Phosphorylation of Human Fen1 by Cyclin-Dependent Kinase Modulates Its Role in Replication Forkregulation

Home , Flap endonuclease

Oncogene (2003) 22, 4301–4313 & 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc ORIGINAL PAPERS Phosphorylation of human Fen1 by cyclin-dependent kinase modulates its role in replication forkregulation

Ghislaine Henneke1,2, Ste´ phane Koundrioukoff1,2 and Ulrich Hu¨ bscher*,1

1Institute of Veterinary Biochemistry and Molecular Biology, University of Zu¨rich, Winterthurerstrasse 190, CH-8057 Zu¨rich, Switzerland

Cyclin-dependent kinase (Cdk) Cdk1–Cyclin A can (Morgan, 1996). The G0-G1 transition is regulated phosphorylate Flap endonuclease 1 (Fen1), a key-enzyme by Cdk4/6–Cyclin D and G1-S transition by Cdk2– of the DNA replication machinery, in late S phase. Cdk1– Cyclin E. Cdk2–Cyclin A and Cdk1–Cyclin A are cyclin A forms a complex in vitro and in vivo with Fen1. activated during S phase. Cdk2–Cyclin A activity rises Furthermore, Fen1 phosphorylation is detected in vivo and throughout S phase whereas Cdk1–Cyclin A activity depends upon Cdks activity. As a functional consequence appears later in correlation with the end of DNA of phosphorylation by Cdk1–Cyclin A in vitro, endo- and synthesis. Finally, the G2-M transition is regulated by exonuclease activities of Fen1 are reduced whereas its Cdk1–Cyclin B (Nigg, 1995; Nurse, 2000). The different DNA binding is not affected. Moreover, phosphorylation Cdk–Cyclin complexes display distinct physiological of Fen1 by Cdk1–Cyclin A abrogates its proliferating cell functions. Cdk4–Cyclin D can phosphorylate the nuclear antigen (PCNA) binding thus preventing stimula- retinoblastoma gene product pRb, which sequesters tion of Fen1 by PCNA. Concomitantly, human cells transcription factors, to allow the progression into expressing the S187A mutant defective for Cdk1–Cyclin S phase (Hinds and Weinberg, 1994). In mitosis, the A phosphorylation accumulate in S phase consistent with a nuclear lamina is disassembled by a direct Cdk1–Cyclin failure in cell cycle regulation through DNA replication. B phosphorylation (Nigg, 1995). DNA replication Our results suggest a novel regulatory role of Cdks onto proteins such as large SV40 T antigen (McVey et al., the end of S phase by targeting directly a key enzyme 1993), DNA polymerase a/primase (Voitenleitner et al., involved in DNA replication. 1997; Schub et al., 2001), DNA polymerase d (Wu et al., Oncogene (2003) 22, 4301–4313. doi:10.1038/sj.onc.1206606 1998; Ducoux et al., 2001), replication protein A (Dutta and Stillman, 1992)and DNA ligase I (Rossi et al., Keywords: replication; cell cycle; Fen1; Cdks; phos- 1999)are all substrates for Cdk–Cyclin (reviewed in phorylation Henneke et al., 2003). In addition to Cdks, other protein-modifying factors such as protein phosphatase 2A (PP2A)and Cdc7-Dbf4 kinase (DDK)have essential functions in triggering S phase and initiating DNA replication (Lin et al., 1998; Johnston et al., 1999, Introduction 2000). Since we recently showed that acetylation regulates flap endonuclease 1 (Fen1)activity through Cyclin-dependent kinases (Cdks)are essential activators the transcriptional coactivator p300 (Hasan et al., 2001), of replication initiation and are also required for once- we addressed the question whether Fen1 could be per-cell cycle control of DNA replication (Vas et al., involved in other post-translational modifications. We 2001). Eukaryotic DNA replication initiates at many found three potential sites of phosphorylation by Cdks replication origins so that several replication forks can (Ser-16, Ser-157 and Ser-187)and we tested if phos- work simultaneously. This allows replication of large phorylation might be a novel regulatory mechanism for Fen1. genomes in a short S phase (Stillman, 1996; Donaldson 0 and Blow, 1999). Entry into S phase and cell-cycle 5 exonuclease-1 or Fen1 is a structure-specific progression are tightly regulated by active kinase metallonuclease that plays an important role in multiple complexes composed by a positive regulatory subunit DNA metabolic processes (Lieber, 1997). In DNA called Cyclin and a catalytic subunit, the Cdk subunit replication, it is required for Okazaki fragment matura- (Fotedar and Fotedar, 1995). The pairs of Cdk–Cyclin tion where it participates in the removal of the displaced found in these complexes are Cdk4/6–Cyclin D, Cdk2– RNA–DNA primers (Bambara et al., 1997). In DNA Cyclin E, Cdk2/1–Cyclin A and Cdk1–Cyclin B repair, Fen1 is required for long-patch base excision repair (BER)(Klungland and Lindahl, 1997; Kim et al., 1998). Genetic analyses of the Fen1 homologue RAD27 *Correspondence: U Hubscher; E-mail: [email protected] 2Both authors contributed equally to this work in Saccharomyces cerevisiae have shown that the gene is Received 20 November 2002; revised 17 March 2003; accepted 20 March not essential for viability (Sommers et al., 1995). 2003 However, yeast Fen1-null mutant yeast strains revealed Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4302 a pleiotropic phenotype including temperature-sensitiv- vitro and in vivo with Cyclin A-dependent kinases. ity for growth, sensitivity to methyl methane sulphonate Moreover, human cells expressing the Fen1 S187A (MMS)and UV radiation (Reagan et al., 1995). mutant defective for phosphorylation by Cdk1–Cyclin Elevated rates of mitotic recombination because of the A accumulate in S phase consistent with a failure in cell presence of destabilized repetitive sequences (Gordenin cycle regulation through DNA replication. Together, et al., 1997; Henricksen et al., 2000). Moreover, Fen1 our results suggest a novel role for Cdks onto the has been proposed to be involved in telomere stability regulation of replication fork functionality through cell (Parenteau and Wellinger, 1999)and nonhomologous cycle phosphorylation of Fen1, a key enzyme in DNA end-joining (NHEJ)(Wu et al., 1999). Recently, Fen1 replication. haploinsufficiency has been demonstrated to lead to rapid tumour progression in mice (Kucherlapati et al., 2002). As chicken cells lacking Fen1 were viable but Results hypersensitive to methylating agents and H2O2, it was suggested that Fen1 functions in BER in vertebrate cells Fen1 is phosphorylated in vitro by Cdk–Cyclin complexes (Matsuzaki et al., 2002). In summary, Fen1 is a multifunctional enzyme in vivo for maintaining genome Since DNA replication proteins are targets of Cyclin- stability. dependent kinases, we investigated whether Cdk–Cyclin As the role of Fen1 in DNA replication and repair is complexes involved in S phase could phosphorylate becoming more clear, several groups are interested to Fen1. Purified Cdk1–Cyclin A, Cdk2–Cyclin A and structurally characterize this enzyme, and the crystal Cdk2–Cyclin E were incubated with 100 and 500 ng structures of two archaeal orthologous Fen1 have so far Fen1 into an ATP-dependent kinase assay. All three been solved (Hosfield et al., 1998; Hwang et al., 1998). Cdk–Cyclin complexes were able to phosphorylate Fen1 For DNA repair and replication, Fen1’s structural as well as the positive control histone H1 (Figure 1a). recognition of the 50 flap is essential. Moreover, we Furthermore, the kinase inhibitor p21Cipl,WAF-1 prevented recently have described that the flexible loop of human phosphorylation at a p21/Fen1 ratio of 1 : 1 suggesting Fen1 is required for flap cleavage activity (Storici et al., that phosphorylation was specific to Cdk kinases 2002). The current model predicts that Fen1 tracks (Figure 1b). From these experiments, we concluded that down the 50 flap until it reaches duplex DNA (Murante Fen1 is a target for Cdk phosphorylation in vitro. et al., 1995)and releases the unannealed single-stranded region at the junction region by endonucleolytic Fen1 is phosphorylated in vivo and its phosphorylation cleavage (Harrington and Lieber, 1994a, b). Fen1 does depends on Cdks activity not cleave bubble substrates, single-stranded 30 flaps, heterologous loop, or holiday junctions whereas DNA To test whether Fen1 is phophorylated in vivo, 293 T substrates containing nicks or small gaps can be cells were transfected with a vector expressing myc- processed by the 50-30 exonuclease activity. Although tagged Fen1. After nuclear extracts preparation, myc- Fen1 acts most efficiently as a flap endonuclease, it is tagged Fen1 protein was immunoprecipitated using the likely that similar enzymatic mechanisms are involved anti-myc-specific antibody and subsequently analysed for its exonuclease activity (Lieber, 1997). Fen1 interacts by Western blot using the anti-Fen 1 antibody. After with several proteins at the replication fork, such as stripping, the same membrane was reprobed with a PCNA (Li et al., 1995; Wu et al., 1996; Jo´ nsson et al., specific antiphosphorylated serine antibody. Western 1998), Dna2 helicase/endonuclease (Budd and Camp- blots analysis clearly showed a band migrating at the bell, 1997), Werner syndrome helicase/exonuclease same position of myc-tagged Fen1 (Figure 2a). No (WRN)(Brosh Jr et al., 2001)and replication protein signal was observed in the negative control. To see A (Biswas et al., 1997). The Dna2 helicase/endonuclease whether phosphorylation of Fen1 was linked to Cdks can physically and genetically interact with RAD27 in activity, myc-tagged Fen1 was expressed in 293T cells budding yeast (Budd and Campbell, 1997), and both treated during 24 h with 40 mm Olomoucine (a specific play a direct role in Okazaki fragment maturation (Bae inhibitor of Cdk1 and Cdk2)(Abraham et al., 1995) and Seo, 2000; Kang et al., 2000). Moreover, since dissolved in DMSO. Cells treated with DMSO alone WRN protein and Fen1 physically interact, an alter- were used as a negative control. Myc-tagged Fen1 was native pathway during DNA replication and/or repair then immunoprecipitated from nuclear extract using the has been proposed (Brosh Jr et al., 2001). The anti-myc antibody. Immunoprecipitations including the interaction of PCNA with Fen1 results in a stimulation DMSO control and Olomoucine treatment were loaded of its endonuclease activity in vitro under physiological three times. Two Western blots were performed, one ionic conditions (Wu et al., 1996). against Fen1 and another against proteins phosphory- In this study, we report that exo- and endonuclease lated on serine. The other part of the samples was used activities of Fen1 are inhibited in vitro upon phosphor- for a silver staining (Figure 2b). The immunoblot using ylation by the Cdk1–Cyclin A complex. Phosphoryla- Fen1 antibody showed that Fen1 was specifically tion of Fen1 abolishes its PCNA binding thus isolated from each extract DMSO and Olomoucine- preventing stimulation of Fen1 by PCNA. Fen1 is treated cells. When the antiphosphorylated serine antiphosphorylated in vivo and phosphorylation is reduced body was used, the signal at the position of myc-tagged by a specific Cdks inhibitor. Fen1 can form a complex in Fen1 was more than 10 times reduced after Olomoucine

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4303

Figure 1 Fen1 is speciﬁcally phosphorylated in vitro by Cdk–Cyclin. (a)1.6 pmol of Cdk2–Cyclin E, Cdk1–Cyclin A and Cdk2–Cyclin A were used with histone H1 (H1)or Fen1 in a kinase assay as outlined in Materials and methods. ( b)11.5 pmol of Fen1 was incubated with 1.6 pmol of Cdk2–Cyclin E, Cdk1–Cyclin A, Cdk2–Cyclin A and increasing amount of p21 in a kinase assay

treatment compared to DMSO (see middle blot in position of myc-tagged Fen1 increased fivefold when the Figure 2b). This result was not because of a difference in cells were in the late S phase compared to the cells in the amount of myc-Fen1 immunoprecipitated as ob- S phase. Moreover, the assumption that the phosphory- served from the silver-stained gel. Taken together these lated form of Fen1 in S phase is likely because of Cdk results indicated that Fen1 is phosphorylated in vivo by activity (Figure 3b)is confirmed by the abolishment of Cdks and that phosphorylation is 10 times reduced upon its phosphorylation status upon Olomoucine treatment, treatment with the Cdks-specific inhibitor Olomoucine. the specific inhibitor of Cdk1 and Cdk2 (Figure 2b). Thus, Fen1 phosphorylation is likely cell cycle regulated Fen1 phosphorylation increases throughout S phase and increased when the cells progress into S phase. Entry into S phase and cell cycle progression are tightly Fen1 binds Cdk1, Cdk2 and Cyclin A in vitro and in vivo regulated by Cdks. The G1-S transition is regulated by Cdk2–Cyclin E. Cdk2–Cyclin A and Cdk1–Cyclin A are Given the fact that Fen1 is a DNA replication protein activated during S phase. Cdk1–Cyclin A activity rises involved in Okazaki fragment processing and a potential throughout S phase whereas Cdk1–Cyclin A activity target for Cdks phosphorylation, we next tested a direct appears later in correlation with the end of DNA interaction between Cdk1, Cdk2 and Cyclin A. More- synthesis (Nigg, 1995; Nurse, 2000). Consequently, we over, Fen1 contains two putative Cyclin–Cdk binding investigated Fen1 phosphorylation during the cell cycle motifs. Indeed, the core sequence ZRXL, where Z and X and more precisely during S phase. 293T cells expressing are conserved basic residues, is present at position 191 transiently a myc-tagged Fen1 were blocked at the G1/ and 260 (Adams et al., 1996). We established pull-down S phase transition upon treatment by hydroxyurea assays using GST–Cdk1, GST–Cdk2 and GST–Cyclin (4 mm). The cells were then released for 3 h to reach A fusion proteins. Purified recombinant Fen1 was S phase, as observed from the flow cytometry analysis incubated with the GST fusion proteins immobilized (Figure 3a). In parallel, cells were released for 6 h and onto glutathione–sepharose beads. Bound Fen1 was were then distributed in late S and G2 phases. Myc- detected by immunoblotting using the anti-Fen1 anti- tagged Fen1 immunoprecipitated from nuclear extract body. Figure 4a shows the interaction of Fen1 with was used for two Western blots, one against Fen1 and Cdk1, Cdk2 and Cyclin A. PCNA was used as a positive another against proteins phosphorylated on serine control, and the DNA binding region of RF-Cp140 (Figure 3b). The immunoblot using Fen1 antibody (RF-C A)and GST alone as negative controls. To showed that Fen1 was specifically isolated. Using the examine whether endogenous Cyclin A-dependent ki- antiphosphorylated serine antibody, the signal at the nases complexes and Fen1 interact in vivo, coimmuno-

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4304

Figure 2 Fen1 is phosphorylated in vivo and its phosphorylation depends on Cdks activity. (a)Myc-tagged Fen1 transiently expressed in 293T cells was immunoprecipitated from nuclear extracts with an anti-myc antibody as outlined in Materials and methods. Transfection with the empty vector was used as a negative control. The left panel shows an immunoblot against Fen1. The right panel shows the same blot probed with an antiphosphorylated serine antibody. (b)Myc-Fen1 was transiently expressed in 293T cells treated with Olomoucine. Cells treated with DMSO were used as a control. The same immunoprecipitations were loaded for two immunoblots against either Fen1 or the phosphorylated serine and were used for silver staining. The negative control (À)was an immunoprecipitation with the mouse antibody IgG1. The input corresponded to 25 ng recombinant Fen1 (expressed in E. coli)

precipitations from the nuclear extract of asynchronous Cdk1 and Cdk2 and the latter three form a tight HeLa cells were performed using the anti-Cyclin A complex with Fen1 in the HeLa nucleus. antibody. Immunoprecipitated complexes were analysed by immunoblotting with anti-Fen1, anti-Cyclin A, anti- Cdk1, anti-Cdk2 and anti-PCNA antibodies. The anti- Phosphorylation of Fen1 by Cdk1–Cyclin A does not Cyclin A antibody was able to coprecipitate Fen1 as well affect its DNA binding but reduces its endonuclease and as Cdk1, Cdk2, Cyclin A and PCNA (Figure 4b). The exonuclease activities Fen1–Cyclin A-dependent kinases complexes were also In DNA replication, Fen1’s structural recognition of a confirmed when the immunoprecipitation was done with 50 flap is essential. The current model predicts that (i) anti-Fen1 antibody (data not shown). These results Fen1 tracks down the 50 flap until it reaches duplex suggested that Fen1 interacts directly with Cyclin A, DNA (Murante et al., 1995)and then (ii)releases the

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4305

Figure 3 Fen1 phosphorylation increases at the end of S phase. (a)Flow cytometry analysis of cell cycle synchronization. ( b) Immunoprecipitations performed with anti-myc antibody were loaded for two immunoblots: one against Fen1 and one against phosphorylated serine. The negative control (À)was an immunoprecipitation with mouse antibody IgG1. A measure at 10 mgof nuclear extract was loaded unannealed single-stranded region at the junction region Having shown that the phosphorylated form of Fen1 by endonucleolytic cleavage (Harrington and Lieber, can be retained onto a flap template, its functional 1994a, b). Considering the two-steps Fen1 mechanism, consequence was next tested. At low-salt concentrations, we first analysed whether phosphorylation of Fen1 Fen1 displays nuclease activity in the absence of its could influence its DNA binding ability. For this study, accessory protein PCNA (Jo´ nsson et al., 1998). Phos- a DNA magnetic beads assay was used. In a preliminary phorylated and unphosphorylated Fen1 were tested in a experiment, we verified that 1.6 pmol of Cdk1–Cyclin A flap cleavage assay including three annealed oligonu- was sufficient to completely phosphorylate up to 11.5 cleotides that form a radioactively labelled flap structure pmol of Fen1 and this phosphorylation was ATP- as a substrate (Figure 5b). Fen1 was first preincubated dependent (data not shown). Therefore, we decided to with Cdk1–Cyclin A (either in the presence or absence use these optimized saturating conditions for all the of ATP)and incubated with the flap substrate. Titra- following experiments to study the effect of phosphor- tions of unphosphorylated (without ATP)and phos- ylation. Then, phosphorylated (Figure 5a and d, lane 1) phorylated (with ATP)Fen1 were performed in order to and unphosphorylated (Figure 5a and d, lanes 2 and 4) quantify the effect of Fen1 phosphorylation (Figure 5b). Fen1 were incubated with nicked (Figure 5d)or flap These results indicated that phosphorylation of Fen1 (Figure 5a)DNA beads. DNA-bound proteins were reduces its endonuclease activity. Addition of Cdk1– detected by Western blot analysis using anti-Fen1 Cyclin A and ATP alone had no effect on Fen1 cleavage antibody indicating a stable DNA–protein interaction. (data not shown)confirming that phosphorylation is To rule out the possibility that DNA binding could be directly involved in reducing Fen1 endonuclease activ- mediated through Cdk1–Cyclin A (Figure 5a and d, lane ity. Indeed, 2.8 ng of phosphorylated Fen1 hydrolysed 4)or ATP (Figure 5a and d, lane 2),flap and nicked 22% of the flap DNA in 15 min instead of 70% with the DNA beads were incubated with Cdk1–Cyclin A and unphosphorylated enzyme (Figure 5c). These data ATP separately. These results revealed that Fen1 DNA indicated a threefold inhibition of endonuclease activity binding was not affected by Cdk1–Cyclin A or ATP. of the phosphorylated Fen1. Since Fen1 is active as a Thus, we concluded that phosphorylation of Fen1 does 50-30 exonuclease at nicks in duplex DNA with lower not influence its DNA binding. efficiency (Lieber, 1997), we also examined whether

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4306 GST–PCNA fusion protein immobilized onto glutathione-sepharose beads. GST was used as a negative control. Bound Fen1 was detected by immunoblot using the anti-Fen1 antibody. As expected, unphosphorylated Fen1 (Figure 6a, lanes 4 and 5)was able to bind PCNA, but on the other hand phosphorylated Fen1 could not bind PCNA (Figure 6a, lane 6). An identical result was obtained when Fen1 was phosphorylated by Cdk2– Cyclin A (data not shown). Fen1–PCNA interaction was not affected when Cdk1–Cyclin A was incubated alone with Fen1 in the absence of ATP (Figure 6a, lane 4)ruling out the possibility of a binding interference because of the presence of the Cdk1–Cyclin A complex. In addition, Figure 6a clearly showed nearly identical levels of Fen1 in lanes 4 and 5, thus excluding any competition among the proteins. These data suggested that phosphorylation of Fen1 by Cdk1–Cyclin A prevents its PCNA binding. Kinetic analysis revealed that PCNA enhances Fen1 binding stability, thus increasing the cleavage efficiency (Wu et al., 1996; Tom et al., 2000). To efficiently Figure 4 Fen1 binds Cdk1, Cdk2 and Cyclin A in vitro and in vivo. stimulate Fen1 activity, the PCNA trimer must encircle (a)2mg GST–Cdk1 fusion protein, GST–Cdk2, GST–Cyclin A and the DNA and must be ‘below’ the flap (Jo´ nsson et al., control GST–proteins were bound on beads and incubated with 1998). It has been shown that PCNA can stimulate Fen1 100 ng Fen1. Interacting Fen1 was detected by Western blot using activity under physiological salt concentrations but not anti-Fen 1 antibody. GST and GST-RF-C A were used as negative controls and GST–PCNA as a positive control. The input under low-salt concentrations (Li et al., 1995). When corresponded to 50 ng recombinant Fen1. (b)Immunoprecipita- PCNA-dependent Fen1 activity was determined at tions with anti-Cyclin A antibody were performed using HeLa cells 125 mm NaCl, Fen1 alone was virtually inactive as nuclear extracts. Immunoprecipitated proteins were tested by expected (Jo´ nsson et al., 1998). Fen1 was preincubated immunoblotting with anti-Fen 1, anti-Cyclin A, anti-Cdk1, anti- with Cdk1–Cyclin A (either in the presence or absence Cdk2 and anti-PCNA. The negative control (À)was an immunoprecipitation with rabbit IgG (VECTOR laboratories). The input of ATP)and then incubated with the linear flap DNA corresponded to 20 mg of nuclear extract substrate. A limiting amount of Fen1 was used (6.72 fmol)such that the flap cleavage was not detectable in the absence of PCNA (Figure 6b and c, lane 1). phosphorylation could affect its exonuclease activity. Addition of PCNA that can be loaded spontaneously Phosphorylated and unphosphorylated Fen1 were ti- onto the flap substrate stimulated as expected, the trated onto a linear duplex DNA containing a nick nuclease activity of unphosphorylated Fen1 (Figure 6b (Figure 5e). Fen1 digests the upstream oligonucleotide and c, lanes 2–4). Incubation of Cdk1–Cyclin A and exonucleolytically in the 50-30 direction, thus releasing ATP alone had no effect on the Fen1 flap cleavage a labelled 50 mononucleotide, which can be resolved and ruling out an interference because of the presence of the quantified (as outlined in Materials and methods). The Cdk1–Cyclin A complex. In addition, our previous data showed that phosphorylation of Fen1 corresponds studies failed to identify PCNA being phosphorylated to a threefold inhibition of the Fen1 exonuclease activity by Cyclin/Cdks (Koundrioukoff et al., 2000), thus (Figure 5f)and this effect was not because of the excluding the contribution of PCNA to the changes in presence of Cdk1–Cyclin A and ATP (data not shown). Fen1–PCNA affinity. Interestingly, PCNA stimulation In summary, our results demonstrated that phosphor- with phosphorylated form of Fen1 could not be detected ylation of Fen1 by Cdk1–Cyclin A reduces its exo- and even in the presence of a large stoichiometric excess of endonuclease activities about threefold. PCNA (2000 fmol)(Figure 6d).These data clearly demonstrated that upon phosphorylation by Cdk1– Phosphorylation of Fen1 by Cdk1–Cyclin A prevents its Cyclin A, Fen1 cannot be stimulated by PCNA any PCNA binding and consequently abolishing PCNA more, which is consistent with its inability to bind PCNA. stimulation Fen1 S187A mutant defective in Cdk1–Cyclin A Fen1 interacts with several proteins at the replication phosphorylation is enzymatically active and colocalizes ´ et al fork including PCNA (Jonsson ., 1998)and it binds with PCNA in the S phase to PCNA (Li et al., 1995)via the PCNA interaction domain called PIP-box motif (Warbrick et al., 1997). To Three potential Cdks consensus phosphorylation sites test whether phosphorylation of Fen1 could affect its were identified in human Fen1 (S16, S157 and S187). PCNA binding, we used pull-down assays (Figure 6a). Next, we performed site-directed mutagenesis on these Purified recombinant Fen1 was phosphorylated in vitro three serines that were substituted into alanines. The by Cdk1–Cyclin A and subsequently incubated with the three mutants were produced in Escherichia coli and

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4307

Figure 5 Phosphorylation of Fen1 by Cdk1–Cyclin A does not affect its DNA binding but reduces its endonuclease and exonuclease activities. A schematic view of the flap (left)and nicked (right)molecules is represented on top. ( a, d)DNA binding of phosphorylated and unphosphorylated Fen1 onto flap and nicked DNA, respectively. DNA binding was measured by using a biotinylated 284 nt construct that was immobilized onto streptavidin-magnetic beads as outlined in Materials and methods. (b)Effect of Fen1 phosphorylation by Cdk1–Cyclin A was analysed by titration (0.28, 0.575, 1.15 and 2.8 ng)into a flap endonuclease assay. ( c) Quantification of flap hydrolyzed products by phosphoimaging The symbols represent: : Kinase reaction in the presence of ATP; m: Kinase reaction in the absence of ATP. (e)Effect of Fen1 phosphorylation by Cdk1–Cyclin A was analysed by titration (0.28, 2.8 and 28 ng)into an exonuclease assay. ( f)Quantification of mononucleotides by phosphoimaging. The symbols are as in ( c)

each tested in a Cdk1–Cyclin A kinase assay (Figure 7a). cells (Figure 7d). Taken together, these results demon- The results indicated that S187 is the only target for strated that the Fen1 S187A mutant is fully active and Cdk1–Cyclin A and confirmed the specificity of Fen1 shows the same subcellular localization as the WT Fen1. phosphorylation by Cdk1–Cyclin A. Moreover, the Fen1 S187A mutant was enzymatically active in a flap Expression of the Fen1 S187A mutant into 293 cells endonuclease assay (Figure 7b), and incubation of the delays the S phase Fen1 S187A into a Cdk1–cyclin A-dependent kinase assay did not affect its endonuclease activity (Figure 7c). The idea that Fen1 phosphorylation increases at the end This control experiment confirmed that the S187A of S phase was tested by carrying out a kinetic analysis mutant could not be phosphorylated by Cdk1–Cyclin on synchronized 293 cells and evaluating the amount of A, therefore strengthening the inhibition of the wild- Fen1 immunoprecipitated as well as its phosphorylation type (WT)Fen1 activity by phosphorylation on S187. status. Compared to the Fen1 S187A mutant, immuno- Besides, the Fen1 S187A mutant colocalized with PCNA precipitation of WT Fen1 results in the similar level of during S phase when it is ectopically expressed in HeLa Fen1 at times 0 and 4 h (Figure 8a, anti-Fen1). The

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4308 in Materials and methods with the WT, the S187A Fen1 mutant and vector alone as negative control (data not shown). The S187A Fen1 mutant caused a two– to threefold accumulation of cells in S phase compared to the WT (Figure 8b). Interestingly, this ﬁnding was observed both for unsynchronized as well as synchronized 293 cells in S phase after a hydroxyurea block followed by a release at 105 and 210 min, suggesting a delay in S phase for the cells expressing the S187A Fen1 mutant.

Discussion

The findings in this paper give evidence that Fen1 is a substrate for Cdks. Fen1 can interact directly with Cyclin A, Cdk1 and Cdk2 in vitro. The interactions were confirmed in vivo while Cyclin-A-dependent kinases form a complex with Fen1 in the HeLa nucleus. Furthermore, Fen1 phosphorylation by Cdk2–Cyclin A and Cdk1–Cyclin A is detected in vitro and depends upon Cdks activity in vivo. The phosphorylation state of Fen1 appears to be cell cycle regulated and increases when the cells progress into S phase. As a functional consequence of phosphorylation, Fen1 exo- and endonuclease activities are reduced in vitro by Cdk1– Figure 6 Phosphorylation of Fen1 by Cdk1–Cyclin A prevents its Cyclin A. The inhibitory effect of Fen1 phosphorylation PCNA binding and consequently abolishing PCNA stimulation. (a) on its enzymatic activity is likely because of a local GST pull-down experiments were performed with phosphorylated and unphosphorylated Fen1 by testing GST (left)and GST-PCNA conformational change rather than a defect in DNA (right). (b, c)Fen1 (0.28 ng)was incubated with ATP ( b)or Cdk1– binding as no difference in ability to bind DNA could be Cyclin A (c), respectively, and included into a PCNA-dependent detected. The primary structure of Fen1 shows three flap cutting assay. Increasing amounts 0, 50, 100 and 200 ng of potential phosphorylation sites by Cdks (S/T P motif): PCNA were tested. (d)Phosphorylated Fen1 (incubation with ATP serines 17, 157 and 187. We identified the serine 187 and Cdk1–Cyclin A)was titrated into a PCNA-dependent flap cutting assay. The flap hydrolyzed products were analysed using a within Fen1 as the only one residue that could be phosphoimager phosphorylated by Cdk1–cyclin A in vitro. It is located into the internal nuclease domain of Fen1. Despite the lack of structural data about this domain, we propose vector alone was always included as a control (data not that phosphorylation of the serine 187 enhances a local shown). The apparent site of the inhibitory Cdk1–Cyclin conformational change that possibly affects the nuclease A phosphorylation in vitro (Figure 7a)was also activity of Fen1. Fen1 endonuclease activity was also observed with the Fen1 S187A in vivo. The Fen1 tested under physiological salt conditions, where enzy- S187A mutant was not phosphorylated in vivo matic activity is completely PCNA-dependent (Jo´ nsson (Figure 8a, compare S187A mutant to the WT at time et al., 1998). The phosphorylated Fen1 could not be 4 h). Furthermore, the WT Fen1 showed an increase of stimulated by PCNA and this correlated with the phosphorylation by the end of S phase (Figure 8a, abolishment of its PCNA binding. This result suggested antiphosphoserine, WT Fen1 panel at time 4 h)in that phosphorylation could also have an effect on the correlation with the expression of Cyclin A (data not folding of the PCNA-binding region. The current in shown). The level of pull-down PCNA was, as expected vitro evidence indicates that the phosphorylation status less efficient at time 4 h (Figure 8a, WT Fen1 panel). of Fen1 by Cdk1–Cyclin A might downregulate the This was likely because of the phosphorylation of WT replication fork functionality by directing the cells to Fen1 in vivo in correlation with its abolishment of exit S phase. Along this line, we observed accumulation PCNA binding after phosphorylation (see also of cells in S phase upon expression of the serine 187 Figure 6a). Fen1 mutant. The level of phosphorylated Fen1 Since phosphorylation by Cdk1–Cyclin A can prevent increased at the end of S phase in correlation with the PCNA–Fen1 interaction and thus inactivating Fen1 expression of Cyclin A and with a lack of interacting nuclease activity at the end of S phase, we also analysed PCNA (Figure 8a). This result is in agreement with the the physiological relevance of the S187A Fen1 mutant abolishment of the interaction between the phosphory- by transient expression in 293 cells, taking into account lated Fen1 and PCNA observed in vitro (Figure 6a). that this mutant was enzymatically fully active in vitro Under such circumstances, one would expect a dramatic (Figure 7b). Transfected cells were analysed as outlined decrease of pulled-down PCNA with the phosphory-

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4309 lated WT Fen1. The moderate decrease of PCNA endonucleolytically cleave (Harrington and Lieber, coimmunoprecipitated (Figure 8a)is likely because of 1994a; Murante et al., 1995; Wu et al., 1996). The the fact that the pool of myc-tagged Fen1 overexpressed presence of PCNA at cleavage site trapped Fen1 at the was not completely phosphorylated. junction and stimulates its enzymatic endonuclease Our results suggest a novel regulatory role of Cdk by activity (Tom et al., 2000), thus removing the entire targeting a key enzyme involved in DNA synthesis. RNA primer. Then, a ligatable nick is created and sealed Taken together, the data presented lead us to propose by the DNA ligase I. Upon phosphorylation by Cdk1– the following model: When DNA synthesis of the cyclin A occurring at the end of S phase, the remaining upstream Okazaki fragment is completed, DNA poly- Fen1 molecules would then dissociate from the DNA merase d displaces both RNA and DNA of the replication site as a result of its inability to interact with downstream fragment, generating an unannealed 50- PCNA and not to be further endonucleolytically flap. Fen1 nuclease prefers a flap substrate, removing stimulated. Then, free PCNA molecules are available the flap at the point of annealing. Fen1 has been to interact with DNA ligase I for sealing the resulting postulated to enter the flap from the 50 unannealed 50- nick substrate. This model would provide a potential end, to track the annealing point, and then to link for the cell cycle control of DNA replication. It

Figure 7 The Fen1 S187A mutant defective for Cdk1–Cyclin A phosphorylation is enzymatically active and colocalizes with PCNA in S phase. (a)WT Fen1 and S16A, S157A, S187A mutants have been tested in Cdk1–Cyclin A kinase assays. ( b)Activity of the WT Fen1 and the S187A Fen1 mutant were analysed by titration (0.25, 2.5, 25 and 250 ng)into a flap endonuclease assay. ( c)Endonuclease activity of the S187A Fen1 was analysed by titration (0.28, 2.8 and 28 ng)after incubation into a Cdk1–Cyclin A-dependent kinase assay (as outlined in Figure 5b, c).( d) Subcellular localization of ectopically expressed myc-Fen1 and native PCNA by immunofluorescence. Myc-Fen1 was visualized in red by using a polyclonal antibody (A-14)and PCNA labelled in green in the same cells using the monoclonal antibody PC 10. DAPI staining in blue located all nuclei. The merge of the PCNA and myc-Fen1 signals in yellow allows identification of positive cells for both markers

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4310 protein A (Dutta and Stillman, 1992), DNA polymerase a/primase (Voitenleitner et al., 1997; Schub et al., 2001) and DNA ligase I (Rossi et al., 1999)have been described. The modification of each protein upon phosphorylation by an active kinase in S phase leads to a negative effect on their function. This observation leaves open the possibility that a synergistic downregulation of replication proteins appears to be neces- sary to terminate DNA replication. Our hypothesis is strongly supported by a very recent finding showing a cell cycle-dependent dynamic association of Cyclin- Cdks complexes with DNA replication proteins (Frouin et al., 2002). Together, these observations would support the hypothesis that cell cycle phosphorylation of Fen1 participates to the downregulation of DNA replication in mammalian cells. Although phosphorylation is required to downregulate DNA replication, we can not exclude that addi- tional post-translational modifications (glycosylation, methylation, sumoylation, ubiquitination, acetylation, polyglutamylation)happen into other DNA metabolic events. This may account for an accurate regulation by protein modifications. The relative importance of basic mechanistic strategies involved at different times during cell cycle seems to be a prerequisite for genome maintenance in human cells.

Materials and methods

Nucleic acids Oligonucleotides for preparing flap and nicked substrates were purchased from Microsynth GmbH (Balgach, Switzerland). Oligomer sequences are listed in Table I (see supplementary information). Prior to annealing, oligonucleotides Ft2_01 and Ft2_05 were labelled at the 50 end using [g-32P] ATP and T4 polynucleotide kinase. Free ATP was removed on a Micro- Spint G-25 column. To generate the substrates, the appro- priate primers were mixed at a 1 : 1 molar ratio in 20 mm Tris- Figure 7 Continued HCl (pH 7.4), 150 mm NaCl, heated to 751C, and slowly cooled to room temperature. The nicked and flap substrates contain, respectively, downstream primers Ft02_05 and Ft02_01 and suggests how Cdks might control the replication fork the upstream primer Ft02_02 annealed to the complementary functionality by targeting a key enzyme in DNA Ft02_04 oligonucleotide. To generate human GST–Cdk1 fusion protein, cDNA encoding WT Cdk1 was PCR-amplified replication. from the plasmid pCDNA3.1/Cdc2-HA (gift from JP Eukaryotic DNA replication is tightly controlled and Magyar). The resulting fragment was cloned into a pGEX- occurs during a restricted period of the cell cycle, the 3X and sequenced to confirm the absence of mutations. The S phase. S phase entry as well as cell cycle progression in primers used were 50-GGAATTGCCCTTGGATCCATG- general is enhanced by the activity of the conserved GAAGATTATACC-30 and 50-CTGGGACGTCCTCGAGT- serine/threonine protein kinases, the Cdks. Despite CACATCTTCTTAATCTG-30. The pET23/C-His-Fen1 was detailed knowledge of general cellular Cdk functions, used as a template for mutagenesis to generate Fen1 muations relatively little is known about their physiological using the following nucleotides: S16A, pGCTGATGTGG- substrates and how Cdks trigger the end of DNA CCCCCGCTGCCATCCGGGAGAATG; S157A, pCTTAT- replication. Our data suggest that cell cycle-dependent CTTGATGCACCCGCTGAGGCAGAGGCCAGC; S187A, pCTGCCTCACCTTCGGCGCCCCTGTGCTAATGCG. phosphorylation is a novel regulatory mechanism for Fen1 by directing cells to exit from S phase and demonstrate that Cdks enhance the end of the DNA Proteins replication by inactivating a protein involved in DNA Human PCNA was produced in E. coli using the plasmid pT7/ replication fork progression. Cell cycle regulation of hPCNA (gift from A Dutta)and purified to homogeneity as other DNA replication proteins such as DNA polymer- described (Schurtenberger et al., 1998). Human replication ase d (Wu et al., 1998; Ducoux et al., 2001), replication protein A was overexpressed in E. coli strain BL21(DE3)

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4311

Figure 8 Expression of S187A mutant into 293 cells delays the S phase. (a)Kinetic analysis on synchronized 293 cells transfected either with S187A mutant (right panel)or WT Fen1 (left panel).Immunoprecipitation of Fen1 and its phosphorylation status were measured by Western blot with the anti-Fen 1 and antiphosphoserine, respectively. PCNA pull-down with the immunoprecipitated Fen1 was detected by the anti-PCNA. (b)Expression of the myc-tagged Fen1 S187A mutant and WT Fen1 into 293 cells. Only transfected cells were analysed and the percentage of cells in G1, S and G2/M stages of the cell cycle was determined by flow cytometry analysis as outlined in Materials and methods harbouring the expression plasmid p11d-tRP-A and purified radiography and quantified on a PhosphorImager using the according to the methods of Henricksen et al. (1994). Human Image-Quant software (Molecular Dynamics). Fen1 was produced in E. coli BL21(DE3)pLysS using the plasmid pET23d and purified as described (Stucki et al., 2001). The purified Cdk–Cyclin complexes were a kind gift from HP DNA binding assays Nasheuer. A biotinylated 284 fragment homologous to M13mp18 DNA from 6140 to 6423 bp was PCR amplified with upstream biotinylated primer BP-6140 (30 nt)and dowstream primer Fen1 assays BP-6423 (25 nt). Biotinylated 284 nt was immobilized onto Endonuclease and exonuclease assays were performed in a streptavidin-magnetic beads (Dynabeads)in buffer A (10 m m final volume of 12.5 ml containing 40 mm Tris-HCl (pH 7.5), Tris-HCl (pH 7.5), 1 mm EDTA, 2 M NaCl), for 2 h 30 min at 10 mm MgCl2,5mm dithiothreitol (DTT), 200 mg/ml bovine room temperature. Unbound products were washed off the serum albumin (BSA), 50 fmol of DNA substrates. After the beads twice with buffer B (10 mm Tris-HCl (pH 7.5), 1 mm addition of phosphorylated or unphosphorylated Fen1, the EDTA, 1 m NaCl), and the duplex DNA was denaturated with reactions were incubated at 301C for 15 min and stopped with 0.1 m NaOH. Nonbiotinylated nucleotides were removed with 2.5 Â stop buffer (95% formamid, 20 mm EDTA, 0.05% (w/v) buffer B. The beads-bound biotinylated 284 fragment was each bromophenol blue and xylene cyanol). For PCNA annealed with D-45 and F-31 or C-30 and F-31 primers (see stimulation, the assays were performed in 13 ml containing Table I)to create the flap and nicked templates, respectively. 50 mm Bis-Tris (pH 6.5), 125 mm NaCl, 10 mm MgCl2,1mm DNA beads bound templates were washed with buffer B and DTT, 0.25 mg/ml BSA, 40 fmol flap substrate and the resuspended in buffer C (30 mm HEPES-NaOH (pH 7.5), 8 mm indicated amounts of PCNA. Reactions were carried out for magnesium acetate, 0.2 mg/ml BSA, 1 mm DTT). Binding 30 min at 301C with phosphorylated or unphosphorylated assays were carried out in a final volume of 16 ml. 50 fmol Fen1 and quenched with stop buffer. The products were volume of DNA templates were coated with 1.5 pmol of RP-A separated onto 15% denaturating gels, visualized by auto- for 2 min on ice and washed with buffer D (30 mm HEPES-

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4312 NaOH (pH 7.5), 5% (v/v) glycerol, 0.1 mg/ml BSA, 1 mm tested. After 45 min at 301C, the reactions were stopped by DTT, 0.01% (v/v)Nonidet P-40, 0.1 m m EDTA). Reaction adding SDS sample buffer and loaded on a 12% SDS– were performed into buffer C with 0.7 pmol of phosphorylated polyacrylamide gel, electrophoresed for 90 min at 120 V, dried or unphosphorylated Fen1 for 1 min at 301C. Then, magnetic and autoradiographed. beads were washed off with buffer D and immediately resuspended in SDS sample buffer. DNA binding proteins were resolved on a 12% SDS–polyacrylamide gels and Immunoﬂuorescence

subsequently transferred to a PVDF membrane. The mem- HeLa cells were transfected with CMV-myc-Fen1 by the CaCl2 brane was probed with rabbit polyclonal anti-hFen1 purified method, incubated for 24 h, blocked for 24 h with 4 mm antibody (produced at the animal facility in the Institute of hydroxyurea and released during 3 h. First, nonionic detergent animal breeding of the University of Zu¨ rich)followed by horse Triton X-100 was used to remove soluble proteins. Then, cells anti-rabbit IgG-horseradish peroxidase (HRP, Sigma)and were fixed in methanol at À201C for 10 min and blocked in 5% Enhanced ChemiLuminescence (ECL, Pierce). BSA, 0.1% Tween in PBS (blocking buffer)for 10 min. Subsequently, the cells were incubated with the A-14 GST pull-down experiments polyclonal anti-myc and PC10 anti-PCNA antibodies in blocking solution. Primary antibodies were visualized by For in vitro binding, 10 ml of glutathione–sepharose beads were Cy.3-conjugated goat anti-rabbit and FITC-conjugated goat incubated with either 2 mg of purified GST-RF-C A or 100 mL anti-mouse antibodies as described (Dahm and Hu¨ bscher, E. coli extract containing overexpressed GST fusion proteins 2002). The coverslips were observed using an epifluorescence 1 for 3 h at 4 C. GST fusion proteins bound on beads were microscope (Olympus BX51)and a microscope digital camera incubated with 100 ng of phosphorylated or unphosphorylated system (Olympus DP50). Fen1. After 1 h at 41C, the beads were washed four times with 50 mm Tris (pH 8), 120 mm NaCl, 0.5% (v/v)Nonidet P-40, 1mm DTT, 1 mm PMSF (binding buffer), and the bound Cell cycle analysis proteins were eluted with 50 mm Tris (pH 6.8), 2 mm EDTA, 1% (w/v) b-mercaptoethanol, 8% (v/v)glycerol and 0.025% In the unsynchronized state, the 293 cells expressing myc- (w/v)bromophenol blue (SDS sample buffer).Proteins were tagged Fen1 were trypsinized 48 h post-transfection. For synchronization, 293 cells were blocked in S phase 24 h after separated onto a 12% SDS–polyacrylamide gels, transferred m onto a nitrocellulose membrane (Micron Separations Inc.), CaCl2 transfection. The hydroxyurea block (4 m )was during and probed with anti-hFen1 rabbit purified antibody followed 24 h followed by a kinetic release of 105 or 210 min. After by horse anti-rabbit IgG-HRP and ECL. trypsinization, the cells were fixed in 70% ethanol, permeabi- lized in PBS containing 0.2% (v/v)Triton and 100 mg/ml RNase A. The cells were then incubated with anti-c-myc Immunoprecipitation and antibodies antibody (9E10)for 60 min at room temperature in blocking Immunoprecipitations from nuclear HeLa were performed as solution (PBS with 2.5 % (w/v)BSA and 0.1% (v/v)Tween described (Koundrioukoff et al., 2000)with anti-Cyclin A 20), washed and incubated with FITC-coupled goat anti- rabbit (Santa Cruz, H-432), anti-hFen1 rabbit purified anti- mouse IgG (ANAWA)in blocking solution for 60 min. The body and anti-c-myc antibody (9E10). The mouse IgG1 cells were stained with 50 ng/ml propidium iodide in PBS. (MOPC-21)from Sigma and rabbit IgG (VECTOR labora- Expression of myc-tagged Fen1 and cell cycle analysis were tories)antibodies were used as negative controls. Proteins were performed by flow cytometry with FACScan (Beckton detected on the nitrocellulose membrane with anti-Cyclin A Dickinson, Moutain View, CA, USA)with the program rabbit or anti-hFen1 rabbit or anti-Cdk2 rabbit (Santa Cruz, CELLQuest (BecktonDickinson). H-298)or anti-PCNA mouse (Santa Cruz, PC10)or anti-Cdk1 mouse (Transduction Laboratories, C12729). A phosphoryl- group-specific antibody (Poly-Z-PS1, Zymed Laboratories) was used to detect protein phosphorylated on serine. Acknowledgements We thank HP Nasheuer for providing purified Cdks complexes, JP Magyar for the Cdk1 clone, R Freire for GST- In vitro kinase assays PCNA plasmid, L Meijer for the gift of Olomoucine, T Hennet In vitro phosphorylation assays were performed with 100 ng of for the use of FACScant flow cytometer, S Hasan and ZO purified Cdk–Cyclin complexes into a final volume of 18 ml Jo´ nsson for stimulating discussions. This work has been containing 40 mm HEPES-NaOH (pH 7.5), 8 mm MgCl2, supported by grants from the Swiss National Science 33.3 mm ATP, 10 mCi of [g-32P]ATP (3000 Ci/mmol; Amersham Foundation (Grant 31-61361.00)to GH and SK, and by the Pharmacia)and the substrates (histone H1, hFen1)to be Kanton of Zu¨ rich to GH and UH.

References

Abraham RT, Acquarone M, Andersen A, Asensi A, Belle´ R, Bae SH and Seo YS. (2000). J. Biol. Chem., 275, 38022–38031. Berger F, Bergounioux C, Brunn G, Buquet-Fagot C, Fagot Bambara RA, Murante RS and Henricksen LA. (1997). D, Glab N, Goudeau H, Goudeau M, Guerrier P, Houghton J. Biol. Chem., 272, 4647–4650. P, Hendriks H, Kloareg B, Lippai M, Marie D, Maro B, Biswas EE, Zhu FX and Biswas SB. (1997). Biochemistry, 36, Meijer L, Mester J, Mulner-Lorillon O, Poulet SA, 5955–5962. Schierenberg E, Schutte B, Vaulot D and Verlhac MH Brosh Jr RM, von Kobbe C, Sommers JA, Karmakar P, (1995). Biol. Cell, 83, 105–120. Opresko PL, Piotrowski J, Dianova I, Dianov GL and Bohr Adams PD, Sellers WR, Sharma SK, Wu AD, Nalin CM VA. (2001). EMBO J., 20, 5791–5801. and Kaelin Jr WG. (1996). Mol. Cell. Biol., 16, Budd ME and Campbell JL. (1997). Mol. Cell. Biol., 17, 6623–6633. 2136–2142.

Oncogene Regulation of DNA replication by phosphorylation of hFenl G Henneke et al 4313 Dahm K and Hu¨ bscher U. (2002). Oncogene, 21, Kucherlapati R. (2002). Proc. Natl. Acad. Sci. USA, 99, 7710–7719. 9924–9929. Donaldson AD and Blow JJ. (1999). Curr. Opin. Genet. Dev., Li X, Li J, Harrington J, Lieber MR and Burgers PM. (1995). 9, 62–68. J. Biol. Chem., 270, 22109–22112. Ducoux M, Urbach S, Baldacci G, Hu¨ bscher U, Lieber MR. (1997). BioEssays, 19, 233–240. Koundrioukoff S, Christensen J and Hughes P. (2001). J. Lin XH, Walter J, Scheidtmann K, Ohst K, Newport J and Biol. Chem., 276, 49258–49266. Walter G. (1998). Proc. Natl. Acad. Sci. USA, 95, Dutta A and Stillman B. (1992). EMBO J, 11, 2189–2199. 14693–14698. Fotedar R and Fotedar A. (1995). Prog. Cell. Cycle Res., 1, Matsuzaki Y, Adachi N and Koyama H. (2002). Nucleic Acids 73–89. Res., 30, 3273–3277. Frouin I, Montecusso A, Biamonti G, Hu¨ bscher U, Spadari S McVey D, Ray S, Gluzman Y, Berger L, Wildeman AG, and Maga G. (2002). EMBO J., 21, 2485–2495. Marshak DR and Tegtmeyer P. (1993). J. Virol., 67, Gordenin DA, Kunkel TA and Resnick MA. (1997). Nat. 5206–5215. Genet., 16, 116–118. Morgan DO. (1996). Curr. Opin. Cell. Biol., 8, 767–772. Harrington JJ and Lieber MR. (1994a). Embo J., 13, Murante RS, Rust L and Bambara RA. (1995). J. Biol. Chem., 1235–1246. 270, 30377–30383. Harrington JJ and Lieber MR. (1994b). Genes Dev., 8, Nigg EA. (1995). BioEssays, 17, 471–480. 1344–1355. Nurse P. (2000). Cell, 100, 71–78. Hasan S, Stucki M, Hassa PO, Imhof R, Gehrig P, Hunziker Parenteau J and Wellinger RJ. (1999). Mol. Cell. Biol., 19, P, Hu¨ bscher U and Hottiger MO. (2001). Mol. Cell., 7, 4143–4152. 1221–1231. Reagan MS, Pittenger C, Siede W and Friedberg EC. (1995). J. Henneke G, Koundrioukoff S and Hu¨ bscher U. (2003). Bacteriol., 177, 364–371. EMBO reports, 4, 252–256. Rossi R, Villa A, Negri C, Scovassi I, Ciarrocchi G, Henricksen LA, Tom S, Liu Y and Bambara RA. (2000). J. Biamonti G and Montecucco A. (1999). EMBO J., 18, Biol. Chem., 275, 16420–16427. 5745–5754. Henricksen LA, Umbricht CB and Wold MS. (1994). J. Biol. Schub O, Rohaly G, Smith RW, Schneider A, Dehde S, Chem., 269, 11121–11132. Dornreiter I and Nasheuer HP. (2001). J. Biol. Chem., 216, Hinds PW and Weinberg RA. (1994). Curr. Opin. Genet. Dev., 38076–38083. 4, 135–141. Schurtenberger P, Egelhaaf SU, Hindges R, Maga G, Hosﬁeld DJ, Mol CD, Shen B and Tainer JA. (1998). Cell, 95, O’Jo´ nsson ZO, May RP, Glatter O and Hu¨ bscher U. 135–146. (1998). J. Mol. Biol., 215, 123–132. Hwang KY, Baek K, Kim HY and Cho Y. (1998). Nat. Struct. Sommers CH, Miller EJ, Dujon B, Prakash S and Prakash L. Biol., 5, 707–713. (1995). Biol. Chem., 270, 4193–4196. Johnston LH, Masai H and Sugino A. (1999). Trends Cell Stillman B. (1996). Science, 274, 1659–1664. Biol., 9, 249–252. Storici F, Henneke G, Ferrari E, Gordenin DA, Hu¨ bscher U Johnston LH, Masai H and Sugino A. (2000). Prog. Cell Cycle and Resnick MA. (2002). EMBO J., 21, 5930–5942. Res., 4, 61–69. Stucki M, O’Jo´ nsson ZO and Hu¨ bscher U. (2001). J. Biol. Jo´ nsson ZO, Hindges R and Hu¨ bscher U. (1998). EMBO J., Chem., 276, 7843–7849. 17, 2412–2425. Tom S, Henricksen LA and Bambara RA. (2000). J. Biol. Kang HY, Choi E, Bae SH, Lee KH, Gim BS, Kim HD, Park Chem., 275, 10498–10505. C, MacNeill SA and Seo YS. (2000). Genetics, 155, Vas A, Mok W and Leatherwood J. (2001). Mol. Cell. Biol., 1055–1067. 21, 5767–5777. Kim K, Biade S and Matsumoto Y. (1998). J. Biol. Chem., 273, Voitenleitner C, Fanning E and Nasheuer HP. (1997). 8842–8848. Oncogene, 14, 1611–1615. Klungland A and Lindahl T. (1997). EMBO J., 16, Warbrick E, Lane DP, Glover DM and Cox LS. (1997). 3341–3348. Oncogene, 14, 2313–2321. Koundrioukoff S, Jo´ nsson ZO, Hasan S, de Jong RN, van der Wu SM, Zhang P, Zeng XR, Zhang SJ, Mo J, Li BQ and Lee Vliet PC, Hottiger MO and Hu¨ bscher U. (2000). J. Biol. MY. (1998). J. Biol. Chem., 273, 9561–9569. Chem., 275, 22882–22887. Wu X, Li J, Li X, Hsieh CL, Burgers PM and Lieber MR. Kucherlapati M, Yang K, Kuraguchi M, Zhao J, Lia M, (1996). Nucleic Acids Res., 24, 2036–2043. Heyer J, Kane MF, Fan K, Russell R, Brown AM, Wu X, Wilson TE and Lieber MR. (1999). Proc. Natl. Acad. Kneitz B, Edelmann W, Kolodner RD, Lipkin M and Sci. USA, 96, 1303–1308.

Oncogene