The mammalian mismatch repair protein MSH2 is required for correct ...

Oncogene (2003) 22, 2110–2120

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

The mammalian mismatch repair protein MSH2 is required for correct MRE11 and RAD51 relocalization and for efficient cell cycle arrest induced by ionizing radiation in G2 phase Annapaola Franchitto1,3,4, Pietro Pichierri1,3,4, Rita Piergentili1, Marco Crescenzi2, Margherita Bignami2 and Fabrizio Palitti*,1 Laboratorio di Citogenetica Molecolare e Mutagenesi – DABAC, Universita` degli Studi della Tuscia, Via S, Camillo de Lellis, 01100 Viterbo, Italy; 2Laboratorio di Tossicologia Comparata ed Ecotossicologia, Istituto Superiore di Sanita`, Viale Regina Elena 299, 00161 Roma, Italy

1

In yeast, MSH2 plays an important role in mismatch repair (MMR) and recombination, whereas the function of the mammalian MSH2 protein in recombinational repair is not completely established. We examined the cellular responses of MSH2-deficient mouse cells to X-rays to clarify the role of MSH2 in recombinational repair. Cell survival, checkpoint functions and relocalization of the recombination-related proteins MRE11 and RAD51 were analysed in embryonic fibroblasts derived from MSH2+/+ and MSH2/ mice, and in MSH2-proficient and deficient mouse colorectal carcinoma cells. Loss of MSH2 function was found to be associated with reduction in cell survival following radiation, absence of either MRE11 or RAD51 relocalization and a higher level of X-ray-induced chromosomal damage specifically in G2-phase cells. Finally, MSH2/ cells showed an inefficient early G2/M checkpoint, being arrested only transiently after irradiation before progressing into mitosis. Consistent with the premature release from the G2-phase arrest, activation of CHK1 was transient and CHK2 was not phosphorylated in synchronized MSH2-null cells. Our data suggest that an active MSH2 is required for a correct response to ionizing radiation-induced DNA damage in the G2 phase of the cell cycle, possibly connecting DSB repair to checkpoint signalling. Oncogene (2003) 22, 2110–2120. doi:10.1038/sj.onc.1206254 Keywords: MSH2; mismatch repair; cell cycle arrest; DSB repair

Introduction The DNA mismatch repair (MMR) system is a highly conserved post replicative editing process that prevents

*Correspondence: F Palitti; E-mail: [email protected] 3 Current address: CNRS, UPR2169, Instabilite´ Geńe´tique et Cancer, Institut Andre´ Lwoff, 7 Rue Guy-Moquet, 94801 Villejuif, France 4 The first two authors contributed equally Received 17 July 2002; revised 19 November 2002; accepted 21 November 2002

mutations (for a review see Kolodner, 1996; Modrich and Lahue, 1996). Germline mutations in MMR genes are associated with predisposition to colorectal and other cancers (for a review see Jiricny and NystromLahti, 2000). While the role of MMR in the correction of replication errors is well established, its involvement in the processing of DNA damage induced by chemical and physical agents is less clear. Thus, whereas inactivation of MMR always leads to tolerance to methylating agents (Karran and Bignami, 1994), either moderate levels of resistance (Aebi et al., 1996; Anthoney et al., 1996) or hypersensitivity to other types of DNA damage has been observed in MMR-defective cells (Fiumicino et al., 2000; Pichierri et al., 2001). For example, MSH2/, MLH1/ and PMS2/ mouse embryo fibroblasts (MEFs) show a modest increase in survival compared to wild-type cells after exposure to ionizing radiation (IR), while conflicting results were observed in other human cellular models (Aquilina et al., 1999; Davis et al., 1998; DeWeese et al., 1998; Fritzell et al., 1997; Xu et al., 2001b; Zeng et al., 2000). In addition, the MMR proteins have been involved in the apoptotic response (Bellacosa, 2001; Berry and Kinsella, 2001; Kaina et al., 2001). The mammalian MSH2 protein forms heterodimers with either MSH3 or MSH6 and these complexes initiate MMR by recognition of base–base mismatches and short insertion/ deletion loops (for a review see Jiricny, 1998; Kolodner and Marsischky, 1999). In Saccharomyces cerevisiae, Msh2 and Msh3 also participate directly in recombinational repair (Alani et al., 1994). The yeast Msh2–Msh3 complex facilitates removal of nonhomologous ends that are formed during gene conversion (Sugawara et al., 1997) or during repair by the single-strand annealing pathway. A role for the Msh2 protein in the suppression of recombination between divergent sequences has also been reported both in yeast and mammalian cells (Chen and Jinks-Robertson, 1999; Ciotta et al., 1998; de Wind et al., 1995; Worth et al., 1994; Zhang et al., 2000). IR induces a large variety of DNA lesions, including double-strand breaks (DSBs), single strand breaks (SSBs), base and sugar damage (Hutchinson, 1985). DSBs are considered to be potentially the most lethal

MSH2 and recombinational repair in G2 phase A Franchitto et al

2111

lesions (Bradley and Kohn, 1979). In S. cerevisiae, the predominant DSB repair pathway is the homologymediated recombinational repair (HR) (Kanaar et al., 1998). HR is also an important repair pathway in mammalian cells (Liang et al., 1998; Sonoda et al., 1998; Takata et al., 1998). The initial step in DSBs repair by HR involves processing of DNA ends to produce 30 single-strand tails that are the substrate for homologous pairing and strand invasion (Kanaar et al., 1998; White and Haber, 1990). The identity of the nuclease in the processing step is still uncertain, but MRE11 is a candidate for this activity (Paull and Gellert, 1998; Trujillo et al., 1998). The RAD51 and RAD54 proteins act during the pairing and the strand exchange steps by forming a nucleoprotein filament along the 30 singlestranded tails. Recombination intermediates produced in this way are then processed further in reactions that involve the resolution of Holliday junctions, DNA synthesis and nick ligation (Baumann and West, 1998). In higher eukaryotes, DSBs are repaired via two pathways: the so-called nonhomologous end-joining (NHEJ) and HR (Kanaar et al., 1998). Vertebrate cells preferentially repair DSBs by HR in the late-S and G2 phases of the cell cycle when an undamaged sister chromatid is available, while NHEJ appears to be confined mainly to the G1/early-S phases (Takata et al., 1998). Among the proteins implicated in DSB repair, MRE11 seems crucial for both HR and NHEJ (Haber, 1998) whereas RAD51 is essential for HR (Kanaar et al., 1998). To investigate the role for mammalian MSH2 in DSBs repair following IR, we studied whether the absence of the MSH2 protein results in an altered response to DNA damage. To this aim, we have used mouse embryonic fibroblasts (MEFs) derived from MSH2 knockout mice and mouse colorectal carcinoma cells either proficient or defective in MSH2 (Humbert et al., 1999). We report that the absence of MSH2 leads to reduced survival and is also associated with aberrant RAD51and MRE11-focus formation in cells irradiated in G2 phase, but not in S phase. Moreover, a higher level of chromosomal damage was found in MSH2/ cells irradiated in the G2 phase of the cell cycle. Interestingly, loss of MSH2 results in premature release from the early-G2/M arrest following IR, with inability to activate properly the checkpoint kinases CHK1 and 2. We suggest that active MSH2 protein is required for both proper early checkpoint function and efficient DSBs repair after IR in mammalian cells.

an additional 12 h. At 4 h after release from the double thymidine block, 80% of the cell population was in the late-S/ G2 of the cell cycle. Distribution in the different stages of the cell cycle was assessed by standard cytofluorimetric techniques, evaluating the DNA content after propidium iodide (PI) staining of ethanol-fixed cells. X-ray irradiation Irradiation was carried out at 371C with a Gilardoni X-ray apparatus (MGL 300/6-D) operating at 300 kV and 6 mA, at a dose rate of 2 Gy/min. After irradiation, cells were recovered in complete medium for the indicated time until harvested. Determination of cell survival and apoptotic cell death To measure cell survival, cells were seeded in complete medium (100 cells/6 cm dish) and treated with X-rays 18 h later. After 1 week, the surviving colonies were stained with Giemsa and counted. The results are presented as a fraction of the relevant control (surviving fraction). For the analysis of the IR-induced apoptotic cell death, MEFs were seeded onto coverslips, irradiated with 4 or 8 Gy and processed at different postirradiation times for the analysis of apoptosis by a TdTmediated end-labelling kit as described by the manufacturer (Roche Biochemicals). Determination of the early-G2 arrest Progression of cells into mitosis from the G2 phase was assessed by analysis of the mitotic indices. Cells were treated either with X-ray (2 or 4 Gy) or UV (20 J/m2) and cultured for different time periods in the presence of 0.02 mg/ml colcemid to accumulate mitotic cells. Slides prepared at each time point were stained with PI and mitotic indices were determined by counting at least 1000 cells per culture. No differences in the mitotic index were detected in untreated samples from either MSH2-proficient or MSH2-deficient cells. Moreover, no differences in the response to colcemid were observed among the different cell lines used (our unpublished results). Western blot analysis

Materials and methods

Cell extracts (20 mg) were denatured and separated on 10 or 7% SDS–polyacrylamide gels according to standard procedure. Proteins were transferred to PVDF membranes (BioRad) and immunoblotting was carried out according to standard procedures. Blots were separately incubated with primary antibodies against CDC2 (mouse monoclonal, Oncogene Research), CHK1 (rabbit polyclonal, Santa Cruz) and CHK2 (mouse monoclonal, BD-Pharmingen) or MRE11 (mouse monoclonal, AbCam). To verify the expression of the MSH2 protein, mouse monoclonal anti-MSH2 antibodies (BD-Transduction Laboratories) were used. Appropriate species-specific HRP-conjugate IgGs were used as secondary antibodies (Amersham). The blots were developed using the ECL detection system (Amersham). Equal loading and transfer were assessed reprobing the blots with anti-b-tubulin antibody (Santa Cruz) and by Ponceau-red staining of the blots.

Cell cultures and synchronization

Analysis of MRE11 and RAD51 foci

All the cell lines were cultured and handled as already described (Pichierri et al. , 2001). For biochemical analyses, cells were enriched in late-S/G2 prior to X-ray irradiation by a double thymidine block. Cells were exposed to 2 mm thymidine for 12 h. The cultures were then washed, allowed to recover in fresh medium for 6 h and then exposed to 2 mm thymidine for

Cells were cultured on coverslips, mock-treated (0 h) or exposed to X-rays (8 Gy) and sampled 1, 2, 4, 8 and 16 h afterwards. In order to analyse MRE11 foci in untreated or irradiated cells shortly after treatment, fixation in a 4% paraformaldehyde-buffered solution was carried out after an in situ fractionation procedure as described (Mirzoeva and Oncogene


2112 Petrini, 2001). Cells were then immediately processed, as previously described (Pichierri et al., 2001), for immunochemical detection of MRE11 or RAD51 using rabbit polyclonal anti-RAD51 or MRE11 antibodies (Oncogene Research) and Alexa546-conjugated mouse anti-rabbit secondary antibody (Molecular Probes). Alternatively, irradiated cultures were processed for simultaneous visualization of RAD51 foci and human proliferating cell nuclear antigen (PCNA) (Santa Cruz Biotechnology). Secondary antibodies (Molecular Probes) (1 h incubation at 371C) were selected to provide the appropriate combinations of species specificity (donkey anti-rabbit or mouse) and colour discrimination (conjugated to either fluorescein or Alexa546). Cells were counterstained with DAPI. For each experimental point, at least 200 nuclei were examined and RAD51 or MRE11 foci scored by eye at a magnification of 100 using a Zeiss epifluorescence microscope. Only nuclei showing >5 foci were considered as positive. Images were captured with a CCD camera and processed by using IPLab image analysis software. Parallel samples, incubated either with goat normal serum or with the secondary antibody, allowed one to discriminate possible artefacts. Recovery of DNA synthesis after X-ray irradiation Cell cultures were irradiated with X-rays (4 Gy) and analysed at different time points (1, 3, 6, and 8 h). To discriminate cells in active DNA synthesis, BrdUrd (30 mg/ml) was added 1 h before harvesting and BrdUrd incorporation in interphase nuclei was evaluated as previously described (Franchitto et al., 1998). To evaluate the percentage of labelled nuclei, at least 500 interphase cells were scored. Only nuclei displaying uniform BrdUrd labelling in the entire nuclear volume were considered as actively replicating. The percentage of cells undergoing DNA synthesis at each time point was calculated as the fraction of treated cells versus mock-treated controls. Chromosome preparations and analysis In order to analyse chromosomal damage induced by X-ray irradiation, cultures were treated with X-rays (2 Gy) and harvested between 3 and 6 h. To label cells in the S phase of the cell cycle, BrdUrd (30 mg/ml) was added to the cultures immediately after irradiation together with colcemid (0.2 mg/ ml) until harvesting. Metaphases were prepared according to standard procedures from colcemid-arrested cells and used for the immunodetection of the BrdUrd incorporation to determine the phase of the cell cycle at which they were treated (i.e. labelled, S phase cells; unlabelled, G2-phase cells) (Palitti et al., 1999). The frequency of chromosomal aberrations was scored in both BrdUrd-labelled and unlabelled metaphases. A total of 100 labelled or unlabelled cells were analysed in detail for chromatid-type aberrations according to the criteria described by Savage (1976).

Colo5#2 in which correction of the phenotype was achieved by transfer of a normal human chromosome #2 bearing the wild-type hMSH2 allele (Humbert et al., 1999). As shown in Figure 1, MSH2 expression was only detected in the MSH2-proficient cells (Colo26) but not in Colo5 or in the MSH2/ MEFs. Both MEFs and colorectal carcinoma cells defective in MSH2 were slightly more sensitive (B1.5-fold) to X-ray irradiation than their MSH2-proficient counterparts (Figure 2a, b). This sensitivity was most evident at the highest doses of X-rays (4 and 8 Gy). As expected, the revertant Colo5#2 cells showed cell survival closely similar to the parental Colo26 cells. These data indicate that, in mammalian cells, loss of MSH2 confers a slight hypersensitivity to X-rays, especially at higher dose levels. Correct MRE11 relocalization in the G2 phase of the cell cycle after IR requires the MSH2 protein Radiation sensitivity is often associated with DNA repair defects, mainly NHEJ or HR (Kanaar et al., 1998). MRE11 can associate with damaged DNA and is involved in its repair either by NHEJ or HR (Kanaar et al., 1998). This factor takes part in multiprotein complexes that can be visualized by immunofluorescence as discrete nuclear foci (Maser et al., 1997). We investigated whether loss of MSH2 could affect MRE11 focus formation. In order to study the early response of the MRE11 protein following DNA damage, we used an in situ fractionation protocol. Accordingly, three distinct localization patterns of MRE11 foci can be detected: pattern I (large aggregates) found in the majority of the unirradiated cells, pattern II (evenly distributed small foci) observed shortly after irradiation and pattern III (large patches) found at later times postirradiation (Mirzoeva and Petrini, 2001). Since the formation of pattern II coincides temporally with DSBs repair, as also

Results Loss of MSH2 confers hypersensitivity to IR We compared the effects of X-rays on the survival of primary MEFs derived from MSH2+/+ or MSH2/ mice. As a complementary cellular system, we used an MMR-proficient mouse colon tumour (Colo26), its MSH2-defective derivative Colo5 and the revertant Oncogene

Figure 1 Western blot analysis of MSH2 expression. Total proteins (20 mg) for each experimental point were loaded onto a 10% SDS–PAGE gel and subjected to Western/immunoblotting. Blots were probed with antibody raised against MSH2. The * indicates aspecific bands


2113

Figure 2 Reduced cell survival of MSH2-deficient cells following IR. MEFs from MSH2 þ / þ and MSH2/ mice (a) or MSH2proficient (Colo26, Colo5#2) and MSH2-deficient (Colo5) mouse colorectal carcinoma cells (b) were exposed to increasing doses of X-rays and analysed for clonogenic survival. Each symbol represents the mean of triplicate independent experiments; the s.e.m. is shown by error bars In some cases, the error bars are smaller than the symbol

demonstrated by colocalization with the H2AX histone (Mirzoeva and Petrini, 2001; Paull et al., 2000; Rogakou et al., 1998), we focused our analysis on this specific MRE11 localization. Both MSH2+/+ and MSHZ/ untreated cells displayed similar numbers of nuclei bearing type-I and -II foci. (Figure 3a, b and our unpublished observation). After 8 Gy the percentage of cells containing MRE11 type-II foci increased in MSH2+/+ cells in a time-dependent manner, whereas in MSHZ/ cells the percentage of these foci did not change significantly up to 8 h postirradiation (F3a, b). In addition, the MRE11 type-II foci that do form in the absence of MSH2 are always associated with PCNA staining (our unpublished observation). The differences in the percentage of nuclei containing foci between the two genotypes cannot be ascribed to changes in the total amount of MRE11 protein, as indicated by Western blotting analysis (Figure 3c). Moreover, it is unlikely that the enhanced apoptotic response of MSH2/ compared to MSH2+/+ MEFs (Figure 3d) is responsible for the altered MRE11 focal localization of these cells. Apoptosis after irradiation is observed at much later times than MRE11 relocalization (16 versus 1–4 h). In addition, the similar timing in the induction of apotosis in the two cell lines, even by staining with the early apoptotic marker Annexin V (Boersma et al., 1996) 3d and data not shown), is not consistent with a selective depletion of the MRE11-positive population of MSH2/ cells (Boersma et al., 1996). Given that the MRE11/RAD50/NBS1 complex is involved in the activation of the S-phase checkpoint after IR (Carney et al., 1998), we evaluated the possibility that IR differently affects the progression of cells through the S phase in MSH2+/+ and MSH2–/– cells. A dramatic decrease in cells replicating their DNA was observed between 3 and 6 h after irradiation but no significant difference between MSH2+/+ and MSH2–/– MEFs was observed in the inhibition of DNA synthesis (Figure 4). The irradiated cells also resumed DNA synthesis in a similar manner, further indicating that the

progression of damaged cells through the S phase is not notably affected by loss of MSH2. These data suggest that the MSH2 protein is required for efficient relocalization of MRE11 protein after IR treatment of G2-phase cells, whereas it is dispensable for a correct activation of the S-phase checkpoint. RAD51 focus formation in the G2 phase of the cell cycle is affected by the absence of MSH2 protein A defect in MRE11 activity could theoretically impair both NHEJ and HR. However, a role for MSH2 in DSB repair has been envisaged only in HR. Thus, to investigate whether the observed incorrect MRE11 relocalization in MSH2-defective cells could be related to DSB repair by HR, we analysed subnuclear relocalization of the RAD51 recombinase, a key enzyme in this process. Compared to wild-type MEFs, untreated MSH2/ cells showed a significantly higher spontaneous RAD51 focus-forming activity (Figure 5a). After 8 Gy of IR exposure, the number of nuclei showing RAD51 foci increased over time and reached the highest values 8 h after treatment in both MSH2+/+ and MSH2/ cells. However, focus formation was somewhat reduced in MSH2/ cells at 4 and 8 h postirradiation (Figure 5a). Given that HR is active in late-S and G2 phases of the cell cycle, experiments were undertaken to determine whether X-rays induce RAD51 foci in both phases. S-phase cells were identified through PCNA immunofluorescence under fixation conditions that, solubilizing unbound PCNA, visualize only the fraction bound to the replication complexes. In wild-type cells, at 4 and 8 h postirradiation, about two-thirds of the nuclei showing RAD51 foci were also PCNA-positive, while the remaining were PCNA-negative. MSH2/ cells showed as many RAD51 foci as wild-type MEFs in PCNApositive nuclei but, in sharp contrast, no foci were detected in MSH2/ cells in PCNA-negative nuclei (F5b, c). As already described by others, the RAD51 Oncogene


2114

Figure 3 Formation of type-II MRE11 foci following IR. MSH2+/+and MSH2/ MEFs were mock-treated (0 h) or exposed to 8 Gy of X-rays and subjected to immunofluorescent staining of MRE11 at different recovery times. The data are the mean7s.e.m. of three independent experiments. (a). MRE11 re-localization in MSH2+/+and MSH2/ cells (b). In untreated cells, type-I MRE11 foci are predominant in both MSH2+/+and MSH2/ cells (indicated by arrowhead in (a)). Following IR, MSH2-proficient but not MSH2deficient cells show type-II MRE11 foci (indicated by the arrowhead in (c)). Analysis of the MRE11 levels shows no difference in protein expression between MSH2+/+and MSH2/ cells (c). Total proteins (20 mg) for each experimental point from mock-treated or 8 Gy-irradiated MSH2+/+and MSH2/ MEFs were loaded onto a 7% SDS–PAGE gel and subjected to Westem/immunoblotting analysis. Blots were probed with the MRE11 antibody. Equal loading and transfer were assessed by ponceau-red staining of the blots. Enhanced apoptotic cell death after IR treatment in MSH2/ MEFs (d). The horizontal reference line represents the mean value of the spontaneous level of apoptosis in the two cell lines. The data are the mean from three independent experiments. Standard errors are not shown but were always o10% of the respective mean value

and MRE11 proteins did not colocalize either in G2- or in S-phase cells after irradiation (Mirzoeva and Petrini, 2001), and the absence of the MSH2 protein did not alter this pattern (data not shown). Thus, as S-phaseassociated RAD51 foci are similar in both cell lines studied, the observed reduction in RAD51 foci in MSH2/ cells is to be ascribed solely to the absence of foci in the G2 phase. Loss of MSH2 is associated with increased levels of chromatid and isochromatid breaks in the G2 phase of the cell cycle following irradiation To verify whether the defects in the response to IRinduced DNA damage in the G2 phase of the cell cycle observed in MSH2-defective cells is accompanied by a higher level of chromosomal damage, MEFs were exposed to 2 Gy of X-rays, a dose that corresponds to Oncogene

B80% of survival in both genotypes (Figure 2). Since HR is thought to take place in late-S or G2 phases of the cell cycle, cells were pulse-labelled with BrdUrd immediately after irradiation, and chromosomal damage was analysed in both BrdUrd-negative (unlabelled, G2-phase cells) and -positive (labelled, S-phase cells) metaphases (Tables 1 and 2, respectively). Because MSH2/ cells did not correctly arrest in G2-phase (Figure 6; see below), chromosomal damage in MSH2+/+ and MSH2/ cells was compared in populations showing similar percentages of labelled mitotic cells (i.e. treated with colcemid for different lengths of time; Palitti et al., 1999). We found a statistically significant increase in the frequency of chromosomal damage in the MSH2/ MEFs, in comparison with wild-type cells. In particular, we observed a sharp enhancement (Bthree-fold) of isochromatid breaks and a more moderate increase in


2115

Figure 4 Analysis of radioresistant DNA synthesis in MSH2proficient and -deficient cell lines. Untreated or X-irradiated (4Gy) MSH2+/+ (closed circles) and MSH2/ (open triangles) MEFs were harvested at the indicated recovery times. Cells were processed for the immunodetection of BrdUrd incorporation as described in ‘Materials and methods’. The data are the mean7s.e.m. of three independent experiments

chromatid breaks (B1.5-fold) after irradiation (Table 1). Evaluation of the chromosomal damage in MSH2+/+ and MSH2/ cells at a second pair of sampling times (6 and 5 h postirradiation, respectively) showed no further increase in the number of chromatid breaks, while the larger induction of isochromatid breaks in MSH2/ cells was conserved (data not shown). Similarly, upon X-irradiation lack of MSH2 protein in the tumour cell line Colo5 was associated with an increased number of isochromatid breaks in comparison with the parental, MSH2-expressing Colo26. Meaningfully, expression of hMSH2 by introduction of the human chromosome 2 in Colo5 cells reduced the frequency of breaks to that observed in Colo26 cells. These data strongly indicate that the higher level of chromosomal damage observed in MSH2-defective cells is because of loss of the MSH2 function. Similar chromosomal damage analyses were performed in cells treated in the late S-phase of the cell cycle (Table 2). As in the previous experiments, we compared populations containing similar numbers of labelled cells. No difference in the frequency of chromatid or isochromatid breaks was found between MSH2+/+ and MSH2/ cells, despite a slight increase

Figure 5 Reduced induction of RAD51 foci in MSH2-deficient cells following IR. Immunofluorescence of RAD51 and PCNA after IR. (a) MEFs were seeded onto coverslips, irradiated with 8 Gy and fixed at different postirradiation times. The data are the mean7s.e.m. of three independent experiments. The percentage of RAD51-positive nuclei showing PCNA staining is plotted versus recovery time (C). Dark bars: RAD51- and PCNA-positive cells; light bars: RAD51-positive and PCNA-negative cells. The data are representative of three different experiments showing similar results. Standard errors are not shown but were always o10% of the respective mean value Oncogene


2116 Table 1 Analysis of the chromosomal damage induced by X-rays in the G2-phase of the cell cycle in MSH2-deficient and proficient unlabelled cells Aberrations per 100 cells Cell line

Genotype

Treatment

Abnormal cells (%)

Labelled cells (%)

Chromatid breaks

MEFs

MSH2+/+

Control 2 Gy+5 h colcemid

5 80

NA 15

5 205

MEFs

MSH2/


7 90

NA 11

Colo26

MSH2+/+


4 63

Colo5

MSH2/


Colo5#2

MSH2+/+


Isochromatid breaks

Chromatid exchanges

Total aberrations (-gaps)

Aberr.ns/cell (mean7s.e)

0 14

0 24

5 243

0.0570.02 2.4370.15

6 285

1 41**

1 10

8 336

0.0870.03 3.3670.19**

NA 10

4 198

0 10

0 13

4 231

0.04+0.02 2.3170.11

4 70

NA 13

2 248

2 31**

0 8

4 287

0.0470.02 2.8770.16*

5 68

NA 8

4 201

0 12

1 8

5 221

0.0570.04 2.2170.12

**Statistically significant Po0.01 (t-test), compared to the wild type *Statistically significant Po0.05 (t-test), compared to the wild type

Table 2

Analysis of the chromosomal damage induced by X-rays in the late-S phase of the cell cycle in MSH2-deficient and proficient labelled cells Aberrations per 100 cellsa

Cell line

Genotype

Treatment

Abnormal cells (%)

Labelled cells (%)

Chromatid breaks

MEFs

MSH2+/+

2 Gy+6 h colcemid Control

6 43

NA 32

8.4 127

MEFs

MSH2/

2 Gy+5 h colcemid Control

9 55

NA 39

10 140

Isochromatid breaks

Chromatid exchanges

Total aberrations (-gaps)

Aberrations/cell (mean7s.e.)

0 2

0 15.6

8.4 144.6

0.0870.01 1.4470.10

0 4.2

2 35.6*

12 187

0.1270.03 1.7970.12

a A total of 500 labelled metaphase cells were scored for each experimental point. *Statistically significant Po0.05 (t-test), compared to the wild-type NA, not applicable; spontaneous chromosomal damage was scored in either labelled or unlabelled cells without any significant difference so the data are referred to 500 metaphase cells either unlabelled and labelled

Figure 6 MSH2-deficient cells arrest only transiently after irradiation in the G2 phase of the cell cycle. MSH2-proficient or -defective cell lines were treated with 2 Gy (a) or 4 Gy (b) of X-rays and recovered for the indicated times. (c) Arrest in the G2 phase of the cell cycle is unaffected in MSH2/ cells following UV radiation. MEFs from MSH2+/+or MSH2/ mice were either mock-treated or exposed to UVC radiation (20 J/m2) and recovered for the indicated times. Mitotic indices were calculated as described in ‘Materials and methods’; the data are the mean7s.e.m. of three independent experiments Oncogene


2117

(Btwofold) in the number of chromatid exchanges obtained in MSH2-defective cells. In conclusion, these data indicate that loss of MSH2 function is associated with increased levels of X-rayinduced isochromatid and chromatid breaks specifically during the G2 phase of the cell cycle. MSH2 is required to sustain G2 arrest induced by IR We observed defective relocalization of the two DSB repair enzymes MRE11 and RAD51 specifically in the G2 phase of the cell cycle. DNA repair and checkpoint response are two interlinked processes (Lowndes and Murguia, 2000), and loss of certain MMR proteins has been correlated to defective G2 accumulation (Davis et al., 1998; Yan et al., 2001). Thus, we investigated whether the defective MRE11 and RAD51 relocalization observed in MSH2/ cells is because of defects in triggering the G2 checkpoint in cells irradiated in the G2 phase. Recently, Xu et al. (2002), demonstrated that the G2 accumulation observed by flow cytometry at late times after irradiation is a pathway completely different from the early G2/M checkpoint response elicited in G2phase treated cells (Xu et al., 2002). For instance, the early G2/M checkpoint response is ATM- and BRCA1dependent, whereas the late G2 accumulation is not (Xu et al., 2001a, 2002). Since we observe defective MRE11 and RAD51 relocalization shortly after treatment (o10 h), we hypothesized a defect in the early rather than in the late G2/M response (Yan et al., 2001). Thus, we analysed the ability of MSH2/ cells irradiated in G2 to progress into mitosis by assessment of the mitotic index. Irradiation with 2 and 4 Gy induced an efficient G2 arrest in all the MSH2-proficient cell lines, as shown by the reduction in the mitotic index shortly after irradiation (F6a, b). The higher dose induced a more prolonged inhibition of cell cycle progression. A similar reduction in the mitotic index was observed in MSH2deficient cells only 1 h after treatment (F6a, b). At later times, however, the observed G2 arrest was released in MSH2-deficient cells. Thus, our data partially agree with those of Yan et al. (2001) on the G2 accumulation of MMR-defective cells, suggesting that in the absence of MSH2, the early G2/M checkpoint is properly activated but prematurely released. The behaviour of MSH2-deficient cells appeared to be specifically attributable to loss of the gene, since the revertant Colo5#2 cells showed G2 arrest duration similar to that observed in the parental cells (F6a, b). Conversely, UV treatment induced an efficient G2 arrest, independent of the presence of a functional MSH2 gene (Figure 6c), suggesting that the role of MSH2 in the early G2/M checkpoint depends on the type of DNA damage. Consistently with the cytological data, in late-S/G2 enriched MSH2+/+ cells CDC2 was inhibited after 4 Gy, as indicated by the presence of its phosphorylated, slowly migrating isoform and by the disappearance of the unphosphorylated form at 90 min postirradiation (Figure 7b). Conversely, MSH2/ cells did not show any apparent difference in the amount of the CDC2

Figure 7 Failure of maintenance of CHK1 phosphorylation and absence of CHK2 activation in MSH2-deficient cells following IR. (a) Analysis of cell cycle distribution of MSH2+/+and MSH2/ MEFs in asynchronous cultures (a, b), immediately after release from the double thymidine block (c, d) and after 4 h from the release when the majority of cells (B80%) were in late-S/G2 (e, f). (b) Synchronized cells were either mock-treated or exposed to 4 Gy of X-rays 4h after the release from the double thymidine block, when in late-S/G2. Blots were probed with antibody raised against CDC2 and sequentially reprobed with Tubulin antibodies. (c) Absence of MSH2 leads to uncorrect CHK1 and CHK2 phosphorylation after IR. -P: phosphorylated isoform. Equal loading and transfer were assessed by ponceau-red staining of the blots. The results shown are representative of those observed in three independent experiments

unphosphorylated isoform between mock-treated and irradiated samples. Maintenance of the CDC2 inhibitory phosphorylation is dependent on the two checkpoint kinases CHK1 and hCDS1/CHK2 (Fumari et al., 1999; Matsuoka et al., 1998; Sanchez et al., 1997). It has been reported that CHK1 initiates the G2 arrest following IR, whereas CHK2 is essential to maintain this arrest (Hirao et al., 2000). Thus, we investigated whether CHK1 and CHK2 were properly activated in MSH2-deficient cells. As expected, IR induced the rapid (30 min) phosphorylation of both CHK1 and CHK2 in MSH2-proficient cells, whereas a transient activation of CHK1 and a complete Oncogene


2118

lack of CHK2 phosphorylation were observed in MSH2-deficient cells (Figure 7c). These results are consistent with the ability of MSH2deficient cells to arrest, but not to maintain, the G2 block after IR and suggest that, in the absence of MSH2, the G2 checkpoint proteins do not properly sustain the DNA damage signal. Discussion We observed that MSH2-defective mouse cells are more radiation sensitive than their wild-type counterparts. An elevated radiosensitivity can be the consequence of either an impaired checkpoint response or a defect in repair of IR-induced DNA damage. IR induces different types of DNA lesions, among these DSBs, which are repaired through the NHEJ or the HR pathways (Kanaar et al., 1998). In vertebrate cells, each of these pathways operates preferentially in a specific phase of the cell cycle: NHEJ in G1/early-S and HR in late-S/G2 (Takata et al., 1998). We found that two of the key proteins in HR, MRE11 and RAD51, were not correctly relocalized after DNA damage in the absence of MSH2, consistently with a role for this mammalian MMR protein in recombination (Alani et al., 1994; Chen and Jinks-Robertson, 1998, 1999; Datta et al., 1996; Elliott and Jasin, 2001; Evans et al., 2000). This defective relocalization was observed in the G2 phase but not in the S phase of the cell cycle, suggesting that the role for MSH2 in the relocalization of MRE11 and RAD51 is somewhat confined to the G2 phase of the cell cycle. Moreover, loss of MSH2 function results in a higher yield of chromosomal damage in the G2 phase of the cell cycle, in particular isochromatid breaks. Specific types of chromosomal damage are associated with inactivation of different HR genes. Inactivation of RAD54 or RAD51 in chicken DT40 cells increases their susceptibility to radiationinduced chromatid breaks, because of a failure of the earlier stages of HR (Sonoda et al., 1998; Takata et al.,

1998). In contrast, MRE11 mutations are associated with elevated induction of isochromatid breaks, possibly deriving from incomplete repair after joint-molecule formation (Yamaguchi-lwai et al., 1999). Thus, our observation of a higher frequency of X-ray-induced isochromatid breaks in MSH2-deficient cells actually suggests that HR starts, but cannot proceed to completion, as in MRE11-null cells. Since MRE11 performs an important role in DNA damage signalling (D’Amours and Jackson, 2002 and references therein), it is possible that in the absence of MSH2 the cellular response to DNA damage is not properly activated and HR does not correctly take place, forcing cells to use NHEJ to repair DSBs. Consistently, it has been reported that NHEJ could serve as backup system for DSB repair when HR is impaired (Fukushima et al., 2001). Such a failure to correctly perform HR-mediated DSB repair in the absence of MSH2 could also correlate with the observed inability to maintain the G2 arrest. Indeed, we found that MSH2-deficient cells arrested correctly at the early G2 checkpoint after IR, but were unable to maintain the arrest and re-entered the cell cycle within 1 h after treatment. Consistent with premature release from the early G2/M checkpoint, both the checkpoint kinases CHK1 and CHK2 were not properly activated in MSH2/ cells, even though CHK1 phosphorylation was detected, as in wild-type cells, immediately after irradiation. CHK1 and CHK2 kinases are both required to activate the G2 checkpoint after IR (Matsuoka et al., 1998; Sanchez et al., 1997). CHK1 appears to be essential in triggering the checkpoint, whereas CHK2 is needed to maintain the arrest (Hirao et al., 2000). Our observations suggest that MSH2 is not necessary to impose this arrested state, arguing against a direct role in the G2 checkpoint, but is required to maintain this arrested state, possibly acting somewhat indirectly in later steps of the G2 checkpoint activation. We did not observe such behaviour following UV irradiation. This suggests that MSH2 plays a role in response to DNA lesions induced by IR but not UV irradiation.

Figure 8 Model for the role of MSH2 in DSB repair by HR in the G2-phase of the cell cycle. Following DNA damage, CHK1 is phosphorylated in both MSH2-proficient and -deficient cells. Cells are arrested at the G2/M border to repair DNA damage. IRinduced base damage, which accumulates in MSH2/ cells (Colussi et al., 2002; DeWeese et al., 1998), could interfere with HR possibly because of increased levels of DNA mismatches. CHK2 is activated in MSH2-proficient cells to sustain G2 arrest, allowing time for mismatch removal, and MSH2 might be required to recognize mismatches and recruit MRE11. Consistent with this possibility, a colocalization of MSH2 and MRE11 has been recently reported (Zink et al., 2002). In the absence of MSH2, MRE11 is not properly relocalized and CHK2 is not activated, leading to premature release of the G2 arrest and elevated yield of isochromatid breaks and chromatid breaks Oncogene


2119

Based on our observations and data reported in the literature, we propose a model to explain the incorrect MRE11 and RAD51 relocalization, the elevated IRinduced chromosomal damage and the premature release from the G2 checkpoint, associated with loss of MSH2. According to this model, isochromatid and chromatid breaks result from the inability of MSH2defective cells to perform an efficient DSB repair via HR (Figure 8). In conclusion, our data support the idea that MSH2 is required both for successful DSB repair in the G2 phase of the cell cycle and for the maintenance of the early-G2 arrest, but not to sense IR-induced DNA damage and

initiate the early phases of HR. Abrogation of these two functions in MSH2/ cells might be responsible for their increased sensitivity towards high IR doses and susceptibility to chromosomal damage.

Acknowledgements We thank Dr P Karran for a critical reading of the manuscript and discussion. The skilful collaboration of Mr Angelo Schinoppi is grateful acknowledged. This work was partially supported by MURST grants and by grants from EC (contract no. FIGH-CT1999-00011).

References Aebi S, Kurdi-Haidar B, Gordon R, Cenni B, Zheng H, Fink D, Christen RD, Boland CR, Koi M, Fishel R and Howell SB. (1996). Cancer Res., 56, 3087–3090. Alani E, Reenan RA and Kolodner RD. (1994). Genetics, 137, 19–39. Anthoney DA, Mcllwrath AJ, Gallagher WM, Edlin AR and Brown R. (1996). Cancer Res., 56, 1374–1381. Aquilina G, Crescenzi M and Bignami M. (1999). Carcinogenesis, 20, 2317–2326. Baumann P and West SC. (1998). Trends Biochem. Sci., 23, 247–251. Bellacosa A. (2001). Cell Death Differ., 8, 1076–1092. Berry SE and Kinsella TJ. (2001). Semin. Radiat. Oncol., 11, 300–315. Boersma AW, Nooter K, Oostrum RG and Stoter G. (1996). Cytometry, 24, 123–130. Bradley MO and Kohn KW. (1979). Nucleic Acids Res., 7, 793–804 Carney JP, Maser RS, Olivares H, Davis EM, Le Beau M, Yates III JR, Hays L, Morgan WF and Petrini JH. (1998). Cell, 93, 477–486. Chen W and Jinks-Robertson S. (1998). Mol. Cell. Biol., 18, 6525–6537. Chen W and Jinks-Robertson S. (1999). Genetics, 151, 1299– 1313. Ciotta C, Ceccotti S, Aquilina G, Humbert O, Palombo F, Jiricny J and Bignami M. (1998). J. Mol. Biol., 276, 705–719. Colussi C, Parlanti E, Degan P, Aquilina G, Bames D, Macpherson P, Karran P, Crescenzi M, Dogliotti E and Bignami M. (2002). Curr. Biol., 12, 912–918. D’Amours D and Jackson SP. (2002). Nat. Rev. Mol. Cell Biol., 3, 317–327. Datta A, Adjiri A, New L, Crouse GF and Jinks-Robertson S. (1996). Mol. Cell. Biol., 16, 1085–1093. Davis TW, Wilson-Van Patten C, Meyers M, Kunugi KA, Cuthill S, Reznikoff C, Garces C, Boland CR, Kinsella TJ, Fishel R and Boothman DA. (1998). Cancer Res., 58, 767– 778. DeWeese TL, Shipman JM, Larrier NA, Buckley NM, Kidd LR, Groopman JD, Cutler RG, te Riele H and Nelson WG. (1998). Proc. Natl. Acad. Sci. USA, 95, 11915–11920. De Wind N, Dekker M, Berns A, RaDNAn M and te Riele H. (1995). Cell, 82, 321–330. Elliott B and Jasin M. (2001). Mol. Cell. Biol., 21, 2671–2682. Evans E, Sugawara N, Haber JE and Alani E. (2000). Mol. Cell, 5, 789–799.

Fiumicino S, Martinelli S, Colussi C, Aquilina G, Leonetti C, Crescenzi M and Bignami M. (2000). Int. J. Cancer, 85, 590– 596. Franchitto A, Pichierri P, Mosesso P and Palitti F. (1998). Mutagenesis, 13, 499–505. Fritzell JA, Narayanan L, Baker SM, Bronner CE, Andrew SE, Prolla TA, Bradley A, Jirik FR, Liskay RM and Glazer PM. (1997). Cancer Res., 57, 5143–5147. Fukushima T, Takata M, Morrison C, Araki R, Fujimori A, Abe M, Tatsumi K, Jasin M, Dhar PK, Sonoda E, Chiba T and Takeda S. (2001). J. Biol. Chem., 276, 44413–44418. Furnari B, Blasina A, Boddy MN, McGowan CH and Russell P. (1999). Mol. Biol. Cell, 10, 833–845. Haber J.E. (1998). Cell, 95, 583–586. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, Yoshida H, Liu D, Elledge SJ and Mak TW. (2000). Science, 287, 1824–1827. Humbert O, Fiumicino S, Aquilina G, Branch P, Oda S, Zijno A, Karran P and Bignami M. (1999). Carcinogenesis, 20, 205–214. Hutchinson F. (1985). Prog. Nucleic Acid Res. Mol. Biol., 32, 115–154. Jiricny J. (1998). Mutat. Res., 409, 107–121. Jiricny J and Nystrom-Lahti M. (2000). Curr. Opin. Genet. Dev., 10, 157–161. Kaina B, Ochs K, Grosch S, Fritz G, Lips J, Tomicic M, Dunkern T and Christmann M. (2001). Prog. Nucleic. Acid Res. Mol. Biol., 68, 41–54. Kanaar R, Hoeijmakers JH and van Gent DC. (1998). Trends Cell. Biol., 8, 483–489. Karran P and Bignami M. (1994). Bioessays, 16, 833–839. Kolodner R. (1996). Genes Dev., 10, 1433–1442. Kolodner RD and Marsischky GT. (1999). Curr. Opin. Genet. Dev., 9, 89–96. Liang F, Han M, Romanienko PJ and Jasin M. (1998). Proc. Natl. Acad. Sci. USA, 95, 5172–5177. Lowndes NF and Murguia JR. (2000). Curr. Opin. Genet. Dev., 10, 17–25. Maser RS, Monsen KJ, Nelms BE and Petrini JH. (1997). Mol. Cell. Biol., 17, 6087–6096. Matsuoka S, Huang M and Elledge SJ. (1998). Science, 282, 1893–1897. Mirzoeva OK and Petrini JH. (2001). Mol. Cell. Biol., 21, 281–288. Modrich P and Lahue R. (1996). Annu. Rev. Biochem., 65, 101–133.

Oncogene


2120 Palitti F, Pichierri P, Franchitto A, Proietti De Santis L and Mosesso P. (1999). Int. J. Radiat. Biol., 75, 621–627. Paull TT and Gellert M. (1998). Mol. Cell, 1, 969–979. Paull TT, Rogakou EP, Yamazaki V, Kirchgessner CU, Gellert M and Bonner WM. (2000). Curr. Biol., 10, 886–895. Pichierri P, Franchitto A, Piergentili R, Colussi C and Palitti F. (2001). Carcinogenesis, 22, 1781–1787. Rogakou EP, Pilch DR, Orr AH, Ivanova VS and Bonner WM. (1998). J. Biol. Chem., 273, 5858–5868. Sanchez Y, Wong C, Thoma RS, Richman R, Wu Z, PiwnicaWorms H and Elledge SJ. (1997). Science, 277, 1497–1501. Savage JR. (1976). J. Med. Genet., 13, 103–122. Sonoda E, Sasaki MS, Buerstedde JM, Bezzubova O, Shinohara A, Ogawa H, Takata M, Yamaguchi-lwai Y and Takeda S. (1998). Embo J., 17, 598–608. Sugawara N, Paques F, Colaiacovo M and Haber JE. (1997). Proc. NatI. Acad. Sci. USA, 94, 9214–9219. Takata M, Sasaki MS, Sonoda E, Morrison C, Hashimoto M, Utsumi H, Yamaguchi-lwai Y, Shinohara A and Takeda S. (1998). EMBO J., 17, 5497–5508. Trujillo KM, Yuan SS, Lee EY and Sung P. (1998). J. Biol. Chem., 273, 21447–21450.

Oncogene

White Cl and Haber JE. (1990). EMBO J., 9, 663–673. Worth Jr L, Clark S, RaDNAn M and Modrich P. (1994). Proc. NatI. Acad. Sci. USA, 91, 3238–3241. Xu B, Kim S and Kastan MB. (2001a). Mol. Cell. Biol., 21, 3445–3450. Xu B, Kim ST, Lim DS and Kastan MB. (2002). Mol. Cell. Biol., 22, 1049–1059. Xu XS, Narayanan L, Dunklee B, Liskay RM and Glazer PM. (2001b). Cancer Res., 61, 3775–3780. Yamaguchi-lwai Y, Sonoda E, Sasaki MS, Morrison C, Haraguchi T, Hiraoka Y, Yamashita YM, Yagi T, Takata M, Price C, Kakazu N and Takeda S. (1999). EMBO J., 18, 6619–6629. Yan T, Schupp JE, Hwang H-S, Wagner MW, Berry SE, Strickfaden S, Veigl ML, Sedwick WD, Boothman DA and Kinsella TJ. (2001). Cancer Res., 61, 8290–8297. Zeng M, Narayanan L, Xu XS, Prolla TA, Liskay RM and Glazer PM. (2000). Cancer Res., 60, 4889–4893. Zhang H, Marra G, Jiricny J, Maher VM and McCormick JJ. (2000). Carcinogenesis, 21, 1639–1646. Zink D, Mayr C, Janz C and Wiesmuller L. (2002). Oncogene, 21, 4788–4800.