Nuclear DNA helicase II (RNA helicase A) interacts with Werner syndrome helicase and stimulates its exonuclease activity.

Nuclear DNA helicase II (NDH II), alternatively named RNA helicase A, is involved in transcription and RNA processing. Here, we report that NDH II interacts with the Werner syndrome helicase WRN, an enzyme associated with premature aging and predisposition to tumorigenesis. NDH II was co-purified with WRN, DNA polymerase delta, and replication protein A (70 kDa) during several steps of conventional column chromatography. Co-immunoprecipitations revealed an association between NDH II, WRN, and polymerase delta. We demonstrate a direct protein-protein interaction between WRN and NDH II that is mediated by the N-terminal double-strand RNA-binding domain II and C-terminal RGG box of NDH II and the N-terminal exonuclease domain of WRN. WRN inhibited the DNA-dependent NTPase and DNA helicase activities of NDH II. On the other hand, the 3' --> 5' exonuclease activity of WRN was increased by the presence of NDH II. NDH II directly stimulated the exonuclease domain of WRN, whereas the exonuclease domain of WRN suppressed the DNA-dependent (but not RNA-dependent) ATPase activity of NDH II. These results suggest that the double-strand RNA-binding domain II and RGG box of NDH II together form a protein-protein interaction surface that contacts the exonuclease domain of WRN. Furthermore, NDH II enhanced the degradation of D-loop DNA by the WRN exonuclease. Taken together, these results suggest that NDH II plays a role in promoting the DNA processing function of WRN, which in turn might be necessary for maintaining genomic stability.

Nuclear DNA helicase II (NDH II 1 ; RNA helicase A) belongs to the DEXH helicase superfamily (1,2) and is able to unwind both double-stranded DNA and RNA in the presence of one of the four ribo-or deoxyribonucleoside triphosphates (3)(4)(5). NDH II is homologous to the Drosophila maleless protein MLE, which participates in sex-specific dosage compensation by enhancing transcription from the single X chromosome of males (6). In addition to transcription (6,7), NDH II has multiple physiological functions, such as in RNA processing and transport (8), in DNA repair (9), and also in tumorigenesis (10). Most recently, NDH II has been found to directly bind to histone ␥-H2AX, which arises from the cellular response to DNA damage at stalled transcription sites (11). Also, NDH II is phosphorylated by DNA-dependent protein kinase, a member of the phosphatidylinositol 3-phosphate kinase family. Phosphorylation by DNA-dependent protein kinase might modify the multiple functions of NDH II in a still unknown manner (12).
Werner syndrome is a rare autosomal recessive genetic disorder manifested by the symptoms of premature aging, such as atherosclerosis, osteoporosis, diabetes mellitus type II, cataracts, and genomic instability with an increased incidence of tumor formation (13). The molecular defect of the syndrome is a defective DNA helicase (WRN) that is homologous to the Escherichia coli RecQ helicase. Unique to all other members of the RecQ helicase family, WRN contains both a 3Ј 3 5Ј helicase and a 3Ј 3 5Ј exonuclease activity. WRN and the other human RecQ helicase homologs, i.e. Bloom syndrome helicase BLM and Rothmund-Thomson helicase RecQL4, are believed to be involved in DNA replication, repair, or recombination to sustain genomic stability. This might be achieved by, for example, facilitating recovery from stalled DNA replication forks, suppressing illegitimate recombination, or maintaining the telomere structure (14). Potentially, WRN can fulfill these multiple roles based on both its exonuclease and helicase activities. Nevertheless, so far, there is no convincing evidence that demonstrates a coordination of these two enzymatic activities. Rather, WRN seems to cooperate with other proteins, such as the Ku antigens (15), the telomeric repeat-binding factor TRF2 (16), BLM (17), and p53 (18), that either stimulate or inhibit the exonuclease activity of WRN. Here, we report that NDH II interacts with WRN and stimulates the exonuclease activity of this protein. This may indicate complementing or overlapping functions of these two helicases in DNA metabolism, possibly for the maintenance of genomic integrity.

EXPERIMENTAL PROCEDURES
Antibodies-The rabbit polyclonal antibody against the N terminus (amino acids (aa) 1-510) of human WRN (antibody 200) was from Abcam (Cambridge, UK). The rabbit polyclonal antibody against the C terminus (aa 1133-1432) of human WRN (H-300) and the goat polyclonal antibody against a C-terminal epitope of human WRN (C- 19) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The mouse monoclonal antibody against the C terminus (aa 1291-1412) of human WRN was from BD Biosciences. The rabbit polyclonal antibody against bovine NDH II was as described previously (2). The mouse monoclonal antibody against the 125-kDa subunit of DNA polymerase ␦ was obtained from M. Lee (New York Medical College, Valhalla, NY). Rat monoclonal antibodies against the 70-kDa subunit of replication protein A (RPA) and the 55-kDa subunit of DNA polymerase ␦ were from H.-P. Nasheuer (National University of Ireland, Galway, Ireland). The mouse monoclonal antibody against RNA polymerase II (8WG16) was from Covance Inc. (Princeton, NJ).
Plasmid Vectors and Recombinant Proteins-Construction of a baculovirus vector for the expression of full-length wild-type WRN carrying a His 6 tag at the N terminus and its purification from Sf9 insect cells through DEAE-cellulose, phosphocellulose P-11, and Ni-NTA-agarose * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
were as described previously (19). The N-terminal exonuclease domain (aa 1-256) and the middle helicase domain (aa 478 -944) of WRN were constructed by PCR synthesis of the corresponding cDNA fragments and their insertion into plasmid pTrcHis2B (Invitrogen) at the XhoI and XbaI restriction sites. The plasmids were expressed in E. coli BL21 to yield a recombinant protein carrying a His 6 tag and a Myc tag at the C terminus. Purification was on Ni-NTA-agarose.
The full-length construct of NDH II carrying a His 6 tag at the N terminus was inserted into the bacmid vector. The resulting baculovirus was expressed in Sf9 insect cells, and the corresponding protein was purified over Ni-NTA-agarose and a poly(rI)⅐poly(rC)-agarose as described previously (20).   [41][42][43]. and the middle part from aa 257 to 1160 were produced as described previously (20) and inserted into the pGEX-2T plasmid vector (Amersham Biosciences) for expression as glutathione S-transferase (GST) fusion proteins.
Co-purification of Human WRN and NDH II-A previously described protocol for the purification of bovine NDH II was followed to observe the possible co-purification of human NDH II and WRN (3). HeLa nuclear extract was prepared from ϳ100 g of cell pellet with 0.35 M NaCl in buffer containing 20 mM potassium phosphate (pH 7.8), 10 mM Na 2 S 2 O 5 , 7 mM ␤-mercaptoethanol, 10% (v/v) glycerol, and 1 mM phenylmethylsulfonyl fluoride. The nuclear extracts were chromatographed on Bio-Rex 70 (50-ml bed volume), DEAE-Sepharose (10 ml), and phosphocellulose P-11 (2 ml), which were equilibrated with buffer containing 20 mM potassium phosphate (pH 7.8), 10 mM Na 2 S 2 O 5 , 7 mM ␤-mercaptoethanol, 10% (v/v) glycerol, 1 mM phenylmethylsulfonyl fluoride, and 50 mM NaCl. All columns were eluted with NaCl gradients from 50 to 500 mM. For chromatography on phosphocellulose, the column was first eluted with a NaCl gradient from 50 to 500 mM and then stepeluted with 500 mM NaCl. The fractions obtained were monitored by Western blotting using polyclonal antibodies against bovine NDH II or WRN (antibody 200), both at a dilution of 1:1000. For Western blotting, proteins from the eluted fractions were precipitated with 10% trichloroacetic acid and resuspended in SDS-PAGE loading buffer. After SDS-PAGE, the proteins were transferred to a Hybond-C Extra nitrocellulose membrane (Amersham Biosciences) using a semidry electroblotter. A biotinylated secondary antibody and a streptavidin-biotinylated horseradish peroxidase complex (both at a dilution of 1:5000) were applied for immunodetection by enhanced chemiluminescence (ECL system, Amersham Biosciences) as described by the manufacturer.
Co-immunoprecipitation-For immunoprecipitation, HeLa cells (ϳ1 g of wet pellet) were suspended in 1 ml of 20 mM Tris-HCl (pH 7.2), 15 mM KCl, 2.5 mM MgCl 2 , 1 mM NaF, 1 mM sodium orthovanadate, 0.05% Nonidet P-40, and 5 g/ml each of the proteinase inhibitors (aprotinin, leupeptin, and pepstatin). The suspension was passed four times through a 25-gauge needle for cell opening, followed by centrifugation at 3000 ϫ g for 5 min at 4°C. The nuclear pellet was suspended in 50 mM Tris-HCl (pH 7.8), 150 mM KCl, 1 mM EDTA, 0.5 mM EGTA, 1 mM NaF, 1 mM sodium orthovanadate, 0.5% Nonidet P-40, and the proteinase inhibitors listed above and then incubated at 4°C for 30 min under rocking, followed by centrifugation at 15,000 ϫ g for 15 min at 4°C. The supernatant (nucleoplasm) was divided into two equal parts and mixed with 10 l of rabbit antiserum against bovine NDH II and the same amount of control rabbit serum or with 10 g of goat (C- 19) or rabbit (H-300) polyclonal antibody against human WRN (see above) and the same amount of control goat or rabbit IgG (Santa Cruz Biotechnology, Inc.). The immunoprecipitation mixtures were rocked for 1 h at 4°C. Then, 50 l of protein A-agarose beads were added to each mixture, followed by incubation for an additional 1 h at 4°C. The beads were finally collected by centrifugation at 15,000 ϫ g for 1 min at 4°C and washed by recentrifugation three times with buffer containing 50 mM Tris-HCl (pH 8.0), 25 mM NaCl, 0.1 mM EDTA, 0.2% Nonidet P-40, and the proteinase inhibitors listed above. Finally, the agarose pellets were dissolved in 50 l of SDS-PAGE loading buffer for Western blotting.
Far-Western Blotting-Recombinant NDH II or WRN proteins were electrophoresed through an SDS-polyacrylamide gel and then transferred to a Hybond-C Extra nitrocellulose membrane as described above. The nitrocellulose membrane with immobilized proteins was incubated for 15 min in Tris-buffered saline (25 mM Tris-HCl (pH 8.0), 140 mM NaCl, and 3 mM KCl) containing 8 M urea. Ten subsequent dilution steps were used to remove the urea. Here, one-third of the buffer was replaced with one-third of Tris-buffered saline without urea; after each step, the membrane was incubated for 10 min at room temperature. Thereafter, the nitrocellulose was incubated for 30 min in binding buffer (25 mM Tris-HCl (pH 8.0), 50 mM NaCl, 1 mM dithiothreitol, and 0.2% Nonidet P-40) supplemented with 1% blocking reagent (Roche Applied Science, Mannheim, Germany). After washing the membrane three times with binding buffer for 5 min each, purified WRN protein or NDH II (ϳ0.2 g/ml) was added to the binding buffer. After incubation with WRN or NDH II for 30 min, the binding of WRN to immobilized NDH II or vice versa was measured by immunodetection with the mouse monoclonal antibody against WRN or the rabbit antibody against NDH II at a dilution of 1:1000.
ATPase Assay-The assay was performed in a mixture (10 l) containing a given amount of WRN or NDH II (see Figs. 4 and 7); 0.2 mM ATP/[␥-32 P]ATP or GTP/[␥-32 P]GTP (ϳ50 cpm/pmol; Amersham Biosciences), and 120 M (nucleotide) M13mp18 single-stranded DNA (ssDNA), activated calf thymus DNA, or poly(rI)⅐poly(rC) double-stranded RNA (dsRNA) in buffer containing 20 mM Tris-HCl (pH 7.5), 3.5 mM MgCl 2 , 0.1 mg/ml bovine serum albumin, and 5 mM dithiothreitol. After incubation at 37°C for 30 min, the reaction was stopped by adding 1 ml of 0.5% activated charcoal (Sigma) suspended in 0.3 M perchloric acid. After vigorously shaking the mixtures for 10 min, activated charcoal was spun down by centrifugation at 15,000 ϫ g for 10 min. The 32 P i released in the supernatant was measured by scintillation counting.
Helicase Assay-The DNA unwinding activity of WRN or NDH II was measured with a partially hybridized M13mp18 ssDNA consisting of a DNA primer (45-mer) with 22 nucleotides in the duplex and 23 nucleotides as a 3Ј-non-complementary extension (3). The reaction mixture (10 l) contained a given amount of WRN or NDH II (see Fig. 5), 45 M (nucleotide) DNA substrate with the primer labeled at the 5Ј-end with [␥-32 P]ATP and T4 polynucleotide kinase, and 3 mM ATP or UTP (for NDH II) as described above for the ATPase assay. Unwinding was indicated by the displacement of the hybridized primer from M13mp18 ssDNA. This was observed after electrophoresis of the DNA through a nondenaturing 15% polyacrylamide gel (3). The radioactive signals were measured by ImageQuant after exposure of the gel to a Phosphor-Imager screen (Amersham Biosciences).
Exonuclease Assay-The exonuclease activity of WRN was measured with the DNA substrate as described above for the DNA helicase assay. Alternatively, a "D-loop" substrate was constructed by three DNA oligonucleotides as depicted in Fig. 8A. The amounts of DNA substrates and the other reaction conditions were the same as those described for the helicase assay. DNA degradation was revealed by electrophoresis of the reaction products through a 7 M urea-14% polyacrylamide gel (20 ϫ 40 cm) buffered with 89 mM Tris base, 89 mM boric acid, and 2 mM EDTA) at 1800 V for 90 min. Radioactive signals were scored by Im-ageQuant (Amersham Biosciences) as described above.

Partial Co-purification of WRN with NDH II from Human
Cells-Using the same method as described previously for the purification of NDH II from calf thymus (3,4), we observed co-purification of WRN and NDH II from human cells (Fig. 1A). Human WRN displayed an overlapping elution peak with NDH II on Bio-Rex 70 and DEAE-Sepharose (Fig. 1B). On phosphocellulose, WRN eluted at Ͼ500 mM NaCl, whereas the bulk of NDH II eluted at ϳ200 mM (Fig. 1B). However, in addition to its elution at a lower salt concentration, there was a considerable amount of NDH II that co-purified with WRN in the high B, interactions between full-length NDH II and WRN examined by Far-Western blotting. NDH II (1 g) was electrophoresed through a 10% polyacrylamide gel and then transferred to a nitrocellulose membrane, followed by Ponceau S staining. Subse-salt fractions on phosphocellulose (Fig. 1B). This result suggests a physical association between these two proteins. We also noticed that full-length recombinant WRN eluted from phosphocellulose at ϳ200 mM NaCl (data not shown). The different elution of recombinant and native WRN on phosphocellulose might have been due to protein complex formation that altered the chromatography of WRN from nuclear extracts of human cells. To address this issue, we looked for other proteins and found that DNA polymerase ␦ and the ssDNAbinding protein RPA co-purified with WRN and NDH II and were also present in the high salt fractions of phosphocellulose (Fig. 1C). This supports the conclusion that both WRN and NDH II might form a nuclear protein complex that is involved in DNA replication and/or repair.
WRN Co-immunoprecipitates with NDH II and DNA Polymerase ␦-The physical association of WRN with NDH II was examined by immunoprecipitation with antibodies directed to WRN or NDH II. As shown in Fig. 2A, WRN from HeLa nuclear extracts co-immunoprecipitated with NDH II. A similar result was also found when immunoprecipitations were carried out with an antibody against WRN (Fig. 2B). Moreover, we observed that the largest subunit (125 kDa) of DNA polymerase ␦ was present in the immunoprecipitates of WRN (Fig. 2B). These results are consistent with the finding that WRN co-purified with NDH II as well as with DNA polymerase ␦ (Fig. 1C).
Full-length NDH II and Its N-terminal dsRBD II and Cterminal RGG Box Bind Directly to WRN-To confirm direct protein-protein interactions between WRN and NDH II, Far-Western blot experiments were performed. NDH II and WRN were expressed with recombinant baculoviruses and purified from infected insect cells. Recombinant WRN (Fig. 3A) bound to NDH II immobilized on nitrocellulose after SDS-PAGE. This was not the case with bovine albumin used as a control (Fig.  3B). These results led to the conclusion that human WRN and NDH II may form physical contacts with each other.
To map the parts of NDH II that support the physical interaction with WRN, we divided the NDH II cDNA into different fragments covering N-terminal dsRBDs I and II (aa 1-256), dsRBD I (aa 1-160), and dsRBD II (aa 161-256); the center of NDH II consisting of the helicase catalytic motifs (aa 257-1160); and the C-terminal regions with the RGG box (aa 953-1269), without the RGG box (aa 953-1160), and with the RGG box alone (aa 1161-1269) (Fig. 3C). These NDH II constructs were subcloned into plasmid pGEX-2T and expressed as GST fusion proteins in bacteria. The recombinant proteins obtained were utilized to examine binding to WRN by Far-Western blotting as described above. These experiments revealed two sites of the NDH II molecule that displayed a binding signal for WRN: one at the N terminus containing dsRBD II (aa 161-256) and the other at the C-terminal RGG box (aa 1161-1269). The combined dsRBDs I and II (aa 1-256) and the expanded C terminus including the RGG box (aa 953-1269) also displayed binding to WRN, in agreement with the presence of dsRBD II or the RGG box, whereas dsRBD I (aa 1-160) alone or the C-terminal region upstream of the RGG box (aa 953-1160) had no detectable affinity for WRN (Fig. 3C). Comparable results were obtained by pull-down experiments with purified GST fusion proteins containing dsRBD I (aa 1-160) or dsRBD II (aa 161-256) of NDH II. Here again, we observed binding of WRN to dsRBD II, but not to dsRBD I (Fig. 3D).
WRN Inhibits the Nucleic Acid-dependent NTPase Activity of NDH II-Because of the physical interaction between WRN and NDH II, it was reasonable to ask whether these two proteins mutually affect each other as nucleic acid-dependent NT-Pases or helicases. This issue was first addressed by examining the nucleic acid-dependent ATPase activity of NDH II and WRN. To this end, we measured the DNA-dependent ATPase levels in increasing amounts of WRN (0 -2.1 pmol) in the absence or presence of NDH II (0.7 or 2.1 pmol) (Fig. 4A). This experiment showed that the absence or presence of NDH II did not significantly change the ATPase levels as measured for increasing amounts of WRN, suggesting that NDH II and WRN may not stimulate each other.
Because WRN and NDH II both hydrolyzed ATP in the presence of DNA, the individual contribution of the two helicases to ATPase activity is difficult to determine. In our initial attempts to address this issue, we exploited the fact that NDH II is able to hydrolyze all four rNTPs or dNTPs (3), whereas WRN uses only ATP or dATP. This difference between the two helicases enabled us to obtain evidence for inhibition of the nucleic acid-dependent GTPase activity of NDH II by WRN (data not shown). To further confirm this, we boiled WRN-or NDH II-containing solutions for 5 min and then observed the influence of heat-inactivated WRN on the DNA-dependent ATPase of native NDH II (Fig. 4B) or, conversely, the effect of heat-denatured NDH II on the DNA-dependent ATPase of native WRN (Fig. 4C). This demonstrated a striking inhibition of the DNA-dependent ATPase activity of NDH II by denatured WRN, whereas no such effect could be detected by adding inactive NDH II to WRN in the DNA-dependent ATPase assay. These results led to the conclusion that heat-stable parts of WRN may interact with the nucleic acid-binding domains of NDH II and thereby decrease its NTPase activity. On the other hand, the NDH II-binding part of WRN is not necessary for its ATPase activity.
NDH II-catalyzed DNA Unwinding Is Inhibited by the Presence of WRN-We next asked whether WRN might also suppress the helicase activity of NDH II as expected from the NTPase assays shown above. Using a partially hybridized double-stranded DNA as substrate, we could show that NDH II unwound DNA at an apparently lower rate compared with WRN (Fig. 5A). To further distinguish the effect of WRN on the DNA helicase activity of NDH II, we utilized UTP as a cofactor, which supports DNA unwinding catalyzed only by NDH II (3), but not by WRN. Under this condition, we observed that the presence of WRN diminished DNA unwinding by NDH II (Fig.  5B). This result is consistent with the above finding that WRN inhibited the DNA-dependent NTPase activity of NDH II.
The Exonuclease Activity of WRN Is Stimulated by NDH II-In addition to its 3Ј 3 5Ј helicase activity, WRN also carries 3Ј 3 5Ј exonuclease activity at its N terminus (19). Although we failed to observe stimulation or inhibition of the nucleic acid-dependent NTPase or helicase activity of WRN by NDH II, we asked whether NDH II affects the exonuclease activity of WRN. As expected, NDH II alone was devoid of any exonuclease activity, whereas under the same conditions, WRN dequently, the membrane was probed with purified full-length WRN. Bovine serum albumin (BSA; 3 g) was loaded as a control. C, WRN interacts with the dsRBD II and RGG box of NDH II. GST fusion proteins encompassing different lengths of NDH II were produced as indicated. All proteins were expressed in E. coli. Bacterial extracts containing the expressed proteins were separated by SDS-PAGE, followed by Far-Western blotting with full-length WRN as a probe. D, GST pull-down assays. Purified GST proteins (5 g each) containing the dsRBD I (aa 1-160) or dsRBD II (aa 161-256) of NDH II were coupled to glutathione-Sepharose (25-l bead volume) and mixed with full-length WRN (1 g) in binding buffer (25 l) containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM dithiothreitol, and 0.2% Nonidet 40. After 30 min, glutathione-Sepharose was spun down, washed, and suspended in SDS-PAGE loading buffer. The WRN pull-down was examined by Western blotting. One-tenth of the WRN input is shown after SDS-PAGE and silver staining. graded DNA in a 3Ј 3 5Ј direction (Fig. 6A) (19). Surprisingly, upon addition of NDH II to WRN, WRN exonuclease activity was significant stimulated (Fig. 6B).
To identify the local region of WRN that is responsible for this functional interaction, we expressed recombinant WRN proteins containing its exonuclease domain (WRN-Exo, aa 1-256) or its helicase domain (WRN-Hel, aa 478 -944). Far-Western blotting showed that the exonuclease domain (but not the helicase domain) of WRN displayed binding affinity for full-length NDH II (Fig. 7A). To examine the effect of the exonuclease domain of WRN on the enzymatic activity of NDH II, we purified WRN-Exo from bacterial extracts on Ni-NTA-agarose based on the presence of a His 6 tag at its C terminus (Fig. 7B, panel a). Surprisingly, WRN-Exo inhibited the ssDNA-dependent ATPase activity of NDH II (Fig. 7B, panel b), similar to the already observed fulllength (inactive) WRN enzyme (Fig. 4B). In contrast, the dsRNAwith 0.7 pmol (E) and 2.1 pmol (f) of NDH II. B, DNA-dependent ATPase activities of NDH II in the absence (Ⅺ) and presence (f) of inactive WRN. WRN was inactivated at 100°C for 5 min before addition to the assay mixtures. C, DNA-dependent ATPase activities of WRN in the absence (Ⅺ) and presence (f) of inactive NDH II. NDH II was heat-inactivated as described above. Unwinding is presented as a phosphorimage (panel a) and was quantified according to the percentage of displaced primer from the annealed substrates (panel b). As a control, the helicase substrate is shown after denaturation at 100°C for 5 min. B, inhibition of NDH II-catalyzed DNA unwinding by WRN. The DNA helicase assay was performed in the presence of increasing amounts of NDH II without (E) and with (•) 2.1 pmol of WRN. UTP (3 mM) was used as energy-providing nucleotide cofactor of NDH II. stimulated ATPase activity of NDH II was barely affected by WRN-Exo (Fig. 7B, panel b) or by full-length WRN (data not shown). Most important, NDH II also stimulated the exonuclease activity of WRN-Exo (Fig. 7C), as already observed for the exonuclease activity of full-length WRN (Fig. 6B). These results favor the conclusion that the exonuclease domain of WRN forms a direct contact with the dsRBD II and RGG box of NDH II and that this probably mediates the observed functional interactions between the two proteins.
NDH II Promotes the Degradation of D-loop DNA by WRN-Next, we examined whether the stimulatory effect of NDH II on the WRN exonuclease is also observed for D-loop DNA reminiscent of a recombination intermediate occurring in vivo (Fig.  8A). WRN was indeed able to degrade the invading primer from the 3Ј-end within an unpaired D-loop region. Moreover, degradation was stimulated by the presence of NDH II (Fig. 8B).
Also with this assay, we attempted to differentiate the contribution of a nucleic acid-binding domain from NDH II to D-loop degradation by WRN. Previously, we were able to remove the C-terminal RGG box of NDH II by limited trypsinmediated proteolysis (20). This experiment was repeated here to obtain an NDH II polypeptide without the RGG box but retaining the N-terminal dsRBDs (as judged from its remaining affinity for poly(rI)⅐poly(rC)-agarose) that were utilized for the purification of full-length NDH II and its deletion products containing the N-terminal dsRBDs (Fig. 8C, panel a) (20).
Truncated NDH II without the RGG box was still able to stimulate the exonuclease activity of WRN (Fig. 8C, panel b). However, addition of dsRNA (for which a dsRBD usually has higher and more specific affinity) apparently suppressed the ability of NDH II to stimulate D-loop degradation by WRN. This result supports the view that the dsRBD II of NDH II significantly contributes to the physical contact with the exonuclease domain of WRN and thereby mediates stimulated degradation of physiologically relevant DNA. DISCUSSION Werner syndrome is a premature aging disease with symptoms similar to some (but not all) of the signs of human aging; it is therefore described as a segmental progeroid syndrome (21). Aging may be due to defects in transcription, DNA repair, or both. Indeed, a function of WRN in transcription is highly likely because the total level of RNA polymerase II-catalyzed transcription decreases considerably in WRN-deficient cells (22). This has been ascribed to the presence of a transcription activation domain in WRN. Recently, a systematic search of the transcription profiles of Werner syndrome cells revealed a reduced transcription of many genes that closely resembles that of normally aging cells (23). Because some of these genes are involved in DNA repair or recombination, the genomic instability in WRN-deficient cells may be attributed also to a defect in transcription. FIG. 6. NDH II stimulates the exonuclease activity of WRN. A, exonuclease assays for WRN and NDH II. The exonuclease assays were performed with the partially annealed substrate as described under "Experimental Procedures" for the DNA helicase assay. Degradation of DNA was observed in the presence of increasing amounts of WRN or NDH II. The lengths of DNA refer to the undigested DNA primer and to the length standards dG 8 and dG 6 . B, NDH II stimulates the exonuclease activity of WRN. The exonuclease activity of WRN (0.7 pmol) was measured under the same conditions described for A in the absence and presence of NDH II (0.7 pmol). nt, nucleotides.
The biochemically observed physical interactions between NDH II and WRN suggest that these two enzymes also cooperate in vivo, e.g. at transcriptionally active domains, replication forks, DNA repair foci, or sites of DNA recombination. The most striking evidence for the presence of NDH II at transcriptionally active sites is its presence in the nucleolus, which depends on active RNA polymerase I (24). Notably, a nucleolar localization of WRN, which also seems to correlate with active rDNA transcription, has been observed (25,26). Alternatively, the association of WRN with DNA polymerase ␦ and RPA speaks for a role in DNA replication (27). Indeed, NDH II and WRN may cooperate at the same subnuclear  5-8). Undigested DNA is shown in lane 1, and addition of NDH II alone is shown in lane 9. C, dsRNA apparently inhibits the stimulation of WRN exonuclease activity by NDH II without an RGG box. Panel a, an RGG box deletion product of NDH II (130 kDa) was obtained by trypsin degradation, followed by purification on poly(rI)⅐poly(rC)-agarose as described previously (20). The purity of full-length NDH II and its RGG box-truncated form was examined by SDS-PAGE and silver staining. Panel b, the exonuclease assays were performed with the D-loop substrate and WRN in the absence (lane 1) and presence (lanes 2 and 3) of NDH II with a deletion of the RGG box. Poly(rI)⅐poly(rC) at different ratios to D-loop DNA (in nucleotides) was added to show the interference with NDH II stimulatory action on WRN exonuclease activity (lanes 4 -6).
locations, such as an rDNA locus, where the two proteins resolve DNA secondary structures that inhibit transcription and/or DNA replication (21). The highly repetitive rDNA sequences are particularly prone to illegitimate recombination, and this may be supervised and counteracted by the combined function of WRN and NDH II to avoid transcription-associated recombination events (28).
At its N terminus, NDH II contains two dsRBDs (dsRBDs I and II); and at its C terminus, it contains an RGG box, which binds both DNA and RNA (20). The WRN exonuclease domain interacts with NDH II via dsRBD II and the RGG box and thereby occludes the two domains from DNA binding, but not from RNA binding. In turn, this leads to inhibition of the DNA helicase and DNA-dependent NTPase activities of NDH II. Furthermore, dsRBD II and the RGG box directly stimulate the exonuclease domain of WRN, possibly in a manner similar to how they stimulate the helicase activity of NDH II. Despite this, the nucleic acid-binding domains, such as the dsRBD and RGG box, may also be responsible for proteinprotein interactions to coordinate the processing of DNA (29). Indeed, the dsRBD of NDH II has been shown to promote protein-protein interactions independent of its binding to RNA (30). In addition, the RGG box of NDH II mediates physical contacts with the SMN (survival of motor neurons) protein, associated with small nuclear ribonucleoproteins, small nucleolar ribonucleoproteins, spliceosomes, and RNA polymerase II (31,32). Possibly via these physical links, NDH II can be recruited to the DNA and RNA processing machinery to play a role in both supervising the genomic integrity and processing RNA. In this respect, NDH II may also be important for the quality control of RNA synthesis, which can be imposed under some genomic stress conditions that damage the transcribed template DNA (33).
According to some recent studies on mice with double knockouts of WRN and telomerase RNA (Terc), the premature aging symptoms of WRN might be related to some defects in telomere metabolism (34,35). The analyses of mitotic chromosomes from Wrn Ϫ/Ϫ /Terc Ϫ/Ϫ mice revealed increased chromosome end-toend fusions as well as an enhanced loss of telomeric lengths. Recently, a significant loss of the single sister telomere was observed in cells deficient in the helicase function of WRN. This implies an involvement of WRN in lagging strand DNA synthesis to copy the G-rich strand at the 3Ј-telomeric end (36). In fact, a functional association of WRN with DNA replication is supported by previous results showing that WRN is present in the DNA replication complex containing proliferating cell nuclear antigen and DNA topoisomerase I (37), where it promotes the traverse of DNA polymerase ␦ through G-rich secondary structures, such as those found at telomeres (38,39). Consistent with these results, association of WRN with DNA polymerase ␦ and RPA was observed in this study, and moreover, the same protein complex also contained NDH II. A recent study suggests that WRN is able to degrade the 3Ј-telomeric singlestrand overhang that is present in the D-loop structure to stabilize the t-loop conformation at chromosomal ends (40). To fulfill this activity, WRN seems to cooperate with other telomeric proteins, such as TRF2. Because NDH II is able to stimulate the degradation of D-loop DNA by WRN in vitro, a similar process might also occur in vivo. This may help to open the telomeric t-loop and thereby facilitate the replication of chromosomal ends.
A recent study suggests that the RecQ helicase of E. coli is involved in the unwinding of DNA ahead of a stalled replication fork, where a ssDNA gap is generated and subsequently coated with a RecA filament to induce an SOS response (41,42). Other RecQ helicases may have a similar role in signaling DNA damage (43), such as the Sgs1 helicase from yeast suggested to produce RPA-coated ssDNA and thereby activate the ATR/ Mec1-mediated DNA damage checkpoint. A previous study has shown that WRN co-localizes with RPA foci in the nucleoplasm after arrest of replication with hydroxyurea (44). In addition to the resolution of a stalled replication fork, e.g. at a site where it collides with a transcription bubble, the cooperated action of WRN and NDH II may generate RPA-coated ssDNA tracts that activate the ATR checkpoint kinase. In turn, this may lead to cell cycle arrest and promote DNA repair. Further studies on this issue are certainly necessary to broaden our understanding of the involvement of NDH II and WRN in the processes of human aging and/or tumorigenesis.