Introduction

Werner syndrome (WS) is an autosomal recessive disease characterized by early onset and increased frequency of many age-related features, including graying and loss of hair, atherosclerosis, cataracts, and cancer. This premature aging phenotype is due to loss of function of the WRN gene product that harbors 3' 5' helicase and 3' 5' exonuclease activities (reviewed in Oshima, 2000). WRN is a member of the RecQ helicase subfamily (reviewed in Chakraverty and Hickson, 1999) that includes BLM and RECQ4 that are defective in the cancer-prone Bloom and Rothmund–Thomson syndromes, respectively. WRN is the only human RecQ member to have exonuclease activity, suggesting that the unique WS phenotype is caused by the combined loss of helicase and exonuclease functions.

Cells from WS patients have increased chromosomal aberrations (Fukuchi et al., 1989), telomere abnormalities (Schulz et al., 1996; Tahara et al., 1997), and dramatic premature replicative senescence (Martin et al., 1970). Introduction of telomerase immortalizes primary WS cells, linking the premature senescence phenotype to defective telomere length maintenance (Wyllie et al., 2000) and suggesting a role for WRN in telomere metabolism. Moreover, telomere erosion and cellular senescence may be involved in cancer etiology and certain aging characteristics (Shay and Wright, 2001; Kim et al., 2002; Maser and Depinho, 2002). Central to telomere structure and maintenance are TRF1 and TRF2, both of which bind specifically to duplex telomeric repeat sequences (Broccoli et al., 1997), but have discrete functions. TRF1 appears to regulate telomeric length by inhibiting telomerase (van Steensel and de Lange, 1997; Smogorzewska et al., 2000; Ancelin et al., 2002). TRF2 specifically protects telomeric ends and thereby suppresses checkpoint activation leading to senescence or apoptosis and prevents end-to-end chromosomal fusions (de Lange, 2002). TRF2 may accomplish this essential telomere protection function by formation and stabilization of looped telomeric structures known as T-loops (Griffith et al., 1999).

Here, we report a physical and functional interaction between WRN and TRF2. TRF2 binds directly to WRN and mediates WRN exonuclease activity specifically on DNA substrates containing telomeric repeat sequences. Our results suggest that TRF2 may recruit WRN during normal telomeric processing, perhaps accounting for the telomeric abnormalities in WS cells. This WRN–TRF2 interaction significantly strengthens the connections between telomere maintenance, cellular senescence, and specific features of human aging including carcinogenesis.

Results

In order to determine whether WRN might be involved in telomere metabolism, we initiated studies on the effects of TRF1 and TRF2 on WRN enzymatic activities on DNA substrates containing telomeric characteristics. The existence of both helicase and exonuclease activities in WRN presents a dilemma in analysing biochemical assays on DNA substrates. Specifically, if both helicase and exonuclease are permitted to occur simultaneously, then such an in vitro experiment cannot ascertain exactly what DNA structure was acted upon even though the original structure is known, and the results can be a complex and sometimes uninterpretable combination of both activities. Another concern is that unwinding of duplex telomeric repeats by WRN helicase activity may disrupt TRF1 and TRF2 binding sites. Our approach is to examine each activity in isolation (by use of mutants or alteration of reaction conditions), thereby clarifying both the DNA structural considerations and data analysis. In this report, we focus on WRN exonuclease activity, but with respect to its physiological function, we obviously consider the combined WRN helicase and exonuclease activities on more extensive DNA structures.

The DNA substrates utilized (Figure 1) contain telomeric repeat sequences (TTAGGG)_n and approximate different telomeric structures. In particular, our telomeric overhang substrate resembles a linear telomere with a 15 nt G-strand 3' overhang; the telomeric D-loop substrate is modeled after part of the T-loop structure where the single-stranded overhang invades the duplex telomeric repeat. For each telomeric substrate, sufficient lengths of telomeric repeats were included to satisfy binding requirements for TRF1 and TRF2 (Bianchi et al., 1999). Proper annealing of telomeric substrates was confirmed by native polyacrylamide gel electrophoresis (PAGE) with and without restriction digestion of substrates at sequences outside telomeric repeats (data not shown). To assess the specific role of telomeric sequences, DNA binding and exonuclease assays were performed on a 51 bp duplex without telomeric repeats (Figure 1).

Figure 1.

Structures and sequences of DNA substrates. The telomeric D loop, telomeric duplex, telomeric 3' overhang, and nontelomeric duplex substrates are shown. Oligomer names are italicized and telomeric repeats are in shaded boxes. Positions of 3' phosphates and 5' radiolabels are indicated by 'p' and asterisks, respectively. Arrows denote sites at which WRN 3' 5' exonuclease activity is measured (in Figures 2c, 3, and 5). For the telomeric D-loop substrate, the invading strand of the D loop is in bold

Full figure and legend (104K)

To determine whether our telomeric substrates were suitable for examining interactions between TRF2 and WRN, a pull-down assay was used to measure DNA binding to His-tagged TRF2 immobilized on Ni-charged magnetic beads. After addition of labeled substrate DNA, the beads were collected and TRF2-dependent DNA depletion from the supernatant was measured after PAGE. In a concentration-dependent manner, TRF2 specifically depletes the D loop, telomeric duplex, and overhang substrates, but not the nontelomeric duplex (Figure 2a). DNA bound to beads via TRF2 was recovered by releasing His-tagged TRF2 from Ni-charged beads with 1 M imidazole buffers (data not shown). In our experiments, this pull-down assay was approximately fourfold more sensitive in detecting TRF2–DNA binding than traditional polyacrylamide gel shift assays (data not shown), presumably due to partial dissociation of TRF2–DNA complexes during electrophoresis. DNase I footprinting was also performed to confirm that TRF2 binds specifically to telomeric repeats. This analysis on the telomeric overhang substrate with excess nonspecific competitor DNA demonstrates that TRF2 binds precisely to the duplex telomeric repeat region (Figure 2b). The position of this footprint also lends support to the idea that TRF2 may bind proximal to the duplex-overhang junction, as suggested by electron microscope studies (Stansel et al., 2001). These experiments not only confirm that our telomeric substrates bind TRF2, but also provide information regarding binding affinity and spatial considerations, key factors for understanding the potential functional interactions between TRF2 and WRN.

Figure 2.

TRF2 binds to telomeric substrates and stimulates their degradation in the presence of WRN. (a) TRF2 binding to DNA substrates (5 fmol each) was measured by pull-down assay as described in Materials and methods. Individual TRF2 concentrations were as follows: 0.2, 0.5, 0.67, 1.0, and 2.0 pmol for telomeric D loop, 0.05, 0.1, 0.2, 0.5, and 1.0 pmol for telomeric overhang, 0.05, 0.1, 0.2, 0.4, and 1.0 pmol for telomeric duplex, and 0.5, 1.0, and 2.0 pmol for nontelomeric duplex. DNA binding to TRF2 is evident by its depletion with increasing TRF2 concentration. (b) DNase I footprinting of TRF2 (12–200 fmol) on the telomeric overhang substrate (5 fmol) was as described in Materials and methods. Nucleotide size markers are on the left and the boundaries of the telomeric duplex region and the TRF2 footprint are indicated by curved and right-angle brackets, respectively. (c) For exonuclease assays, DNA substrates (5 fmol each) were preincubated with or without TRF2 (15–200 fmol) followed by the addition of WRN and incubation for 10 min at 37°C. For the D-loop substrate, the WRN concentration was 4 fmol; for the other substrates, WRN was 90 fmol. The DNA products were heat-denatured and degradation of radiolabeled strands was examined by denaturing PAGE. The two bands in the untreated telomeric D-loop reactions correspond to the labeled DL80P/80 nt and INV5'A15/36 nt (invading strand) oligomers; only the invading strand is subject to WRN exonuclease activity

Full figure and legend (273K)

WRN exonuclease activity on these substrates was then examined with or without TRF2; WRN helicase activity was prevented by withholding ATP. Importantly, TRF2 alone has no detectable exonuclease activity. Using limiting WRN, TRF2 significantly increases the amount and extent of degradation from the 3' end of the invading strand of the D-loop substrate (Figure 2c, left). Since WRN alone binds, unwinds, and degrades D-loop structures with extreme efficiency (Orren et al., 2002), we reasoned that TRF2's effect might be enhanced on substrates that were lower affinity structures for WRN, but still contained duplex telomeric repeats. In support of this idea, WRN alone acted weakly on the single-stranded 3' end of the G-strand of the telomeric overhang substrate, but the addition of TRF2 greatly increased degradation of the first few nucleotides (Figure 2c, top right). When the 80 bp telomeric duplex was examined, there was little or no exonuclease activity using WRN or TRF2 alone; however, WRN and TRF2 together facilitated extensive stepwise degradation from the 3' end (Figure 2c, middle right). Notably, the TRF2 concentrations that enhanced exonuclease digestion in these experiments were in the range needed for binding to the same substrates (compare Figures 2a and c). In contrast, the nontelomeric duplex substrate was not detectably degraded by WRN with TRF2 under identical conditions (Figure 2c, bottom right). Thus, the combination of TRF2 and WRN consistently increased exonucleolytic degradation of substrates containing sufficient telomeric character to mediate TRF2 binding.

Since TRF2's positive effect on exonuclease activity was maximally observed on the telomeric duplex, this substrate was utilized to investigate its specificity. First, this substrate was incubated with or without TRF2 and either wild-type WRN or a mutant protein (WRN-E84A) that lacks 3' 5' exonuclease activity. Only the combination of wild-type WRN and TRF2 allowed degradation (Figure 3a), confirming that the 3' 5' exonuclease facilitated by TRF2 is inherent to WRN. Next, the effects of TRF1 and TRF2 on WRN exonuclease were directly compared. In these experiments, the WRN concentration was fixed, while TRF2 or TRF1 concentrations were varied. As above, WRN alone does not detectably digest this substrate, but addition of increasing concentrations of TRF2 promoted degradation by WRN (Figure 3b). Again, the range of concentration over which TRF2 mediates WRN exonuclease activity closely corresponds to TRF2's binding affinity for this telomeric duplex substrate (compare Figures 2a and 3b). Unlike TRF2, TRF1 does not facilitate WRN exonuclease activity (Figure 3c, top), even though TRF1, over this range of concentration, binds readily to this substrate by the pull-down assay (Figure 3c, bottom). Finally, the effect of TRF2 was examined on another 3' 5' exonuclease, exonuclease III (ExoIII) of Escherichia coli. In stark contrast to its positive effect on WRN exonuclease, TRF2 completely inhibited degradation of the telomeric duplex by ExoIII (Figure 3d). Taken together, these results demonstrate the highly specific enhancement of WRN's 3' 5' exonuclease activity by TRF2 on telomeric substrates.

Figure 3.

Specific enhancement of WRN exonuclease by TRF2. For (a–d), exonuclease reactions containing telomeric duplex substrate (5 fmol) were carried out for 10 min at 37°C and reaction products were analysed as described in Figures 2c. (a) Telomeric duplex was preincubated with or without TRF2 (0.75 pmol), then incubated with wild-type WRN or exonuclease-deficient WRN-E84A (30 fmol each). (b) Telomeric duplex was preincubated with TRF2 (0–750 fmol), then incubated with wild-type WRN (240 fmol). (c) (top) Telomeric duplex was incubated with TRF1 (0.2, 0.4, 1.0, 2.0 pmol) or TRF2 (0.075 or 0.15 pmol), then with wild-type WRN (240 fmol). Bottom: binding of telomeric duplex to TRF1 (0.4, 1.0, 2.0, or 5.0 pmol) was measured by pull-down assay as described in Experimental procedures. (d) Telomeric duplex was preincubated with or without TRF2 (0.2 pmol), then incubated with ExoIII. ExoIII concentrations are in units of activity (U) as defined by the manufacturer

Full figure and legend (168K)

The stimulatory effect of TRF2 could be mediated by its alteration of the DNA substrate or by a direct interaction with WRN (in the presence or absence of DNA). Immunoprecipitations were performed to test for a direct interaction between purified WRN and TRF2. An antibody raised against TRF2 effectively bound TRF2 alone and did not immunoprecipitate WRN nonspecifically (Figure 4). Importantly, when both proteins were incubated together, WRN was immunoprecipitated efficiently by TRF2, attesting to their direct physical interaction, even in the absence of DNA. Our observation of a direct interaction between TRF2 and WRN supports recent reports of coimmunoprecipitation and colocalization of these proteins in vivo (Johnson et al., 2001; Opresko et al., 2002).

Figure 4.

Coimmunoprecipitation of TRF2 and WRN. TRF2 (7.7 pmol) and/or WRN (0.3 pmol) were incubated with anti-TRF2 antibody followed by addition of Protein A Sepharose beads. The supernatant (S) and immunoprecipitated (IP) fractions were split and proteins in parallel samples were analysed by Western blotting. Pure WRN and TRF2 were loaded as markers (Mkrs). Parallel membranes were probed with anti-WRN (top panel) or anti-TRF2 (bottom) antibodies. Immunoprecipitation of WRN only occurs when TRF2 is present, indicating a direct interaction between the proteins

Full figure and legend (66K)

This TRF2–WRN interaction suggests that TRF2 might direct WRN to telomeric DNA, a possibility tested as follows. Telomeric overhang substrate and an 80-fold excess of unlabeled competitor DNA were added simultaneously to reactions containing preincubated WRN and TRF2, and exonuclease activity was measured after 10 and 30 min. Reactions containing WRN alone showed minimal digestion at 10 and 30 min (Figure 5). TRF2 markedly increases both the initial amount and extent of WRN degradation of the telomeric overhang even in the presence of excess competitor. These experiments indicate that the TRF2 can preferentially guide WRN to DNA containing-duplex telomeric repeats even when confronted with a large excess of nontelomeric DNA. After 30 min with TRF2, WRN had degraded extensively into the substrate (Figure 5), even well into the duplex region to which TRF2 binds at comparable concentrations (see Figure 2b), suggesting that following its recruitment, WRN activities may even destabilize or displace TRF2.

Figure 5.

WRN is preferentially directed to telomeric DNA by TRF2. After preincubation of WRN (30 fmol) and TRF2 (200 or 500 fmol) for 5 min at 4°C, telomeric overhang substrate (5 fmol, 0.25 ng) and salmon sperm DNA (20 ng) were added concurrently and the mixture incubated at 37°C for 10 or 30 min. Exonuclease activity was analysed as described in Materials and methods. Nucleotide size markers are shown on the right

Full figure and legend (189K)

Discussion

Our results indicate that TRF2 recruits WRN specifically to telomeric DNA substrates. This recruitment is mediated by TRF2 binding to telomeric repeats and a direct physical interaction between TRF2 and WRN. These interactions facilitate the 3' 5' exonuclease activity of WRN on three different telomeric substrates. This TRF2–WRN interaction is very specific, as another 3' 5' exonuclease (ExoIII) is severely inhibited by TRF2, and TRF1 can bind telomeric substrates, but does not influence WRN exonuclease. Intuitively, TRF2 binding to DNA might be expected to inhibit exonuclease activity in that region, as observed for ExoIII. The fact that TRF2 enhances WRN exonuclease activity on each of these structures containing telomeric repeats indicates both cooperativity and considerable flexibility in positioning of WRN on DNA by TRF2.

This study is the first to show TRF2's positive effect on the WRN exonuclease and that it parallels TRF2 binding to telomeric DNA as well as directly to WRN itself. However, both helicase and exonuclease activities of WRN are probably necessary for its proper function at telomeres or elsewhere. Here, we have not addressed TRF2's effect on WRN helicase activity, due to (1) possible disruption of TRF2 binding sites by WRN unwinding capacity and (2) WRN's inability in the presence or absence of TRF2 to unwind the telomeric overhang, telomeric duplex, and nontelomeric duplex substrates (data not shown). The latter is not surprising, as WRN alone cannot unwind blunt-ended duplexes (Machwe et al., 2002) and has limited ability to unwind long duplexes (Brosh et al., 1999). Another recent study (Opresko et al., 2002) has examined TRF2's effect on WRN unwinding on much shorter DNA substrates. Intriguingly, their study suggests that TRF2 can stimulate WRN helicase activity on short-forked substrates (22 bp duplex) with two telomeric repeats, although these substrates could not bind TRF2 in a polyacrylamide gel shift assay. Their experiments did not show an effect of TRF2 on WRN exonuclease on short-forked substrates containing telomeric repeats (Opresko et al., 2002), in marked contrast to our findings. Since our data clearly show that TRF2 specifically facilitates WRN exonuclease activity on three different (and significantly larger) substrates containing telomeric repeats, we suggest that DNA structural considerations and perhaps TRF2-binding dynamics are responsible for the differences between our findings and those reported by Opresko and colleagues. Our demonstration that TRF2 binding to telomeric repeats mediates WRN enzymatic activity is novel and may underlie recruitment of WRN to telomeric structures in vivo. Perhaps on appropriate (or physiological) telomeric structures, TRF2 could simultaneously enhance both the helicase and exonuclease activities of WRN.

This functional interaction between TRF2 and WRN implies their coordinated action in telomere metabolism, consistent with the telomeric defects and premature cellular senescence of WS cells. TRF2 protects telomeric ends from end joining and checkpoint activation pathways (de Lange, 2002), putatively by formation of the aforementioned T-loops (Griffith et al., 1999). T-loop structures may also be intermediates in the recombinational ALT (alternative lengthening of telomeres) pathway (Lundblad, 2002). At first glance, a telomeric role for WRN's 3' 5' exonuclease activity seems at odds with prevailing notions about telomeric processing and the accelerated telomere shortening of WS cells, but we can envision several scenarios for the coordinated function of TRF2 with WRN helicase/exonuclease. First, WRN could be recruited to telomeres by TRF2 to promote or reverse telomeric recombination by the ALT pathway, as WRN colocalizes with TRF2 in ALT-utilizing cells (Johnson et al., 2001; Opresko et al., 2002) and can act on specific recombination intermediates (Constantinou et al., 2000; Orren et al., 2002). Perhaps ALT or similar telomere maintenance pathways go awry when WRN is absent, resulting in accelerated telomere loss. Second, TRF2 may orient WRN for proper G-strand processing within the context of a linear or looped telomeric structure. Although the opposite directionality of WRN indicates that it is not the 5' 3' exonuclease that performs C-strand degradation following leading strand replication to generate the 3' telomeric overhangs, the G strand may require processing as part of the optimally protected form of telomeric end structure or for correct orientation of G-strand overhang proteins (hPOT1, telomerase). In line with this reasoning, TRF2 overexpression in cells not expressing telomerase leads to overall telomere shortening, including significant G-strand resection, without loss of protection until telomere length is well below the normal length at senescence (Ancelin et al., 2002; Karlseder et al., 2002). G-strand trimming requires a 3' 5' exonuclease; our data suggest that WRN, through coordinated interactions with TRF2, might carry out such controlled degradation. Without WRN, such processing might be lost, causing the appearance or persistence of end structures that destabilize telomeres and activate strand break checkpoints pathways that trigger senescence and apoptosis. Lastly, WRN may disrupt T-loop structures during the S phase, allowing replication to proceed towards chromosome ends. We have shown that WRN has extremely high affinity for D-loop structures such as those within the T loop (Orren et al., 2002). WRN's physical interaction with TRF2 (putatively bound at the T loop) may further increase the disruption of this structure by its combined helicase and exonuclease activities, an idea supported by TRF2's enhancement of WRN exonuclease on the 3' end of the invading strand of our telomeric D-loop substrate. Interactions between WRN and several replication factors, including DNA polymerase , PCNA, and RPA (Brosh et al., 1999; Lebel et al., 1999; Kamath-Loeb et al., 2000) may deliver WRN to TRF2 and telomeric DNA structures upon replication fork blockage. An inability to disrupt telomeric structures may delay and perhaps truncate replication, explaining the slow S phase and accelerated telomere shortening in WS cells, respectively (Poot et al., 1992; Schulz et al., 1996). Regardless of the mechanism, loss of WRN function likely increases occurrence of free telomeric ends that could either join to form chromosomal fusions or activate checkpoint pathways and trigger senescence or apoptosis. Notably, a single shortened (or unprotected) telomere is sufficient to elicit the latter responses (Hemann et al., 2001). We favor the last hypothesis, because it fits nicely with the WS phenotypes, the structural and protective functions of TRF2, the ability of WRN to disrupt and degrade unusual DNA structures including D loops efficiently, and the proposed roles for the RecQ family in resolving replication blockage (Chakraverty and Hickson, 1999). In this regard, WRN may also help overcome replication blocks outside of telomeric regions.

Telomere maintenance is essential for proper cell division and genomic stability. Unprotected telomeres can cause end-to-end chromosomal fusions that, in turn, initiate cycles of breakage and rejoining, possibly accounting for the rampant chromosomal aberrations observed in tumor cells (Maser and Depinho, 2002). Unprotected telomeres can also trigger cellular senescence and apoptosis that may be involved in eliciting aging characteristics (Shay and Wright, 2001; Kim et al., 2002). Thus, time-dependent appearance of unprotected telomeres may drive carcinogenesis and certain aging processes. Our finding of functional interactions TRF2 and WRN strongly supports these inter-relationships. By this reasoning, telomere defects in WS individuals may cause both increased cancer frequency from elevated levels of genetic instability and premature aging characteristics from more frequent occurrences of cellular senescence and apoptosis. By extension, the specific aging phenotype of WS individuals may point to certain normal aging characteristics in which telomere dynamics may play a role.

Materials and methods

Proteins

ExoIII was obtained from New England Biolabs. His-tagged wild-type and mutant WRN proteins were overexpressed and purified as described (Orren et al., 1999), except that 0.1% Nonidet P-40 (NP40) was included in all liquid chromatography buffers. The WRN-E84A mutant containing a glutamate to alanine substitution mutation at a conserved residue in the exonuclease domain lacks exonuclease activity, but retains helicase activity (Machwe et al., 2002). Baculovirus containing constructs for His-tagged TRF1 and TRF2 were provided by T de Lange (Rockefeller University). TRF1 and TRF2 were individually overproduced and purified as reported previously (Bianchi et al., 1997), except that TRF2 was eluted in buffer containing 300 mM imidazole, 500 mM NaCl, 20 mM Tris-HCl (pH 8.0), and 5 mM -mercaptoethanol (BME).

DNA substrates

Designations and sequences of oligonucleotides (purchased from Integrated DNA Technologies and Operon) are depicted (in the context of substrate structure) in Figure 1. For the D-loop substrate, DL80-P and DL80-D oligomers contained 3' phosphates to block WRN exonuclease at the blunt ends. DL80-P, INV5'A15, G77telo, and 51top oligonucleotides were 5'-labeled with [-³²P]ATP and T4 polynucleotide kinase, 3' phosphatase-free (Roche Molecular Biochemicals) and unincorporated radionucleotides were removed using standard methods. Labeled oligomers were annealed with partially or fully complementary unlabeled oligomers by heating at 90°C and slow cooling to 25°C. After native PAGE, annealed substrates were isolated using a gel extraction kit (Qiagen). Telomeric D-loop substrate was constructed by a two-step annealing process detailed previously (Orren et al., 2002).

Pull-down DNA-binding assay

Immobilization of His-tagged TRF1 and TRF2 on NiNTA-charged magnetic beads (Qiagen) was utilized to measure protein-specific binding of DNA substrates by a pull-down assay. All steps were carried out at 4°C. Aliquots (10 l) of bead suspension were equilibrated with bead buffer (40 mM Tris (pH 8), 50 mM NaCl, 4 mM MgCl₂, 10% glycerol, 0.05% Tween-20, and 5 mM BME), then incubated with TRF1 (0.4–5.0 pmol) or TRF2 (0.05–2 pmol) for 30 min with mixing. After two washes with bead buffer, beads were mixed with labeled DNA substrates (5 fmol) for 30 min, then collected by magnetic force. The supernatants were removed and analysed by native PAGE. After visualization of radiolabeled species using a phosphorimager (Molecular Dynamics), DNA binding was assessed by monitoring depletion of DNA substrates from the supernatant by TRF1 or TRF2 bound to the beads. No binding of DNA to beads was observed in the absence of protein. Binding of TRF1 and TRF2 to beads and recovery of DNA from TRF1- or TRF2-treated beads after elution with high imidazole (1 M) buffer were routinely monitored by SDS–PAGE and native PAGE, respectively.

DNase I footprinting

TRF2 (0–200 fmol) was incubated with telomeric overhang substrate (5 fmol) and salmon sperm DNA (10 ng) for 5 min at 4°C followed by a 2 min incubation at 37°C with DNase I (0.02 U). Reactions were stopped with equal volumes of formamide dyes (95% formamide, 20 mM EDTA, 0.1% bromophenol blue (BPB), and 0.1% xylene cyanol). DNA products were heat-denatured, separated by denaturing 14% PAGE, and visualized after gel drying by phosphorimaging.

Exonuclease assay

DNA substrates (5 fmol) were preincubated for 5 min at 4°C with or without TRF1 (0.2–2 pmol) or TRF2 (0.015–0.750 pmol), then incubated for 10 min (unless otherwise indicated) at 37°C with wild-type WRN (4–240 fmol) or WRN-E84A (30 fmol) in WRN buffer (40 mM Tris-HCl (pH 8.0), 4 mM MgCl₂, 0.1 mg/ml bovine serum albumin, 0.1% NP40, and 5 mM dithiothreitol). Where indicated, salmon sperm DNA (20 ng) was added concurrently with telomeric overhang substrate. ExoIII reactions (TRF2) were carried out likewise in the manufacturer's buffer. DNA products were analysed as described for DNase I footprinting.

Immunoprecipitation

Purified WRN (0.3 pmol) and/or TRF2 (7.7 pmol) were incubated in WRN buffer (50 l) for 1 h at 4°C, followed by 1 h of incubation with 2 g of anti-TRF2 mouse monoclonal antibody (Upstate Biotechnology). Next, 50 l of equilibrated Protein A Sepharose suspension (Amersham-Pharmacia) was added, followed by incubation at 4°C for 1 h. Protein A beads were collected by centrifugation and supernatants were removed. The beads were washed (3 200 l) with WRN buffer; bound proteins were eluted by boiling the beads in WRN buffer (30 l) supplemented with SDS (2%), glycerol (10%), BPB (0.1%), and BME (350 mM). Proteins in the supernatant and immunoprecipitated samples were separated by SDS–PAGE and detected by Western blotting using rabbit anti-TRF2 and goat anti-WRN (Santa Cruz Biotechnology) with donkey anti-rabbit-HRP and donkey anti-goat-HRP secondary antibodies, respectively, followed by ECL detection (Amersham-Pharmacia).

References

1. Ancelin K, Brunori M, Bauwens S, Koering C-E, Brun C, Ricoul M, Pommier J-P, Sabatier L and Gilson E. (2002). Mol. Cell. Biol., 22, 3474–3487. | Article | PubMed | ISI | ChemPort |
2. Bianchi A, Smith S, Chong L, Elias P and de Lange T. (1997). EMBO J., 16, 1785–1794. | Article | PubMed | ISI | ChemPort |
3. Bianchi A, Stansel RM, Fairall L, Griffith JD, Rhodes D and de Lange T. (1999). EMBO J., 18, 5735–5744. | Article | PubMed | ISI | ChemPort |
4. Broccoli D, Smogorzewska A, Chong L and de Lange T. (1997). Nat. Genet., 17, 231–235. | Article | PubMed | ISI | ChemPort |
5. Brosh RM, Orren DK, Nehlin JO, Ravn PH, Kenny MK, Machwe A and Bohr VA. (1999). J. Biol. Chem., 274, 18341–18350. | Article | PubMed | ISI | ChemPort |
6. Chakraverty RK and Hickson ID. (1999). Bioessays, 21, 286–294. | Article | PubMed | ISI | ChemPort |
7. Constantinou A, Tarsounas M, Karow JK, Brosh RM, Bohr VA, Hickson ID and West SC. (2000). EMBO Rep., 1, 80–84. | Article | PubMed | ISI | ChemPort |
8. de Lange T. (2002). Oncogene, 21, 532–540. | Article | PubMed | ISI | ChemPort |
9. Fukuchi K, Martin GM and Monnat RJ. (1989). Proc. Natl. Acad. Sci. USA, 86, 5893–5897. | Article | PubMed | ChemPort |
10. Griffith JD, Comeau L, Rosenfield S, Stansel RM, Bianchi A, Moss H and de Lange T. (1999). Cell, 97, 503–514. | Article | PubMed | ISI | ChemPort |
11. Hemann MT, Strong MA, Hao L-Y and Greider CW. (2001). Cell, 107, 67–77. | Article | PubMed | ISI | ChemPort |
12. Johnson FB, Marciniak RA, McVey M, Stewart SA, Hahn WC and Guarente L. (2001). EMBO J., 20, 905–913. | Article | PubMed | ISI | ChemPort |
13. Kamath-Loeb AS, Johannson E, Burgers PMJ and Loeb LA. (2000). Proc. Natl. Acad. Sci. USA, 97, 4603–4608. | Article | PubMed | ChemPort |
14. Karlseder J, Smogorzewska A and de Lange T. (2002). Science, 295, 2446–2449. | Article | PubMed | ISI | ChemPort |
15. Kim S-H, Kaminker P and Campisi J. (2002). Oncogene, 21, 503–511. | Article | PubMed | ISI | ChemPort |
16. Lebel M, Spillare EA, Harris CA and Leder P. (1999). J. Biol. Chem., 274, 37795–37799. | Article | PubMed | ISI | ChemPort |
17. Lundblad V. (2002). Oncogene, 21, 522–531. | Article | PubMed | ISI | ChemPort |
18. Machwe A, Xiao L, Theodore S and Orren DK. (2002). J. Biol. Chem., 277, 4492–4504. | PubMed |
19. Martin GM, Sprague CR and Epstein CJ. (1970). Lab. Invest., 23, 86–92. | PubMed | ISI | ChemPort |
20. Maser RS and DePinho R. (2002). Science, 297, 565–569. | Article | PubMed | ISI | ChemPort |
21. Opresko PL, von Kobbe C, Laine J-P, Harrigan J, Hickson ID and Bohr VA. (2002). J. Biol. Chem., 277, 41110–41119. | Article | PubMed | ISI | ChemPort |
22. Orren DK, Brosh RM, Nehlin JO, Machwe A, Gray MD and Bohr VA. (1999). Nucleic Acids Res., 27, 3557–3566. | Article | PubMed | ISI | ChemPort |
23. Orren DK, Theodore S and Machwe A. (2002). Biochemistry, 41, 13483–13488. | Article | PubMed | ISI | ChemPort |
24. Oshima J. (2000). Bioessays, 22, 894–901. | Article | PubMed | ISI | ChemPort |
25. Poot M, Hoehn H, Runger TM and Martin GM. (1992). Exp. Cell Res., 202, 267–273. | Article | PubMed | ISI | ChemPort |
26. Schulz VP, Zakian VA, Ogburn CE, McKay J, Jarzebowicz AA, Edland SD and Martin GM. (1996). Hum. Genet., 97, 750–754. | Article | PubMed | ISI | ChemPort |
27. Shay JW and Wright WE. (2001). Rad. Res., 155, 188–193.
28. Smogorzewska A, van Steensel B, Bianchi A, Oelmann S, Schaefer MR, Schnapp G and de Lange T. (2000). Mol. Cell Biol., 20, 1659–1668. | Article | PubMed | ISI | ChemPort |
29. Stansel RM, de Lange T and Griffith JD. (2001). EMBO J., 20, 5532–5540. | Article | PubMed | ISI | ChemPort |
30. Tahara H, Tokutake Y, Maeda S, Kataoka H, Watanabe T, Satoh M, Matsumoto T, Sugawara M, Ide T, Goto M, Furuichi Y and Sugimoto M. (1997). Oncogene, 15, 1911–1920. | Article | PubMed | ISI | ChemPort |
31. Van Steensel B and de Lange T. (1997). Nature, 385, 740–743. | Article | PubMed | ChemPort |
32. Wyllie FS, Jones CJ, Skinner JW, Haughton MF, Wallis C, Wynford-Thomas D, Faragher RGA and Kipling D. (2000). Nat. Genet., 24, 16–17. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Titia de Lange for providing constructs for overproduction of TRF1 and TRF2 and for helpful comments. This work was supported in part by Grant NS-008900 from the Ellison Medical Foundation to DKO.