A RAD18-UBC13-PALB2-RNF168 axis mediates replication fork recovery in BRCA1-deficient cancer cells

Abstract

BRCA1/2 proteins function in genome stability by promoting repair of double-stranded DNA breaks through homologous recombination and by protecting stalled replication forks from nucleolytic degradation. In BRCA1/2-deficient cancer cells, extensively degraded replication forks can be rescued through distinct fork recovery mechanisms that also promote cell survival. Here, we identified a novel pathway mediated by the E3 ubiquitin ligase RAD18, the E2-conjugating enzyme UBC13, the recombination factor PALB2, the E3 ubiquitin ligase RNF168 and PCNA ubiquitination that promotes fork recovery in BRCA1- but not BRCA2-deficient cells. We show that this pathway does not promote fork recovery by preventing replication fork reversal and degradation in BRCA1-deficient cells. We propose a mechanism whereby the RAD18-UBC13-PALB2-RNF168 axis facilitates resumption of DNA synthesis by promoting re-annealing of the complementary single-stranded template strands of the extensively degraded forks, thereby allowing re-establishment of a functional replication fork. We also provide preliminary evidence for the potential clinical relevance of this novel fork recovery pathway in BRCA1-mutated cancers, as RAD18 is over-expressed in BRCA1-deficient cancers, and RAD18 loss compromises cell viability in BRCA1-deficient cancer cells.

Graphical Abstract

Figure 1.

RAD18 and UBC13 mediate fork recovery in BRCA1- but not BRCA2-deficient cells, and stalled replication forks are targeted for nucleolytic degradation. (A) Western blot of whole extract in U2OS cells depleted of BRCA1, RAD18, and/or UBC13 (left), Western blot of nuclear extract in U2OS cells depleted of BRCA1, RAD18 and/or UBC13 (RIGHT). (B) Fork recovery fiber assay scheme (TOP). Cells were labeled with 20uM IdU for 15 min, treated with HU for 2 h, and released into CldU for 15 min. Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA1, RAD18 and/or UBC13 knockdown in U2OS cells (bottom). Mean with SEM shown, N = 3, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, **P < 0.01, ***P < 0.001. (C) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA2 and/or RAD18 knockdown in U2OS cells (bottom). Mean with SEM shown, N = 2, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, ns = not significant. (D) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA2 and/or UBC13 knockdown in U2OS cells. Mean with SEM shown, N = 2, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, ns = not significant. (E) Fork degradation fiber assay scheme (top). Cells were labeled with 20 uM IdU for 30 min, followed by incubation with 200 uM CldU for 30 min, and then treated with 4 mM HU for 5 h. IdU tract and CldU tract lengths were measured on contiguous red–green fibers upon BRCA1, SMARCAL1, ZRANB3 and/or HLTF knockdown in U2OS WT (grey) and RAD18 KO (blue) cells (bottom). Each dot represents a CldU/IdU ratio from a single DNA fiber tract, blue line represents median value, N = 3, >150 fiber tracts quantified per sample for each independent experiment, Statistics: Kruskal–Wallis test followed by Dunn's multiple comparison test, **P < 0.01, ***P < 0.001, ****P < 0.0001. (F) Fork degradation fiber assay scheme (top). IdU tract and CldU tract lengths were measured on contiguous red-green fibers upon BRCA1, SMARCAL1, ZRANB3 and/or HLTF knockdown in U2OS WT (gray) or UBC13-depleted (blue) cells (BOTTOM). Each dot represents a CldU/IdU ratio from a single DNA fiber tract, blue line represents median value, N = 3, >150 fiber tracts quantified per sample for each independent experiment, Statistics: Kruskal–Wallis test followed by Dunn's multiple comparison test, ****P < 0.0001, ns = not significant.

Figure 2.

MRE11, but not DNA2, degrades reversed replication forks in BRCA1-deficient cancer cells lacking RAD18 or UBC13. (A) Fork degradation fiber assay scheme (TOP). Cells were labeled with 20uM IdU for 30 min, followed by incubation with 200 uM CldU for 30 min, and then treated with 4 mM HU for 5 h ± 50 uM Mirin ± 30 uM C5. IdU tract and CldU tract lengths were measured on contiguous red-green fibers upon BRCA1 knockdown in U2OS WT (gray) and RAD18 KO (blue) cells (BOTTOM). Each dot represents a CldU/IdU ratio from a single DNA fiber tract, blue line represents median value, N = 3, >150 fiber tracts quantified per sample for each independent experiment, Statistics: Kruskal–Wallis test followed by Dunn's multiple comparison test, **P = 0.01, ****P < 0.0001, ns = not significant. (B) Representative images of a replication fork (top left) and a reversed replication fork (bottom left) captured with EM. Percentages of reversed replication forks were quantified in BRCA1-deficient U2OS WT or RAD18 KO cells upon knockdown of UBC13 under untreated (UT) conditions (gray), with HU alone (red) or with HU + 30 uM Mirin (blue), N = 1. See Supplementary Figure S2D for the second biological repeat and Supplementary Figure S2E for quantification.

Figure 3.

PCNA ubiquitination and PALB2 promote fork recovery in BRCA1-deficient cells. (A) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA1 knockdown in 293T WT and K164 mutant cells (bottom). Mean with SEM shown, N = 4, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, *P < 0.05, ns = not significant. (B) Western blot of chromatin bound RAD18, BRCA1, PCNA and ubiquitinated PCNA upon knockdown of BRCA1 and/or RAD18 under NT conditions or after treatment with 4 mM HU for 2 h. N = 3, Representative blot shown. (C) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA1, BRCA2, RAD51 or PALB2 knockdown or treatment with RAD51 inhibitor (RAD51i, B02) at a concentration 27 uM in U2OS WT cells. N = 2, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, *P < 0.05, ns = not significant. (D) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) upon BRCA1 and/or RAD52 knockdown. N = 2, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, ns = not significant.

Figure 4.

RNF168 promotes fork recovery in BRCA1-deficient cells and mediates RAD18 recruitment to chromatin upon HU treatment. (A) Fork recovery fiber assay scheme (top). Quantification of stalled (red bars) and restarted (blue bars) replication forks upon BRCA1, RAD18, UBC13, PALB2 or RNF168 knockdown in U2OS WT cells. N = 2, >150 fiber tracts quantified per sample for each independent experiment, Statistics: unpaired t-test, **P < 0.01, *P < 0.05, ns = not significant. (B) Representative images of RAD18 foci by immunofluorescence microscopy upon knockdown of BRCA1 and/or RNF168 in U2OS WT cells in non-treated (NT) conditions or treatment with 4 mM HU for 2 h (left). Quantification of RAD18 foci per cell upon knockdown of RAD18, BRCA1, and/or RNF168 in U2OS WT cell under NT conditions (gray bars) or upon HU treatment (red bars). N = 3, >100 cells quantified per sample for each independent experiment. Statistics: one-way ANOVA, ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, ns = not significant.

Figure 5.

Loss of RAD18 or mutation of the K164 PCNA residue compromises cell viability in BRCA1-deficient cells. (A) Representative images of clonogenic survival assays upon knockdown of BRCA1 and/or UBC13 in U2OS WT or U2OS RAD18 KO cells (left). Quantification of % cell viability by clonogenic survival assay upon knockdown of BRCA1 (gray bars) and/or UBC13 (red bars) in U2OS WT or U2OS RAD18 KO cells (blue bars) (right). Mean with SEM shown, N = 4, Statistics: unpaired t-test, *P < 0.05, **P < 0.01, ns = not significant. (B) Western blot of U2OS WT or U2OS RAD18 KO cells depleted of BRCA1 and/or UBC13. (C) Quantification of MTS proliferation assay measuring proliferation relative to Day 0 upon knockdown of BRCA1 (gray dotted line) and/or UBC13 (red solid and dotted lines) in U2OS WT (gray solid line) or U2OS RAD18 KO cells (blue solid and dotted lines). Mean with SEM shown, N = 3, Statistics: two-way ANOVA followed by Bonferroni, ****P < 0.0001, significance shown for comparisons between WT versus RAD18 KO siBRCA1; WT versus WT siUBC13; and WT versus WT siBRCA1/siUBC13.

Figure 6.

RAD18 is overexpressed in BRCA1-deficient human ovarian tumors. (A) RAD18 mRNA expression from The Cancer Genome Atlas (TCGA, Nature 2012) for WT (gray, N = 425), BRCA1-mutated (red, N = 15) and BRCA2-mutated (blue, N = 18) breast cancers (left). RAD18 mRNA expression from TCGA (Nature 2011) for WT (gray, N = 250), BRCA1-mutated (red, N = 37), and BRCA2-mutated (blue, N = 31) ovarian cancers (right). Each dot corresponds to a single tumor sample, Mean with SEM indicated by the black line and error bars, Statistics: Unpaired t-tests, P values reported. (B) Representative images of RAD18 immunohistochemistry (IHC) in WT, BRCA1-mutated, and BRCA2-mutated ovarian tumors (LEFT). Quantification of raw RAD18 IHC scores (average of intensity and quantity of staining) in WT (gray), BRCA1-mutated (red), and BRCA2-mutated (blue) ovarian cancer samples from both primary (P) and metastatic (M) sites (RIGHT). Each dot corresponds to a single tumor sample, Mean with SEM indicated by the black line and error bars, Statistics: Mann–Whitney tests, *P < 0.05, **P < 0.01. (C) Proposed model of fork recovery in BRCA1-deficient cancer cells. SMARCAL1, ZRANB3 and HLTF facilitate formation of reversed replications forks, which are targeted for extensive nucleolytic degradation by MRE11 and EXO1 in BRCA1-deficient cells treated with HU. A RAD18-UBC13-PALB2-RNF168 axis, along with PCNA ubiquitination mediates fork recovery from in this genetic background. Upon loss of RAD18 or UBC13 in BRCA1-deficient cells, reversed replication forks are still degraded by the MRE11 and EXO1 nucleases. Targeting the RAD18, UBC13 and PCNA ubiquitination pathway compromises fork recovery in these cells and leads to decreased cell viability, as well as increased sensitivity to replication stress. Created with BioRender.com.