Fig 2: Accumulation of lagging strand gaps in BRCA-deficient cells.a, b BrdU alkaline comet assay showing that PRIMPOL depletion reduces, but does not abolish replication stress-induced ssDNA accumulation in BRCA2-knockout HeLa cells, indicating the presence of PRIMPOL-independent gaps. Representative micrographs (a) and quantifications (b) are shown. At least 45 nuclei were quantified for each condition. The mean values are represented on the graph, and the p-values (t-test, two-tailed, unpaired) are listed at the top. A schematic representation of the assay conditions is also presented. RT-qPCR experiments confirming PRIMPOL depletion are presented in Supplementary Fig. 6a. c, d SIRF assay showing that, upon replication stress, LIG1 knockdown induces PAR chain formation in wild-type HeLa cells, but not in BRCA1 or BRCA2-depleted HeLa cells upon treatment with a low dose of HU (0.4 mM), indicating the prevalence of lagging strand gaps in these cells. Representative micrographs (scale bar represents 10 µm) (c) and quantifications (d) are shown. At least 35 positive cells were quantified for each condition. Center line indicates the median, bounds of box indicate the first and third quartile, and whiskers indicate the 10th and 90th percentile. The p-values (Mann–Whitney test, two-tailed) are listed at the top. A schematic representation of the SIRF assay conditions is also presented. Western blots confirming the co-depletion are shown in Supplementary Fig. 6d. e, f SIRF assay showing that Pola inhibition suppresses the increased CHAF1A retention in BRCA1 or BRCA2-depleted HeLa cells. Representative micrographs (scale bar represents 10 µm) (e) and quantifications (f) are shown. At least 35 positive cells were quantified for each condition. Center line indicates the median, bounds of box indicate the first and third quartile, and whiskers indicate the 10th and 90th percentile. The p-values (Mann–Whitney test, two-tailed) are listed at the top. A schematic representation of the SIRF assay conditions is also presented. g, h SIRF assay showing that RPA1 co-depletion restores HIRA levels to the same levels in wild-type and BRCA1 or BRCA2-knockodown HeLa cells. Representative micrographs (scale bar represents 10 µm) (g) and quantifications (h) are shown. At least 30 positive cells were quantified for each condition. Center line indicates the median, bounds of box indicate the first and third quartile, and whiskers indicate the 10th and 90th percentile. The p-values (Mann–Whitney test, two-tailed) are listed at the top. A schematic representation of the SIRF assay conditions is also presented. Western blots confirming the co-depletion are shown in Supplementary Fig. 6h.

Fig 3: DNA binding domain B (DBD-B) in RPA70 is phosphorylated by Aurora B kinase.a Three RPA subunits RPA70, RPA32, and RPA14 form a heterotrimer and harbor multiple oligosaccharide/oligonucleotide (OB) domains. A, B, C, and D are DNA binding domains (DBDs). OB-F and the wh-domain are two protein interaction domains. The N-terminus of RPA32 (shown in red square) is hyperphosphorylated by multiple kinases including ATM, DNA-PK, CDK, and ATR and the known sites of phosphorylation are denoted in the insert. Arg-382 and Ser-384 in the putative Aurora B kinase motif are also noted in RPA70. The structure of human RPA OB-domains and connecting linkers were generated using AlphaFold and aligned to the trimerization core observed in the crystal structure (PDB:1L1O). b Sequence alignment of residues in DBD-B reveal a conserved putative phosphorylation motif for Aurora kinase in eukaryotes. A sequence logo representation of the conservation is also shown. c In vitro phosphorylation of recombinant RPA with Aurora B, followed by MS-MS analysis, shows residue Ser-384 in RPA70 as the sole site of phosphorylation. Extracted Ion Chromatograms (EICs) for unphosphorylated and phosphorylated tryptic peptide containing Ser-384 in the WT-RPA70 along with EIC of corresponding point mutations are shown. Alanine substitution of the Ser-384 phosphosite, or perturbation of the Aurora B recognition motif through a cancer-associated Gln substitution at Arg-382, results in loss of phosphorylation.

Fig 4: A phosphomimetic S384D substitution minimally influences the DNA binding properties of RPA but induces configurational changes.a Contacts between Arg-384 and the ssDNA are shown in the crystal structure of DBD-A, B (PDB:1JMC). Ser-384 does not directly contact the ssDNA, but positions Phe-386 to promote a key base stacking interaction with the base. RPA or RPAS384D binding to DNA was measured by fluorescence anisotropy using either (b) 5'-FAM-(dT)20 or (c) 5'-FAM-(dT)40 ssDNA oligonucleotides. In both cases, high-affinity stoichiometric binding is observed for RPA and RPAS384D and no significant changes in the binding behavior are observed. Stopped flow kinetic analysis of ssDNA binding was performed with (d) RPA or (e) RPAS384D by monitoring the change in intrinsic tryptophan fluorescence as a function of ssDNA (dT)35 concentration. f Plot of the kobs versus ssDNA concentration yields kon and koff values. RPAS384D has a slower kon (1.9 ± 0.1 × 1010 M-1 s-1) and faster koff (65 ± 7 s-1) compared to RPAWT (kon = 5.2 ± 0.3 × 1010 M-1 s-1 and koff = 39 ± 4 s-1). KD values extracted from these measurements show RPAWT binding to ssDNA with ~5-fold higher affinity (KD = 0.75 ± 0.4 nM) compared to RPAS384D (KD = 3.4 ± 0.2 nM). g A Förster Resonance Energy Transfer (FRET) experiment was developed using two fluorescent versions of RPA. RPA was site-specifically labeled on either DBD-A with Cy3 or DBD-D with Cy5. Equimolar ratios of both fluorescent RPA were mixed with ssDNA (dT)97 in a stopped flow. Changes in Cy5 fluorescence were monitored by exciting Cy3 at 535 nm. Assembly of multiple RPA on the long ssDNA substrate results in a high FRET signal (green trace). In the absence of ssDNA, no enhancement in fluorescence is observed. h RPA filaments were pre-formed on ssDNA (dT)97 using the Cy5- and Cy3-labeled RPA and facilitated exchange activity was measured by challenging the RPACy3-RPACy5-ssDNA assembly with mixing against unlabeled RPA or RPAS384D. RPA exchanges the fluorescent-RPA faster (kFE = 1.8 s-1) compared to RPAS384D (kFE = 0.9 s-1). Data are presented as ±SEM from three independent experiments.

Fig 5: Aurora B phosphorylates RPA at Ser-384 specifically in mitosis.a Western blot analysis of asynchronous (Async) HCT116 cells, nocodazole-arrested mitotic cells, and arrested cells released (3 h) into G1 phase. Blots were probed with monoclonal phospho Ser-384 RPA70 (pS384-RPA70) (custom antibody, Genscript) and total RPA70 antibodies. Vinculin was used as a loading control. Mitotic arrest was confirmed using phospho-Ser10-histone H3 specific antibody (pS10-HH3). * Indicates a slower migrating form of RPA70 enriched in mitotic cells. b Similar mitosis-specific phosphorylation of RPA70 at Ser-384 was also observed in 293 T cells using western blot analysis. All subsequent studies were carried out in HCT116 cells. c RPA70 phosphorylation at Ser-384 was assessed in different phases of the cell cycle by synchronization at G1/S boundary using double thymidine block followed by release into S and G2 phases. Phosphorylation in mitosis was assessed as described in a). Tubulin and vinculin were used as loading controls. Specific phosphorylation of RPA70 at Ser-384 is observed only during mitosis. d Aurora B was selectively inhibited in mitotic cells with short-term treatment (45 min) of 3 µM AZD1152 and results in loss of RPA70 Ser-384 phosphorylation. Inhibition of Aurora B was confirmed by loss of Ser10-Histone H3 phosphorylation. Asynchronous cells and vehicle-treated (0.1% DMSO) mitotic cells were used as controls. e Cells were treated with vehicle (Veh; 0.1% DMSO) or 10 ng/ml SN-38 for 21 h to induce replication stress. DNA damage does not induce RPA70 phosphorylation at Ser-384. All blots in this figure are representative of at least three independent replicates.