POT1 Stimulates RecQ Helicases WRN and BLM to Unwind Telomeric DNA Substrates * □ S

Defects in human RecQ helicases WRN and BLM are responsible for the cancer-prone disorders Werner syndrome and Bloom syndrome. Cellular phenotypes of Werner syndrome and Bloom syndrome, including genomic instability and premature senescence, are consistent with telomere dysfunction. RecQ helicases are proposed to function in dissociating alternative DNA structures during recombinationand/orreplicationattelomericends.Herewereport that the telomeric single-strand DNA-binding protein, POT1, stronglystimulatesWRNandBLMtounwindlongtelomericforked duplexes and D-loop structures that are otherwise poor substrates forthesehelicases.Thisstimulationisdependentonthepresenceof telomeric sequence in the duplex regions of the substrates. In contrast, POT1 failed to stimulate a bacterial 3 (cid:2) –5 (cid:2) -helicase. We find that purified POT1 binds to WRN and BLM in vitro and that full-length POT1 (splice variant 1) precipitates a higher amount of endogenous WRN protein, compared with BLM, from the HeLa nuclear extract. We propose roles for the cooperation of POT1 with RecQ helicases WRN incubated for 1 h °C.Wellswerewashed,andsecondaryantibody(1:10,000,anti-rabbit IgG-horseradish peroxidase; Vector Laboratories) was and incubated fo r 1 h at 37 °C.After washes, bound WRN, BLM, or APE1 was detected with o -phenylenediamine dihydrochloride followed by termination with 3 M H 2 SO 4 . Absorbance was read at 490 nm. Values were normalized to the positive control (signal from wells coated with WRN, BLM, or APE1). To ensure that the interaction was not mediated by DNA, control reactions included either 5 (cid:1) g/ml DNase I (Calbiochem) or 10 (cid:1) g/ml ethidium bromide during the binding step. To determine thedissociationconstant( K d )fortheWRN-POT1v2andBLM-POT1v2 complexes, the fraction of immobilized POT1v2 bound by WRN or BLM was calculated, and the data were analyzed by Hill plot and Scat-chard binding theory, as described previously (30). For this analysis and theplotsinFig.5, D and E ,therelativeabsorbancevalueswerecorrected for background in the BSA alone control.

Deficiencies in human RecQ helicases, WRN, BLM, and RecQL4, are responsible for genomic instability disorders, namely Werner syndrome (WS), 4 Bloom syndrome (BS), and Rothmund-Thomson syndrome, respectively. WS patients display the early onset of many age-associated pathologies including graying and loss of hair, wrinkling of the skin, cataracts, type II diabetes, osteoporosis, cardiovascular disease, and a high frequency of sarcomas (1,2). BS patients exhibit severe growth retardation, immunodeficiency, sun-induced facial erythema, and a greatly increased predisposition to a wide range of cancers that develop early in life (3). Cells derived from WS and BS patients display increased chromosomal rearrangements and deletions, defects in DNA replication and homologous recombination, and decreased replicative life spans (reviewed in Refs. 2 and 3).
Cellular phenotypes of WS and BS include features consistent with telomere dysfunction. Cells deficient in BLM helicase display increased telomere associations between homologous chromosomes (4). Cells lacking WRN, or those that express a dominant-negative WRN mutant, exhibit increased telomere loss (5) and loss of telomeres from single sister chromatids (6). The forced expression of telomerase rescues both WS and BS primary fibroblasts from premature replicative senescence (7,8). Furthermore, mice singly null for Wrn appear normal; however, the late generation mice null for both Wrn and Terc have shortened telomeres and develop pathologies resembling human WS (9). Mutations in Wrn and Blm in Terc null mice also enhance and accelerate the development of phenotypes associated with telomere dysfunction (10). Collectively, these studies indicate that the RecQ helicases WRN and BLM are important for proper telomere maintenance and function.
Telomeres are protein-DNA complexes that protect the ends of linear chromosomes, and consequences of telomere dysfunction include chromosome end fusions and genomic instability, apoptosis, or senescence (reviewed in Ref. 11). Human telomeres consist of 5-15 kb of TTAGGG tandem repeats and terminate in a 3Ј single strand (ss) G-rich tail that serves as the substrate for telomerase-mediated elongation in an open telomere state. In one proposed capped or closed state, this tail loops back and invades the telomeric duplex tract, forming an intratelomeric D-loop and a large lasso-like t-loop structure (12). This t-loop is stabilized by a protein complex that includes telomere repeat binding factors (TRF) 1 and 2, which bind duplex (TTAGGG) n DNA and regulate telomere length and access of the 3Ј tail to telomerase and/or nucleases (13,14). Human POT1 (protection of telomeres-1) protein (15) is also an important regulator of telomere length (16 -18) and binds specifically to telomeric ssDNA (19,20). POT1 associates with TRF1 and TRF2 through the interaction with other telomeric proteins including TPP1 (formerly termed PIP1, PTOP, or TINT1) and TIN2 (reviewed in Refs. 21 and 22). Collectively, these proteins cooperate with telomeric DNA structure in proper telomere maintenance and capping.
The precise roles of RecQ helicases in telomere maintenance are unclear; however, studies suggest they likely function in recombination and/or replication at telomeric ends. In budding and fission yeast, RecQ helicases function in an alternative pathway for lengthening of telomeres (ALT) that occurs via recombination in telomerase-negative cells (23)(24)(25). WRN and BLM have been found to associate with telomeres in human ALT cell lines (26), and WRN was found at telomeres in S-phase human primary fibroblasts (6). WRN and BLM are 3Ј-5Ј DNA helicases that are capable of dissociating telomeric D-loop structures (4,26). The WRN protein also contains a 3Ј-5Ј-exonuclease that cooperates with the helicase activity to release the invading tail of a telomeric D-loop (26). In addition, these helicases unwind G-quadruplex DNA structures that form readily in telomeric sequences (27). Therefore, the RecQ helicases are thought to participate in telomere maintenance by resolving complex structures at telomeric ends to facilitate DNA replication or recombination pathways.
The WRN and BLM helicases are poorly processive and likely function with co-factor proteins in a complex at the telomeres. TRF2 protein physically binds to WRN and BLM and stimulates these helicases to unwind relatively short duplex substrates that are telomeric or nontelomeric (4,28). However, these helicases require replication protein A (RPA) to completely unwind longer duplex substrates in vitro (29 -31). RPA interacts physically with WRN and BLM (29,32) and binds to the partially unwound single strands to prevent their reannealing upon helicase dissociation. POT1 resembles RPA in that it also interacts with ssDNA via oligonucleotide/oligosaccharide binding (OB) folds, but it differs in having high specificity for telomeric DNA sequences (15,19,20,33). Therefore, we asked whether POT1 could facilitate WRN and BLM unwinding of longer telomeric substrates. We tested two POT1 splice variants: the full-length version (v1) and a less abundant C-terminal truncated version (v2) (19). Both POT1 variants stimulated the WRN and BLM helicases to unwind telomeric forked duplex and D-loop structures that were poor substrates for these helicases alone. In contrast to RPA, optimal stimulation was dependent on the presence of telomeric sequence in the duplex regions of the substrates. Furthermore, we observed that WRN and BLM physically interact with POT1.

EXPERIMENTAL PROCEDURES
Proteins-Recombinant histidine-tagged wild type WRN protein and the exonuclease-dead WRN mutant (X-WRN, E84A) were purified using a baculovirus/insect cell expression system as described previously (26,34). Recombinant histidine-tagged BLM was overexpressed in Saccharomyces cerevisiae and purified as described previously (35). Recombinant human POT1 variant 1 and variant 2 proteins were purified using a baculovirus/insect cell expression system as described previously (19). Splice variant 1 is the full-length protein, and variant 2 comprises the N-terminal half, including the two OB folds responsible for DNA binding (19,36). The purity of these protein preparations was analyzed by SDS-PAGE and Coomassie staining as shown in Fig. 1A. Human RPA was kindly provided by Dr. Mark Kenny (Albert Einstein Cancer Center, New York). Bacterial UvrD helicase was a generous gift from Dr. Steve Mattson (University of North Carolina, Chapel Hill). Human recombinant APE1-purified protein was provided by Dr. David Wilson (National Institute on Aging, Baltimore) (37).
DNA Substrates-The forked duplex substrates consisted of 15-mer ssDNA tails followed by 34 bp of duplex DNA that contained either the (TTAGGG) 4 sequence or nontelomeric sequence and were constructed as described previously (28,31). Substrates were radiolabeled at the 5Ј-end of the forked side. The D-loop Tel DS substrate containing the (TTAGGG) 4 sequence in the 33-bp duplex portion of the invading strand was constructed by annealing the BT, BB, and invading strand (INV) oligonucleotides as described previously (26). Variant D-loops were similarly constructed, and all D-loops contained a 5Ј-end radiolabeled INV strand. The D-loop Tel SS contained the (TTAGGG) 4 sequence in the displaced single strand region of the D-loop instead and was constructed by annealing the BT, BBmx, and INVtel oligonucleotides. The D-loop Non Tel contained only nontelomeric sequence and was constructed by annealing the BT, BBmx, and INVmx strands. The following oligonucleotides were used: INVtel, 5Ј-CGTGACCAGGAC-GTGAGTCTGGAGTGCAGAGGGTTAGGGTTAGGGTTAGGG-TTAGGGTTAGGGACAATCATCCTGACTGCAGACCGAGCT-TGA; INVmx, 5Ј-CACCATCCAGTTCTCTTTTGAGAACTGGAT-GGTGTATCACATTGCGTTGATGGGACCGTTAACGCTC; and BBmx, 5Ј-TCAAGCTCGGTCTGCAGTCAGGATGATTGTGAGC-GTTAACGGTCCCATCAACGCAATGTGATATCTGCACTCGA-GACTCACGTCCTGGTCACG.
Helicase and Exonuclease Reactions-Reactions were performed in standard reaction buffer (31), unless otherwise indicated. DNA substrate and protein concentrations were as indicated in the figure legends. The reactions were initiated by the addition of WRN or BLM protein and were incubated at 37°C for 15 min. For the reactions containing WRN (total volume ϭ 30 l), a 10-l aliquot was mixed with formamide stop dye and run on a 14% denaturing polyacrylamide gel. The remainder of the reactions containing WRN (20 l) or the reactions containing BLM helicase (20 l) were added to 10 l of 3ϫ native stop dye supplemented with 75 g/ml proteinase K and a 10ϫ molar excess of unlabeled competitor oligonucleotide (28). The products were deproteinized for 30 min at 37°C and were then run on 8 or 12% native polyacrylamide gels as indicated in the figure legends. Products were visualized using a PhosphorImager, and quantitation was performed using ImageQuant software (Amersham Biosciences).
For the reactions with WRN on the forked telomeric substrate, the percent of unreacted substrate and the percent of each major product band in the native gels (as defined by displaced oligonucleotide size, Fig.  2, A and E) were calculated as a fraction of the total radioactivity in the reaction lane. For the reactions with WRN and the various D-loop substrates, the percent of displaced near full-length products (longest) was calculated as a function of total displaced products in the lane (sum of the longest product band and the collective shorter product bands) in the native gels (Fig. 3A). For the reactions with BLM helicase or X-WRN, the percent of total product displacement was quantitated as a fraction of total radioactivity in the reaction lane as described previously (28). Values were corrected for background in the no enzyme control and heat-denatured substrate lanes.
ELISA Detection of Protein Interactions-ELISA was conducted as described previously (30), with some modification. The blocking and binding steps were performed in phosphate-buffered saline containing 3% BSA and 0.1% Tween 20. Wells were coated with 75 ng of purified POT1v1 or POT1v2 diluted in carbonate buffer (50 l) or with BSA as a background control by incubation for 2 h at 37°C. Independent wells were also coated with WRN, BLM, or APE1 as positive controls. After blocking, various concentrations of WRN, BLM, or APE1 (negative control) protein were added (50 l) to the wells coated with POT1v1 or POT1v2 (see Fig. 5 legend) and incubated for 2 h at 37°C. Following washes, primary antibody (1:1,500, anti-rabbit IgG against WRN (Novus); 1:1000, anti-rabbit IgG against BLM (Abcam); or 1:100 antirabbit IgG against APE1 (Trevigen)) was added and incubated for 1 h at 37°C. Wells were washed, and secondary antibody (1:10,000, anti-rabbit IgG-horseradish peroxidase; Vector Laboratories) was added and incubated for 1 h at 37°C. After washes, bound WRN, BLM, or APE1 was detected with o-phenylenediamine dihydrochloride followed by termination with 3 M H 2 SO 4 . Absorbance was read at 490 nm. Values were normalized to the positive control (signal from wells coated with WRN, BLM, or APE1). To ensure that the interaction was not mediated by DNA, control reactions included either 5 g/ml DNase I (Calbiochem) or 10 g/ml ethidium bromide during the binding step. To determine the dissociation constant (K d ) for the WRN-POT1v2 and BLM-POT1v2 complexes, the fraction of immobilized POT1v2 bound by WRN or BLM was calculated, and the data were analyzed by Hill plot and Scatchard binding theory, as described previously (30). For this analysis and the plots in Fig. 5, D and E, the relative absorbance values were corrected for background in the BSA alone control.
GST-POT1-Sepharose Pull-down Assay-GST-tagged POT1v1 and POT1v2 were expressed in baculovirus/insect cell cultures as described previously (19). 500-ml cultures were harvested and resuspended in 8 ml of 25 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, including protease inhibitors. Samples were lysed by sonication and incubated with 20 g/ml DNase I (Calbiochem) for 30 min on ice. Samples were then cleared by centrifugation.
GST alone or GST-POT1v1 or -POT1v2 from 1 ml of lysates was bound to glutathione beads and then incubated with HeLa nuclear extracts (400 ul) prepared as described previously (28,38). After washing, total protein was eluted in sample buffer by boiling and was analyzed by SDS-PAGE and Western blot analysis. Loading of the GST-tagged POT1 variants were determined by Amido Black staining. Membranes were probed with a mouse monoclonal anti-WRN antibody (1:500, BD  -22). The substrate was incubated with either 5 nM X-WRN protein alone (lanes 2, 9, 16, and 20) or together with increasing concentrations of POT1v1 (5, 15, or 40 nM) (lanes 3-5 or 10 -12, respectively) or 8 nM RPA (lanes 17 and 21) for 15 min at 37°C. The reactions were run on a 12% native gel. OE, heat-denatured substrate. C, the percent displaced substrate was calculated for the reactions in A (lanes 2-5), as described under "Experimental Procedures," and plotted as a function of the molar ratio of POT1v1 to X-WRN. D, the telomeric forked duplex (0.5 nM) was incubated with either 5 nM X-WRN alone (f) or together with 40 nM POT1 v1 (•) at 37°C, and reaction aliquots were terminated at 0, 0.5, 1, 2, 4, 8, and 16 min. The reactions were analyzed on a 12% native gel, and percent displacement was calculated and plotted as a function of time.

POT1 Stimulation of the WRN Helicase
Activity-To determine whether POT1 could promote WRN helicase unwinding, we first tested full-length POT1 (splice variant 1). Coomassie-stained gels of the purified proteins used for this study, and the subsequent experiments described in this paper, are shown in Fig. 1A as described previously (19,28). For the helicase substrate we used a 34-bp forked duplex structure that contained 15-mer ssDNA tails at one end (Fig. 1B). Both the WRN helicase and exonuclease are active on this substrate (31). Therefore, we first tested an exonuclease-dead WRN mutant (X-WRN) to measure directly the effects on the WRN unwinding activity. X-WRN contains a point mutation (E84A) that inactivates the exonuclease activity (39). The duplex region of the telomeric fork consists of the (TTAGGG) 4 sequence followed by 10 bp of unique sequence, whereas the telomeric sequence is scrambled in the mixed fork (Fig. 1B). Minimal displacement of either the telomeric forked duplex (Fig. 1B, lane 2) or the mixed forked duplex (lane 9) was achieved by X-WRN alone. However, the incubation of X-WRN with increasing concentrations of full-length POT1v1 resulted in up to a 10-fold increase in displacement of the telomeric fork, from 3 to 35% (Fig. 1B, lanes 2-5, and C). However, POT1v1 did not simulate X-WRN to unwind the mixed forked duplex (Fig. 1B, lanes 10 -12), in contrast to RPA (lane 21). Time course analyses indicated that an 8-fold molar excess of POT1v1 increased the estimated rate of X-WRN strand displacement dramatically by 10-fold, from 0.01 bp/min/X-WRN (monomer) to 0.1 bp/min/X-WRN (monomer), of the telomeric fork (Fig. 1D). These data indicate that POT1v1 directly stimulates the WRN helicase to unwind the 34-bp telomeric substrate.
Next we examined the activity of wild type WRN in the presence of full-length POT1v1. On the 34-bp forked duplex, the WRN helicase initiates unwinding at the forked end, whereas the 3Ј-5Ј-exonuclease initiates digestion at the blunt end (31) (Fig. 7C). Therefore, the reactions were run both on a native gel to visualize the unwound products ( Fig. 2A) and on a denaturing gel (Fig. 2B) to better visualize the exonuclease products. As reported earlier (31), the WRN helicase displaced shortened versions of the DNA strand ( Fig. 2A, lane 2), because of the action of the exonuclease activity (Fig. 2B, lane 2). Once the DNA strand of the forked duplex is unwound, further digestion is limited because WRN alone is largely inactive on ssDNA (39) and does not digest telomeric ssDNA under the reaction conditions used here (supplemental Fig. 1). After incubation of the telomeric fork with WRN (2.5 nM), we observed the loss of the shorter products in favor of longer products as a function of POT1v1 concentration (2.5, 7.5, and 20 nM) (Fig. 2, A and  B, lanes 5-7). The most prominent products resulted from digestion termination at the G runs of the telomeric sequence yielding either 39-, 33-, and/or 27-nt products (Fig. 2, A and B, see the nucleotide sequence of the top strand). POT1v1 stimulated the WRN helicase to displace longer versions of the DNA strand (39 and 33 nt) and decreased the extent of WRN exonuclease digestion. In contrast, T4 single strandbinding protein (gp32) failed to stimulate the WRN helicase, as reported previously (data not shown and see Ref. 29). RPA was added as a positive control and promoted WRN displacement of the full-length strand (Fig.  2, A and B, lane 4). Unlike POT1, RPA binds ssDNA without sequence specificity and thus can pre-load on the substrate and coat both unwound strands.
To determine whether POT1v1 directly inhibits the catalytic activity of the WRN exonuclease, we conducted the reactions in the absence of helicase activity by omitting ATP. Under these conditions, the exonu-clease digested further until the duplex was shortened to thermally unstable lengths (Fig. 2, A and B, lane 3). In the absence of WRN helicase activity, POT1v1 did not significantly alter the pattern of WRN exonuclease products (Fig. 2, A and B, lanes 8 -10). This suggests that POT1v1 does not directly inhibit the WRN exonuclease activity, but rather alters the pattern of digested products by facilitating the WRN helicase. This is consistent with previous reports for RPA (31).
Consistent with the X-WRN results, POT1v1 required telomeric sequence in the substrate to promote the helicase activity of the wild type WRN. Although RPA stimulated the WRN helicase to unwind the full-length DNA strand of the mixed fork (Fig. 2C, lane 4), increasing concentrations of POT1v1 did not alter the pattern of WRN displaced products (Fig. 2C) or digested products (Fig. 2D), either in the presence or absence of ATP.
The POT1 C Terminus Is Not Required for WRN Helicase Stimulation-POT1 variant 2 has the same affinity and specificity for telomeric ssDNA as the full-length variant 1 (19) but is missing the C-terminal domain that interacts with TPP1 and perhaps other proteins (Fig. 2G). Therefore, we next tested the ability of v2 to facilitate the WRN helicase. Similar to v1, increasing concentrations of POT1v2 resulted in an increase in displaced longer products and a decrease in the shortest products (Fig. 2, E and F, lanes 4 -6). The stimulation was specific to the telomeric substrate (Fig. 2, E, lanes 11-13, and F, lanes 10 -12). Next we compared the extent of WRN stimulation by an 8-fold molar excess of either POT1v1 or POT1v2. The bands in the native gels (Fig. 2, A and E) representing the unreacted telomeric substrate and the most prominent displaced products (unwound DNA strands 49, 39, 33, or Յ27 nucleotides long) were quantitated and calculated as a percent of total radioactivity in the lane (Fig. 2H). These results show that POT1v1 or POT1v2 increased the percent of unwound full-length product (49 nt) by at least 57%, 2.3-or 3-fold, respectively, compared with WRN alone (Fig. 2H). Release of a 39-mer product occurs when digestion proceeds through the 3Ј-terminal 10 bp of unique sequence but stops at the telomeric (TTAGGG) 4 repeats (see substrate sequence in Fig. 2A). The percent of 39-mer unwound products was increased 3-or 4-fold in the presence POT1v1 or POT1v2, respectively. Incubation of WRN with either POT1v1 or POTv2 also decreased the percent of the shortest products displaced (Ͻ27 nt) by 2.5-or 4.6-fold, respectively. POT1v2 was somewhat more effective at promoting the WRN helicase. However, the data indicate that the C-terminal protein-protein interaction domain specific to variant 1 is not required for WRN helicase stimulation and that the POT1 DNA-binding domain (OB folds) is sufficient for this activity.
POT1 Promotes WRN-mediated Release of Intact Telomeric Tails from a D-loop Structure-Telomeric ends are proposed to be capped by a t-loop, which involves the formation of a D-loop structure (12). A similar D-loop may also be an important intermediate in the recombination-based ALT pathway. Therefore, we tested the action of the WRN helicase and exonuclease on a telomeric D-loop in the presence of POT1v1. D-loop structures were designed to contain the (TTAGGG) 4 sequence in either the duplex regions with the invading strand (Tel ds) or in the single strand region (Tel ss). A third D-loop consisted only of nontelomeric sequence (Fig. 3A). The invading strand in all three D-loops formed a 33-bp duplex region, similar to the forked duplexes, and was 5Ј-radiolabeled. The WRN exonuclease acted at the 3Ј-end of the invading tail (Fig. 3C, lanes 2, 7, and 12), and shortened versions of this strand were released by the WRN helicase (Fig. 3B, lanes 2, 8, and  14), as reported previously (26). For the Tel ds D-loop, the prominent exonuclease products corresponded to digestion termination at the G 3 runs within the telomeric repeats (Fig. 3C, lanes 1-5), whereas promi-nent termination sites were not observed with the other D-loop substrates (Fig. 3C, lanes 6 -15).
The addition of increasing POT1v1 concentrations to the reactions containing the Tel ds D-loop and WRN resulted in the disappearance of shorter products and an increase in the longest unwound product (near full length) (Fig. 3B, lanes 1-6). An 8-fold molar excess of POT1v1 over WRN increased the proportion of displaced near full-length product by 3.3-fold, from 17 to 56%, as a function of total displaced products (Fig.  3D). POT1v1 had less of an effect on WRN activity with other D-loop substrates. An 8-fold molar excess of POT1v1 over WRN altered the proportion of displaced longest products from 59 to 74% (1.2-fold) for Tel ss and from 54 to 78% (1.4-fold) for the nontelomeric D-loop, as a function of total products (Fig. 3). In the absence of ATP, and thus no WRN helicase activity, POT1v1 did not detectably alter the WRN exonuclease activity on any of the D-loop substrates (supplemental Fig. 2). In addition, a 7-fold molar excess of the RAD51 protein, which also interacts with WRN 5 and binds to ssDNA, failed to alter the pattern of WRN helicase and exonuclease products on the Tel ds D-loop (supplemental Fig. 3). Therefore, POT1v1 was more effective in stimulating the WRN helicase to release longer invading tails when the D-loop substrate contained telomeric sequence in the duplex region.

POT1 Variants Facilitate BLM Unwinding Specifically of the Tel ds D-loop-
We next tested the ability of the POT1 variants to stimulate unwinding by the BLM RecQ helicase. We observed that the BLM helicase by itself released the invading strand from all three D-loop substrates but was particularly poor at displacing the Tel ds D-loop (Fig. 4,  A, D, and E). For the Tel ds D-loop we observed a striking increase in the displacement of the invading strand by the BLM helicase upon addition of POT1v1 or POT1v2 (Fig. 4, A and B). An 8-fold molar excess of POT1v1 or POT1v2 over BLM helicase dramatically increased the % displacement from 4 to 25 or 41%, respectively. This corresponds to a 6and 10-fold stimulation of BLM helicase activity by POT1v1 and POT1v2, respectively. Nearly complete displacement (80%) was achieved with increased concentrations of both BLM and POT1v2 (Fig.   4, A, lanes 12-15, and B). Furthermore, as observed for X-WRN (Fig. 1), an 8:1 molar ratio of POT1v1 to BLM resulted in a 10-fold increase in displacement of the 34-bp telomeric fork, from 3 to 36% displacement (data not shown). We next tested a bacterial 3Ј-5Ј-helicase, UvrD, to determine whether POT1 could stimulate unwinding by an unrelated helicase, simply by binding to the released DNA strands. Both POT1v1 and v2 failed to stimulate UvrD helicase to unwind the Tel ds D-loop (Fig. 4C) or the 34-bp telomeric fork (data not shown). Next, we tested BLM helicase activity on the Tel ss and nontelomeric D-loops, in the presence of the POT1 variants. The BLM helicase was more active on the Tel ss and nontelomeric D-loops, 66 and 29% displacement, respectively, compared with Tel ds (Fig. 4, D and E). We observed similar results with X-WRN (data not shown). However, neither the addition of POT1v1 nor v2 significantly increased the percent of strand displacement by the BLM helicase, in contrast to RPA (Fig. 4, D and E).
POT1 Variants Interact Physically with BLM and WRN-Because both POT1 and RPA facilitate unwinding by WRN and BLM helicases, and RPA binds the RecQ helicases (reviewed in Ref. 40), we predicted that POT1 may also interact physically with WRN and BLM. ELISAs were conducted by coating the wells of microtiter dishes with BSA (control) or POT1 variants followed by incubation with increasing concentrations of either WRN or BLM protein. Bound WRN or BLM was detected using antibodies against each, respectively, and colorimetric analysis. Fig. 5, A-C shows the signals observed in control reactions compared with those obtained at the highest concentrations of WRN and BLM tested in Fig. 5, D-E. The signals obtained with an 8:1 molar ratio of helicase to POT1 variant were dramatically increased above background (BSA alone) and approached those in the positive controls (wells coated with WRN or BLM) Fig. 5, A and B. In contrast, that addition of WRN or BLM to wells coated with BSA yielded signals close to background (Fig. 5, A and B). As a negative control, we tested for binding of the POT1 variants to an unrelated, nontelomeric protein. For this we chose human APE1, a protein that functions in base excision repair for which we previously established ELISA conditions (41). The addition of an 8-fold molar excess of purified APE1 to wells coated with POT1v1 or POT1v2 yielded signals similar to background that were greatly reduced compared with the positive controls (Fig. 5C). These data attest to the specificity of the interactions detected by ELISA between WRN and the POT1 variants and between BLM and POT1v2.
To assess further the affinity of the POT1 variants for the RecQ helicases, we first tested WRN protein and observed that the colorimetric signal increased as a function of increasing WRN concentrations for both POT1 variants. However, wells coated with POT1v2 yielded a consistently higher and saturable signal upon incubation with WRN, at the same concentrations compared with POT1v1 (Fig. 5D). Therefore, we also tested for BLM binding to POT1v2. The binding curves for both helicases started to plateau after equal molar POT1v2 and WRN or POT1v2 and BLM concentrations, whereas WRN binding to POT1v1 did not reach saturation at the tested ratios (Fig. 5, D and E). To test for the possibility that the interaction was mediated by DNA, an experimental point within the linear range of the binding curves (1:1 molar ratio) was repeated in the presence of DNase I. No significant reduction in the signal was detected for any of the interactions (Fig. 5, D and E). To determine the strength of the physical interactions between both helicases and POT1v2, the data in Fig. 5, D and E, were analyzed by Hill plots, and apparent dissociation constants (K d ) were determined using the Scatchard binding theory as described previously (30). The apparent dissociation constants for the POT1v2-WRN and POT1v2-BLM interactions were 16 and 12 nM, respectively. These analyses revealed a moderate affinity between purified recombinant POT1v2 and the RecQ helicases WRN and BLM and a weaker but detectable interaction between WRN and POT1v1.
We used a pull-down assay to test for the association of a glutathione S-transferase (GST)-POT1v1 and -POT1v2 with the endogenous human RecQ helicases in HeLa nuclear extracts (NE). Glutathione beads were bound with GST-POT1v1, GST-POT1v2, or the GST protein and incubated with HeLa NE (400 l), and the bound proteins were analyzed by SDS-PAGE and Western blot. Amido Black staining of the membrane indicated the relative amounts of GST-POT1v1, GST-POT1v2, and GST control proteins that were successfully precipitated (Fig. 6); these proteins are indicated by arrows on the membrane. Additional bands on the membrane represent proteins that were co-precipitated by binding the GST-tagged proteins, and the intense lower bands are likely GST alone. POT1 has been reported to interact with a number of proteins (42). Western blot analysis of the membrane with antibodies against WRN indicated that WRN co-precipitated with GST-POT1v1 and to a lesser extent with GST-POT1v2 but not detectably with the GST protein control (Fig. 6). Comparison with the input indicates that ϳ15% of the WRN protein was precipitated with GST-POT1v1. These results differ from the ELISA data with purified proteins, which showed that WRN binds more strongly to POT1v2. One possible explanation is that the C-terminal domain of POT1v1, which is absent in POTv2, may enhance interaction with WRN in a cellular context, either by directly binding WRN or indirectly via interaction with other proteins. The C-terminal domain of POT1v1 has been reported to interact with proteins (42) that would be present in the HeLa nuclear extract but not in the in vitro ELISA experiment. The presence of other nuclear proteins may influence the interactions between the RecQ helices and POT1. The membrane was then stripped and re-probed for BLM protein.
Under these conditions, BLM co-precipitated weakly with both GST-POT1v1 and GST-POT1v2 but not detectably with GST protein alone.

DISCUSSION
The POT1 (protection of telomeres) protein is a human telomerespecific single-stranded DNA-binding protein, and the only one identified thus far. Here we report that both POT1 splice variants 1 and 2 strongly stimulated the RecQ helicases WRN and BLM to unwind telomeric substrates that were poorly unwound by these enzymes alone. For simplicity, where similar results were achieved with both POT1 variants, we will use POT1 to represent both. We observed that POT1 stimulated the WRN helicase to displace longer versions of the DNA strand from a 34-bp telomeric forked duplex and decreased the extent of WRN 3Ј-5Ј-exonuclease digestion of this strand. POT1 did not influence WRN activities on a nontelomeric forked duplex. POT1 promoted both WRN and BLM to release the invading tail of a D-loop more effectively when there was telomeric sequence in the duplex portion, compared with telomeric sequence in the ssDNA region. We observed that POT1 altered the pattern of unwound, digested products to favor WRN dissociation of longer strands both on the telomeric forked duplex and the Tel ds D-loop (Figs. 1-3). We considered the possibility that POT1 inhibited WRN exonuclease activity on the telomeric substrates, similar to the TRF2 inhibition of WRN digestion observed previously with the Tel ds D-loop (26). However, we favor a mechanism whereby POT1 limits the WRN exonuclease progression by increasing the rate of WRN helicase-mediated strand displacement. As reported previously (31), the WRN helicase influences the WRN exonuclease by altering the DNA structure, making it a poor substrate for the exonuclease (i.e. converting dsDNA to ssDNA). Thus, the rate of duplex displacement affects the extent to which the substrate is degraded by the exonuclease. Under the reaction conditions used here, we did not detect any WRN digestion of telomeric ssDNA, and POT1 did not alter WRN digestion in the absence of the WRN helicase (Fig. 1, and supplemental Figs. 1 and 2). However, POT1 dramatically increased the percent of telomeric substrates unwound by BLM or an exonuclease-dead WRN mutant and directly increased the rate of X-WRN strand displacement by 10-fold ( Figs. 1 and 4).
A model for the modulation of WRN activities by POT1 on the telomeric forked duplex is shown in Fig. 7 and can be extended to the Tel ds D-loop. The WRN helicase and exonuclease cooperate to decrease the length of the duplex until it becomes unstable and melts. If the duplex remains stable after WRN dissociation, the partially unwound strands reanneal (Fig. 7D). Then additional cycles of enzyme binding, activity, and dissociation are required (Fig. 7, C and D), resulting in the release of shorter products (Fig. 7E); the major products in the system studied here were 33 and 27 nucleotides. POT1 may bind the partially unwound G-rich telomeric strand as WRN unwinds (Fig. 7, F and G), thereby preventing reannealing upon WRN dissociation. POT1 interaction with WRN may facilitate loading of POT1 onto the partially unwound G-rich single strand. In a subsequent cycle, unwinding can resume at the point where the enzyme previously dissociated, thereby increasing the rate of duplex displacement.
The data suggest two levels at which the POT1 variants may stimulate the RecQ helicases. First, because both POT1 variants facilitate WRN and BLM, the DNA-binding domain of POT1 (which is identical in v1 and v2) is important for the helicase stimulation. Consistent with this, substrates that lacked the DNA-binding sequence for POT1 within the duplex region showed limited or no stimulation (Figs. 1-4). For the D-loop with telomeric sequence in ssDNA region (such that POT1 could pre-load), the helicase products were similar in the presence or absence of POT1 (Figs. 3 and 4). This suggests that the ability of POT1 to coat one strand as the duplex is unwound by WRN and BLM is critical for the stimulation. RPA, on the other hand, can bind both ssDNA strands nonspecifically as they are unwound, which may account for the higher levels of stimulation with RPA. Second, protein-protein interactions between POT1 and the helicases may contribute. Both POT1 variants interacted with purified WRN in vitro by ELISA, although POT1v2 bound WRN with higher affinity than POT1v1, and similarly bound to BLM (Fig. 5). Most interestingly, POT1v1 was more effective at precipitating endogenous WRN from nuclear extracts than POT1v2. This suggests the POT1v1 C-terminal domain may enhance interaction with WRN in a cellular context, either by directly binding WRN or indirectly via interaction with other proteins. Alternatively, the C terminus of POT1v2, which differs from POT1v1 (Fig. 2G) (36), may interact with other nuclear proteins that could interfere with binding by WRN and BLM. Such interactions with the POT1 variants would be absent in our in vitro assays with recombinant purified proteins. In a cellular context, BLM may associate more weakly or less stably with the POT1v1 protein complex compared with WRN, because less of the endogenous BLM protein was precipitated by GST-POT1v1 than endogenous WRN (Fig.  6). Nevertheless, the interaction between POT1 and WRN or BLM likely contributes to the stimulation, because we observed that POT1 failed to stimulate the bacterial 3Ј-5Ј-helicase UvrD.
Another critical telomeric factor, TRF2, was also reported to stimulate WRN and BLM helicase activity but via a distinctly different mechanism than POT1. Unlike POT1, TRF2 binds duplex telomeric DNA and was found to stimulate dissociation of both telomeric and nontelomeric substrates that were effectively unwound by WRN and BLM alone (4,28). In contrast, we observed that TRF2 did not facilitate either WRN or BLM in unwinding of the 34-bp telomeric forked duplex (28). The action of the WRN helicase unwinding at the forked end and WRN exonuclease digestion at the blunt end shortens the duplex to unstable lengths (C), so that the remaining duplex melts upon enzyme dissociation (E). If WRN does not sufficiently decrease the duplex length before dissociation, the partially unwound strands will reanneal (D), and multiple rounds of WRN binding, unwinding/digestion activity, and dissociation are required for melting. POT1 binds the TTAGGGTTAG sequence of the top strand revealed during unwinding (F) and can prevent strand reannealing in the event of WRN dissociation (G). The remaining duplex can then be rapidly unwound upon WRN binding (H). Shown are the major WRN products that result in the absence (E) or presence of POT1 (H). The physical interaction between WRN and POT1 may contribute to recruitment of the enzymes to the substrate. Numbers indicate the nucleotide positions of the 49-mer top strand; the telomeric sequence is indicated by a thick line, and WRN and POT1 are represented by an oval or circle, respectively. TRF2, unlike POT1, does not bind telomeric ssDNA (43) and thus cannot coat the strands as they are unwound by the helicases. This further supports the model that POT1 coating of the partially unwound strand is critical for promoting the helicases. Thus, our data indicate that the RecQ helicases WRN and BLM require RPA or POT1 to unwind long telomeric duplex substrates, but partnership with POT1 is likely to be more important at the telomeres.
Increasing evidence supports roles for RecQ helicases WRN and BLM in dissociating alternate DNA structures or intermediates during DNA replication and recombination at the telomere. Both BLM and WRN associate with telomeres during S-phase in immortalized telomerase-deficient cells that used the recombination-based ALT pathway (4,26). WRN also associates with telomeres in S-phase primary fibroblasts (6). In human ALT cells, WRN and BLM localize to nuclear foci that contain TRF1, TRF2, telomeric DNA (26,44), and POT1 (45). Resolution of the telomeric D-loop/t-loop structure is required during progression of the DNA replication machinery to the chromosome end, and the dissociation of intra-telomeric and/or inter-telomeric D-loop intermediates in the ALT process is required to complete repair (46,47). Here we demonstrate that POT1 stimulates WRN and BLM helicases to unwind a 34-bp telomeric forked duplex and a 33-bp telomeric duplex D-loop structure that were otherwise poorly unwound by the helicase activities alone. The only other protein identified thus far that stimulates WRN and BLM to completely unwind these substrates has been RPA (Figs. 2 and 4) (26,28). However, RPA has not been detected as one of the core components of the human telomeric protein complex that functions in regulating and maintaining telomere structure and function. Since POT1 is a primary member of this telomeric complex along with TRF1, TRF2, TPP1, TIN2, and Rap1 (42), the functional interaction between POT1 and WRN or BLM is likely to be more pertinent to DNA metabolic pathways at the telomeres.
We propose that the ability of POT1 to coat the ssDNA telomeric tail as it is released by the RecQ helicases during unwinding of telomeric D-loops may be important for protection of the telomeric tail and for proper reformation of the telomere end. POT1 is essential for preventing telomere loss in Schizosaccharomyces pombe (15). Knock down of POT1 in some human cell lines results in loss of the telomeric 3Ј tail, chromosome end fusions, and apoptosis or senescence (48,49), suggesting POT1 directly protects the telomere 3Ј tail. POT1 also functions in generating the correct sequence of the 5Ј-recessed end of the C-rich strand, indicating possible roles in regulating nucleolytic processing of the telomere end (50). Cells defective in WRN helicase activity exhibited deletions of telomeres from sister chromatids that were replicated by lagging strand DNA synthesis, the strand that contains the 3Ј tail (6). The ability of POT1 to bind the tail, as it is revealed by the action of the RecQ helicases unwinding a telomeric D-loop, may facilitate the proper processing of the telomere end and prevent inappropriate strand invasion events by the released tail. An alternative, but not mutually exclusive, possibility is that RecQ helicases may cooperate with POT1 to dissociate G-quadruplex structures that can form in the G-rich telomeric ssDNA strands.
In summary, we propose that the RecQ helicases WRN and BLM may cooperate with POT1 in the proper dissociation of telomeric end structures during DNA replication and/or recombination. The physical interaction between WRN and BLM suggest they may cooperate in common pathways (38). We have demonstrated that POT1 specifically stimulates WRN and BLM helicases to resolve telomeric substrates and that these helicases require RPA or POT1 to unwind long telomeric duplexes. The ability of POT1 to coat the telomeric 3Ј tail, as it is revealed by these helicases, may be important for maintaining the integrity of this tail and the telomere structure.