HRad17 Colocalizes with NHP2L1 in the Nucleolus and Redistributes after UV Irradiation*

The rad17 gene ofSchizosaccharomyces pombe plays an important role as a checkpoint protein following DNA damage and during DNA replication. The human homologue of S. pombe rad17,Hrad17, was recently identified, but its function has not yet been established. Using the yeast two-hybrid system, we determined that HRad17 can interact with a nucleolar protein, NHP2L1. This interaction was also demonstrated biochemically, in human cells. Immunofluorescence studies revealed that HRad17 and NHP2L1 colocalize to the nucleolus, and immunogold labeling further resolved the location of NHP2L1 to the dense fibrillar component of the nucleolus. Interestingly, the localization of HRad17 in the nucleolus was altered in response to UV irradiation. These results provide some insight into the DNA damage and replication checkpoint mechanisms of HRad17.

Cell cycle checkpoint proteins ensure that events of the cell cycle progress in an orderly fashion. Whenever DNA is damaged or incompletely replicated, cell cycle checkpoint proteins act to delay cell cycle progression until the aberrant DNA is repaired or replication is completed. Without this protective delay, cell division would occur in the presence of damaged or unreplicated DNA, which would result in genetic mutations or cell death (1)(2)(3).
Genetic data in yeast and recent biochemical studies in yeast and human suggest that an early step in the DNA damage checkpoint response may involve the activation of Rad3p/ATM/ ATR and Chk1 kinases. Phosphorylation of Chk1 does not occur when checkpoint rad genes are inactive, placing chk1 downstream of rad genes in the cascade of events (16). Furthermore, the Chk1 kinase has been shown to phosphorylate Cdc25c, and it is known that when phosphorylated on Ser 216 , Cdc25c is sequestered, and thus inhibited, by 14-3-3 (17)(18)(19). This scenario reveals a potential link between the checkpoint Rad proteins and Cdk regulation through Cdc25 (20,21).
In the last few months, there have been three independent reports of the cloning of Hrad17, the human homologue of the Rad17 gene from S. pombe (11,22,23). The HRad17 protein has a significant amino acid identity with S. pombe Rad17p and has been demonstrated to interact physically with HRad1 but not with HRad9 (11). More recently, we showed that Hrad17 is highly expressed in human testis and tumor tissues (22), and Li et al. (23) confirmed the cell-cycle checkpoint functionality of Hrad17, using a complementation assay in S. pombe rad17 mutants. In the present work, we investigated the intracellular localization of HRad17, and we report that it predominantly localizes to the nucleolus. To further investigate the possible biological roles of Hrad17, yeast two-hybrid screening was performed to identify potential HRad17 binding partners. We found that HRad17 interacts and colocalizes with NHP2L1, the human homologue of an essential S. cerevisiae nuclear protein, NHP2 (24). In addition, the localization of HRad17 was found to be sensitive to UV irradiation, as the protein then translocated out of the nucleoli and assumed a more diffuse nuclear distribution. These results suggest a functional link between HRad17 and NHP2L1 and implicate the nucleolus as an important site in the DNA damage checkpoint response. vector pACT-2 (CLONTECH). The nucleotide sequence of these constructs was confirmed by automated sequencing.
The human testis matchmaker cDNA library was purchased from CLONTECH. The procedures for yeast two-hybrid screening and elimination of false positives were performed exactly according to the manufacturer's instructions. Briefly, six million yeast transformants were screened, and 20 positive clones were selected based on His3 and lacZ reporter gene expression. After cycloheximide counterselection and yeast mating to eliminate false positives, a total of six clones remained positive. Plasmids of these six positive clones were isolated from yeast strain CG1945 and transformed into Escherichia coli strain KC8. Automated sequencing allowed for the identification of two of these six clones as NHP2L1 and two others as KIP. We then proceeded with yeast cotransformation experiments to confirm the two-hybrid interaction.
pQE and FLAG Expression Constructs-Hrad17 was cut from the pAS-2-Hrad17 construct by BamHI and SalI and then cloned into pQE32 (Qiagen, Hilden, Germany). The ORF region of NHP2L1 was amplified by PCR and cloned into pQE30 (Qiagen). For cloning NHP2L1 and KIP coding sequences in frame into vector pFLAG-CMV2 (Kodak, New Haven, CT), PCR was used to amplify the ORFs corresponding to each gene. The amplified fragments were cloned into the NotI and SalI sites of pFLAG-CMV2. The primers were as follows: for NHP2L1, 5Ј-TATGCGGCCGCGATGACTGAGGCTGATGTGAAT-3Ј and 5Ј-GCGG-TCGACTTAGACTAAGAGCCTTTCAAT-3Ј, and for KIP, 5Ј-AATGCG-GCCGCGATGGGGGGCTCGGGCAGTCGC-3Ј and 5Ј-TGCGTCGACT-CACAGGACAATCTTAAAGGA-3Ј. The nucleotide sequence of these constructs was confirmed by automated DNA sequencing.
Antibodies-Monoclonal antibodies against HRad17 (31E9) were developed by immunizing mice with His 6 -HRad17 and screening the culture supernatants with recombinant HRad17 protein using ELISA. NHP2L1 polyclonal antiserum (R86) was raised in rabbits against two synthetic peptides derived from the amino acid sequences, KQLRKGA-NEATKTLNRG and SQLKQQIQSIQQSIERLLV from NHP2L1. The polyclonal antiserum was further purified by Affi-10 gel (Bio-Rad) column coupled with His-tagged NHP2L1. After extensive washing with 10 mM Tris (pH 7.5) and 10 mM Tris (pH 7.5), 500 mM NaCl, the antibodies were eluted with 100 mM glycine (pH 2.5) and 100 mM triethylamine (pH 11) and dialyzed with PBS for two days.
Transfection Experiments-For cultures used in immunoprecipitation assays, HeLa cells were grown to approximately 60 -80% confluence on 10-cm dishes. 20 g of DNA in 40 l of Superfect reagent (Qiagen) were added to the cells. For cultures used for indirect immunofluorescence detection experiments, HeLa cells were seeded onto glass coverslips, placed in 3-cm dishes, and transfected with 1-2 g of DNA in 5 l of Superfect reagent. Following incubation at 37°C for 2-3 h, the DNA-Superfect complex was replaced with Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. 48 h after transfection, the cells were either fixed for microscopic analysis (immunofluorescence) or lysed, and cell extracts were prepared for immunoprecipitation.
Immunocytochemistry-For indirect immunofluorescence, cells grown on coverslips were fixed in cold methanol (Ϫ20°C) for 10 min, washed in PBS, and incubated with polyclonal R86 antiserum or monoclonal 31E9 antibody at 37°C for 1 h. After rinsing with PBS, antirabbit or anti-mouse antibodies conjugated with rhodamine or fluorescein isothiocyanate (Jackson Immuno Research, West Grove, PA) were applied at 37°C for 30 min. Subsequently, DNA was labeled with 4Ј,6-diamidino-2-phenylindole, and coverslips were mounted in antifade solution (Molecular Probes Inc., Eugene, OR). For double immunofluorescence labeling, both primary or secondary antibodies were mixed and applied together. Confocal images were recorded using a LSM 410 confocal laser scanning microscope (Carl Zeiss, Germany) and printed with the Fujix Pictrography 3000 color printer (Fujifilm, Japan) using Adobe PhotoShop software (Adobe Systems, Mountain View, CA).
For immunogold labeling, HeLa cells were fixed for 30 min with 0.1% glutaraldehyde in Zamboni fixative (2% paraformadehyde and 1:6 dilution of a saturated picric acid solution in 0.1 M PBS). Samples were then cryoprotected and frozen in liquid nitrogen as described (25). Ultrathin cryosections were prepared and incubated with the primary antiserum for 30 min, followed by incubation for 30 min with goat anti-rabbit IgG and gold conjugates of 10 nm gold particle size (10 g/ml). Sections were thoroughly washed with PBS and stained for 2 min in 1% neutral uranyl acetate and then 2 min in 4% unbuffered uranyl acetate. Samples were mounted in a thin film of 1.25% methylcellulose (Fluka Chemical Co., Donkokomo, NY) and examined with a JEOL 100-cx transmission electron microscope (JEOL USA, Peabody, MA).
UV Irradiation-UV doses were delivered with a single pulse using the UV Stratalinker 2400 (Stratagene, La Jolla, CA). Culture medium was removed prior to UV irradiation and replaced immediately after treatment. In most of these experiments, a dose of 40 J/m 2 was used, and the cells were observed after an additional 2 h of culture.

RESULTS
The Subcellular Localization of HRad17-To address the subcellular distribution of HRad17, monoclonal antibodies (mAbs) against HRad17 were generated. The mAb 31E9 was found to recognize a protein of 75 kDa by Western blotting of HeLa total cell lysates (Fig. 1A, lane 1) and immunoprecipitation (lanes 2 and 3). This corresponds to the predicted molecular mass of HRad17. The two major bands below Hrad17 come from the IgG used for immunoprecipitation, as confirmed by an experiment where a control antibody (25G10) was used (not shown). The 31E9 antibody also recognizes a single 75-kDa band in extracts of insect cells expressing HRad17 (data not shown). Immunostaining analysis of HeLa cells with the 31E9 antibody showed that HRad17 localized to the nucleus, exhibiting a punctate pattern with concentration mostly in the nucleoli (Fig. 1B).
Identification of an HRad17 Interacting Protein, NHP2L1-Using the yeast two-hybrid method with HRad17 as the bait protein, approximately six million human testis cDNA library transformants were screened, and a total of six colonies were found that were HIS3 and lacZ positive. Two of them were identified by sequencing as NHP2L1, representing two different mRNA transcripts, 0.7 and 1.3 kilobases, as described previously by others (26). Another two colonies were identified as CIB/KIP (27,28). The specificity of the interaction between KIP/NHP2L1 and HRad17 was further confirmed by cotransformation of HRad17 with NHP2L1 and KIP into yeast strain CG1945. In addition, pACT-2-PCNA and pACT-2-Chk1 were constructed as controls. The result showed that NHP2L1 and KIP can specifically interact with HRad17 using the yeast two-hybrid method ( Fig. 2A).
To investigate whether NHP2L1 and KIP interact with HRad17 in mammalian cells, immunoprecipitation experiments were carried out. HeLa cells were transfected with Nterminal FLAG-tagged NHP2L1 or KIP. Cell lysates were immunoprecipitated with either 31E9 or anti-FLAG mAbs and analyzed by Western blotting with corresponding antibodies. Immunoprecipitation with the 31E9 anti-HRad17 antibody followed by Western blotting with anti-FLAG antibody revealed Hrad17 and NHP2L1 in the same immunocomplex (Fig. 2C). The opposite experiment, immunoprecipitation with anti-FLAG antibody and Western blotting with 31E9, also showed an association between HRad17 and NHP2L1 (Fig. 2D). As was the case for Fig. 1, the two major bands seen in Fig. 2 (C and D) come from the IgG used for immunoprecipitation, as confirmed by a set of experiments where lysates were substituted for lysis buffer alone (not shown). These results demonstrate that HRad17 and NHP2L1 can interact in HeLa cells. However, we were unable to detect co-immunoprecipitation of HRad17 and KIP (Fig. 2, C and D).
Two New Homologues of NHP2L1 in S. cerevisiae and Caenorhabditis elegans-Although NHP2 shares some similarity with NHP2L1, a search of the GenBank TM data base revealed an open reading frame from S. cerevisiae even more closely related to NHP2L1. S. cerevisiae NHP2 and human NHP2L1 have only 33% identity (Fig. 3), but this novel S. cerevisiae ORF (GenBank TM accession number S15037) is 69% identical to human NHP2L1. Therefore, the homology of NHP2L1 to this new gene is likely to be much more relevant than its homology to NHP2. Similarly, we found a C. elegans sequence (Swiss-Prot number Q21568) that is 77% identical to NHP2L1, suggesting that this is a highly conserved protein.
NHP2L1 Is a Nucleolar Protein-Because HRad17 and NHP2L1 could interact in HeLa cells, we set out to examine whether NHP2L1 localized in the nucleolus. First, HeLa cells were transfected with FLAG-tagged NHP2L1. Immunofluorescent labeling with the anti-FLAG antibody showed that NHP2L1 indeed located to the nucleoli (Fig. 4A). Second, a GFP-NHP2L1 fusion construct was made to transfected into HeLa cells. Again, the NHP2L1 green signal was found to reside in the nucleoli (data not shown). Third, a rabbit polyclonal antiserum, R86, was raised against NHP2L1 using two NHP2L1 synthetic peptides, and an affinity-purified polyclonal antibody was prepared. The affinity-purified antibody specifically recognized one band in the nuclear preparation of HeLa cells, whose size corresponds to that of NHP2L1 (Fig. 4E). Staining of HeLa cells with this affinitypurified antibody predominantly localized to the nucleoli, with only fainter, punctate labeling in other nuclear regions (Fig. 4B). Control experiments using preimmunized rabbit serum and synthetic peptides for immunocompetition did not show a nucleolar signal (Fig. 4, C and D). These results demonstrate that NHP2L1 is a nucleolar protein.
NHP2L1 Is Concentrated in the Dense Fibrillar Component of the Nucleolus-To more precisely determine where HRad17 and NHP2L1 resided within the nucleolus, immunogold labeling of ultrathin cryosections prepared from HeLa cells was performed. The anti-HRad17 monoclonal antibody 31E9 was incompatible with our electron microscopy fixation procedures. However, immunogold labeling with polyclonal anti-NHP2L1 antiserum R86 showed that NHP2L1 is largely concentrated in the dense fibrillar component of the nucleolus (Fig. 5). No signal was found in C, Cell lysates prepared as in B were immunoprecipitated with mAb 31E9 and analyzed by Western blot using mAb M2. HRad17 and FLAG-NHP2L1 were found in the same complex. D, cell lysates prepared as in B were immunoprecipitated with mAb M2 and analyzed by Western blot using mAb 31E9. As shown in B, HRad17 and FLAG-NHP2L1 were found in the same complex. IP, antibody used for immunoprecipitation; WB, antibody used for Western blotting. the same region when preimmunized serum was used.
HRad17 Colocalizes with NHP2L1-To test whether HRad17 and NHP2L1 colocalized in the nucleolus, double staining experiments were performed. Using simultaneous labeling of HeLa cells with anti-HRad17 and anti-NHP2L1 antibodies followed by rhodamine and fluorescein isothiocyanateconjugated secondary antibodies, large overlapping areas of staining could be visualized in the nucleoli. These are shown in yellow in the superimposed confocal images (Fig. 6).
Redistribution of HRad17 after UV Irradiation-To investigate whether the localization of HRad17 may be responsive to DNA damage, HeLa cells were irradiated with UV light (40 J/m 2 ) and cultured for another 2 h. Interestingly, HRad17 staining was no longer seen in the nucleoli, although the nucleolar structures could still be detected by phase contrast microscopy. Instead, we observed only the punctate nuclear labeling pattern after UV treatment (Fig. 7B). The total amount of HRad17 protein appeared unchanged (Fig. 7A), however, suggesting that the disappearance of HRad17 nucleolar labeling was caused by protein redistribution rather than degradation. This effect was also observed with lower UV doses, but complete disappearance of Hrad17 nucleolar labeling occurred only after longer post-irradiation incubation periods (e.g. 16 h at 20 J/m 2 ). In control experiments, MPP10, a human U3 small nucleolar ribonucleoprotein component (29) still localized to the nucleolus after UV irradiation (Fig. 7, C and D). Finally, redistribution of HRad17 was also be observed when the cells were treated with methyl methanesulfonate, mitomycin C, adriamycin, or a combination of ␥ irradiation and cisplatinum, 2 suggesting that this translocation of HRad17 out of the nucleolus indeed accompanies DNA damage. In contrast, FIG. 3. Amino acid sequence alignment of human, C. elegans, and S. cerevisiae NHP2L1 proteins with S. cerevisiae NHP2. Alignments were determined using the GCG Prettyplot program. Gaps were introduced to optimize the alignment and shown as dashes. C. elegans and S. cerevisiae amino acid sequences were retrieved from the Swiss-Prot and GenBank TM data bases (accession numbers Q21568 and S15037, respectively). ce, C. elegans; hs, Homo sapiens; sc, S. cerevisiae.  the localization of NHP2L1 remained nucleolar when these DNA damage stimuli were used (data not shown). DISCUSSION We recently isolated an Hrad17 cDNA clone on the basis of its high levels of expression in human testis and tumor tissues (22). In the present work, we used the yeast two-hybrid method to identify interacting proteins. Our screen revealed two HRad17 interacting proteins, NHP2L1 and KIP. KIP is related to the phosphatase calcineurin B and was reported to bind to DNA-dependent protein kinase (28), which is a relative of Rad3p/ATM/ATR kinases, with a C-terminal domain homologous to phosphatidylinositol 3-kinase and is required for V(D)J rearrangements of immunoglobulin genes, recombination, and repair of radiation-induced double-stranded breaks (30). Therefore, KIP may participate in the regulation of DNA doublestrand break repair via the regulation of phosphorylation and dephosphorylation processes (28). Although HRad17 and KIP could clearly interact within the confines of the yeast twohybrid system, we could not detect such an association in human cells by co-immunoprecipitation. Although this does not rule out that an in vivo interaction may occur in human cells, we concentrated our efforts on the other Hrad17-binding protein, NHP2L1.
NHP2L1 is a putative human homologue of the NHP2 gene in S. cerevisiae (26). NHP2 is related to the high mobility group (HMG) proteins based on the contents of basic and acidic amino acids, and is essential in yeast (24). Typically, HMG proteins have a molecular mass that is Ͻ30 kDa and are highly charged. HMG domains can bind to a variety of non-B-DNA structures, such as B-Z DNA junction and platinated DNA (31,32). An RNA binding motif that was hypothesized to deliver additional activity to the ribosome was also found in NHP2 (33). No obvious HMG domain or RNA-binding motif can be identified in human NHP2L1 protein. Two other homologues of human NHP2L1 were identified in C. elegans and S. cerevisiae (Fig. 3). To assess the importance of the new yeast NHP2L1, we used the single step method (34) to create a null mutant of NHP2L1 in S. cerevisiae. After screening more than 200 transformants containing HIS3 or TRP1 marker genes, however, we were unable to obtain a null mutant of NHP2L1 in S. cerevisiae, 2 suggesting that NHP2L1 may be an essential gene in yeast.
HRad17 and NHP2L1 were found to localize to the nucleolus. The nucleolus is the site of rDNA transcription, processing of rDNA transcripts, and formation of preribosomal particles (35). It can be divided into three compartments, the fibrillar centers (FCs), the dense fibrillar component (DFC), and the granular component. Early work showed that the FC is a reservoir for rDNA and RNA polymerase I (36,37). The DFC is thought to be the likely site of rRNA transcription, because nascent Br-UTPlabeled rRNA was detected at the boundary of FC and DFC (38,39). However, another line of evidence attributed the site of rRNA transcription to the FC region (40). Later rRNA processing and preribosome assembly take place in the granular component region (35). Immunogold labeling of NHP2L1 showed that it resides primarily in the DFC region of nucleolus (Fig. 4), suggesting that the function of NHP2L1 may be related to rRNA transcription or processing. We were unable to determine the precise location of Hrad17 within the nucleolus. However, its localization in the nucleolus suggests it may play a role in one or more nucleolar processes.
Recently, it has been reported that nucleolar function may be involved in the aging process. In aging S. cerevisiae cells, the Sir complex, consisting of Sir2p, Sir3p, and Sir4p, is found to relocalize from the telomere to the nucleolus, and the number of the extrachromosomal rDNA circles is dramatically increased (41)(42)(43)(44). In humans, the Werner syndrome is an autosomal recessive disorder characterized by premature aging. The Werner syndrome protein, which encodes an ATP-dependent DNA helicase, was found to localize within the nucleolus in human cells (45). The S. pombe Rad17p protein is required for DNA damage and replication checkpoints. If HRad17 plays a similar checkpoint function, its localization raises the possibility of a heretofore unsuspected role for the nucleolus in the DNA damage response.
DNA damage induced by UV light has been well studied, and a variety of DNA repair mechanisms are triggered to restore the genomic integrity in most organisms (46). Given the rapidity of the cellular response to genome damage one would expect that both the damage sensing mechanisms and the immediate downstream elements of the signal transduction pathway would be constitutively present within the nucleus. Indeed, both ATM and Hrad1, two human proteins with known cell cycle checkpoint functionalities (7,47), display constant protein levels after such DNA damaging stresses as UV or ␥ irradiation or treatment with the radiomimetic drug neocarzinostatin (47,48). Our finding of a stable level of Hrad17 protein upon UV irradiation fits well with these observations and suggests that perhaps the entire apparatus of DNA damage sensing and repair is constitutively present in the nucleus, circumventing the need for time-consuming protein synthesis in the response to DNA damage. However, ATM and Hrad1 display punctate nuclear labeling patterns excluding nucleoli in the basal state, and this pattern did not change when the cells were treated with DNA damaging agents (49). 2 This contrasts with the observations reported here of a strong Hrad17 staining in the nucleoli, in addition to fainter, punctate nuclear labeling. Most striking is the disappearance of Hrad17 nucleolar labeling following DNA damage, apparently because of a redistribution of Hrad17 protein to its extranucleolar pool. The significance of this redistribution remains to be determined, but it may be noteworthy that other proteins involved in DNA repair have been shown to redistribute in response to UV irradiation, including the Xeroderma pigmentosum type G protein and BRCA1 (46,50).
One plausible function of HRad17 in the cell cycle checkpoint process is suggested by its similarity to RF-C components. Human RF-C is a complex composed of five proteins of molecular masses 36, 37, 38, 40, and 140 kDa. During DNA replication, the RF-C complex recognizes primed DNA at the 3Ј-OH of the pre-Okazaki fragment and recruits trimeric PCNA molecules onto DNA in the presence of ATP. Subsequent ATP hydrolysis catalyzed by RF-C is required for polymerase ␦ to join the complex and initiate chain elongation (51). Therefore, HRad17 may interact with the RF-C complex and interfere with its replication activity, causing cell cycle arrest. In fact, we have found HRad17 and RFC140 in the same immunocomplex generated with Hrad17 antibody 31E9 and RFC140 monoclonal antibodies from Dr. Bruce Stillman, Cold Spring Harbor. 2 In summary, we characterized the nuclear localization of the putative checkpoint protein Hrad17 as an overall dotted pattern with prominent staining in the nucleoli. The nucleolar portion of Hrad17 redistributed upon treatment with DNA damaging agents indicating a possible role of this organelle and of Hrad17 in the DNA damage and replication checkpoint pathways.