All EMBO Press journals Open Access as of 1 January 2024 - read the FAQs

12 October 2006
Free access

The structure‐specific endonuclease Mus81–Eme1 promotes conversion of interstrand DNA crosslinks into double‐strands breaks

The EMBO Journal
25: 4921 - 4932
Repair of interstrand crosslinks (ICLs) requires multiple‐strand incisions to separate the two covalently attached strands of DNA. It is unclear how these incisions are generated. DNA double‐strand breaks (DSBs) have been identified as intermediates in ICL repair, but enzymes responsible for producing these intermediates are unknown. Here we show that Mus81, a component of the Mus81–Eme1 structure‐specific endonuclease, is involved in generating the ICL‐induced DSBs in mouse embryonic stem (ES) cells in S phase. Given the DNA junction cleavage specificity of Mus81–Eme1 in vitro, DNA damage‐stalled replication forks are suitable in vivo substrates. Interestingly, generation of DSBs from replication forks stalled due to DNA damage that affects only one of the two DNA strands did not require Mus81. Furthermore, in addition to a physical interaction between Mus81 and the homologous recombination protein Rad54, we show that Mus81−/− Rad54−/− ES cells were as hypersensitive to ICL agents as Mus81−/− cells. We propose that Mus81–Eme1‐ and Rad54‐mediated homologous recombination are involved in the same DNA replication‐dependent ICL repair pathway.


A DNA interstrand crosslink (ICL) covalently connects the two complementary strands of the DNA double helix, thereby blocking important DNA transactions, such as transcription and replication, that require unwinding of the two DNA strands. Because they cause such a dramatic block to accessing genetic information, ICL‐inducing agents are extremely cytotoxic (Dronkert and Kanaar, 2001). ICL‐inducing agents, such as mitomycin C, nitrogen mustards, platinum compounds and psoralens, are more cytotoxic to proliferating cells compared to nondividing cells and therefore they are widely used in chemo‐ and phototherapy of cancers and skin diseases.
Owing to the nature of ICLs, mechanism(s) for their repair are complex. ICLs damage both DNA strands at the same, or very close, nucleotide positions. Therefore, repair mechanisms involving a simple excision followed by templated resynthesis are not sufficient. In Escherichia coli and Saccharomyces cerevisiae, ICL repair requires nucleotide excision repair (NER), homologous recombination and translesion DNA synthesis (Dronkert and Kanaar, 2001; McHugh et al, 2001). In these organisms, NER seems to be involved in generating the incision(s) near the ICL. Homologous recombination on the other hand has several roles in ICL repair (Dronkert and Kanaar, 2001). One important role for homologous recombination is the repair of DNA double‐strand breaks (DSBs), which can result from ICL processing in S. cerevisiae cells (McHugh et al, 2001). Another role, documented for E. coli RecA‐mediated homologous recombination in vitro, is the generation of the substrate for a second round of strand incisions by NER enzymes (Cheng et al, 1991). Alternative roles proposed for homologous recombination repair or homology‐directed repair in ICL repair include mechanisms involving break‐induced replication, single‐strand annealing or the generation of substrates for translesion DNA synthesis (Dronkert and Kanaar, 2001; Niedernhofer et al, 2005).
In vertebrate cells, homologous recombination and translesion DNA synthesis are involved in ICL repair as well (Dronkert and Kanaar, 2001). However, an interesting difference between S. cerevisiae and higher eukaryotes is the role of the NER proteins in ICL repair. Most, if not all, NER proteins in S. cerevisiae cells are involved in ICL repair as deduced from the ICL hypersensitivity of the respective mutants. By contrast, in mammalian cells, mutations in XPF and ERCC1 confer extreme ICL sensitivity, but mutations in other genes essential for NER, including XPA, XPG and CSB, are not dramatically ICL hypersensitive (Dronkert and Kanaar, 2001; De Silva et al, 2002). This suggests that XPF‐ERCC1, a heterodimeric structure‐specific endonuclease (de Laat et al, 1998), plays a central role in ICL repair that is largely independent of NER. In addition to an NER‐independent role in ICL repair, XPF‐ERCC1 functions in at least two subpathways of homology‐directed DNA repair. Both in S. cerevisiae and mammalian cells, XPF‐ERCC1 is involved in DSB repair, through single‐strand annealing, and in homologous gene targeting (Paques and Haber, 1999; Adair et al, 2000; Sargent et al, 2000; Niedernhofer et al, 2001; Langston and Symington, 2005).
Unraveling the mechanism(s) of ICL repair requires answers to two central questions; what are the intermediates in ICL repair at the DNA level and how are they generated? Recently, it has become clear that one pivotal intermediate that can arise during the repair of an ICL is a DSB (Akkari et al, 2000; De Silva et al, 2000; Niedernhofer et al, 2004; Rothfuss and Grompe, 2004). The formation of this DSB intermediate requires DNA replication, suggesting that a stalled replication forks at the site of the ICL is recognized and processed by a structure‐specific endonuclease into a DSB. Owing to the ICL hypersensitivity of cells mutated in the XPF–ERCC1 complex and the biochemical properties of this complex, it has been suggested that XPF–ERCC1 would convert ICLs to DSBs. However, recent studies have demonstrated that this conversion is XPF–ERCC1 independent (De Silva et al, 2000; Niedernhofer et al, 2004), thus leaving the question of how DSBs are generated from ICLs partially unanswered.
Recently, a structure‐specific endonuclease, Mus81–Eme1, with amino‐acid sequence similarity to XPF‐ERCC1 has been identified in yeast and mammalian cells (Heyer, 2004). In yeast, mus81 mutants exhibit sensitivity to hydroxyurea, UV‐light and methyl methanesulfonate but not to ionizing radiation (Interthal and Heyer, 2000; Doe and Whitby, 2004; Doe et al, 2004). This sensitivity profile is consistent with a role for Mus81–Eme1 in processing stalled DNA replication forks. The biochemical properties of the enzyme complex are also consistent with its involvement in forming DSBs at DNA structures resembling replication forks. Like XPF–ERCC1, Mus81–Eme1 also cleaves branched DNA structures (Boddy et al, 2001; Chen et al, 2001; Constantinou et al, 2002; Kaliraman et al, 2001; Ciccia et al, 2003; Ogrunc and Sancar, 2003; Whitby et al, 2003). However, XPF–ERCC1 prefers three‐way branched junctions containing two single‐stranded DNA arms, whereas Mus81–Eme1 has a preference for three‐way junctions that are more double‐stranded in nature, such as 3′‐flap and replication fork‐like structures (Heyer, 2004). Mus81 and Eme1 mutant embryonic stem (ES) cells, which have recently become available, display hypersensitivity to ICL‐inducing agents such as mitomycin C and cisplatin (Abraham et al, 2003; McPherson et al, 2004).


Mus81 is involved in processing ICLs into DSBs

To test whether the Mus81–Eme1 structure‐specific endonuclease is involved in processing ICLs into DSBs, we first constructed mouse ES cells lacking Mus81 (Supplementary Figure 1). Next, we analyzed ICL‐induced DSB formation in Mus81‐proficient and ‐deficient ES cells. Proliferating ES cells were treated with different doses of mitomycin C for 24 h and ICL‐induced DSBs were detected using PFGE (Figure 1A). As we previously demonstrated, mitomycin C treatment resulted in an increase in broken DNA in wild‐type ES cells (Niedernhofer et al, 2004). The increase was dose dependent (Figure 1A) and was not owing to DNA fragmentation during apoptosis (Supplementary Figure 2). In ES cells lacking ERCC1, mitomycin C‐induced DSBs were observed (Figure 1A). However, cells lacking Mus81 mitomycin C failed to induce DSBs, even at the highest dose of mitomycin C used (Figure 1A). Similarly, in response to another ICL‐inducing agent, cisplatin, wild‐type ES cells showed a dose‐dependent increase in broken DNA that was not observed in Mus81−/− ES cells (Figure 2A). We conclude that the structure‐specific endonuclease Mus81–Eme1 is involved in processing ICLs into DSBs.
Figure 1. Analysis of mitomycin C‐induced DSB formation in wild‐type, Ercc1−/− and Mus81−/− ES cells. (A) Using PFGE, DSB formation was analyzed. Cells of the indicated genotype were treated with increasing concentrations of mitomycin C for 24 h, collected into agarose plugs and their DNA was separated by size on an agarose gel. Under the electrophoresis conditions used, high molecular weight genomic DNA remains in the well, whereas lower molecular weight DNA fragments (several Mbp to 500 kbp) migrate into the gel and are compacted into a single band. (B) Cell cycle profiles of wild‐type and Mus81−/− ES cells after continuous treatment with mitomycin C for 24 h. Using a FACscan, the cell cycle profile of cells pulse labeled with BrdU was analyzed by total DNA content as determined by propidium iodide (PI) staining (x‐axis) and replication status as determined by BrdU incorporation (y‐axis). Cells were either untreated or incubated with 2.0 μg/ml mitomycin C. (C) Time course of mitomycin C‐induced DSB formation in wild‐type and Mus81−/− ES cells. Cells were treated with 1.0 μg/ml mitomycin C, indicated by (+). The untreated control sample is indicated by (−).
Figure 2. Analysis of cisplatin‐induced DSB formation in wild‐type and Mus81−/− ES cells. (A) Cells of the indicated genotype were treated with increasing concentrations of cisplatin for 24 h and their DNA was analyzed by PFGE. (B) Clonogenic survival curve of wild‐type and Mus81−/− ES cells in response to increasing doses of cisplatin. (C) Cell cycle profiles of wild‐type and Mus81−/− ES cells after continuous treatment with increasing doses of cisplatin for 24 h. Bi‐parameter (BrdU and PI) FACscan plots are shown.

Mus81 operates in S phase

Given the biochemical activity of Mus81–Eme1 on splayed arm DNA substrates and the lack of mitomycin C‐induced DSBs in Mus81−/−, the role of Mus81–Eme1 in ICL repair is likely the conversion of ICL‐stalled replication forks into DSBs. Thus, Mus81–Eme1 would function during S phase. Consistent with this notion, culturing the cells in the continuous presence of mitomycin C or cisplatin resulted in their accumulation in S phase (Figures 1B and 2C, respectively). If Mus81–Eme1 acts in S phase on stalled replication forks, then the induction of DSBs by mitomycin C should be slow, in contrast to DSB induction by agents that directly act on DNA such as ionizing radiation. Indeed, in wild‐type ES cells cultured in the presence of 1.0 μg/ml mitomycin C, DSB induction became apparent after around 12–18 h and subsequently increased over time (Figure 1C). Again, no DSBs were induced in Mus81−/− cells, even after 30 h of incubation with mitomycin C.
If Mus81–Eme1 functions on ICL‐stalled replication forks, then active replication would be required to detect Mus81‐dependent, ICL‐dependent DSBs. To test this premise, we blocked replication by the addition of thymidine, which resulted in accumulation of the cells in S phase (Figure 3A and B). When mitomycin C was added, no increase in DSBs was detected, either in wild‐type or Mus81−/− ES cells (Figure 3C). In normal growth conditions, without added thymidine to block replication, Mus81‐dependent DSBs were detected upon the addition of mitomycin C. Furthermore, when cells treated with mitomycin C, under conditions of thymidine‐blocked replication (Figure 4A), were allowed to resume replication by incubating them in media lacking thymidine and mitomycin C, Mus81‐dependent DSBs were detected (Figure 4C, compare lanes 7 and 14). Based on the results of the PFGE analysis, the cell cycle analysis and the hypersensitivity of Mus81−/− cells to both mitomycin C and cisplatin (Figures 2B and 7C, respectively), we conclude that Mus81–Eme1 is involved in DSB formation when DNA replication forks are blocked by an ICL.
Figure 3. Inhibition of replication suppresses Mus81‐dependent DSB formation in response to ICLs. (A) Schematic representation of the experimental protocol. Replication in wild‐type and Mus81−/− ES cells was inhibited by incubating them in media containing 20 mM thymidine for 2.5 h. Next, the cells were treated with 2 μg/ml mitomycin C and 20 mM thymidine for 10 h (III). Control cells were either untreated (I), treated with 20 mM thymidine for 2.5 h (II) or 12.5 h (IV), or treated with 2 μg/ml mitomycin C for 10 h (V). (B) Cell cycle profiles of cells treated as described in panel (A) were determined by bi‐parameter (BrdU and PI) FACS analysis. (C) Wild‐type and Mus81−/− ES cells were treated as described above and the DSB formation was analyzed by PFGE.
Figure 4. Mus81‐dependent generation of mitomycin C‐induced DSBs occurs during S phase. (A) Schematic representation of the experimental protocol. Wild‐type and Mus81−/− ES cells were incubated in 5 mM thymidine for 12 h. For the last 6 h of the incubation, mitomycin C was added to a final concentration of 2 μg/ml. Control cells were incubated with thymidine‐containing media only. Next, the cells were washed twice with PBS, incubated in fresh medium to allow resumption of replication, and collected at the indicated times. (B) Cell cycle profiles of cells treated as described in panel (A), shown as BrdU incorporation versus PI plots. (C) The relative amount of broken DNA in wild‐type and Mus81−/− ES cells treated as described in panel (A) was assessed by PFGE.
In the experiments described above, the cells were continuously incubated in the presence of mitomycin C. This resulted in accumulation of the cells in S phase (Figure 1B) and allowed us to observe the involvement of Mus81 in converting ICLs to DSBs. In contrast, when cells were incubated for 1 h in the presence of mitomycin C and subsequently placed in fresh media without mitomycin C, the cells did not accumulate in S phase, irrespective of their genotype (Figure 5A). Instead, 12 h after the mitomycin C pulse most cells were in late S phase, whereas G2, G1 and early S phase cells were detected by 36 h. Under these conditions, DNA breaks were observed in both wild‐type and Mus81−/− ES cells (Figure 5B).
Figure 5. Analysis of DSB formation after pulse treatment of cells with mitomycin C. (A) ES cells of the indicated genotype were treated for 1 h with mitomycin C and continued to be incubated in media without mitomycin C for the indicated amount of time before their cell cycle profile was determined by FACS analysis. The cell count versus PI staining and BrdU incorporation versus PI profile are shown. (B) Wild‐type and Mus81−/− ES cells were treated with mitomycin C for 1 h. Cells were incubated in media without mitomycin C for the indicated amount of time and the amount of broken DNA was determined by PFGE.

Mus81‐mediated DSB formation is DNA lesion selective

Next, we asked whether Mus81–Eme1 is required, in general, to produce DSBs under conditions where DNA damage leads to replication fork stalling. Therefore, instead of using mitomycin C, UV light was used to stall replication through DNA damage induction (Courcelle and Hanawalt, 2001; Branzei and Foiani, 2005). Wild‐type, Mus81+/− and Mus81−/− ES cells were treated with increasing doses of UV light and the cells were analyzed for DSB formation after 4 h. After the treatment, the cells accumulated in S phase, just as was observed after treatment with the ICL‐inducing agents (Figure 6C). A dose‐dependent increase in broken DNA was observed, irrespective of whether Mus81 was functional (Figure 6A). Consistent with this observation, Mus81 was not required for cell survival in response to UV‐light treatment (Figure 6B). By contrast, Mus81−/− cells were hypersensitive to ICL‐inducing agents (Figures 2B and 7C, and McPherson et al, 2004; Dendouga et al, 2005). We conclude that not all stalled replication forks are equivalent and that the replication fork cleavage activity of Mus81–Eme1 depends on the lesion that causes the stalling.
Figure 6. Analysis of UV‐light‐induced DSB formation in wild‐type and Mus81−/− ES cells. (A) UV‐light‐induced DSB formation in wild‐type, Mus81+/− and Mus81−/− ES cells as analyzed by PFGE. (B) Survival curve in response to UV light for wild‐type, and Mus81−/− ES cells. Xpa−/− ES cells served as a control. (C) Cell cycle profiles of wild‐type and Mus81−/− ES cells 4 h after treatment increasing doses of UV light. The BrdU incorporation versus PI profiles are shown.
Figure 7. Analysis of relationship between Mus81 and Rad54 with respect to ICL repair. (A) Immunoprecipitation (IP) of Rad54 and Mus81. Using an anti‐HA‐antibody, HA‐tagged Rad54 protein was precipitated from HA‐tagged Rad54 knock‐in ES cells. The precipitated material was analyzed by immunoblotting using antibodies against Mus81, HA and Rad54. As a negative control, Rad54+/− cells (cell line #18) were used, because the cell line is isogenic to the Rad54HA/− cell line, except for the HA‐tag on Rad54. (B) Identification of the Mus81 and Rad54 proteins in the input material for the immunoprecipitation. Immunoblots using whole‐cell extracts representing 1% of the material used for the immunoprecipitation. As controls for the identification of Mus81 and Rad54, extracts from Mus81−/− and Rad54−/− ES cells were used. Signals from nonspecific proteins are indicated by an asterix. (C) Comparison of mitomycin C sensitivity of wild‐type, Mus81−/−, Rad54−/− and Mus81−/− Rad54−/− ES cells. ES cells of the indicated genotype were treated with increasing doses of mitomycin C for 1 h after which the medium was refreshed. Colonies were fixed, stained and counted after 5–8 days. Error bars indicate the standard error of the mean. (D) Analysis of mitomycin C‐induced DSB formation in wild‐type, Rad54−/−, Mus81−/− and Mus81−/− Rad54−/− ES cells using PFGE.

Physical and genetic interactions between Mus81 and Rad54

A one ended‐DSB such as generated by Mus81–Eme1 from stalled replication forks is an ideal substrate for the initiation of homologous recombination as a next step in ICL repair. Consistent with this notion we observed a reduction of mitomcyin C‐induced sister chromatid exchanges (SCEs) in Mus81−/− cells compared to wild‐type ES cells. Wild‐type cells treated with 0.2 μg/ml mitomycin C displayed 40.8±4.0 SCE per metaphase, whereas Mus81−/− cells showed 31.5±1.9 SCE per metaphase. Interestingly, the spontaneous level of SCEs was already slightly, but significantly (P<0.01) reduced in Mus81−/− cells compared to wild type from 9.9±0.8 to 7.2±0.5 SCEs per metaphase. Possibly, repair of replication forks stalled due to endogenous DNA damage is less likely to proceed through a DSB intermediate in the absence of Mus81.
In addition to reduced mitomycin C‐induced SCEs level in Mus81−/− ES cells, evidence for a link between Mus81 and homologous recombination is also provided by the interaction between S. cerevisiae Mus81 and the homologous recombination protein Rad54 in a two‐hybrid assay and in co‐immunoprecipitation experiments (Interthal and Heyer, 2000). We asked whether a physical and genetic interaction exists between mouse Mus81–Eme1 and Rad54. Whole‐cell extracts were prepared from an ES cell line that carries a Rad54 knockout allele and an HA‐tagged Rad54 knock‐in allele, which expresses HA‐tagged and fully functional Rad54 protein from the endogenous promoter (Tan et al, 1999). HA‐tagged Rad54 protein was precipitated with immobilized anti‐HA antibodies. Immunoblotting of the precipitated samples was used to detect the presence of Mus81, the HA epitope and Rad54 (Figure 7A and B). Co‐immunoprecipitation of Mus81 with Rad54 was detected. In contrast, Mus81 was not detected when the precipitation was performed using extracts prepared from an isogenic ES cell line in which HA‐tagged Rad54 was absent. The interaction is likely protein‐mediated, because DNA in the extracts was digested with DNase I before the immunoprecipitation.
The physical interaction between Mus81–Eme1 and Rad54 is consistent with a function of these proteins in the same ICL repair pathway. The DNA intermediate from which homologous recombination during ICL repair would be initiated is the DSB generated by Mus81–Eme1. Therefore, inactivating mutations in Mus81 should be epistatic to mutations in Rad54 in the context of ICL repair. To test this premise, we generated Mus81−/− Rad54−/− double knockout ES cells and compared their degree of mitomycin C sensitivity to that of either of the single mutants. Rad54−/− ES cells were about three‐fold more sensitive to mitomycin C than wild‐type ES cells (Figure 7C and Essers et al, 1997). The Mus81−/− ES cells displayed a seven‐fold increase in mitomycin C sensitivity. The Mus81−/− Rad54−/− double knockout ES cells were as sensitive to mitomycin C as the Mus81−/− cells (Figure 7C). We conclude that Mus81 is epistatic to Rad54 with respect to repair of ICLs. Consistent with this notion, mitomycin C‐induced DSBs were observed in Rad54−/− ES cells, but not in Mus81−/− Rad54−/− ES cells (Figure 7D).


A DNA ICL covalently links both strands of the DNA double helix and thus its repair requires incisions not only on both sides of the crosslink but also in both DNA strands. Previously, DSBs have been identified as intermediates in ICL repair. Here, we identify Mus81–Eme1 as a structure‐specific endonuclease involved in converting ICLs to DSBs in a DNA replication‐dependent manner.
ICL‐inducing agents cause very heterogeneous types of DNA distortions (Dronkert and Kanaar, 2001). Therefore, recognition of ICLs posses a problem because recognition based on chemical and three‐dimensional structure would require multiple recognition proteins. Instead, cells probably rely on detection methods that do not require direct recognition of the ICL, such as ICL‐induced transcription or replication stalling. For cells in S phase, a replication fork stalled by an ICL can provide a branched DNA structure that triggers the required strand cleavages. However, classical excision repair pathways such as NER or base excision repair alone are not sufficient for ICL repair because these pathways have evolved to cleave only one of the two DNA strands. While it has been established that ICLs are converted into DSBs in a replication‐dependent manner (Akkari et al, 2000; De Silva et al, 2000; Niedernhofer et al, 2004; Rothfuss and Grompe, 2004), the identity of nucleases responsible for this conversion had not been determined.
Mammalian cells contain at least two structure‐specific endonucleases that cleave branched DNA structures: XPF‐ERCC1 and Mus81–Eme1 (Heyer et al, 2003; Heyer, 2004). XPF‐ERCC1, first identified for its function in NER, prefers three‐way branched junctions containing two single‐stranded DNA arms, whereas Mus81–Eme1 cleaves three‐way junctions with at least two double‐stranded arms, such as 3′ flaps and structures resembling replication forks. Ercc1−/− cells are extremely sensitive to ICL‐inducing agents, but the XPF–ERCC1 complex is not involved in the generation of ICL‐induced DSBs and probably plays a role in another step of ICL repair (Niedernhofer et al, 2004). We have generated Mus81−/− ES cells to address whether Mus81 was part of the structure‐specific endonuclease complex responsible for DSB formation after treatment with crosslinking agents. Mus81−/− cells, as well as Eme1−/− cells, are hypersensitive to mitomycin C and cisplatin (Figures 2B and 7C, and Abraham et al, 2003; McPherson et al, 2004; Dendouga et al, 2005). Culturing wild‐type ES cells in the continuous presence of an ICL‐inducing agent results in the accumulation of the cells in S phase (Figures 1B and 2C) and in DSB formation that increases in a time‐ and dose‐dependent manner (Figures 1C, A and 2A). By contrast, no such increase in DSB formation occurs in the absence of Mus81, even through the cells do accumulate in S phase. Inhibition of DNA replication suppresses Mus81‐dependent DSBs in response to mitomycin C (Figure 3), whereas resuming replication resulted in their formation (Figure 4). Taken together, the biochemical activity of the Mus81–Eme1 complex, the mitomycin C hypersensitivity of Mus81−/− and Eme1−/− cells, and the lack of ICL‐induced DSB formation in Mus81−/− S phase cells suggest that Mus81–Eme1 is a structure‐specific endonuclease involved in cleaving one of the branched arms of replication forks stalled by ICL lesions.
While Mus81 is involved in cleaving replication forks stalled at ICLs, it does not seem to be responsible for dealing with DNA damage‐associated replication stalling in general. After treatment with UV light, Mus81−/− cells accumulate a similar amount of DSBs as wild‐type cells (Figure 6). Encountering DNA damage that affects only one strand, such as induced by UV light, will cause problems for a replicative polymerase but may not stop a replicative helicase. The incomplete replicated regions resulting in this case are apparently not substrates for Mus81. However, the DNA structures created, and the challenges to restarting replication, when a fork encounters an ICL are likely to be much different. Because the integrity and movement of a replication fork is determined by the presence and movement of a replicative helicase, halting this enzyme, as an ICL will, will likely cause complete disruption of fork movement. DNA synthesis will stop on both strands and the complex assemblies of replication proteins may disassociate from each other and their DNA templates.
While the results of our experiments reveal the involvement of Mus81–Eme1 in converting ICLs into DSBs, they also demonstrate the importance of the assay to detect these DSBs. When cells are continuously exposed to mitomycin C, they accumulate in S phase and Mus81‐dependent DSBs can be detected (Figure 1). On the other hand, when they are treated with a short pulse of mitomycin C, cells do not accumulate in S phase and while DSBs are detected they are not Mus81‐dependent (Figure 5), consistent with the results of a previous study (Dendouga et al, 2005). It is possible that unrepaired or partially repaired ICL damage, for example gaps or crosslinks between sister chromatids, are still present in G2 cells. Once such cells go through mitosis, this damage can impair proper chromosome segregation, for example as evidenced by an increase in anaphase bridging especially under conditions of reduced ICL repair (Niedernhofer et al, 2004), and lead to mechanical breakage of DNA molecules, both in wild‐type and Mus81‐deficient cells. Alternatively, the Mus81‐independent breaks might occur in the subsequent S phase. The conversion of mitomycin C mono‐adducts to ICLs is slow (Warren et al, 1998). Therefore, the ratio of ICLs to mono‐adducts might be low in the pulse treatment assay because cells do not accumulate in S phase. When DNA‐containing mono‐adducts will be replicated in the subsequent S phase, the mono‐adducts could trigger Mus81‐independent DNA cleavage in a manor analogous to UV‐light‐induced DNA damage (Figure 6).
Our results indicate that Mus81–Eme1 is involved in cleaving a replication fork stalled by an ICL to produce a DSB, which itself is a genotoxic intermediate that has to be repaired before it causes further damage to the genome (Dronkert and Kanaar, 2001; McHugh et al, 2001). In mammalian cells, DSBs can be repaired through two mechanistically distinct pathways; homologous recombination and non‐homologous end joining (van Gent et al, 2001). Homologous recombination mutants are hypersensitive to ICL‐inducing agents, suggesting that recombination is the main pathway involved in DSB‐associated ICL repair (McHugh et al, 1999; De Silva et al, 2000). The one‐ended DSBs created by Mus81–Eme1 cleavage at a stalled replication fork (Supplementary Figure 3) are substrates for homologous recombination rather than nonhomologous end joining (Cromie et al, 2001). Thus, it would be advantageous if Mus81–Eme1 endonuclease would be directly linked to the recombination machinery for efficient processing of the DSBs it generates. Indeed, Mus81 has first been identified in S. cerevisiae through a two‐hydrid interaction with the homologous recombination protein Rad54 (Interthal and Heyer, 2000). Here we show, in mammalian cells, that the Mus81 and Rad54 proteins interact (Figure 7A) and that they are genetically involved in the same ICL survival pathway (Figure 7C). Furthermore, we show that mitomycin C‐induced SCEs are reduced in the absence of Mus81. These results are consistent with the notion that the Mus81‐generated DSBs are further processed by Rad54‐mediated homologous recombination. The additional sensitivity of the Mus81−/− ES cells compared to Rad54−/− ES cells indicates that alternative pathways exists for repair of the ICL‐induced DSBs that do not involve Rad54, consistent with the mild homologous recombination defect of Rad54−/− ES cells (Essers et al, 1997; Dronkert et al, 2000). One such alternative pathway might involve the Rad54 paralog Rad54B (Wesoly et al, 2006). Both Rad54 paralogs interact with Rad51, which provides the catalytic core of homologous recombination. Rad51 assembles into nucleoprotein filaments on single‐stranded DNA and promotes homology recognition and DNA strand exchange, which can result in repair of DSBs (Wyman et al, 2004). In a possible scenario to limit the genotoxicity of the DSB intermediate in ICL repair, Rad51 could assemble on single‐stranded DNA arising at the ICL‐stalled DNA replication fork, before the DSB has occurred. As Rad54 in mammalian cells interacts with Rad51 upon the induction of DNA damage (Tan et al, 1999), it is possible that it is the Rad51 nucleoprotein filament that attracts Rad54. This notion has previously been proposed as a mechanism to target Rad54, which acts on the duplex DNA, to the intact homologous template DNA (Mazin et al, 2000; Solinger et al, 2001). As Rad54 interacts with Mus81, the structure‐specific endonuclease would only get in the proximity of the stalled replication fork when the critical DSB repair proteins are already in place.
Based on the results described above, we propose a model for a Mus81‐dependent ICL repair pathway (Supplementary Figure 3). ICL lesions prevent DNA unwinding required for processive replication and thereby induce stalling of replication forks. The exact structure of DNA strands at an ICL stalled fork is not known. A replicative helicase may stop some base pairs ahead of the crosslinked nucleotides. The presence of single‐stranded gaps in the nascent strands will depend on the resulting mis‐coordination between leading and lagging strands. Regression of the newly synthesized strands or other recombination‐mediated DNA strand exchanges may be needed to assure that the yet to be replicated parental and new daughter DNA strands remain associated for fork recovery after ICL removal. Many scenarios can be envisioned that would result in branched DNA structures that match the in vitro nuclease activity of Mus81–Eme1 for cleavage to a DSB. Although potentially dangerous, DSBs are needed to remove ICLs and may facilitate subsequent repair processes. A DSB may serve as an exit point for stalled replication proteins as well as a release for accumulated positive DNA supercoiling, which inhibits most DNA‐binding proteins. In both cases, DSB formation would promote recruitment of repair enzymes, including incision nuclease(s). The ICL is still in place after Mus81 cleavage (Supplementary Figure 3) and therefore additional strand incisions are required for its removal. The hypersensitivity of Ercc1−/− and XPF mutant cells to ICL‐inducing agents highlights the importance of this complex for ICL repair. The XPF–ERCC1 endonuclease can incise ICL‐containing DNA in vitro (Bessho et al, 1997; Kuraoka et al, 2000). This activity is consistent with a role in incising DNA at one side of an ICL, in the unreplicated DNA region of the cleaved fork shown in Supplementary Figure 3. Subsequently, rotating the crosslinked base out of the helix creates in effect, a single‐stranded gap (Kaye et al, 1980; Matsumoto et al, 1989; De Silva et al, 2000). This idea is supported by the observation that ICL‐induced single‐stranded breaks and single‐stranded gaps are decreased in XPF‐ and ERCC1‐deficient cell lines (De Silva et al, 2000; De Silva et al, 2002; Rothfuss and Grompe, 2004). Homologous recombination between the daughter DNA molecules could re‐establish a replication fork. The Mus81‐generated DSB would be processed such that its 3′ end invades and pairs with the other daughter molecule to create a primer for the polymerase. DNA synthesis over the site of the crosslink would likely require a translesion polymerase to restore the duplex and eventually result in re‐establishing a complete replication fork.
The coordinated cooperation of several DNA repair pathways is clearly required during Mus81‐dependent repair of ICLs during S phase. Additional complexities to understanding ICL repair are provided by the possibility that other Mus81‐independent ICL repair pathways must exist. For example, the ICL sensitivity of Ercc1−/− ES cells is much greater than that of Mus81−/− cells (Figure 7C). Outside of S phase, ICLs cannot be detected by stalled DNA replication, yet they will still be highly toxic owing to their interference with transcription. Recently, compelling evidence has been provided for the existence of a G1 phase ICL repair in S. cerevisiae, which is dependent on NER and a translesion DNA polymerase (Sarkar et al, 2006).

Materials and methods

Cell lines

The cell lines used in this study are described in Supplementary Table 1. The method of de novo isolation of ES cells is described previously (Essers et al, 2000). The generation of Mus81−/− ES cells is described in the Supplementary data. Subconfluent cultures of ES cells were treated with DNA‐damaging agents or thymidine as indicated.

Detection of DSBs by pulse‐field gel electrophoresis

Subconfluent cultures of wild‐type, Mus81+/−, Mus81−/−, Rad54−/−, Rad54−/− Mus81−/− and Ercc1−/− ES cells were treated with mitomycin C or cisplatin for 24 h, unless otherwise indicated. In case of UV‐light treatment, cells were washed with 10 ml phosphate‐buffered saline (PBS), exposed to UV light and incubated in fresh medium to allow repair and cell cycle progression for 4 h. Cells were harvested after trypsinization and agarose plugs containing 106 cells were prepared with a CHEF‐disposable plug mold (Bio‐Rad). The cells were lysed by incubation of the plugs in 1 mg/ml proteinase K in 100 mM EDTA, 0.2% sodium deoxycholate, 1% sodium lauryl sarcosine for 48 h at 37°C and then washed repetitively with 10 mM Tris–HCl, pH 8.0, 100 mM ETDA. Electrophoresis was performed for 23 h at 13°C through 0.9% agarose in Tris–borate–EDTA buffer using a Biometra Rotaphor apparatus with the following parameters: interval, 30–5 s log; angle, 120°–110° linear; 180–120 V log). The DNA was stained with ethidium bromide and visualized using a Typhoon 9200 scanner (Amersham Pharmacia Biotech). The electrophoresis conditions were specifically designed to compact lower molecular weight DNA fragments (several Mbp to 500 kbp) into a single band, while keeping high molecular weight genomic DNA in the well. The lower molecular weight DNA fragments are the results of DSBs in the chromosomal DNA. Thus, the assay allows broken DNA to be readily detected. However, in the context of DSBs arising during DNA replication stalling, the assay has limited sensitivity and is not quantitative. When a DSB occurs at a DNA replication fork, the broken DNA is still attached to the template chromosome. To detect DSBs in the assay, two relatively closely (several Mbp) spaced independent DSBs have to occur.

Flow cytometric analysis

Before collection after the various indicated treatments, ES cells were incubated with 10 μM BrdU for 5 min at 37°C, harvested by trypsynization and fixed overnight with 70% ethanol at 4°C (Smits et al, 2000). After the ethanol was washed away, the cells were treated with 0.1 N HCl containing 0.5 mg/ml pepsin (Merck) for 20 min at room temperature. Next, the cells were treated with 2 N HCl for 12 min at 37°C, followed by the addition of borate buffer (pH 8.5). The cells were washed with PBS containing 0.5% Tween‐20 and 0.1% BSA and incubated with FITC‐conjugated anti‐BrdU antibodies (Becton Dickinson) for 1 h at 4°C. After washing, the cells were counterstained with a solution containing PI (10 μg/ml) and RNase (10 μg/ml) for 30 min at 37°C. The cells were analyzed on a fluorescence‐activated cell sorter (Becton Dickinson) using CellQuest software.

Colony survival assays and SCE analysis

Sensitivity of ES cells to increasing doses of mitomycin C was determined as described previously (Budzowska et al, 2004). Briefly, ES cells of the indicated genotypes were plated in 60 mm dishes, at various dilutions. After overnight incubation, cells were treated with mitomycin C for 1 h. Subsequently, the cells were washed twice with PBS and incubated with fresh medium for 6–8 days. Cells were fixed, stained and colonies were counted. All experiments were performed in triplicate. Similar protocols were followed to determine sensitivities to cisplatin and UV light. Analysis of SCEs was carried out as described (Dronkert et al, 2000).


Subconfluent cultures of ES cells were harvested after trypsinization and lysed in IP lysis buffer (20 mM Tris–HCl, pH 7.5, 1 mM EDTA, 0.5% Triton X‐100, 150 mM NaCl, 10% glycerol, protease inhibitors Complete (Roche)) for 1 h on ice. The lysate was centrifugated for 1 h at 4°C at 46 000 r.p.m. using SW60 rotor (Beckman) and the supernatant was collected into a new tube. To get rid of DNA, 2 mM MgCl2 and DNase I were added and the mixture was incubated at room temperature for 15 min. An anti‐HA antibody attached to an agarose matrix (Roche) was added to the lysate followed by an overnight incubation at 4°C. The precipitate was washed four times with Wash buffer (20 mM Tris–HCl, pH 7.5, 2 mM EDTA, 200 mM NaCl, 0.5 % Triton X‐100).

Supplementary data

Supplementary data are available at The EMBO Journal Online (


We thank Veronique Smits for helpful suggestions and Cecile Beerens, Maurice Kunen and Ellen van Drunen for technical assistance. This work was supported by grants from the Dutch Cancer Society (KWF), the Netherlands Organization for Scientific Research (NWO) and the European Commission.

Supporting Information


Abraham J, Lemmers B, Hande MP, Moynahan ME, Chahwan C, Ciccia A, Essers J, Hanada K, Chahwan R, Khaw AK, McPherson P, Shehabeldin A, Laister R, Arrowsmith C, Kanaar R, West SC, Jasin M, Hakem R (2003) Eme1 is involved in DNA damage processing and maintenance of genomic stability in mammalian cells. EMBO J 22: 6137–6147
Adair GM, Rolig RL, Moore‐Faver D, Zabelshansky M, Wilson JH, Nairn RS (2000) Role of ERCC1 in removal of long non‐homologous tails during targeted homologous recombination. EMBO J 19: 5552–5561
Akkari YM, Bateman RL, Reifsteck CA, Olson SB, Grompe M (2000) DNA replication is required to elicit cellular responses to psoralen‐induced DNA interstrand cross‐links. Mol Cell Biol 20: 8283–8289
Bessho T, Mu D, Sancar A (1997) Initiation of DNA interstrand cross‐link repair in humans: the nucleotide excision repair system makes dual incisions 5′ to the cross‐linked base and removes a 22‐ to 28‐nucleotide‐long damage‐free strand. Mol Cell Biol 17: 6822–6830
Boddy MN, Gaillard PH, McDonald WH, Shanahan P, Yates III JR, Russell P (2001) Mus81–Eme1 are essential components of a Holliday junction resolvase. Cell 107: 537–548
Branzei D, Foiani M (2005) The DNA damage response during DNA replication. Curr Opin Cell Biol 17: 568–575
Budzowska M, Jaspers I, Essers J, de Waard H, van Drunen E, Hanada K, Beverloo B, Hendriks RW, de Klein A, Kanaar R, Hoeijmakers JH, Maas A (2004) Mutation of the mouse Rad17 gene leads to embryonic lethality and reveals a role in DNA damage‐dependent recombination. EMBO J 23: 3548–3558
Chen XB, Melchionna R, Denis CM, Gaillard PH, Blasina A, Van de Weyer I, Boddy MN, Russell P, Vialard J, McGowan CH (2001) Human Mus81‐associated endonuclease cleaves Holliday junctions in vitro. Mol Cell 8: 1117–1127
Cheng S, Sancar A, Hearst JE (1991) RecA‐dependent incision of psoralen‐crosslinked DNA by (A)BC excinuclease. Nucleic Acids Res 19: 657–663
Ciccia A, Constantinou A, West SC (2003) Identification and characterization of the human mus81–eme1 endonuclease. J Biol Chem 278: 25172–25178
Constantinou A, Chen XB, McGowan CH, West SC (2002) Holliday junction resolution in human cells: two junction endonucleases with distinct substrate specificities. EMBO J 21: 5577–5585
Courcelle J, Hanawalt PC (2001) Participation of recombination proteins in rescue of arrested replication forks in UV‐irradiated Escherichia coli need not involve recombination. Proc Natl Acad Sci USA 98: 8196–8202
Cromie GA, Connelly JC, Leach DR (2001) Recombination at double‐strand breaks and DNA ends: conserved mechanisms from phage to humans. Mol Cell 8: 1163–1174
de Laat WL, Appeldoorn E, Jaspers NG, Hoeijmakers JH (1998) DNA structural elements required for ERCC1–XPF endonuclease activity. J Biol Chem 273: 7835–7842
De Silva IU, McHugh PJ, Clingen PH, Hartley JA (2000) Defining the roles of nucleotide excision repair and recombination in the repair of DNA interstrand cross‐links in mammalian cells. Mol Cell Biol 20: 7980–7990
De Silva IU, McHugh PJ, Clingen PH, Hartley JA (2002) Defects in interstrand cross‐link uncoupling do not account for the extreme sensitivity of ERCC1 and XPF cells to cisplatin. Nucleic Acids Res 30: 3848–3856
Dendouga N, Gao H, Moechars D, Janicot M, Vialard J, McGowan CH (2005) Disruption of murine Mus81 increases genomic instability and DNA damage sensitivity but does not promote tumorigenesis. Mol Cell Biol 25: 7569–7579
Doe CL, Osman F, Dixon J, Whitby MC (2004) DNA repair by a Rad22–Mus81‐dependent pathway that is independent of Rhp51. Nucleic Acids Res 32: 5570–5581
Doe CL, Whitby MC (2004) The involvement of Srs2 in post‐replication repair and homologous recombination in fission yeast. Nucleic Acids Res 32: 1480–1491
Dronkert ML, Beverloo HB, Johnson RD, Hoeijmakers JH, Jasin M, Kanaar R (2000) Mouse RAD54 affects DNA double‐strand break repair and sister chromatid exchange. Mol Cell Biol 20: 3147–3156
Dronkert ML, Kanaar R (2001) Repair of DNA interstrand cross‐links. Mutat Res 486: 217–247
Essers J, Hendriks RW, Swagemakers SM, Troelstra C, de Wit J, Bootsma D, Hoeijmakers JH, Kanaar R (1997) Disruption of mouse RAD54 reduces ionizing radiation resistance and homologous recombination. Cell 89: 195–204
Essers J, van Steeg H, de Wit J, Swagemakers SM, Vermeij M, Hoeijmakers JH, Kanaar R (2000) Homologous and non‐homologous recombination differentially affect DNA damage repair in mice. EMBO J 19: 1703–1710
Heyer WD (2004) Recombination: Holliday junction resolution and crossover formation. Curr Biol 14: R56–R58
Heyer WD, Ehmsen KT, Solinger JA (2003) Holliday junctions in the eukaryotic nucleus: resolution in sight? Trends Biochem Sci 28: 548–557
Interthal H, Heyer WD (2000) MUS81 encodes a novel helix‐hairpin‐helix protein involved in the response to UV‐ and methylation‐induced DNA damage in Saccharomyces cerevisiae. Mol Gen Genet 263: 812–827
Kaliraman V, Mullen JR, Fricke WM, Bastin‐Shanower SA, Brill SJ (2001) Functional overlap between Sgs1–Top3 and the Mms4–Mus81 endonuclease. Genes Dev 15: 2730–2740
Kaye J, Smith CA, Hanawalt PC (1980) DNA repair in human cells containing photoadducts of 8‐methoxypsoralen or angelicin. Cancer Res 40: 696–702
Kuraoka I, Kobertz WR, Ariza RR, Biggerstaff M, Essigmann JM, Wood RD (2000) Repair of an interstrand DNA cross‐link initiated by ERCC1–XPF repair/recombination nuclease. J Biol Chem 275: 26632–26636
Langston LD, Symington LS (2005) Opposing roles for DNA structure‐specific proteins Rad1, Msh2, Msh3, and Sgs1 in yeast gene targeting. EMBO J 24: 2214–2223
Matsumoto A, Vos JM, Hanawalt PC (1989) Repair analysis of mitomycin C‐induced DNA crosslinking in ribosomal RNA genes in lymphoblastoid cells from Fanconi's anemia patients. Mutat Res 217: 185–192
McHugh PJ, Gill RD, Waters R, Hartley JA (1999) Excision repair of nitrogen mustard‐DNA adducts in Saccharomyces cerevisiae. Nucleic Acids Res 27: 3259–3266
Mazin AV, Bornarth CJ, Solinger JA, Heyer WD, Kowalczykowski SC (2000) Rad54 protein is targeted to pairing loci by the Rad51 nucleoprotein filament. Mol Cell 6: 583–592
McHugh PJ, Spanswick VJ, Hartley JA (2001) Repair of DNA interstrand crosslinks: molecular mechanisms and clinical relevance. Lancet Oncol 2: 483–490
McPherson JP, Lemmers B, Chahwan R, Pamidi A, Migon E, Matysiak‐Zablocki E, Moynahan ME, Essers J, Hanada K, Poonepalli A, Sanchez‐Sweatman O, Khokha R, Kanaar R, Jasin M, Hande MP, Hakem R (2004) Involvement of mammalian Mus81 in genome integrity and tumor suppression. Science 304: 1822–1826
Niedernhofer LJ, Essers J, Weeda G, Beverloo B, de Wit J, Muijtjens M, Odijk H, Hoeijmakers JH, Kanaar R (2001) The structure‐specific endonuclease Ercc1–Xpf is required for targeted gene replacement in embryonic stem cells. EMBO J 20: 6540–6549
Niedernhofer LJ, Lalai AS, Hoeijmakers JH (2005) Fanconi anemia (cross)linked to DNA repair. Cell 123: 1191–1198
Niedernhofer LJ, Odijk H, Budzowska M, van Drunen E, Maas A, Theil AF, de Wit J, Jaspers NG, Beverloo HB, Hoeijmakers JH, Kanaar R (2004) The structure‐specific endonuclease Ercc1–Xpf is required to resolve DNA interstrand cross‐link‐induced double‐strand breaks. Mol Cell Biol 24: 5776–5787
Ogrunc M, Sancar A (2003) Identification and characterization of human MUS81–MMS4 structure‐specific endonuclease. J Biol Chem 278: 21715–21720
Paques F, Haber JE (1999) Multiple pathways of recombination induced by double‐strand breaks in Saccharomyces cerevisiae. Microbiol Mol Biol Rev 63: 349–404
Rothfuss A, Grompe M (2004) Repair kinetics of genomic interstrand DNA cross‐links: evidence for DNA double‐strand break‐dependent activation of the Fanconi anemia/BRCA pathway. Mol Cell Biol 24: 123–134
Sargent RG, Meservy JL, Perkins BD, Kilburn AE, Intody Z, Adair GM, Nairn RS, Wilson JH (2000) Role of the nucleotide excision repair gene ERCC1 in formation of recombination‐dependent rearrangements in mammalian cells. Nucleic Acids Res 28: 3771–3778
Sarkar S, Davies AA, Ulrich HD, McHugh PJ (2006) DNA interstrand crosslink repair during G1 involves nucleotide excision repair and DNA polymerase zeta. EMBO J 25: 1285–1294
Smits VA, van Peer MA, Essers MA, Klompmaker R, Rijksen G, Medema RH (2000) Negative growth regulation of SK‐N‐MC cells by bFGF defines a growth factor‐sensitive point in G2. J Biol Chem 275: 19375–19381
Solinger JA, Lutz G, Sugiyama T, Kowalczykowski SC, Heyer WD (2001) Rad54 protein stimulates heteroduplex DNA formation in the synaptic phase of DNA strand exchange via specific interactions with the presynaptic Rad51 nucleoprotein filament. J Mol Biol 307: 1207–1221
Tan TL, Essers J, Citterio E, Swagemakers SM, de Wit J, Benson FE, Hoeijmakers JH, Kanaar R (1999) Mouse Rad54 affects DNA conformation and DNA‐damage‐induced Rad51 foci formation. Curr Biol 9: 325–328
van Gent DC, Hoeijmakers JH, Kanaar R (2001) Chromosomal stability and the DNA double‐stranded break connection. Nat Rev Genet 2: 196–206
Warren AJ, Maccubbin AE, Hamilton JW (1998) Detection of mitomycin C‐DNA adducts in vivo by 32P‐postlabeling: time course for formation and removal of adducts and biochemical modulation. Cancer Res 58: 453–461
Wesoly J, Agarwal S, Sigurdsson S, Bussen W, Van Komen S, Qin J, Van Steeg H, Van Benthem J, Wassenaar E, Baarends WM, Ghazvini M, Tafel AA, Heath H, Galjart N, Essers J, Grootegoed JA, Arnheim N, Bezzubova O, Buerstedde JM, Sung P, Kanaar R (2006) Differential contributions of mammalian Rad54 paralogs to recombination, DNA damage repair and meiosis. Mol Cell Biol 26: 976–989
Whitby MC, Osman F, Dixon J (2003) Cleavage of model replication forks by fission yeast Mus81–Eme1 and budding yeast Mus81–Mms4. J Biol Chem 278: 6928–6935
Wyman C, Ristic D, Kanaar R (2004) Homologous recombination‐mediated double‐strand break repair. DNA Repair 3: 827–833

Information & Authors


Published In

The EMBO Journal cover image
Read More
The EMBO Journal
Vol. 25 | No. 20
18 October 2006
Table of contents
Pages: 4921 - 4932

Article versions

Submission history

Received: 3 January 2006
Accepted: 16 August 2006
Published online: 12 October 2006
Published in issue: 18 October 2006


Request permissions for this article.


  1. homologous recombination
  2. interstrand crosslink
  3. Rad54
  4. stalled replication forks
  5. structure‐specific endonucleases



Katsuhiro Hanada
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Magda Budzowska
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Mauro Modesti
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Alex Maas
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Claire Wyman
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Department of Radiation Oncology Erasmus MC Rotterdam The Netherlands
Jeroen Essers
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Department of Radiation Oncology Erasmus MC Rotterdam The Netherlands
Roland Kanaar* [email protected]
Department of Cell Biology & Genetics Erasmus MC Rotterdam The Netherlands
Department of Radiation Oncology Erasmus MC Rotterdam The Netherlands


Corresponding author. Department of Cell Biology & Genetics, Erasmus University, Medical Genetics Center, Erasmus MC, PO Box 2040, 3000 CA Rotterdam, The Netherlands. Tel.: +31 10 408 7186; Fax: +31 10 408 9468; E-mail: [email protected]
These authors contributed equally to this work

Metrics & Citations



Download Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.

Citing Literature

View Options

View options


View PDF

Get Access







Copy the content Link

Share on social media