Deoxyribonucleic acid (DNA) topoisomerases are essential for removing the supercoiling that normally builds up ahead of replication forks. The camptothecin (CPT) Top1 (topoisomerase I) inhibitors exert their anticancer activity by reversibly trapping Top1–DNA cleavage complexes (Top1cc’s) and inducing replication-associated DNA double-strand breaks (DSBs). In this paper, we propose a new mechanism by which cells avoid Top1-induced replication-dependent DNA damage. We show that the structure-specific endonuclease Mus81-Eme1 is responsible for generating DSBs in response to Top1 inhibition and for allowing cell survival. We provide evidence that Mus81 cleaves replication forks rather than excises Top1cc’s. DNA combing demonstrated that Mus81 also allows efficient replication fork progression after CPT treatment. We propose that Mus81 cleaves stalled replication forks, which allows dissipation of the excessive supercoiling resulting from Top1 inhibition, spontaneous reversal of Top1cc, and replication fork progression.
Top1 (topoisomerase I) regulates DNA topology during replication, transcription, and chromatin remodeling (Champoux, 2001; Wang, 2002). It removes DNA torsional stress (supercoiling) by forming cleavage complexes (Top1–DNA cleavage complexes [Top1cc’s]) in which one DNA strand is cleaved by covalent binding of Top1 to a 3′ DNA phosphate. After DNA relaxation, Top1cc’s reverse rapidly, and Top1 is released as the DNA religates. The plant alkaloid camptothecin (CPT) and its clinical derivatives, topotecan and irinotecan, are highly selective Top1 inhibitors that reversibly trap Top1cc by binding at the enzyme–DNA interface (Hsiang et al., 1985; Pommier et al., 2010). At pharmacological concentrations, they kill cancer cells in a replication-dependent manner (Holm et al., 1989; Hsiang et al., 1989), as replication forks collide with the stabilized Top1cc’s. Such collisions are because Top1cc’s do not reverse fast enough ahead of moving replication forks (Pommier et al., 2006), resulting in DNA double-strand ends (DSEs; Hsiang et al., 1989; Strumberg et al., 2000).
The repair of Top1-mediated DNA damage requires both the excision of the covalently attached Top1 from the DNA and the repair of the DSE resulting from replication. A specialized pathway for the excision of Top1cc hinges on TDP1 (tyrosyl-DNA phosphodiesterase I), an enzyme conserved from yeast to humans, which hydrolyzes the covalent bond between Top1 and the 3′ DNA end (Yang et al., 1996; Pouliot et al., 1999; Interthal et al., 2005; Dexheimer et al., 2008a). Genetic studies also indicate that Top1cc can be repaired by other TDP1-independent pathways involving the endonucleases Rad1XPF-Rad10ERCC1 and Mus81-Mms4EME1 (Liu et al., 2002; Vance and Wilson, 2002; Deng et al., 2005; Ciccia et al., 2008; Zhang et al., 2011).
Mus81-Eme1 is a heterodimeric endonuclease that acts preferentially on DNA substrates mimicking stalled replication forks and nicked Holliday junctions in vitro (Interthal and Heyer, 2000; Fricke et al., 2005; Ciccia et al., 2008; Ehmsen and Heyer, 2008). It cleaves such structures in the duplex region adjacent to the branched point (Ehmsen and Heyer, 2009). Mus81-Eme1 has been shown to play a critical role in both replication fork rescue and homologous recombination. It converts collapsed replication forks into DNA double-strand breaks (DSBs; Hanada et al., 2006, 2007; Franchitto et al., 2008; Froget et al., 2008; Shimura et al., 2008) and is considered a back-up system for the resolution of replication fork stalling in the absence of RecQ helicases (Kaliraman et al., 2001; Mullen et al., 2001; Doe et al., 2002; Trowbridge et al., 2007; Franchitto et al., 2008; Shimura et al., 2008). Mus81 is also required for the proper completion of homologous recombination, both during meiosis (Smith et al., 2003; Jessop and Lichten, 2008; Oh et al., 2008) and mitosis (Blais et al., 2004; Roseaulin et al., 2008).
In the present study, we explore the role of Mus81 in the cellular response of human cells to Top1 inhibitors. We show that Mus81 is not involved in the direct excision of Top1cc but rather participates in the repair and recovery of damaged replication forks by incision of the stalled replication forks.
Mus81-deficient cells are hypersensitive to CPT
To assess the potential role of Mus81 in the cellular responses to Top1 inhibitors, we compared survival of wild-type (WT) and Mus81-deficient (Mus81−/−) human HCT116 colon carcinoma cells (Fig. 1 A) after CPT treatment (Holm et al., 1989). Cells were exposed to CPT for 1 h, and CPT sensitivity was assessed by clonogenic survival assays. Fig. 1 B shows that Mus81-deficient cells are significantly more sensitive than WT cells to pharmacological concentrations (0.01, 0.1, and 1 µM) of CPT, which selectively target replicating cells (Huang et al., 2010). At high concentrations of CPT (10 µM), which exceed those achieved in cancer therapy and for which cell lethality is primarily related to transcription-induced damage (Sordet et al., 2009; Zhang et al., 2011), >98% of the WT cells were killed, and the impact of Mus81 knockdown was minimal (Fig. 1 B). These results demonstrate the involvement of Mus81 in the cellular response to Top1 inhibitors under conditions in which Top1 inhibitors primarily target S-phase cells.
To investigate whether the hypersensitivity of Mus81−/− cells resulted from an enrichment in S-phase cells, we determined the cell cycle distributions of WT and Mus81−/− cells by BrdU incorporation assays. WT and Mus81−/− cells showed similar cell cycle profiles, both in the absence and presence of 1 µM CPT for 1 h (Fig. 1 C). Thus, Mus81 deficiency did not result in a higher proportion of S-phase cells, which excludes the possibility that more cells in S phase could account for the increased sensitivity to CPT.
Mus81 does not excise Top1cc’s
Because Mus81-Eme1 is a 3′ flap endonuclease, it has been postulated that it would cleave 5′ from the Top1cc and participate in the excision of trapped Top1 from the DNA (Liu et al., 2002; Vance and Wilson, 2002; Deng et al., 2005). To determine whether Mus81 is involved in Top1cc removal, we analyzed the formation of Top1cc’s in WT and Mus81−/− cells treated with various CPT concentrations. Top1cc’s were detected by Western blotting of cellular DNA fractions with an anti-Top1 antibody (Miao et al., 2006; Zhang et al., 2011). Top1cc levels were similar in WT and Mus81−/− cells upon CPT treatment (Fig. 2, A and B). In addition, Top1cc reversed efficiently in both WT and Mus81−/− cells after CPT removal (Fig. 2 B). Consistent with these results, Mus81 silencing in MDA-MB-231 breast cancer cells did not impair the formation or reversal of Top1cc in response to CPT (Fig. S1, A and B). Under the same experimental conditions, TDP1 silencing resulted in higher Top1cc levels, both during and after CPT treatment (Fig. S2; Miao et al., 2006).
Alkaline elution assays were used to detect Top1cc’s as DNA–protein cross-links (DPCs; Miao et al., 2006) in WT and Mus81−/− cells treated with CPT. DPCs formed similarly in WT and Mus81−/− cells (Fig. 2 C, left, top curves). 30 min after CPT removal, reversal of DPC was partial but was not significantly different in WT and Mus81−/− cells (Fig. 2 C). DPCs fully reversed after 60 min in both WT and Mus81−/− cells (Fig. 2 C, right). Together, these results demonstrate that Mus81 is not involved in the excision of Top1 from the DNA.
Mus81 generates DNA DSBs in CPT-treated cells
To elucidate the role of Mus81 in the cellular response to Top1-induced DNA damage, we next analyzed the influence of Mus81 on the formation of Top1-induced DNA DSBs. WT and Mus81−/− cells were exposed to various concentrations of CPT for 1 h, and the DSB marker γ-H2AX (Rogakou et al., 1998; Furuta et al., 2003) was examined by immunofluorescence. CPT-induced γ-H2AX levels were greatly reduced in Mus81-deficient cells (Fig. 3 A), indicating that Mus81 is involved in the formation of DSBs in CPT-treated cells. This result was further supported by neutral comet assays showing that Mus81−/− cells accumulated less DSBs than WT cells in response to CPT (Fig. 3 B). Time-course analysis showed reduced γ-H2AX induction in Mus81-deficient cells at both short and long time points after CPT treatment (Fig. 3 C).
Mus81−/− cells also demonstrated a lower and delayed induction of other DNA damage response signals in response to CPT, including phosphorylation of Chk2 at threonine 68 (ATM [ataxia telangiectasia mutated] substrate) and phosphorylation of RPA2 (Fig. 3 C). However, phosphorylation of Chk1 at serine 317 (ATR [ataxia telangiectasia and Rad3 related] substrate) was induced similarly in Mus81−/− and WT cells upon CPT treatment, indicating reduced DSBs but efficient activation of the replication checkpoint in Mus81−/− cells (Fig. 3 C).
siRNA-mediated depletion of Mus81 in MDA-MB-231 cells also resulted in a marked reduction of CPT-induced γ-H2AX (Fig. S1, C and D). As Mus81 functions as a heterodimer with Eme1 (Ciccia et al., 2008), we measured γ-H2AX induction upon CPT exposure in HCT116 cells transfected with Eme1-targeting siRNAs. Eme1 silencing resulted in a marked reduction in CPT-induced γ-H2AX, recapitulating the phenotype of Mus81−/− cells (Fig. S1 E). Collectively, these results demonstrate that Mus81 is involved in the formation of DSBs in CPT-treated cells.
Mus81-dependent DNA DSBs form at replication foci
CPT-induced DSBs result from interference of stabilized Top1cc with either replication or transcription machineries (Hsiang et al., 1989; Ryan et al., 1991; Strumberg et al., 2000; Pommier et al., 2006; Sordet et al., 2009). To investigate whether Mus81-dependent DNA breaks are related to replication (Seiler et al., 2007) or transcription (Sordet et al., 2009; Zhang et al., 2011) in CPT-treated cells, we next analyzed γ-H2AX induction in replicating and nonreplicating cell subpopulations. WT and Mus81−/− cells were labeled with the thymidine analogue 5-ethynyl-2′-deoxyuridine (EdU; to detect replicating cells) before and during CPT treatment. EdU and γ-H2AX costaining showed that γ-H2AX was induced in both replicating (EdU positive) and nonreplicating (EdU negative) cells, with higher γ-H2AX levels being observed in the replicating cells (Fig. 4 A). Mus81 deficiency was associated with significantly lower γ-H2AX levels in replicating cells but did not affect γ-H2AX induction in nonreplicating cells (Fig. 4, A and B).
The occurrence of Mus81-dependent DNA breaks in replicating cells is consistent with the cleavage of stalled DNA replication forks by Mus81 (see Introduction). To address this possibility, we performed single-cell analyses of EdU and γ-H2AX foci. In WT cells treated with CPT, 68% of the replication foci (labeled with EdU) colocalized with γ-H2AX foci (Fig. 4, C and D; and Fig. S3, top), which is consistent with the preferential induction of DSBs by CPT in replication foci (Seiler et al., 2007). On the other hand, in Mus81-deficient cells, the fraction of damaged replication foci (i.e., those EdU foci colocalizing with γ-H2AX foci) was significantly decreased (39%; Fig. 4, C and D; and Fig. S3, bottom), and the fraction of γ-H2AX foci outside the EdU foci was significantly increased (Fig. 4, C and E; and Fig. S3). These results indicate the selective induction of Mus81-dependent DSBs at replication foci.
Mus81 promotes replication fork progression after Top1 inhibition
To study the effects of Mus81 on replication fork progression, we analyzed replication in single DNA molecules using DNA combing (Conti et al., 2007; Seiler et al., 2007). Incorporation of the thymidine analogues iododeoxyuridine (IdU) and chlorodeoxyuridine (CldU) was visualized on stretched DNA fibers prepared from untreated or CPT-treated WT and Mus81−/− cells. Cells were labeled with IdU and CldU for 30 min each, and CPT was added during the last 20 min of the IdU pulse (Fig. 5 A). This protocol allows the analysis of both replication fork slow down (by comparing replication fork velocity in the presence and in the absence of CPT) and replication fork recovery (by comparing fork velocity during CPT treatment and after CPT removal). In both WT and Mus81−/− cells, CPT induced a marked reduction in replication fork velocity (Fig. 5, B and C; and Fig. S4). After CPT removal, CldU incorporation was measured in both WT and Mus81−/− cells (Fig. 5, B and C). In WT cells, replication fork velocity resumed to ∼50% of its mean value before CPT (Fig. 5 B, left), whereas, in Mus81−/− cells, there was no detectable recovery upon CPT removal (Fig. 5 B, right). These results show that Mus81 is associated with recovery of replication fork velocity after CPT removal.
CPT and its clinically used derivatives topotecan and irinotecan are primarily cytotoxic to S-phase cells at pharmacological concentrations (Holm et al., 1989; Hsiang et al., 1989), as they induce replication-associated DSBs that are primarily processed by homologous recombination (Eng et al., 1988; Nitiss and Wang, 1988; Vance and Wilson, 2002; Hochegger et al., 2006; Nakamura et al., 2010). At least some of those DSBs have been shown to result from the conversion of Top1-associated single-strand breaks into DSBs by replication runoff (Fig. 6 A; Hsiang et al., 1989; Tsao et al., 1993; Shao et al., 1999; Strumberg et al., 2000).
Our study proposes an alternative by showing that the structure-specific endonuclease Mus81-Eme1 is responsible for the generation of DSBs at stalled replication forks in response to Top1 trapping. This conclusion is based on our findings that Mus81 inactivation either by stable knockout (Hiyama et al., 2006) or siRNA reduces CPT-induced DSBs selectively in replicating cells and at replication foci (Fig. 4). Our conclusion is consistent with the known biochemical activities of Mus81-Eme1, which preferentially processes substrates mimicking replication forks (Kaliraman et al., 2001; Doe et al., 2002; Fricke et al., 2005; Ehmsen and Heyer, 2008). The Mus81-dependent DSBs are not lethal but are associated with DNA repair. Indeed, inactivation of Mus81 results in more cell killing (Fig. 1 B) and less DSBs (Figs. 3 and 4). Mus81-dependent DNA breakage has also been observed in mammalian cells upon treatment with cisplatin (Hanada et al., 2006) and with the replication fork-blocking drugs aphidicolin and hydroxyurea (Hanada et al., 2007; Franchitto et al., 2008; Froget et al., 2008; Shimura et al., 2008). These studies have established that Mus81 can convert a stalled replication fork into a DSB. Here, we show the involvement of Mus81 in the formation of replication-associated DSBs after CPT treatment rather than in the direct excision of Top1 from the DNA. Our findings are best explained by the possibility that Top1 inhibition leads to replication fork stalling, which is resolved by Mus81-dependent DNA cleavage (Fig. 6, B–H). This interpretation is consistent with an independent publication proposing that Top1 inhibition by CPT in yeast leads to supercoiling accumulation ahead of replication forks, leading to replication fork stalling (Koster et al., 2007).
Because Mus81 is responsible for only a fraction of the replication-associated DSBs after Top1 inhibition (Fig. 4, C and D), it is plausible that Top1-mediated replication-associated DSBs are produced both by Mus81-dependent and Mus81-independent mechanisms. Thus, it is likely that Top1 trapping induces replication-associated DSBs either by replication runoff (Fig. 6 A) or by Mus81-dependent cleavage of replication forks stalled by positive supercoils (Fig. 6 C). Replication fork cleavage by Mus81-Eme1 could allow the dissipation of the excessive superhelical tension (Fig. 6, C and D), which could resolve the topological block resulting from Top1 deficiency ahead of the fork. This hypothesis is supported by genetic experiments in yeast, showing that mutations in both Mus81 and Mms4Eme1 results in growth defects in top1Δ strains (Mullen et al., 2001). It is also supported by the recent findings that Top1 deficiency in yeast, murine, and human cells is associated with genomic instability (Christman et al., 1993; Miao et al., 2007; Tuduri et al., 2009). Top1 deficiency in yeast is associated with enhanced genomic instability in the ribosomal DNA cluster, suggesting that Top1 may be particularly important in highly transcribed regions (Christman et al., 1993). In mammalian cells, Top1 deficiency leads to an accumulation of stalled replication forks and DNA breakage at replication sites (Miao et al., 2007; Tuduri et al., 2009). Because Top1 is required for efficient DNA relaxation during replication (Kim and Wang, 1989; Koster et al., 2007), Top1-deficient cells probably accumulate unresolved supercoiling in replicating DNA, which further leads to replication fork stalling and breakage.
We also show that Mus81 is required for efficient replication fork progression after CPT removal. Parallel observations have been made in cells treated with hydroxyurea or aphidicolin (Hanada et al., 2007; Shimura et al., 2008), indicating that Mus81-induced DSBs act as resolution intermediates for replication fork recovery. DSBs arising during replication are repaired primarily by homologous recombination (Takata et al., 1998; Arnaudeau et al., 2001; Rothkamm et al., 2003; Sonoda et al., 2006). Thus, Mus81-dependent DSBs could support strand invasion and initiate sister chromatid exchange. Mus81 is also required for the resolution of Holliday junctions (Blais et al., 2004; Fricke et al., 2005; Ehmsen and Heyer, 2008; Jessop and Lichten, 2008; Oh et al., 2008; Roseaulin et al., 2008) and could therefore participate in both initiation and completion of homologous recombination during S phase (Fig. 6 E). Alternatively, Mus81-dependent fork cleavage could enable replication progression by allowing retraction of the replication machinery (Fig. 6, D–H), possibly in association with the Bloom helicase. After nucleolytic resection of the 5′ end of the DNA, annealing of the two newly synthesized DNA strands would initiate replication fork regression, allowing reannealing of the two parental DNA strands and gap filling by DNA polymerase.
In conclusion, our study demonstrates that the Mus81-Eme1 endonuclease is involved in the repair of Top1-mediated DNA damage by promoting the cleavage and restoration of stalled replication forks. This mechanism may not be limited to CPT-induced DNA damage, as Top1 can be trapped by a variety of endogenous DNA lesions, such as oxidized bases, mismatches, abasic sites, adducts, and strand breaks (Pourquier and Pommier, 2001; Pommier et al., 2006; Dexheimer et al., 2008b). Mus81 is therefore a novel and important determinant of the cellular response of cancer cells to Top1 inhibitors and replication fork stalling by Top1 deficiency and excessive supercoiling.
Materials and methods
Cell lines and drugs
Human MDA-MB-231 breast adenocarcinoma cells were obtained from the Developmental Therapeutics Program (National Cancer Institute) and were maintained in RPMI 1640 supplemented with 10% fetal bovine serum. Human HCT116 colorectal carcinoma cells and HCT116 Mus81−/− cells (Hiyama et al., 2006) were grown in DME supplemented with 10% fetal bovine serum. CPT was obtained from the Drug Synthesis and Chemistry Branch (National Cancer Institute).
Clonogenic survival assay
After drug treatment, cells were seeded in 6-well plates at a density of 50, 250, or 500 cells/well and incubated for 10–13 d to allow colony formation. Colonies were fixed with methanol and stained with 0.05% methylene blue (Sigma-Aldrich) for 30 min. The surviving fraction was calculated by dividing the number of colonies in treated wells by the number of colonies in untreated wells.
Alkaline elution assays
DPCs were detected using alkaline elution as previously described (Covey et al., 1989). In brief, cells were radiolabeled overnight with 0.02 µCi/ml [14C]thymidine and chased with radioisotope-free medium 4 h before drug treatment. Cell aliquots were placed in ice-cold HBSS and irradiated with 30 grays to break the DNA. Cells were layered onto polyvinylchloride-acrylic copolymer (protein adsorbing) filters and lysed with LS-10 (2 M NaCl, 0.2% sarkosyl, and 0.04 M disodium EDTA, pH 10). DNA was eluted from the filters with tetrapropylammonium hydroxide-EDTA. After elution, filters were incubated for 1 h at 65°C with 1 M HCl and an additional hour at room temperature in the presence of 0.3 M NaCl. Radioactivity in fractions and filters was measured with a liquid scintillation analyzer (2200A Tri Carb Scintillation Analyzer; Packard Instruments).
BrdU incorporation assay
Cells were pulse labeled with 50 µM BrdU (EMD) during the last 10 min of CPT treatment. Cells were harvested by trypsinization, fixed in 70% ice-cold ethanol, and stored at −20°C. Cells were incubated for 30 min at room temperature in 2 N HCl–0.5% Triton X-100 to allow DNA denaturation. The medium was neutralized in 0.1 M sodium borate, pH 8.5, and cells were washed twice in PBS containing 0.5% Tween 20 and 0.5% bovine serum albumin. Cells were incubated for 1 h at room temperature with an FITC-conjugated anti-BrdU antibody (BD) and were treated with 0.5 mg/ml RNase A and 50 µg/ml propidium iodide. Samples were analyzed with a flow cytometer (FACScan; BD) using the CellQuest software (BD).
Detection of Top1cc’s
After drug treatment, cells were lysed in a reagent (DNAzol; Invitrogen), and genomic DNA was prepared according to the manufacturer’s instructions. Samples were sonicated briefly to shear the DNA. Varying concentrations of DNA were spotted on transfer membranes (Immobilon-FL; Millipore) using a slot-blot manifold (GE Healthcare). Membranes were blocked in blocking buffer (Odyssey; LI-COR Biosciences) and incubated successively with a C21 anti-Top1 antibody (gift from Y.-C. Cheng, Yale University, New Haven, CT) and with the goat anti–mouse secondary antibody (IRDye 800CW; LI-COR Biosciences). Both antibodies were diluted in blocking buffer. Membranes were imaged with the Odyssey Infrared Imaging System (LI-COR Biosciences).
For immunofluorescence assays, cells were grown in chamber slides (Lab-Tek; Thermo Fisher Scientific). The staining for γ-H2AX was performed as previously described (Zhang et al., 2011). After drug treatment, cells were fixed and permeabilized by a 20-min incubation at room temperature in 2% paraformaldehyde and an overnight incubation at 4°C in 70% ethanol. Slides were blocked in 8% bovine serum albumin and stained successively with anti–γ-H2AX antibody (Abcam) and with a fluorescent secondary antibody (anti–mouse Alexa Fluor 488; Invitrogen). Both antibodies were diluted in 1% bovine serum albumin. Slides were mounted in mounting medium containing DAPI (Vectashield; Vector Laboratories) and visualized using a confocal microscope (Eclipse TE300; Nikon). Fluorescent signals in individual cells were quantified using Photoshop CS5 (Adobe).
For the simultaneous detection of γ-H2AX and replication foci, cells were labeled with 30 µM EdU for 90 min. 1 µM CPT was added during the last 60 min of the EdU pulse. γ-H2AX staining was performed as described in the previous paragraph, and EdU was detected with a flow cytometry assay kit (Click-iT EdU Alexa Fluor 647 Flow Cytometry Assay; Invitrogen) according to the manufacturer’s instructions. Confocal images were sequentially acquired with ZEN (2009, SP1; Carl Zeiss) software on a confocal system (LSM 510; Carl Zeiss) with an inverted microscope (Axio Observer.Z1; Carl Zeiss) and a UV laser tuned to 364 nm, a 25-mW argon visible laser tuned to 488 nm, and a 5-mW HeNe laser tuned to 633 nm. A 63× Plan Apochromat 1.4 NA oil immersion objective was used. Emission signals after sequential excitation of DAPI, Alexa Fluor 488, and Alexa Fluor 633 by the 364-, 488-, or 633-nm laser lines were collected with a band pass 385–470, band pass 505–550, or long pass 650 filter, respectively, using individual photomultipliers. Images were acquired at room temperature, and the mounting medium was Vectashield with DAPI. Images were adjusted using Photoshop and combined using Illustrator (Adobe).
Neutral comet assay
DNA DSBs were detected using the neutral comet assay according to the gel electrophoresis kit protocol (CometAssay; Trevigen; Zhang et al., 2008). Comet tail moments were measured with CometScore 1.5 software (TriTek Corporation).
Whole-cell extracts were obtained by homogenization of cell pellets in lysis buffer (1% SDS and 10 mM Tris-HCl, pH 7.4) supplemented with proteases and phosphatases inhibitors (Roche). Proteins were separated by SDS-PAGE electrophoresis and immunoblotted with the following antibodies: anti-Mus81 (Abcam), anti–γ-H2AX, antiphospho-RPA2 (S4/8; Novus Biologicals), antiphospho-Chk1 (S317), antiphospho-Chk2 (T68), anti–glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Cell Signaling Technology), and anti–β-actin (Sigma-Aldrich).
Cells seeded in 6-well plates or LabTek chambers were transfected with targeting or nontargeting siRNAs using transfection reagent (DharmaFECT; Thermo Fisher Scientific). Mus81-targeting, Eme1-targeting, and nontargeting siRNAs were products of Thermo Fisher Scientific. TDP1-targeting siRNAs were purchased from QIAGEN.
DNA replication profiling by molecular combing
Molecular combing was performed as previously described (Conti et al., 2007; Seiler et al., 2007). In brief, asynchronous exponentially growing cells were labeled sequentially with the thymidine analogues IdU and CldU for 30 min each. To study replication fork arrest and restart, 1 µM CPT was added during the last 20 min of the IdU pulse. At the end of the CldU pulse, cells were harvested by trypsinization and embedded in low-melting agarose plugs. Agarose plugs were treated with proteinase K and then melted at 70°C in the presence of 4-morpholinepropanesulfonic acid, pH 6.5. After β-agarase digestion, DNA was combed on silanized surfaces (MicroSurfaces, Inc.). DNA fibers were denatured in 0.5 M NaOH and probed with the following primary antibodies: mouse anti-BrdU fluorescein isothiocyanate (IdU specific; BD) and rat anti-BrdU (CldU specific; Accurate Chemical and Scientific Corp.). After labeling with fluorescent secondary antibodies (anti–mouse Alexa Fluor 488 and anti–rat Alexa Fluor 594; Invitrogen), slides were mounted in Vectashield mounting medium. Images were captured with the AttoVision software (BD) using an epifluorescence microscope (Pathway; BD). Replication signals were measured using ImageJ (National Institutes of Health) with custom-made modifications. Measurements were converted from micrometers to kilobases according to a constant stretching factor (1 µm = 2 kb). Fork velocities were calculated by dividing the length of a replication signal (in kilobases) by the labeling time (in minutes).
All results are presented as means ± SEM or SD. Differences between samples were assessed using the Mann–Whitney test or the Student’s t test, depending on the distribution of the sample. The normal distribution of a sample was tested with the Kolmogorov–Smirnov test. All analyses were conducted with Prism 5.0 (GraphPad Software). P-values were two sided and considered statistically significant when <0.05.
Online supplemental material
Fig. S1 shows the effect of Mus81 and Eme1 silencing on the cellular response to CPT. Fig. S2 shows the effect of TDP1 silencing on the formation and reversal of CPT-induced Top1cc’s. Fig. S3 shows additional microscopy images of replication foci (EdU foci) and γ-H2AX foci in CPT-treated WT and Mus81−/− cells. Fig. S4 shows the frequency distribution of replication fork velocities in WT and Mus81−/− cells.
We thank Drs. Shar-Yin Huang, Elisabetta Leo, and Benu Brata Das for helpful discussions during the conduction of the study. We thank Drs. Céline Douarre, Mathieu Métifiot, Christophe Redon, and Asako Nakamura for technical assistance. We also thank Dr. Susan Garfield for providing the information of image acquisition.
This work was supported by the Center for Cancer Research (National Cancer Institute Intramural Research Program, National Institutes of Health).
M. Regairaz and Y.-W. Zhang contributed equally to this paper.