Inter-telomeric recombination is present in telomerase-positive human cells

Abstract

Immortal cells require a mechanism of telomere length control in order to divide infinitely. One mechanism is telomerase, an enzyme that compensates the loss of telomeric DNA. The second mechanism is the alternative lengthening of telomeres (ALT) pathway. In ALT pathway cells, homologous recombination between telomeric DNA is the mechanism by which telomere homeostasis is achieved. We developed a novel homologous recombination reporter system that is able to measure inter-telomeric recombination in a sensitive manner. We asked the fundamental question if homologous recombination between different telomeres is present in telomerase-positive cells. In this in vitro study, we showed that homologous recombination between telomeres is detectable in ALT cells with the same frequency as in cells that utilize the telomerase pathway. We further described an ALT cell clone that showed peaks of recombination which were not detected in telomerase-positive clones. In telomerase-positive cells the frequency of inter-telomeric recombination was not increased by shortened telomeres or by a fragile telomere phenotype induced with aphidicolin. ALT cells, in contrast, responded to aphidicolin with an increase in the frequency of recombination. Our results indicate that inter-telomeric recombination is present in both pathways of telomere length control, but the factors that increase recombination are different in ALT and telomerase-positive cells.

Introduction

Linear chromosomes contain repetitive hexameric sequences (TTAGGG in mammals) at their end, known as telomeres.¹ Telomeres form a loop-like structure (t-loop) that is protected by the shelterin complex. This shelterin complex is a macromolecular structure containing several telomere binding proteins that block DNA damage signaling, which would otherwise elicit from a linear chromosomal end.² One important function of telomeres is to serve as an expendable DNA buffer for the end replication problem.³ The DNA polymerase is unable to replicate the very end of the chromosome during lagging strand synthesis, which results in the loss of telomeric DNA if compensatory mechanisms are not present.

So far two of these compensatory mechanisms are known to overcome the end-replication problem in immortal cells. The first and most frequent mechanism involves telomerase, an enzyme that adds telomeric repeats to chromosomal ends.⁴ The second mechanism capable of achieving telomere homeostasis is the alternative lengthening of telomeres (ALT) pathway.⁵ Due to the lack of a specific ALT marker, the diagnosis of ALT is made when the telomerase pathway is firmly ruled out. Characteristic features of the ALT pathway are the lack of detectable telomerase activity and a heterogeneous pattern of telomere length, usually ranging from very short (< 1 kb) to abnormally long (> 20 kb).⁶ Furthermore, ALT cells contain ALT-associated promyelocytic leukemia nuclear bodies (PML) bodies, complexes consisting of PML protein plus telomeric DNA, telomere binding proteins such as TRF1 and TRF2 and proteins involved in DNA recombination (e.g., RAD50, RAD51, RAD52, MRE11, NBS1, BLM and WRN).⁷ Yet another characteristic of the ALT pathway is recombination between telomeres from sister chromatids (T-SCE), which is detected by Co-FISH analysis.^8,9 There is ample evidence that homologous recombination is involved in telomere maintenance in ALT cells both in yeast and in human cell models.¹⁰ Telomerase-negative yeast cells maintain telomeres via RAD52 and Kluyveromyces lactis cells transformed with tagged telomeric circles, obtaining long telomeres that show integration and amplification of the tag.¹¹ In human ALT cells, tagged telomeres show copy switching from one telomere to another, which was not observed in a telomerase-positive cell line.¹² Finally, ALT telomeres can harbor non-canonical repeats at the base of their telomere, which is suggestive of a recombination process that has taken place.¹³

Several individual mechanisms are proposed how telomeres are elongated in ALT cells.¹⁴ In the unequal T-SCE model, one telomere is elongated at the expense of the other sister telomere that gets shorter. If a proper segregation mechanism is in place then a cell population with long telomeres would emerge, whereas the daughter cells with the short ends would eventually succumb to death. In another model telomeric DNA is synthesized via homologous recombination-dependent DNA replication.¹⁵ Through this mechanism telomeric DNA is copied from a donor telomere to the recipient telomere, wherein the source of the telomeric template can be different. Via a break-induced replication process, a telomere from another chromosome can serve as a template leading to the copying of sequence from one telomere to another, resulting in a net increase in telomeric DNA.¹² Another potential source of template DNA is extrachromosomal telomeric DNA, which is abundantly present in ALT cells.^7,16-19 A third possibility is that the t-loop structure of the telomere itself might prime telomere polymerization.⁷

There are several indications that homologous recombination processes are not only restricted to ALT cells. The trimming of telomeres is a mechanism of telomere shortening that sets an upper telomere length limit.²⁰ This mechanism involves the resolution of recombination intermediate structures and requires recombination proteins like Rad52 and Mre11 in yeast. Extrachromosomal t-circles generated as product of a telomeric recombination process are used as a marker of telomere trimming.²¹ These t-circles have been detected in somatic as well as in telomerase-positive mammalian cells. Their presence indicates that recombination processes are occurring despite the absence of other markers of ALT activity. In general, homologous recombination is a response of the cell to a DNA double-strand break (DSB). DNA double-strand breaks arise to a significant amount from stalled replication forks.²² There are several possibilities, exogenous but also endogenous, as to how replication fork stalling arises. Among the endogenous sources are unusual DNA structures and fragile chromosomal regions, which are known to cause replication fork instability.^23,24 Examples of unusual DNA structures are DNA with nicks and gaps or DNA with tandem repeats that give rise to cruciform structures.^23,25 In addition, G-rich DNA is able to form G-G Hoogsteen base paired G4 DNA (g-quadruplexes), which represents yet another hindrance for DNA replication.^26,27 Due to its repetitive G-rich nature, it is therefore reasonable that telomeres have been identified as fragile DNA sites recently.²⁸ Sfeir et al. report replication fork stalling under experimental conditions in mammalian cells that lack TRF1. TRF1 as well as the two helicases RTEL1 and BLM are required for efficient telomeric replication. Furthermore it was shown that the DNA polymerase inhibitor aphidicolin, which induces gaps and breaks at common fragile DNA sites, also induces telomere fragility.²⁹ Taken together, these data suggest that the replication machinery encounters many hurdles when it comes to the replication of the telomere. DNA double-strand breaks due to replication fork stalling might therefore trigger homologous recombination in order to restore telomeric integrity. This would apply to ALT cells and telomerase-positive cells alike.

The present study used a novel homologous recombination reporter system that is able to quantify inter-telomeric recombination in a sensitive manner. We addressed the question whether homologous recombination between different telomeres is observed in telomerase-positive cells. We further looked at different conditions that influence inter-telomeric recombination and how telomerase-positive and ALT cells responded to these conditions. Our results showed that homologous recombination between different telomeres occurred in ALT, but also in telomerase-positive cells, with the same frequency. We describe that recombination led to an increase in reporter gene copy number in an ALT cell clone and to a lesser extent also in telomerase-positive cell clones. Inducing a fragile telomere phenotype with aphidicolin lead to a moderate increase in the rate of recombination in ALT cells but not in telomerase-positive cells. Exposure to a G4 ligand drastically increased recombination in ALT and telomerase-positive cells. Our results demonstrate for the first time that homologous recombination between different telomeres is not an exclusive characteristic of ALT cells.

Results

Cell clones with the telomeric reporter system

The basic outline of our subtelomeric and telomeric reporter construct is shown in Figure 1A and B. Each of the reporter plasmids isolated do not confer to puromycin resistance, since the coding sequence of the resistance gene is disrupted by telomeric DNA. When both reporter tags are present in the same cell, a recombination event between the two reporter constructs can occur due to the sequence homology provided by the telomere they are embedded in. The newly generated recombination molecule then contains the full-length puromycin resistance gene on the same telomeric end (Fig. 1C). Following transcription the intervening telomeric sequence between the splice donor and splice acceptor site is removed, thus leading to puromycin resistance in these cells (Fig. 1D).

Figure 1. The inter-telomeric recombination reporter system. (A) The subtelomeric tag is composed of a CMV-zeocin resistance gene followed by a SV40 promoter. The puromycin resistance gene is disrupted at its 3′end (and divided into puro 5′and puro 3′) by a sequence of 800 bp of telomeric DNA flanked by a splice donor (SD) and a splice acceptor site (SA). The construct was transfected into cells after linearization with AflII in order to induce a chromosomal healing event. In order to confirm that the very 5′ end of the plasmid got lost upon integration a PCR assay was performed using the SP6 and AflII-SA oligo. The probes used for Southern blot analysis to determine copy number and position of the tag were either the 5′puro probe or the full-length puromycin probe. (B) The telomeric tag has 800 bp of TTAGGG sequence at its 5′end followed by a splice acceptor site (SA) and the truncated puro 3′ gene. Cells were selected with hygromycin after transfection with this construct. When linearized with KpnI the construct is flanked by 800 bp of telomeric DNA at both sites allowing a chromosomal healing event. (C) Homologous recombination which involves the subtelomeric and the telomeric tag leads to co-localization of both tags on the same chromosome. (D) Due to the splice donor and the splice acceptor site the intervening telomeric DNA is removed rendering the cells now resistant to puromycin. Digestion of genomic DNA with XcmI and ScaI and subsequent Southern blot analysis with a puromycin- and hygromycin-specific probe allows the detection of both tags as a co-localizing signal.

Several control constructs were generated in order to ensure the specificity of our reporter system (Fig. S1). In order to generate stable clones harboring both reporter tags at the intended position, we first transfected SW 39 and SW 26 cells with the linearized subtel-tag plasmid. Southern blot analysis for single copy insertions showed that approximately 40% of the clones containing one copy of the subtel-tag (data not shown). These clones were further analyzed for the appropriate subtelomeric position of the tag by Bal31 digestion and subsequent Southern blot analysis. The diagnostic procedure is illustrated in Figure S2A. As shown in Figure 2A and B, clones with the tag located at the telomeric end are sensitive to Bal31 digestion. The puromycin-specific smear showed progressive reduction in size as a function of Bal31 incubation. In contrast, clones with subtel-tags at a non-telomeric position were Bal31-resistant even with an incubation time that exceeds 1h (data not shown). Since the telomere length of the tagged end is heterogeneous and sometimes hard to distinguish from background activity, we further confirmed the subtelomeric localization of the tag in clone SW 26 subtel-15 and clone SW 39 subtel-17 by a telomeric pull-down technique (Fig. S2B). As shown in Figure 2A and B, a smear was detectable in the fraction of isolated telomeres in SW 26 subtel-15 and SW 39 subtel-17 cells. The length of the smear corresponded to the length of the telomeric sequence attached to the tag and was long in the ALT cell clone but short in the telomerase-positive clone. This confirmed that tagged telomeres are processed according to the engaged pathway of telomere length control in SW 39 and SW 26 cells. In contrast, the supernatant fraction that is deprived of telomeres showed no signal when hybridized with the 5′puro probe. For clones with an internal position of the subtel-tag, no signal was retrieved in the telomeric pull-down fraction, but only in the supernatant fraction (data not shown). The presence of an intact reporter cassette on the subtel-tag would potentially allow cells to gain resistance via sister chromatid exchange. We therefore confirmed that the subtelomeric reporter lost parts of the very 5′end of the linearized plasmid upon integration into the genome. For this purpose, a PCR analysis as well as a plasmid rescue experiment was done. As shown in Figure S3A, no PCR signal was obtained for SW 26 subtel-15 and SW 39 subtel-17, suggesting that the site was lost upon vector integration. In order to obtain a DNA sequence from the vector integration site, a plasmid rescue experiment was performed. Sequence analysis of the retrieved plasmids showed that in SW39 subtel-17 the splice acceptor site and parts of the puro 3′gene were lost. With regard to the chromosomal integration site, T2AG₃ sequence variants like TTGGGG were present, suggesting that integration into a subtelomeric region has occurred. A similar result was obtained for the SW 26 subtel-15 clone. Again, the splice acceptor site and a small part of the puro 3′gene were lost, and the upstream sequence contained subtelomeric sequence arrays (data not shown).

Figure 2. Copy number and position of the telomeric reporter tags. (A) A SW 26 clone tagged with a single copy of the subtelomeric reporter was analyzed for the correct position. Bal31 digested DNA of clone subtel-15 was hybridized to a full-length puromycin probe. As indicated on the gel (brackets), a faint smear which decreases in size as a function of the Bal31 incubation time appears. The single band at 4.6 kb represents the puro 3′gene which also hybridizes to the full-length probe. The telomeric pull-down analysis isolates telomeres after XcmI digestion and shows a smear in the pull-down fraction, whereas no signal is detectable in the supernatant fraction. The reason for using XcmI in place of the HindIII restriction enzyme was to obtain a high molecular weight smear, which makes visualization of the tagged telomeres easier. (B) A SW 39 clone with a single copy of the subtelomeric tag was analyzed for the position of the tag by Bal31 digestion and telomeric pull-down analysis. Similar to the SW 26 clone, telomere length of the tagged telomere declined with increasing Bal31 exposure (brackets). For confirmation, telomeres were isolated after a HindIII digestion. Again a smear was detected in the pull-down fraction when hybridized to a full-length puro probe, whereas no signal was obtained after hybridizing the supernatant fraction. (C) The SW 26 subtel-15 clone was transfected with the telomeric tag and clones with a limited copy number of the tel-tag were further analyzed for their position. The clones tel-2 (total seven copies) and tel-4 (total two copies) showed a faint heterogeneous smear in the fraction of isolated telomeres. In the supernatant fraction three bands were detected for tel-2 and one band for tel-4, indicating that some copies in these clones where located at a non-telomeric position. AflII digestion generated only a faint smear, which is hardly discernible from background due to the heterogeneous length of the telomeres. Digestion of DNA with an enzyme that cuts outside the construct is leaving more non non-telomeric DNA attached to the isolated telomeres, therefore producing a high molecular and therefore less heterogeneous smear on conventional gel electrophoresis. As shown for the tel-4 clone, digestion with XcmI results in one band in the supernatant fraction and a smear in the pull-down fraction, indicating one internal and one telomeric copy of the tel-tag. (D) The SW 39 subtel-17 clone was transfected with the tel-tag and a number of individual clones with a limited number of copy insertions were isolated. Clones with a smear in the pull-down fraction and a lack of signal in the supernatant fraction were selected for subsequent experiments (tel-8, tel-12, tel-17). The three clones chosen had a different telomere length attached to the tel-tag. (E) A plasmid rescue experiment confirmed an intra-telomeric insertion of the tel-tag in one SW 26 and three SW 39 clones. Plasmid DNA was digested with a six-enzyme mix sparing telomeric DNA. For the SW 39 clones the resulting band of 800 bp corresponds to the size of the telomeric DNA flanking the tel-tag construct at the 5′ end. For the SW 26 clone a deeper intra-telomeric insertion was observed with a band of 1.4 kb length after 6-enzyme digestion. Hybridizing the gel with a telomere specific probe confirmed the presence of telomeric DNA in the retrieved plasmids.

In order to set up a complete reporter system, SW 39 subtel-17 and SW 26 subtel-15 clones were then transfected with the KpnI linearized tel-tag. As shown in Figure S3B, the vast majority of the SW 26 subtel-15/tel-tag clones showed multiple copies. Only two clones showed a limited number of reporter construct insertions. Clone tel-4 had two copies, and clone tel-2 had seven copies of the tag. As shown in Figure 2C, tel-4 had one copy at an internal position (single band) and one copy at a telomeric position (smear). Clone tel-2 had three copies at an internal and the remaining copies at a telomeric position. Since the smear in the pull-down experiment, which results from a single tagged telomere, is hard to visualize in ALT cells, we repeated the experiment with a restriction enzyme that cuts outside the tag, thereby leaving a bigger portion of chromosomal DNA attached to the tagged telomere. This results in a smear of higher molecular weight. As shown in Figure 2C, digestion with XcmI results in a high molecular weight smear in the pull-down fraction for clone tel-4 and a single band in the supernatant fraction, confirming the results from the experiment with the AflII digestion. For SW 39 subtel-17/tel-tag transfected cells, more than 60% of the clones had a single copy of the tel-tag present (Fig. S3C). The position of the tag was further confirmed to be telomeric in clones tel-8, tel-12 and tel-17 by telomeric pull-down analysis (Fig. 2D). Based on the criteria of a low copy number and the telomeric position of the tags, one ALT cell clone and three telomerase-positive cell clones were further analyzed. In order to confirm that integration did not destroy the splice acceptor site a PCR analysis (Fig. S3D) as well as a plasmid rescue experiment was done (Fig. 2E). The splice acceptor-specific PCR assay showed that due to the introduction of the tel-tag, a band becomes detectable in all subtel/tel-tagged clones. In addition, DNA retrieved from a plasmid rescue experiment was digested with a six-enzyme mix and subjected to a conventional TRF analysis. As shown in Figure 2E, telomeric sequences upstream of the tel-tag were rescued in the three SW 39 clones and the single SW 26 clone. The size of the telomeric fragment was 800 bp long, which matched the size of the TTAGGG₃ fragment upstream of the splice acceptor site in the tel-tag construct (Fig. 1B). In the SW 26 cell clone, the retrieved telomeric sequence was 1.4 kb long, suggesting a deeper intratelomeric insertion of the tag. In summary the diagnostic experiments identified three telomerase-positive and one ALT cell clone with the subtel-tag located in a subtelomeric region and the telomeric tag located intratelomerically, with at least 800 bp of TTAGGG₃ sequence upstream of the reporter tag.

Inter-telomeric homologous recombination in telomerase-positive and ALT cells

In order to measure the frequency of inter-telomeric homologous recombination in cells using different pathways of telomere length control, ALT and telomerase-positive cells were subjected to a recombination experiment. Each individual experiment was done with a minimum of three 35 cm tissue culture plates containing 1 × 10⁷ cells per plate. The results of at least seven independent experiments showed that the frequency of homologous recombination was the same in ALT and telomerase-positive cells (Fig. 3). In SW 26 subtel-15/tel-4, the mean number of colonies obtained per 1 × 10⁷ cells was 2.6 (range 0–35.5). In clone tel-2, the frequency of recombination was lower with a mean of 1 colony/1 × 10⁷ cells (range 0–0.3). For the three SW 39 subtel-17 clones the mean number of colonies per 1 × 10⁷ cells was 1.1 (range 0–3.5), 5.2 (range 0–28) and 2.6 (range 0–5) for clone number 8, 12 and 17, respectively. The frequency of recombination remained constant and did not increase with the number of experiments performed; ruling out that contamination of the culture with accumulating puromycin-resistant clones was taking place. Since puromycin could potentially force these cells to undergo recombination due to the selection pressure, we did a titration experiment. Cells were cultured for 1 PD each, with increasing concentrations of 125 ng/ml, 250 ng/ml and 500 ng/ml puromycin. The two lower concentrations of puromycin still allowed the cells to proliferate. Again, no difference in the frequency of telomeric recombination was observed, suggesting that our assay is selecting out those cells that already have undergone a recombination event rather than forcing cells into recombination due to the selection applied. Since the number of recombinant colonies varied between individual experiments, we performed a recombination experiment over several PDs. For this purpose 1 × 10⁷ cells from the tel-4 clone were kept in culture continuously and analyzed for the frequency of recombination every 4 PDs. As shown in Figure 4A, no puromycin-resistant colonies were detected in the ALT cell clone for most of the PDs, which is in part attributed to the lower number of cells used per experiment. At PD 63.6, a peak was observed with a number of 11 colonies/1 × 10⁷ cells followed by a burst of recombination for the following 15 PDs with a maximum of 80 colonies/1 × 10⁷ cells. Subsequently a decline was observed in the following PDs. Telomere length analysis over the range of 86,9 PDs showed no change in the median telomere length in this ALT cell clone (Fig. S4A). The frequency of recombination in the telomerase-positive cell clones did not show such variability (Fig. 4B and C). SW 39 subtel-17/tel-12 was followed for 81,5 PDs. For most of the PDs the frequency of recombination was below 10 colonies/1 × 10⁷ cells, with the exception of one experiment at PD 30. However, this elevated rate of recombination was only seen at one PD and did not persist over a certain period of time, like we did observe for the ALT cell clone. For the SW39 subtel-17/tel-17 clone no peak of recombination was observed over a range of 41,6 PDs. In order to address if homologous recombination increases as telomeres become shorter, we overexpressed a dominant-negative hTERT protein in the SW 39 subtel-17/tel-17 clone. Cells were monitored for levels of telomerase expression and telomere length. Recombination experiments were scheduled before cells entered crisis. In one dominant-negative hTERT subclone (#2), crisis was reached before the clone could be analyzed for recombination. Approximately 5 × 10⁶ cells went through crisis, and numerous colonies were obtained thereafter. As shown in Figure 5A and B, dominant-negative hTERT overexpression blocks telomerase activity to a various extent in different subclones. The enzyme activity ranged from 63% (subclone #2 after crisis) to 7% (subclone #8) compared with the parental cell clone. The higher telomerase activity in subclone #2 suggests that telomerase reactivation occurs after cells have emerged from crisis. The corresponding TRF analysis in these dominant-negative hTERT clones showed a shortening in telomere length in all clones compared with the parental clone (Fig. 5B). Even post-crisis clone #2 with a higher enzyme activity had shorter telomeres. For the recombination assay a total of 3 × 10⁷ cells were seeded per experiment, and blasticidin selection was stopped immediately before the experiment was performed. Despite the presence of reduced telomerase activity and very short telomeres, the frequency of homologous recombination was not increased (0–0.5 colonies/1 × 10⁷ cells) compared with the parental cell clone (Table 1).

Figure 3. The frequency of homologous recombination in ALT and telomerase-positive cell clones. A series of independent experiments with two ALT cell clones (SW 26 subtel-15/tel-4 and /tel-2) and three telomerase-positive cell clones (SW 39 subtel-17/tel-8, /tel-12 and /tel-17) showed that the mean number of recombinant clones per 1 × 10⁷ cells was the same.

Figure 4. The frequency of recombination over multiple population doublings. (A) SW 26 subtel-15/tel-4 ALT cells showed a low frequency of recombination for most of the PDs. A peak of recombination occurred between PD 67.5 and PD 83 followed by a drop in the number of recombinant clones to the baseline level thereafter. In (B and C) the frequency of recombination over several PDs in two telomerase-positive cell clones is shown. In contrast to the ALT cell line such a peak of recombination was not observed in SW 39 subtel-17/tel-12 and SW 39 subtel-17/tel17 cells.

Figure 5. Overexpression of a dominant-negative hTERT protein shortens telomeres in a telomerase-positive cell clone containing the reporter construct. (A) Overexpression of a dominant-negative hTERT protein shows inhibition of telomerase activity in subclones of SW 39 subtel-17/tel-17 cells. All clones showed a reduced telomerase activity with the exception of subclone #2, which showed higher enzyme activity. Experiments with this subclone were done after cells went through crisis, which is often associated with reexpression of telomerase. (B) Inhibition of telomerase leads to shortening of telomeres in all subclones as detected by TRF analysis. Due to the presence of critically short telomeres clones entered crisis soon after recombination experiments were done with the exception of subclone #2. In all subclones used for recombination experiments the median telomere length was shorter compared with the parental clone. The distinct bands present in all subclones represent interstitial telomeric DNA, which is typical for this cell line.

Table 1. Telomerase inhibition in SW 39 clones.

Clone SW 39 subtel-17	Telomere length	Telomerase activity	Crisis	t-HR colonies/1 × 107 cells
tel-17	3.6 kb	100%	-	2.5
tel-17 dnhTERT#2	1.7 kb	61%	post	0
tel-17 dnhTERT#3	3.0 kb	14%	pre	0
tel-17 dnhTERT#5	2.8 kb	24%	pre	0.5
tel-17 dnhTERT#8	3.0 kb	7%	pre	0

SW 39 clones with the recombination reporter system infected with a dominant-negative hTERT (dn) protein. Expression of dnhTERT leads to a reduced telomerase activity measured by q-PCR and to the shortening of telomeres. The frequency of inter-telomeric homologous recombination was low, regardless if cells were pre- or post- telomere length mediated crisis.

Recombination molecules in puromycin-resistant clones

Cells obtained puromycin resistance only via a homologous recombination event involving the telomeric sequences of both reporter tags. As a result, both parts of the puromycin gene come together on one telomeric end. We therefore did a Southern blot analysis to demonstrate tag-co-localization and to look for a change in copy number of the tags after a recombination process has taken place. Digestion of genomic DNA with XcmI and ScaI in cells that have not recombined results in a smear after Southern blot analysis when hybridized with the subtelomeric probe. In contrast a single band is expected when hybridized with the telomeric probe. However, if both tags come together at the same telomeric end, a single band is expected after subtelomeric and telomeric probe hybridization. In addition, both signals have to be on the same position on both gels, demonstrating co-localization of the tags.

Southern blot analyses of puromycin-resistant clones are shown in Figure 6A and B. For the parental cell clone SW 26 subtel-15/tel-4 and SW 39 subtel-17/tel-17 a smear was detected after puromycin hybridization due to the subtelomeric localization of the tag (see also Fig. 2A and B). The hygromycin-probed membrane showed two bands in the ALT cell line and one band in the telomerase-positive cell line as expected (see also Fig. 2C and D). However, the puromycin-resistant clones revealed a different picture. All SW 26 clones now showed distinct bands instead of a smear when hybridized with a puromycin-specific probe. Furthermore, one additional band appeared between 6 and 8 kbp length, suggesting that a gain in copy number of a subtel-tag has occurred. When the membrane was hybridized with the hygromycin-specific probe, two more copies of the telomeric tag were detected in the recombinant clones, which were both located between 6 and 8 kbp length. These newly gained tel-tag copies co-localized with the subtel-tag copies, showing that puromycin resistance was in fact associated with co-localization of both reporter tags on one chromosomal end.

Figure 6. Recombination molecules in puromycin resistant SW 26 ALT and SW 39 telomerase-positive clones. (A) Individual puromycin-resistant clones were isolated and analyzed for copy number and co-localization of the two reporter tags. In SW 26 cells the copy number of the subtel- and the tel-tag increased after cells became puromycin-resistant. Hybridization with a puromycin-specific probe as well as with a hygromycin-specific probe showed more copies compared with the parental cell clone. Co-localization of the puromycin-resistance gene and the hygromycin-resistance gene is present when both bands overlap at the same position on the gel. This occurred between 6 and 8 kbp length and was seen only for the newly gained copies of the tel-tag. (B) Puromycin-resistant clones in the telomerase-positive cell line showed either no change or an increase in copy number of the subtel-reporter tag. The copy number of the tel-tag remained unchanged in 5 out of 12 clones or was increased in 5 out of 12 clones. In two clones the hygromycin sequence was lost after recombination. Co-localization of the subtelomeric and the telomeric tag detected by puromycin and hygromycin hybridization was observed in 9 out of 12 clones.

The recombinant SW 39 clones presented also individual bands in the puromycin hybridized gel, with 8 out of 12 clones having an additional sub-telomeric tag gained. Resistant clone r3, r4, r11 and r12 retained one copy. Hybridization for the telomeric tag shows that 5/12 clones have gained at least one copy (r1, r2, r8, r9, r10), and 5/12 clones still have the same number of copies after the recombination process (r4, r5, r6, r7, r12). Two clones (r3 and r11) did not show a hygromycin-specific signal, suggesting a loss of the hygromycin resistance gene. With regard to co-localization clones, r3, r4 and r11 did not show a puromycin-hygromycin match. Since recombination can occur over a shorter stretch of DNA just involving the 3′puro part of the tag, we stripped the hygromycin probed blot and reprobed it with a puromycin-specific probe that detects the 3′puro part of the gene (Fig. 1A). Indeed, these clones showed a band that co-localized with the 5′puromycin gene, again confirming that all puromycin-resistant clones have to have a co-localization of at least both parts of the resistance gene (data not shown). Table 2 summarizes the data from the recombination molecule analysis. In contrast to the ALT cell clone, which consistently showed a gain in copy number in all recombinant clones, the three SW39 clones revealed a more heterogeneous pattern. A gain, as well as no change or even a loss of reporter tag copies were detected as a consequence of the recombination process.

Table 2. Copy number changes in resistant clones.

clones	copy number	SW 26 subtel-15/tel-4	SW 39 subtel-17/tel-8	SW 39 subtel-17/tel-12	SW 39 subtel-17/tel-17
subtelomeric tag	+ − 0	100% − -	− − 100%	66% − 34%	17% − 83%
telomeric tag	+ − 0	100% − −	− − 100%	42% 16% 42%	− − 100%

Changes of reporter gene copy numbers in SW 26 and SW 39 puromycin resistant cell clones compared with the parental cell clone. In SW 26 cells homologous recombination was associated with a gain (+) in copy number for the subtelomeric and the telomeric tag. In SW 39 cells either a gain (+), no change (0) or (−) in copy number for the subtelomeric reporter was observed. For the telomeric reporter even a loss of copy number was observed in the tel-12 clone.

Telomeric recombination induced by a fragile telomere phenotype

Telomeres are recognized as fragile chromosomal sites defined by their susceptibility to low doses of aphidicolin (APC).^28,30 We investigated whether ALT and telomerase-positive human cells respond differently with regard to telomeric recombination to the induction of a fragile telomere phenotype. For this purpose, cells were incubated with 1.6 μM of aphidicolin followed by a recombination assay. Aphidicolin is hindering DNA replication by blocking DNA polymerase activity, thereby creating DNA nicks and gaps. As shown in Table 3, the addition of aphidicolin increased the frequency of homologous recombination 1.8-fold in SW 26 cells. Since previous experiments showed that SW 26 cells have peaks of recombination, we looked at the effect of aphidicolin when cells were either in a low or in a high state of recombination. As a cut off to discriminate between the low and the high state of recombination, a value 1.5-fold of the mean frequency of recombination was used. We found this increase in recombination being present in cells with a low (1.6-fold) as well as in cells with a high state of recombination (1.8-fold) at approximately the same order of magnitude. In order to see whether telomerase activity alters this moderate increase of recombination, we infected SW 26 subtel-15/tel-4 cells with a retrovirus containing the hTERT gene together with an IRES-Blast cassette. Two individual colonies as well as a pooled population of hTERT-infected cells were used for subsequent experiments. As shown in the TRAP analysis in Figure 7A, infected cells expressed telomerase activity (hTERT clone #4 more than hTERT clone #3) but still showed the presence of long and heterogeneous telomeres typical for ALT cells (Fig. 7B). This finding is in agreement with a previous observation showing that telomerase expression does not always alter the telomere phenotype when expressed in ALT cells.³¹ With regard to telomeric recombination, we found a lower rate of recombination in the pooled cells and in the individual clones expressing telomerase activity compared with the ALT cell clone without telomerase activity (Table 3). When adding aphidicolin to the hTERT-expressing SW 26 subtel-15/tel-4 cells, no increase in recombination was observed. In the pool of infected cells, the frequency of recombination in the presence of APC was 0,33 colonies/1 × 10⁷ cells. For clone #3 (0,2 colonies/1 × 10⁷ cells) and clone #4 (0,3 colonies/1 × 10⁷ cells) similar results were obtained. When adding aphidicolin to telomerase-positive cells, no increase in the frequency of recombinant was observed. These results suggest that the presence of telomerase activity can suppress intertelomeric recombination in ALT cells and in cells with a fragile telomere phenotype induced by aphidicolin.

Table 3. Telomeric recombination induced by aphidicoline.

Cell clone	Frequency of recombination	− APC	+ APC	− TmPy	+ TmPy
SW 26 subtel-15/tel-4	mean	2,6	4,8	1,9	7,6
	low recombination (≤ 3,9)	2,5	4
	high recombination (> 3,9)	57,5	107
+hTERT pool	mean	0,3	0,3
+ hTERT clone 3	mean	0,2	0,2
+ hTERT clone 4	mean	0,3	0,3
SW 26 subtel-15/tel-2	mean	2,6	6,0
SW 39 subtel-17/tel-12	mean	4,3	4,4	1,6	10,1
SW 39 subtel-17/tel-17	mean	2,6	2,9

Telomeric recombination in SW 26 and SW 39 cells exposed to aphidicoline and TmPy4 (expressed as number of colonies/1 × 10⁷ cells). Aphidicoline lead to a moderate but consistent increase in the frequency of homologous inter-telomeric recombination in two ALT cell clones. This increase was observe in cells that were in a lower state of recombination (cut off 1.5-fold of the mean frequency of recombination) as well as in cells that were at a recombinogenic peak. Reconstitution of telomerase activity in SW 26 cells abolished APC induced recombination. In addition no increase in recombination with APC was seen SW 39 cell clones. The G4 DNA stabilizing substance TmPy4 however increased the frequency of inter-telomeric recombination in SW 26 and SW 39 cells.

Figure 7. Expression of telomerase in an ALT cell clone. (A) Overexpression of hTERT in SW 26 cells leads to the reconstitution of telomerase activity. Compared with the parental cell clone, tel-4 hTERT#3 and tel-4 hTERT#4 have more TRAP activity. (B) Despite the presence of telomerase activity in ALT cells, telomere length remains heterogeneous. Notably, the signal from very short telomeres appeared reduced in the telomerase-expressing subclones compared with the parental ALT cell line.

We next went on to see if homologous recombination between telomeres is increasing when a G4 ligand is imposed to the cell. Telomeric sequences are known to form g-quadruplex structures in vitro,^32,33 and there are hints that they also form in vivo.³⁴ We incubated SW 26 subtel-15/tel-4 and SW 39 subtel-17/tel-12 cells with the G4 ligand TmPy4. As shown in Table 3 homologous recombination within telomeres was elevated in SW 26 and in SW 39 cells when incubated with TmPy4 compared with the untreated control cells (4- and 6-fold, respectively). The maximum frequency of recombination detected for the ALT cell clone was 85 colonies/1 × 10⁷ cells (mean: 7.6 colonies/1 × 10⁷ cells) and 65 colonies/1 × 10⁷ cells in the telomerase-positive cell line (mean: 10.1 colonies/1 × 10⁷ cells).

Discussion

Our data show for the first time, that inter-telomeric homologous recombination is not only detected in ALT but also in telomerase-positive cells. The concept that homologous recombination is the mechanism by which ALT cells maintain their telomeres is based on the observation that tagged telomeres show tag dispersal to other telomeric ends.¹² This copy switching of tags was not observed in telomerase-positive cells. However, following a tagged telomere in a pool of cells inherits the potential for a bias. In order to become detectable with Southern blot or FISH technique, either many cells must have undergone a recombination event simultaneously, or cells have gained a growth advantage after recombination. This bears the risk that a significant number of recombination events are missed due to the low sensitivity of the method of detection. DNA replication through telomeres is posing a challenge for the DNA replication machinery. Stallments of the replication fork and subsequent repair mechanisms, like homologous recombination strive to resume replication and hence are expected to occur physiologically.³⁵ One would expect that replication difficulties and, as a consequence, homologous recombination occur in both cell lines alike. Therefore our finding, albeit not unexpected, formally proves that recombination is present in telomerase-positive immortal human cells. More recently Neumann et al. showed, with the same reporter system they used before, that inter-telomeric recombination is also present in normal somatic cells in a mouse model. This finding supports the idea that ALT activity is an inherent element of normal telomere biology.³⁶

Extensive molecular analysis showed that our reporter system is located at the base of the telomere, and we therefore cannot assess the frequency of recombination downstream of the tags. Therefore our data do not preclude that the frequency and pattern of recombination is different further to the telomeric end. Our results are in line with the finding that recombination in regions closer than 10 kb to the subtelomere-telomere border are observed in human cells.³⁷ Sister chromatid exchange rates in the subtelomeric region are elevated compared with other regions of the genome in an EBV immortalized human cell line. In addition sequence analysis of subtelomeres reveals a high interindividual variability with duplications and deletions, suggesting that this is a region of active recombination.^38,39 Furthermore, subtelomeres have been shown to be more susceptible to DNA double-strand breaks in yeast and mammalian cells.^40,41

The number of recombinants measured consecutively in one ALT cell clone was very variable. Due to the fact that we had only one clone with a reasonably low number of reporter tag copies it is difficult to draw a conclusion out of this experiment. Nevertheless we observed bursts of recombination lasting over several PDs, resulting in a drastic increase in the number of puromycin-resistant clones. This observation is similar to the observation of telomere dynamics described by the seminal paper of Murnane et al.⁴² Herein the authors report that marked telomeres in ALT pathway cells undergo gradual as well as rapid shortenings. These shortenings are followed by rapid elongations, which do not correlate with the length of the telomere itself. Such de-repressions of a basal level of ALT activity are occurring, e.g., under experimental situations where mutant telomeric sequences can activate ALT features even in the presence of telomerase.⁴³ Mutant telomeric sequences most likely bind less proteins of the shelterin complex. The presence of sequence variability at the base of telomere in ALT cells makes it likely that shelterin function is impaired. This would eventually lead to a dysfunctional telomere with a less repressive chromatin structure.^44,45 We did not observed such variability in recombination in telomerase-positive cell clones; however, more data from experiments with other ALT cell clones would be needed to link this clearly to underlying biological processes.

There is experimental evidence that short telomeres are recombinogenic in yeast and mammalian cells, including human cancer cells.^9,46,47 The supporting working hypothesis deals with the idea that short telomeres are getting uncapped, hence leading to recombination in the absence of telomerase. Our result that short telomeres do not increase inter-telomeric recombination challenges this concept. In four subclones of a tagged telomerase-positive clone, the expression of a dominant-negative hTERT protein showed inhibition of telomerase activity. Despite the fact that telomerase inhibition was not complete, these subclones had significantly shortened telomeres. One explanation for the contradiction with previous reports could be that the remaining levels of telomerase activity are sufficient to prevent recombination. One could speculate that short telomeres bind more TRF1 and TRF2, which further bind more TRF1 interacting nuclear factor 2 (TIN2). TIN2 has been shown to bind heterochromatin protein 1 (HP1), which increases telomere cohesion.⁴⁸ Such cohesive telomeric ends are more efficiently elongated by telomerase, suggesting that even very low levels might suffice to maintain telomere length control. It is also noteworthy that previous reports dealt with cells that either lack any telomerase activity or with cells expressing a mutant form of the enzyme.^46,47 Another explanation is that these cells can engage telomeric sister chromatid exchange. Morrish et al. looked at the p/q ratio of telomeres by FISH analysis in mammalian cells, an assay that captures the effect of various forms of recombination. Our assay is only capable of measuring inter-telomeric recombination, hence the presence of other types of recombination in critically short ends cannot be ruled out.

Analysis of recombination molecules showed a complex picture. If a break induced replication process (BIR) would be the only mechanism of recombination one would expect that the copy number of the subtel-tag is unchanged but an increase in the number of the tel-tags would occur. Instead, we observed an increase in the number for both tags in SW 26 cells, which makes it likely that not just a break-induced replication process has taken place. In addition the pattern of bands observed on Southern blot analysis was quite similar for all SW 26 clones with only minor variations in size. This suggests that a single mechanism of recombination was engaged. Since we picked clones from different plates a contamination with satellite colonies can be ruled out. The recombination molecules generated by the SW 39 clones showed more variation. Here copy number amplifications as well as no change or even deletions of copies were observed. Incubation of human cells with aphidicolin induced a fragile telomere phenotype in human fibroblasts expressing hTERT and Hela 1.3 cells.²⁸ Our experiments showed that aphidicolin induces homologous recombination in ALT cells but not in telomerase-positive cells or in ALT cells expressing telomerase. The magnitude of this effect was small (approximately 1.8-fold) but is in the same order of magnitude as the increase of fragile telomeres reported by Sfeir et al. ALT cells are known to have nicks and gaps within their telomeres and by adding aphidicolin more lesions are expected to be generated.²⁵ Telomerase itself might have a protective effect. Nicks and gaps that lead to DNA double-strand breaks might get repaired by the enzyme as end processive mechanisms act on the broken end. In addition SW 39 cells have very short telomeres unlike BJ-hTERT or Hela 1.3 cells used by Sfeir et al. One would expect that longer telomeres pose a greater challenge to telomere replication than short ends.

The finding that homologous recombination between non-sister telomeres is present not only in ALT, but also in telomerase-positive cells has several implications. It adds to the problem that the diagnosis of ALT is already hampered by the fact that no specific marker for ALT exists, and an accurate diagnosis of the ALT pathway is based on the presence of telomere maintenance without detectable telomerase activity.⁴⁹ The smaller the number of specific ALT characteristics the more difficult it is to find a specific diagnostic marker for ALT. As a result the biological relevance of ALT in tumor specimen is hard to assess. When current diagnostic tools are applied, a contamination with telomerase expressing stromal cells and tumors with ALT and telomerase activity present at the same time lead to a false-negative ALT diagnostic. Inter-telomeric recombination is a mechanism whereby telomeres can potentially be elongated. In telomerase-positive cells, inter-telomeric recombination can therefore be a source of treatment resistance when telomerase activity is blocked in tumor cells. The fact that we did not see a vast increase in recombination in cells expressing a dominant-negative hTERT protein, is most likely due to the remaining levels of telomerase activity, which might be sufficient to prevent inter-telomeric recombination. However, if potent antitelomerase inhibitors become available in the clinical setting, this might become an issue of concern. A G4 DNA ligand is yet another therapeutic possibility to disrupt telomere homeostasis. The fact that it strongly induces recombination underscores the necessity that each telomere-directed therapy requires a careful evaluation for the potential of homologous recombination being a mechanism of treatment resistance.

A central question remains whether inter-telomeric recombination is able to elongate telomeres in telomerase-positive cells. In fact it has recently been shown that features of both pathways can be present in one cell, which underscores the importance of this question.⁵⁰ Our observation, that recombinant clones showed a gain in copy numbers of the subtelomeric but also the telomeric tag suggests that this gain in DNA might also be associated with an actual gain of telomeric DNA. If this observation is extended to a clinical scenario, the mere application of a telomerase inhibitor might not suffice for a telomere targeted therapy. The impact of telomerase not only on tumor cell but also on tumor stem cell viability illustrates that blocking telomerase activity is a multi-facetted treatment strategy.⁵¹ However if telomeric recombination is a pathway that tumor cells use to escape anti-telomerase therapy, more sophisticated approaches would be required which attack both pathways of telomere length control at the same time. One example for such a treatment approach is targeting the Ku70/80 protein heterodimer, which is involved in non-homologous end joining (NHEJ) of DNA double-strand breaks. Inhibiting Ku70/80 function in telomerase-positive cells leads to telomere deprotection, resulting in telomere shortening, end-to-end fusions and genetic instability.⁵² In ALT cells telomere length is unchanged when Ku70/80 is depleted, but the levels of t-circles are drastically reduced, which is associated with a robust growth inhibition.⁵³ Treatment strategies which tackle one target and block both telomere length control mechanisms are clearly advantageous over a combined strategy where each individual pathway is targeted simultaneously. The increasing complexity of connected and redundant pathways of telomere homeostasis, underscore the importance of basic research to elucidate these individual mechanisms. A telomeric recombination reporter system is a valuable tool not only with regard to elucidating the mechanism of ALT but also to assess the efficacy of potential drugs that target the ALT pathway.

Material and Methods

Subtelomeric and telomeric reporter constructs

For the subtelomeric reporter construct (subtel-tag) the SV40-puromycin resistance gene was isolated from the pBabe-Puro retroviral vector by HindIII and ClaI restriction enzyme digestion (New England Biolabs). The isolated fragment was ligated into the polylinker region of the pSV-Zeo plasmid (Invitrogen). After obtaining a clone which carried the puromycin resistance gene, a BcgI digestion was done (New England Biolabs). BcgI is a single cutter in the 3′part of the purmomycin resistance gene. After isolating the linear fragment an oligo was ligated into the plasmid containing a splice donor site (SD) followed by a short polylinker region (XbaI, KpnI, ClaI, AflII) and a splice acceptor site (SA) at its end. The polylinker of the new plasmid was digested with KpnI and ClaI and an 800 bp KpnI-TTAGGG_n-ClaI fragment taken out of the pSXNeo1.6T2AG₃ plasmid was inserted.⁵⁴ In order to generate a chromosomal healing vector the sub-tel tagging plasmid was linearized with AflII (Fig. 1A).^54,55 For the telomeric reporter construct (tel-tag) a 800 bp SmaI-TTAGGG_n-KpnI fragment was isolated from the pSXNeo1.6T2AG3 plasmid and was dropped into the pCDNA3.1 Hygro (−) vector (Invitrogen) via a blunted SacI and a KpnI cloning site (pCDNA3.1 Hygro_TTAGGG). The newly formed construct was opened with HindIII. After Klenow blunting the end, the construct was digested with KpnI followed by gene cleaning of a 6.4 kb fragment. This fragment now contained the SV40-hygromycin gene followed by the ampicillin resistance gene and 800 bp of telomeric sequence. A fragment containing the TTAGGG_800bp—splice acceptor—3′puromycin resistance gene was isolated from the sub-tel tagging construct by digestion with BamHI and KpnI. The BamHI site was blunt ended by Klenow treatment. After gene cleaning, the fragment was ligated with the 6.4 kb pCDNA3.1 Hygro_TTAGGG fragment via a blunt end (HindIII and BamHI) and sticky (KpnI) ligation. In order to generate a chromosomal healing vector, the construct was linearized with KpnI, which generates a plasmid that is flanked by two telomeric sequences at each end (Fig. 1B). The control constructs are either versions of the subtel-tag or isolated fragments from both tags (Fig. S1).

Cell culture and clones with subtelomeric and telomeric tags

SW 26 ALT cells and SW 39 telomerase-positive cells were grown as previously described.⁵⁶ In order to place the subtel-tag into the intended subtelomeric position, cells were transfected with 1 μg of AflII linearized plasmid (Fugene, Roche) following selection with 400 μg/ml zeocin (Invitrogen). Individual clones were isolated. To identify clones with a single copy of the subtelomeric tag, genomic DNA was isolated and digested with KpnI followed by Southern blot analysis with a puromycin-specific probe. A single band visible on the Southern blot corresponds to a single copy of the tag. Once clones with a single copy were identified the subtelomeric position was verified. For this purpose a Bal31 digestion and a telomeric pull-down assay was performed. Once a clone with a single copy of the tag at a subtelomeric position was identified, this clone was transfected for a second time in order to introduce the telomeric tag. For this transfection 1 μg of a KpnI linearized tag tel-tag was used following selection with 300 μg/ml hygromycin (Invitrogen). Again, genomic DNA from individual clones was used to screen for single copy insertions. DNA was digested either with AflIII or ScaI enzyme followed by Southern blot analysis with a hygromycin-specific probe. Clones containing a limited number of copies were further analyzed for the correct endchromosomal position with a telomeric pull-down assay.

In order to induce a fragile telomere phenotype, cells were incubated with aphidicolin (Sigma). Log phase growing SW 26 and SW 39 tagged clones were incubated with 1.6 μM of aphidicolin (APC) under ambient culture conditions. Under these conditions, cells were still proliferating. Lower concentrations of APC down to 0.4 μM were giving the same results. After 16 h, cells were washed three times in PBS and cultured for one population doubling (PD) in cell culture media. After cells divided once, the culture medium was replaced by a puromycin (Invitrogen) containing media (500 ng/ml) and kept under selection until colonies emerged. The g-quadruplex stabilizing agent TmpY4 (Calbiochem) was added to log phase growing cells at a concentration of 25 μM for a period of 48h. After 48 h cells were washed three times in PBS and after one PD puromycin selection followed.

Bal31 digestion

In order to identify the subtelomeric localization of the reporter plasmid, 20 μg of genomic DNA was digested for variable time with 3U of Bal31 in a 1× buffer provided by the manufacturer at 30°C (New England Biolabs). The reaction was heat inactivated for 10min at 65°C in the presence of 20 mM EGTA. After phenol chloroform purification, the DNA was precipitated and resuspended in TE buffer. Ten μg of Bal31 digested DNA was subjected to restriction enzyme digestion with 100 U of HindIII followed by gel electrophoresis and Southern blot analysis with a puromycin- or hygromycin-specific probe.

Telomeric pull-down assay

In order to detect an endchromosomal position of the subtel-tag and the tel-tag a telomere pull-down analysis was performed as described previously (Fig. S2).^19,57 Forty μg of genomic DNA were digested with HindIII for the subtel-tag analysis and with AflII for the tel-tag analysis. Subsequently, the DNA was subjected to an annealing reaction to a biotin-labeled C-rich oligo (AATCCC)₆ (Metabion) at declining temperatures of 80, 65, 55, 45 and 35°C for 30 min each in 1 × SSC/0.1% Triton X-100. Next streptavidin-coated magnetic beads (20 μl of a 10 mg/ml stock solution) (New England Biolabs) that were pre-washed (1 × SSC/1% Triton X-100, coated with 5× Denhardt’s for 30 min at RT) and resuspended in 1× SSC were added to the reaction mixture. The samples were incubated on a rotator overnight at 4°C. To recover DNA attached to telomeric sequences, the beads were collected with a magnet, and the supernatant was separated. The beads were gently washed twice in ice cold 1× SSC/1% Triton X-100. To elute the telomeric DNA, the beads were brought into suspension with 50 μl of 1× TE buffer and incubated at 65°C for 10 min. DNA obtained from the pull down as well as DNA from the supernatant was analyzed by Southern blot hybridization with a probe specific to puromycin or hygromycin.

Southern blot analysis

Digested DNA samples were run on a 0.8% agarose gel in 1× TAE buffer at 1–2 V/cm overnight. After nicking the DNA in 0.25 M HCl the gel was denatured in 0.5 M NaOH/1.0 M NaCl for 2 × 20 min at RT followed by neutralization in 0.5 M Tris pH 7.5/1.5 M NaCl twice for 20 min each. The DNA was blotted onto a nylon membrane (Hybond N membrane, Amersham) by capillary transfer in 20× SSC buffer overnight. After crosslinking the membrane was prehybridized in a low temperature hybridization buffer containing 50% formamide, 5× Denhardt’s solution, 5× SSPE, 0.2% SDS, 100 μg/ml denatured salmon sperm DNA and 10% dextran sulfate at 42°C for 30 min. Hybridization was done overnight at 42°C in a roller tube with a p53-labeled DNA probe. A random prime DNA labeling kit was used for probe labeling using 50 μCi of dCTP (Roche). Prior to hybridization the probe was purified with a Sephadex column, denatured at 95°C for 5 min and stored on ice until further used.

Splice acceptor site specific PCR

Two hundred ng of undigested genomic DNA was amplified with 12.5 pmol of an AflII-splice acceptor oligo (5′-ttaagcccttagggtctg-3′) and a SP-6 oligo (5′-gctatttaggtgacactataga-3′) in a reaction volume of 25 μl containing 200 μM of dNTPs, 2.5 U of failsafe enzyme mix in 1× fail safe buffer H according to the manufacturers protocol (Epicenter). PCR conditions included an initial denaturation step of 95°C for 3 min with subsequent denaturations for 30 sec, annealing at 57°C for 30 sec and elongation at 68°C for 1 min for a total of 39 cycles. PCR products were visualized on an ethidium bromide stained gel.

Plasmid rescue procedure

A plasmid rescue procedure was done in order to confirm the subtelomeric and telomeric insertion by obtaining sequence information about the insertion site. Ten μg of genomic DNA was digested with HindIII (subtel-tag) or BglII (tel-tag) in order to rescue the reporter and the flanking chromosomal DNA lying between the tag and the next HindIII or BglII restriction site. After phenol-chloroforme-IAA purification and sodium acetate precipitation, the DNA was religated in 1× ligation buffer with 10 U of T4 DNA ligase in a volume of 300 μl at RT for a minimum of 5 h (Roche). After DNA ligation, the samples were once more phenol-chloroform-IAA purified, precipitated again and resuspended in 5 mM TRIS-HCl pH 8.5. A total amount of 200 ng of DNA was electroporated into DH5a competent cells following the manufacturer’s protocol (Invitrogen). Individual colonies were isolated after zeocin (50 μg/ml) or ampicillin (100 ng/ml) (Invitrogen) selection and subjected to mini prep analysis. The obtained plasmid DNA was sequenced wherein for sequencing the SP-6 primer was used. An intratelomeric localization of the tag was confirmed if T2AG₃ DNA was rescued together with the reporter. Therefore the DNA was digested with a six-enzyme mix (AluI, CfoI, HaeI, HinfI, MspI, RsaI) followed by a conventional TRF analysis to detect any telomeric sequence rescued together with the plasmid.

Retroviral infection

A retrovirus was generated as described previously.⁹ The pWZL-Blast-hTERT and pWZL-Blast-D869A dominant-negative hTERT plasmid was a gift from J. Shay and W. Wright at UT Southwestern Medical Center Dallas, TX. The supernatant of blasticidin-selected PA317 cells was used to infect SW 26 and SW 39 tagged cells. After selection with 5 μg/ml of blasticidin (Invitrogen) individual clones were picked and kept under selection throughout the experiment. Since the median telomere length in SW 39 cells is short (3.5 kb) the dominant-negative hTERT infected cells were used for a recombination experiment as soon as enough cells (3 × 10⁷) were available. One clone entered crisis due to short telomeres before an experiment could be done. Most of the cells from this clone were lost during crisis but the remaining cells were used for an experiment as soon as the culture was restored. For the pWZL-Blast-hTERT infected cells, single clones but also the pool of infected cells were used for experiments. Transient transfections were done using 10 μg of the pWZL-Blast-hTERT plasmid transfected with Fugene reagent (Roche).

TRF, TRAP and telomerase qPCR analysis

Nucleic acid for TRF analysis was obtained with the Purgene DNA extraction kit (Gentra Systems). Plasmid DNA was obtained by a mini prep procedure. The amount of DNA obtained was measured with a photometer (Eppendorf). For telomere length analysis a total amount of 2 μg of genomic DNA was digested with a six-enzyme mix (Alu I, Cfo I, Hae I, Hinf I, Msp I and Rsa I) (New England Biolabs) at 37°C overnight. Digested DNA was run on a 0.8% agarose gel in 1× TAE buffer at 1.4 V/cm. After denaturing (0.5 M NaOH, 1.5 M NaCl) for 20 min the gel was rinsed in H₂O for 10 min, dried at 56°C for 3 h and finally neutralized (1.5 M NaCl, 0.5 M Tris). After neutralization, hybridization with a P32-labeled (TTAGGG)4 oligonucleotide (Metabion) was performed at 42°C for 16 h. After washing the gel once in 2 × SSC (1× = 0.15 M NaCl, 0.015 M Na citrate, pH 7.0) for 15 min at room temperature followed by two washes with 0.1 × SSC/0.1% SDS for 10 min at room temperature, the gel was exposed to a phosphor screen and analyzed on a phosphor imager (Cyclone Storage Phosphor System, Perkin Elmer).

The mean telomere length was calculated as previously described.⁵⁸ Telomerase activity was measured using a Trapeze kit (Integrin) following the basic protocol of the manufacturer. For each assay, the equivalent of 1,000 cells was used, and PCR amplification was done for 30 cycles. The telomerase quantitative RT PCR was done following the protocol published by Herbert et al.⁵⁹

Homologous recombination assay and of recombination products

The frequency of homologous recombination was measured by seeding 1 × 10⁷ cells on a 30 cm cell culture dish (Becton Dickinson). For each experiment, a minimum of three dishes was used with the exception of the continuous experiments. After 10–12 h, cells were adherent, and the tissue culture medium was replaced by a culture medium containing 500 ng/ml puromycin. In case of a recombination event colonies emerged after approximately 7 d, which were stained after they have grown to a macroscopically visible size (150–1,000 cells). Colonies were stained after fixation of the plate with methanol for 15 min followed by the application of a 5% Giemsa stain for 1 h at RT. The number of colonies per 1 × 10⁷ cells represented the number of detectable recombination events between the sub-tel and the tel-tag and represents the frequency of recombination.

For analysis of the recombination molecules puromycin-resistant clones were harvested by ring cloning after they have grown to a size of 50–70 cells/colony. Clones were propagated in puromycin, and after they reached a number of 1–3 × 10⁷ cells, half of them were frozen, and the other half was used for DNA extraction. In order to analyze the recombination molecules genomic DNA was digested with the restriction enzyme XcmI, which is a non-cutting enzyme for both tags. After XcmI digestion, the samples were digested again with ScaI, which cuts in the tel-tag (Fig. 1C). Samples were divided equally and run on two separate gels under the same conditions followed by the hybridization with a purmomycin- or a hygromycin-specific probe. Co-localization of both tags was diagnosed by the appearance of the puromycin- and hygromycin-specific bands at the same place on both membranes.

List of reagents

• Restriction enzymes (New England Biolabs):

• HindIII (R0104S), ClaI (R0197S), XbaI (R0145S), KpnI (R0142S), AflII (R0520S), SacI (R0156S), AluI (R0137S), CfoI (R6241), HaeI (R0107S), HinfI (R0155S), MspI (R0106S), RsaI (R0167S)

• Plasmids (Invitrogen):

• pSV-Zeo V860-20

• pCDNA3.1 Hygro (−) V870-20

• Reagents and consumables:

• Fugene Transfection reagent (Roche) 05061377001

• Zeocin (Invitrogen) ant-zn-1

• Hygromycin (Invitrogen) 10687-010

• Aphidicolin (Sigma) A0781-1MG

• Puromycin (Invitrogen) A11138-02

• TmpY4 (Calbiochem) 613560-25MG

• Bal31 (New England Biolabs) M0213S

• streptavidin-coated magnetic beads (New England Biolabs) S1420S

• Hybond N membrane (Amersham) RPN82N

• DNA random prime labeling kit (Roche) 11004760001

• T4 DNA ligase (Roche) 10481220001

• DH5a competent cells (Invitrogen) 18258-012

• Blasticidin (Invitrogen) R210-01

• Purgene DNA extraction kit (Gentra Systems) 158422

• Trapeze kit (Intergen) S7700

• Failsafe PCR kit (epicenter) FS99060

Glossary

Abbreviations:

ALT

alternative lengthening of telomeres

APBs

ALT-associated PML bodies

APC

aphidicolin

BLM

Bloom helicase

EBV

Epstein Barr virus

Co-FISH

chromosomal orientation fluorescence in situ hybridization

NBS

Nijmegen breakage syndrome

NHEJ

non-homologous end joining

HP1

heterochromatin protein 1

PDs

population doublings

PML

promyelocytic leukemia

Rad 50

51, 52 radiation-sensitive protein

RTEL1

regulator of telomere elongation

SV40

Simian virus 40

TIN2 TRF1

interacting nuclear protein 2 t-loop, telomeric loop TRF-1/2 telomere repeat binding factors ½

T-SCE

telomeric sister chromatid exchange

WRN

Werner helicase

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplemental Materials

Supplemental materials may be found here: 
www.landesbioscience.com/journals/cc/article/25136