Introduction

Werner syndrome (WS) is an autosomal recessive genetic disease demonstrating a wide range of phenotypic abnormalities, including characteristics of premature aging (Salk 1982). Werner syndrome results from mutations within the WRN protein, which is a member of the RecQ DNA helicase family and has both a 3′ to 5′ helicase activity and a 3′ to 5′ exonuclease activity (Brosh and Bohr 2002; Oshima 2000; Shen and Loeb 2000). WRN has also been reported to have a 5′ to 3′ exonuclease activity on 3′ overhangs (Suzuki et al. 1999). Cells isolated from individuals with WS demonstrate shortened life-span in culture (Martin et al. 1970), as well as an increased rate of DNA rearrangements, including translocations, deletions, and dicentrics (Fukuchi et al. 1989; Gebhart et al. 1988; Grigorova et al. 2000; Salk et al. 1981; Scappaticci et al. 1982; Stefanini et al. 1989). Thus, WS cells demonstrate chromosome instability similar to cells with defects in other RecQ helicases (van Brabant et al. 2000).

Chromosome instability in cells deficient in WRN could occur through a variety of mechanisms. WRN binds to a number of proteins involved in recombination, including replication protein A (RPA; Brosh et al. 1999; Shen et al. 1998), PCNA, topoisomerase I (Lebel et al. 1999), and DNA polymerase δ (Szekely et al. 2000), and colocalizes with RAD51 (Sakamoto et al. 2001). Consistent with these observations, WS cells have been demonstrated to have a defect in homologous recombination (Prince et al. 2001; Saintigny et al. 2002). Combined with the observations that WS cells have a prolonged S phase (Poot et al. 1992) and WRN colocalizes with RPA in cells arrested in S phase with hydroxyurea (Constantinou et al. 2000), these results suggested that WS cells have a defect in resolving stalled replication forks (Constantinou et al. 2000; Franchitto and Pichierro 2002; Shen and Loeb 2000). WRN has also been demonstrated to bind to DNA-PKcs (Karmakar et al. 2002; Yannone et al. 2001), Ku (Cooper et al. 2000; Li and Comai 2000), and FEN-1 (Brosh et al. 2001), proteins involved in non-homologous end joining. In addition, the activity of WRN is influenced by binding to DNA-PKcs and Ku (Cooper et al. 2000; Karmakar et al. 2002). Thus, in addition to a role in homologous recombination, WRN may also have a role in the non-homologous end-joining pathway for double-strand break (dsb) repair. The lack of sensitivity to most DNA-damaging agents has led to the observation that WS is distinct from many other types of chromosome instability syndromes that involve defects in DNA repair (Gebhart et al. 1988; Stefanini et al. 1989). However, WRN-deficient cells show sensitivity to 4-nitroquinoline 1-oxide (4NQO; Gebhart et al. 1988; Stefanini et al. 1989), crosslinking agents (Poot et al. 2001), and a slight sensitivity to ionizing radiation (Grigorova et al. 2000; Saintigny et al. 2002; Yannone et al. 2001).

In addition to an apparent role in DNA recombination/repair, WRN also has a role in telomere maintenance. Telomeres are composed of a 6-bp repeat sequence and associated proteins that form a cap that protects chromosome ends (Blackburn 2001; McEachern et al. 2000). Telomeric repeat sequences are added on by the enzyme telomerase, which compensates for the loss of DNA from the ends of chromosomes during cell division. In humans, telomerase is active in germ line cells, but inactive in most somatic cells (de Lange 1995). As a result, telomeres become progressively shorter in somatic cells, which has been proposed to limit cellular life-span by promoting replicative senescence (Harley 1995). Consistent with this hypothesis, transfection of human fibroblasts with the catalytic subunit of telomerase greatly extends their life-span (Bodnar et al. 1998). A primary function of telomeres is to prevent chromosome fusion, as illustrated by the massive increase in chromosome fusion observed in somatic cells that fail to senesce and therefore continue to undergo telomere shortening (Counter et al. 1992). Increased chromosome fusion is also seen in cells with mutations in genes that affect telomere function, including DNA-PKcs, Ku, and TRF2 (Bailey et al. 1999; Gilley et al. 2001; Goytisolo et al. 2001; Hsu et al. 2000; van Steensel et al. 1998). The ability of cells to properly maintain telomeres is therefore a critical factor in maintaining chromosome stability.

The role of telomeres in cell senescence led to the proposal that the short life-span of WS cells in culture might be due to accelerated telomere shortening (Martin 1994). Although WS cells do show accelerated telomere shortening, the premature senescence in WS cells was found to occur when their telomeres were longer than the normal control cells (Schulz et al. 1996). However, a subsequent study found that WS cells at senescence have telomeres that are similar in length to normal senescent cells (Choi et al. 2001). Combined with the fact that WS cells can be immortalized by expression of telomerase (Choi et al. 2001; Ouellette et al. 2000; Wyllie et al. 2000), these results suggest that WS cells have a defect in telomere maintenance that leads to premature senescence. A role for WRN in telomere maintenance is also indicated by its association with Ku and DNA-PKcs, since the Ku/DNA-PK complex has been demonstrated to be located at telomeres and to be essential for proper telomere function (Bailey et al. 1999; Boulton and Jackson 1998; Espejel et al. 2002; Gilley et al. 2001; Hsu et al. 1999, 2000). In addition, WRN has been found to bind to TRF2 (Opresko et al. 2002), which plays an essential role in maintaining the capping function of telomeres and preventing chromosome fusion (van Steensel et al. 1998). Thus, deficiencies in WRN could influence chromosome stability through its role in both DNA repair and telomere maintenance.

We previously investigated the consequences of spontaneous telomere loss in clone B3 of the EJ-30 human tumor cell line, which has a Herpes Simplex Virus thymidine kinase (HSV-tk) selectable marker gene integrated immediately adjacent to a telomere (Fouladi et al. 2000; Lo et al. 2002a). Loss of the HSV-tk gene was found to be associated with the formation of inverted repeats, large duplications, and prolonged periods of instability specific to the "marker" chromosome containing the telomeric HSV-tk gene. The absence of a telomere and the presence of anaphase bridges more than 20 generations after the loss of the telomere demonstrated that this chromosome instability involved breakage/fusion/bridge (B/F/B) cycles (Lo et al. 2002a). In the present study, the influence of WRN on telomere function was investigated using a clone of B3, B3-WrnDN1, that contains an inducible WRN gene, K577M-WRN, that contains a mutation previously reported to function in vivo as a dominant-negative (Wang et al. 2000). Cultures of B3-WrnDN1 with and without expression of K577M-WRN were then analyzed for the frequency of telomere loss, the rate of loss of the telomeric HSV-tk gene, and chromosome fusion.

Materials and methods

Cell culture

The EJ-30 cell line (obtained from Dr. William Dewey, University of California, San Francisco) was subcloned from the EJ bladder cell carcinoma cell line, which is also named MGH-U1 (O'Toole et al. 1983). Clone B3 was isolated from EJ-30 following transfection with the pNCT-tel plasmid and contains a single copy of the plasmid integrated at the end of the short arm of chromosome 16 (Fouladi et al. 2000). B3 and all B3-derived subclones were grown in α-MEM (UCSF Cell Culture Facility) containing 5% fetal calf serum (Hyclone), 5% donor calf serum with iron (Hyclone), and 1 mM l-glutamine (UCSF Cell Culture Facility). Cell lines derived from clone B3 that contain inducible WRN genes were established using the RetroTet-ART inducible system that contains a promoter that is responsive to tetracycline or doxycycline. To develop these cell lines, clone B3 was infected with both the RTRg(−)cd8 retrovirus containing a tetracycline-responsive repressor and an RTAb(+) retrovirus containing a tetracycline-responsive activator. Identification of clones containing the RTRg(−)cd8 retrovirus was accomplished by the presence of CD8 using a CD8-specific antibody. Following infection, 200 cells were replated in 100-mm tissue culture dishes and the cells incubated for 2 weeks at 37°C. CD8-positive colonies were then detected by first washing the cells at 4°C with phosphate-buffered saline (PBS; UCSF Cell Culture Facility) containing 0.25% bovine serum albumin (BSA; Sigma), followed by incubation at 4°C for 15 min in 1 ml PBS containing 0.25% BSA and a 1:100 dilution of CD8-specific antibody (Pharmingen). The cells were then washed at 4°C with PBS containing 0.25% BSA and screened for CD8 expression by analysis with a fluorescence microscope using a lens immersed directly within the growth medium. Clones expressing CD8 were then isolated, and one of these clones was infected with the RTAb(+) retrovirus. Forty clones selected at random from the RTAb(+)-infected cells were then tested for expression of the activator in RTAb(+) by individually infecting them with the HRSpuro-gfp retrovirus and screened for GFP expression in the presence of 1 μg/ml doxycycline using a B-2E/C FITC filter on a Nikon E600 Eclipse fluorescence microscope. Clone B3rr, that showed high levels of GFP in the presence of 1 μg/ml doxycycline but no GFP in the absence of doxycycline, was then infected with the HRSpuro retrovirus containing the wild-type WRN or K577M-WRN genes (kindly provided by Dr. Junco Oshima, University of Washington). K577M-WRN contains a K to M mutation at amino acid 577 in the WRN protein, which inactivates helicase activity and functions in vivo as a dominant-negative in mice (Wang et al. 2000). Expression of K577M-WRN in all subsequent experiments was regulated by the addition of 1 μg/ml doxycycline (Clontech) to the medium.

Immediately prior to use, 4NQO (Sigma) dissolved in acetone was diluted in medium. One hundred and fifty cells plated on 100-mm dishes were allowed to attach and then were treated for 1 h with the various concentrations of 4NQO. After treatment, the cells were washed twice with growth medium and grown for 2 weeks prior to staining and counting colonies.

Viral preparation and infection

Retroviruses were generated by transient transfection of Amphotropic 293 cells (Clontech) with the retrovirus vectors using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Infection of cells with retroviruses was accomplished by incubation of the cells with medium harvested from the Amphoteric 293 cells 36–72 h after transfection and filtered with a 0.45-μm Acrodisc filter (Gelman). Prior to infection, 8 μg/ml Polybrene (Sigma) was added to the medium containing the retrovirus vector. After 6 h, the medium was replaced with growth medium, and after an additional 24 h the cells were analyzed for expression of transgenes.

Southern blot analysis

Genomic DNA purified as previously described (Murnane et al. 1985) was digested with restriction enzymes according to the manufacturer's instructions. Genomic DNA was fractionated by agarose gel electrophoresis using standard protocols, depurinated by treatment with 0.25 M HCl for 30 min, and transferred in 0.5 M NaOH onto a charged nylon Hybond-N+ membrane (Amersham) using a vacuum transfer apparatus (Pharmacia). Prehybridization for 3 h and hybridization overnight were performed at 65°C as previously described (Lo et al. 2002b). For analysis of plasmid sequences hybridization was performed with the pNCT-Δ probe, which is identical to pNCT-tel except that it does not contain telomeric repeat sequences. For analysis of average telomere length, hybridization was performed with an 800-bp fragment of telomeric repeat sequences isolated from pNCT-tel. The average length of terminal restriction fragments was estimated from the mean length of the hybridization signal as determined by densitometry.

Reverse transcriptase-polymerase chain reaction (RT-PCR)

Total RNA was isolated from cell cultures according to the manufacturer's instructions using the StreptaPrep Total RNA Miniprep kit (Stratagene). The RT-PCR was then performed using a SuperScript One-step RT-PCR kit with Platinum Taq (Invitrogen) using the manufacturer's instructions. The RT-PCR conditions involved 1 cycle at 45°C for 30 min and then 94°C for 2 min. This was followed by 35 cycles at 94°C for 30 s, 56°C for 30 s, and 72°C for 5 min, with the final cycle of 72°C for 10 min. The primer sequences used for RT-PCR were 5′-GGTGCACCCAAATTTGTCTAGA-3′ and 5′-TGACAGCAACATTATCTCTT-3′. As a control, RT-PCR was also performed simultaneously for the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase using the primers 5′-CATGTGGGCCATGAGGTCCACCAC-3′ and 5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′.

Chromosome and telomere analysis

Chromosome analysis was performed following the addition of 0.1 μg/ml colcemid to the medium for 1.5–2 h at 37°C. The cells were then trypsinized, pelleted at 500 g for 5 min, the supernatant removed, 2 ml 0.56% KCl added, and incubated for 30 min at 37°C. The cells were then spread on a slide and dried at room temperature prior to analysis.

Telomere analysis was performed using the Telomere PNA FISH kit/Cy3 (Dako) as described by the manufacturer. Peptide nucleic acid (PNA) probes are highly sensitive and can detect telomeres less than 150 bp in length (Lansdorp et al. 1996). Briefly, the metaphase preparations were treated with 3.7% formaldehyde in TRIS-buffered saline (TBS; Sigma) for 2 min, washed with TBS, and incubated with Pretreatment Solution for 10 min. The slide is then treated with a cold ethanol series consisting of 70%, 86%, and 95% ethanol. After the slide is dried, 10 μl telomere PNA probe labeled with Cy3 was added to the slide and a coverslip placed on top. The slide was denatured in a preheated incubator at 80°C and placed in the dark for 30 min at room temperature. The slide was rinsed with Wash Solution at 65°C, put through the cold ethanol series again, and air dried. The slides were then prepared for fluorescent microscopy by addition of 50 μl mounting solution and addition of a coverslip. Chromosomes without detectable telomeres were scored visually.

Fluctuation analysis and mutant selection

For analysis of the rate of loss of the HSV-tk gene, B3-WrnDN1 was first plated at 100 cells per well in a 24-well culture dish in medium with or without 1 μM doxycycline. These cultures were expanded to more than 106 cells, after which time 105 were plated in triplicate into medium containing 50 μM ganciclovir (a kind gift from Hoffmann/La Roche) and 400 μg/ml G418 (Gibco) for identification of cells that had lost the HSV-tk gene and retained the neo gene. After an additional 4 weeks in culture with and without doxycycline, the cultures were again plated into medium containing ganciclovir and G418. The rate of telomere loss was determined as previously described (Fouladi et al. 2000) from the frequency of G418r/ganr cells in the different populations by the method of Luria and Delbruck (Luria and Delbruck 1943) as modified by Capizzi and Jameson (1972).

Telomerase activity

Relative telomerase activity was determined using the Telo TAAGGG Telomerase PCR ELISA assay as described by the manufacturer (Roche), using measurements taken with a spectrophotometer at a wavelength of 450 nm.

Results

Development of cell lines

The role of WRN in telomere loss and chromosome instability in human cells was investigated using clone B3 of the human EJ-30 bladder cell carcinoma cell line (Fouladi et al. 2000). Clone B3 contains a selectable HSV-tk gene located immediately adjacent to a telomere on chromosome 16 (Fig. 1). Clone B3 was established by transfection with the linearized pNCT-tel plasmid that contains telomeric repeat sequences on one end. The integration of a plasmid containing telomeric repeat sequences on one end results in the "seeding" of a new telomere on the end of a broken chromosome (Barnett et al. 1993; Farr et al. 1991; Hanish et al. 1994). The seeded telomeres are elongated in culture and their length and dynamics become similar to the other telomeres in the cell (Sprung et al. 1999a, b). Selection with ganciclovir for cells that lose the HSV-tk gene provides a method for determining the rate of telomere loss, and for isolation and analysis of the types of events resulting from telomere loss (Fouladi et al. 2000; Lo et al. 2002a, b).

Fig. 1.
figure 1

Structure of the plasmid sequences integrated adjacent to a telomere in clone B3 of the EJ-30 tumor cell line. The pNCT-tel plasmid linearized with NotI seeded the formation of a new telomere on the end of chromosome 16p following integration. The location of the cellular DNA, plasmid vector (amp/ori), neo gene, Herpes Simplex Virus thymidine kinase (HSV-tk) gene, and telomeric repeat sequences are shown. The location of BamHI (Bm) and ClaI (Cl) restriction sites used for Southern blot analysis, and the size of the restriction fragments, are also shown

Full size image

To investigate the influence of WRN on telomere stability, we established a clone of B3 that contains the RetroTet ART tetracycline-inducible system (Rossi et al. 1998). To reduce the leakiness of the more traditional tetracycline-inducible system that utilizes only an activator, this system utilizes an inducible promoter that responds to both a tetracycline-responsive activator and a tetracycline-responsive repressor. The activator is active in the presence of tetracycline, while the repressor is active in the absence of tetracycline. A clone of B3, B3rr, was identified that contains both the activator and repressor and that expresses GFP in the presence of doxycycline following infection with a retrovirus containing the GFP gene on the inducible promoter (see Materials and methods). A clone of B3rr, B3-WrnDN1, was then established that contains the WRN gene with the K577 M mutation (K577M-WRN) under the control of the tetracycline-inducible promoter. K577M-WRN has been demonstrated to function as a dominant-negative in vivo in mice (Wang et al. 2000). Inducible expression of K577M-WRN in clone B3-WrnDN1 was confirmed by RT-PCR (Fig. 2). One of the PCR primers used for RT-PCR is specific for the viral vector and therefore does not amplify the endogenous WRN protein. B3-WrnDN1 showed no detectable expression of the transduced K577M-WRN without doxycycline present, while expression of K577M-WRN was easily detectable following growth for 2 weeks in the presence of doxycycline.

Fig. 2.
figure 2

RT-PCR demonstrating inducible expression of K577M-WRN in B3-WrnDN1. Total RNA isolated from cultures of clone B3-WrnDN1 grown with (+) or without (−) doxycycline was analyzed by RT-PCR using primers specific for the K577M-WRN transgene. Primers for the glyceraldehyde-3-phosphate dehydrogenase housekeeping gene were used as a control (lower band). Molecular weight markers (S) and their sizes are shown

Full size image

To determine whether expression of K577M-WRN produced a phenotype similar to WS, we next analyzed cultures of B3-WrnDN1 expressing K577-WRN for sensitivity to 4NQO. Similar to WS cells (Gebhart et al. 1988; Stefanini et al. 1989), cultures of B3-WrnDN1 expressing K577M-WRN demonstrated increased sensitivity to 4NQO. After treatment with various concentrations of 4NQO, B3-WrnDN1 cells grown in medium with doxycycline showed decreased survival compared to B3-WrnDN1 cells grown in medium without doxycycline (Fig. 3). The decrease in survival eventually levels off, typical of toxicity of 4NQO in WS cells (Prince et al. 1999), but may also reflect the lack of expression of the K577M-WRN transgene in all of the cells in the population. Control cultures consisting of the parental clone B3 and a clone of B3 expressing wild-type WRN also showed no increase in sensitivity to 4NQO when grown in the presence of doxycycline. These results demonstrate that K577-WRN can function as a dominant-negative in human cells.

Fig. 3.
figure 3

Sensitivity to 4-nitroquinoline 1-oxide (4NQO) in B3-WrnDN1 cultures with and without expression of K577M-WRN. Cultures of B3-WrnDN1 grown with (filled diamonds) and without (filled circles) doxycycline were plated in various concentrations of 4NQO and the number of colonies after 2 weeks was determined relative to the number of colonies in the absence of 4NQO. As controls, cultures of the parental B3 cell line (filled triangles), or B3 expressing wild-type WRN (filled squares), were also cultured in doxycycline and analyzed for the presence of telomeres after 2–3 weeks

Full size image

Loss of detectable telomeres and chromosome fusion in cells expressing K577M-WRN

The influence of K577M-WRN expression on telomere stability in B3-WrnDN1 was investigated by FISH using a telomere-specific PNA probe. Telomere-specific PNA probes are highly sensitive, and have been demonstrated to detect telomeres less than 150 bp in length (Lansdorp et al. 1996). In the initial experiment, cultures of B3-WrnDN1 were grown for 4 weeks with and without doxycycline in the growth medium, and metaphase spreads were analyzed for the presence of telomeric repeat sequences on chromosome ends. Chromosomes without detectable telomeres were rare in the cultures grown without doxycycline present in the culture medium (0 in100 metaphase spreads analyzed), while telomeres were often undetectable on one (Fig. 4a) or both (Fig. 4b) chromatids in the cultures grown with doxycycline (20 in 100 metaphase spreads analyzed). Consistent with this observation, no dicentric chromosomes were observed in the cultures grown without doxycycline (0 in 100 metaphase spreads analyzed), while dicentric chromosomes (Fig. 4c) were detected in the cultures grown with doxycycline present (4 in 100 metaphase spreads analyzed). This increase in the frequency of dicentric chromosomes demonstrates that the events leading to loss of detectable telomeres in B3-WrnDN1 cultures expressing K577M-WRN result in the inability of the telomere to prevent chromosome fusion.

Fig. 4a–c.
figure 4

Analysis of the loss of detectable telomeres and chromosome fusion in B3-WrnDN1 cultures expressing K577M-WRN. Cultures of B3-WrnDN1 grown with or without doxycycline were analyzed by FISH using a telomere-specific peptide nucleic acid (PNA) probe (yellow). Chromosomes were counterstained with DAPI. Cells expressing K577M-WRN showed an increase in the number of chromosomes with no detectable telomere on one (a) or both (b) sister chromatids, as well as an increase in the number of dicentric chromosomes (c). Arrows show the location of undetectable telomeres on a single chromatid (a) or both chromatids (b), or the location of centromeres in a dicentric chromosome (c)

Full size image

Time course of the loss of detectable telomeres in cells expressing K577M-WRN

The frequency of chromosomes without detectable telomeres was investigated at various times after addition of doxycycline to induce the expression of K577M-WRN. A continual increase in the number of chromosomes without detectable telomeres was observed during growth of B3-WrnDN1 cultures in the presence of doxycycline, eventually reaching an average of one undetectable telomere per cell after 28 days (Fig. 5). In contrast, B3-WrnDN1 cultures grown without doxycycline showed no increase in chromosomes without detectable telomeres during this same time period. Similarly, control cultures consisting of the parental clone B3 and a clone of B3 expressing wild-type WRN also showed no increase in the number of undetectable telomeres when grown in the presence of doxycycline (Fig. 5). The rate of recovery of telomeres in B3-WrnDN1 cultures grown for 4 weeks in doxycycline was investigated following the removal of doxycycline. The B3-WrnDN1 cultures grown without doxycycline for 3 weeks showed a continuous reduction in the number of chromosomes without telomeres from 0.3 to 0.1 per cell, while cells grown in the presence of doxycycline during this same period showed an increase in the number of chromosomes without detectable telomeres.

Fig. 5a, b.
figure 5

The rate of loss of detectable telomeres and their recovery in B3-WrnDN1 cultures with and without expression of K577M-WRN. a The metaphase chromosomes of cultures of B3-WrnDN1 were grown with (filled diamonds) and without (filled circles) doxycycline and metaphase chromosomes were analyzed after 1–4 weeks for the presence of telomeres by FISH using a telomere-specific PNA probe. As controls, cultures of the parental B3 cell line (filled triangles) or B3 expressing wild-type WRN (filled squares) were also cultured in doxycycline and analyzed for the presence of telomeres after 2 and 3 weeks. b Metaphase chromosomes of cultures of B3-WrnDN1 grown with doxycycline for 4 weeks were subsequently grown with (filled diamonds) and without (filled circles) doxycycline for various times and metaphase chromosomes were analyzed for the presence of telomeres by FISH using a telomere-specific PNA probe

Full size image

Consequences of telomere loss

To investigate the consequences of telomere loss resulting from the expression of K577M-WRN, we studied the loss of the HSV-tk gene located at the telomere on the short arm of chromosome 16. Clones of EJ-30 have a relatively high spontaneous rate of loss of the HSV-tk gene (10−4 events per cell per generation; Fouladi et al. 2000), indicating that EJ-30 is similar to many other human tumor cells that have telomere instability (Gisselsson et al. 2001; Sprung et al. 1999a). To determine whether expression of K577M-WRN causes a further increase in the rate of telomere loss, we performed fluctuation analysis by growing populations of B3-WrnDN1 with and without doxycycline and monitored the frequency of cells in the population that were resistant to both ganciclovir and G418 (G418r/ganr). Selection with G418 was used simultaneously with ganciclovir to insure that the neo gene was retained, since some portion of the plasmid is required for subsequent analysis of the types of recombination events involved. Consistent with the cytogenetic analysis, the B3-WrnDN1 cells grown in doxycycline demonstrated a 4.7-fold increase in the rate of appearance of G418r/ganr cells as compared with the same cells grown without doxycycline (Table 1). The actual rate of telomere loss is even higher, since G418r/ganr cells represent only those events that do not involve the loss of the neo gene, which represent only a small fraction of the total (Fouladi et al. 2000). Thus, in view of the already high rate of spontaneous loss of telomeres in the EJ-30 tumor cell line (Fouladi et al. 2000), expression of K577M-WRN results in a large increase in the rate of telomere loss.

Table 1. Influence of K577M-WRN on the rate of telomere loss in B3-WrnDN1. (Dox Doxycycline)
Full size table

Southern blot analysis was performed to investigate the nature of the rearrangements involved in the loss of function of the HSV-tk gene in the G418r/ganr subclones (Fig. 6). Loss of the HSV-tk gene in B3 (Fouladi et al. 2000), like mouse ES cells (Lo et al. 2002a, b), was previously found to result in either the addition of a telomere to the site of the break or sister chromatid fusion and B/F/B cycles leading to chromosome instability. The telomeric pNCT-tel plasmid contains a single BamHI site in the center of the plasmid (Fig. 1). Therefore, as previously reported (Fouladi et al. 2000), digestion of genomic DNA from clone B3 produces two bands with BamHI, a 4.5-kb band that contains the internal portion of the plasmid, and a larger band that contains the terminal restriction fragment (Fig. 6a). The larger band is generally diffuse in appearance, due to heterogeneity in the length of the telomere in different cells in the population. Digestion of genomic DNA from clone B3 with BamHI and ClaI, which cuts off the telomeric repeat sequences, produces the 4.5-kb internal band as well as a 4.0-kb band containing the HSV-tk gene (Fig. 6b). Similar to the parental cell line, G418r/ganr subclones that have a telomere added at a new location in the HSV-tk gene show two bands following digestion with BamHI alone, although the diffuse band containing the telomeric repeat sequences varies in size in different subclones. However, with BamHI and ClaI digestion, the 4.0-kb band is missing in these subclones as a result of the loss of the ClaI site adjacent to the telomere, and instead is replaced by a larger diffuse band containing the telomeric repeat sequences. G418r/ganr subclones that contain fusions involving the HSV-tk gene are missing the diffuse band with either BamHI digestion alone or with digestion with both BamHI and ClaI (Fouladi et al. 2000). Instead, these subclones show a new non-diffuse band that differs in length in different subclones. By these criteria, two G418r/ganr subclones selected from cultures grown without doxycycline showed rearrangements consistent with sister chromatid fusions (subclones 2 and 3), while three subclones showed no apparent changes, possibly due to point mutations in the HSV-tk gene (subclones 1, 4, and 5). By comparison, two G418r/ganr subclones selected from cultures grown with doxycycline showed rearrangements consistent with sister chromatid fusion (subclones 1 and 2), one with the addition of a telomere at a new location in the HSV-tk gene (subclone 3) and two with no detectable change (subclones 4 and 5). Thus, a significant portion of the G418r/ganr subclones from cultures expressingWRN-K577 M show no detectable changes in the structure of the HSV-tk gene, possibly due to an increased rate of small deletions associated with loss of WRN function (Fukuchi et al. 1989). The results therefore demonstrate that WRN deficiency increases the rate of both point mutations (base changes and/or small deletions) and chromosome rearrangements resulting from telomere loss.

Fig. 6a, b.
figure 6

Southern blot analysis of the types of chromosome rearrangements found in B3-WrnDN1 cultures with and without expression of K577M-WRN. Genomic DNA isolated from five ganr/G418r subclones of B3-WrnDN1 grown for 3 weeks with doxycycline and five subclones isolated from of B3-WrnDN1 grown without doxycycline was digested with BamHI (a) or BamHI and ClaI (b). Similar digestions of genomic DNA isolated from the parental clone B3 are shown as controls. Hybridization was performed with the pNCT-Δ probe, which is identical to pNCT-tel except that it does not contain telomeric repeat sequences. BamHI cuts the telomeric plasmid once near the center, while ClaI cuts between the plasmid and the telomeric repeat sequences (see Fig. 1). As with clone B3, the genomic DNA digested with both BamHI and ClaI in subclones 4 and 5 (+ doxycycline), and lanes 1, 4, and 5 (− doxycycline) also showed two bands of 4.0 and 4.5 kb (distinguishable at shorter exposures), and therefore showed no detectable change in the HSV-tk gene. In contrast, subclone 3 (+ doxycycline) shows a heterogeneous band following digestion with both BamHI and ClaI, indicating the presence of a telomere at a new location in the HSV-tk gene. Molecular size markers consisting of Lambda bacteriophage DNA HindIII restriction fragments are shown

Full size image

Analysis of telomerase activity and average telomere length in cells expressing K577M-WRN

Telomerase activity was determined in cultures of B3-WrnDN1 grown with and without doxycycline to determine whether the increased rate of telomere loss in cells expressing K577M-WRN was associated with direct effects on telomerase activity. B3-WrnDN1 cultures grown for 4 weeks in the presence of doxycycline had 20-fold less relative telomerase activity than B3-WrnDN1 cultures grown without doxycycline in two separate experiments (1.4 and 1.4 versus 25.3 and 32.0, respectively). Despite this reduced telomerase activity, by visual inspection, no apparent shortening was detectable on most telomeres by FISH analysis with the telomere-specific PNA probe (Fig. 4). To determine whether changes in telomere length are involved, Southern blot analysis was performed to determine the average length of terminal restriction fragments using a telomere-specific probe. Using this approach, no change in average telomere length was detectable by Southern blot analysis after growth for 8 weeks in the presence of doxycycline (Fig. 7). Thus, the ability of telomerase to maintain telomere length alone does not appear to explain the increased rate of telomere loss in cells expressing K577M-WRN.

Fig. 7.
figure 7

Influence of K577M-WRN expression on average telomere length in clone B3-WrnDN1. Genomic DNA from clone B3-WrnDN1 grown with (+) or without (−) doxycycline was digested with RsaI and HinfI, and Southern blot analysis was performed using hybridization with a telomeric repeat-specific probe. Molecular size markers consisting of Lambda bacteriophage HindIII restriction fragments are shown

Full size image

Discussion

The results presented here demonstrate an increase in the rate of telomere loss in human tumor cells expressing a mutant K577M-WRN protein that does not possess helicase activity. These results demonstrate that K577M-WRN functions as a dominant-negative for telomere function, similar to its ability to function as a dominant-negative in sensitizing cells to 4NQO (Fig. 3). A previous study found that K577M-WRN can also function in vivo as a dominant-negative in mice (Wang et al. 2000), although another study did not observe a dominant-negative effect for K577M-WRN in sensitization to cis-platinum (Saintigny et al. 2002). An increase in telomere loss was evident from both FISH analysis with a telomere-specific PNA probe and by analysis of the rate of loss of the telomeric HSV-tk gene. This telomere loss appears to be a stochastic process, since no change in average telomere length was evident in cells expressing K577M-WRN by Southern blot analysis. The sudden loss of a telomere could result from several different mechanisms, including: (1) an increase in the rate of dsbs or decreased repair of dsbs in subtelomeric or telomeric regions, (2) an increased rate of recombination in subtelomeric or telomeric regions, or (3) a loss of end capping function. An increase in the rate of dsbs could occur due to problems encountered during DNA replication or failure to resolve stalled replication forks, consistent with the association of WRN protein with the RPA, PCNA, and topoisomerase I proteins involved in DNA replication and recombination (Brosh et al. 1999; Constantinou et al. 2000; Lebel et al. 1999; Saintigny et al. 2002; Shen et al. 1998). Alternatively, a decrease in the efficiency of dsb repair would be consistent with the association of the WRN protein with the Ku and DNA-PKcs proteins involved in non-homologous end joining (Bailey et al. 1999; Boulton and Jackson 1998; Espejel et al. 2002; Gilley et al. 2001; Hsu et al. 1999, 2000). Chromosome breaks within the subtelomeric region could result in direct inactivation of the HSV-tk gene, while breaks that occur within the telomeric repeat sequences could result in inactivation of the HSV-tk gene following degradation of the end of the chromosome if the remaining telomeric repeat sequences are too short to form a functional telomere.

The loss of sufficient telomeric repeat sequences to form a functional telomere could also occur through recombination. The high rate of chromosome rearrangements and deletions in WS cells (Fukuchi et al. 1989; Gebhart et al. 1988) results in an increase in "pop out" events (Prince et al. 2001) that could explain the extensive variability previously reported in the length of telomeres in WS B-lymphoblastoid cell lines (Tahara et al. 1997). The colocalization of WRN with RPA at stalled replication forks (Constantinou et al. 2000), and the suppression of hyper-recombination in WS cells by transfection with the RusA resolvase or a dominant-negative Rad51 (Saintigny et al. 2002), led to the suggestion that WS cells fail to resolve stalled replication forks (Constantinou et al. 2000; Saintigny et al. 2002). Therefore, hyper-recombination within the telomeric repeat sequences due to problems encountered during DNA replication may result in large deletion events and loss of telomere function.

The loss of a telomere in the B3-WrnDN1 cell line could also be a secondary event that results from the inability of the telomere to properly form a cap that protects the end of the chromosome. Although the reduced levels of telomerase activity did not result in a change in telomere length on most chromosomes, it could influence the capping function that has been proposed for telomerase (Zhu et al. 1999). In this scenario, complete loss of the telomere and inactivation of the HSV-tk gene would occur only after breakage of dicentric chromosomes that fused despite the presence of long tracks of telomeric repeat sequences. Chromosome fusion without the loss of most of the telomeric repeat sequences could occur as a result of the loss of the single-stranded tail on the end of the chromosome, which has been proposed to be necessary for the formation of T-loops that protect the ends of chromosomes (Stansel et al. 2001). Chromosome fusion resulting from the loss of the single-stranded tail has been previously seen with deficiencies in DNA-PKcs (Bailey et al. 1999) and TRF2 (van Steensel et al. 1998). Consistent with this possibility, WRN has been demonstrated to bind TRF2, which promotes WRN helicase activity (Opresko et al. 2002). However, although we observed an increase in dicentric chromosomes following expression of the K577M-WRN mutation in the B3-WrnDN1 cell line, no telomeric repeat sequences were observed at the junctions between these dicentric chromosomes (Fig. 4c). Thus, complete loss of the telomere appears to occur prior to chromosome fusion, indicating that stochastic events and not loss of capping function are involved.

Regardless of the mechanism involved, telomere loss due to a deficiency in WRN could have severe consequences for chromosome stability, since telomere loss can lead to extensive chromosome rearrangement for many cell generations by initiating B/F/B cycles (Fouladi et al. 2000; Lo et al. 2002a, b). Although it has previously been shown that telomerase expression can prevent senescence in primary fibroblasts from individuals with WS, the continued chromosome instability in these cell lines has led to the suggestion that telomere dysfunction is not responsible for instability in WS cells (Choi et al. 2001; Ouellette et al. 2000). However, this conclusion was based upon the assumption that telomere loss in WS cells occurs primarily through progressive shortening of telomeric repeat sequences, not by stochastic mechanisms. In view of the results presented here, it is clear that telomere loss in WRN-deficient cells can occur through stochastic mechanisms. Therefore, even in cells expressing telomerase, telomere loss could be a significant source of chromosome instability in WRN-deficient cells. Consistent with this possibility, the types of rearrangements associated with telomere loss in WS cells reflects those observed in chromosomes undergoing B/F/B cycles after the loss of a telomere. In addition to sister chromatid fusion and DNA amplification, B/F/B cycles can also result in non-reciprocal translocations and fusions with other chromosomes (Fouladi et al. 2000; Lo et al. 2002a, b) that have been reported to be increased in WS cells (Fukuchi et al. 1989; Gebhart et al. 1988; Grigorova et al. 2000; Salk et al. 1981; Scappaticci et al. 1982; Stefanini et al. 1989). WRN can therefore be added to the growing list of proteins that are known to play a role in maintaining telomere structure and preventing chromosome fusion. In view of the complexity of this process, it is not surprising that tumor cells commonly demonstrate deficiencies in telomere function that can contribute to the chromosome instability associated with tumor cell progression.