CHK2 regulates MUS81-dependent DSBs in response to replication stress and BRCA2 deficiency

Eva Malacaria; Carolina Figlioli; Anita Palma; Masayoshi Honda; Pasquale Valenzisi; Marialuisa Casella; Serena Camerini; Fabio Pucci; Sara Rinalducci; Maurizio Semproni; Maria Spies; Annapaola Franchitto; Pietro Pichierri

doi:10.1093/nar/gkag766

. 2026 Aug 13;54(15):gkag766. doi: 10.1093/nar/gkag766

CHK2 regulates MUS81-dependent DSBs in response to replication stress and BRCA2 deficiency

Eva Malacaria ^1,^b, Carolina Figlioli ^2,^b, Anita Palma ³, Masayoshi Honda ⁴, Pasquale Valenzisi ⁵, Marialuisa Casella ⁶, Serena Camerini ⁷, Fabio Pucci ^8,⁹, Sara Rinalducci ¹⁰, Maurizio Semproni ¹¹, Maria Spies ¹², Annapaola Franchitto ¹³, Pietro Pichierri ^14,^15,^✉

¹ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

² Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

³ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

⁴ Department of Biochemistry and Molecular Biology, University of Iowa Carver College of Medicine, 51 Newton Road, Iowa City, IA 52242, United States

⁵ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

⁶ Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

⁷ Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

⁸ Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

⁹ Istituto Nazionale Biofisica e Biostrutture (INBB), Viale delle Medaglie d’Oro, 305, Rome 00136, Italy

¹⁰ Proteomics lab, Department of Ecology and Biology, Università della Tuscia, Largo dell’Università, Viterbo 01100, Italy

¹¹ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

¹² Department of Biochemistry and Molecular Biology, University of Iowa Carver College of Medicine, 51 Newton Road, Iowa City, IA 52242, United States

¹³ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

¹⁴ Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy

¹⁵ Istituto Nazionale Biofisica e Biostrutture (INBB), Viale delle Medaglie d’Oro, 305, Rome 00136, Italy

^✉

To whom correspondence should be addressed. Email: pietro.pichierri@iss.it

The first two authors should be regarded as Joint First Authors.

Roles

Fabio Pucci: Formal analysis, Investigation

PMCID: PMC13469066 PMID: 42598853

Abstract

MUS81 is a structure-specific endonuclease that processes DNA intermediates during mitosis and in S-phase following replication stress. It plays a crucial role in cleaving deprotected reversed forks in BRCA2-deficient cells. However, how MUS81 is regulated during replication stress in human cells remains unknown. Our study reveals that CHK2 binds to the MUS8–EME2 complex in S-phase through the FHA domain and positively regulates the formation of DSBs in response to persistent or pathological replication stress. Mechanistically, our cellular and biochemical data identify a dual-site phosphorylation regulatory mechanism involving priming at the CDK2 site S95 followed by CHK2-dependent modification of the activatory site S97. Phosphorylation of MUS81 does not affect recruitment to the stalled forks but is crucial for the replication stress-dependent association with SLX4 in S-phase. At deprotected forks, in BRCA2-depleted cells, the CHK2–MUS81 complex assembles downstream of fork reversal and degradation, and CHK2-dependent phosphorylation is essential for replication fork recovery and viability. Together, our findings elucidate a novel regulatory mechanism of the MUS81 complex in S-phase and reveal a previously unrecognized role of the ATM–CHK2 axis in responding to persistent replication fork arrest or fork deprotection in the absence of BRCA2.

Graphical Abstract

For image description, please refer to the figure legend and surrounding text.

Introduction

The replication stress response is essential for maintaining genome integrity and acts as a barrier to cancer development in human cells [1–4]. A key function of this response is to protect replication forks from excessive degradation and collapse [3, 5]. To achieve this, perturbed replication forks are targeted by various proteins that coordinate checkpoint activation with fork remodelling and protection during prolonged arrest [6–8]. A critical step in fork protection is the formation of reversed forks [8]. This reversal process involves several enzymes, notably SMARCAL1, as well as the homologous recombination factors RAD51 and BRCA2 [7]. BRCA2 plays a central role in homologous recombination by promoting RAD51 nucleofilament assembly [9] and is frequently mutated in familial breast cancers [10, 11]. BRCA2 and RAD51 are also required to protect reversed forks from unscheduled degradation by the nucleases such as MRE11 and EXO1 [7, 8]. Loss of BRCA2 impairs RAD51-mediated fork protection, leading to pathological degradation of nascent DNA strands. This degradation ultimately triggers recombination-based replication restart via break-induced replication pathways [12–14].

Pathological degradation and cleavage of reversed forks have been linked to the sensitivity of BRCA2-deficient cells to cisplatin and PARP inhibitors [15–17]. Therefore, understanding the mechanism underlying this process and its regulation is highly relevant for cancer therapy. However, the regulatory pathways governing pathological fork recovery in the absence of BRCA2 remain poorly understood.

In BRCA2-deficient cells, a critical step in the processing of unprotected replication fork is the cleavage of a flap DNA intermediate by the MUS8–EME2 complex, acting downstream of MRE11/EXO1 activity [13, 18]. The function of MUS81 in BRCA2-deficient cells is essential for replication fork recovery and for promoting cell viability under both normal and perturbed replication conditions [13, 18].

MUS81 is a structure-specific endonuclease that cleaves recombination and replication DNA intermediates, generating Double Strand Breaks (DSBs). In yeast, MUS81 forms a heterodimeric complex with Eme1(Mms4), while in human cells, MUS81 can also heterodimerize with EME2, particularly in S-phase to process stalled replication forks [19]. MUS81-dependent cleavage needs to be tightly regulated throughout the cell cycle to prevent unscheduled DSB formation during S-phase and to ensure proper resolution of DNA structures in G2/M [20]. In human cells, the activity of the MUS81–EME1 complex during G2/M is regulated by CDK1, PLK1, and CK2 [20, 21]. In contrast, little is known about how the MUS81 complex is regulated during replication, especially in response to replication stress and under pathological conditions such as BRCA2 deficiency.

In fission yeast, Mus81/Eme1 function is regulated by the kinases Cdc2^CDK1 and Chk1 in response to DNA damage, while the checkpoint kinase Cds1^CHK2 negatively regulates MUS81 during S-phase by interacting with it through its FHA domain [22–25]. In human cells, both CHK1 and CHK2 are activated in response to replication stress. However, CHK1 is considered the functional homologue of Cds1, despite CHK2 being its structural counterpart [26]. Moreover, MUS81 is required to generate DSBs and support viability in cells depleted of the ATR-CHK1 checkpoint [27].

Fork reversal and degradation generate free DNA ends that may activate CHK2 via ATM, in addition to ATR. Nevertheless, the involvement of CHK2 in the pathological processing of perturbed forks upon replication stress, or in the absence of BRCA2, remains poorly characterized. It is also unknown whether a crosstalk between MUS81 and CHK2 exists in human cells.

Therefore, in this study, we combined biochemical and single-cell assays to investigate whether CHK2 regulates the formation of MUS81-dependent DSBs at replication forks, and to determine whether CHK2 and MUS81 functionally interact during sustained replication arrest or in the absence of BRCA2.

Our study demonstrates that human CHK2 and MUS81 physically interact through the FHA domain, and that the I157T mutation significantly impairs this interaction, whereas the R117A mutation does not. We further show that CHK2 contributes to the regulation of MUS81-dependent DSBs formation at replication forks during prolonged arrest or destabilization due to BRCA2 deficiency. In addition, we identify specific MUS81 residues phosphorylated by CHK2 that are functionally required for DSB formation at reversed replication forks in BRCA2-deficient cells.

These findings uncover a critical regulatory mechanism governing the MUS81 complex at demised replication forks and indicate that CHK2 activity, and its interaction with MUS81, controls the cleavage of reversed forks in the absence of BRCA2.

Materials and methods

Cell lines and culture conditions

The MRC5SV40 cells were maintained in DMEM (Life Technologies) supplemented with 10% foetal bovine serum (FBS; Boehringer Mannheim) and incubated at 37°C in a humidified 5% CO₂ atmosphere. Stable MRC5SV40 shMUS81 were described in Palma et al. [21]. The wild-type form of RNAi-resistant MUS81 ORF cloned into the pEF1a-IRES-NEO plasmid was subjected to Site-Directed mutagenesis (SDM, Quickchange II XL-Stratagene) to introduce the T86 or S95, S97 mutations. Sequence-verified plasmids were then transfected into MRC5 shMUS81 cells by the Neon Nucleofector (Life technologies). To generate shBRCA2 inducible cell lines, MRC5SV40 were transduced with lentivirus expressing an shBRCA2 cassette under the control of a doxycycline-regulated promoter at 0.5 of multiplicity of infection (Dharmacon SmartVector inducible lentivirus, sequence code V3SH7669-225202141; [28]). After puromycin selection at 300 ng/ml, a single clone was selected and used throughout the study.

To obtain MUS81 knockout, two RNP complex obtained with Alt-R CRISPR–Cas system (IDT) specific to exon 1 of MUS81 were nucleofected using the Neon Trasfection System. For targeting of MUS81 gene the oligos were 5′TCCTACAGCACTTCGGAGAC3′ and 5′CATGCCCCGGACTCACCATC3′ [28].

Non-inducible MRC5 shBRCA2 were obtained infecting MRC5 with MISSION® Lentiviral Transduction Particles NM_000059 by Sigma–Aldrich; cells were selected by 1 µg/ml puromycin (Gibco, A11138-03).

AAVS1 EF1A alfaTAG-MUS81 FLAG plasmid was ordered on VectorBuilder.com.

Transfections

RNA interference against MUS81 was performed using FlexiTube HsMUS81 6 (Qiagen) at final concentration of 10 nM using Lullaby 48 h before to perform experiments. BRCA2 was depleted with ON-TARGETplus Human BRCA2-SMARTpool (Dharmacon) at 25 nM using Interferin (Polyplus) and experiments were performed after 48 h of transfection. RNA interference against EME1/2 was performed with SMARTpool siRNAs (Dharmacon, geneID 197342).

The efficiency of protein depletion was monitored by western blotting 72 h after transfection. MRC5, MRC5 shMUS81, shMUS81, and shMUS81 S97A/D cell lines were transfected with pcDNA3-HA-Chk2 (wild type or with CHK2 mutation) using Neon transfection system 48 h prior to perform experiments. MUS81 WT and mutants were transfected in MRC5 MUS81^KO and MRC5 MUS81^KO shBRCA2 48 h prior to perform experiments.

Chemicals

Hydroxyurea (HU) was added to culture medium at 2 mM from stock solutions 20 mM prepared in phosphate-buffered saline solution (PBS) to induce DNA replication arrest or DNA damage. The inhibitor of MRE11 exonuclease activity (MIRIN, Calbiochem) was added to DMEM 15 min prior HU at 50 µM. Nocodazole (NOC; Sigma–Aldrich) was used at 0.5 μg/ml. The CHK2 inhibitor (Bml-277, Calbiochem) was used at 10 µM. CldU (Sigma–Aldrich) was dissolved in sterile water at 200 mM stock solution and used at 50 μM. IdU (Sigma–Aldrich) was dissolved in sterile DMEM as a stock solution 2.5 mM and stored at −20°C.

Phosphorylation site prediction

For the prediction of MUS81 phosphorylation sites, the MUS81 protein sequence was scanned using the GPS 6.0 software with a medium threshold that consist in a <6% of false identification rate (GPS 6.0 phosphorylation prediction).

Phosphopeptide enrichment

Purification of phosphopeptides was then performed according to Thingholm et al. [29]. Briefly, tryptic peptides were diluted five-fold in dihydroxybenzoic acid (DHB) buffer [350 mg/ml DHB, 80% (v/v) acetonitrile (ACN), 2% (v/v) TFA] and applied to TiO₂ beads (200 μg) pre-equilibrated in 50% ACN. The sample was then washed once in DHB buffer, before being washed two times with wash buffer [80% ACN (v/v), 2% TFA (v/v)] to remove the DHB. The sample was finally eluted with 25 μl of 2.5% ammonium hydroxide solution (pH ≥ 10.5) and immediately neutralized with 2.5 μl of formic acid (FA). All buffers used ultrapure water and were made fresh on the day of experimentation.

LC–ESI–CID/ETD–MS/MS

The TiO₂-enriched samples were analysed using a split-free nano-flow liquid chromatography system (EASY-nLC II, Proxeon, Odense, Denmark) coupled to a 3D-ion trap (model AmaZon ETD, Bruker Daltonik, Germany) equipped with an online ESI nano-sprayer (the spray capillary was a fused silica capillary, 0.090 mm o.d., 0.020 mm i.d.). A sample volume of 15 μl was loaded by the autosampler onto a homemade 2 cm fused silica precolumn (100 μm I.D.; 375 μm O.D.; Reprosil C18-AQ, 5 μm, Dr. Maisch GmbH, Ammerbuch-Entringen, Germany). Sequential elution of peptides was accomplished using a flow rate of 300 nl/min and a linear gradient from Solution A (2% ACN; 0.1% FA) to 50% of Solution B (98% ACN; 0.1% FA) in 40 min over the precolumn in-line with a homemade 15 cm resolving column (75 μm I.D.; 375 μm O.D.; Reprosil C18-AQ, 3 μm, Dr. Maisch GmbH, Ammerbuch-Entringen, Germany). To identify phosphorylation sites, two types of peptide fragmentation were carried out in parallel in the mass spectrometer: (i) collision induced dissociation (CID); (ii) electron transfer dissociation (ETD). When CID was used, a MS² was automatically performed on the three most intense mass spectrometry (MS) ions, and MS³ was triggered if one of the top three MS² peaks corresponded with neutral loss of 98.0, 49.0, 32.7 m/z. For ETD experiments, the reaction time was set to 100 ms using a reactant ICC of 500 000 allowing a maximum accumulation time for the reactant ion of 10 ms. Acquired MS/MS spectra were processed in DataAnalysis 4.0, and submitted to Mascot search program (Matrix Science, London, UK). The following parameters were adopted for database searches: SwissProt database; taxonomy=homo sapiens; peptide mass tolerance of ±0.3 Da; fragment mass tolerance of ±0.3 for CID ions and of ±1.3 Da for ETD ions; enzyme specificity trypsin with one missed cleavages considered; fixed modifications: carbamidomethyl (C); variable modifications: oxidation (M) and phosphorylation (ST). Phosphopeptide identifications were accepted if the Mascot score was over the 95% confidence limit based on the ‘identity’ score of each peptide. A delta ion score was calculated of all phosphopeptides containing more than one serine or threonine residues by taking the difference between the two top ranking Mascot ion scores. Phosphorylation site assignments with a delta score >5 were automatically accepted [30]. All fragmentation spectra with delta score ≤5 were manually inspected as to whether the phosphorylation sites were unambiguously determined or not.

Hydrolysis ‘in situ’ and mass spectrometry analysis

The MUS81 experiment was prepared under three different conditions (untreated, +4 h HU, and +4 h HU + inhibitor), with four replicates for each condition, for a total of 12 samples. Each replicate was resolved on a polyacrylamide gel electrophoresis. A protein band of interest at ~60 kDa was excised from each lane and subjected to in-gel protein digestion. The gel pieces were washed and destained, treated with 10 mM dithiothreitol (DTT) in 50 mM ammonium bicarbonate (AMBIC) for 45 min at 56°C and with 55 mM iodoacetamide in 50 mM AMBIC for 30 min at Room Temperature (RT) in the dark. In-gel digestion was performed by incubation with 0.033 µg/μl Glu-C (Thermo Fisher Scientific) for 45 min in ice and then overnight at 25°C. The following day, an additional peptide extraction step was performed using a mixture of 50% ACN and 2.5% FA to maximize peptide recovery. The resulting peptide mixture was dried in a Speedvac, dissolved in a solution containing 5% of ACN and 0.1% of FA, and injected into an Ultimate 3000 UHPLC (Dionex, Thermo Fisher Scientific, CA, USA) coupled with an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific, CA, USA). Peptides were desalted on a trap column (Acclaim PepMap 100 C18, Thermo Fisher Scientific) and then separated on a 35-cm long silica capillary (Silica Tips, MSWil, The Netherlands), packed in-house with a C18 resin, 1.9 μm as particle size and 100 Å as pore size (Michrom BioRe- sources, C A, USA). The analytical separation was developed by a 60-min gradient of buffer A (5% of ACN and 0.1% of FA) and buffer B (95% of ACN and 0.1% of FA). Buffer B increased from 5% to 8% in 5 min, up to 32% in 35 min, then to 80% in further 5 min, and followed by 4 min and 10 min long washing and equilibrating steps, respectively.

MS analysis was performed using parallel reaction monitoring (PRM) acquisition method picking the m/z values relative to the precursor and product ions, as shown in the Table 1. Precursor and fragment ions were selected based on previously data-dependent acquisition results.

Table 1.

List of settings used for PRM analysis

Peptide	State	MH⁺	Charge	Precursor (m/z)	Product (m/z)
80–99	Unmodified	2160.03	z = 4	540.76	562.28
80–99	Unmodified	2160.03	z = 5	432.81	422.22
	Phosphorylated	2240.02	z = 3	747.34	632.83
	Phosphorylated	2240.02	z = 4	560.76	585.29
80–110	Unmodified	3280.59	z = 5	656.92	770.41
	Phosphorylated	3360.58	z = 4	840.90	770.41
	Phosphorylated	3360.58	z = 5	448.80	770.41

Open in a new tab

Raw data were searched against MUS81 sequence using the Proteome Discoverer 2.4 version (Thermo Fisher Scientific, CA, USA) by using semi-specific Glu-C cleavage with a maximum of three missed cleavages, 10 ppm and 0.02 Da tolerance for precursor and fragment ions, respectively, carbamidomethyl (C) as fixed modification and oxidation (M), acetylation (N-Term), Met-loss (M), and phosphorylation (S,T,Y) as variable modifications. MS/MS spectra were also manually inspected. Freestyle 1.8 software (Thermo Fisher Scientific) was employed with the PPD detection algorithm to measure the area of the transitions between precursor and their relative fragment ions as described in Table 1. The ratio between the summed areas of the phosphorylated peptide transitions and the summed areas of the phosphorylated and unmodified peptide transitions were calculated for each replicate. These values were then averaged across the replicates.

All MS data were deposited to the ProteomeXchange Consortium via the MassIVE repository with the dataset identifier MassIVE MSV000101714

Production and purification of GST-fused fragments

The FHA domain of CHK2 was amplified by PCR from the pcDNA3-HA-CHK2 plasmid containing the full-length wild-type or mutant CHK2 ORF (a gift from Dr Domenico Delia’s lab). PCR-amplified DNA was cloned into the GST-pGEX-2TK plasmid (Stratagene). GST-FHA fragment was expressed into E.coli BL21PlysS at 30°C for 4 h. Bacterial pellets were lysed in BER reagent (Pierce) supplemented with DNase and lysozyme, as indicated by the manufacturer, and purified by incubation with GSH-magnetic beads (Promega) after extensive washings, as reported previously [21]. Quantification of the magnetic-beads-bound fragments was performed after sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and Coomassie staining against serial dilutions of purified bovine serum albumin (BSA). Magnetic beads-bound GST-FHA fragments were then used as substrates for pull-down and in vitro kinase assays.

Western blot analysis

Blots were incubated with the primary antibodies in Table 2.

Table 2.

List of antibodies used in the study

Antibody	Source	Cat.#	WB diluition	PLA diluition
MUS81	Santa Cruz Biotechnologies	SC-376661	–	1/300
MUS81	Santa Cruz Biotechnologies	SC-47692	1/1000	–
CHK2	Santa Cruz Biotechnologies	SC-9064	1/1000	–
CHK2	Cell Signaling	3440S	–	1/100
LAMIN B1	Abcam	16048	1/10 000	–
BRCA2	Bethyl	A303434A	1/1000	–
TOPOIIb	SantaCruz	25330	1/1000	–
GAPDH	Millipore	MAB374	1/1000	–
HA tag	Cell Signaling	3724S	1/1000	1/150
SLX4	Bethyl	A302-269A	1/2000	–
pT86CHK2	Cell Signaling	2661S	1/2000
SMARCAL1	Abcam	AB154226	1/2000	1/100
eGFP	Santa Cruz Biotechnologies	SC-9996	1/1000	–
pS1981ATM	Cell Signaling	13050S	–	1/100
Cyclin A	Santa Cruz Biotechnologies	SC-271682	1/1000	–
EME2	Santa Cruz Biotechnologies	SC-164270	1/1000	–
ATM	Cell Signaling	92 356	1/1000
Biotin	Bethyl	A150-109A	–	1/2000
Biotin	Jackson Immunoresearch	200-002-211	–	1/2000
Alfa-TAG	NanoTag	N1583	–	1/100
FLAG	Sigma	F7425	–	1/2000
FLAG	Sigma	F1804	1/1000	1/100
EME1	Santa Cruz Biotechnologies	SC-53275	1/1000	–

Open in a new tab

After incubations with horseradish peroxidase-linked secondary antibodies (Jackson ImmunoResearch, 1:30 000), the blots were developed using the chemiluminescence detection kit ECL-Plus (Amersham) according to the manufacturer’s instructions. Quantification was performed on blot acquired by ChemiDoc XRS+ (Bio-Rad) using Image Lab software, and values shown on the graphs represent a normalization of the protein content evaluated through Lamin B1 or GAPDH immunoblotting.

Immunofluorescence

Cells were grown on 35-mm coverslips and harvested at the indicated times after treatments. For immunofluorescence, after further washing with PBS, cells were pre-extracted with 0.5% Triton X-100 and fixed with 3% paraformaldehyde (PFA)/2% sucrose at RT for 10 min. After blocking in 3% BSA for 15 min, staining was performed with mouse monoclonal anti-phospho-ATM Ser1981 diluted in a 1% BSA/0.1% saponin in PBS solution, for 1 h at 37°C in a humidifier chamber. After extensive washing with PBS, Alexa Fluor® 488 conjugated-goat anti-mouse was used at 1:200 and applied for 1 h at RT followed by counterstaining with 0.5 mg/ml of 4,6-diamidino-2-phenylindole (DAPI).

In situ PLA assay for protein–protein interaction

The in situ PLA NaveniFlex (Navinci Diagnostics) was performed according to the manufacturer’s instructions. The in situ PLA MUS81–CHK2 was performed with DuoLink, Merck was performed according to the manufacturer’s instructions. Samples were mounted in Prolong Gold anti‐fade reagent with DAPI (blue). Images were acquired randomly using Eclipse 80i Nikon Fluorescence Microscope, equipped with a Video Confocal (ViCo) system. Where indicated, proximity-ligation assay (PLA) spot counts were obtained using the ScanR software. At least 1500 cells were counted for each condition. Cells were gated according to the cell cycle phase that was identified by plotting the total intensity of DAPI on the mean intensity of EdU labelling generating a mask used in all the experimental points.

Quantitative in situ analysis of protein interactions at DNA replication forks (SIRF assay)

Exponential growing cells were seeded onto microscope chamber slide. The day of experiment, cells were incubated with 125 µM EdU for 8 min and treated as indicated. After treatment, cells were pre-extracted in 0.5% Triton X-100 for 5 min on ice and fixed with 3% PFA, 2% sucrose in PBS 1X for 15 min at RT. Cells were permeabilized in 0.25% Triton X-100 in PBS for 15 min at RT. Click reaction cocktail was added into slides (30 μl/well) and incubated at room temperature 30 min in a humid chamber protected from light. Slides were washed with PBS and blocked in 3% BSA/PBS for 20 min. The EdU detection was applied the Click-iT™ EdU Alexa Fluor™ Imaging Kit (Invitrogen) using 5 mM Biotin-Azide for 30 min at RT. The concentrations of primary antibodies were indicated in Table 2. The negative controls were obtained by using each single primary antibody or KO cells. Samples were incubated with secondary antibodies conjugated with PLA probes MINUS and PLUS: the PLA probe anti‐mouse PLUS and anti‐rabbit MINUS (Duolink®, Sigma–Aldrich). The incubation with all antibodies was accomplished in a humidified chamber for 1 h at 37 °C. Next, the PLA probes MINUS and PLUS were ligated using two connecting oligonucleotides to produce a template for rolling‐cycle amplification. After amplification, the products were hybridized with red fluorescence‐labelled oligonucleotide. Samples were mounted in reagent with DAPI (blue).

For p1981ATM SIRF 10.000 cells were plated onto chamber slides followed by 48 h of doxycycline induction. The day of experiment, cells were incubated with 125 µM EdU for 8 min, followed 2 mM HU as indicated. After treatments, cells were fixed with 2% PFA/PBS for 15 min at RT. Cells were permeabilized in 0.25% Triton X-100 in PBS for 15 min at RT. Click reaction cocktail was added into slides (30 μl/well) and incubated at room temperature 30 min in a humid chamber protected from light. Slides were washed with PBS and blocked in 3% BSA/PBS for 20 min. After washing with PBS, cells were incubated with the primary antibodies. The primary antibodies were indicated in the Table 2 at the end of the paragraph. Samples were incubated with secondary antibodies conjugated with PLA probes MINUS and PLUS: the PLA probe anti‐mouse PLUS and anti‐rabbit MINUS (OLINK Bioscience). The incubation with all antibodies was accomplished in a humidified chamber for 1 h at 37°C. Next, the PLA probes MINUS and PLUS were ligated using two connecting oligonucleotides to produce a template for rolling‐circle amplification. After amplification, the products were hybridized with red fluorescence‐labelled oligonucleotide. Samples were mounted in Prolong Gold anti‐fade reagent with DAPI (blue). Images were acquired randomly using Eclipse 80i Nikon Fluorescence Microscope, equipped with a Video Confocal (ViCo) system.

DNA fibre assay

Cells were pulse-labelled with 50 µM CldU for 20 min, washed twice with PBS, and treated with HU 2 mM as reported in the experimental schemes. After treatments cells were washed twice with PBS and incubated for 30 min in fresh medium with 250 μM IdU. DNA fibres were prepared and analysed to evaluate fork restart. For immunodetection of labelled tracks, the following primary antibodies were used: rat anti-CldU/BrdU (Abcam) and mouse anti-IdU/BrdU (Becton Dickinson).

Chromatin isolation

Cells (4 × 10 ⁶ cells/ml) were resuspended in buffer A [10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.34 M sucrose, 10% glycerol, 1 mM DTT, 50 mM sodium fluoride, protease inhibitors (Roche)]. Triton X-100 (0.1%) was added, and the cells were incubated for 5 min on ice. Nuclei were collected in pellet by low-speed centrifugation (4 min, 1300 × g, 4°C) and washed once in buffer A. Nuclei were then lysed in buffer B [3 mM ethylenediaminetetraacetic acid (EDTA), 0.2 mM EGTA, 1 mM DTT, protease inhibitors]. Insoluble chromatin was collected by centrifugation (4 min, 1700 × g, 4°C), washed once in buffer B + 50 mM NaCl, and centrifuged again under the same conditions. The final chromatin pellet was resuspended in 2× Laemmli buffer and sonicated for 15 s in a Tekmar CV26 sonicator using a microtip at 25% amplitude.

Neutral comet assay

DNA breakage induction was evaluated by Comet assay (single-cell gel electrophoresis) in non-denaturing conditions. Briefly, dust-free frosted-end microscope slides were kept in methanol overnight to remove fatty residues. Slides were then dipped into molten low melting point (LMP) agarose at 0.5% and left to dry. Cell pellets were resuspended in PBS and kept on ice to inhibit DNA repair. Cell suspensions were rapidly mixed with LMP agarose at 0.5% kept at 37°C and an aliquot was pipetted onto agarose-covered surface of the slide. Agarose-embedded cells were lysed by submerging slides in lysis solution [30 mM EDTA, 0.1% sodium dodecyl sulphate (SDS)] and incubated at 4°C, 1 h in the dark. After lysis, slides were washed in Tris Borate EDTA (TBE) 1× running buffer (Tris 90 mM; boric acid 90 mM; EDTA 4 mM) for 1 min. Electrophoresis was performed for 20 min in TBE 1× buffer at 1 V/cm. Slides were subsequently washed in distilled H₂O and finally dehydrated in ice-cold methanol. Nuclei were stained with GelRed (1:1000) and visualized with a fluorescence microscope (Zeiss), using a 60× objective, connected to a CCD camera for image acquisition. At least 200 comets per cell line were analysed using CometAssay IV software (Perceptive instruments) and data from tail moments processed using Prism software. Apoptotic cells (smaller Comet head and extremely larger Comet tail) were excluded from the analysis to avoid artificial enhancement of the tail moment.

Human MUS81–EME1 purification

Mammalian expression vector containing N-terminus FLAG-tagged MUS81 [21], MUS81S97A and MUS81S97D (this study), and N-terminus c-MYC tagged EME1 (Origene) were used as vector. HEK293T cells were grown in the high glucose Dulbecco’s Modified Eagle’s Medium (Gibco) supplemented with 10% FBS (GIBCO), 1 mM sodium pyruvate, 1% penicillin, and 1% streptomycin at 37°C in the presence of 5% CO₂. Cells were transiently transfected for 48 h using jetPRIME (Avantor) using same amount of MUS81 and EME1 vectors. Cells were washed with PBS (Gibco), then resuspended in the ice-cold lysis buffer [50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 10% glycerol, and 1 mM phenylmethanesulfonyl fluoride], and incubated at 4°C for 30 min. After centrifugation, clarified cell lysate was mixed with M2 anti-FLAG agarose beads (Sigma–Aldrich) and then incubated at 4°C for 2 h. Beads were then washed with lysis buffer, resuspended in the elution buffer [lysis buffer with 150 µg/ml of 3× FLAG peptides (Sigma–Aldrich)], and incubated at 4°C for 30 min. Eluted protein was immediately divided into small aliquots and preserved at −80°C.

MUS81–EME1 nuclease assay

Twenty nanomolar of purified Cy3-labelled fork DNA with 30-mer leading gap [32] mixed with the indicated concentrations of MUS81–, MUS81S97A–, and MUS81S97D–EME1 complex in 20 μl of standard reaction buffer, containing 30 mM Tris–acetate (pH 7.5), 100 mM NaCl, 5 mM magnesium chloride, 2 mM DTT, and 0.1 mg/ml BSA and incubated for 10 min at 25°C. Reactions were stopped and deproteinized by the addition of 10 μl of stop solution containing 0.5% (v/v) SDS, 0.5 mg/ml proteinase K, 20% glycerol, and 0.1% Orange-G and incubation at room temperature for 30 min. The reaction products were separated by electrophoresis on a 10% (29:1) native polyacrylamide gel in 1× TAE buffer containing 40 mM Tris–acetate (pH 8.0) and 1 mM EDTA for 3 h at 100 V. The gel was visualized and quantified using the ChemiDoc MP imaging system (BIO-RAD) by monitoring Cy3.

Viability assay

For viability assay, MRC5 MUS81^KO cells were transiently transfected with MUS81^WT and its mutant forms; 8 h post-transfection, cells were silenced for BRCA2. MRC5 were then seeded in 96-well plate, 750 cells/well, and 24 h later cells were treated with increasing doses of HU.

Cells were grown for 7 days in a CO₂ incubator at 37°C. Cell Counting Kit-8 (MedChem Express, HY-K0301) was used to test the cell response to HU treatments. Cells were incubated with 10 μl/100 μl well for 2 h and absorbance was measured at 450 nm using a luminescence microplate reader (Victor).

Results

CHK2–MUS81 interaction in response to replication stress involves the FHA domain of CHK2 and is disrupted by the I157T CHK2 mutation

To investigate the functional correlation between MUS81 and CHK2 in human cells, we first examined their potential physical association by performing immunoprecipitation experiments in HEK293T cells transiently transfected with GFP-MUS81 and HA-CHK2 expression plasmids. Given that MUS81 is involved in the resolution of DNA intermediates in S-phase during replication stress and in mitosis, cells were treated with HU or NOC to enrich for S-phase and mitotic population, respectively [21]. MUS81 was detected in anti-HA immunoprecipitates from asynchronous and mitotic cells, but HU treatment resulted in a five-fold increase in MUS81–CHK2 complex formation compared to untreated cells, and two-fold compared to NOC-treated cells (Fig. 1A).

To further validate the MUS81–CHK2 interaction, we used MUS81 KO MRC5 cells stably complemented with Flag-tagged wild-type MUS81 (MUS81^wt), which express MUS81 at near-endogenous levels [28]. CHK2 was readily detected in anti-Flag-MUS81 immunoprecipitates from MRC5^wt cells by an anti-MUS81 antibody, but not in immunoprecipitation from parental MUS81 KO cells lacking MUS81 expression (Fig. 1B). It is worth nothing that although the MUS81 appeared enriched in NOC-treated cells (see input) the amount of CHK2 found in complex with MUS81 was reduced in agreement with the Co-immunoprecipitation (co-IP) from HEK293T cells.

Next, we sought to confirm the MUS81–CHK2 interaction at the single-cell level using the PLA. PLA has several advantages over the CoIP since it represents an interaction taking place within the 40-nm range, which often corresponds to a direct protein–protein interaction. Furthermore, since PLA signals can be visualized in individual cells, quantitative imaging can correlate the presence of PLA spots with DNA content determining during which cell cycle stage the MUS81–CHK2 interactions occur. To this aim, MUS81 KO cells complemented with the Flag-tagged MUS81, which also contains an ALFA-tag for IF detection, were treated for 4 or 24 h with 2 mM HU and the PLA was performed using anti-ALFA-tag and anti-CHK2 antibodies. During automated analysis, PLA spots were counted in cells sorted by DNA content corresponding to S-phase. This approach allowed us to distinguish between true accumulation of spots induced by HU and accumulation resulting from cell cycle effects. As shown in Fig. 1C, some CHK2–MUS81 PLA spots were already detectable in S-phase of unchallenged cells. The number of PLA spots increased little after 4 h of HU but it was greatly enhanced after 24 h consistent with the expected rise of MUS81-dependent DSBs. Few, if any, PLA spots were detected when only one antibody was used in the assay, confirming the specificity of the observed CHK2–MUS81 interactions. These findings demonstrate that MUS81 and CHK2 are physically associated in cells, and this association increases under conditions that promote fork collapse, as evidenced by the rise in PLA spots after prolonged HU treatment. Furthermore, this interaction is detectable with endogenous proteins, supporting its physiological relevance.

In human cells, MUS81 can heterodimerize with either EME1 or EME2 [33]. While MUS81–EME1 complex is present throughout the cell cycle, interaction with EME2 is thought to occur predominantly during S-phase [19]. Despite the robust increase in CHK2–MUS81 complex formation after 24 h of HU, both EME1 and EME2 were detected in anti-CHK2 immunoprecipitates (Supplementary Fig. S1A). However, CoIP cannot discriminate between events taking place in or outside S-phase and thus we exploited in situ PLA coupled with quantitative imaging to analyse association between CHK2 and EME2, which should be the MUS81 partner in response to HU. Since no commercially available antibody against EME2 performs reliably in PLA, we transiently transfected MRC5 cells with Flag-tagged EME2. As shown in the western blot of Fig. 1D, the ectopic EME2 was expressed at roughly the same level as the endogenous EME2 (upper band versus lower band in the doublet). PLA spot counts were performed in S-phase and in G2/M-phase gated populations. Consistent with the CHK2–MUS81 PLA, the number of CHK2-EME2 PLA spots increased over time in S-phase in response to HU (Fig. 1D). Of note, although much more reduced in number, interaction between CHK2 and MUS81 was observed to increase, in a time-dependent manner, also in G2/M-phase gated cells (Fig. 1D). Thus, large part of the MUS81–CHK2 complex involves EME2 and occurs in S-phase cells. Consistent with this result, RNAi-mediated knockdown of EME1 failed to affect the number of CHK2–MUS81 PLA-positive nuclei at 24 h of HU treatment, while depletion of EME2, although suboptimal, reduced the number of PLA-positive nuclei indicating a diminished interaction (Supplementary Fig. S1B).

Then, we wanted to assess directly the prevalence of the MUS8–EME2 dimer in S-phase after 24 h of HU-induced replication stress. To do this, we performed MUS8–EME2 in situ PLA and analysis by quantitative imaging. As explained for the CHK2-EME2 PLA, we overcome the absence of reliable anti-EME2 antibodies for use in IF by using the Flag-tag of transiently expressed EME2. As shown in Fig. 1E, MUS8–EME2 PLA spots in S-phase gated cells were extremely increased by treatment with 24 h of HU, indicating enhanced interaction under replication stress conditions.

After confirming that CHK2 forms a complex with MUS8–EME2 under physiological conditions, we next investigated the functional relevance of this interaction. Prolonged replication fork arrest induced by HU is known to trigger MUS81-dependent cleavage of stalled forks and unresolved replication intermediates [34–36]. Since HU-induced replication stress enhances MUS81–CHK2 interaction, we examined whether CHK2 kinase activity promotes the formation of MUS81-dependent DSBs. To this end, we performed a neutral Comet assay on cells treated with HU for 24 h, in the presence or absence of the CHK2-specific inhibitor BML-277 (Fig. 1F). In cells expressing MUS81 HU treatment led to a marked accumulation of DSBs, which was significantly reduced upon CHK2 inhibition, indicating that CHK2 activity contributes to DSB formation under replication stress (Fig. 1F). In contrast, MUS81-depleted cells (shMUS81) showed minimal DSB accumulation following HU treatment, and this was unaffected by CHK2 inhibition, as expected (Fig. 1F). These findings suggest that CHK2 kinase activity functionally promotes MUS81-dependent DSB formation at stalled replication forks.

In yeast, Cds1 interacts with Mus81 via its FHA1 protein-docking domain [37]. To determine whether a similar mechanism operates in human cells, we investigated whether the interaction between CHK2 and MUS81 following replication stress depends on the CHK2-FHA domain. We used a mutant form of CHK2 carrying the I157T substitution in the FHA domain, a mutation known to severely impair protein–protein interactions without affecting CHK2 activation [38, 39]. Immunoprecipitation assays were performed in HEK293T shMUS81 cells co-transfected with Flag-MUS81 and HA-CHK2^wt or HA-CHK2^I157T. Following 24 h of HU, MUS81 was readily detected in anti-HA immunoprecipitates from cells expressing CHK2^wt, but not in mock-transfected controls (Fig. 2A). However, the interaction was abrogated almost completely in cells expressing the CHK2-I157T mutant, indicating that the FHA domain is required for MUS81 binding.

To further support the CoIP data, we analysed the formation of a CHK2–MUS81 complex by PLA in cells expressing the Flag-tagged wild-type or I157T CHK2 protein. Cells were treated with 2 mM HU for 24 h and PLA spots counted in S-phase gated cells by quantitative imaging. Although CHK2–MUS81 PLA spots were detected already in S-phase cells in the absence of treatment, they were greatly increased after 24 h of HU, as expected (Fig. 2B). Expression of the CHK2-I157T FHA mutant significantly reduced the number of PLA spots between MUS81 and CHK2, reversing the HU-induced accumulation. However, this mutant did not prevent the interaction between MUS81 and CHK2 in unchallenged S-phase cells (Fig. 2B). Given that PLA is a signal amplification technique, the absence of effect of the I157T mutant for the interaction with MUS81 in untreated conditions suggests that a very low level or metastable association with MUS81 does not involve the FHA domain of CHK2. In contrast, formation of a stable complex in response to HU requires either a functional FHA domain of CHK2 or a distinct CHK2 partner bound through the FHA domain. These findings were further supported by GST pulldown assays, in which the wild-type CHK2 FHA domain efficiently pulled down MUS81 from nuclear extracts, whereas the I157T mutation substantially reduced although did not abrogate binding (Supplementary Fig. S2A). Notably, another FHA domain residue, R117, is also critical for phospho-dependent interactions with known CHK2 partners, such as BRCA1 [39]. Surprisingly, mutation of this residue (CHK2-R117A) did not disrupt the interaction with MUS81 (Supplementary Fig. S2B), suggesting that the CHK2–MUS81 association may largely occur independently of phospho-dependent recognition, in contrast to the mechanism described in yeast.

Then, we investigated whether the interaction between the FHA-CHK2 domain and MUS81 is required to promote DNA breakage upon replication stress. To this aim, we performed a neutral Comet assay in MRC5 cells transfected with HA-tagged CHK2^wt, CHK2^I157T, or CHK2^R117A (Fig. 2C). Since both mutations severely impair CHK2 binding to multiple substrates, but only the I157T mutation disrupts its interaction with MUS81, we reasoned that a specific defect in CHK2–MUS81 complex formation would selectively suppress DSB formation in the CHK2^I157T-expressing cells. As predicted, overexpression of CHK2^I157T abolished DSB accumulation following HU-induced replication stress, whereas overexpression of CHK2^wt or CHK2^R117A did not affect DSB formation compared to control (empty vector-transfected) cells (Fig. 2C).

Together, these results demonstrate that MUS81 and CHK2 form a functional complex in response to replication stress-induced fork collapse. This interaction is mediated by the FHA domain of CHK2 and is specifically disrupted by the I157T mutation, but not by the R117A mutation. Furthermore, both pharmacological inhibition of CHK2 and disruption of its interaction with MUS81 via the I157T mutation impair DSB formation under replication stress, highlighting the importance of the CHK2–MUS81 axis in processing stalled replication forks.

Formation of the MUS81–CHK2 complex and CHK2 activity promotes DSBs and targets deprotected replication forks in the absence of BRCA2

The MUS81 complex targets replication forks following prolonged stalling and is also responsible for cleaving deprotected reversed forks in the absence of BRCA2 or other replication caretakers [12, 13, 27, 35, 40, 41]. Thus, we wanted to assess whether CHK2 associated to MUS81 was also in this condition. To this end, we performed CHK2–MUS81 in situ PLA followed by spot counts in S-phase gated nuclei by quantitative imaging using MUS81 KO MRC5 cells complemented with the ALFA-tagged wild-type MUS81 and stably depleted of BRCA2 with an shRNA (Fig. 3A). As shown in Fig. 3B, loss of BRCA2 greatly stimulated the formation of CHK2–MUS81 PLA spots in S-phase already in untreated cells. Interestingly, the number of PLA spots was increased at 4 h of HU treatment in the absence of BRCA2 and was two-fold higher at 24 h of HU if compared with the BRCA2-proficient parental cells (Fig. 3B). Since HU promoted association of CHK2 with the MUS8–EME2 complex in BRCA2-proficient cells (Fig. 1D), we investigated whether the same hold true in the absence of BRCA2. Thus, we performed CHK2-EME2 PLA at 4 or 24 h of HU treatment. Of note, the number of CHK2-EME2 PLA spots in S-phase cells observed in the absence of BRCA2 was constantly lower than that observed in BRCA2-proficient cells (Fig. 3C). Likewise, the number of MUS8–EME2 PLA spots detected in S-phase cells in the absence of BRCA2 after 4 h of HU was significantly lower than observed in BRCA2-proficent cells (Supplementary Fig. S3) although MUS81-dependent DSBs are higher in the absence of BRCA2 at early time-point after HU. These results indicate that the interaction between CHK2 and MUS81 is stimulated in the absence of BRCA2 in S-phase and in response to HU but suggest that, at deprotected forks, loss of BRCA2 might stimulate the formation of the MUS81–EME1 complex.

To determine whether the MUS81–CHK2 interaction contributes to DSB formation in BRCA2-deficient cells, we assessed the impact of CHK2 inhibition and impaired MUS81–CHK2 binding on this process. Specifically, we used the CHK2 inhibitor BML-277 and overexpression of the FHA-domain mutant CHK2-I157T to evaluate the roles of CHK2 activity and its association with MUS81, respectively (Fig. 4A and B). As expected, treatment with HU for 2 h induced a significant increase in DSBs in BRCA2-deficient cells, as detected by neutral Comet assay. However, co-treatment with the CHK2-Dox inhibitor markedly reduced DSB formation. Similarly, overexpression of the FHA CHK2-I157T mutant also significantly suppressed the accumulation of MUS81-dependent DSBs in BRCA2-depleted cells under the same conditions. Of note, both CHK2^I157T expression or MUS81 depletion reduced DSB formation, although the effect of CHK2-I157T appeared slightly more pronounced (Supplementary Fig. S4A). Consistent with a role for CHK2 in modulating MUS81 function at deprotected forks, we observed elevated levels of CHK2 activation (phosphorylation at T68) as well as ATM activation (phosphorylation at S1981) in BRCA2-deficient cells, particularly after 2 h of HU treatment (Supplementary Fig. S4B and C). To investigate whether the activation of the ATM–CHK2 axis in the absence of BRCA2 originates from stalled replication forks, we analysed the recruitment of active, S1981-phosphorylated, ATM at nascent DNA using SIRF assay in doxycycline-inducible shBRCA2 cells. Consistent with timing of the formation of MUS81-dependent DSBs, SIRF analysis revealed early recruitment of active ATM to replication forks in BRCA2-deficient cells (+Dox), with signal intensity increasing as early as 2 h of HU treatment and remaining sustained at all other time-points (Fig. 4C). In contrast, BRCA2-proficient cells (−Dox) showed elevated levels of fork-associated pS1981-ATM only at 4 h of HU treatment, which declined over time (Fig. 4C). These findings suggest that loss of BRCA2 leads to persistent ATM activation at stalled forks. Importantly, analysis of pS1981-ATM foci in chromatin-enriched fractions from MUS81-depleted cells indicated that ATM activation occurs independently of MUS81-mediated DSB formation irrespective of the BRCA2 status of the cells (Supplementary Fig. S4C).

In BRCA2-deficient cells, MUS81 activity is essential for cleaving regressed forks and enabling replication restart [13]. To investigate whether CHK2 contributes to this process, we assessed the impact of CHK2 inhibition or disruption of its interaction with MUS81 on replication fork recovery in BRCA2-depleted cells. Cells were either treated with the CHK2 inhibitor (CHK2i) or transfected with FHA-domain CHK2 mutants: I157T, which affects MUS81 binding and reduces DSBs, or R117A, which does not (Fig. 2 and Supplementary Fig. S2). Replication fork recovery was quantified by DNA fibre assay, measuring the number of stalled forks following sequential labelling with CldU and IdU. As previously reported, replication fork recovery was impaired in BRCA2-deficient cells and was largely dependent on MUS81 [13]. CHK2 inhibition in these cells led to a significant increase in the number of stalled forks after recovery from HU treatment, closely resembling the phenotype observed upon MUS81 depletion (Fig. 4D). Interestingly, expression of the FHA CHK2-I157T mutant also resulted in a similar accumulation of stalled forks, indicating that the disruption of CHK2–MUS81 interaction hampers fork restart. In contrast, the FHA CHK2-R117A mutant, which retains MUS81 binding and DSBs, had no effect on fork recovery. These findings indicate that CHK2 activation and its interaction with MUS81 are essential for the early formation of DSBs in BRCA2-depleted cells and for promoting subsequent replication fork recovery.

In BRCA2-deficient cells, the MUS81 complex targets deprotected forks after their reversal [13]. To determine whether the CHK2–MUS81 complex in BRCA2-deficient cells occurs downstream of fork reversal, we performed CHK2/MUS81 PLA after transient depletion of SMARCAL1, a key enzyme involved in fork reversal [42]. As expected, transfection with siBRCA2 oligos led to increased CHK2–MUS81 complex formation early after HU treatment compared to cells transfected with control (siCTRL) oligos (Fig. 4E). Co-depletion of SMARCAL1 and BRCA2 greatly reduced the MUS81–CHK2 interaction at all time-points, although SMARCAL1 depletion did not abolish the interaction completely (Fig. 4E).

To confirm the requirement of fork reversal, we pre-treated cells with Olaparib, a PARP inhibitor known to prevent this process [43]. Of note, Olaparib pre-treatment completely suppressed CHK2–MUS81 complex formation, as detected by PLA, in BRCA2-depleted cells at 2 and 4 h of HU treatment (Supplementary Fig. S5A). In BRCA2-deficient cells, MUS81 functions downstream of pathological degradation of reversed forks by MRE11/EXO1 and requires RAD52 [13, 14, 44, 45]. Consistent with this, CHK2–MUS81 interaction did not occur in HU-treated BRCA2-depleted cells treated with either Mirin or RAD52i (Supplementary Fig. S5B).

In the absence of a functional MUS81 complex, BRCA2-deficient cells accumulate increased number of resected but unbroken reversed forks, a phenotype that correlates with impaired replication restart [13]. We observed a similar phenotype when the CHK2–MUS81 interaction was disrupted, either by the I157A mutation or by CHK2 inhibition (Fig. 4D). Previously, we demonstrated that ectopically expressed bacterial RuvA, an avid binder of HJs, accumulates on chromatin under conditions that promote reversed fork formation [45]. Therefore, we used chromatin-bound RuvA as a surrogate marker of the presence of reversed forks in cells stably expressing a doxycycline-regulated shBRCA2 cassette. These cells were transiently transfected with CHK2 mutants, alone or in combination with MUS81-targeting siRNA, and treated with HU to induce replication fork stalling (Fig. 4F). The HU treatment induced chromatin accumulation of RuvA in BRCA2-proficient cells (WT cells; −Dox), consistent with the presence of HJ-like structures, such as reversed forks. Depletion of BRCA2 by DOX treatment slightly reduced RuvA chromatin association, while co-depletion of MUS81 restored an increased RuvA accumulation, consistent with a buildup of uncleaved reversed forks (Fig. 4F). Interestingly, in BRCA2-depleted cells, expression of the CHK2^I157T FHA mutant, which fails to bind MUS81, also led to an increase in chromatin-bound RuvA, but this effect was not further enhanced by MUS81 depletion.

Together, these results support a model in which the ATM–CHK2 signaling axis promotes MUS81-dependent DSB formation and that timely assembly of the CHK2–MUS81 complex is critical for the proper function of MUS81 at deprotected replication forks in BRCA2-deficient cells.

Furthermore, these results indicate that, in BRCA2-deficient cells, formation of the CHK2–MUS81 complex occurs after both fork reversal and degradation. This interaction likely facilitates proper targeting and processing of reversed forks downstream of their pathological resection.

Dual phosphorylation of MUS81 at S95 and S97 licenses DSB formation under replication stress

The FHA domain of CHK2 mediates recognition of its phosphorylated substrates [46], raising the possibility that MUS81 could be a direct CHK2 target in human cells. To investigate this, we performed an in vitro phosphorylation assay using the N-terminal region of human MUS81 (amino acids 76–206) as a substrate. This fragment encompasses a largely unstructured region that includes residues required for SLX4 binding [47, 48], as well as several putative CHK2 phosphorylation sites [21] and sites found phosphorylated in the public database Uniprot [49]. The purified GST-tagged MUS81 N-terminal fragment was incubated with recombinant CHK2 kinase, separated by SDS–PAGE, and analysed by MS/MS (Fig. 5A). Qualitative MS identified two major phosphorylation sites, T86 and S97, within the MUS81 fragment (Table 3 in Fig. 5 A, and Supplementary Figs S6 and S7). Remarkably, one of these sites, S97, is among the sites found phosphorylated in a phosphoproteomic screening in cancer cells [49], and reported in Uniprot. We next performed quantitative MS of the full-length MUS81 expressed in HEK293T cells comparing untreated condition, 24 h of 2 mM HU treatment, and 24 h of HU treatment with the CHK2i (Fig. 5B). Surprisingly, MS/MS analysis did not identify phosphorylation at T86 or at S97 but easily identified phosphorylation at S95 (Supplementary Figs S8 and S9). Phosphorylation at S95 was detected even under unperturbed conditions, was slightly increased after 24 h of HU treatment but barely reduced by CHK2 inhibition (Fig. 5B). Of note, also S95 was among the residues found phosphorylated in the phosphoproteomic screen performed in cancer cells that detected pS97 of MUS81 and matches the consensus sequence of Cyclin/CDKs [49]. Consistent with in silico prediction and MS data, a MUS81 fragment containing S95 and S97 was found phosphorylated in vitro by immunopurified CDK2/Cyclin A complex (Supplementary Fig. S10). Thus, we decided to investigate the functional relevance of S95 and S97 phosphorylation omitting T86 since this latter was identified in our CHK2 in vitro kinase assay but not in the full-length protein in the cells or by others.

To assess the functional relevance of MUS81 phosphorylation during replication stress, we performed neutral Comet assays in MUS81 KO MRC5 cells expressing wild–type MUS81, non–phosphorylable mutants (S95A, S97A), phosphomimetic mutants (S95D, S97D), or the nuclease–dead D307A mutant as further internal control (Fig. 5C). As expected, wild–type MUS81 restored HU–induced DSB formation, whereas both S95A and S97A mutants failed to do so, indicating that phosphorylation at these residues is required for MUS81 activation under replication stress. Conversely, the phosphomimetic mutants displayed elevated DSB levels even in untreated conditions, and S95D further increased DSB accumulation upon HU exposure, suggesting that phosphorylation at these sites is sufficient to prime MUS81 for nuclease activation.

To dissect the upstream kinase dependency of these events, we next evaluated DSB formation in the presence of a CHK2 inhibitor (CHK2i) (Fig. 5D). CHK2 inhibition reduced HU–induced DSBs in cells expressing wild–type MUS81 but had no effect on the unphosphorylable S95A or S97A mutants. Notably, CHK2 inhibition still decreased DSB levels in the single phosphomimetic mutantsmost prominently S95D—whereas the double phosphomimetic S95/97DD mutant, predicted to be refractory to CHK2 regulation, was completely insensitive to CHK2i and even showed a slight increase in DSBs (Fig. 5D). These findings, together with biochemical evidence that CHK2 phosphorylates S97 in vitro, MS data showing constitutive S95 phosphorylation in cells (consistent with a putative CDK2 site), and phosphoproteomic reports identifying both S95 and S97 in cancer cells, suggest that dual phosphorylation at these residues contributes to the fully licensing of MUS81-dependent DSBs under persistent replication stress.

We next focused on S97, which is the site targeted in vitro by CHK2, and evaluated whether the loss of phosphorylation or the presence of constitutive phosphorylation mimicked by the S97D mutation might affect the cleavage of a forked DNA substrate in vitro. To this aim, we co-expressed the MUS81–EME1 complex as a proxy of the physiological functionality of the MUS81 endonuclease also because co-expression of the MUS81 and EME2 resulted in an even less stochiometric complex because of the very different expression levels of the two subunits. As shown in Supplementary Fig. S11A and B, the immunopurified MUS81–EME1 complex is very active on a forked DNA used as substrate. We tested the cleavage of the DNA substrate using MUS81 complexes containing either wild-type MUS81, the S97A mutation, or the S97D mutation. We found that the inability to phosphorylate S97 (S97A) partially reduced endonuclease activity (from ~95% to <80% at 20 nM), while the S97D mutant showed a wild-type-like cleavage (Fig. 6A). Since the spontaneous level of S97 phosphorylation is expected to be very low as suggested by the MS data, this in vitro assay suggests that phosphorylation, as mimicked by the S97D mutation, only stimulates mildly the activity of the MUS81 endonuclease. Thus, we next evaluated whether phosphorylation of MUS81 at S95 or at S97 affected the recruitment of MUS81 to stalled forks using the SIRF assay. As shown in Fig. 6B, the wild-type MUS81 is recruited to perturbed forks in a time-dependent manner, with the higher number of SIRF spots observed at 24 h when DSBs are accumulating. Notably, either the S95A or the S97A, although markedly defective for the DSB formation, were equally or better recruited at perturbed forks if compared to the wild-type, whereas the phosphomimetic S97D showed a much higher level of SIRF spots at all time-points (Fig. 6B). The apparent improved ability of the DSB-defective S97A MUS81 mutant to be recruited to perturbed forks prompted us to evaluate the association with SLX4, which is known to stimulate nuclease function of MUS81 complex [48].

The in situ MUS81–SLX4 PLA was performed in MUS81 KO cells stably re-expressing wild-type MUS81 or the S97A mutant and transiently expressing HA–SLX4, either after treatment with 2 mM HU for 24 h or under untreated conditions. As shown in Fig. 6C, PLA spot counts in S-phase gated cells revealed a clear increase in MUS81–SLX4 interaction in wild-type cells upon HU-induced replication arrest. In contrast, HU failed to stimulate MUS81–SLX4 PLA spot formation in S-phase cells expressing the mutant MUS81 protein. Notably, although the HU-induced increase in MUS81–SLX4 association in S-phase was abolished by the non-phosphorylable S97A mutation, SLX4 recruitment to perturbed forks was independent of MUS81 S97 phosphorylation status (Supplementary Fig. S12). Altogether, these data indicate that phosphorylation of MUS81 at S95/S97 is required to promote HU-induced DSBs and suggest that MUS81–SLX4 complex formation in S-phase, a step that correlates with DSB induction during prolonged fork arrest, requires phosphorylation at S97. In contrast, phosphorylation at S97 is dispensable for MUS81 recruitment to stalled forks and has little effect on nuclease activity in vitro, indicating that phosphorylation primarily controls productive complex activation rather than fork targeting per se.

MUS81 phosphorylation promotes recovery of BRCA2-deficient cells upon replication stress

We next focused on the role of CHK2-dependent and independent phosphorylation of MUS81 on the formation of DSBs at deprotected forks. To assess whether phosphorylation influenced DSB formation at deprotected replication forks, we performed a neutral Comet assay following 2 h of HU in MUS81 KO MRC5 cells stably depleted of BRCA2 and re-expressing MUS81 wild-type or the different single phosphorylation mutants, S95A, S97A, S95D, S97D, or the double phosphomimetic S95/97DD (Fig. 7A). As expected, few DSBs were detectable at 2 h of HU in BRCA2-proficient cells but these DSBs were increased in the absence of BRCA2. Although the level of DSBs in BRCA2-proficient versu deficient cells was different, it was reduced in the absence of MUS81, confirming the genetic dependencies observed by others and our group [13, 28]. Expression of the S95A or S97A unphosphorylable MUS81 mutant reduced significantly DSBs irrespective of the presence of BRCA2, but much more clearly in cells depleted of BRCA2 where even 2 h of HU induced many DSBs (Fig. 7A). Expression of the phosphomimetic S95D MUS81 mutant was sufficient to induce a striking amount of DSBs irrespective of the BRCA2 status of the cells, and this phenotype was even more pronounced in cells expressing the S97D MUS81 (Fig. 7A). Of note, the level of DSBs detected at 2 h of HU in the cells expressing the phosphomimetic MUS81 mutants roughly matches that detected at 24 h of HU, indicating that the targeting of the stalled forks is very deregulated in these mutant cells.

Neutral Comet results suggest that phosphorylation of MUS81 at S95 and S97 is required for DSBs formation in BRCA2-depleted cells. Given that DSBs are necessary to facilitate the recovery of stalled replication forks, under these conditions, we assessed replication fork stalling using a DNA fibre assay after HU treatment. For this, we compared the number of stalled forks after recovery from 2 h of HU in shBRCA2 cells KO for MUS81 or complemented with the wild-type or the unphosphorylable S95A or S97A MUS81 (Fig. 7B). Consistent with previous reports [13], BRCA2-defective cells efficiently restarted transiently stalled replication forks, while the absence of MUS81 greatly impaired fork restart with persistence of about the 30% of stalled forks. Most remarkably, expression of the S95A MUS81 mutant resulted in a severe defect in fork restart since more than the 50% of perturbed forks remined stalled. The expression of the S97A MUS81 mutant affected fork restart with the number of stalled forks during recovery matching what seen in the absence of MUS81.

Importantly, CHK2 inhibition in BRCA2-depleted cells expressing the S97A MUS81 did not further increase fork stalling, supporting the conclusion that S97 is a critical CHK2 target site (Supplementary Fig. S13). These findings suggest that dual-site phosphorylation of MUS81 is critical for controlling MUS81-dependent DSB formation and facilitates fork restart in BRCA2-deficient cells.

MUS81 cleaves intact forks in wild-type cells and partially regressed forks in BRCA2-depleted cells, facilitating recovery from genotoxic stress or replication fork deprotection [13, 34]. To assess whether loss of the CHK2-dependent S97 phosphorylation was required for the survival of BRCA2-deficient cells following replication stress, we performed a survival assay after HU treatment. MUS81 KO cells or shBRCA2 MUS81 KO cells were transfected with the wild-type, the S97A, or S97D MUS81 mutant. These cells were treated with increasing concentrations of HU and analysed for cell survival. In the absence of BRCA2, loss of MUS81 reduced the cell survival in HU, as reported previously [13] (Fig. 7C). Interestingly, expression of S97A MUS81 sensitized cells to HU even in BRCA2-proficient cells, which were further sensitized to HU by concomitant depletion of BRCA2 (Fig. 7C). In contrast, the phosphomimetic S97D MUS81 mutant did not sensitize cells to HU, independently of the presence of BRCA2. In contrast, expression of the S97A mutant, but not of the S97D mutant of MUS81 sensitized BRCA2-depleted cells to HU (Supplementary Fig. S14A and B). Similar to the expression of the unphosphorylable S97A MUS81 mutant, the expression of the FHA-I157T CHK2 mutant increased HU sensitivity in BRCA2-depleted cells, even in the presence of endogenous MUS81, compared to cells expressing wild-type CHK2 (Supplementary Fig. S15).

Thus, genetic ablation of MUS81, disruption of its interaction with CHK2, or prevention of its phosphorylation is sufficient to induce comparable hypersensitivity to HU in BRCA2-depleted cells.

Discussion

In this work, we identify CHK2 as a positive regulator of the MUS81 endonuclease complex during prolonged or pathological replication fork stalling. We demonstrate that CHK2 forms a complex with MUS8–EME2 through its FHA domain and that subsequent phosphorylation at S97, in concert with phosphorylation at S95, a CDK2-consensus site, is required to license DSB formation at stalled replication forks. In BRCA2-deficient cells, the ultimate outcome of this dual-site and CHK2-dependent MUS81 complex regulation is replication fork recovery and increased viability under replication stress. Altogether, our findings define a previously unrecognized two-tier regulatory mechanism in which loading of MUS81 complex at perturbed forks is independent of phosphorylation, but dependent on fork remodelling, while the dual phosphorylation determines the targeting of the substrate by regulating the interaction with the SLX4 scaffold (Fig. 8).

Because the formation of DSBs at replication forks can produce gross chromosomal rearrangements, MUS81 endonuclease activity is tightly regulated in S-phase [33]. In yeast, the Mus81 complex is inhibited during normal replication or following transient replication arrest by the checkpoint kinase Cds1, while persistence of DNA damage or in the absence of the RecQ helicase Rqh1 can trigger Mus81 activation acting through a combination of checkpoint and cell-cycle kinases [22, 24, 25, 50, 51]. Of note, in yeast Mus81 the interaction with Cds1 is FHA-dependent but also phosphorylation-dependent [37]. Our study indicates that this interaction occurs through the FHA domain of CHK2 also in humans. Our findings also indicates that the FHA integrity is important for the interaction with MUS81 since the FHA-I157T CHK2 variant reduces the association with MUS81, although not completely, suggesting that MUS81 and CHK2 may also interact through other domains. At this stage, we do not know whether the interaction between MUS81 and CHK2 is phospho-dependent also in human cells. However, the absence of functional effects of the FHA-R117A CHK2 variant on MUS81-dependent outputs, such as DSBs or fork restart in BRCA2-deficient cells, combined with a wild-type level of coimmunoprecipitation, argue against a phospho-dependent binding. Indeed, the FHA-R117A mutation is known to abrogate phospho-dependent interactions such as that with BRCA1. Notably, FHA domains are known for their broad specificity and versatility. They can bind not only canonical phosphopeptides but also non-canonical and even non-phosphorylated sequences [52, 53]. Interestingly, this evolutionary conservation of the biochemical determinants of the MUS81–CHK2 interaction, combined with the divergent functional outcome, likely reflects the more frequent formation of intermediates requiring MUS81 complex resolution during replication of complex genomes, favouring a positive control, rather a negative one by the checkpoint, as well as the existence of two checkpoint kinases in human cells, CHK1 and CHK2, and two different regulatory subunits of MUS81, EME1, and EME2. One of the central findings of our study is that the CHK2-dependent regulation of MUS81 mainly involves the EME2-containing complex, which is consistent with EME2 being the relevant MUS81 partner in S-phase [19, 36, 54]. Several observations support this conclusion: first of all, combinatory binary PLA identifies proximal interaction between MUS81 and CHK2, EME2 and CHK2, and MUS81 and EME2 in S-phase cells upon replication stress; second, MUS81-dependent DSBs in response to replication stress are not affected by loss of EME1, but are reduced by EME2 silencing although the siRNA used was not very efficient; third, phosphorylation of EME1 by CDK1/PLK1 and of MUS81 by CK2 at serine 87 restricts the activation of the SLX4–MUS81–EME1 complex to late G2/M, preventing inappropriate cleavage events in S-phase [21, 55]. One limitation of our study is that most of conclusions are drawn from PLA experiments. Although PLA is not a formal proof of the existence of a stable complex, this technique can only read interactions within 40 nm, which is a distance likely to occur between direct or almost-direct partners, and the multiple PLA pairs suggest for the concurrence of the proximal interactions between the three proteins. More rigorous biochemical assay will be needed to assess the composition of the complex and the specific binding regions.

Strikingly, in BRCA2-deficient cells, the CHK2–MUS81 association is preserved, and it is increased by HU, yet the CHK2-EME2 PLA spots are not increased accordingly. This is paralleled by the lack of a consistent increase in the association between MUS81 and EME2 at 4 h of HU, when loss of BRCA2 triggers a substantial amount of MUS81-dependent DSBs that are not observed in the presence of BRCA2. The simplest explanation of this discordance with results obtained in BRCA2-proficient cells is that the MUS81 complex assembled at de-protected forks is not MUS8–EME2 but MUS81–EME1. Currently, limitations in the reliability of available anti-EME1 antibodies in PLA prevented a formal testing of this hypothesis. However, our Co-IP data show that CHK2 can co-immunoprecipitate EME1, although in a lesser amount than EME2, in cells treated with 24 h of HU. These findings indicate that assembly of both MUS81–CHK2 complexes may occur as a response to sustained replication fork stalling. However, in both cases, CHK2 may preferentially regulate MUS81 when it is engaged at stalled forks. Although the MUS8–EME2 complex was primarily implicated in the response to deprotected forks when BRCA2 is depleted [13], earlier studies from our group and others have also implicated the MUS81–EME1 complex in responses to CHK1 inhibition or oncogene-induced replication stress [27, 40, 41], suggesting that distinct MUS81-containing complexes may form in S-phase under pathological conditions. Interestingly, the number of MUS81 SIRF spots, representing its recruitment to stalled forks, is more abundant than the number of PLA spots representing interaction with CHK2. This is consistent with a transient, hit-and-run mode of action by which only a small, dynamic, subpopulation of fork-engaged MUS81 complexes is being targeted by CHK2 at any given time. Furthermore, our combined MUS81 SIRF assay, MUS8–EME2 or MUS81–CHK2 PLA together argue against formation of a stable complex.

While the C-terminal domain of MUS81 is primarily involved in the interaction with EME1 or 2 and contributes to the DNA binding surfaces of the complex, the N-terminal region is largely unstructured and contributes to the interaction with SLX4. Our biochemical data indicate that CHK2 can target the N-terminal region of MUS81 at two residues, T86 and S97. Of these two N-terminal CHK2 consensus residues, only S97 has been reported to be phosphorylated in the cell, together with the adjacent S95 [49]. Notably, we failed to identify pS97 in the full-length MUS81 protein expressed in the cell, however, our quantitative proteomics experiments reliably identified pS95, which was detectable even in untreated cells, mildly increased in cells treated with HU for 24 h but unaffected by the CHK2i. The inability to identify pS97 could be consistent with a small proportion of fork-engaged MUS81 complexes targeted by CHK2 in the cell, as suggested by our combined SIRF and PLA data, and with a regulatory modification. In contrast, phosphorylation of pS95 may represent a cell-cycle-related modification consistent with a classic S-P CDK consensus site, most likely CDK2 in S-phase in agreement with the observed increase in HU-treated samples and the in vitro kinase assay data. Nonetheless, both sites prove to be functionally relevant during prolonged fork stalling. Indeed, loss-of-function (S95, S97A) and gain-of-function (S95D, S97D, S95/97DD) MUS81 mutant expression in MUS81 KO cells reveal a dual-site mechanism and a possible hierarchy of the two phosphorylation events: either alanine substitution prevents HU-induced DSB formation, while each phosphomimetic substitution causes elevated levels of DSBs even in unstressed cells, especially for S97D. CHK2 inhibition reduces DSBs in cells expressing wild-type MUS81 but has no effect in S95A or S97A. Conversely, the single phosphomimetic mutants retain partial CHK2i sensitivity, most prominently the S95D, while the double phosphomimetic S95/97DD is fully refractory to CHK2 inhibition. These genetic findings, together with the biochemical observation that S97A only mildly affects nuclease activity in vitro, while S97D is as active as a wild-type, support a two-tier model of MUS81 activation in S-phase at stalled forks. We can speculate that the pS95 is CDK2-regulated event that can ‘prime’ the subsequent modification of S97 by CHK2 only when stalled forks need to be cleaved avoiding unwanted DNA breakage. Either alanine substitution collapses the activation module: S95A by removing the ‘priming’ mark and disrupting the subsequent CHK2-dependent event, S97A by directly removing the activating site. This scenario is consistent with the observation that S95A phenocopies CHK2i despite S95 itself being CHK2i-independent and is consistent with pS95 being more abundant and pS97 transient and/or much more limited and difficult to detect by quantitative MS. Interestingly, this priming logic is reminiscent of the well-characterized CDK-primed events, including those by PLK1, and of the regulation reported in the yeast Mus81/Mms4 complex [33, 56].

Mechanistically, S95 and S97 lie within the N-terminal region of MUS81 that mediates interaction with SLX4. We find that loss of phosphorylation does not affect MUS81 complex recruitment to perturbed forks, however S97A mutation abrogates the HU-dependent stimulation of the association of MUS81 with SLX4 in S-phase. SLX4 is known to alleviate the auto-inhibitory interaction between the hairpin–helix–hairpin domain of MUS81 and EME1 [48], and given the structural similarities [57], likely also between MUS81 and EME2. Thus, CHK2-mediated phosphorylation of a pS95 ‘primed’ MUS81 could induce a conformational change in the N-terminal region promoting association with SLX4, which is independently recruited to stalled forks. The formation of the MUS81/SLX4 complex would release the autoinhibitory locking and license nuclease activity. The constitutive DSB-promoting function of S95/97DD combined with its CHK2i resistance may be consistent with a phospho-locked open configuration. The proposed regulatory role of S97 phosphorylation with a specific requirement for MUS81–SLX4 association also explains the observed wild-type activity of S97A-MUS81 or S97D-MUS81 in vitro where SLX4 is absent.

The MUS81 complex is essential for resolving deprotected forks induced by the loss of BRCA2 [13], but its activity must be restricted to forks that have undergone degradation and resection to prevent inappropriate cleavage of intact replication structures and consequent genome rearrangements. Our findings indicate that CHK2-dependent regulation carries out this role precisely. The MUS81–CHK2 interaction is greatly reduced when fork reversal is impaired, and when MRE11 is prevented from binding to or degrading the reversed fork, indicating that CHK2 is recruited only after substrate formation. This is consistent with the reported observation that, upon loss of BRCA2, MUS81 functions downstream of the CtIP, MRE11, and EXO1 nucleases to cleave flapped DNA structures at reversed replication forks [13]. Moreover, phosphorylated ATM can be detected at stalled forks, independent of MUS81, suggesting that ATM–CHK2 signalling is activated prior to DSB formation in BRCA2-deficient cells. This points to an upstream regulatory role for ATM–CHK2 in controlling MUS8–EME2 function.

Notably, active ATM, phosphorylated at S1981, can also be detected at stalled replication forks in BRCA2-proficient cells, despite the absence of detectable DSBs. This is consistent with the role of the DNA end-binding factors, such as KU, which may interact with reversed forks [58].

In the absence of MUS81-dependent cleavage, extensively resected forks fail to properly restart [13]. Consistent with this, disruption of CHK2-dependent regulation of MUS81, through the S95A, S97A, or FHA I157T mutation, impairs cleavage of stalled forks, leading to their accumulation and severely compromising replication fork restart and cellular recovery from replication stress.

Mutations in ATM or CHK2 are known to predispose individuals to breast cancer, and the ATM–CHK2 pathway has been shown to support MUS81-dependent DSB formation in response to cisplatin treatment in human breast cancer models [59]. Although, the CHK2-I157T missense mutation variant, initially associated to Li–Fraumeni syndrome [38, 60], is now classified as a SNP, it remains linked to increased breast cancer risk (odds ratio = 1.4), particularly in carriers who also harbour pathogenic mutations of CHK2 or BRCA1/2 [31]. Our data demonstrate that this variant selectively disrupts CHK2 interaction with MUS81 and impairs replication fork recovery, aligning with previous reports of compromised binding to other checkpoint proteins, such as BRCA1. Beyond its clinical reclassification, I157T thus represents a clean separation-of-function allele to dissect the role of CHK2 in regulating endonuclease activity under replication stress, and may behave as a low-penetrance modifier of PARPi response in BRCA2-mutant tumours.

In conclusion, this study reveals that activation of MUS8–EME2 or also MUS81–EME1 at persistently-arrested replication forks in wild-type cells or in deprotected forks in BRCA2-deficient cells, is governed by an integrated two-tier mechanism: a phospho-independent MUS81 docking step, controlled by fork remodelling and resection at least in BRCA2-deficient cells, and a coordinated CDK/CHK2 dual-site phosphorylation switch on the MUS81 N-terminus that licenses DSB formation mediated by the association with SLX4. Because activation of MUS81 at deprotected forks has been linked to PARP inhibitor sensitivity [15, 16], and because inactivating CHK2 mutations recur in breast cancer [61, 62], our findings may have direct translational implications for PARPi response biomarkers and for the rational design of DDR-targeting combinations. They argue that tumours harbouring CHK2-I157T or other CHK2 hypomorph mutations, particularly when combined with BRCA1/2 alterations, may exhibit altered PARPi response curves that warrant prospective clinical evaluation.

Supplementary Material

gkag766_Supplemental_File

gkag766_supplemental_file.pdf^{(2.3MB, pdf)}

Acknowledgements

We are grateful to Drs Domenico Delia and Giacomo Buscemi for the gift of the pcDNA3-HA-Chk2 plasmid. We are grateful to Prof. Massimo Lopes for scientific discussion. The authors gratefully acknowledge Dr. Andreas Jakob (Bruker Daltonics, Bremen, Germany) for his expert technical assistance with proteomics data analysis. Graphical abstract Created in BioRender. Pichierri, P. (2026) https://BioRender.com/2a5ae29.

Author contributions: A.P. performed all the biochemical experiments and initial functional characterization of the MUS81–CHK2 complex formation. E.M. and C.F. performed the analysis of fork recruitment in vivo, analysis of ATM–CHK2 activation, analysis of RuvA chromatin recruitment, DNA fibre assays, and most of the functional analysis of the MUS81 phosphorylation mutants. M.S. assisted with the purification of protein fragment in bacteria and with Co-IP. P.V. contributed to the analysis of DSBs formation in CHK2i-treated cells. S.R. performed the analysis of MUS81 phosphorylation in vitro and MS to identify phosphorylated residues. S.C., M.C., and F.P. performed the quantitative proteomics to detect MUS81 phosphorylation in the cells. M.H. performed in vitro nuclease assay. A.P., E.M., C.F., S.C., M.H., and S.R. analysed data and contributed to designing the experiments and writing the paper. A.F., M.H., M.S., and P.P. designed experiments, analysed data, supervised research, and wrote the paper. All authors approved the paper. E.M. and C.F. contributed equally to this work and should be regarded as joint first authors.

Contributor Information

Eva Malacaria, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Carolina Figlioli, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Anita Palma, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Masayoshi Honda, Department of Biochemistry and Molecular Biology, University of Iowa Carver College of Medicine, 51 Newton Road, Iowa City, IA 52242, United States.

Pasquale Valenzisi, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Marialuisa Casella, Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Serena Camerini, Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Fabio Pucci, Core Facilities Service, Proteomics Unit, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy; Istituto Nazionale Biofisica e Biostrutture (INBB), Viale delle Medaglie d’Oro, 305, Rome 00136, Italy.

Sara Rinalducci, Proteomics lab, Department of Ecology and Biology, Università della Tuscia, Largo dell’Università, Viterbo 01100, Italy.

Maurizio Semproni, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Maria Spies, Department of Biochemistry and Molecular Biology, University of Iowa Carver College of Medicine, 51 Newton Road, Iowa City, IA 52242, United States.

Annapaola Franchitto, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy.

Pietro Pichierri, Genome Stability Group, Section of Advanced Studies on DNA Metabolisms and Environmental Stress, Department of Environment and Health, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome 00161, Italy; Istituto Nazionale Biofisica e Biostrutture (INBB), Viale delle Medaglie d’Oro, 305, Rome 00136, Italy.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was supported by the Italian Ministry of Health ‘Ricerca Finalizzata’ to P.P. (Grant No. RF-2016-02362022), to the Italian Ministry of Health ‘Ricerca Finalizzata’ (PNRR-MAD-2022-12376183), the NIH 2R01CA232425 to M.S., and by the Associazione Italiana per la Ricerca sul Cancro to P.P. (IG No. 21428) and to A.F. (IG No. 19971). Funding to pay the Open Access publication charges for this article was provided by the AIRC 21428.

Data availability

The datasets generated and analyzed during the current study are available at request through submission on public repositories. The proteomics data are available in the ProteomeXchange via the MassIVE repository with the dataset identifier MassIVE MSV000101714 and with identifier PXD064334.

References

1. Osborn AJ, Elledge SJ, Zou L. Checking on the fork: the DNA-replication stress-response pathway. Trends Cell Biol. 2002;12:509–16. 10.1016/S0962-8924(02)02380-2. [DOI] [PubMed] [Google Scholar]
2. Hills SA, Diffley JFX. DNA replication and oncogene-induced replicative stress. Curr Biol. 2014;24:R435–44. 10.1016/j.cub.2014.04.012. [DOI] [PubMed] [Google Scholar]
3. Berti M, Vindigni A. Replication stress: getting back on track. Nat Struct Mol Biol. 2016;23:103–9. 10.1038/nsmb.3163. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Branzei D, Foiani M. The checkpoint response to replication stress. DNA Repair. 2009;8:1038–46. 10.1016/j.dnarep.2009.04.014. [DOI] [PubMed] [Google Scholar]
5. Malacaria E, Honda M, Franchitto A et al. Physiological and pathological roles of RAD52 at DNA replication forks. Cancers. 2020;12:402. 10.3390/cancers12020402. [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Cimprich KA, Cortez D. ATR: an essential regulator of genome integrity. Nat Rev Mol Cell Biol. 2008;9:616–27. 10.1038/nrm2450. [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Quinet A, Lemaçon D, Vindigni A. Replication fork reversal: players and guardians. Mol Cell. 2017;68:830–3. 10.1016/j.molcel.2017.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Neelsen KJ, Lopes M. Replication fork reversal in eukaryotes: from dead end to dynamic response. Nat Rev Mol Cell Biol. 2015;16:207–20. 10.1038/nrm3935. [DOI] [PubMed] [Google Scholar]
9. San Filippo J, Sung P, Klein H. Mechanism of eukaryotic homologous recombination. Annu Rev Biochem. 2008;77:229–57. [DOI] [PubMed] [Google Scholar]
10. Lord CJ, Ashworth A. BRCAness revisited. Nat Rev Cancer. 2016;16:110–20. 10.1038/nrc.2015.21. [DOI] [PubMed] [Google Scholar]
11. Tutt A, Ashworth A. The relationship between the roles of BRCA genes in DNA repair and cancer predisposition. Trends Mol Med. 2002;8:571–6. 10.1016/S1471-4914(02)02434-6. [DOI] [PubMed] [Google Scholar]
12. Schlacher K, Christ N, Siaud N et al. Double-strand break repair-independent role for BRCA2 in blocking stalled replication fork degradation by MRE11. Cell. 2011;145:529–42. 10.1016/j.cell.2011.03.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Lemaçon D, Jackson J, Quinet A et al. MRE11 and EXO1 nucleases degrade reversed forks and elicit MUS81-dependent fork rescue in BRCA2-deficient cells. Nat Commun. 2017;8:860. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Mijic S, Zellweger R, Chappidi N et al. Replication fork reversal triggers fork degradation in BRCA2-defective cells. Nat Commun. 2017;8:859. 10.1038/s41467-017-01164-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Chaudhuri AR, Callen E, Ding X et al. Replication fork stability confers chemoresistance in BRCA-deficient cells. Nature. 2016;535:382–7. 10.1038/nature18325. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Rondinelli B, Gogola E, Yücel H et al. EZH2 promotes degradation of stalled replication forks by recruiting MUS81 through histone H3 trimethylation. Nat Cell Biol. 2017;19:1371–8. 10.1038/ncb3626. [DOI] [PubMed] [Google Scholar]
17. Dungrawala H, Bhat KP, Le Meur R et al. RADX promotes genome stability and modulates chemosensitivity by regulating RAD51 at replication forks. Mol Cell. 2017;67:374–86. 10.1016/j.molcel.2017.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Lai X, Broderick R, Bergoglio V et al. MUS81 nuclease activity is essential for replication stress tolerance and chromosome segregation in BRCA2-deficient cells. Nat Commun. 2017;8:15983. 10.1038/ncomms15983. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Pepe A, West SC. MUS8–EME2 promotes replication fork restart. Cell Rep. 2014;7:1048–55. 10.1016/j.celrep.2014.04.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Wild P, Matos J. Cell cycle control of DNA joint molecule resolution. Curr Opin Cell Biol. 2016;40:74–80. 10.1016/j.ceb.2016.02.018. [DOI] [PubMed] [Google Scholar]
21. Palma A, Pugliese GM, Murfuni I et al. Phosphorylation by CK2 regulates MUS81/EME1 in mitosis and after replication stress. Nucleic Acids Res. 2018;46:5109–24. 10.1093/nar/gky280. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Giaccherini C, Scaglione S, Coulon S et al. Regulation of MUS81–EME1 structure-specific endonuclease by Eme1 SUMO-binding and Rad3ATR kinase is essential in the absence of Rqh1BLM helicase. PLoS Genet. 2022;18:e1010165. 10.1371/journal.pgen.1010165. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Boddy MN, Lopez-Girona a, Shanahan P et al. Damage tolerance protein Mus81 associates with the FHA1 domain of checkpoint kinase Cds1. Mol Cell Biol. 2000;20:8758–66. 10.1128/MCB.20.23.8758-8766.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Kai M, Boddy MN, Russell P et al. Replication checkpoint kinase Cds1 regulates Mus81 to preserve genome integrity during replication stress. Genes Dev. 2005;19:919–32. 10.1101/gad.1304305. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Dehé P-M, Coulon S, Scaglione S et al. Regulation of MUS81–EME1 Holliday junction resolvase in response to DNA damage. Nat Struct Mol Biol. 2013;20:598–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Ciccia A, Elledge SJ. The DNA damage response: making it safe to play with knives. Mol Cell. 2010;40:179–204. 10.1016/j.molcel.2010.09.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Murfuni I, Basile G, Subramanyam S et al. Survival of the replication checkpoint deficient cells requires MUS81-RAD52 function. PLoS Genet. 2013;9:e1003910. 10.1371/journal.pgen.1003910. [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Blandino F, Malacaria E, Figlioli C et al. Phosphorylation status of MUS81 is a modifier of olaparib sensitivity in BRCA2-deficient cells. Nucleic Acids Res. 2023;51:6723–37. 10.1093/nar/gkad470. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Thingholm TE, Jørgensen TJD, Jensen ON et al. Highly selective enrichment of phosphorylated peptides using titanium dioxide. Nat Protoc. 2006;1:1929–35. 10.1038/nprot.2006.185. [DOI] [PubMed] [Google Scholar]
30. Wang Y-T, Tsai C-F, Hong T-C et al. An informatics-assisted label-free quantitation strategy that depicts phosphoproteomic profiles in lung cancer cell invasion. J Proteome Res. 2010;9:5582–97. 10.1021/pr100394u. [DOI] [PubMed] [Google Scholar]
31. Apostolou P, Papasotiriou I. Current perspectives on CHEK2 mutations in breast cancer. Breast cancer. 2017;9:331–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Honda M, Razzaghi M, Gaur P et al. The RAD52 double-ring remodels replication forks restricting fork reversal. Nature. 2025;641:512–9. 10.1038/s41586-025-08753-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Matos J, West SC. Holliday junction resolution: regulation in space and time. DNA Repair. 2014;19:176–81. 10.1016/j.dnarep.2014.03.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Hanada K, Budzowska M, Davies SL et al. The structure-specific endonuclease Mus81 contributes to replication restart by generating double-strand DNA breaks. Nat Struct Mol Biol. 2007;14:1096–104. 10.1038/nsmb1313. [DOI] [PubMed] [Google Scholar]
35. Franchitto A, Pirzio LM, Prosperi E et al. Replication fork stalling in WRN-deficient cells is overcome by prompt activation of a MUS81-dependent pathway. J Cell Biol. 2008;183:241–52. 10.1083/jcb.200803173. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Técher H, Koundrioukoff S, Carignon S et al. Signaling from MUS8–EME2-dependent DNA damage elicited by Chk1 deficiency modulates replication fork speed and origin usage. Cell Rep. 2016;14:1114–27. 10.1016/j.celrep.2015.12.093. [DOI] [PubMed] [Google Scholar]
37. Boddy MN, Lopez-Girona a, Shanahan P et al. Damage tolerance protein Mus81 associates with the FHA1 domain of checkpoint kinase Cds1. Mol Cell Biol. 2000;20:8758–66. 10.1128/MCB.20.23.8758-8766.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Bell DW, Varley JM, Szydlo TE et al. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science. 1999;286:2528–31. 10.1126/science.286.5449.2528. [DOI] [PubMed] [Google Scholar]
39. Li J, Williams BL, Haire LF et al. Structural and functional versatility of the FHA domain in DNA-damage signaling by the tumor suppressor kinase Chk2. Mol Cell. 2002;9:1045–54. 10.1016/S1097-2765(02)00527-0. [DOI] [PubMed] [Google Scholar]
40. Murfuni I, Nicolai S, Baldari S et al. The WRN and MUS81 proteins limit cell death and genome instability following oncogene activation. Oncogene. 2012;32:610–20. 10.1038/onc.2012.80. [DOI] [PubMed] [Google Scholar]
41. Forment JV, Blasius M, Guerini I et al. Structure-specific DNA endonuclease mus81/eme1 generates DNA damage caused by chk1 inactivation. PLoS One. 2011;6:e23517. 10.1371/journal.pone.0023517. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Rickman K, Smogorzewska A. Advances in understanding DNA processing and protection at stalled replication forks. J Cell Biol. 2019;218:1096–107. 10.1083/jcb.201809012. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Berti M, Ray Chaudhuri A, Thangavel S et al. Human RECQ1 promotes restart of replication forks reversed by DNA topoisomerase I inhibition. Nat Struct Mol Biol. 2013;20:347–54. 10.1038/nsmb.2501. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Hengel SR, Malacaria E, Folly da Silva Constantino L et al. Small-molecule inhibitors identify the RAD52-ssDNA interaction as critical for recovery from replication stress and for survival of BRCA2 deficient cells. eLife. 2016;5:e14740. 10.7554/eLife.14740. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Malacaria E, Pugliese GM, Honda M et al. Rad52 prevents excessive replication fork reversal and protects from nascent strand degradation. Nat Commun. 2019;10:1–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Durocher D, Taylor I, Sarbassova D et al. The molecular basis of FHA domain:phosphopeptide binding specificity and implications for phospho-dependent signaling mechanisms. Mol Cell. 2000;6:1169–82. 10.1016/S1097-2765(00)00114-3. [DOI] [PubMed] [Google Scholar]
47. Nair N, Castor D, Macartney T et al. Identification and characterization of MUS81 point mutations that abolish interaction with the SLX4 scaffold protein. DNA Repair. 2014;24:131–7. 10.1016/j.dnarep.2014.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Wyatt HDM, Laister RC, Martin SR et al. The SMX DNA repair tri-nuclease. Mol Cell. 2017;65:848–60. 10.1016/j.molcel.2017.01.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Zhou H, Di Palma S, Preisinger C et al. Toward a comprehensive characterization of a human cancer cell phosphoproteome. J Proteome Res. 2013;12:260–71. 10.1021/pr300630k. [DOI] [PubMed] [Google Scholar]
50. Lambert S, Mizuno K, Blaisonneau J et al. Homologous recombination restarts blocked replication forks at the expense of genome rearrangements by template exchange. Mol Cell. 2010;39:346–59. 10.1016/j.molcel.2010.07.015. [DOI] [PubMed] [Google Scholar]
51. Doe CL, Ahn JS, Dixon J et al. MUS81–EME1 and Rqh1 involvement in processing stalled and collapsed replication forks. J Biol Chem. 2002;277:32753–9. 10.1074/jbc.M202120200. [DOI] [PubMed] [Google Scholar]
52. Almawi AW, Matthews LA, Guarné A. FHA domains: phosphopeptide binding and beyond. Prog Biophys Mol Biol. 2017;127:105–10. 10.1016/j.pbiomolbio.2016.12.003. [DOI] [PubMed] [Google Scholar]
53. Matthews LA, Selvaratnam R, Jones DR et al. A novel non-canonical forkhead-associated (FHA) domain-binding interface mediates the interaction between Rad53 and Dbf4 proteins. J Biol Chem. 2014;289:2589–99. 10.1074/jbc.M113.517060. [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Pepe A, West SC. Substrate specificity of the MUS8–EME2 structure selective endonuclease. Nucleic Acids Res. 2014;42:3833–45. 10.1093/nar/gkt1333. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Duda H, Arter M, Gloggnitzer J et al. A mechanism for controlled breakage of under-replicated chromosomes during mitosis. Dev Cell. 2016;39:740–55. 10.1016/j.devcel.2016.11.017. [DOI] [PubMed] [Google Scholar]
56. Gallo-Fernández M, Saugar I, Ortiz-Bazán MÁ et al. Cell cycle-dependent regulation of the nuclease activity of MUS81–EME1/Mms4. Nucleic Acids Res. 2012;40:8325–35. 10.1093/nar/gks599. [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Hua Z, Fang Q, Zhang D et al. Crystal structure of the human MUS8–EME2 complex. Structure. 2022;30:743–52. 10.1016/j.str.2022.02.015. [DOI] [PubMed] [Google Scholar]
58. Dhoonmoon A, Nicolae CM, Moldovan G-L. The KU-PARP14 axis differentially regulates DNA resection at stalled replication forks by MRE11 and EXO1. Nat Commun. 2022;13:5063. 10.1038/s41467-022-32756-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Lv J, Qian Y, Ni X et al. Feedback regulation of methyl methanesulfonate and ultraviolet-sensitive gene clone 81 via ATM/Chk2 pathway contributes to the resistance of MCF-7 breast cancer cells to cisplatin. Tumour biol. 2017;39:1010428317694307. [DOI] [PubMed] [Google Scholar]
60. Sancar A, Lindsey-Boltz L, Unsal-Kaçmaz K et al. Molecular mechanisms of mammalian DNA repair and the DNA damage checkpoints. Annu Rev Biochem. 2004;73:39–85. 10.1146/annurev.biochem.73.011303.073723. [DOI] [PubMed] [Google Scholar]
61. Stolarova L, Kleiblova P, Janatova M et al. CHEK2 germline variants in cancer predisposition: stalemate rather than checkmate. Cells. 2020;9:2675. 10.3390/cells9122675. [DOI] [PMC free article] [PubMed] [Google Scholar]
62. Boonen RACM, Wiegant WW, Celosse N et al. Functional analysis identifies damaging CHEK2 missense variants associated with increased cancer risk. Cancer Res. 2022;82:615–31. 10.1158/0008-5472.CAN-21-1845. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gkag766_Supplemental_File

gkag766_supplemental_file.pdf^{(2.3MB, pdf)}

Data Availability Statement

[B1] 1. Osborn AJ, Elledge SJ, Zou L. Checking on the fork: the DNA-replication stress-response pathway. Trends Cell Biol. 2002;12:509–16. 10.1016/S0962-8924(02)02380-2. [DOI] [PubMed] [Google Scholar]

[B2] 2. Hills SA, Diffley JFX. DNA replication and oncogene-induced replicative stress. Curr Biol. 2014;24:R435–44. 10.1016/j.cub.2014.04.012. [DOI] [PubMed] [Google Scholar]

[B3] 3. Berti M, Vindigni A. Replication stress: getting back on track. Nat Struct Mol Biol. 2016;23:103–9. 10.1038/nsmb.3163. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B4] 4. Branzei D, Foiani M. The checkpoint response to replication stress. DNA Repair. 2009;8:1038–46. 10.1016/j.dnarep.2009.04.014. [DOI] [PubMed] [Google Scholar]

[B5] 5. Malacaria E, Honda M, Franchitto A et al. Physiological and pathological roles of RAD52 at DNA replication forks. Cancers. 2020;12:402. 10.3390/cancers12020402. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B6] 6. Cimprich KA, Cortez D. ATR: an essential regulator of genome integrity. Nat Rev Mol Cell Biol. 2008;9:616–27. 10.1038/nrm2450. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B7] 7. Quinet A, Lemaçon D, Vindigni A. Replication fork reversal: players and guardians. Mol Cell. 2017;68:830–3. 10.1016/j.molcel.2017.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B8] 8. Neelsen KJ, Lopes M. Replication fork reversal in eukaryotes: from dead end to dynamic response. Nat Rev Mol Cell Biol. 2015;16:207–20. 10.1038/nrm3935. [DOI] [PubMed] [Google Scholar]

[B9] 9. San Filippo J, Sung P, Klein H. Mechanism of eukaryotic homologous recombination. Annu Rev Biochem. 2008;77:229–57. [DOI] [PubMed] [Google Scholar]

[B10] 10. Lord CJ, Ashworth A. BRCAness revisited. Nat Rev Cancer. 2016;16:110–20. 10.1038/nrc.2015.21. [DOI] [PubMed] [Google Scholar]

[B11] 11. Tutt A, Ashworth A. The relationship between the roles of BRCA genes in DNA repair and cancer predisposition. Trends Mol Med. 2002;8:571–6. 10.1016/S1471-4914(02)02434-6. [DOI] [PubMed] [Google Scholar]

[B12] 12. Schlacher K, Christ N, Siaud N et al. Double-strand break repair-independent role for BRCA2 in blocking stalled replication fork degradation by MRE11. Cell. 2011;145:529–42. 10.1016/j.cell.2011.03.041. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B13] 13. Lemaçon D, Jackson J, Quinet A et al. MRE11 and EXO1 nucleases degrade reversed forks and elicit MUS81-dependent fork rescue in BRCA2-deficient cells. Nat Commun. 2017;8:860. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B14] 14. Mijic S, Zellweger R, Chappidi N et al. Replication fork reversal triggers fork degradation in BRCA2-defective cells. Nat Commun. 2017;8:859. 10.1038/s41467-017-01164-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B15] 15. Chaudhuri AR, Callen E, Ding X et al. Replication fork stability confers chemoresistance in BRCA-deficient cells. Nature. 2016;535:382–7. 10.1038/nature18325. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B16] 16. Rondinelli B, Gogola E, Yücel H et al. EZH2 promotes degradation of stalled replication forks by recruiting MUS81 through histone H3 trimethylation. Nat Cell Biol. 2017;19:1371–8. 10.1038/ncb3626. [DOI] [PubMed] [Google Scholar]

[B17] 17. Dungrawala H, Bhat KP, Le Meur R et al. RADX promotes genome stability and modulates chemosensitivity by regulating RAD51 at replication forks. Mol Cell. 2017;67:374–86. 10.1016/j.molcel.2017.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B18] 18. Lai X, Broderick R, Bergoglio V et al. MUS81 nuclease activity is essential for replication stress tolerance and chromosome segregation in BRCA2-deficient cells. Nat Commun. 2017;8:15983. 10.1038/ncomms15983. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B19] 19. Pepe A, West SC. MUS8–EME2 promotes replication fork restart. Cell Rep. 2014;7:1048–55. 10.1016/j.celrep.2014.04.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B20] 20. Wild P, Matos J. Cell cycle control of DNA joint molecule resolution. Curr Opin Cell Biol. 2016;40:74–80. 10.1016/j.ceb.2016.02.018. [DOI] [PubMed] [Google Scholar]

[B21] 21. Palma A, Pugliese GM, Murfuni I et al. Phosphorylation by CK2 regulates MUS81/EME1 in mitosis and after replication stress. Nucleic Acids Res. 2018;46:5109–24. 10.1093/nar/gky280. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B22] 22. Giaccherini C, Scaglione S, Coulon S et al. Regulation of MUS81–EME1 structure-specific endonuclease by Eme1 SUMO-binding and Rad3ATR kinase is essential in the absence of Rqh1BLM helicase. PLoS Genet. 2022;18:e1010165. 10.1371/journal.pgen.1010165. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B23] 23. Boddy MN, Lopez-Girona a, Shanahan P et al. Damage tolerance protein Mus81 associates with the FHA1 domain of checkpoint kinase Cds1. Mol Cell Biol. 2000;20:8758–66. 10.1128/MCB.20.23.8758-8766.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B24] 24. Kai M, Boddy MN, Russell P et al. Replication checkpoint kinase Cds1 regulates Mus81 to preserve genome integrity during replication stress. Genes Dev. 2005;19:919–32. 10.1101/gad.1304305. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B25] 25. Dehé P-M, Coulon S, Scaglione S et al. Regulation of MUS81–EME1 Holliday junction resolvase in response to DNA damage. Nat Struct Mol Biol. 2013;20:598–603. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B26] 26. Ciccia A, Elledge SJ. The DNA damage response: making it safe to play with knives. Mol Cell. 2010;40:179–204. 10.1016/j.molcel.2010.09.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B27] 27. Murfuni I, Basile G, Subramanyam S et al. Survival of the replication checkpoint deficient cells requires MUS81-RAD52 function. PLoS Genet. 2013;9:e1003910. 10.1371/journal.pgen.1003910. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B28] 28. Blandino F, Malacaria E, Figlioli C et al. Phosphorylation status of MUS81 is a modifier of olaparib sensitivity in BRCA2-deficient cells. Nucleic Acids Res. 2023;51:6723–37. 10.1093/nar/gkad470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B29] 29. Thingholm TE, Jørgensen TJD, Jensen ON et al. Highly selective enrichment of phosphorylated peptides using titanium dioxide. Nat Protoc. 2006;1:1929–35. 10.1038/nprot.2006.185. [DOI] [PubMed] [Google Scholar]

[B30] 30. Wang Y-T, Tsai C-F, Hong T-C et al. An informatics-assisted label-free quantitation strategy that depicts phosphoproteomic profiles in lung cancer cell invasion. J Proteome Res. 2010;9:5582–97. 10.1021/pr100394u. [DOI] [PubMed] [Google Scholar]

[B31] 31. Apostolou P, Papasotiriou I. Current perspectives on CHEK2 mutations in breast cancer. Breast cancer. 2017;9:331–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B32] 32. Honda M, Razzaghi M, Gaur P et al. The RAD52 double-ring remodels replication forks restricting fork reversal. Nature. 2025;641:512–9. 10.1038/s41586-025-08753-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B33] 33. Matos J, West SC. Holliday junction resolution: regulation in space and time. DNA Repair. 2014;19:176–81. 10.1016/j.dnarep.2014.03.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B34] 34. Hanada K, Budzowska M, Davies SL et al. The structure-specific endonuclease Mus81 contributes to replication restart by generating double-strand DNA breaks. Nat Struct Mol Biol. 2007;14:1096–104. 10.1038/nsmb1313. [DOI] [PubMed] [Google Scholar]

[B35] 35. Franchitto A, Pirzio LM, Prosperi E et al. Replication fork stalling in WRN-deficient cells is overcome by prompt activation of a MUS81-dependent pathway. J Cell Biol. 2008;183:241–52. 10.1083/jcb.200803173. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B36] 36. Técher H, Koundrioukoff S, Carignon S et al. Signaling from MUS8–EME2-dependent DNA damage elicited by Chk1 deficiency modulates replication fork speed and origin usage. Cell Rep. 2016;14:1114–27. 10.1016/j.celrep.2015.12.093. [DOI] [PubMed] [Google Scholar]

[B37] 37. Boddy MN, Lopez-Girona a, Shanahan P et al. Damage tolerance protein Mus81 associates with the FHA1 domain of checkpoint kinase Cds1. Mol Cell Biol. 2000;20:8758–66. 10.1128/MCB.20.23.8758-8766.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B38] 38. Bell DW, Varley JM, Szydlo TE et al. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science. 1999;286:2528–31. 10.1126/science.286.5449.2528. [DOI] [PubMed] [Google Scholar]

[B39] 39. Li J, Williams BL, Haire LF et al. Structural and functional versatility of the FHA domain in DNA-damage signaling by the tumor suppressor kinase Chk2. Mol Cell. 2002;9:1045–54. 10.1016/S1097-2765(02)00527-0. [DOI] [PubMed] [Google Scholar]

[B40] 40. Murfuni I, Nicolai S, Baldari S et al. The WRN and MUS81 proteins limit cell death and genome instability following oncogene activation. Oncogene. 2012;32:610–20. 10.1038/onc.2012.80. [DOI] [PubMed] [Google Scholar]

[B41] 41. Forment JV, Blasius M, Guerini I et al. Structure-specific DNA endonuclease mus81/eme1 generates DNA damage caused by chk1 inactivation. PLoS One. 2011;6:e23517. 10.1371/journal.pone.0023517. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B42] 42. Rickman K, Smogorzewska A. Advances in understanding DNA processing and protection at stalled replication forks. J Cell Biol. 2019;218:1096–107. 10.1083/jcb.201809012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B43] 43. Berti M, Ray Chaudhuri A, Thangavel S et al. Human RECQ1 promotes restart of replication forks reversed by DNA topoisomerase I inhibition. Nat Struct Mol Biol. 2013;20:347–54. 10.1038/nsmb.2501. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B44] 44. Hengel SR, Malacaria E, Folly da Silva Constantino L et al. Small-molecule inhibitors identify the RAD52-ssDNA interaction as critical for recovery from replication stress and for survival of BRCA2 deficient cells. eLife. 2016;5:e14740. 10.7554/eLife.14740. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B45] 45. Malacaria E, Pugliese GM, Honda M et al. Rad52 prevents excessive replication fork reversal and protects from nascent strand degradation. Nat Commun. 2019;10:1–19. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B46] 46. Durocher D, Taylor I, Sarbassova D et al. The molecular basis of FHA domain:phosphopeptide binding specificity and implications for phospho-dependent signaling mechanisms. Mol Cell. 2000;6:1169–82. 10.1016/S1097-2765(00)00114-3. [DOI] [PubMed] [Google Scholar]

[B47] 47. Nair N, Castor D, Macartney T et al. Identification and characterization of MUS81 point mutations that abolish interaction with the SLX4 scaffold protein. DNA Repair. 2014;24:131–7. 10.1016/j.dnarep.2014.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B48] 48. Wyatt HDM, Laister RC, Martin SR et al. The SMX DNA repair tri-nuclease. Mol Cell. 2017;65:848–60. 10.1016/j.molcel.2017.01.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B49] 49. Zhou H, Di Palma S, Preisinger C et al. Toward a comprehensive characterization of a human cancer cell phosphoproteome. J Proteome Res. 2013;12:260–71. 10.1021/pr300630k. [DOI] [PubMed] [Google Scholar]

[B50] 50. Lambert S, Mizuno K, Blaisonneau J et al. Homologous recombination restarts blocked replication forks at the expense of genome rearrangements by template exchange. Mol Cell. 2010;39:346–59. 10.1016/j.molcel.2010.07.015. [DOI] [PubMed] [Google Scholar]

[B51] 51. Doe CL, Ahn JS, Dixon J et al. MUS81–EME1 and Rqh1 involvement in processing stalled and collapsed replication forks. J Biol Chem. 2002;277:32753–9. 10.1074/jbc.M202120200. [DOI] [PubMed] [Google Scholar]

[B52] 52. Almawi AW, Matthews LA, Guarné A. FHA domains: phosphopeptide binding and beyond. Prog Biophys Mol Biol. 2017;127:105–10. 10.1016/j.pbiomolbio.2016.12.003. [DOI] [PubMed] [Google Scholar]

[B53] 53. Matthews LA, Selvaratnam R, Jones DR et al. A novel non-canonical forkhead-associated (FHA) domain-binding interface mediates the interaction between Rad53 and Dbf4 proteins. J Biol Chem. 2014;289:2589–99. 10.1074/jbc.M113.517060. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B54] 54. Pepe A, West SC. Substrate specificity of the MUS8–EME2 structure selective endonuclease. Nucleic Acids Res. 2014;42:3833–45. 10.1093/nar/gkt1333. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B55] 55. Duda H, Arter M, Gloggnitzer J et al. A mechanism for controlled breakage of under-replicated chromosomes during mitosis. Dev Cell. 2016;39:740–55. 10.1016/j.devcel.2016.11.017. [DOI] [PubMed] [Google Scholar]

[B56] 56. Gallo-Fernández M, Saugar I, Ortiz-Bazán MÁ et al. Cell cycle-dependent regulation of the nuclease activity of MUS81–EME1/Mms4. Nucleic Acids Res. 2012;40:8325–35. 10.1093/nar/gks599. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B57] 57. Hua Z, Fang Q, Zhang D et al. Crystal structure of the human MUS8–EME2 complex. Structure. 2022;30:743–52. 10.1016/j.str.2022.02.015. [DOI] [PubMed] [Google Scholar]

[B58] 58. Dhoonmoon A, Nicolae CM, Moldovan G-L. The KU-PARP14 axis differentially regulates DNA resection at stalled replication forks by MRE11 and EXO1. Nat Commun. 2022;13:5063. 10.1038/s41467-022-32756-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B59] 59. Lv J, Qian Y, Ni X et al. Feedback regulation of methyl methanesulfonate and ultraviolet-sensitive gene clone 81 via ATM/Chk2 pathway contributes to the resistance of MCF-7 breast cancer cells to cisplatin. Tumour biol. 2017;39:1010428317694307. [DOI] [PubMed] [Google Scholar]

[B60] 60. Sancar A, Lindsey-Boltz L, Unsal-Kaçmaz K et al. Molecular mechanisms of mammalian DNA repair and the DNA damage checkpoints. Annu Rev Biochem. 2004;73:39–85. 10.1146/annurev.biochem.73.011303.073723. [DOI] [PubMed] [Google Scholar]

[B61] 61. Stolarova L, Kleiblova P, Janatova M et al. CHEK2 germline variants in cancer predisposition: stalemate rather than checkmate. Cells. 2020;9:2675. 10.3390/cells9122675. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B62] 62. Boonen RACM, Wiegant WW, Celosse N et al. Functional analysis identifies damaging CHEK2 missense variants associated with increased cancer risk. Cancer Res. 2022;82:615–31. 10.1158/0008-5472.CAN-21-1845. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

CHK2 regulates MUS81-dependent DSBs in response to replication stress and BRCA2 deficiency

Eva Malacaria

Carolina Figlioli

Anita Palma

Masayoshi Honda

Pasquale Valenzisi

Marialuisa Casella

Serena Camerini

Fabio Pucci

Sara Rinalducci

Maurizio Semproni

Maria Spies

Annapaola Franchitto

Pietro Pichierri

Roles

Abstract

Graphical Abstract

Graphical Abstract.

Introduction

Materials and methods

Cell lines and culture conditions

Transfections

Chemicals

Phosphorylation site prediction

Phosphopeptide enrichment

LC–ESI–CID/ETD–MS/MS

Hydrolysis ‘in situ’ and mass spectrometry analysis

Table 1.

Production and purification of GST-fused fragments

Western blot analysis

Table 2.

Immunofluorescence

In situ PLA assay for protein–protein interaction

Quantitative in situ analysis of protein interactions at DNA replication forks (SIRF assay)

DNA fibre assay

Chromatin isolation

Neutral comet assay

Human MUS81–EME1 purification

MUS81–EME1 nuclease assay

Viability assay

Results

CHK2–MUS81 interaction in response to replication stress involves the FHA domain of CHK2 and is disrupted by the I157T CHK2 mutation

Figure 1.

Figure 2.

Formation of the MUS81–CHK2 complex and CHK2 activity promotes DSBs and targets deprotected replication forks in the absence of BRCA2

Figure 3.

Figure 4.

Dual phosphorylation of MUS81 at S95 and S97 licenses DSB formation under replication stress

Figure 5.

Figure 6.

MUS81 phosphorylation promotes recovery of BRCA2-deficient cells upon replication stress

Figure 7.

Discussion

Figure 8.

Supplementary Material

Acknowledgements

Contributor Information

Supplementary data

Conflict of interest

Funding

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases