Requirements for the nucleolytic processing of DNA ends by the Werner syndrome protein-Ku70/80 complex.

Werner syndrome (WS) is an inherited disease characterized by premature onset of aging, increased cancer incidence, and genomic instability. The WS gene encodes a protein with helicase and exonuclease activities. Our previous studies indicated that the Werner syndrome protein (WRN) interacts with Ku, a heterodimeric factor of 70- and 80-kDa subunits implicated in the repair of double strand DNA breaks. Moreover, we demonstrated that Ku70/80 strongly stimulates and alters WRN exonuclease activity. In this report, we investigate further the association between WRN and Ku70/80. First, using various WRN deletion mutants we show that 50 amino acids at the amino terminus are required and sufficient to interact with Ku70/80. In addition, our data indicate that the region of Ku80 between amino acids 215 and 276 is necessary for binding to WRN. Then, we show that the amino-terminal region of WRN from amino acid 1 to 388, which comprise the exonuclease domain, can be efficiently stimulated by Ku to degrade DNA substrates, indicating that the helicase domain and the carboxyl-terminal tail are not required for the stimulatory process. Finally, using gel shift assays, we demonstrate that Ku recruits WRN to DNA. Taken together, these results suggest that Ku-mediated activation of WRN exonuclease activity may play an important role in a cellular pathway that requires processing of DNA ends.

Werner syndrome (WS) 1 is a rare disorder characterized by the early appearance of many diseases characteristics of human aging such as atherosclerosis, osteoporosis and diabetes mellitus type II (1)(2)(3). In addition, WS patients are prone to many types of soft tissue tumors (4). Cells from WS patients display a shortened replicative life span and elevated levels of chromosomal abnormalities, including insertions, deletions and translocations (5,6). The gene responsible for WS has been cloned and encodes a protein of 1,432 amino acids (WRN) (7). A nuclear localization signal has been identified at its carboxylterminal end (8), and all the described WS mutations result in a predicted truncated protein that fails to enter the nucleus. The central region of WRN is homologous to a seven-motif domain found in helicases from a wide variety of organisms, including bacteria (recQ), yeast (Sgs1), and human (Bloom syndrome (BLM), RecQL) (7,9,10). On the other hand, the amino-terminal region of WRN is quite unique among this family of helicases because it contains a sequence that resembles the exonuclease domain of Escherichia coli RNA polymerase I and RNase D (11,12). Indeed, studies with the purified recombinant protein have shown that WRN displays both 3Ј-to 5Ј-exonuclease and 3Ј-to 5Ј-helicase activities (13)(14)(15).
We have shown recently that Ku interacts with and alters the specificity of the WRN exonuclease (16). These results indicated that in the presence of Ku, WRN degradation of 3Ј-recessed strand of a partial DNA duplex is strongly stimulated. In addition, Ku alters the specificity of WRN so that the blunt end DNA duplex and 3Ј-protruding DNA are also hydrolyzed by WRN exonuclease. Ku is a 70-and 80-kDa heterodimer that binds to the ends of double strand DNA and translocates along the DNA in an ATP-independent manner, allowing several Ku heterodimers to bind to a single DNA molecule (17)(18)(19)(20). In higher eukaryotes, Ku has been implicated in the metabolism of DNA ends, including the repair of double strand DNA breaks (DSBs) and the V(D)J recombination process (21,22). In mammalian cells, Ku appears to be the first protein to recognize DSBs. Upon binding to the ends of broken DNA, Ku interacts with the DNA-dependent protein kinase catalytic subunit, stabilizes the interaction of this catalytic subunit with DNA, and enhances its protein kinase activity (18,21,23,24). The subsequent steps in the end-joining process are then thought to require the recruitment of specific activities necessary for the processing and ligation of DNA ends (25)(26)(27)(28). Although factors such as XRCC4 and DNA ligase IV have been shown to be required for the ligation step, the processing of DNA ends is still poorly understood, and the identity of the nucleases involved in this process remains to be defined. Interestingly, the finding that Ku interacts with and stimulates the activity of the WRN exonuclease suggests that WRN may be one of the factors that function in the processing of broken DNA ends. In this study, to understand better the functional relevance of this finding, we have characterized the interaction of WRN with Ku and DNA ends. First, we show that Ku interacts with the extreme amino terminus of WRN, from amino acid 1 to 50, whereas the central region of Ku80, from amino acid 215 to 276, is necessary for binding to WRN. Then, we demonstrate that Ku can efficiently stimulate the exonuclease activity of an amino-terminal mutant WRN (WRN⌬ C388; amino acids 1-388), indicating that the helicase and carboxyl-terminal domains of WRN are dispensable for Kustimulated processing of DNA ends. Moreover, using electrophoretic mobility shift assays we show that WRN forms a strong complex with Ku on DNA.

EXPERIMENTAL PROCEDURES
Protein Purification-Recombinant FLAG-WRN and a series of FLAG-WRN deletion mutants were purified as described in Ref 16. Baculoviruses expressing recombinant FLAG epitope-tagged WRN were used to infect Sf9 cells. Whole cell lysates were prepared in lysis buffer (10 mM Hepes pH 7.5, 100 mM NaCl, 1.5 mM MgCl 2 , 0.5% Nonidet P-40). FLAG-WRN was purified through a DEAE-cellulose column and by affinity chromatography on anti-FLAG resin. Histidine-tagged Ku70 and Ku80 were coexpressed in Sf9 cells, and the Ku70/80 heterodimer was purified by metal affinity resin (Talon, CLONTECH Biotech) and DNA cellulose chromatography. The plasmid containing the WRN cDNA with a substitution at amino acid 82 (Asp 3 Ala; WRN82A) was constructed using a polymerase chain reaction-based strategy. The open reading frame with amino acids 1-82A, and 82A-311 were amplified by polymerase chain reaction using the following DNA primers: 5Ј-CGGCCATATGAGTGAAAAAAAATTGGA-3Ј/5Ј-TGGCCACTCCAT-GGCAAATCCCACC-3Ј and 5Ј-TGGTGGGATT TGCCATGGAGTGGC-3Ј/5Ј-AAAGGATCCATGGGCCTCAGTTCAGTCTCA-3Ј. The amplified fragments were digested with NdeI and NcoI and subcloned into pR-SET-A vector. An NdeI-EcoRV fragment containing a substitution at position 82 (Asp 3 Ala) was then cloned into pcDNA-FLAG-WRN. Full-length WRN cDNA containing the D82A substitution was subcloned into the NdeI-NotI sites of pVL1392-FLAG-WRN. Recombinant DNA was purified and transfected into Sf9 cells with linearized baculovirus DNA (BaculoGold, PharMingen) to generate the recombinant baculovirus expressing FLAG epitope-tagged WRN82A.
Protein Binding Assays-EcoRI-digested full-length Ku80 cDNA fragments were subcloned in-frame into the EcoRI site of pCDNA3.1-HisC. This allows different subdomains of Ku80 to be expressed in vitro under the control of the T7 RNA polymerase promoter. These constructs were then subjected to in vitro transcription-translation in the presence of [ 35 S]methionine (NEN Life Science Products) using a rabbit reticulocyte lysate translation system (Promega). 2 g of FLAG-WRN was immobilized on FLAG beads and then incubated with 15 l of in vitro translated Ku80 mutants and 135 l of buffer containing 10 mM Hepes pH 7.5, 100 mM NaCl, 1.5 mM MgCl 2 , 0.5 mM EDTA, 10% glycerol, and 0.2% Nonidet P-40. After extensive washes, resin-bound proteins were released by boiling in SDS-sample buffer and electrophoresed on SDS-polyacrylamide gel followed by autoradiography.
Mutant WRNs were generated by insertion of the appropriate cDNA sequences into either pGEX-2T or pGEX-3X vectors and expressed as GST fusion proteins in E. coli [(GST-WRN(1-50), GST-WRN(50 -749), and GST-WRN(150 -749); numbers represent amino acid residues spanning the fusion protein)]. 4 g of each purified GST fusion protein and FLAG-WRN were immobilized on the appropriate affinity beads (glutathione or FLAG resins) and then incubated with 500 l of cell lysates prepared from Sf9 cells expressing Ku80 (the concentration of Ku80 in the extract was estimated at ϳ3 ng/l). After extensive washing, bound proteins were eluted from the beads using BCO buffer (1 M KCl, 10 mM Tris-HCl pH 7.5, 1 mM EDTA, 5% glycerol, 1 mM dithiothreitol) containing a mixture of protease inhibitors, and analyzed by SDS-PAGE and Western blotting with anti-Ku80 antibodies. In a similar assay, 4 g of each WRN GST fusion protein and FLAG-WRN were immobilized on beads and then incubated with 500 l of lysates from Sf9 cell coinfected with baculoviruses expressing Ku70 and Ku80. After extensive washes, the bound proteins were released from the beads using BCO buffer, and then analyzed by SDS-PAGE and silver staining.
Electrophoretic Mobility Shift Assay-40-mer (B1) (GATTTC-CCGCTAGCAATATTCTGCAGCCAAGCTTCCGCGC) was labeled us-ing [ 32 P]ATP and T4 polynucleotide kinase and then annealed to a partially complementary 46-mer (A3). 32 P-Labeled DNA (80 fmol, 200,000 cpm) was incubated with increasing amounts (100 -300 fmol) of Ku and full-length WRN in 10 l of buffer (10 mM Tris-HCl pH 7.5, 80 mM NaCl, 4 mM KCl, 2 mM EDTA, and 10% glycerol) at 25°C for 10 min. Then, the samples were resolved by electrophoresis through a 4% polyacrylamide gel at 10 V/cm in the cold room. The gels were dried on Whatman 3MM paper and subjected to autoradiography. For gel supershift assays, 200 fmol of Ku and 200 fmol of WRN were first preincubated with DNA substrates at 25°C for 10 min under the conditions described above. Then, 200 ng of goat polyclonal Ku80 antibodies (Santa Cruz), goat polyclonal ␤-actin antibodies (Santa Cruz), or rabbit polyclonal WRN antibodies (Novus), was added to the appropriate reaction mixture. After a 10-min incubation at 25°C, the samples were loaded on a polyacrylamide gel and processed as described above.

Mapping the Interaction Domains between WRN and
Ku80 -We have shown previously that the amino-terminal region of WRN interacts with Ku80 (16). To identify the minimal amino acid domain of WRN necessary for this proteinprotein interaction, we performed pull-down assays using GST-WRN fusion proteins containing different segments of the amino-terminal region of WRN (GST-WRN(1Ϫ50), GST-WRN(50Ϫ749), and GST-WRN(150Ϫ749)) ( Fig. 1A). In addition, full-length FLAG-WRN was used in the assay as positive control. Each GST fusion protein and FLAG-WRN was immobilized on either glutathione or FLAG beads and used in pulldown assays with lysates from Sf9 cells infected with a recombinant baculovirus expressing Ku80. After extensive washes, the bound products were resolved on a SDS-PAGE and analyzed by Western blotting with antibodies against Ku80. As expected, Ku80 interacted with FLAG-tagged WRN (Fig. 1B These results indicate that the region of WRN between amino acids 1 and 50 is necessary and sufficient for binding to Ku80. To demonstrate further that this region of WRN interacted with an integral Ku70/80 heterodimer, each of the immobilized GST fusion proteins and FLAG-WRN was incubated with an Sf9 lysate coinfected with two baculoviruses expressing Ku70 and Ku80. After extensive washes, the bound proteins were analyzed with SDS-PAGE and silver staining. The results of this experiment confirmed that the amino-terminal domain (amino acids 1-50) is sufficient for binding to the Ku70/80 heterodimer (Fig. 1C, lane 3). Identical results were obtained when purified proteins were mixed in solution prior to pull down (not shown). To complement this analysis and identify the region of Ku80 necessary for the interaction with WRN, various deletion mutants of Ku80 were expressed in an in vitro reticulocyte lysate system in the presence of [ 35 S]methionine ( Fig. 2A, upper panel). The in vitro translated products were then incubated with either FLAG-tagged WRN ( Fig. 2A, lower panel) or a control FLAG-tagged hepatitis C virus polymerase (not shown) that were immobilized on FLAG beads. Bound proteins were resolved on a SDS-PAGE and analyzed by autoradiography (Fig.  2B). This experiment revealed that deletion of the carboxylterminal region of Ku80, from amino acid 277 to 732, did not affect the binding to WRN (lanes 2-4); however, further deletion to amino acid 216 completely abolished the interaction (lane 1). Deletion of the amino terminus of Ku80, from amino acids 1 to 325, also abrogated the interaction with WRN (lane 5). The FLAG-tagged hepatitis C virus polymerase did not interact with any of the Ku80 mutants used in this assay (data not shown). Thus, these results indicate that the central region of Ku80, from amino acids 215 to 276, is important for mediating the interaction with WRN.
Ku70/80 Alters the Exonuclease Activity of WRN⌬C388 -WRN exonuclease activity resides near its amino terminus (13,29), whereas the helicase domain is located in the central portion of the protein (7, 14, 15). A recent study also found that the amino-terminal domain of WRN, from amino acids 1 to 333, is sufficient for the degradation of DNA substrates with a 3Ј-overhang (13). Because our results indicate that this region of WRN also binds to Ku80 (see Fig. 1), we were interested in testing whether WRN⌬C388, a mutant form of WRN containing the first 388 amino acids and missing the helicase domain, is sufficient for Ku-dependent enhancement of DNA hydrolysis. FLAG epitope-tagged mutant WRN (WRN⌬C388) was produced using a baculovirus expression system and then used in exonuclease assays with several DNA substrates in the presence or absence of Ku70, Ku80, or Ku70/80. WRN⌬C388 weakly hydrolyzed the 3Ј-recessed strand of a partial DNA  ). B, blunt-ended, 5Ј-32 P-labeled 20mer (A1)/20-mer (A2) DNA substrate was used in exonuclease assays under the conditions described above. C, 3Ј-protruding, 5Ј-32 P-Labeled 46-mer (A3)/20-mer (A1) DNA substrate was used in exonuclease assays as described in A. D, 5Ј-32 P-Labeled 5Ј-20-mer single strand DNA substrate was used in exonuclease assays as described in A.
duplex and the 3Ј-ends of blunt end DNA (Fig. 3, A and B, lane  4). Importantly, in the presence of Ku70/80, the exonuclease activity of WRN⌬C388 was strongly stimulated (lane 7). The dramatic increase in exonuclease activity depended on the presence of Ku70/80 because either Ku70 or Ku80 alone failed to stimulate WRN⌬C388 (lanes 5 and 6). More significantly, in the presence of the Ku70/80 heterodimer, WRN⌬C388 efficiently digested 3Ј-protruding ends and single strand DNA substrates (Fig. 3, C and D, lane 7).
To confirm that the DNA degradation is catalyzed by WRN and not by a contaminating activity present in the Ku preparation, we performed the exonuclease assay in the presence of WRN82A, a point mutant of WRN lacking exonuclease activity (13). As shown in Fig. 4 5-8), suggesting that the observed DNA degradation depends on the presence of a functional WRN exonuclease.
In conclusion, Ku stimulates and alters the exonuclease activity of WRN lacking the helicase domain, suggesting that DNA unwinding is not required in this process.
Ku Stimulates WRN Exonuclease Activity in an ATP-independent Manner-WRN contains amino acid regions possessing nucleic acid-dependent ATPase and helicase activities (7,14). It has also been shown that the helicase activity of WRN is completely dependent on the hydrolysis of ATP, and ATP significantly stimulates the activity of the WRN exonuclease (29,30). To investigate further the mechanism of stimulation of the WRN exonuclease activity by Ku, we then asked whether ATP is required for this stimulatory process. As shown in Fig. 5, in the presence of Ku, the addition of ATP had no effect on the activity of either full-length WRN or WRN⌬C388, indicating that Ku-mediated stimulation of WRN exonuclease activity does not require ATP hydrolysis.
Ku Recruits WRN to Double Strand DNA-Ku has high specificity for DNA substrates, it binds tightly to the ends of double strand DNA, and has equal affinity for 5Ј-protruding, 3Ј-protruding, and blunt end DNA (18,31). On the other hand, WRN binds very weakly to double strand DNA (32,33). Therefore, we performed electrophoretic mobility shift analysis to determine whether Ku facilitates the recruitment of WRN to DNA through protein-protein interactions. Upon incubation of a 40/46-mer DNA duplex with purified Ku, we observed the formation of two distinct complexes (Fig. 6, lanes 4 -6; complexes 1 and 2), most likely representing either one or two Ku70/80 heterodimers binding to one DNA molecule. In contrast, WRN failed to bind to DNA, as determined by the absence of any retarded complex on the gel (lanes 2 and 3). However, when both WRN and Ku were incubated with the DNA substrate, in addition to the complexes observed previously (complexes 1 and 2), a slower migrating band appeared in the gel (lanes 7-12; complex 3). This band, as well as the Ku⅐DNA complex 2, was supershifted by anti-Ku80 antibodies (lane 13). Moreover, the addition of WRN antibodies supershifted the putative WRN⅐Ku⅐DNA complex, whereas the Ku⅐DNA complexes were unaffected (lane 14). Antibodies against an unrelated protein (␤-actin) had no effect on the relative mobility of any of the observed complexes (lane 15). Altogether, these results indicate that WNR can form a stable complex on DNA in the presence of Ku, further supporting the idea that Ku recruits WRN to DNA ends through specific protein-protein interaction.

DISCUSSION
Previous studies demonstrated that WRN possesses an intrinsic 3Ј-to 5Ј-exonuclease activity that digests 3Ј-recessed DNA substrates (13,15). More recently, our work has indicated that WRN interacts with Ku70/80 heterodimer, and Ku70/80 stimulates and alters the exonuclease activity of WRN, so that it can hydrolyze 3Ј-recessed, 5Ј-recessed, and blunt end DNA substrates efficiently (16). Because WRN does not bind to double strand DNA and has only a weak affinity for single strand DNA (32), our finding raises the possibility that Ku may function as a bridging factor that brings WRN in close proximity to its DNA substrate. Ku-mediated recruitment of WRN may therefore represent a critical step in the targeting and subsequent activation of WRN exonuclease on specific DNA sites. In this study, to understand better the functional relationship between Ku and WRN, we have performed a detailed analysis of the interactions between these two factors. Our analysis shows that the interaction between WRN and Ku80 requires a stretch of 50 amino acids at the extreme amino terminus of WRN. These results are in apparent conflict with the data by Cooper et al. (34) showing that Ku interacts with the carboxyl-terminal region of WRN. Under our experimental conditions, we have never detected a direct interaction between the carboxyl-terminal domain of WRN and Ku. As we have suggested previously (16), because the carboxyl terminus of WRN can dimerize with the full-length protein (34), it is possible that cellular WRN may have mediated the interaction observed in their affinity purification experiment.
Our analysis of Ku80 mutants has also allowed the identification of a region between amino acids 215 and 276 as an important domain required for Ku80 binding to WRN. This region is distinct from the regions of Ku80 which have been found to be involved in interactions with Ku70 and a DNA-dependent protein kinase catalytic subunit (35)(36)(37).
Because the region of WRN which binds to Ku80 is proximal to the WRN exonuclease domain, we reasoned that the aminoterminal region of WRN might be sufficient for Ku stimulation of the exonuclease activity. To address this point, we investigated the effect of Ku on the exonuclease activity of WRN⌬C388, a form of WRN which is missing the helicase domain and the carboxyl-terminal region. Our results indicate that although WRN⌬C388 poorly digests 3Ј-recessed ends and blunt end DNA duplexes, in the presence of Ku it efficiently digests 3Ј-recessed, 5Ј-recessed ends and blunt end DNA duplexes as well as single strand DNA. The observed activity is strictly dependent on the presence of a functional WRN because Ku70/80 does not stimulate WRN82A, a mutant protein lacking exonuclease activity. In addition, our results show that ATP is dispensable for this stimulatory process. Thus, Kustimulated WRN exonuclease activity is physically and functionally separable from the helicase domain and the carboxyl terminus of WRN and does not require ATP hydrolysis. To investigate how WRN can stably associate with its DNA substrate, we then performed gel shift assays. These experiments show that although WRN by itself has undetectable DNA binding affinity, in the presence of Ku it forms a strong complex with DNA. These results imply that protein interactions between WRN and Ku play a key role in the recruitment of WRN to DNA ends and suggest a potential function in the processing of DSBs. In eukaryotic cells, DSBs produced by external sources such as ionizing radiations or some chemical agents undermine the integrity of the genome, and if not repaired, they can lead to cell death (21,27). In mammals, the Ku heterodimer, DNA-dependent protein kinase catalytic subunit, and the Ligase4⅐Xrcc4 complex are involved the repair of double strand DNA breaks by nonhomologous end joining (23, 24,38,39). In addition, because improper ligation can be potentially deleterious to the cell, several studies have suggested that nucleolytic activities are necessary in the repair process to trim the DNA before the ends are rejoined (40,41). These nucleases must either recognize the broken DNA ends or be recruited to the site of DNA damage by specific protein-protein interactions. Mre11, a 3Ј-to 5Ј-exonuclease and endonuclease, has been implicated in one or more DNA repair pathways (21,41). Mre11, Rad50, and NBS1, an Xrs2p-like protein whose loss leads to Nijmegen breakage syndrome, bind cooperatively to DNA and form a distinct protein-DNA complex (42). This complex displays several enzymatic activities, including partial unwinding of the DNA duplex and cleavage of the 3Ј-protruding end of double strand DNA. Recently, it has also been shown that Mre11 can facilitate the joining of mismatched ends, suggesting that Mre11 may direct DNA end joining at sites of microhomology (43). Nevertheless, whether Mre11 plays a role in nonhomologous end joining remains unclear. WRN does not bind directly to double strand DNA; however, it interacts with Ku and is recruited to DNA ends by specific protein-protein interactions. Therefore, it is tempting to speculate that WRN, like Mre11, may participate in a specialized form of DNA repair which requires the proper processing of DNA ends before they are rejoined. Within the last few years it has become apparent that several distinct pathways of repairing DSBs operate in mammalian cells (21,40). Thus, it is likely that some of these pathways, although sharing a subset of common components, may also utilize a number of unique and more specialized factors. WRNϪ/Ϫ fibroblasts are sensitive to 4-nitroquinoline 1-oxide but are insensitive to other DNA-damaging agents such as ionizing radiation (44,45), suggesting that if WRN is directly involved in the processing of damaged DNA, it may be required for the repair of a specific subset of DNA lesions. A later study on the ligation of linear plasmid DNA in lymphoblast cells from WS patients showed an increased rate of deletion at the ligation site but no reduction in end joining (46). These data may therefore indicate a subtle defect in the accuracy of the DSBs repair machinery. Further support for the idea that WRN and Ku may be part of a common pathway comes from studies of Ku80 mutant mice. WRNϪ/Ϫ fibroblasts and cells from Ku80 knockout mice display interesting similarities, such as a higher degree of chromosomal instability characterized by DNA breakage, translocation, and rearrangements (47)(48)(49). In addition, Ku80 knockout mice show signs of premature aging, such as osteopenia, atrophic skin, and cancer (48). Thus, both Ku80 and WRN appear to influence the rate of senescence. Interestingly, WRN is localized in the nucleolus of human cells (50,51), raising the possibility that its function may be required for a specific nucleolar process. On the other hand, it is also possible that WRN is sequestered and kept inactive in the nucleolus, ready to be released when it is needed, as observed for some cell cycle regulatory factors (52). Our goal is now to determine the specific process in which the WRN nucleolytic activity is required. The answer to this and other questions on the function of the WRN will provide important clues on the biological activities involved in the process of aging.