 by the Mann–Whitney test (middle). Western blots of p53 and GAPDH (right).

Fig. 3. R-loops-dependent DNA damages at repetitive sequences.

a Genome-wide sequencing of DNA damage sites in HCT116 and BKO cells. ChIP by γH2AX antibody was used for sequencing analysis. The plot shows the relative ChIP-seq reads of γH2AX at indicated genomic regions in BKO versus control HCT116 cells. Data were from one experiment. b HCT116 and BKO cells were used for γH2AX-ChIP-qPCR analysis. Data are shown as the percentage of input DNA in γH2AX antibody at the sequences of (peri-)centromere, rDNA, telomere, and control p21 (CDKN1A) regions (mean ± SEM, n = 3, ***P < 0.001, ns: no significant by two-tailed unpaired Student’s t-test). c γH2AX-ChIP-qPCR in wild-type and ICF LCL cells. γH2AX-ChIP-qPCR was analyzed at the sequences of (peri-)centromere of chromosome and intergenic region downstream of SNRPN. Data are expressed as percentage of input DNA (mean ± SEM, n = 3, *, *** P < 0.05, 0.001, ns: no significant by two-tailed unpaired Student’s t-test). d γH2AX-ChIP-qPCR analysis in ICF LCLs after infection with retrovirus of empty HA-vector, HA-RNaseH1-WT, or HA-RNaseH1-D209N. Data are expressed as percentage of input DNA (mean ± SEM, n = 3, *, ***P < 0.05, 0.001, ns: no significant by two-tailed unpaired Student’s t-test).

Fig. 4. XPG and XPF cause centromeric DNA breaks in BKO and ICF cells.

a, b HCT116 and BKO cells were infected with shXPG and shXPF lentivirus subsequently. After recovery and selection, cells were analyzed by γH2AX IF staining and γH2AX-ChIP-qPCR analysis. a IF staining images (scale bar, 10 μm, left). Fluorescent intensity was quantitated (n > 150) from three independent experiments, ***P < 0.001 by the Mann–Whitney test (upper right). Western blots of XPG, XPF, γH2AX, H2AX, and Actin (bottom right). b Data of γH2AX-ChIP-qPCR are expressed as % input (mean ± SEM, n = 3, *, ***P < 0.05, 0.001, ns: no significant by two-tailed unpaired Student’s t-test). c, d ICF LCLs were infected by LacZ and shXPG and shXPF lentivirus. After selection, cells were analyzed by γH2AX IF staining and γH2AX-ChIP-qPCR analysis. c IF staining images of γH2AX (Scale bar,10 μm, left). Fluorescence Intensity was quantitated (n > 100) from three independent experiments and relative intensity is expressed, ***P < 0.001 by the Mann–Whitney test (middle). Western blots of XPG, XPF, γH2AX, H2AX, 53BP1, and Actin (right). d Data of γH2AX-ChIP-qPCR analysis are expressed as described in Fig. 3b (mean ± SEM, n = 3, *, **P < 0.05, 0.01, ns: no significant by two-tailed unpaired Student’s t-test). e Effect of XPG/XPF knockdown on chromatin segregation. ICF LCLs with and without XPG and XPF knockdown were arrested in mitosis by nocodazole treatment overnight. Cells were released from nocodazole arrest for 70 min and stained with Hoechst. Representative images of chromatin lagging and bridge errors are shown on the left. Percentages of cells with anaphase bridges and lagging chromosomes are shown in the histogram. Error bars are shown in means ± SEM, n = 3. *, *** indicate P < 0.05 and 0.001, respectively, by two-tailed unpaired Student’s t-test.

Fig. 4. XPG and XPF cause centromeric DNA breaks in BKO and ICF cells.

a, b HCT116 and BKO cells were infected with shXPG and shXPF lentivirus subsequently. After recovery and selection, cells were analyzed by γH2AX IF staining and γH2AX-ChIP-qPCR analysis. a IF staining images (scale bar, 10 μm, left). Fluorescent intensity was quantitated (n > 150) from three independent experiments, ***P < 0.001 by the Mann–Whitney test (upper right). Western blots of XPG, XPF, γH2AX, H2AX, and Actin (bottom right). b Data of γH2AX-ChIP-qPCR are expressed as % input (mean ± SEM, n = 3, *, ***P < 0.05, 0.001, ns: no significant by two-tailed unpaired Student’s t-test). c, d ICF LCLs were infected by LacZ and shXPG and shXPF lentivirus. After selection, cells were analyzed by γH2AX IF staining and γH2AX-ChIP-qPCR analysis. c IF staining images of γH2AX (Scale bar,10 μm, left). Fluorescence Intensity was quantitated (n > 100) from three independent experiments and relative intensity is expressed, ***P < 0.001 by the Mann–Whitney test (middle). Western blots of XPG, XPF, γH2AX, H2AX, 53BP1, and Actin (right). d Data of γH2AX-ChIP-qPCR analysis are expressed as described in Fig. 3b (mean ± SEM, n = 3, *, **P < 0.05, 0.01, ns: no significant by two-tailed unpaired Student’s t-test). e Effect of XPG/XPF knockdown on chromatin segregation. ICF LCLs with and without XPG and XPF knockdown were arrested in mitosis by nocodazole treatment overnight. Cells were released from nocodazole arrest for 70 min and stained with Hoechst. Representative images of chromatin lagging and bridge errors are shown on the left. Percentages of cells with anaphase bridges and lagging chromosomes are shown in the histogram. Error bars are shown in means ± SEM, n = 3. *, *** indicate P < 0.05 and 0.001, respectively, by two-tailed unpaired Student’s t-test.

Fig. 5. The regulation of (peri-)centromeric R-loops by XPG/XPF and transcription in BKO cells.

a, b Levels of (peri-)centromeric R-loops by DRIP analysis. a HCT116 and BKO cells and b cells with or without XPG/XPF knockdown. All DNA samples were untreated (-) or treated (+) with RNase H before immunoprecipitation by S.9.6 antibody. qPCR values of DNA-RNA hybrids at (peri-)centromere sequences of chromosomes are normalized by IgG control. Data are expressed relative to that in HCT116 cells and presented as mean ± SEM of three independent experiments (*, **P < 0.05, 0.01, ns: no significant by two-tailed unpaired Student’s t-test, RNH: RNase H). c HCT116 and BKO cells were used for Pol II-ChIP-qPCR analysis. Data are shown as the percentage of input DNA in Pol II antibody versus IgG control at (peri-)centromere sequences of chromosomes regions (mean ± SEM, n = 3, *, **, ***P < 0.05, 0.01 0.001, ns: no significant by two-tailed unpaired Student’s t-test).

Fig. 6. The functional mutation of DNMT3b in ICF increases XPG/XPF accessibility to (peri-)centromeric R-loops.

a, b Levels of (peri-)centromeric R-loops by DRIP analysis. a Wild-type and ICF LCL cells. b Cells with or without XPG/XPF knockdown. All DNA samples were untreated (–) or treated (+) with RNase H followed by S.9.6 antibody immunoprecipitation for qPCR analysis at (peri-)centromere sequences of chromosomes. qPCR values of DNA-RNA hybrids are normalized by IgG control. Data are expressed relative to that in wild type cells and presented as mean ± SEM of three independent experiments (*, **P < 0.05, 0.01, ns: no significant by two-tailed unpaired Student’s t-test, RNH: RNase H). c Pol II-ChIP-qPCR analysis in wild-type and ICF LCLs. Data are shown as % of input DNA in Pol II antibody versus IgG control at (peri-)centromere sequences of chromosomes regions (mean ± SEM, n = 3, *P < 0.05, ns: no significant by two-tailed unpaired Student’s t-test).

Fig. 7. PARP1-mediated EJ repair in ICF cells.

a ICF cells were fixed for γH2AX/ACA, Rad51/ACA, and 53BP1/ACA IF co-staining. Representative confocal co-localized signals (scale bar, 10 μm). Manders’ coefficient of co-localized γH2AX/ACA, Rad51/ACA, and 53BP1/ACA was acquired by confocal imaging as described in the methods. Examples of co-localized signals of γH2AX/ACA and 53BP1/ACA are shown under the Merged image. Co-localized signals are quantitated (n > 100) and expressed in relative values of Manders’ coefficient, ***P < 0.001 by the Mann–Whitney test. b ICF LCLs were treated with NU7441 (5 μM) or AZD2281 (5 μM) for 6 h and fixed for 53BP1/ACA IF co-staining. Representative images are shown (scale bar, 10 μm). The fluorescent intensity of 53BP1 was quantitated and relative intensity is expressed (n > 100), *, ***P < 0.001 by the Mann–Whitney test. Manders’ coefficient of co-localized 53BP1/ACA was quantified and shown. Examples of co-localized signals of 53BP1/ACA are shown under the images. Quantitation data of co-localized signal are expressed (n > 150), *, ***P < 0.05, 0.001 by the Mann–Whitney test. c and d ICF LCLs were treated with AZD2281 (5 μM) for 6 h for Comet and 53BP1-ChIP-qPCR analysis. c Comet tail moments in cells (n = 300) were analyzed by CometScore (***P < 0.001 by Mann–Whitney test). d Data of 53BP1-ChIP-qPCR are shown as the percentage of input DNA at the centromere sequence of chromosome 1 from two independent experiments. e Effect of XPG/XPF knockdown on AZD2281 sensitivity by viability assay. ICF LCLs with or without shXPG/XPF knockdown were treated with AZD2281 (5 μM) for 72 h. The percentage of the cell viability from three independent experiments is shown, mean ± SEM, ***P < 0.001 by two-tailed unpaired Student’s t-test.