PRIMPOL-Mediated Adaptive Response Suppresses Replication Fork Reversal in BRCA-Deficient Cells

Abstract

Acute treatment with replication-stalling chemotherapeutics causes reversal of replication forks. BRCA proteins protect reversed forks from nucleolytic degradation, and their loss leads to chemosensitivity. Here, we show that fork degradation is no longer detectable in BRCA1-deficient cancer cells exposed to multiple cisplatin doses, mimicking a clinical treatment regimen. This effect depends on increased expression and chromatin loading of PRIMPOL and is regulated by ATR activity. Electron microscopy and single-molecule DNA fiber analyses reveal that PRIMPOL rescues fork degradation by reinitiating DNA synthesis past DNA lesions. PRIMPOL repriming leads to accumulation of ssDNA gaps while suppressing fork reversal. We propose that cells adapt to repeated cisplatin doses by activating PRIMPOL repriming under conditions that would otherwise promote pathological reversed fork degradation. This effect is generalizable to other conditions of impaired fork reversal (e.g., SMARCAL1 loss or PARP inhibition) and suggests a new strategy to modulate cisplatin chemosensitivity by targeting the PRIMPOL pathway.

Keywords: ATR; BRCA; DNA damage; DNA replication; PRIMPOL; adaptive response; replication fork repriming; replication fork reversal; replication stress response; ssDNA gaps.

Graphical abstract

Figure 1

Treatment with a Cisplatin Pre-dose Abolishes Nascent DNA Degradation in BRCA1-Deficient Cells (A) Cell survival of UW and UW+BRCA1 cells upon 6 days of chronic treatment with the indicated doses of cisplatin. Means ± SEM (n = 3). Statistics: two-way ANOVA followed by Bonferroni test. ∗∗p < 0.01, ∗∗∗p < 0.001. (B) Schematic of the DNA fiber assay and representative DNA fiber images of UW cells. IdU (red) was added for 20 min followed by CldU (green) for 60 min ± 150 μM cisplatin ± 50 μM mirin (added concomitantly with CldU labeling). Scale bar: 25 μm. (C) Dot plot and median of IdU tract lengths in UW and UW+BRCA1 cells ± 150 μM cisplatin ± 50 μM mirin (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. (D) Schematic of the DNA fiber analysis with multiple cisplatin doses. Cells were treated with the cisplatin pre-dose (50 μM for 1 h). After 24 h, cells were treated with the second cisplatin dose (challenging dose, 150 μM) added concomitantly with CldU for 1 h. (E and F) Dot plot and median of IdU tract lengths in UW (E) and U2OS cells depleted for BRCA1 (siBRCA1) (F) ± 150 μM cisplatin ± pre-dose (n = 3). ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. See also Figure S1.

Figure 2

PRIMPOL Rescues Nascent Strand Degradation in BRCA1-Deficient Cells (A–C) PRIMPOL mRNA (A) and protein (B and C) expression 24 h after treatment with 0 or 50 μM cisplatin in UW and UW+BRCA1 cells. (A) Means ± SEM (n = 5) are shown and presented relative to untreated UW+BRCA1 cells. Representative western blot (B) and quantifications (C) from three independent experiments. Statistics: two-way ANOVA followed by Bonferroni test. ns, non-significant, ∗p < 0.05. (D) Chromatin-bound PRIMPOL in UW cells ± 150 μM cisplatin ± 50 μM mirin ± pre-dose. A representative western blot from three independent experiments is shown. Whole-cell extracts of PRIMPOL KO cells confirm the PRIMPOL antibody specificity. (E) Expression of PRIMPOL after siRNA (siPRIMPOL) knockdown in UW cells 24 and 48 h after transfection (top). Dot plot and median of IdU tract lengths in siPRIMPOL or siRNA control (siCT) UW cells ± 150 μM cisplatin ± pre-dose (bottom) (n = 3). ns, non-significant, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. (F) Dot plot and median of IdU tract lengths in siCT or siPRIMPOL UW cells treated with ± 30 J/m² UVC ± pre-dose (10 J/m²) (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. (G) Dot plot and median of CldU/IdU ratios in siCT or siPRIMPOL UW cells treated with 4 mM HU for 2 h ± pre-dose (1 mM HU for 2 h) (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. See also Figure S2.

Figure 3

ATR Activity Controls PRIMPOL-Mediated Adaptive Response (A) Percentage of UW+BRCA1 and UW cells positive for chromatin-bound RPA detected by flow cytometry 24 h after treatment with 0 or 50 μM cisplatin. Means ± SEM (n = 3). Statistics: two-way ANOVA followed by Bonferroni test. ns, non-significant, ∗∗p < 0.01. (B) p-Chk1 (S345, green), total Chk1 (red), p-RPA32 (S33, green), and total RPA (red) expression in UW+BRCA1 and UW cells 24 h upon 0 or 50 μM cisplatin. Simultaneous detection of phosphorylated form and total protein bands is shown in p-Chk1/Chk1 and p-RPA/RPA. A western blot representative of three independent experiments is shown. (C) Schematic for the DNA fiber assay with the ATR inhibitor (ATRi) VE-821. 62.5 nM VE-821 was added 1 h prior to treatment with the pre-dose and removed from the media 4 h before performing the DNA fiber assay (top). Dot plot and median of IdU tract lengths in UW cells ± 150 μM cisplatin ± pre-dose ± ATRi (bottom) (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. (D and E) PRIMPOL mRNA (D) and protein (E) levels 24 h upon 0 or 50 μM cisplatin (pre-dose) ± ATRi. (D) Means ± SEM. Three independent biological replicates are shown and presented as fold change between 50 μM cisplatin and untreated samples. Representative western blot (E, top) and quantification (E, bottom) from four independent experiments. Statistics: two-way ANOVA followed by Bonferroni test. ns, non-significant, ∗p < 0.05. See also Figure S3.

Figure 4

PRIMPOL Primase Activity Rescues Nascent Strand Degradation in BRCA1-Deficient Cells (A) PRIMPOL overexpression in UW cells upon transfection with WT (wild-type), AxA (catalytic dead) and CH (primase dead only) V5-PRIMPOL constructs (top). Dot plot and median of IdU tract lengths in UW cells overexpressing the different constructs ± 150 μM cisplatin (bottom) (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. (B) Schematic for detection of ssDNA gaps using the ssDNA-specific S1 nuclease upon multiple treatments with cisplatin (top). Dot plot and median of CldU tract lengths in UW cells ± 150 μM cisplatin ± pre-dose ± S1 nuclease (bottom) (n = 3). ns, non-significant, ∗p < 0.05, ∗∗∗∗p < 0.0001. (C) Representative electron micrograph of a replication fork with internal ssDNA gaps behind the fork indicated by the arrows. Scale bar: 500 nm (left). Magnified internal ssDNA gap. Scale bar: 100 nm (right). P: parental strand, D: daughter strand. (D) Percentage of replication forks with 1, 2 or ≥3 internal ssDNA gaps in UW cells ± 150 μM cisplatin ± pre-dose ± mirin. Means ± SEM (n = 3). Statistics: one-way ANOVA followed by Bonferroni test. ∗∗p < 0.01. See also Figure S4 and Table S1.

Figure 5

PRIMPOL Overexpression Is Linked to Decreased Fork Reversal (A) Representative electron micrograph of a reversed replication fork. Scale bar: 200 nm. Magnified four-way junction at the reversed fork. Scale bar: 20 nm. P, parental strand; D, daughter strand; R, reversed arm. (B) Schematic for the EM assay in UW cells treated with multiple doses of cisplatin (top). Percentage of reversed replication forks in UW cells ± 150 μM cisplatin ± pre-dose ± mirin. Means ± SEM (n = 3). Statistics: one-way ANOVA followed by Bonferroni test. (bottom). ns, non-significant, ∗p < 0.05. (C) Percentage of reversed forks (left) and forks with internal gaps (right) in U2OS cells treated with 150 μM cisplatin ± pre-dose. Average ± SEM (n = 3). ∗p < 0.05. (D) Schematic for the EM assay in U2OS cells overexpressing V5-PRIMPOL (top left). Expression of V5-PRIMPOL in U2OS cells (PRIMPOL in green, V5 in red, PRIMPOL/V5 shows simultaneous detection of both bands). PCNA was used a loading control (top right). Percentage of reversed forks (bottom left) and forks with internal gaps (bottom right) in U2OS mock-treated or overexpressing V5-PRIMPOL treated with 150 μM cisplatin. Means ± SEM (n = 3). Statistics: unpaired t test. ∗∗p < 0.01. See also Tables S1, S2, and S3.

Figure 6

Depletion of Fork Reversal Factors Leads to Accumulation of PRIMPOL-Dependent ssDNA Gaps (A) Schematic for detection of ssDNA gaps using the ssDNA-specific S1 nuclease upon treatment with cisplatin (150 μM). (B) Expression of RAD51 and PRIMPOL after siRNA (siRAD51 and siPRIMPOL) knockdown in UW cells 48 h after transfection (top). Dot plot and median of CldU tract lengths upon treatment with 150 μM cisplatin in UW cells depleted for RAD51, PRIMPOL, or RAD51/PRIMPOL ± S1 nuclease (bottom) (n = 3). ns, non-significant, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. (C) Expression of SMARCAL1 48 h after depletion with siRNA (siSMARCAL1) and PRIMPOL upon addition of doxycycline (DOX) in U2OS siBRCA1 cells stably expressing a DOX-inducible shPRIMPOL (shPRIMPOL) (top). Dot plot and median of CldU tract lengths upon treatment with 150 μM cisplatin in U2OS siBRCA1 cells depleted for SMARCAL1, PRIMPOL, or SMARCAL1/PRIMPOL ± S1 nuclease (bottom) (n = 2). ns, non-significant, ∗∗∗∗p < 0.0001. (D) Dot plot and median of CldU tract lengths upon treatment with 150 μM cisplatin in U2OS cells depleted for SMARCAL1, PRIMPOL, or SMARCAL1/PRIMPOL ± S1 nuclease (bottom) (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. (E) Dot plot and median of CldU tract lengths upon treatment with 150 μM cisplatin in U2OS cells depleted for PRIMPOL (shPRIMPOL) ± PARPi and ± S1 nuclease (n = 3). ns, non-significant, ∗∗∗∗p < 0.0001. See also Figure S5.

Figure 7

Impact of PRIMPOL on Cell Survival (A) Cell survival of UW and UW+PRIMPOL cells upon 6 days of chronic treatment with ATR inhibitor (ATRi, VE-821, 250 nM) and the indicated doses of cisplatin. Means ± SEM (n = 3). Statistics: two-way ANOVA followed by Bonferroni test comparing UW+ATRi versus UW+PRIMPOL+ATRi. ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (B) Expression of PRIMPOL after siRNA (siPRIMPOL) in UW+BRCA1 and UW cells (top). Cell count in PRIMPOL-depleted UW+BRCA1 and UW cells. Means ± SEM (n = 3) (bottom). Statistics: two-way ANOVA followed by Bonferroni test comparing UW siCT versus UW siPRIMPOL. ∗∗∗p < 0.001. (C) Doubling time of DT40 cell mutants over the course of approximately 40 passages post-transfection with the smarcal1 targeting construct. Each circle represents an individual measurement of doubling time derived from three independent experiments, repeated in duplicate and on multiple independently derived clones (two primpol −/−; three smarcal1 +/−; and five primpol −/− smarcal1 +/−). Means and SD plotted as line and whiskers. Statistics: Kruskal-Wallis test for difference between smarcal1 +/− and primpol −/− smarcal1 +/−. ns, non-significant, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (D) Proposed model. Fork repriming and reversal are two alternative mechanisms by which cells deal with cisplatin-induced DNA lesions (top). Upon treatment with a single cisplatin dose, fork reversal is the most frequent event. However, reversed forks are targeted by nucleases in a BRCA-deficient background leading to nascent DNA degradation (middle). Multiple treatments with cisplatin lead to an ATR-dependent upregulation of PRIMPOL, shifting the balance toward repriming events while suppressing fork reversal. Alternatively, depletion of fork reversal factors, such as RAD51 and SMARCAL1, also favors PRIMPOL-dependent repriming in both BRCA-deficient and -proficient cells (bottom). See also Figures S6 and S7.