Telomere Maintenance Without Telomerase

Telomere maintenance without telomerase

Victoria Lundblad*,1

1Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas, TX 77030, USA

Recombination-dependent maintenance of telomeres, ®rst possibility that a potentially similar pathway operating discovered in budding yeast, has revealed an alternative in human cells, referred to as ALT (for Alternative pathway for telomere maintenance that does not require Telomere Lengthening), might contribute to neoplastic the enzyme telomerase. Experiments conducted in two transformation. This review discusses the investigations budding yeasts, S. cerevisiae and K. lactis, have shown in budding yeast that have helped elucidate this recombination can replenish terminal G-rich telomeric recombination-dependent process, followed by a dis- tracts that would otherwise shorten in the absence of cussion of the possible mechanisms that have been telomerase, as well as disperse and amplify sub-telomeric proposed as a result of the yeast studies. A brief repeat elements. Investigation of the genetic requirements summary of the ALT pathway, and a discussion about for this process have revealed that at least two dierent whether ALT proceeds through a similar recombina- recombination pathways, de®ned by RAD50 and tion-dependent pathway, concludes this review. The RAD51, can promote telomere maintenance. Although reader is also referred to other excellent reviews on this critically short telomeres are very recombinogenic, topic (Biessmann and Mason, 1997; Kass-Eisler and recombination among telomeres that have only partially Greider, 2000; Reddel et al., 2001; Kraus et al., 2001). shortened in the absence of telomerase can also contribute to telomerase-independent survival. These observations provide new insights into the mechanism(s) Discovery of a telomerase-independent pathway for by which recombination can restore telomere function in telomere maintenance yeast, and suggest future experiments for the investigation of potentially similar pathways in human cells. Several observations during the 1980's provided early Oncogene (2002) 21, 522 ± 531. DOI: 10.1038/sj/onc/ clues that recombination could potentially maintain 1205079 telomeric repeats at chromosome ends. Linear plasmids that lacked a functional telomere, but contained Keywords: telomerase; EST1; ALT; recombination; homology to sub-telomeric regions, could acquire a break-induced replication functional chromosome terminus in a homology- dependent process (Dunn et al., 1984). Rescue of these linear plasmids was proposed to involve strand Linear chromosomes must have a mechanism to ensure invasion of the sub-telomeric region of a chromosome that their termini are fully duplicated when the by the terminus of the linear plasmid, followed by chromosome itself is replicated. For most eukaryotes, replicative copying of the donor onto the end of the the ends of nuclear chromosomes are maintained by plasmid (Figure 1a). The concept that a genetic the reverse transcriptase telomerase. Telomerase adds recombination intermediate could be converted into a species-speci®c telomeric repeats onto chromosome replication fork had already been proposed as a termini, thereby counterbalancing sequence loss that mechanism for replication of the ends of the linear would ensue as a result of semi-conservative DNA bacteriophage T4 genome (Luder and Mosig, 1982). replication by the conventional DNA replication Formosa and Alberts (1986) provided biochemical machinery (for reviews, see Greider, 1996; McEachern support for the contribution of recombination proteins et al., 2000). This not only ensures complete replication to the formation of a replication fork, and further of the genome, but the telomeric repeats that are suggested that homology-initiated DNA replication synthesized by telomerase provide a binding site for the might also apply to the telomere end replication host of telomere-speci®c proteins that prevent recogni- problem. Similar recombination-based models for tion of chromosome termini as double strand breaks telomere maintenance had also been elaborated by (for reviews, see Lundblad, 2000; Blackburn, 2001). Bernards et al. (1983) and Walmsley et al. (1984). However, increasing interest has been focused on a Direct demonstration that recombination could main- recombination-dependent mechanism for replenishing tain the natural termini of linear chromosomes, however, telomeric repeats. This pathway, ®rst discovered in awaited the analysis of telomerase-defective yeast budding yeast, is receiving attention due to the strains. The discovery of the S. cerevisiae EST1 gene (Lundblad and Szostak, 1989), later shown to encode a subunit of the telomerase holoenzyme (Hughes et al., *Correspondence: V Lundblad; E-mail: [email protected] 2000), allowed examination of the consequences of a Telomere maintenance without telomerase V Lundblad 523 essential genes (as a result of extensive terminal deletions) and/or chromosome transmission failures resulting from end-to-end fusions that trigger a break- age-fusion-bridge cycle. linear plasmid critically short telomere Although the majority of yeast cells die in the absence resection resection of telomerase, Lundblad and Blackburn (1993) observed that a small subpopulation could escape the lethal consequences of a telomerase defect. The rare survivors 3’ 3’ recovered from an est1-D strain not only regained the ability to grow but also displayed dramatic changes at strand exchange strand exchange their chromosomal termini, due to global ampli®cation of both telomeric and sub-telomeric repeat sequences. These extensive rearrangements arose as a result of recombination, as the appearance of survivors was extension of extension of invading strand invading strand blocked if the est1-D strain was also defective for RAD52, which is responsible for the majority of homologous recombination events in yeast. Survivors could be grouped into two general classes based on the second strand second strand pattern of ampli®cation of telomeric and/or sub- synthesis synthesis telomeric sequences. Although survivors were recovered as healthy growing clones arising from a culture of mostly inviable cells, continued single colony propagation of both classes of survivors revealed variable growth final products final products patterns: sub-clones could be isolated in which a senescence phenotype re-appeared, although survivors readily reappeared again from these senescing sub- clones. Highly dynamic changes in telomere structure, linear healed plasmid with sequences rapidly lost or acquired at individual Figure 1 Two alternative mechanisms by which break-induced telomeres of survivor strains, was also observed, replication can capture a telomere. (a) A linear plasmid that providing a possible explanation for the growth contains homology to the sub-telomeric Y' repeat can initiate variability. Finally, neither est17 strains early in strand invasion at a chromosomal Y' element. Subsequent senescence nor subsequently isolated survivors displayed replication results in the nonreciprocal translocation of intact Y' and terminal G-rich telomeric repeats onto the end of the linear a genome-wide hyper-recombination phenotype. How- plasmid. (b) Strand invasion to initiate break-induced replication ever, this analysis did not exclude the possibility that in a can occur between a critically short telomere that still retains cell that lacked telomerase, unreplicated telomeres could telomeric repeat sequences and another terminus, thereby themselves become hyper-recombinogenic, a point that elongating the shortened telomere. Staggered alignment of the invading and donor strands (not depicted here) could also result will be re-addressed later in this review. in termini that are longer than chromosome ends maintained by Although all survivors recovered from an est1-D telomerase strain shared common features, such as RAD52- dependence and highly dynamic changes at their telomeres, they could be grouped into two classes telomere replication defect in an easily experimentally based on notable dierences in their telomeric manipulated organism. A haploid yeast strain deleted for structure. One class of survivors was distinguished by the est1-D gene exhibited gradual telomere shortening extensive ampli®cation of sub-telomeric repeats (called and an accompanying progressive decline in viability, Y' elements). In most wild type strains of S. cerevisiae, dubbed yeast cellular senescence. This established an approximately 2/3 of the telomeres contain tandem experimental paradigm for the relationship between arrays of one to four Y' elements, separated by short telomere replication and limits on replicative capacity, stretches of *50 to 100 bp of G-rich telomeric repeats which was recapitulated in human cells several years ago (Figure 2a). In a telomerase-pro®cient strain, exchanges (Bodnar et al., 1998). In addition, the rate of loss of a of Y' elements between telomeres had been shown to non-essential reporter chromosome increased, in parallel occur at low frequency, resulting in the occasional with the appearance of est17 senescence (Lundblad and dispersal of these repeats among chromosome ends Szostak, 1989). Since chromosome loss is the most (Horowitz and Haber, 1985). However, in one class of probable outcome of a DNA double strand break survivors, Y' copy number increased dramatically (Kramer and Haber, 1993; Sandell and Zakian, 1993), (5200-fold in some survivor strains), forming extended this suggested that cell death in an est1-D cell occurs tandem arrays at many telomeres (Figure 2a). In when telomeres shorten suciently that the ends of addition, telomeres that did not have any Y' elements chromosomes resemble double strand breaks. A more in the pre-survivor est17 strain acquired Y' elements. recent study has expanded the potential lethal events that Sub-telomeric repeat ampli®cation was also accompa- occur in an est1-D cell (Hackett et al., 2001). Their nied by a correspondingly substantial increase in the analysis indicates that cell death may result from loss of total amount of telomeric DNA (due to the short

Oncogene Telomere maintenance without telomerase V Lundblad 524 further suggested that conversion to a telomerase- independent survivor was a multi-step process that involved multiple rounds of recombination, occurring during the growth of an est17 culture, with each bene®cial recombination event providing a slight selective advantage. This model implied that there was a gradient of telomere function, such that any reduction in telomere length could reduce the replicative capacity of the est17 culture, and any recombination event that restored function to even a single telomere could contribute to improved growth. Sub- sequent work showed that a strain with a partially defective telomerase, resulting in stably short telomeres but no senescence, had a severe growth disadvantage relative to a telomerase-pro®cient strain, as assessed by liquid competition experiments (Morris and Lundblad, 1997). Alternative models, which propose that critically short telomeres are primarily responsible for triggering senescence and serving as substrates for recombination, are considered later in this review.

Figure 2 Telomeric structure in telomerase-defective survivors. (a) S. cerevisiae chromosome ends consist of telomeres that have one to four copies of the sub-telomeric repeat element Y' (which are bracketed by short stretches of telomeric repeats) as well as telomeres that lack Y' elements. One class of telomerase-defective survivors (type I) is characterized by extensive ampli®cation and dispersal of Y' elements to most telomeres, although the terminal G-rich tract remains short. In a second class of survivors (type II), the G-rich terminal tract is itself elongated by recombination. (b) Survivors of a telomerase defect in K. lactis exhibit extended telomeres, similar to the type II survivors of S. cerevisiae. Note that K. lactis telomeres do not have sub-telomeric repeats bracketed by telomeric repeats

stretches of telomeric tracts bracketing each Y' element), with up to 4% of the genome composed of telomeric repeat DNA in some survivors. However, despite the overall increase in telomeric DNA, the terminal G-rich repeat tract remained short (and did not appear to undergo further shortening) in this class of survivors. This analysis also identi®ed a second class of est17 survivors, with a distinctively dierent telomeric restriction pattern: survivors in this group were characterized by changes in the multiple sizes of telomeric restriction bands, relative to the parental strain. Survivors in this second class did not exhibit extensive sub-telomeric repeat ampli®cation, although many, if not all, telomeres appeared to acquire at least partial Y' sequences. The full Figure 3 Models for recombination-mediated exchanges be- characterization of the nature of the telomeric changes tween telomeres. (a) Recombination between the terminal G-rich tract of a telomere that lacks Y' elements and the internal G-rich occurring in this class of survivors came from subsequent tract of a Y'-containing telomere results in conversion of the non- work by Teng and Zakian (1999), discussed in more Y'-containing telomere to one that now has both sub-telomeric detail in the next section. repeats and additional telomeric G-rich repeats. (b) Integration of Based on the above observations, Lundblad and an extrachromosomal Y' element restores a terminally short Blackburn (1993) proposed that, in the absence of telomere, and also converts it to a Y'-containing telomere. (c) Exchanges between G-rich terminal tracts, analogous to the telomerase, recombination could maintain G-rich nonreciprocal translocation in part A, could also maintain telomeric repeats and restore telomere function. We telomeres

Oncogene Telomere maintenance without telomerase V Lundblad 525 For the class of survivors that displayed Y' lactis telomeres do not have tandem arrays of sub- ampli®cation, Y' elements were proposed to be telomeric repeats interspersed with G-rich telomeric acquired by either a nonreciprocal translocation event repeat sequences (note that, unlike telomeric repeats, (Figure 3a), or by integration of extrachromosomal Y' sub-telomeric repeat elements show little conservation circles into a critically short telomere (Figure 3b). The from species to species). The absence of sub-telomeric latter model was based on the demonstration that Y' repeats meant that a class of survivors analogous to the elements, which contained a weak origin of replication S. cerevisiae survivors that exhibited extensive Y' (Chan and Tye, 1983), could be propagated as ampli®cation was not recovered in K. lactis. Instead, extrachromosomal plasmids (Horowitz and Haber, the terminal G-rich telomeric tract was greatly elongated 1985). Expansion of sub-telomeric Y' elements to in telomerase-defective K. lactis survivors (Figure 2b). multiple telomeres in this class of est17 survivors Dierent telomeres in post-senescent survivors exhibited would have two consequences. First, this would variable lengths, often with tract lengths longer than increase the extent of homology between telomeres, telomeres of a telomerase-pro®cient strain. However, thereby enhancing subsequent recombination events these elongated telomeres still exhibited the same gradual and continuing to drive the process. In addition, the shortening observed during the initial senescence of a ampli®ed arrays of Y'-associated telomeric repeats newly generated telomerase-defective strain, suggesting would increase the amount of G-rich telomeric repeat that the recombination mechanism that was replenishing DNA present at each chromosome, which provides a telomeric repeats was not as ecient as that provided by reservoir of G-rich telomeric repeat DNA for recombi- telomerase. As with S. cerevisiae survivors, K. lactis nation onto the extreme terminus of chromosomes. survivors also displayed an unstable growth phenotype, Recombination between terminal G-rich telomeric manifested as secondary episodes of transient senes- repeats, without Y' ampli®cation, also appeared to be cence. capable of maintaining the terminal G-rich tract These observations led McEachern and Blackburn (Figure 3c). This was inferred from the observation (1996) to present a model for telomere repeat array that even survivors that exhibited extensive ampli®ca- lengthening (Figure 1b) similar to that proposed by tion of Y' elements still possessed individual telomeres Dunn et al. (1984) for the healing of linear plasmids composed only of G-rich repeats, with no newly (Figure 1a). They proposed that the telomere short- acquired Y' element. These G-rich tracts therefore ening that occurs in the absence of telomerase results in appeared to be maintained by telomere ± telomere loss of the essential `capping' function of a telomere, so recombination. Thus, the key feature of the overall that chromosome ends are now perceived as double model was that the acquisition of G-rich telomeric strand breaks. Processing of these exposed termini by repeats through recombination was responsible for DNA repair enzymes could produce a 3' single restoring telomere function and viability. stranded overhang that could invade another chromo- Subsequent work from several laboratories has somal telomere. Extension of this invading strand, shown that the appearance of survivors is not speci®c followed by second strand synthesis, would result in a to yeast strains that are defective for EST1 function. non-reciprocal translocation, with sequences from the Yeast strains lacking any telomerase subunit (Est1p, donor chromosome transferred to the `uncapped' Est2p, Est3p or TLC1), or defective in the telomerase chromosome. These gene conversion events could be recruitment function of Cdc13p, all produce survivors the result of invasion of a 3' strand that still retains in a recombination-dependent manner that is indis- homology with a telomeric repeat tract on another tinguishable from that observed for est17 survivors chromosome; staggered alignment of the invading and (Singer and Gottschling, 1994; Lendvay et al., 1996; donor strands would generate a telomeric tract longer Teng and Zakian, 1999). than that observed in telomerase-plus cells. Although the primary structural change observed in K. lactis survivors was extension of the telomeric tract Recombination-mediated telomere maintenance can array, gene conversion events in sub-telomeric regions elongate terminal telomeric G-rich repeats of K. lactis survivors were also observed. Since 11 of the 12 K. lactis chromosome termini share homology in Analysis of the consequences of a telomerase defect in a the sub-telomeric region, this indicates that strand related budding yeast, Kluveromyces lactis, revealed invasions could initiate well into sub-telomeric regions. additional insights regarding how recombination could However, McEachern and Blackburn pointed out that maintain telomeres. In K. lactis, loss of telomerase the majority of the recombination events that they function had the same consequences as in S. cerevisiae: observed must depend on retention of at least a few gradual attrition of telomeric repeats, accompanied by a telomeric repeats at the time that chromosome termini senescence phenotype (McEachern and Blackburn, undergo recombination-mediated elongation. 1995). Survivors could also be recovered from senescent Through a careful molecular analysis, Teng and cultures of telomerase-de®cient K. lactis strains, through Zakian (1999) showed that the second class of S. a mechanism that was again dependent on RAD52 cerevisiae survivors identi®ed by Lundblad and Black- function (McEachern and Blackburn, 1996). However, burn (1993) exhibited many of the properties displayed analysis of telomere structure in these survivors un- by survivors recovered from K. lactis. Telomeres in this covered some notable dierences. Unlike S. cerevisiae, K. second class of S. cerevisiae survivors, which they re-

Oncogene Telomere maintenance without telomerase V Lundblad 526 named type II survivors (while survivors with extensive senescence is observed (McEachern and Iyer, 2001; Y' ampli®cation were labeled type I survivors), were Rizki and Lundblad, 2001). However, additional shown to terminate with very long, heterogeneous observations indicate critically short telomeres may tracts of G-rich telomeric repeats (Figure 2a). As had also be preferred substrates for at least one type of been previously observed in K. lactis, these extended recombination (Teng et al., 2000). Thus, both varia- telomeric tracts were not stably maintained: they tions of this model are likely to contribute to displayed the slow attrition of a telomerase-defective telomerase-independent telomere maintenance. strain, as well as more abrupt changes that were An initial clue that recombination-mediated ex- reminiscent of previous observations showing that changes were important for cell survival even early in elongated telomeric tracts could undergo rapid trunca- the propagation of a telomerase-de®cient strain came tion (Li and Lustig, 1996; Bucholc et al., 2001). Thus, from observations of the consequence of loss of although type I and type II survivors both retain RAD52 function. A rad527 mutation not only blocked terminal G-rich repeat tracts, terminal sequences are the appearance of survivors, but in addition, an est1-D greatly elongated in type II survivors. Notably, the rad52-D strain also displayed a greatly accelerated short tracts that are retained on the telomeres of type I senescence phenotype (Lundblad and Blackburn, 1993). survivors are apparently more stable than the con- Whereas an est1-D strain at the 25 generation time tinually shortening type II telomeric tracts (although it point was indistinguishable from an EST1 strain, an cannot be ruled out that additional shortening of a est1-D rad52-D strain displayed an obvious growth type I telomere might be lethal and thus would not defect by 25 generations and was completely inviable contribute to the signal on a telomeric Southern blot). by *50 generations (depicted schematically in Figure Teng and Zakian also studied in detail the growth 4). This strongly suggested that recombination was properties of the two classes of survivors. In their strain contributing to telomere function even in a newly background, all type II survivors exhibited a stable generated telomerase-defective strain, during the initial growth phenotype, and when propagated in culture, type period of telomere shortening. II survivors also had a selective advantage over type I As a reciprocal test of this premise, Rizki and survivors. The growth disadvantage of type I survivors Lundblad (2001) examined the consequences of genetic might be due ± at least in part ± to the burden of alterations that increased the frequency of genetic replicating a genome increased by up to 10%, due to the exchanges between telomeric repeat DNA. This work extensive ampli®cation of Y' and telomeric repeat showed that defects in mismatch repair ± which relieve sequences. However, type I survivors also showed a high a block to recombination between imperfectly matched propensity for conversion to type II survivors, suggesting sequences ± enhanced the proliferative capacity of that telomeres in the absence of telomerase may be telomerase-defective strains in both S. cerevisiae and K. substrates for several dierent types of recombination lactis. Using liquid competition assays, the eects of a events. The eventual outcome may be dictated by a mismatch repair mutation were detected during the balance between the susceptibility of an under-replicated very early stages of growth of a telomerase-defective telomere to dierent types of recombination events, yeast strain, even before a senescence phenotype was versus the selective advantage of dierent recombinant evident (Figure 4). Increased survival rates were shown telomeric structures. In other words, telomeres in budding to be due to elimination of the anti-recombination yeast may be more recombinogenic for the types of events activity of the mismatch repair pathway, rather than an that result in type I survivors, but the selective growth indirect eect of increased mutation rates. Notably, advantages of type II survivors allows this class of enhancement of survival of telomerase-defective K. survivors to dominate. This may have implications for the analysis of potential intermediates for similar telomere maintenance pathways in human cells.

Do critically short telomeres promote the recombination-dependent pathway?

The original model by Lundblad and Blackburn proposed a stochastic process of recombination events at telomeres, with selection pressure for viability ensuring that there would be a cumulative acquisition of telomeric sequences. A variant of this model is that telomeres become more recombinogenic in the absence of telomerase. A further extension of the model is that recombination occurs in a burst, triggered by critically Figure 4 Schematic representation of the relative proliferative short telomeres, rather than a succession of cumulative capacity of a telomerase-defective strain (est27), a telomerase- exchanges. Recent studies from several dierent defective strain that is also eliminated for the major pathway for recombination (est27 rad527) or a telomerase-defective strain in laboratories suggest that short telomeres are in fact which homeologous recombination between chromosome ends is recombinogenic, even before critically shortening or increased, due to a defect in mismatch repair (est27 msh27)

Oncogene Telomere maintenance without telomerase V Lundblad 527 lactis cells was also dependent on elimination of study by McEachern and Iyer (2001), greatly elongated mismatch repair. Since K. lactis telomeres are com- telomeric tracts were not observed in short telomere posed of perfect telomeric G-rich repeats, this strongly strains with partial loss of telomerase function, despite suggested that recombination between partially mis- the increased telomeric gene conversion rate. This matched sub-telomeric regions could contribute to contrasts with the greatly elongated telomeric tracts survival in the absence of telomerase. displayed by survivors of a telomerase null mutation More direct support for the potential contribution of (McEachern and Blackburn, 1996). However, one recombination between sub-telomeric repeats to telo- alternative to a graded recombination response is that merase-independent survival came from studies by partially uncapped telomeres are healed by a dierent McEachern and Iyer (2001). Recombination between recombination mechanism than that utilized by critically telomeres was monitored by following phenotypically short telomeres. silent changes in the telomeric repeat tract or the URA3 gene placed in the sub-telomeric region of one telomere. In a telomerase-defective strain, recombina- A summary of potential molecular mechanisms tion-mediated exchanges were greatly elevated, by up to 1000-fold. Notably, the URA3 gene exhibited rapid Work from several dierent labs in the last several years dispersal to most, or even all, of the 12 telomeres, has expanded our understanding of the genetic require- reminiscent of the ampli®cation of Y' elements ments for telomerase-independent telomere maintenance observed in S. cerevisiae survivors (although note that in yeast. The ®rst step in this direction came from in these experiments, URA3 was not bracketed by collaborative work from the Greider and Haber telomeric repeats). Even in cells in which telomerase laboratories that investigated the eects of a large set was still partially active (i.e. telomeres were shortened, of DNA recombination mutations on the ability of a but not suciently to confer a senescence phenotype), tlc17 strain to form survivors (Le et al., 1999). Their gene conversion rates were increased by 100 to 200- results demonstrated that at least two dierent recombi- fold, again arguing that recombination at telomeres nation pathways, de®ned by RAD50 and RAD51, were can be substantially enhanced before chromosome ends involved in telomerase-independent survival. Both become terminally short. The authors proposed that rad507 tlc17 and rad517 tlc17 strains gave rise to the enhanced recombination rates of stably short survivors, but a triply mutant rad507 rad517 tlc17 strain telomere mutants re¯ects a weakened ability of a short was as defective as a rad527 tlc17 strain at generating telomere to provide a telomeric `cap' that protects post-senescent survivors. These two pathways correlated chromosome ends from DNA repair activities that with the two classes of survivors recovered from S. normally act on double strand breaks. cerevisiae: type I survivors required RAD51 function However, a strikingly dierent conclusion was reached whereas type II survivors were dependent on RAD50 when the telomeric changes occurring during the function (Teng et al., 2000; Chen et al., 2001). However, appearance of type II survivors were analysed in detail the correlation between recombination pathway and (Teng et al., 2000). Monitoring of telomere length at survivor class may not be absolute, as type II survivors multiple time points during the outgrowth of a tlc17 could form in a rad507 mutant background, at least at strain in culture showed that telomeres gradually low frequency, but could not be stably maintained (Chen shortened during the pre-survivor period, as previously et al., 2001). Additional analysis has shown that other observed for a telomerase defect, and then abruptly members of the RAD51 epistasis group (RAD54, RAD55 lengthened; this correlated with a sudden switch to an and RAD57) were also required for the appearance of improved growth rate of the culture. When an type I survivors (Le et al., 1999; Chen et al., 2001). Based individually marked telomere in an already established both on information about genetic requirements and type II survivor was monitored, the previously observed observations of the structural alterations that occur at continuous shortening of type II telomeres was inter- telomeres, four general models have been proposed for spersed with abrupt elongation events, increasing in size how recombination can maintain telomeres. The next in 1 to 2 kb segments of presumably telomeric repeat four sections discuss these models. tract DNA. Based on these observations, Zakian and colleagues argued against an incremental process of telomere lengthening and instead proposed that a one- Break-induced replication (BIR) step event was responsible. They speculated that the substrate for this event could be extra-chromosomal BIR is a one-ended recombination event that initiates a telomeric repeat DNA, generated by intra-chromatid replication fork (Figure 1). Strand invasion of an intact recombination. chromosome by the broken end (via regions of One simple way to reconcile these two sets of homology shared between the reporter and donor observations is to consider a model in which there is a chromosomes) is followed by replicative copying of graded increase in recombination that correlates with the donor sequences. If replication continues to the end of severity of the telomeric length defect. In other words, the donor chromosome, this will result in the shortened telomeres exhibit enhanced susceptibility to nonreciprocal transfer of chromosomal DNA ± as recombination-mediated exchanges, but critically short well as terminal telomeric DNA sequences ± to the site telomeres are even more recombinogenic. Notably, in the of the double strand break. Thus, through BIR, a

Oncogene Telomere maintenance without telomerase V Lundblad 528 broken chromosome can be healed by capturing a Integration of extrachromosomal DNA telomere from the donor (Bosco and Haber, 1998). Note that a key feature of this replicative mechanism is Integration of extra chromosomal DNA could provide an that the broken chromosome is not repaired at the alternative mechanism for sudden alterations in the size of expense of the donor chromosome, as would be the telomeric and sub-telomeric tracts, via excision and case if telomere capture occurred as the result of a reintegration of chromosomal DNA, as an interchromo- nonreciprocal crossover. For a more complete discus- somal recombination process (Figure 3b). If the excised sion of BIR, see Kraus et al. (2001). DNA also contains an origin of replication, as has been Direct demonstration that BIR can be utilized as a demonstrated for Y' elements (Chan and Tye, 1983), this response to a telomerase defect has come from the provides a rapid means of expansion and dispersal of both recent work by Greider and colleagues (Hackett et al., sub-telomeric and telomeric DNA. This may provide 2001). Their analysis of the molecular events occurring either an alternative, or even an additional, route to the in late generation est17 cultures showed that not only production or maintenance of type I survivors, in telomeric repeat DNA was lost: much more substantial addition to the BIR pathway discussed above. Reintegra- terminal deletions, extending for many kilobases, also tion by recombination between telomeric DNA present occurred. If left unrepaired, these extensive terminal on the Y' circle and terminal telomeric tracts on the deletions would presumably lead to lethality, either due recipient chromosome, as depicted in Figure 1b, would to the loss of essential genes or deleterious eects on also provide a means of replenishing the chromosome chromosome transmission. However, these deletions terminus. Notably, in an S. cerevisiae strain that has few could be healed by break-induced replication, via Y' elements, with no tandem arrays or bracketing internal exchanges with other chromosomes through short telomeric repeat tracts, type I survivors are rare (Huang et stretches of homology shared by the donor and al., 2001), possibly due to the inability to generate terminally deleted chromosomes. This indicates that a extrachromosomal Y' circles. primary mechanism for the generation of telomerase- Whether excised DNA consisting of simple telomeric independent survivors relies on BIR. Although the BIR DNA is also capable of participating in recombination- events recovered by Hackett et al. (2001) involved mediated reintegration events is unknown. However, exchanges between non-homologous chromosomes Lustig and colleagues (Bucholc et al., 2001) have (because of their experimental design, which dictated proposed a model to explain rapid shortening of what they recovered), healing by BIR can also occur elongated telomeric repeat tracts, termed TRD (Telo- through intrachromosomal translocations (Bosco and meric Rapid Deletion), such that a potential product of Haber, 1998) or unequal sister chromatid exchange. this model is excised telomeric DNA. Since rapid Multiple studies by the Haber laboratory, using HO- truncations are observed in type II survivors (Teng and induced double strand breaks, have de®ned the genetic Zakian, 1999), this argues that TRD is probably requirements for BIR, as well as enhancer sites that operative in type II survivors strains. If reintegration of can facilitate BIR events (Signon et al., 2001; Malkova telomeric DNA can contribute to survivor formation, et al., 2001; Kraus et al., 2001). These studies have this may be a mechanism employed by type II survivors, demonstrated that a RAD51-independent pathway can possibly as a maintenance mechanism of this survivor promote BIR, using non-random sites to initiate the state once formed. Intriguingly, human ALT cells, which process. However, RAD51-dependent pathway(s) for appear to maintain telomeres by a non-telomerase BIR are also available (discussed in Kraus et al., 2001). mechanism, contain a novel variant of promyelocytic This suggests that survivors that exhibit extensive ampli®cation of sub-telomeric elements (type I) proceed by a RAD51-dependent BIR pathway (Figure 1a), whereas survivors that expand the terminal telomeric tract (type II) are promoted by a RAD51- independent, RAD50-dependent process (Figure 1b). Thus, both classes of survivors may establish and/or maintain their telomeres through BIR. However, one genetic observation about the requirements for BIR that does not correlate well with the requirements of the survivor pathway is the role of the Sgs1p helicase. Although Sgs1p is not required for conventional BIR (Signon et al., 2001), the appearance of type II survivors is eliminated in telomerase- defective strains that also lack SGS1 (Huang et al., 2001; Cohen and Sinclair, 2001; Johnson et al., 2001). One possible explanation, proposed by Louis and colleagues (Huang,etal. 2001) for the telomere-speci®c SGS1 requirement may be due to a need to disrupt G- Figure 5 Rolling circle replication (a) or t-loop mediated quartets and other secondary structures that single- extension (b) provide two mechanisms for rapid expansion of stranded G-rich telomeric DNA can form. the G-rich tract of a telomere

Oncogene Telomere maintenance without telomerase V Lundblad 529 leukemia (PML) bodies that is composed of telomeric addition, such structures potentially regulate elongation DNA, as well as telomere binding proteins and proteins by telomerase, since a sequestered 3' terminus would not involved in DNA repair and replication (including be accessible to telomerase. The 3' terminus of the Rad51p and Rad52p) (Yeager et al., 1999). However, invading G-rich telomeric strand could also potentially whether ALT-associated PML bodies are an obligatory initiate DNA replication, thereby providing a telomerase- intermediate in either the establishment or maintenance independent means for elongation of this terminus. steps of the ALT pathway, rather than a byproduct of Formation of t-loops does in fact resemble a recombina- elongated telomeres, has not yet been resolved. tion intermediate, and bears substantial resemblance to the interchromosomal BIR process depicted in Figure 1b. Furthermore, Rad50p (along with other components of Rolling circle replication the Rad50/Mre11/Nbs1 protein complex) have been invoked as players in either the formation and/or Rolling circle DNA replication has received increasing maintenance of t-loop structures (Zhu et al., 2000), attention as a potential mechanism for rapid expansion suggesting that t-loops might be the initiating structure of the G-rich telomeric tract (Figure 5a). Invasion of a for type II survivor formation. However, although the telomeric tract by an extrachromosomal telomeric structure of a t-loop is certainly suggestive of a replicative circle could potentially establish a replication fork that model for extension of the chromosome terminus, this initiates rolling circle replication, thereby allowing presents somewhat of a paradox: t-loops have been extensive elongation of the telomeric tract. Such a proposed as a mechanism to provide chromosome end model is particularly appealing when considering protection (Grith et al., 1999), whereas recombination- mechanisms that would promote the abrupt elongation mediated telomere maintenance is thought to occur in observed during the initial appearance of type II response to a loss of end protection. survivors. As pointed out by Teng et al. (2000), Although t-loops have not been identi®ed so far in conventional gene conversion events between critically yeast, detection of these telomeric structures in short telomeres does not provide an easy explanation evolutionary unrelated organisms such as mammals, for the sudden elongation observed at the termini of ciliates and trypanosomes suggests that this is a type II telomeres, whereas rolling circle replication conserved feature of eukaryotic telomeres. However, could provide one potential route for rapid elongation. although trypanosome t-loops are small (*1 kb) Rolling circle replication has been well characterized relative to the average size of mammalian t-loops, this as a replicative mechanism used by small genomes such still substantially exceeds the size of telomeric tracts in as single-stranded bacteriophages and bacterial plas- yeast, particularly the even shorter tracts that occur mids (for a review, see Novick, 1998). These replicons when telomerase is lacking. Therefore, a potential role depend on speci®c initiator proteins which both melt for t-loops in rescuing yeast senescence and promoting open and nick the initiator region, thereby providing a survivors remains speculative at this time. 3' OH that can serve as a primer for the initiation of DNA synthesis. If rolling circle replication does in fact occur at telomeres, the 3' terminus of the G-rich strand Potential parallels to an ALT (Alternative Telomere of the telomere could provide priming activity. Lengthening) pathway in human cells However, this model leaves open many questions, such as what activities are required to initiate the process As mentioned at the start of this review, interest in the (which involves melting of both the duplex telomeric yeast survivor pathway has been enhanced by the tract and the invading duplex extrachromosomal DNA, possibility that recombination may contribute to telo- regulation of assembly of the replisome, etc.). In mere maintenance in human cells as well. If so, this may addition, proving that recombination-mediated elonga- have particular signi®cance to processes that promote tion is templated by a single extrachromosomal neoplasia. In fact, although most immortalized human telomeric DNA circle (Figure 5a), as opposed to cells, including tumor cells, have substantially up- integration of multiple extrachromosomal circles regulated telomerase (Shay and Bacchetti, 1997), a subset (Figure 3b), will be experimentally challenging. of immortalized cells do not display high levels of telomerase activity (Bryan et al., 1995, 1997). Further- more, telomeres in these cells show a strikingly dierent Elongation of a t-loop pro®le, displaying heterogeneous telomeres that range from very short to lengths that are elongated by up to t-loops provide an alternative to the rolling circle 50 kb. The absence of detectable telomerase enzyme replication model (Figure 5b). t-loops are intrachromo- activity, coupled with the striking telomeric pro®le, has somal structures that are formed when the 3' single-strand led to considerable speculation that ALT may proceed by G-rich extension at the end of the chromosome invades the same mechanism(s) that promote telomere elongation the duplex telomeric region (Grith et al., 1999; Munoz- in K. lactis survivors or type II survivors in S. cerevisiae. Jordan et al., 2001). These structures have been proposed The speci®c genetic requirements that promote the to provide an architectural means of protecting chromo- ALT pathway have not yet been elucidated, due to the some ends from degradative and/or DNA repair activ- diculties of genetic analysis in human cells, nor has an ities, by masking the telomeric single strand terminus. In inducible system for ALT, where the intermediate steps

Oncogene Telomere maintenance without telomerase V Lundblad 530 in the pathway can be analysed, been established. In Observations from yeast may provide some insight into addition, the steady attrition of telomeric DNA observed the potential complexities of these experiments. The in K. lactis and type II S. cerevisiae survivors have not genomic alterations that occur in telomerase-defective been observed. However, numerous studies have shown survivors can be very stable, even when telomerase is that human chromosome termini are subject to enhanced restored. Introduction of EST1 back into an est1-D levels of recombination, as monitored by sequence survivor that exhibits extensive Y' ampli®cation results in polymorphisms in sub-telomeric regions (see Wilkie et elongation of the terminal telomeric tract back to wild al., 1991; Baird et al., 2000 for several examples). type length, as well as a substantial reduction in Y' copy Additional work indicates that terminal deletions found number (Lundblad and Blackburn, 1993; Teng and on human chromosomes can be healed by capture of sub- Zakian, 1999). However, even with extensive propaga- telomeric and telomeric sequences from another chro- tion, the distribution of Y' elements does not return to its mosome (Meltzer et al., 1993). A study by Reddel's starting point: each telomere retains at least one Y' laboratory (Dunham et al., 2000) was designed to ask element in an EST1 strain that has a previous history as a speci®cally whether ALT telomeres participated in telomerase-defective survivor (V Lundblad, unpublished recombination, by monitoring the transmission of a results). When this strain is converted back to est1-D,it selectable marker inserted into telomeric sequences in again exhibits telomere shortening and senescence, but in either ALT or telomerase-plus immortalized cells. This this newly senescent strain, survivors appear more readily marker spread to other chromosome ends much more and at earlier time points. This suggests that the retained frequently in ALT cells than in telomerase-plus cells, single Y' elements prime the re-appearance of survivors. suggesting that the telomeres in ALT cells (or a subset, Similarly, following reintroduction of telomerase into a such as the shortest telomeres?) were more recombino- type II survivor, telomeres return to a roughly wild type genic than telomeres maintained by telomerase. Notably, telomeric restriction pattern only very slowly (Teng and when the same marker was placed in a sub-telomeric Zakian, 1999). This suggest that even if telomeres are no location, dispersal to other chromosomes was not longer highly recombinogenic after the reintroduction of observed, in contrast to the results obtained in K. lactis telomerase, the stable extended telomeric tracts still decay by McEachern and Iyer (2001). only very slowly. It is also worth noting that recombination may Extending these observations to human cells, this contribute to telomere maintenance and potentially suggests that using telomere length as a monitor of tumor formation, even without conversion to the full whether a recombination-promoted pathway for telo- ALT pathway. In particular, any genetic alteration that mere maintenance is still operating in the presence of enhances recombination rates at telomeres ± such as telomerase may not be ideal. This monitors the product, defects in mismatch repair ± may partially relieve the rather than the presence of an active mechanism. The telomere maintenance barrier to proliferation, thereby solution may be to develop assays that directly monitor contributing to the mechanism of oncogenesis. the frequency of recombination events at telomeres. One such assay, using a reporter gene that marks a telomeric tract, has been described above (Dunham et al., 2000). An Can the ALT pathway and telomerase-mediated telomere alternative approach relies on the presence of natural maintenance co-exist in the same cell? variants in the telomeric repeat tract. Human telomerase synthesizes TTAGGG repeats onto chromosome termini A number of recent papers have been directed at the (Morin, 1989). However, the proximal end of this repeat above question, by examining the consequences of tract contains variant telomeric repeats such as reconstitution of telomerase activity in ALT cells as a TGAGGG and TCAGGG, which can be monitored by result of ectopic expression of telomerase subunits a PCR-based assay (Baird et al., 1995). Since the (Perrem et al., 2001; Cerone et al., 2001; Ford et al., distribution of these variant repeats is highly variable, 2001; Grobelny et al., 2001) or by fusion of ALT cells this assay may provide a semi-quantitative means of with telomerase-positive tumor cell lines (Perrem,etal. monitoring the extent of recombination at human 2001). However, the conclusions from these four chromosome ends. groups were somewhat contradictory. Two groups reported that when telomerase was ectopically expressed, short telomeres were preferentially elongated, Summary and future possibilities but the ALT pathway was not repressed (Grobelny et al., 2001; Cerone et al., 2001). In contrast, Shay and Experiments using budding yeasts have provided both Wright and colleagues, using the same ectopic speci®c models and genetic requirements that dictate the expression approach, observed rapid reduction of the frequency of recombination-dependent exchanges be- longest telomeres and concluded that ALT was tween telomeres in cells that lack telomerase. These inhibited (Ford et al., 2001). Reddel and co-workers observations have implicated speci®c recombination did not observe inhibition of ALT when telomerase proteins and also highlighted other alterations, such as was ectopically expressed, but when ALT cells were defects in mismatch repair, that may similarly in¯uence fused with telomerase-positive tumor cells, telomeres the survival of human cells that fail to express telomerase. showed rapid initial decreases in length, followed by Additional studies, not discussed in this review, suggest more gradual shortening (Perrem et al., 2001). that alterations in components of telomeric chromatin

Oncogene Telomere maintenance without telomerase V Lundblad 531 can also in¯uence either the types of recombination events Acknowledgments that occur at the telomere (Teng et al., 2000) or the I thank Jim Haber and Jenny Hackett for explaining the frequency with which the survival pathway appears subtleties of the genetic requirements of BIR to me, Sara (Chandra and Lundblad, unpublished observations). Evans for excellent editorial assistance and ®gure prepara- All of these examples in yeast provide a possible roadmap tion, and Lou Zumstein, Rachel Cervantes and Erin Pennock for critical reading of the manuscript. Work in for experiments that may help elucidate the ALT the author's laboratory is supported by grants from the pathway. NIH and the Ellison Medical Foundation.

References

Baird DM, Coleman J, Rosser ZH and Royle NJ. (2000). Le S, Moore JK, Haber JE and Greider CW. (1999). Am.J.Hum.Genet.,66, 235 ± 250. Genetics, 152, 143 ± 152. Baird DM, Jereys AJ and Royle NJ. (1995). EMBO J., 14, Lendvay TS, Morris DK, Sah J, Balasubramanian B and 5433 ± 5443. Lundblad V. (1996). Genetics, 144, 1399 ± 1412. Bernards A, Michels PA, Lincke CR and Borst P. (1983). Li B and Lustig AJ. (1996). Genes Dev., 10, 1310 ± 1326. Nature, 303, 592 ± 597. Luder A and Mosig G. (1982). Proc. Natl. Acad. Sci. USA, Biessmann H and Mason JM. (1997). Chromosoma, 106, 63 ± 79, 1101 ± 1105. 69. Lundblad V. (2000). Mutat. Res., 451, 227 ± 240. Blackburn EH. (2001). Cell, 106, 661 ± 673. Lundblad V and Blackburn EH. (1993). Cell, 73, 347 ± 360. Bodnar AG, Ouellette M, Frolkis M, Holt SE, Chiu C-P, Lundblad V and Szostak JW. (1989). Cell, 57, 633 ± 643. Morin GB, Harley CB, Shay JW, Lichtsteiner S and MalkovaA,SignonL,SchaeferCB,NaylorML,TheisJF, Wright WE. (1998). Science, 279, 349 ± 352. Newlon CS and Haber JE. (2001). Genes Dev., 15, 1055 ± Bosco G and Haber JE. (1998). Genetics, 150, 1037 ± 1047. 1060. Bryan TM, Englezou A, Dalla-Pozza L, Dunham MA and McEachern MJ and Blackburn EH. (1995). Nature, 376, Reddel RR. (1997). Nat. Med., 3, 1271 ± 1274. 403 ± 409. Bryan TM, Englezou A, Gupta J, Bacchetti S and Reddel McEachern MJ and Blackburn EH. (1996). Genes Dev., 10, RR. (1995). EMBO J., 14, 4240 ± 4248. 1822 ± 1834. Bucholc M, Park Y and Lustig AJ. (2001). Mol. Cell. Biol., McEachern MJ and Iyer S. (2001). Mol. Cell., 7, 695 ± 704. 21, 6559 ± 6573. McEachern MJ, Krauskopf A and Blackburn EH. (2000). Cerone MA, Londono-Vallejo JA and Bacchetti S. (2001). Annu. Rev. Genet., 34, 331 ± 358. Hum. Mol. Genet., 10, 1945 ± 1952. Meltzer PS, Guan XY and Trent JM. (1993). Nature Genet., Chan CS and Tye BK. (1983). J. Mol. Biol., 168, 505 ± 523. 4, 252 ± 255. Chen Q, Ijpma A and Greider CW. (2001). Mol. Cell. Biol., Morin GB. (1989). Cell, 59, 521 ± 529. 21, 1819 ± 1827. Morris DK and Lundblad V. (1997). Curr. Biol., 7, 969 ± 976. Cohen H and Sinclair DA. (2001). Proc. Natl. Acad. Sci. Munoz-Jordan JL, Cross GA, de Lange T and Grith JD. USA, 98, 3174 ± 3179. (2001). EMBO J., 20, 579 ± 588. Dunham MA, Neumann AA, Fasching CL and Reddel RR. Novick RP. (1998). Trends Biochem. Sci., 23, 434 ± 438. (2000). Nat. Genet., 26, 447 ± 450. Perrem K, Colgin LM, Neumann AA, Yeager TR and Dunn B, Szauter P, Pardue ML and Szostak JW. (1984). Cell, Reddel RR. (2001). Mol. Cell. Biol., 21, 3862 ± 3875. 39, 191 ± 201. Reddel RR, Bryan TM, Colgin LM, Perrem KT and Yeager Ford LP, Zou Y, Pongracz K, Gryaznov SM, Shay JW and TR. (2001). Radiat. Res., 155, 194 ± 200. Wright WE. (2001). J. Biol. Chem., 276, 32198 ± 32203. Rizki A and Lundblad V. (2001). Nature, 411, 713 ± 716. Formosa T and Alberts BM. (1986). J. Biol. Chem., 261, Sandell LL and Zakian VA. (1993). Cell, 75, 729 ± 739. 6107 ± 6118. Shay JW and Bacchetti S. (1997). Eur. J. Cancer, 33, 787 ± Greider CW. (1996). Annu. Rev. Biochem., 65, 337 ± 365. 791. GrithJD,ComeauL,Rosen®eldS,StanselRM,BianchiA, Signon L, Malkova A, Naylor ML, Klein H and Haber JE. Moss H and de Lange T. (1999). Cell, 97, 503 ± 514. (2001). Mol. Cell. Biol., 21, 2048 ± 2056. Grobelny JV, Kulp-McEliece M and Broccoli D. (2001). Singer MS and Gottschling DE. (1994). Science, 266, 404 ± Hum. Mol. Genet., 10, 1953 ± 1961. 409. Hackett JA, Feldser DM and Greider CW. (2001). Cell, 106, Teng CS, Chang J, McCowan B and Zakian AV. (2000). Mol. 275 ± 286. Cell., 6, 947 ± 952. Horowitz H and Haber JE. (1985). Mol. Cell. Biol., 5, 2369 ± Teng SC and Zakian VA. (1999). Mol. Cell. Biol., 19, 8083 ± 2380. 8093. Huang P, Pryde FE, Lester D, Maddison RL, Borts RH, Walmsley RW, Chan CS, Tye BK and Petes TD. (1984). Hickson ID and Louis EJ. (2001). Curr. Biol., 11, 125 ± 129. Nature, 310, 157 ± 160. Hughes TR, Evans SK, Weilbaecher RG and Lundblad V. Wilkie AO, Higgs DR, Rack KA, Buckle VJ, Spurr NK, (2000). Curr. Biol., 10, 809 ± 812. Fischel-Ghodsian N, Ceccherini I, Brown WR and Harris Johnson FB, Marciniak RA, McVey M, Stewart SA, Hahn PC. (1991). Cell, 64, 595 ± 606. WC and Guarente L. (2001). EMBO J., 20, 905 ± 913. Yeager TR, Neumann AA, Englezou A, Huschtscha LI, Kass-Eisler A and Greider CW. (2000). Trends Biochem. Sci., Noble JR and Reddel RR. (1999). Cancer Res., 59, 4175 ± 25, 200 ± 204. 4179. Kramer KM and Haber JE. (1993). Genes Dev., 7, 2345 ± 2356. Zhu XD, Kuster B, Mann M, Petrini JH and Lange T. Kraus E, Leung WY and Haber JE. (2001). Proc. Natl. Acad. (2000). Nat. Genet., 25, 347 ± 352. Sci. USA, 98, 8255 ± 8262.

Oncogene