The ends of linear eukaryotic chromosomes, the telomeres, fulfill unique and essential functions in genome integrity. First and above all, they represent the ends of the DNA molecules. In general, non-telomeric double-strand DNA breaks (DSBs) are not tolerated and are rapidly repaired.⇓
This scenario, however, does not apply to telomeres (Muller, 1938; McClintock, 1939). Indeed, telomeric chromosome ends are bound by a complex cast of factors that protect chromosome ends from degradation or fusion (d'Adda di Fagagna et al., 2004). Second, owing to the semi-conservative synthesis of DNA, a DNA end cannot be completely duplicated (Watson, 1972; Olovnikov, 1973). To avoid continuous sequence loss from the telomeres in dividing cells, special mechanisms have evolved. In most eukaryotic organisms, the solution to the problem involves a ribonuclear complex, called telomerase, that is minimally composed of a reverse transcriptase catalytic protein subunit and an RNA subunit that is used as the template (Greider and Blackburn, 1987). In order to accommodate these two functions, the overall telomere structure is dynamic and undergoes dramatic changes during the cell cycle particularly in S phase (Blackburn, 2001; Vega et al., 2003; Smogorzewska and de Lange, 2004). Telomere structure and function have been most intensely studied in model systems such as budding yeast and in human cell lines. Therefore, we chose these organisms to illustrate similarities and differences in telomere composition and structure.
Telomeric DNA structures are highly conserved amongst eukaryotes. In virtually all organisms, chromosomes end with an array of short direct repeats. For example, human telomeres comprise 2-20 kb of (TTAGGG)n repeats. The species-specific repeats are related to each other, each being composed of a G-rich and a C-rich strand. Moreover, the directionality of the repeats is conserved such that the 3′ end of the chromosome is on the G-rich strand. In addition, the G-rich strand is longer than the C-rich strand and forms a single-strand extension at the very ends of the chromosomes (also called G-tails). For species that have relatively long G-tails, such as humans, there is evidence that the terminal G-tails invade and displace G-strands in the double-stranded portion of the telomere, yielding a configuration that has been dubbed the t-loop (Griffith et al., 1999). Budding yeast telomeres, composed of irregular (TG)1-6TG2-3-repeats, are about 300 bp long and their G-tails normally are short (Larrivee et al., 2004). However, this terminal DNA arrangement is dynamic and changes during the cell cycle: longer G-tails can be detected on yeast telomeres during S phase, when telomeres are replicated (Wellinger et al., 1993).
With few exceptions, the telomeric repeat tract lengths are not precise but rather centered around a genetically determined mean. Thus, in dividing cell populations, the actual tract length at individual telomeres can fluctuate around the average mean size. Importantly, an adequate, but not precisely determined minimal tract of telomeric repeats is essential for functional telomeres in all species. In certain plant and insect species, much longer (≥1 kb) and complex repeats are found at chromosome ends. Although not discussed in detail here, the mechanisms of telomeric DNA maintenance in these latter organisms clearly are related to those active in organisms that have the shorter repeats (Pardue and DeBaryshe, 2003).
Telomere-associated proteins identified in various experimental systems clearly serve similar functions. For example, although not orthologues, the primary double-strand repeat binding proteins TRF1 and TRF2 of humans and the major telomere binding protein of budding yeast, Rap1p, share similar DNA-binding motifs and serve as binding platforms for a number of additional telomere-associated proteins. Moreover, the human orthologue of this yeast Rap1p protein, hRAP1, associates with telomeres through an interaction with TRF2. The major G-tail-binding proteins (Cdc13p in budding yeast, POT1 in mammals) also contain related protein motifs and are viewed as functional equivalents.
The first major function of the telomeric complex is to protect the chromosome ends from degradation. Furthermore, telomeres prevent chromosomal DNA ends from being recognized as DNA damage. In yeast, Cdc13p is essential for these functions, since after its experimentally induced inactivation, telomeric DNA is degraded (Garvik et al., 1995). Although an initial report suggests a similar role for POT1 in humans (Veldman et al., 2004), the phenotypes induced by a complete absence of this protein are not known yet. Therefore, proteins associating with the G-tails appear essential for preventing enzymes from degrading telomeric DNA or processing the DNA ends such that they would be suitable for DSB repair. Proteins associated with the double-strand portion of the repeats are also important for this end-protection function. In particular, the human TRF2 protein is crucial for maintaining a normal chromosomal end structure because it collaborates in the processes leading to t-loops (van Steensel et al., 1998). However, it is less clear how telomeric proteins physically protect the DNA from DNA-processing events, such as those brought about by stochastic nuclease attack. In certain ciliates, the only system where this has been addressed directly, the binding of Pot1-related proteins to telomeres can protect the DNA from degradation, at least in vitro (Gottschling and Zakian, 1986). We know even less about how the proteins prevent telomeres from being detected as DNA damage. This issue is further compounded by the paradoxical finding that proteins that are involved in sensing and processing damaged DNA are also intimately associated with telomeric chromatin. For example, orthologues of the Ku complex, ATM/ATR proteins and the Mre11 complex clearly localize to telomeres and/or have specific functions in maintaining functional telomeres in yeasts and humans (d'Adda di Fagagna et al., 2004). However, there is increasing evidence that these proteins may act on telomeres primarily during S phase, when telomeric DNA is replicated (see below). Their functions outside S phase are poorly understood.
Outside S phase, the telomere complex adopts a fold-back structure that is thought to be organized in a specialized chromatin structure through associations between telomeric and subtelomeric binding proteins. The result is a tight chromatin domain that extends into nucleosomal DNA and represses transcriptional activity in telomere-proximal areas, an effect that has been dubbed telomere position effect (TPE) (Gottschling et al., 1990; Baur et al., 2001). Related to this issue, telomeric DNA in mammals is in fact not exclusively bound by telomere-binding proteins; a significant portion of telomere distal repeat DNA is organized in nucleosomes.
The second major problem arising at telomeres is how to maintain telomeric repeat tracts at functional lengths in dividing cell populations. The most prevalent solution involves the telomerase holoenzyme, a reverse transcriptase that uses the free 3′ end of the G-tails as primer to add species-specific telomeric G-strand repeats by using a short region on its associated RNA as a template (Greider and Blackburn, 1987; Greider and Blackburn, 1989; Kelleher et al., 2002; Smogorzewska and de Lange, 2004). The C-rich telomeric strand is synthesized by components of the conventional replication machinery that is involved in lagging-strand synthesis. Furthermore, telomeric G-strand and C-strand syntheses may be coordinated and temporally co-regulated (Price, 1997; Diede and Gottschling, 1999; Adams Martin et al., 2000). Consistent with this notion, synthesis of telomeric repeats occurs specifically in late S phase (Marcand et al., 2000). Therefore, an emphasis of recent research has been on how this telomere-specific end replication becomes actively engaged and how it is regulated to maintain telomeres at lengths that are within the species-specific window.
Progress towards those ends has been made in several areas. With few exceptions, the evidence argues against replication initiating at telomeres. Therefore, replication forks should move from the interior of the chromosome towards the ends. Given the length and conserved directionality of the double-strand telomeric repeats, the C-strand will always be synthesized by lagging-strand synthesis and the G-strand by leading-strand synthesis. Although it is not known exactly how far the conventional fork remains intact on telomeres, the outcome at the very ends of the two daughter chromatids is thought to be different: a blunt end on the leading-strand ends and a recessed end at the lagging-strand ends. These two types of end may either be sensed differently or require different levels of processing (Bailey et al., 2001; Parenteau and Wellinger, 2002). Irrespective of these considerations, it is important to retain the intimate association between the conventional DNA-replication machinery with telomeric repeat replication and, hence, telomere length regulation (Chakhparonian and Wellinger, 2003). The passage of the replication fork through the telomeric repeats may destabilize the telomeric chromatin and remove t-loop structures, allowing telomerase and perhaps other enzymes access to telomeres.
Second, the molecular details of the transition from conventional replication to end-replication are just beginning to be understood. Conceivably, telomere-associated DNA-repair proteins could get directly involved at this step. For example, the presumed blunt ends on the leading-strand chromatids could be recognized as damage, engaging the associated signaling and processing proteins such as Ku, the Mre11 complex and ATM/ATR. The result would be the generation of the required G-tails and binding of the G-tail-associated proteins. In yeast cells, the G-tail-binding protein Cdc13p is required to orchestrate and regulate end replication by telomerase and C-strand synthesis (Evans and Lundblad, 2000; Qi and Zakian, 2000). The link between the DNA-repair proteins and end replication is further emphasized by the fact that, in yeast, the Ku proteins directly bind the RNA component of telomerase and play a crucial role in the association of the enzyme with telomeres (Fisher et al., 2004). At the point of engaging productive end replication, the regulation of further interactions becomes crucial: the telomerase enzyme must gain access to its substrate, and specific telomerase subcomponents, such as Est1p in yeast, must be associated with Cdc13p and telomerase in an active fashion. If these intricate reorganizations at the telomeres do not occur in the restricted temporal window in S phase, productive end replication may not be established. In yeast, for example, not all telomeres are elongated in every cell cycle and those telomeres with relatively short repeat tracts have a better chance of completing the required steps successfully (Teixeira et al., 2004).
Third, for telomere lengths to stay within certain limits, an active end elongation complex must be halted – not a small feat. The precise molecular mechanisms involved in this regulatory step are nebulous. However, in yeast and mammals, the telomeric proteins involved in assembling the complexes on the double-strand repeats are crucial for this regulation. Indeed, they somehow progressively inhibit telomeric repeat synthesis as the overall length of the particular telomere increases (Marcand et al., 1997; van Steensel and de Lange, 1997; Shore, 2001; Smogorzewska and de Lange, 2004). In addition to merely stopping the elongation, a functional telomeric complex must be re-generated following synthesis. At the level of DNA, in mammalian cells this may involve the creation of the t-loop structures.
For all of these above steps, one can expect that post-translational modifications of the central proteins play important, albeit as yet unknown, regulatory roles. The observed mean lengths of telomeric repeat tracts thus will be the net result of multiple processes occurring at telomeres, including those affecting conventional replication, end replication, end-processing steps, accidental degradation and even occasional losses of large portions of the tracts by a recombination-based mechanism called telomere rapid deletion (TRD) (Lustig, 2003). All of these mechanisms may be subject to their proper regulatory events, which renders the study of overall telomere length quite complex. However, given the established importance of telomeres for genome stability as well as the strong links between telomere biology, human disease and aging, there really is no end in sight for research on telomeres.
- © The Company of Biologists Limited 2005