Telomere Capping Structure

Even though there are many proteins involved in telomere biology, there is only a small number of specialized telomere-binding proteins currently known to protect chromosome ends from degradation and from end-to-end joining events. Yeast double stranded DNA (dsDNA) telomeric repeats are bound by Rap1

(repressor/activator site binding protein), a major component of telomeric chromatin (Conrad et al, 1990; Lustig et al, 1990). Rap1 is a large protein of 827 amino acid residues (Shore & Nasmyth, 1987) that contains two distinct but structurally similar Myb-like DNA-binding domains between residues 361 and 596 (Henry et al, 1990;

Konig et al, 1996). A structured linker between the domains and a long C-terminal tail contribute to the binding specificity (Konig et al, 1996). Rap1 also has a BRCT (BRCA1 C-terminal) domain in the N-terminal region, which is well conserved among Rap1-related proteins in other organisms.

Through a C-terminal domain Rap1 recruits two proteins, Rif1 and Rif2 (Rap1 interacting factor 1 and 2) (Hardy et al, 1992a; Hardy et al, 1992b; Kyrion et al, 1992; Wotton & Shore, 1997) (Fig. 1), which, together with Rap1, establish a protein counting mechanism regulating telomere length (properties and mechanism of this

Introduction

complex are described in Section 3.5). Rap1 C-terminal domain binds also proteins mainly involved in heterochromatin formation, Sir3 (silent information regulator 3) and Sir4 (Moretti et al, 1994; Cockell et al, 1995). Sir2, a nicotinamide adenine dinucleotide (NAD+)-dependent histone deacetylase, forms a complex with Sir3 and Sir4 by interacting mainly with Sir4 (Strahl-Bolsinger et al, 1997). Sir3 and Sir4 interact not only with Rap1 but also with the N-termini of histones H3 and H4, which are hypoacetylated by Sir2 (Hecht et al, 1995; Hoppe et al, 2002). Thus, the Sir2-Sir3-Sir4 complex spreads over the nucleosomes proximal to the telomeres from the nucleation sites with Rap1 at the telomeres.

The Cdc13 protein recognizes telomeric ss DNA and binds the G-overhang with high affinity and sequence specificity (Lin & Zakian, 1996; Nugent et al, 1996).

The Cdc13 DNA-binding domain is thought to comprises a single OB

(oligonucleotide/oligosaccharide binding) fold augmentedby an unusually large loop for DNA recognition (Mitton-Fry et al, 2002), even though computational analysis predicts the existence of multiple OB-folds (Theobald & Wuttke, 2004). Cdc13 performs its essential function at telomeres by protecting chromosome ends from degradation, and loss of its function leads to extensive resection of the C-strand and a Rad9-mediated cell cycle arrest (Weinert & Hartwell, 1993; Garvik et al, 1995;

Lydall & Weinert, 1995; Booth et al, 2001). Two essential Cdc13-associated proteins, Stn1 and Ten1, also contribute to this capping activity, and the lethality of the cdc13Δ strain can be rescued if Stn1 protein is artificially delivered to telomeres by fusion with the DNA binding domain of Cdc13 (Grandin et al, 1997; Grandin et al, 2001; Pennock et al, 2001) (for the function of this complex in telomerase

recruitment and telomere replication see Session 3.3 and 3.4).

In humans, six proteins, TRF1, TRF2, hRap1, TIN2, TPP1, and POT1, form the so-called shelterin complex, which is a constitutive component of human telomeres (reviewed in (de Lange, 2005)). The related TRF1 and TRF2 proteins bind double-stranded telomeric repeats as preformed homodimers (Bianchi et al, 1997), whereas POT1 binds to the single-stranded telomeric 3′ overhang (Fig. 1). TIN2 connects, through protein-protein interactions, TRF1, TRF2, and TPP1 (Kim et al, 1999; Ye & de Lange, 2004; Ye et al, 2004a). TPP1 binds, in addition, POT1, thus recruiting POT1 also to the double-stranded part of telomeres (Liu et al, 2004).

Human Rap1, the ortholog of yeast Rap1, associates with telomeres through

Introduction

interaction with TRF2 (Li et al, 2000; O'Connor et al, 2004). POT1 has also been reported to interact directly with TRF2 (Loayza & De Lange, 2003; Yang et al, 2005).

Many examples are available in which perturbation of shelterin components disturbs telomere length regulation, by interfering with telomere capping, telomerase action, or possibly telomere heterochromatin states (reviewed in (Hug & Lingner, 2006)).

It is suggested that the G-tail loops back, invades duplex telomeric DNA forming what is called a t-loop, which provides an elegant architectural solution for telomere capping. The model for a t-loop came from in vitro observation that telomeric repeat binding TRF2 can remodel linear telomeric DNA into large duplex loops (Griffith et al, 1999). Furthermore, electron microscopy analysis of psoralen cross-linked telomeric DNA purified from human and mouse cells revealed abundant large loops with a size distribution consistent with their telomeric origin (Griffith et al, 1999). t-loops seem to be conserved feature of telomere structure. These

structures have also been found in the germline DNA (found in the micronucleus) of the hypotrichous ciliate Oxytricha fallax (Murti & Prescott, 1999), in the flagellate protozoan Trypanosoma brucei (Munoz-Jordan et al, 2001), and in Pisum sativum (peas)(Cesare et al, 2003). Experimental evidence for fold-back structures exists in S. cerevisiae, but the overhang status in these structures is unknown

(Strahl-Bolsinger et al, 1997; de Bruin et al, 2001). Very recently t-loops and homologous recombination dependent t-circles were observed in yeast Kluyveromyces lactis telomere mutant strain (Cesare et al, 2007). This strain, however, contains mutant telomeres that are poorly bound by Rap1, resulting in a telomere uncapping

phenotype and significant elongation of the telomeric DNA. Thus, even though the elongated telomeres allowed the isolation and examination of native yeast telomeric DNA by electron microscopy, it is not shown that wild-type K. lactis telomeres form such structures. t-loops are proposed to be precursors of t-circles formed in this strain, and might be formed due to uncapping of the telomeres.

Sequestration of the 3’-OH terminus in a t-loop can mask the G-tail from a variety of detrimental activities, but there is evidence that its binding by single-stranded DNA (ssDNA) proteins is sufficient for protection. As mentioned above the micronucleus of hypotrichous ciliates have t-loops, however their macronuclear telomeres do not (Murti & Prescott, 1999). The macronucleus contains highly amplified gene-sized DNA molecules at a remarkable copy number of 2 10⁷ per

Introduction

nucleus and an average size of just a few kilobases. Rather than containing long telomeres with t-loops, these DNA molecules are capped by ultrashort telomeres with termini that are hidden inside a tenaciously bound protein complex (Horvath et al, 1998).