The PIN1-p38-CtIP signalling axis protects stalled replication forks from deleterious degradation

Abstract

Human CtIP plays a critical role in homologous recombination (HR) by promoting the resection of DNA double-strand breaks. Moreover, CtIP maintains genome stability through protecting stalled replication forks from nucleolytic degradation. However, the upstream signalling mechanisms governing the molecular switch between these two CtIP-dependent processes remain largely elusive. Here, we show that phosphorylation of CtIP by the p38α stress kinase and subsequent PIN1-mediated CtIP cis-to-trans isomerization is required for fork stabilization but dispensable for HR. We found that stalled forks are degraded in cells expressing non-phosphorylatable CtIP or lacking PIN1-p38α activity, while expression of a CtIP trans-locked mutant overcomes the requirement for PIN1-p38α in fork protection. We further reveal that Brca1-deficient mammary tumour cells that have acquired PARP inhibitor (PARPi) resistance regain chemosensitivity after PIN1 or p38α inhibition. Collectively, our findings identify the PIN1-p38-CtIP signalling pathway as a critical regulator of replication fork integrity.

Graphical Abstract

Figure 1.

CtIP cis-to-trans isomerization protects stalled forks from nucleolytic degradation. (A) Fork degradation was evaluated upon HU treatment in U2OS cells depleted of endogenous CtIP and stably expressing either GFP empty vector (ev), or siCtIP-resistant GFP–CtIP-wild-type (wt), –S276A, –T315A, and –S276A/T315A variants. Representative DNA fiber images are shown (top). (B) Fork degradation was evaluated upon HU treatment in U2OS cells depleted of endogenous CtIP and stably expressing siCtIP-resistant GFP–CtIP-wt or –S276A/T315A variants. In addition, cells were either mock-treated or treated with the DNA2 inhibitor NSC-105808 (2 μM, simultaneously with HU). (C) Fork degradation was evaluated upon HU treatment in U2OS cells inducibly expressing siCtIP-resistant GFP–CtIP-wt or –S276A variants and depleted of endogenous CtIP alone, or co-depleted of CtIP and SMARCAL1. (D) Fork degradation was evaluated upon HU treatment in U2OS cells depleted of endogenous CtIP and inducibly expressing siCtIP-resistant GFP–CtIP-wt, –S276A, –P277A, or –S276A/P277A (trans-locked) variants. (A–D) Box and whisker plots of IdU/CldU-tract length ratios for individual replication forks are shown. Numbers indicated above the individual plots represent the mean ratios ± standard deviation. Statistical significance was calculated using the Kruskal–Wallis test with Dunn’s multiple comparisons (****P ≤ 0.0001, **P ≤ 0.01, ns, nonsignificant). Schematics of the CldU/IdU pulse-labelling protocol are shown (top). (E) Metaphase spread analysis upon HU treatment of U2OS cells depleted of endogenous CtIP and inducibly expressing siCtIP-resistant GFP–CtIP-wt, –S276A, or –S276A/P277A (trans-locked) variants. Chromatid breaks, fusions, and radials were scored. Total chromosomal aberrations per metaphase are shown. The mean with standard deviation of biological triplicates is shown. Statistical significance was calculated using one-way ANOVA test (****P ≤ 0.0001, **P ≤ 0.01, ns, non-significant).

Figure 2.

CtIP isomerization by PIN1 promotes fork stability but is dispensable for HR. (A) Fork degradation was evaluated upon HU treatment in U2OS cells pre-treated with the PIN1 inhibitor KPT-6566 (10 μM, 1 h before labelling) alone or in combination with either the Mre11 inhibitor Mirin (25 μM, simultaneously with HU) or the DNA2 inhibitor NSC-105808 (2 μM, simultaneously with HU). (B) Fork degradation was evaluated upon HU treatment in U2OS cells depleted of endogenous CtIP and inducibly expressing siCtIP-resistant GFP–CtIP-wt, –S276A, or –S276A/P277A (trans-locked) variants. In addition, cells were either mock-treated or treated with the PIN1 inhibitor KPT-6566 (10 μM, 1 h before labelling). (C) Fork degradation was evaluated upon HU treatment in U2OS cells inducibly expressing siCtIP-resistant GFP–CtIP-wt, –S276A, or –S276A/P277A (trans-locked) variants and depleted of endogenous CtIP alone, or co-depleted of CtIP and BRCA1. (A–C) Box and whisker plots of IdU/CldU-tract length ratios for individual replication forks are shown. Numbers indicated above the individual plots represent the mean ratios ± standard deviation. Statistical significance was calculated using the Kruskal–Wallis test with Dunn’s multiple comparisons (****P ≤ 0.0001, ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05, ns, non-significant). Schematics of the CldU/IdU pulse-labelling protocol are shown (top). (D) HR efficiency was evaluated in U2OS DR-GFP cells depleted for endogenous CtIP and transfected with either empty vector (ev) or indicated siCtIP-resistant FLAG-CtIP constructs. Cells were co-transfected with the I-SceI expression plasmid and harvested at 48 h post-transfection and analysed by the flow cytometry for GFP signal. Data are shown as percentage of GFP-positive cells. Values represent the mean ± standard deviation of three independent experiments. Statistical significance was calculated using the unpaired t-test (****P ≤ 0.0001, ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05, ns, non-significant). (E) Electrophoretic mobility of recombinant CtIP-wt and -S276A either non-treated or treated with λ phosphatase. (F) Endonuclease assay with recombinant MRN complex and either phosphorylated CtIP-wt or phosphorylated S276A variant on a 5′ end-labelled 70 bp-long double-stranded DNA substrate blocked at both ends with streptavidin. The quantitation (cleavage, %) is an average from three independent experiments. Schematic of the substrate and endonucleolytic cleavage is shown (top).

Figure 3.

HU-activated p38α phosphorylates CtIP at S276 and facilitates CtIP-dependent fork protection. (A) Myc-Trap of HEK293T cells transfected with Myc–p38α. Whole-cell lysates (input) and immunoprecipitates were analysed by western blotting using specific antibodies. (B) GFP-Trap of U2OS cells inducibly expressing GFP–CtIP-wt and treated with HU (2 mM, 4 h). Where indicated, cells were treated with the p38α PROTAC NR-11c (1 μM, 24 h before HU). Whole-cell lysates (input) and immunoprecipitates were analysed by western blotting using specific antibodies. Densiometric quantification of CtIP–pS276 band in the GFP-Trap for four independent experiments is shown on the right (% indicate CtIP–pS276 band intensity versus CtIP band intensity). Statistical significance was calculated using the unpaired t-test (****P ≤ 0.0001, **P ≤ 0.01, *P ≤ 0.05, ns, non-significant). (C) GFP-Trap of HEK293T cells co-transfected with GFP–PIN1 and indicated FLAG–CtIP variants. Twenty-four hours post-transfection, cells were either mock-treated or treated with the p38α inhibitor PH-797804 (1 μM) for 24 h. Whole-cell lysates (input) and immunoprecipitates were analysed by western blotting using specific antibodies. (D) Fork degradation was evaluated upon HU treatment in U2OS cells either treated with the p38α inhibitor PH-797804 (1 μM, 24 h before labelling) or with the p38α PROTAC NR-11c (1 μM, 24 h before labelling). Western blotting of lysates from the same experiment is shown below. (E) Fork degradation was evaluated upon HU treatment in U2OS cells pre-treated with the p38α inhibitor PH-797804 (1 μM, 24 h before labelling) alone or in combination with either the Mre11 inhibitor Mirin (25 μM, simultaneously with HU) or the DNA2 inhibitor NSC-105808 (2 μM, simultaneously with HU). (F) Fork degradation was evaluated upon HU treatment in U2OS cells inducibly expressing siCtIP-resistant GFP–CtIP-wt or –S276A/P277A trans-locked mutant and depleted of endogenous CtIP. In addition, cells were either mock-treated or treated with p38α inhibitor PH-797804 (1 μM, 24 h before labelling) or with the p38α PROTAC NR-11c (1 μM, 24 h before labelling). (D–F) Box and whisker plots of IdU/CldU-tract length ratios for individual replication forks are shown. Numbers indicated above the individual plots represent the mean ratios ± standard deviation. Statistical significance was calculated using the Kruskal–Wallis test with Dunn’s multiple comparisons (****P ≤ 0.0001, **P ≤ 0.01, *P ≤ 0.05, ns, non-significant). Schematics of the CldU/IdU pulse-labelling protocol are shown (top).

Figure 4.

PIN1 and p38α activities are required for CtIP accumulation at stalled replication forks. (A) CtIP SIRF assay in U2OS cells pulsed-labelled with EdU (25 μM) for 10 min followed by treatment with HU (2 mM) for 4 h. Where indicated cells were treated with the PIN1 inhibitor KPT-6566 (10 μM, 1 h before EdU labelling). (B) CtIP SIRF assay in U2OS cells pulsed-labelled with EdU (25 μM) for 10 min followed by treatment with HU (2 mM) for 4 h. Where indicated cells were treated with the p38α inhibitor PH-797804 (1 μM, 24 h before EdU labelling). (C) GFP–CtIP SIRF assay in U2OS cells inducibly expressing siCtIP-resistant GFP–CtIP-wt, –S276A, or –S276A/P277A trans-locked mutant and depleted of endogenous CtIP. Cells were pulsed-labelled with EdU (25 μM) for 10 min followed by treatment with HU (2 mM) for 4 h. (D) GFP–CtIP SIRF assay in U2OS cells inducibly expressing siCtIP-resistant GFP–CtIP-wt or –S276A/P277A trans-locked mutant and depleted of endogenous CtIP. In addition, cells were either mock treated or treated with the PIN1 inhibitor KPT-6566 (10 μM, 1 h before EdU labelling) or the p38α inhibitor PH-797804 (1 μM, 24 h before EdU labelling). Cells were pulsed-labelled with EdU (25 μM) for 10 min followed by treatment with HU (2 mM) for 4 h. (A–D) Dot plots show the number of PLA foci and the median from at least 120 EdU-positive cells. Statistical significance was calculated using the Kruskal–Wallis test with Dunn’s multiple comparisons (****P ≤ 0.0001, **P ≤ 0.01, ns, non-significant). Representative images are shown on top of each figure. Scale bars: 10 μm.

Figure 5.

PIN1 or p38α inhibition impairs CtIP accumulation at stalled forks and overcomes Olaparib resistance in Brca1^−/− tumour cells. (A) CtIP SIRF assay in KB1P-G3 (Trp53^−/−; Brca1^−/−) and (Trp53^−/−; Brca1^−/−; H2afx^−/−) cells, either mock-treated or treated with the PIN1 inhibitor KPT-6566 (10 μM, 1 h before EdU labelling), or with the p38α inhibitor PH-797804 (1 μM, 24 h before labelling). Cells were pulse-labelled with EdU (25 μM) for 10 min followed by treatment with HU (8 mM) alone or in combination with the PIN1 or p38α inhibitors for 6 h. Dot plots show the number of PLA foci and the median from at least 150 EdU-positive cells. Statistical significance was calculated using the Kruskal–Wallis test with Dunn’s multiple comparisons (****P ≤ 0.0001, **P ≤ 0.01, ns, non-significant). Representative images are shown on the right. Scale bars: 10 μm. (B) Colony formation assay was performed in same cells as in panel (A), either mock-treated or treated with the PIN1 inhibitor KPT-6566 (2.5 μM) and with the PARPi Olaparib (75 nM) for 10 days. (C) Colony formation assay was performed in same cells as in panel (A), either mock-treated or treated with the p38α inhibitor PH-797804 (10 μM) and with the PARPi Olaparib (75 nM) for 10 days. (B and C) Plotted values are mean ± standard deviation of three biological replicates. Statistical significance was calculated using the unpaired t-test (****P ≤ 0.0001, **P ≤ 0.01, ns, non-significant). Representative images are shown (top).

Figure 6.

Schematic model depicting the role of PIN1–p38α–CtIP signalling axis in fork protection. During unperturbed S-phase, CDK2-mediated phosphorylation of T315 promotes PIN1 binding to CtIP. In response to replication stress, p38α kinase phosphorylates CtIP at S276. Subsequently, PIN1 catalyses the cis-to-trans isomerization of the pS276–P277 peptide bond, ensuring accumulation of CtIP at stalled forks. Ultimately, this phosphorylation-isomerization cascade promotes CtIP-dependent protection of nascent DNA from DNA2-mediated nucleolytic processing, thereby maintaining genome stability. Created in BioRender (https://BioRender.com/pfv2qy0).