Ionizing radiation triggers chromatin-bound kin17 complex formation in human cells.

The human DNA-binding (HSA)kin17 protein cross-reacts with antibodies raised against the stress-activated Escherichia coli RecA protein. We show here that (HSA)kin17 protein is directly associated with chromosomal DNA as judged by cross-linking experiments on living cells. We detected increased amounts of DNA-bound (HSA)kin17 protein 24 h after gamma irradiation, with 2.6-fold more (HSA)kin17 molecules after 6 Gy of irradiation (46,000-117,000 molecules). At this time we observed that highly proliferating RKO cells displayed the concentration and co-localization of (HSA)kin17 and replication protein A in nucleoplasmic foci. Our results suggest that 24 h post-irradiation (HSA)kin17 protein may localize at the sites of unrepaired DNA damages. RKO clones expressing an (HSA)KIN17 antisense transcript (RASK.5 and RASK.13 cells) revealed that reduced (HSA)kin17 protein levels are correlated with a decrease in clonogenic cell growth and cell proliferation, as well as an accumulation of cells in early and mid-S phase. Taken together our observations support the idea that (HSA)kin17 protein is a DNA maintenance protein involved in the cellular response to the presence of DNA damage and suggest that it helps to overcome the perturbation of DNA replication produced by unrepaired lesions.

Ionizing radiation (IR) 1 induces a large range of DNA damage, including DNA double-strand breaks (DSBs), which represent a major threat to the integrity of mammalian genomes through chromosomal breakages and rearrangements (1). In mammalian cells, DSBs are repaired either by the homologous recombinational repair or by nonhomologous end joining (2,3). DSB repair pathways are usually characterized by the sequestration of many factors into discrete nuclear foci at the sites of DNA lesions and until completion of DSB repair (4). Some of the proteins belonging to these pathways act as a sensor for DNA damage or are involved in cell cycle checkpoints. This is the case for the histone H2AX, 53BP1, RPA, Rad51, BRCA1, or the Mre11-Rad50-Nbs1 nuclease complex (4 -8). For instance, the tumor suppressor gene BRCA1, previously involved in the regulation of the replication checkpoint and transcription-coupled repair, forms a multiprotein complex with Mre11-Rad50-Nbs1 and other proteins following irradiation, termed as BASC (BRCA1-associated surveillance complex), which may serve as a sensor of DNA lesions (8,9).
In this cascade of IR-induced proteins forming nuclear foci, we characterized here the HSA kin17 protein. Murine MMU kin17 protein was identified on the basis of a cross-reactivity with antibodies raised against the Escherichia coli RecA protein, a key enzyme in homologous recombination and recombinational repair of damaged DNA (10,11). This cross-reactivity stemmed from a sequence homology stretching over 39 amino acids highly conserved during evolution (12). This domain is located in the carboxyl-terminal region of the E. coli RecA protein, a region involved in the regulation of DNA binding (13). Recent data show that kin17 proteins are highly conserved during evolution. 2 In particular, a homologous protein has been identified in the yeast Schizosaccharomyces pombe. This conservation from yeast to human points to an essential role of kin17 proteins. For instance, the expression of kin17 proteins is preserved in the phylogeny of the brain of higher vertebrates (14,15). To date, major features of the kin17 protein are its abilities (i) to bind in vitro to double-stranded DNA and preferentially to DNA with a curved topology stretching over illegitimate recombination junctions (16,17), (ii) to complement the functions of the H-NS (histone-like nucleoid structure) protein transcription factor in deficient bacterial strains in controlling gene expression (18), and (iii) to be a stress-activated protein recruited during the cellular response to ionizing radiation or UVC (19,20). In particular, UVC irradiation induced a stabilization of MMU KIN17 mRNA from 80 min to more than 8 h in mouse fibroblasts (21). Interestingly, ⌬XPA mouse cells, which are unable to repair UVC-induced DNA damage, accumulated MMU KIN17 mRNA at lower doses of UVC (5-10 J/m 2 ) than repair-proficient mouse fibroblasts (20 -30 J/m 2 ), suggesting that DNA damage per se is required for the stabilization of MMU KIN17 mRNA (21).
In human cells, the HSA KIN17 gene is localized on chromosome 10 at position p15-p14. HSA KIN17 transcripts were ubiquitously found at low levels in all human organs tested, displaying an expression profile akin to that of housekeeping genes. Heart, skeletal muscle, and testis displayed the highest HSA KIN17 mRNA levels compared with the other tissues ana-* This work was supported by Electricité de France Contract 8702. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: CEA-DSV-DRR, Laboratoire de Génétique de la Radiosensibilité, BP  lyzed (12). HSA kin17 protein expression is associated with the differentiation program of human keratinocytes cultivated in the in vitro reconstructed skin model (22). Mouse and human kin17 proteins are able to arrest cell proliferation of human tumor cells when their expression is transiently increased after transfection (23). 3 Only the immortalized but not tumorigenic HEK 293 cell line tolerates ectopic expression of either MMU kin17 or HSA kin17 proteins and can be propagated in mass culture for several weeks. However, the constitutive overexpression of MMU kin17 protein in HEK 293 cells entailed major growth defects and nuclear abnormalities (23). We show here for the first time that the HSA kin17 protein is mainly present in the nuclear compartment associated with nuclear structures in human cells. We demonstrate that a fraction of nuclear HSA kin17 protein is directly associated with DNA. HSA kin17 protein is localized in discrete nuclear foci spread throughout the nucleoplasm. Strikingly, ␥ irradiation induces an increase in the DNA-bound fraction of HSA kin17 protein that correlates with the appearance of foci containing both HSA kin17 protein and replication protein A (RPA) 24 h after IR. To ascertain the requirement of HSA kin17 in the immortalized phenotype, we have expressed HSA KIN17 antisense transcript in RKO carcinoma cells. These RKO cells expressing the antisense HSA KIN17 (called RASK cells) displayed markedly reduced proliferation rates associated with a defect in S phase progression. All of our observations support the idea that HSA kin17 protein is involved in the cellular re-3 D. S. F. Biard, unpublished data.
FIG. 1. HSA kin17 protein binds directly to chromosomal DNA in vivo. In vivo cross-linking with formaldehyde (1%, 3 min) was performed in HeLa cells. After cross-linking, the nuclei were isolated and lysed, and DNA-protein complexes were purified by equilibrium centrifugation in two consecutive cesium chloride gradients. The proteins were analyzed onto a 10% SDS-PAGE. A, purification of DNA-protein complex from proliferating HeLa cells after the first equilibrium centrifugation revealed both HSA kin17 and PCNA in the same fractions. B, DNase-treated fractions 6 -8 were pooled, recentrifuged on an isopycnic cesium chloride gradient, fractionated, and analyzed as above. C, as in B with HeLa cells treated with 0.2 M nocodazole for 16 h. D, DMS treatment demonstrates a direct binding of HSA kin17 to DNA. Purified DNA-protein complexes from a second CsCl gradient were treated with DMS according to the procedure described under "Materials and Methods." Lane 1, mock treated cells ϩ formaldehyde cross-link. sponse to the presence of DNA damage and suggest that it may help to overcome the DNA replication arrest produced by unrepaired lesions.

MATERIALS AND METHODS
Cell Cultures-Human colorectal carcinoma RKO cells were obtained from M. F. Poupon. Human cervical carcinoma HeLa cells were obtained from E. May. The cell lines were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum, 100 units/ml of penicillin, and 100 g/ml of streptomycin, under 5% CO 2 .
Cloning of EBV vectors carrying HSA KIN17 cDNA in an antisense orientation has been performed as described elsewhere (23). Transfected RKO cells carrying EBV plasmids were propagated in culture in the presence of 500 g/ml hygromycin B (Invitrogen). We used the following vectors: pEBVCMV (pB482) and pEBVCMVas HSA Kin17 (pB399as). Transfection experiments were carried out using Lipo-fectAMINE 2000 (Invitrogen). RKO clones transfected with the pB399as EBV vector carrying the HSA KIN17 cDNA-SV40 polyadenylation signal cartridge in an antisense orientation were termed RASK, for RKO antisense HSA KIN17 cDNA. Control clones carrying the pB482 plasmid were named R482.
The cells were irradiated using a 137 Cs source (IBL 637 CisBio International) with a dose rate of 1.97 Gy/min. For clonogenic cell growth, the RKO cells were seeded as indicated in the table and cultivated for 2 weeks in the presence of hygromycin B. Growing clones were fixed with 4% paraformaldehyde and stained with methylene blue, and the clones containing more than 50 cells were counted. Each experiment was done three times.
Monoclonal Antibodies against HSA kin17 Protein and ELISA-Monoclonal antibodies (mAb K3, mAb K31, mAb K36, and mAb K58) were obtained after inoculation of recombinant His-tagged human HSA kin17 protein (His 6 -HSA kin17) purified by metal chelation and heparin column chromatography from baculovirus-infected Sf9 Spodoptera frugiperda cell extracts and injected in mice as described previously. 4 We used either hybridoma supernatants (mAb K36) or IgG anti-HSA kin17 protein (Ig K36; Ig K58) purified from ascites fluid. Purified immunoglobulin from rabbit polyclonal antibody anti-His 6 -HSA kin17 (IgG 77P) was obtained as described elsewhere. 4 A conventional two-site immunometric assay (sandwich immunoassay) based on two specific monoclonal antibodies recognizing nonoverlapping epitopes was developed essentially as described by Grassi et al. (24). We chose the mAb K3 and mAb K31 monoclonal antibodies, the FIG. 3. Detection of HSA kin17 protein relocalization by indirect immunofluorescence 24 h after irradiation. RKO cells were seeded 3 days before irradiation and fixed with acetone/methanol 24 h after irradiation (2 and 6 Gy). HSA kin17 protein was revealed using the monoclonal antibody mAb K36 and a Cy2 TM -conjugated affinity-purified goat anti-mouse IgG. Two digitized images of representative cells/ dose are shown at a magnification of 500ϫ.
FIG. 4. RKO cells displaying low levels of HSA kin17 protein fail to accumulate HSA kin17 protein after ionizing radiation. A, 7 weeks after transfection and selection in the presence of 500 g/ml hygromycin B, RASK and R482 cells were fixed, and endogenous HSA kin17 was revealed using the mAb K36 antibody as described for Fig. 3. The arrows indicate an extranuclear structure close to the nuclei. B, 10 weeks after transfection, 0.2 ϫ 0.10 6 cells were seeded in 6-cm dishes in the presence of 500 g/ml hygromycin B and irradiated 3 days later at 1.97 Gy/min. 24 h later, the cells were trypsinized, counted, and lysed with RIPA buffer as indicated under "Materials and Methods." The protein of 60,000 cells was loaded onto a 10% SDS-polyacrylamide gel. mAb K3 being conjugated to acetylcholinesterase as reporter enzyme (24).
Indirect Immunofluorescence Staining-The cells were plated at 5,000 cells/cm 2 on glass coverslips and treated. At the indicated times after treatment, the cells were fixed for 5 min in 70% acetone, 30% methanol at Ϫ20°C. The primary antibodies were diluted in the incubation buffer B (0.5% Tween 20, 12% bovine serum albumin, 0.036% NaN 3 in PBS) and incubated for 45 min. The following antibodies were used: hybridoma supernatant mAb K36 anti-HSA kin17 (diluted by half), purified Ig K36 anti-HSA kin17 (450 ng/ml), purified rabbit immunoglobulin IgG 77P against HSA kin17 (5 g/ml), and anti-RPA70 (directed against the 70-kDA subunit; mouse monoclonal NA13 antibody, 500 ng/ml; Oncogene Research Products). Primary antibodies were revealed with either Cy2 TM -conjugated affinity-purified goat anti-mouse IgG or Cy3 TM -conjugated affinity-purified goat anti-rabbit IgG (Jackson Laboratories, Inc.; 2 g/ml). The cells were counterstained with 4Ј,6-diamino-2-phenylindole (4 g/ml). Immunofluorescence staining was viewed using a Zeiss Axiophot 2 epifluorescence microscope coupled to a cooled Sensys 1400 camera from Photometrics monitored by the Zeiss KS300 3.0 program. The use of a CCD camera-based imaging system allows high resolution and a wide dynamic range for acquiring and analyzing fluorescent staining. Representative fields for each cell line are presented. Irradiation experiments were reproduced more than 10 times under different culture conditions.
Protein Extraction and Western Blot-The cells were seeded at 5,000 cells/cm 2 3 days before trypsinization and treated as indicated. The cells were lysed using RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Igepal TM , 0.1% SDS, 0.1% sodium deoxycholate, protease inhibitor mixture from Roche Molecular Biochemicals) or buffer N (50 mM Tris-HCl, pH 7.9, 150 mM NaCl, 1% Igepal TM , 1 mM EDTA, protease inhibitor mixture from Roche Molecular Biochemicals), as indicated in the legends of the figures. The lysates were kept on ice for 30 min with the buffer N, and soluble proteins recovered after centrifugation (20,000 ϫ g for 15 min) were quantified by Bradford assay (Bio-Rad) and analyzed by ELISA or Western blot. The remaining pellets were denatured in Laemmli sample buffer, boiled for 10 min at 100°C, and analyzed by Western blot.
Purified IgG Ig K36 and Ig K58 were used at a concentration of 25 ng/ml. Other antibodies used were anti-p53 protein (hybridoma supernatant mAb DO-7 antibody diluted to 1:2000). DO-7 antibody was kindly provided by Dr. E. May. Anti-PCNA (mouse monoclonal PC10 antibody diluted to 50 pg/ml; Novo Castra) and anti-RPA70 (25 ng/ml) were also used.
In Vivo Cross-linking and Preparation of DNA-Protein Complexes-The procedure has been described by Göhring and Fackelmayer (25). Briefly, HeLa cells were washed once with PBS and incubated for 3 min at 37°C in Dulbecco's modified Eagle's medium (without serum) containing 1% formaldehyde. Cross-linked cells were recovered by centrifugation (5 min at 750 ϫ g). The cells were resuspended in buffer RSB (10 mM Tris-HCl, 10 mM NaCl, and 3 mM MgCl 2 , pH 8.0) and homogenized in a chilled Dounce homogenizer. The nuclei were collected by centrifugation (8 min, 750 ϫ g), and the unbound proteins were extracted in buffer E (10 mM Tris-HCl, 10 mM Na 2 S 2 O 5 , 1 M NaCl, 0.1% Igepal TM , 1 mM EDTA-KOH, and 0.5 mM phenylmethylsulfonyl fluoride, pH 8.0). After extraction, the nuclei were pelleted as above, resuspended in 0.1 M NaCl, and lysed at a final concentration of 2% sodium sarkosyl. The samples were layered on a preformed CsCl step gradient (3 ml of 1.75 g/ml CsCl solution, 3.5 ml of 1.5 g/ml CsCl solution, and 3 ml of 1.3 g/ml CsCl solution). All CsCl solutions were prepared in 20 mM Tris-HCl, 1 mM EDTA-KOH, and 0.5% sodium sarkosyl. After centrifugation for 24 h at 200,000 ϫ g at 20°C and fractionation from the top, DNA-protein complexes sediment at a density of ϳ1.3 g/ml.
Aliquots of the fractions were sonicated, and the density of individual fractions was determined. For RNase digestion, the pooled fractions were dialyzed against 20 mM Tris-HCl, pH 8.0, 50 mM NaCl, and 2 mM EDTA before 5 g of DNase-free RNase was added for 30 min at room temperature. Solid CsCl was added to a density of 1.5 g/ml in a final volume of 5 ml, and the sample was recentrifuged for 72 h at 250,000 ϫ g at 20°C. The gradient was fractionated and analyzed. The DNA-containing fraction was briefly sonicated and desalted by gel filtration on Sephadex G25 columns. Cross-links were cleaved by incubation at 100°C for 10 min in Laemmli sample buffer and analyzed on SDS-polyacrylamide gels.
Additional dimethylsulfate (DMS) treatment was done after the second CsCl gradient as follows. The DNA-containing fraction was sheared by sonication and then desalted over a Sephadex G25 column into 30 mM sodium phosphate, pH 7.4, 2.5 mM EDTA, 2.5 mM EGTA. After desalting, 1 ⁄10 volume of 10.5 M DMS was added for 18 h on ice, then for 8 h at 45°C. 1 ⁄10 volume of (fresh) 0.2 M NaBH 4 was then added for 30 min on ice. After incubation, SDS was added to 1% final concentration, and the sample was boiled for 60 min. After boiling, the samples were treated once more with NaBH 4 at 4°C, and DNA-protein complexes were analyzed by SDS-PAGE.
Flow Cytometry Analysis of BrdUrd Incorporation-The cells were plated at a density of 5,000 (control R482) or 10,000 (RASK) cells/cm 2 3 days before irradiation. 24 h later, the cells were pulse-labeled with 30 M BrdUrd for 15 min, washed in PBS, and collected by trypsinization. The cells were resuspended in PBS and fixed with ethanol (75%). BrdUrd-labeled cells were detected as described by Bensaad et al. (26). Briefly, the nuclei were isolated following treatment with pepsin 0.5% in 30 mM HCl for 20 min, and cellular DNA was partially denatured with 2 N HCl for 20 min at 37°C. After extensive washing, the cells were incubated successively with rat anti-BrdUrd antibodies for 1 h at room temperature and with fluorescein isothiocyanate-conjugated goat antirat IgG secondary antibody for 30 min at room temperature. Then they were washed again twice in PBS and stained with 25 g/ml propidium iodide for 20 min at room temperature. The data were collected using a FACsort flow cytometer (BD PharMingen).

RESULTS
Association of Human HSA kin17 Protein with Chromosomal DNA in Vivo-Mazin et al. (16,17) demonstrated that in vitro mouse MMU kin17 protein recognizes DNA, particularly doublestranded DNA with a curved topology. In agreement with these results, we have observed that most HSA kin17 protein was localized in nuclei of different cultured human cell lines (data not shown). We then tested whether in vivo endogenous human HSA kin17 protein could be associated with chromosomal DNA. We used a method based on a limited cross-linking of living HeLa cells with formaldehyde to stabilize DNA-protein interactions prior to the extraction (25). This method minimized the formation of nonspecific cross-links and further excluded noncross-linked material by two consecutive cesium chloride density gradient centrifugations. Equilibrium density gradient centrifugation separates cellular components according to their density. Under these experimental conditions, DNAprotein complexes exhibited a density near to that of native chromatin (1.4 g/ml). Covalent bonds introduced by formaldehyde are reversible by boiling in SDS-containing buffers, thus allowing the analysis of proteins by SDS-PAGE. After a first centrifugation, we observed a co-migration of HSA kin17 and PCNA proteins as components of a high molecular weight complex displaying similar densities (Fig. 1A). Fractions 6 -8 containing DNA were pooled and purified by a second isopycnic centrifugation. Most HSA kin17 protein was detected in fractions corresponding to DNA-protein complexes, suggesting that HSA kin17 protein was bound in vivo to DNA in HeLa cells (Fig. 1B).
We asked whether HSA kin17 protein is associated with the chromatin in a cell cycle-dependent manner. HeLa cells were treated with a microtubule poison (nocodazole) to trigger an anaphase arrest and to trap cells into mitosis (27). HeLa cells arrested in G 2 -M retained HSA kin17 tightly associated with the chromatin structure, as did mock treated cells (Fig. 1C). Therefore, HSA kin17 remains associated with the chromatin structure in both proliferating or G 2 -M-arrested cells. Hence, in vivo association of HSA kin17 protein with DNA structures was independent of the cell cycle. We next examined whether HSA kin17 protein was directly bound to DNA. We used DMS to convert heat-reversible methylene bonds induced by formaldehyde to stable DNAprotein covalent bonds that are resistant to boiling in SDScontaining buffers. Under these conditions, if HSA kin17 protein is directly associated with the chromatin, the DNAprotein complex is stabilized and cannot be separated by SDS-PAGE. Conversely, if HSA kin17 protein is linked to pro-teins of the chromatin, covalent methylene bonds are reversed by boiling, and the protein can migrate in SDS-PAGE. After DMS treatment the HSA kin17 band disappeared, showing that in vivo most Hs kin17 protein is directly linked to DNA (Fig. 1D, lane 2).
␥ Irradiation Increases the Amount of HSA kin17 Protein Bound to DNA-Because HSA kin17 is a chromatin-bound protein, we next asked whether ␥ irradiation could preferentially increase the HSA kin17 fraction tightly associated with DNA. First, experiments were carried out to estimate the kinetics of induction of the total HSA kin17 protein content in RKO cells. RKO cells were seeded 3 days before irradiation and ␥-irradiated at 50% of confluence to avoid any effect caused by serum stimulation. Under these conditions, about 25% of the cells were in S phase as determined by BrdUrd pulse incorporation and flow cytometry analysis (see Fig. 5A and data not shown). Total proteins were recovered, and the HSA kin17 protein level was assessed at different times following irradiation. Although a tremendous induction of p53 was always observed 6 and 24 h after irradiation, we usually observed an increased HSA kin17 protein level only 24 and 48 h after irradiation ( Fig. 2A). At these times, most of the RKO cells were arrested in the G 2 phase, as evidenced by flow cytometry (see Fig. 5E and data not shown). Under these conditions, the PCNA protein level remained unchanged ( Fig. 2A). Second, we assessed the level of HSA kin17 protein anchored to the DNA structure 24 h after IR. In this approach, we used a lysis buffer containing 1% Igepal TM (buffer N) to discrimi-nate between cytoplasmic and soluble nuclear proteins (detergent-soluble fraction) and nuclear proteins highly anchored to DNA (DNA-bound fraction). We observed a dosedependent increase in the DNA-bound HSA kin17 protein levels 24 h after ␥ irradiation starting with a dose of 0.5 Gy (Fig. 2B). No significant induction was noted in the detergent-soluble fraction. Therefore, ␥ irradiation mainly induced DNA-bound HSA kin17 protein. The p53 protein level increased in both fractions, suggesting that ␥ irradiation induced both DNA-bound p53 as well as detergent-soluble p53 (Fig. 2B). We also noted a slight increase in the PCNA content in the soluble fraction as well as an increase in the insoluble fraction only at 10 Gy. A similar increase in DNAbound PCNA protein has already been reported at doses higher than 10 Gy (28).
To confirm that IR increased the level of HSA kin17 protein

FIG. 5-continued
Ionizing Radiation Induces DNA-bound HSA kin17 Protein tightly associated with DNA, the proteins were recovered with buffer N containing increased ionic strengths (0.15, 0.5, or 1 M NaCl), and the amount of HSA kin17 protein in the detergentsoluble fraction was quantified by ELISA. To quantify the amount of HSA kin17 protein/cell, we used a two-site immunometric assay (sandwich immunoassay) based on two monoclonal antibodies (24). HSA kin17 recombinant protein allowed us to calibrate the ELISA (data not shown). The irradiation experiments were performed with RKO cells and with derived clones carrying a pEBVCMV vector (see below). We observed an enhanced recovery of the soluble HSA kin17 protein with increased ionic strength of buffer N (0.5 or 1 M NaCl) as compared with 0.15 M NaCl (Fig. 2C). Considering that 100% of HSA kin17 molecules were recovered with 1 M NaCl after irradiation at 6 Gy, we obtained only 73% at 0.5 M NaCl and 35% at 0.15 M NaCl. At 1 M NaCl we reached roughly the same percentage recovery as that obtained with an RIPA buffer. This indicated that increased amounts of HSA kin17 tightly associated with DNA were observed after damage to DNA (Fig. 2C). At higher ionic strength, the number of HSA kin17 molecules increased 2.6-fold 24 h after irradiation at 6 Gy (46,000 -117,000 molecules). A similar result was obtained with the parental RKO cell lines, implying that the expression of the viral EBNA-1 protein did not interfere with the HSA kin17 response (data not shown).
Localization of HSA kin17 Protein in Large Nucleoplasmic Foci Following ␥ Irradiation-Because several nuclear proteins involved in DNA repair/DNA damage recognition pathways concentrate into nuclear foci after irradiation, we performed immunocytochemical staining using the mAb K36 antibody to determine HSA kin17 sublocalization in RKO cells. The cells were seeded 3 days before irradiation to avoid serum stimulation of proliferation-associated proteins such as HSA kin17. In nonirradiated RKO cells, HSA kin17 showed a weak and diffuse staining pattern throughout the nucleoplasm (Fig. 3). Twenty-four hours after irradiation, enhanced HSA kin17 protein levels were clearly detected at 2 and 6 Gy. At these times, HSA kin17 protein coalesces into large foci that might correspond to the HSA kin17 fraction tightly associated with DNA.
Isolation of RASK Cells Expressing an Antisense HSA KIN17 Transcript-To evaluate the importance of foci-forming HSA kin17 protein during cell proliferation and the cellular response to ionizing radiation, we generated several clones displaying low levels of HSA kin17 protein.
In a first step, we analyzed endogenous HSA kin17 protein levels in different human cells by ELISA and Western blot. We conclude that endogenous HSA kin17 protein is tightly associated with the DNA, whatever the cell line used (data not shown). We also observed that human carcinoma cells, such as RKO cells, exhibited the greatest number of HSA kin17 molecules/cell as compared with either normal human fibroblasts or other tumoral cells. 4 For this reason, we decided to reduce the HSA kin17 protein level in RKO cells. RKO clones were isolated after transfection of the pEBVCMVas HSA KIN17 vector (pB399as) carrying a HSA KIN17 cDNA-SV40 polyadenylation signal cartridge in an antisense orientation, followed by subse-quent hygromycin B selection. These clones were termed RASK (RKO antisense HSA KIN17. From the 60 clones isolated, half died rapidly, and the others grew very slowly. After several weeks of cultivation, three RASK clones were selected and characterized in more detail (RASK.1, RASK.5, and RASK.13). Three clones carrying the pEBVCMV vector (pB482) were selected at random and used as controls (R482.1, R482.2, and R482.3).
Because RASK cells expressing the antisense HSA KIN17 cDNA were usually unstable, we systematically assessed HSA kin17 protein levels by either immunocytochemical staining or Western blot. As judged by indirect immunofluorescence, more than 95% of HSA kin17 protein was essentially localized in nucleoplasmic speckles of diameters ranging from 0.1 to 2 m in proliferating R482 cells (Fig. 4A). In highly proliferating cells, the greater number of HSA kin17 foci inside nuclei of R482 cells led to an intense nuclear staining. We further noted that  a few R482 cells presented the staining of an extranuclear structure close to the nuclei (indicated by arrows in Fig. 4A).
In the parental RKO cells, we currently detected an extranuclear structure close to the nuclei and probably in the vicinity of the nucleosome, which concentrated the HSA kin17-specific fluorescence signal. This indicated that the expression of a viral nuclear protein (EBNA-1) coded by EBV vectors could not account for the observed distribution of the HSA kin17 protein. Seven weeks after transfection and hygromycin B selection, the production of antisense HSA KIN17 mRNA in RASK cells appeared to decrease strikingly the number of HSA kin17 nucleoplasmic foci, leading to a weak diffuse nuclear staining (Fig. 4A). Curiously, RASK cells maintained one or two extranuclear stained structures/cell, suggesting that HSA kin17 protein presented here was certainly very stable (arrows in Fig. 4A). Western blot analysis of HSA kin17 protein also revealed a dramatic decrease in the HSA kin17 basal level (70 -80% less) in the three antisense clones selected, as compared with controls (Fig. 4B). This reduction was specific for HSA kin17 protein because the expression of PCNA remained unchanged. Interestingly, RASK cells failed to induce HSA kin17 protein 24 h after irradiation, as did control clones. Therefore, we used RASK cells to study the biological consequences of a reduced HSA kin17 protein content in a human tumor cell.
Early and Mid-S Phase Accumulation of RASK Cells-We asked whether reduced HSA kin17 protein levels affect cell proliferation. Plating efficiencies of the different clones were assessed after seeding the same number of control and RASK cells/cm 2 . Under these culture conditions, RASK cells exhibited a markedly decreased proliferation rate, with plating efficiencies 15-fold lower than those observed for control clones (data not shown). When cells were plated at different densities to account for their specific plating efficiencies, we also observed a dramatic decrease of growing clones in both RASK clones (Table I). Therefore, decreased levels of HSA kin17 protein strongly affected cell growth.
This decreased cell proliferation observed in RASK cells that express the antisense HSA KIN17 mRNA could stem from alteration of their cell cycle. Therefore, we analyzed the cell cycle of BrdUrd pulse-labeled RASK clones by flow cytometry analysis. The cells were seeded 3 days before irradiation (6 Gy) and analyzed 24 h later. The cells were pulse-labeled for 15 min with 30 M BrdUrd. Incorporation of BrdUrd into cellular DNA was measured by fluorescence-activated cell sorter analysis, and the percentage of BrdUrd-positive cells corresponded to S phase cells actively synthesizing DNA (Fig. 5).
Comparison of the cell cycle of RASK cells versus control cells revealed a low number of RASK.5 cells in the G 1 phase (45 Ϯ 2%) as compared with control (65 Ϯ 2%) and an enhanced proportion of BrdUrd-positive cells with a DNA content between 2 and 4 N (34 Ϯ 1% versus 24 Ϯ 2%) ( Fig. 5 and Table II). RASK.5 cells accumulated in early and mid-S phase, but only a few cells were detected in late S phase (Fig.  5C). This suggested that low HSA kin17 protein levels resulted in better entry into the S phase, but replicating cells were hampered in their progress to the S phase. We also detected an elevated percentage of cells in the G 2 phase (20 Ϯ 1%) as compared with control cells (12 Ϯ 1%). We obtained similar results using RASK.13 cells (data not shown). These cell cycle modifications could explain the reduced proliferation rates observed in the RASK cells. R482.1 cells and RASK.5 cells displayed a G 2 arrest 24 h after 6 Gy, with 65 Ϯ 9% and 61 Ϯ 4% of cells in G 2 phase, respectively, indicating that HSA kin17 protein is not essential for the ␥-ray-induced G 2 arrest (Fig.  5, F versus H).
Endogenous HSA kin17 Co-localized with the RPA Protein 24 h after Irradiation-Because HSA kin17 is (i) associated with DNA and (ii) induced after IR at later times post-irradiation (24 h), we hypothesized that after irradiation HSA kin17 protein could be associated with remaining DNA lesions. This prompted us to compare the intracellular localization of HSA kin17 with a protein known to participate in the DNA repair processes.
In mock-irradiated proliferating RKO cells, RPA was uniformly distributed throughout the nucleoplasm as a dispersed and punctate pattern that corresponds to replication foci (Fig. 6A). After ␥ irradiation, RPA concentrated in nuclear foci of very strong intensity and almost all the bright RPA foci co-localized with HSA kin17 (Fig. 6, B and C). These results suggested that HSA kin17 and RPA could cooperate in response to IR-induced DNA lesions. However, RPA and HSA kin17 foci of strong intensity were never detected soon after irradiation at 6 Gy (3 h and 6 h) in proliferating RKO cells nor at later times post-irradiation in slowly proliferating RKO cells (data not shown). Co-immunoprecipitation experiments were unsuccessful in demonstrating a strong physical interaction between endogenous HSA kin17 and RPA-70 proteins in RKO cells. We assumed that (i) HSA kin17 and RPA belong to a same high molecular weight complex without direct interaction between them, (ii) only a small fraction of both proteins participate in the same nuclear foci, and (iii) constitutive amounts of both HSA kin17 and RPA proteins were too low to be detected under our experimental conditions.

DISCUSSION
The well conserved kin17 proteins are DNA-binding proteins activated in response to ionizing and UVC irradiations (19,20). Prior studies were mainly performed at the mRNA level. Recently, the production of large amounts of human recombinant HSA kin17 protein made it possible to obtain a panel of new monoclonal antibodies and to develop biochemical approaches. This also affords us the opportunity to quantify endogenous HSA kin17 protein levels in different human cells using ELISA. Normal, immortalized, or tumoral cells present a wide range of HSA kin17 protein levels. Although proliferating normal human fibroblasts elicited a low level, non-small cell lung carcinoma cells (H1299) and colorectal carcinoma cells (RKO) display the highest level. 5 Therefore, high HSA kin17 protein levels are observed during carcinogenesis that may be a consequence of uncontrolled proliferation or genomic instability. Alternatively, elevated levels could be required during the process of cancer development. Immunocytochemical staining performed with the mAb K36 antibody identified HSA kin17 as a protein mainly localized in nuclei with a staining pattern resembling those observed for other proteins involved in DNA replication. This observation is consistent with the previously reported localization of kin17 protein in HeLa cells using antibodies directed against the mouse MMU kin17 protein and raised in rabbits (12). Interestingly, certain tumor-derived cells concentrate a part of the HSA kin17 protein in a dense fluorescence focus neighboring to the nucleus, particularly in resting cells. At present, we have no explanation for this observation.
We identify here HSA kin17 as a protein tightly associated in vivo with the chromosomal DNA. We show that HSA kin17 protein exists in the cells as both a soluble fraction and a DNAbound fraction. Other nuclear proteins involved in the DNA metabolism, such as PCNA, exhibited a similar distribution. IR triggers a redistribution of HSA kin17 protein from a soluble form to DNA-bound complexes comparable with that observed for PCNA (28). However, whereas an increased level of the PCNA insoluble fraction was observed at higher doses of irradiation (e.g. 10 Gy in our experiment), the DNA-bound HSA kin17 fraction increases at a lower dose (0.5 Gy). This enhanced amount of HSA kin17 bound to DNA observed after irradiation coincides with the appearance of large intranuclear focal sites of HSA kin17 scattered throughout the nucleus. We show that HSA kin17 and RPA proteins co-localized in large foci 24 h after irradiation at 6 Gy. IR-induced RPA/ HSA kin17 foci were observed only in highly proliferating cells, suggesting that DNA replication is required.
The heterotrimeric RPA protein is a single-stranded DNAbinding protein required for DNA replication, recombination, nucleotide excision repair, DSB repair, and transcriptional regulation (29). RPA is a crucial component of the early stage of nucleotide excision repair, because it binds synergistically with XPA to damaged single-strand DNA, allowing the subsequent recruitment of the other repair factors at the site of DNA damage (30,31). RPA is also involved in the gap-filling step of nucleotide excision repair, which requires PCNA, RF-C, and DNA polymerase ␦/⑀ (32). RPA forms discrete foci after irradiation in many cell lines. The rate of dispersion of foci-forming proteins is usually compared with the rate of DSB repair. Kinetic experiments of DSB repair have previously shown a biphasic response with a fast component for repair of most breaks (half-time ranging from 20 to 30 min) followed by a slow component for repair of the remaining breaks (90 -300 min) (33). MacPhail and Olive (34) have shown that the extent of IR-induced RPA foci increased linearly between 8 and 24 h of incubation after irradiation, suggesting that RPA foci concentrated after the completion of DNA repair at sites of unrepairable DNA damage. Golub et al. (35) have demonstrated that RPA co-localized with the DNA recombinational protein Rad51 through its 70-kDa subunit. The highest number of Rad51-RPA co-localizations was observed 1 day following irradiation (5 Gy) of mouse fibroblasts. Furthermore, 30 h after irradiation RPA co-localizes with Rad51 in micronuclei (36). Taken together these data indicate that RPA foci are associated with unrepaired DNA damage and/or with DNA sites that are unable to replicate. The idea that a fraction of HSA kin17 protein colocalized with RPA foci at these sites of damaged DNA is further supported by the observation that HSA kin17 protein co-purifies with proteins of the replication complex of human cells. 5 If HSA kin17 protein acts at the site of DNA replication, in particular of damaged DNA, a lowered HSA kin17 protein level might impede cell proliferation and decrease resistance to IR. The antisense strategy used here to constitutively decrease the HSA kin17 protein level confirmed this idea. The overexpression of an antisense HSA KIN17 transcript led to a 75% decrease in the HSA kin17 protein level, which correlates with a reduced cell growth and increased radiosensitivity. 6 Flow cytometry analysis of BrdUrd incorporation showed an accumulation of HSA kin17 antisense cells in early and mid-S phase and a subsequent increase in the number of cells in the G 2 phase. The increased number of cells in the G 2 phase may correspond to RASK cells undergoing DNA repair. We hypothesized that a premature entry of RASK cells into the S phase, as evidenced by the low number of cells in the G 1 phase, may lead to an accumulation of DNA damage and a subsequent arrest in the G 2 phase. The accumulation of RASK cells in early and mid-S phase indicates that HSA kin17 protein participates in DNA replication. Indeed, ongoing experiments revealed that HSA kin17 protein co-purifies with RPA and PCNA proteins in elution fractions corresponding to DNA replication complexes. 5 Our results show that human PCNA and HSA kin17 proteins are components of the same set of replication proteins activated by ionizing radiation. The decrease in the intracellular concentration of HSA kin17 protein affects the cell cycle, apparently by interfering with proteins localized at damaged DNA sites unable to replicate. The RASK cells described here compared with other human cells unable to repair doublestrand breaks will be used to test this idea further and to shed some light on the molecular role played by the human HSA kin17 protein.