Human DNA helicase HELQ participates in DNA interstrand crosslink tolerance with ATR and RAD51 paralogs

Mammalian HELQ is a 3′–5′ DNA helicase with strand displacement activity. Here we show that HELQ participates in a pathway of resistance to DNA interstrand crosslinks (ICLs). Genetic disruption of HELQ in human cells enhances cellular sensitivity and chromosome radial formation by the ICL-inducing agent mitomycin C (MMC). A significant fraction of MMC sensitivity is independent of the Fanconi anaemia pathway. Sister chromatid exchange frequency and sensitivity to UV radiation or topoisomerase inhibitors is unaltered. Proteomic analysis reveals that HELQ is associated with the RAD51 paralogs RAD51B/C/D and XRCC2, and with the DNA damage-responsive kinase ATR. After treatment with MMC, reduced phosphorylation of the ATR substrate CHK1 occurs in HELQ-knockout cells, and accumulation of G2/M cells is reduced. The results indicate that HELQ operates in an arm of DNA repair and signalling in response to ICL. Further, the association with RAD51 paralogs suggests HELQ as a candidate ovarian cancer gene.

D NA interstrand crosslinks (ICLs) are particularly toxic because they disrupt genetic information on both strands, potently inhibiting DNA replication and transcription. DNA ICLs are produced by many environmental agents including photoactivatable furocoumarins present in some plants, aldehydes from industrial sources and endogenous agents, including products of lipid peroxidation and nitric oxide 1,2 . Compounds that form ICLs are widely used in cancer chemotherapy and include platinum-based compounds, mitomycin C (MMC), psoralens, nitrogen mustards and derivatives (melphalan and chlorambucil) and nitrosoureas such as BCNU 1,2 . Multiple pathways are implicated in various regulated steps of ICL repair, including many proteins associated with pathways of homologous recombination, the Fanconi anaemia (FA) network, mismatch repair, nucleotide excision repair and translesion DNA synthesis [3][4][5][6][7][8] .
The RAD51 paralogs (RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3) are a group of related proteins with distant similarity to the RAD51 protein, which mediates homologous DNA pairing during recombination. Inactivation of any RAD51 paralog confers substantial cellular sensitivity to agents causing ICL 9,10 . However, it has been difficult to obtain mechanistic insights into how the RAD51 paralogs operate in ICL repair 10,11 . Models for the role of the RAD51 paralogs are diverse and range from loading of RAD51 on DNA 5,12 to post synaptic functions in RAD51 filament dissociation 13,14 , to direct involvement in Holliday junction resolution 15 . In fact, there may be several functions for the paralogs. Furthermore, protein complexes of subsets of the RAD51 paralogs exist 10,11 , but specific functions of these subcomplexes remain unclear.
Helicase-like enzymatic activities are predicted to be useful for several steps in ICL repair. Two of the FA proteins, BRIP1 (also known as FANCJ) and FANCM contain conserved helicase motifs and can translocate on DNA in an ATP-dependent fashion 5 . However, specific roles for these and other helicases in ICL repair have yet to be identified.
The HELQ gene, originally designated HEL308 (ref. 16), is a candidate for participation in ICL repair. HELQ is a DNA helicase that translocates in the 3 0 -5 0 direction and can displace oligonucleotides of 70 nt or more from DNA 16 . The homologue from Solfolubus solfataricus is powerful enough to readily overcome a streptavidin-biotin interaction 17 . On the basis of model substrates in vitro, it has been proposed that HELQ opens up the parental strands at a blocked DNA replication fork and remodels nascent lagging strand intermediates to facilitate the loading of subsequent factors required for DNA damage processing or restart of DNA replication 18 . However, the in vivo function of mammalian HELQ remains to be explored.
Here we report that genetic disruption of HELQ enhances the sensitivity of human cells to ICL-inducing agents, including MMC. The enhanced sensitivity is independent of the FA pathway. Sister chromatid exchange (SCE) frequency and sensitivity to UV radiation or topoisomerase inhibitors is unaltered. Analysis of interacting protein partners revealed that HELQ is associated with the RAD51 paralogs RAD51B/C/D and XRCC2, and with the DNA damage-responsive kinase ATR. Reduced phosphorylation of the ATR substrate CHK1 occurs in HELQ-knockout cells treated with MMC, and accumulation of G 2 /M cells is reduced. The results indicate that HELQ operates in DNA repair and cell cycle checkpoint signalling in response to ICL. The association with RAD51 paralogs and the frequent copy number loss of HELQ in ovarian cancers suggests HELQ as a candidate ovarian cancer predisposition gene.

Results
Inactivation of HELQ sensitizes human cells to ICL damage. To explore the biological function of HELQ, the human gene was disrupted in the U2OS osteosarcoma cell line by targeting the first exon with specific zinc finger nucleases (ZFNs). A heterozygous knockout (HELQ þ / À ) cell line (10H3) was isolated, and two knockout (HELQ À / À ) cell lines (5G6 and 2G9) were obtained by retargeting 10H3 with the same ZFNs. The ZFN disruptions introduced frameshift mutations early in the open reading frame (Fig. 1).
HELQ-deficient cells were examined for sensitivity to treatment with DNA-damaging agents in comparison with wild-type and heterozygous cells. The HELQ-deficient U2OS cells were sensitive to the ICL-inducing agents MMC and cisplatin ( Fig. 2a and Supplementary Fig. S1A), but were not consistently sensitized to bleomycin, etoposide, camptothecin, temozolomide, hydroxyurea, olaparib or UV irradiation ( Fig. 2b and Supplementary Fig. S1B). This suggests a specific function for HELQ in the processing of ICL. Further, MMC-induced radial chromosome formation was moderately increased in HELQ À / À cells ( Supplementary Fig. S1C). Radial chromosome formation after treatment with ICL-inducing agents is a cytological hallmark of FA cells with defects in ICL processing 19 .
HELQ and FA pathway ICL sensitivity are additive. To test whether the function of HELQ in ICL processing is dependent on the FA pathway, we used short interfering RNA (siRNA) knockdown of FANCD2 gene expression in wild-type and HELQ À / À cells. FANCD2-knockdown cells were more sensitive than HELQ À / À cells. A knockdown of FANCD2 in HELQ À / À cells resulted in even higher sensitivity to MMC (Fig. 3a). HELQ therefore appears to have a role in ICL processing that is separate from the FA pathway. Consistent with this notion, the canonical FANCD2 monoubiquitination response to MMC treatment was still observed in HELQ À / À cells (Fig. 3b).

HELQ interacts with the BCDX2 complex and ATR.
To determine molecular pathways relevant to HELQ, we searched for associated proteins in vivo. HELQ was stably expressed as a Flag and HA epitope fusion (eHELQ) in HeLa S3 cells. eHELQ was purified from the nuclear extract by immunoprecipitation sequentially with anti-Flag and anti-HA antibodies 20,21 (Fig. 4a). The HELQ complex was separated on a gradient gel and proteins from gel sections were identified by liquid chromatography-mass spectrometry. With high-ranking confidence, we identified components of the 'BCDX2 complex' consisting of RAD51B, RAD51C, RAD51D and XRCC2. The DNA damage-responsive kinase ATR and its partner ATRIP were also identified as binding partners (Supplementary Table S1). In a control Flag-HA purification from HeLa S3 cells transfected with empty control vector, neither BCDX2 nor ATR was present. Further, we performed parallel experiments in the same system using Flag-HA-tagged POLN 22,23 as bait, and detected neither RAD51 paralogs nor ATR.
There are five RAD51 paralogs that form two distinct subcomplexes in mammalian cells, denoted as BCDX2 and CX3 (composed of RAD51C and XRCC3) 9,24 . All members of the BCDX2 subcomplex were detected, but no XRCC3 was identified in HELQ coimmunoprecipitates, indicating the absence of CX3. The presence of these proteins in the HELQ complex was confirmed by immunoblotting with specific antibodies (Fig. 4a).
To independently test the association of HELQ with the BCDX2 complex, similar experiments were performed to capture proteins associating with XRCC2. XRCC2 was stably expressed as a Flag and HA epitope fusion (eXRCC2) in HeLa S3 cells. The eXRCC2 was sequentially immunoprecipitated with anti-FLAG and anti-HA antibodies and examined by silver staining and mass spectrometry (Fig. 4b). The most abundant specific proteins in this XRCC2 complex as indicated by the number of unique peptides were XRCC2 (25), HELQ (47), RAD51D (23), RAD51B (22) and RAD51C (18) (Supplementary Table S2). RAD51B, RAD51C, RAD51D and HELQ were verified by immunoblotting to be present in the XRCC2 complex (Fig. 4b). This also represents the first detection of endogenous HELQ in human cells. Neither the eHELQ or eXRCC2 complexes included RAD51, or DNA polymerase POLN, previously proposed as a HELQinteracting partner 25 (Supplementary Table S1).  HELQ interactions persist in cells exposed to MMC. We then asked whether HELQ-BCDX2 and HELQ-ATR interactions persist in cells treated with MMC. Flag and HA epitope tagged HELQ (eHELQ) was purified from nuclear extracts of cells treated without MMC or with MMC for 18 h. Similar amounts of ATR and BCDX2 were coimmunoprecipitated with HELQ before and after MMC treatment (Fig. 4c). These interactions were sustained after incubation with the DNase/RNase Benzonase, indicating that the HELQ complex can be isolated in the absence of intact polynucleotides (Fig. 4c).
To confirm these interactions in another cell line, and to exclude an epitope-tag-specific interaction, green fluorescent protein (GFP)-tagged HELQ was coexpressed with DsRed-Monomer-tagged RAD51B in HEK293T cells. After immunoprecipitation with GFP antibody from whole-cell extract, DsRed-Monomer-tagged RAD51B and endogenous ATR were identified, in the presence or absence of Benzonase. The GFP-tag alone did not interact with DsRed-Monomer-tagged RAD51B (Fig. 4d). A K365M mutation in the Walker A consensus ATPbinding site of HELQ lacks DNA helicase activity 16 . HELQ K365M retained the ability to interact with ATR and DsRed-Monomer-tagged RAD51B (Fig. 4d). The interaction between GFP-tagged HELQ and DsRed-Monomer-tagged RAD51B was also confirmed in U2OS cells ( Supplementary Fig. S2). We also coexpressed GFP-tagged HELQ and Flag-tagged XRCC2 in HEK293T cells, and confirmed that those proteins were coimmunoprecipitated with anti-GFP antibody. Moreover, endogenous ATR and XRCC2 were coimmunoprecipitated with GFPtagged HELQ (Fig. 4e).
It has been proposed that RAD51 paralogs may have a role in facilitating RAD51 assembly at the sites of DSBs and promoting intersister chromatid recombinational repair 9 . However, HELQ disruption did not affect MMC-induced RAD51 foci formation ( Supplementary Fig. S4A). Further, neither spontaneous nor  MMC-induced SCE frequencies were affected by HELQ disruption (Fig. 5). The lack of reduction of spontaneous SCE frequency is consistent with results obtained for RAD51 paralog disruption in rodent cells, which have normal or only modestly reduced frequencies of spontaneous SCE [26][27][28] . However, damage-induced SCE is reduced in Rad51d-knockout MEFs 29 .
HELQ has a role in promoting CHK1 activation. ATR kinase activity is triggered during ICL repair; the mechanism of this activation is complex and actively studied 30,31 . As HELQ À / À cells were sensitive to DNA-crosslinking agents and HELQ interacts with ATR, we asked whether HELQ status influences activation of the ATR-mediated DNA damage checkpoint after MMC treatment. To this end, U2OS (HELQ þ / þ ), 10H3 (HELQ þ / À ), 5G6 (HELQ À / À ) and 2G9 (HELQ À / À ) cells were incubated with MMC. Using CHK1 Ser345 phosphorylation as a marker, we found that MMC-induced CHK1 phosphorylation was significantly reduced in HELQ À / À cells (Fig. 6a). CHK1 phosphorylation signals to arrest mammalian cells in the G 2 /M phase of the cell cycle after DNA damage by MMC 32,33 . To measure G 2 /M arrest, cells were incubated either in the absence or presence of MMC, and their cell cycle profile was analysed. Twenty-four hours after MMC treatment, HELQ À / À cells had a lower G 2 /M population (17.1%) than wild-type (22.7%) (Fig. 6b). This finding was confirmed by an independent analysis method ( Supplementary Fig. S3). The results indicate that HELQ has a role in the activation of G 2 /M arrest after MMC treatment.
On the basis of model substrates in vitro, it has been proposed that HELQ opens up the parental strands at a blocked DNA replication fork and remodels nascent lagging strand intermediates to facilitate the loading of subsequent factors required for DNA damage processing or restart of DNA replication 18 .
At a DNA lesion present on only one DNA strand (such as a UV radiation-induced cyclobutane pyrimidine dimer), DNA replication forks are stalled and single-stranded DNA (ssDNA) is generated, which is required for ATR activation 34 . We examined global formation of RPA foci after treatment of growing cells with MMC, and found that it was not significantly affected in HELQ-knockout cells ( Supplementary Fig. S4B). Generation of ssDNA at replication forks blocked at monoadducts may not be the function of HELQ that promotes CHK1 phosphorylation. This is consistent with the unchanged sensitivity of HELQ-deficient cells to UV radiation or strand-breaking agents. However, when a DNA replication fork stalls at an ICL, processing to generate ssDNA may be impaired, and HELQ may have a more specific function in this context. A global assay of ssDNA formation would not detect such a defect, as a maximum of only about 5% of the adducts caused by MMC can be ICLs 35 .

Discussion
This study reveals that HELQ is involved in cellular resistance to ICLs in human cells, and places HELQ within a complex associated with the RAD51 paralogs. This is consistent with the cellular sensitivity to agents causing ICLs, which is conferred by loss of RAD51 paralogs 9,10 . The specific role of the RAD51 paralogs remains elusive, and it has been difficult to obtain mechanistic insights into how they operate in ICL repair 5,10,11 . We found that human HELQ only interacts with one of the two RAD51 paralog subcomplexes, BCDX2, whereas the 'CX3 complex' involving RAD51C and XRCC3 is not tightly associated. BCDX2 and CX3 may have independent functions in DNA repair. In chicken DT40 cells, a double RAD51B RAD51D mutant is no more sensitive to cisplatin than RAD51B single mutants, whereas an XRCC3 RAD51D mutation affecting both subcomplexes had an additive effect on the sensitivity 36 . Our findings indicate that HELQ has a BCDX2-related function affecting sensitivity to ICLs. The unchanged SCE frequency in HELQ-defective cells does not rule out a role for HELQ (or the RAD51 paralogs) in homologous recombination. SCE are thought to arise via homologous recombination to overcome DNA replication blocks. As previously suggested 27,37 , restart of stalled replication forks by homologous recombination may occur primarily by a noncrossover mode of Holliday junction resolution that does not produce SCE as cytologically identifiable events.
An association of HELQ and RAD51 paralogs is consistent with the phenotypic similarity of several egg-shell morphology mutants of Drosophila 38 . Mus301, a Drosophila homologue of HELQ, was isolated as a 'spindle-shaped' egg-shell morphology mutant called spn-C. Other similar spindle morphology mutants with female sterility include spn-A (Rad51) and the two Drosophila Rad51 paralogs, spn-B (Xrcc3-like) and spn-D (Rad51C-like) as well as okra (Rad54) 39 . In Caenorhabditis elegans, the HELQ homologue is termed helq-1 and its disruption confers sensitivity to ICL-inducing agents 40 . Apparently HELQ-1 together with RFS-1, the C. elegans RAD51 paralog, aids in the disassembly of RAD51 filaments from double-stranded DNA during meiosis 14,41 . The normal frequency of loading of RAD51 suggests a possible function for HELQ following RAD51 filament assembly and strand invasion in homologous recombination. A late role for some mammalian RAD51 paralogs in recombination has also been proposed, perhaps in resolution or disassembly of the recombination machinery 13 .
The RAD51 paralogs also have connections with a group of FA-related proteins, which are involved in repair and tolerance of ICL 42 . The HELQ-associated protein RAD51C is tentatively designated as FANCO 43 , and a homozygous XRCC2 mutation is present in an FA patient 44 . In the present experiments, HELQ was not epistatic to the downstream target of the FANC core complex, FANCD2 (Fig. 3a), and FANCD2 was not identified in the HELQ complex (Supplementary Table S1). Other FA-related   pathway proteins are not epistatic with the FA pathway in ICL repair, including SLX4 (also known as FANCP) and FAN1 (FANCD2/FANCI-associated nuclease 1) [45][46][47] . HELQ is a candidate FA-related gene, but like SLX4 and FAN1, HELQ may have functions that only partially overlap with the FA pathway. ATR is involved in both CHK1 phosphorylation and FANCD2 mono-ubiquitination 48 , but those pathways can proceed independently 49 . Although RAD17 and TOPBP1 are required for ATR-mediated CHK1 phosphorylation, chicken cells lacking RAD17 or human cells treated with siRNA-targeting TOPBP1 are able to sustain FANCD2 monoubiquitination 31,49 . FANCD2 was ubiquitinated after MMC treatment in HELQ-knockout cells (Fig. 3b). This indicates that HELQ is involved in ATR-mediated CHK1 activation but not in ATR-mediated FANCD2 monoubiquitination. A direct HELQ-ATR association may be important for CHK1 phosphorylation. The helicase activity-deficient HELQ K365M mutant was still immunoprecipitated with ATR in vivo (Fig. 4d). ATR was identified in the HELQ complex but not in the XRCC2 complex, which suggests that the association of HELQ with ATR does not require the BCDX2 complex (Fig. 4a,b).
The connection described here between HELQ and the RAD51 paralogs suggests that HELQ is a candidate gene for loss or mutation in sporadic or inherited ovarian cancer. Germ line mutations predisposing to ovarian cancer have been found in the RAD51C 50-53 and RAD51D genes 54 . Current data from The Cancer Genome Atlas (TCGA) shows that the frequency of heterozygous loss of HELQ in ovarian cancer is comparable to RAD51C and RAD51D (Table 1). There are currently seven serous ovarian cancer patient samples listed with HELQ homozygous deletion, which is more than found for the RAD51 paralogs, except RAD51D (Table 1). Several other mutations and polymorphisms in HELQ have been identified in various sequenced cancers (Fig. 7). A common intronic polymorphism in HELQ was recently identified as a locus significantly associated with age at natural menopause 55 . It will be worthwhile to determine the functional effect of this polymorphism on ovarian recombination function. The capacity for homologous recombination must be sustained in order to maintain human ovarian function with age 56 , indicating that alterations in HELQ may be relevant to both cancer and aging.

Methods
Experimental procedures for cell culture, antibodies, plasmid constructs, mass spectrometry and immunofluorescence are in the supplementary methods.
ZFN-mediated gene targeting. CompoZr Custom ZFNs designed to target exon 1 of HELQ were purchased from Sigma-Aldrich. pZFN1 and pZFN2 plasmids encoding ZFNs-targeting HELQ were transfected to U2OS by Amaxa Cell Line Nucleofector Kit V with programme X-001. The HELQ-specific ZFN binding (upper case letters) and cutting site (lower case) was 5 0 -CACCGCGGCCGAGCTC GTGCCCGGAGATGAGGGGAAAGAG-3 0 . The protocol to deliver ZFNs for the gene knockout was provided by Sigma-Aldrich. After the transfection, single-cell clones were sorted into 96-well plates by flow cytometry. Genomic DNA isolated from individual clones was amplified by PCR with ZFN primers F (5 0 -TTCCATA TGGGCTTAGCGTC-3 0 ) and R (5 0 -GTCCGGGTTTGTATCACCAC-3 0 ), and then each PCR product was used for the Cel-I Nuclease Mismatch assay (Surveyor). Targeted genomic DNA sequence was confirmed after TA-cloning of each Cel-I-positive PCR product. Confirmed clones (heterozygous knockout clones) were isolated and retargeted with the same HELQ-specific ZFNs to target the second allele of HELQ. The targeted genomic DNA sequence of HELQ-knockout cell lines was confirmed by direct sequencing of PCR products amplified with ZFN primers F and R, and sequencing after TA-cloning of the PCR products.
Survival experiments. For the ATPlite assay (PerkinElmer), 1,250 cells were dispensed per well into white 96-well plates and incubated 24 h before inducing DNA damage. The cells were incubated with DNA damage-inducing agents for the indicated time. After incubation, the cells were immediately lysed and assayed for ATPlite luminescence as described in the manufacturer's instructions. For the clonogenic assay, 1.0 Â 10 5 cells were plated in 60-mm culture plates and incubated for 24 h before DNA damage induction. Groups of plates were exposed to indicated doses of MMC for 1 h or UVC. After making a dilution series for each group, cells were returned to the incubator until colonies could be detected in the samples (7 to 14 days), and then were fixed, stained and scored for survival.
Affinity purification of HELQ and XRCC2 complexes. HeLa S3 cells stably expressing FLAG-HA epitope-tagged HELQ (eHELQ) or XRCC2 (eXRCC2) were grown to 1.0 Â 10 6 cells per ml as 9 or 12 l of suspension cultures 20,21 . For the MMC treatment, MMC (final 100 ng ml À 1 ) was added to the suspension culture 18 h before collection. The supernatant, nuclear extract and chromatin fractions were prepared from the cells, and the HELQ and XRCC2 complexes were immunoprecipitated from the nuclear extracts by incubating with M2 anti-FLAG agarose gel for 4 h with rotation in the presence or absence of 50 U ml À 1 Benzonase Nuclease (Novagen). After an extensive wash with buffer 0.1B (100 mM KCl, 20 mM Tris-HCl (pH 8.0), 5 mM MgCl 2 , 10% glycerol, 1 mM PMSF, 0.1% Tween-20, 10 mM b-mercaptoethanol), the bound proteins were eluted from M2 agarose by incubation for 60 min with 0.2 mg ml À 1 of FLAG peptide (Sigma-Aldrich) in the same buffer. A total of 100 ml of FLAG antibodyimmunoprecipitated material was further purified by immunoprecipitation with protein A sepharose (Pharmacia) conjugated with anti-HA 12CA5 antibody. The bound proteins were washed with 0.1B and eluted with 17 ml of 0.1 M Glycine-HCl (pH 2.5). After the elution, pH was neutralized with 3 ml of 1 M Tris-HCl (pH 8.0). To verify all proteins found in each complex by immunoblotting, only one gel was used per complex. Each membrane was cut horizontally into sections to immunoblot specifically for proteins of various sizes. Owing to the size similarity of certain proteins, such as RAD51B, RAD51C, RAD51D and XRCC2, some membrane sections were stripped using Thermo Restore Western Blot Stripping Buffer and reblotted multiple times for subsequent protein identification.   HELQ À / À ) collected 24 h after exposing to an indicated amount of MMC. SMC1 was used as a loading control. (b) G 2 /M populations in U2OS and HELQ À / À U2OS cells in response to MMC treatment. HELQ À / À and HELQ þ / þ U2OS cells were exposed to 30 ng ml À 1 MMC for 1 h and then collected at 24 h. The cells were pulse-labelled with EdU before collecting, fixed and stained. Cell cycle distributions were determined by flow cytometry and presented as the mean of two independent experiments (Expt) 1 and 2 for U2OS and three independent experiments 1, 2 and 3 for HELQ À / À U2OS cells (Expt 1 and 2 for 5G6 and Expt 3 for 2G9). The figure illustrates data for Expt 1 (see also Supplementary Figure S3). Full-size immunoblots can be found in Supplementary Fig. S5. siRNA Transfection. The FANCD2-specific Stealth RNAs (designated 'siFD2', 5 0 -AAUGAACGCUCUUUAGCAGACAUGG-3 0 ) (Invitrogen) and ON-TAR-GETplus Non-Targeting siRNAs as a negative control designated 'siC' (Thermo Scientific) were used. The siRNAs were introduced into wild-type or HELQknockout U2OS cells. Twenty-four hours before transfection, cells were plated in a six-well plate at 2.0 Â 10 5 cells per well. For each well, 5 pmol of siRNAs was diluted into 250 ml of Opti-MEM (Invitrogen). In a separate tube, 5 ml of Lipofectamine RNAiMAX reagent (Invitrogen) was diluted into 250 ml of Opti-MEM and incubated at room temperature for 10 min. The Lipofectamine RNAiMAX dilution was added into the diluted siRNA duplex and incubated at room temperature for 20 min. Before the transfection, the medium was replaced with fresh 2.5 ml of DMEM supplemented with 10% foetal bovine serum for each well. The Lipofectamine RNAiMAX-siRNA complex was added dropwise to the cells and incubated at 37°C. After 24 h, the cells were washed, trypsinized and plated with fresh DMEM medium supplemented with 10% foetal bovine serum and 1% penicillin-streptomycin (Invitrogen). To measure the levels of proteins, whole-cell crude extracts were prepared 48 h after the RNA transfection and analysed using immunoblotting. To analyse MMC sensitivity, the cells were treated with the indicated amount of MMC 48 h after the transfection.  6 g ml À 1 colcemid was added to the culture. The cells were trypsinized, rinsed with PBS, exposed to 0.075 M KCl for 15 min at 37°C, fixed in 3:1 methanol:glacial acetic acid at 48 h post BrdU treatment and spread on glass slides. The cells were stained with 0.5 Â SSC containing 2 mg ml À 1 Hoechst 33258 at 3 to 7 days after spreading cells. The cells were then washed with 2 Â SSC and 0.5 Â SSC, exposed to UVA light (Sylvania 15W F15T8/350BL) at a distance of 10 cm for 1 h without drying the cells and incubated with 0.5 Â SSC at 60°C for 10 min. The cells were Giemsa stained, and 35 metaphases per sample were analysed using a BX41 Olympus microscope with Â 100 oil objective and analysed to identify SCE number in each metaphase.