ZFP281-BRCA2 prevents R-loop accumulation during DNA replication

Abstract

R-loops are prevalent in mammalian genomes and involved in many fundamental cellular processes. Depletion of BRCA2 leads to aberrant R-loop accumulation, contributing to genome instability. Here, we show that ZFP281 cooperates with BRCA2 in preventing R-loop accumulation to facilitate DNA replication in embryonic stem cells. ZFP281 depletion reduces PCNA levels on chromatin and impairs DNA replication. Mechanistically, we demonstrate that ZFP281 can interact with BRCA2, and that BRCA2 is enriched at G/C-rich promoters and requires both ZFP281 and PRC2 for its proper recruitment to the bivalent chromatin at the genome-wide scale. Furthermore, depletion of ZFP281 or BRCA2 leads to accumulation of R-loops over the bivalent regions, and compromises activation of the developmental genes by retinoic acid during stem cell differentiation. In summary, our results reveal that ZFP281 recruits BRCA2 to the bivalent chromatin regions to ensure proper progression of DNA replication through preventing persistent R-loops.

Fig. 1. ZFP281 is required for proper S phase progression in mouse ES cells.

a Western blot showing successful depletion of ZFP281 in mouse ES cells by CRISPR-Cas9. α-TUBULIN was used as a loading control. Four independent experiments show similar results. b Growth curve showing numbers of WT (yellow), ZFP281 KO-1 (green), and ZFP281 KO-2 (red) mouse ES cells counted at different time intervals after plating. Mean ± SEM from three independent experiments. c Flow cytometry analysis showing the percentage of WT, ZFP281 KO-1, and ZFP281 KO-2 ES cell at different cell cycle phases. Three independent experiments show similar results. d Western blot analysis (upper panel) showing the levels of Cyclin A2, Cyclin E1 and Cyclin C in WT, ZFP281 KO-1, and ZFP281 KO-2 ES cells. Histone H3 was used as a loading control. Gray values (lower panel) of Cyclin A2, Cyclin E1 and Cyclin C protein bands were determined by Image J and normalized to the loading control. Mean ± SEM from three independent experiments. Two-tailed, unpaired Student’s t tests were performed. (Cyclin A2, WT vs. ZFP281 KO-1, p = 0.0005; WT vs. ZFP281 KO-2, p = 0.0003. Cyclin E1, WT vs. ZFP281 KO-1, p = 0.0001; WT vs. ZFP281 KO-2, p = 0.0001. Cyclin C, WT vs. ZFP281 KO-1, p = 0.6566; WT vs. ZFP281 KO-2, p = 0.0500.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. e EdU staining pattern (upper left) of WT and ZFP281 KO ES cells, DNA was counterstained using DAPI; Representative images showing the EdU staining pattern (upper right) of mouse ES cells at early, middle and late S phase; Percentage of WT (total n = 422) and ZFP281 KO ES (total n = 338) cells at different stages of S phase. Three independent experiments show similar results. Source data are provided as a Source Data file.

Fig. 2. ZFP281 binds to nascent DNAs.

a Co-immunofluorescence showing co-localization of ZFP281 and PCNA in mouse ES cells. DNA was counterstained using DAPI. Three independent experiments show similar results. b Profile plot showing normalized pixel intensity of ZFP281 (green) and PCNA (red) corresponding to the region marked with white lines. c Western blot analysis (upper panel) showing the level of chromatin-bound ZFP281 in mouse ES cells after release from APH treatment for 0, 1, 2, 3 and 4 hrs respectively. Histone H3 was used as a loading control. Gray values (lower panel) of ZFP281 protein bands were determined by Image J and normalized to the loading control. Mean ± SEM from three independent experiments. Two-tailed, unpaired Student’s t tests were performed. (DMSO vs. R0, p < 0.0001; DMSO vs. R1, p = 0.0128. DMSO vs. R2, p = 0.1565. DMSO vs. R3, p < 0.0001. DMSO vs. R4, p < 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. d Western blot analysis (upper panel) showing the levels of total, chromatin-bound and soluble PCNA in WT and ZFP281 KO ES cells. Histone H3 and MEK2 were used as loading controls. Gray values (lower panel) of PCNA protein bands were determined by Image J and normalized to the loading control. Mean ± SEM from three independent experiments. Two-tailed, unpaired Student’s t tests were performed. (Total PCNA, WT vs. ZFP281 KO, p = 0.0740; Chromatin-bound PCNA, WT vs. ZFP281 KO, p < 0.0001; Soluble PCNA, WT vs. ZFP281 KO, p < 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. e Reciprocal immunoprecipitation analyses showing the interaction between ZFP281 and PCNA. The experiments were performed three times with similar results. f The representative images of ZFP281/PCNA-PLA foci in DMSO and APH treated ES cells (left panel). PLA using IgG and PCNA antibodies showed no background (n = 50). Quantification of the number of PLA puncta per nucleus in DMSO (n = 222) and APH (n = 171) treated ES cells (right panel). Error bars represent 95% confidence intervals. Two-tailed, unpaired Student’s t tests were performed. IgG + PCNA vs. ZFP281 + PCNA, p < 0.001; ZFP281 + PCNA (DMSO) vs. ZFP281 + PCNA (APH) p < 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. g BrdU re-ChIP assays showing that ZFP281 binds to nascent DNAs at the tested regions. Mean ± SEM from three independent experiments. h BrdU IP-qPCR analyses showing BrdU incorporation defects at the Gata6, Pou3f2, Kctd1 and Lhx4 loci in ZFP281 KO cells after release from APH treatment. Mean ± SEM from three independent experiments.

Fig. 3. ZFP281 depletion leads to aberrant R-loop accumulation.

a Representative immunofluorescence images with the S9.6 antibody showing the levels of R-loop in WT, ZFP281 KO-1, and ZFP281 KO-2 ES cells. DNA was counterstained using DAPI. Three independent experiments show similar results. b Fluorescence intensities of R-loop loci in WT (n = 48), ZFP281 KO-1 (n = 66) and ZFP281 KO-2 (n = 53) ES cells. Three independent experiments show similar results. Error bars represent 95% confidence intervals. Two-tailed, unpaired Student’s t tests were performed. (WT vs. ZFP281 KO-1, p < 0.0001; WT vs. ZFP281 KO-2, p < 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. c Immunofluorescence imaging of H3S10P and H3T3P in WT, ZFP281 KO, and ZFP281 KO with RNase H1 overexpression ES cells. White arrows showing the interphase cells positive for H3S10P. d Percentage of WT (total n = 369), ZFP281 KO (total n = 380) and ZFP281 KO with RNase H1 overexpression (total n = 290) ES cells in mitosis and interphase. Mean ± SEM from three independent experiments. Two-tailed, unpaired Student’s t tests were performed. (Mitosis, WT vs. ZFP281 KO, p = 0.0574; ZFP281 KO vs. ZFP281 KO + RNase H1, p = 0.5526. Interphase, WT vs. ZFP281 KO, p = 0.0020; ZFP281 KO vs. ZFP281 KO + RNase H1, p = 0.0016.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. e Western blot analysis (upper panel) showing the levels of chromatin-bound γH2A.X and PCNA in WT, ZFP281 KO, and ZFP281 KO with RNase H1 overexpression ES cells after release from APH treatment for 0, 0.5 and 1 h respectively. Histone H3 was used as a loading control. Gray values (middle and lower panels) of γH2A.X and PCNA protein bands were determined by Image J and normalized to the loading control. Mean ± SEM from three independent experiments. Multiple t tests were performed. (γH2A.X in WT and ZFP281 KO, DMSO, p < 0.0001; R0, p = 0.0005; R0.5, p < 0.0001; R1, p = 0.0035. γH2A.X in ZFP281 KO and ZFP281 KO + RNase H1, DMSO, p < 0.0001; R0, p = 0.0003; R0.5, p < 0.0001; R1, p = 0.0001. PCNA in WT and ZFP281 KO, DMSO, p < 0.0001; R0, p = 0.0002; R0.5, p < 0.0001; R1, p < 0.0001. PCNA in ZFP281 KO and ZFP281 KO + RNase H1, DMSO, p < 0.0001; R0, p < 0.0001; R0.5, p < 0.0001; R1, p < 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant.

Fig. 4. ZFP281 interacts with BRCA2.

a Purification and mass spectrometry analyses of the ZFP281-associated proteins. Clonal cell lines expressing FLAG-tagged ZFP281 were generated in 293 Flp-in-TRex cells and the ZFP281 associated proteins were purified using the FLAG-affinity purification method and analyzed by SDS-PAGE, silver staining and mass spectrometry. At least two independent biological repeats with similar results. b Confirmation of the interaction of ZFP281 with EMSY, BRCA2, and QSER1 by endogenous immunoprecipitations. c Size exclusion chromatography of ES nuclear extracts demonstrated that the ZFP281, BRCA2, EMSY and QSER1 co-elute at ~ 2 MDa (fractions 9–12). d, e Different ZFP281 truncated proteins were expressed with an N-terminal FLAG tag in 293 T cells, and the interactions of BRCA2, EMSY, QSER1 with these ZFP281 truncated proteins were examined by FLAG immunoprecipitations, followed by western blot analyses. f, g Different BRCA2 truncated proteins were expressed with an N-terminal FLAG tag in 293 T cells, and the interactions of ZFP281, EMSY, QSER1 with these BRCA2 truncated proteins were examined by FLAG immunoprecipitations, followed by western blot analyses. At least three biological replicates were performed in b, c, e and g with similar results.

Fig. 5. ZFP281 is required for the recruitment of BRCA2 to the bivalent regions in mouse ES cells.

a Pie chart showing the percentage of BRCA2 peaks overlapping with a transcription start site (TSS), residing within a gene (inside), or upstream or downstream of the nearest gene. b G/C rich sequence is overrepresented in BRCA2 peaks. Statistical significance of the over-representation and the percentage of the G/C rich sequence in BRCA2 peaks are shown. p value from hypergeometric test. c Average occupancy plots of ZFP281, BRCA2, H3K4me1 and H3K4me3 at the BRCA2 and ZFP281 co-bound regions (left panel) and the BRCA2 only regions (right panel) in mouse ES cells. Shown are ±5 kb of the center of BRCA2 peaks. d Heat maps of the binding profiles in mouse ES cells for BRCA2, ZFP281, H3K4me3, and H3K27me3 are shown at the BRCA2 and ZFP281 co-bound regions, which were partitioned into two groups: the H3K4me3 and H3K27me3 co-bound (bivalent) group, and the H3K4me3 only (active) group. Shown are ±50 kb of the center of BRCA2 peaks. e BRCA2 occupancy log2 fold change after ZFP281 KO was measured at the BRCA2 and ZFP281 co-bound regions. Shown are ±50 kb around the center of BRCA2 peaks. f, g DRIP-qPCR showing that the levels of R-loop are reduced at the tested bivalent regions (labeled with red color) after ZFP281 knockout (f) or AID mediated BRCA2 knockdown (g), while remains unchanged at the tested active Pura and Hspd1 loci (labeled with green color). DRIP-qPCR analyses in samples pre-treated with RNase H were used as negative controls. Mean ± SEM from three independent experiments. Multiple t tests were performed. (f, Gata6, p = 0.0009; Kctd1, p = 0.0006; Pou3f2, p = 0.0092; Lhx4, p = 0.0069; Meis2, p = 0.0170; Hoxa1, p = 0.0051; Pura, p = 0.1255; Hspd1, p = 0.5828. g Gata6, p < 0.001; Kctd1, p = 0.0002; Pou3f2, p = 0.0003; Lhx4, p = 0.0012; Meis2, p = 0.0002; Hoxa1, p = 0.0005; Hspd1, p = 0.0611.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. h Reciprocal immunoprecipitation analyses showing the interaction between ZFP281 and DDX5. i Immunoprecipitation analyses showing that the interaction between BRCA2 and DDX5 was compromised after ZFP281 knockout. The experiments in h and i were performed three times show similar results.

Fig. 6. BRCA2 also requires PRC2 for its binding to the bivalent chromatin.

a Heat maps of the BRCA2 binding profile in WT and EED KO ES cells. Shown are ±50 kb of the center of BRCA2 peaks. b ChIP-qPCR showing that the occupancies of BRCA2 are reduced at the tested bivalent regions (labeled with red color) in EED KO ES cells, while remains unchanged at the tested active Pura and Hspd1 loci (labeled with green color). c ChIP-qPCR showing that the occupancies of ZFP281 remains unchanged after EED KO. b, c The HEMO gene (Hba2) serves as a negative control for ChIP-qPCR. Mean ± SEM from three independent experiments. Multiple t tests were performed. (b, Hba2, p = 0.3989; Hoxb13, p = 0.0006; Gata6, p = 0.0019; Meis2, p < 0.0001; Pou3f2, p = 0.0012; Vrtn, p = 0.0023; Lhx4, p = 0.0055; Pura, p = 0.1818; Hspd1, p = 0.0564. c Hba2, p = 0.1929; Hoxb13, p = 0.3279; Gata6, p = 0.1113; Meis2, p = 0.1445; Pou3f2, p = 0.0102; Vrtn, p = 0.0657; Lhx4, p = 0.2136; Hspd1, p = 0.5652.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. d Endogenous immunoprecipitation analyses showing the interaction of SUZ12 with ZFP281 and BRCA2. Three independent biological repeats with similar results. e RT-qPCR showing that the induction of Hoxb5 and Hoxb1 by RA is impaired in ZFP281 KO ES cells. Results shown are technical replicates from representative biological replicates. Mean ± SEM from three independent experiments. Two-tailed, unpaired Student’s t tests were performed. (Hoxb1 and Hoxb5 in NonT shRNA, DMSO vs. RA24, p < 0.0001. Hoxb1 in RA24, NonT shRNA vs. ZFP281 shRNA, p < 0.0001; NonT shRNA vs. BRCA2 shRNA, p < 0.0001. Hoxb5 in RA24, NonT shRNA vs. ZFP281 shRNA, p < 0.0001; NonT shRNA vs. BRCA2 shRNA, p = 0.0001.) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, n.s. = not significant. f Cartoon model showing that ZFP281 functions together with BRCA2 to facilitate DNA replication through preventing R-loop accumulation in the bivalent chromatin (upper panel). ZFP281 depletion impairs the binding of BRCA2 to the bivalent chromatin, leads to aberrant R-loop accumulation and subsequent replication defects (middle panel). RNase H1 overexpression in ZFP281 knockout cells is able to rescue the DNA replication defects caused by ZFP281 knockout (lower panel).