Telomerase and chromosome end replication

Telomere length homeostasis requires telomerase, a cellular reverse transcriptase, which uses an internal RNA moiety as a template for the synthesis of telomere repeats. Telomerase elongates the 3’ ends of chromosomes, whereas the complementary strand is filled in by conventional DNA polymerases. In human, telomerase is ubiquitously expressed only during the first weeks of embryogenesis and subsequently downregulated in most cell types. Correct telomere length setting is crucial for long-term survival. Telomere shortening suppresses tumor formation through limiting the replicative potential of cells. On the other side the telomere length reserve must be sufficient to avoid premature cellular senescence and acceleration of age-related disease.

Structure and function of telomerase

The telomerase holoenzyme consists of an RNA moiety (TR), which is associated with several protein subunits. The catalytic protein subunit is referred to as TElomerase Reverse Transcriptase (TERT). It bears structural similarity with reverse transcriptases from retroelements and provides the active site. TERT extends chromosome ends by iterative reverse transcription of its RNA template. Following the addition of each telomeric repeat, the RNA template and the telomeric substrate translocate, resetting their relative position in the active site, which involves disruption and reformation of base-pairs. We assessed the telomerase reaction cycle in yeast by probing base-pair accessibility of the RNA template by dimethyl sulfate (DMS) in the absence and presence of DNA substrates and during elongation. DMS‑methylation at position N‑1 of A and N‑3 of C in RNA occurs readily in single-stranded RNA but the bases become protected upon base pair formation or protein binding. The modified positions are detected as stop sites during primer extension by reverse transcriptase. Our analysis demonstrates that the length of the RNA/DNA hybrid is kept constant at 7 base pairs upon primer binding and elongation (Fig. 1). Thus, new base pair formation at the 3’ end of the substrate during elongation coincides with disruption of base pair interactions at the other side of the template. Presumably, this circumvents the generation of an exceedingly high energy barrier for translocation and dissociation.

Figure 1: The telomerase reaction cycle.
The telomerase reverse transcriptase (schematically depicted in blue) reverse transcribes the telomerase RNA template (red) for the synthesis of telomeric repeats. Telomere binding involves base-pairing with the RNA template. During substrate elongation, base-pair formation at the DNA 3’ end coincides with base-pair disruption further 5’ on the DNA. Limitation of RNA-DNA duplex length should facilitate primer dissociation or translocation.

Telomere length homeostasis: in vivo analysis of telomere replication

Telomerase activity is regulated at individual chromosome ends in cis, to achieve telomere length homeostasis. This regulation involves telomeric proteins and perhaps higher order DNA structures. A precise understanding of telomere length homeostasis has been hampered by the lack of assays that delineate the non-uniform telomere extension events of single chromosome molecules.

To study the mechanisms that control telomerase access and extension efficiency we developed for Saccharomyces cerevisiae a system to measure telomerase activity at nucleotide resolution at single chromosome end molecules. In this method, a telomerase‑negative strain is mated with a telomerase-positive wild type strain. After the first zygotic cell cycle, telomere elongation is determined by telomere‑PCR, which involves cloning and sequencing of marked telomeres derived from the telomerase-negative parent. Telomere elongation is detected due to divergence of telomeric sequences, which is dependent on yeast telomerase activity. Our results demonstrate for the first time that telomerase does not act on every telomere in each cell cycle. Instead, it exhibits an increasing preference for telomeres as their lengths decline. The frequency of telomere extension increases steadily as a function of telomere length, from roughly 6‑8% at 300 nucleotide-long telomeres to 42‑46% at 100 nucleotides. Deletion of the telomeric proteins Rif1 or Rif2 gives rise to longer telomeres by increasing the frequency of elongation events. The extent of telomere elongation, on the other hand, is not regulated by Rif1 and Rif2. Recombination is contributing to telomere maintenance at an estimated frequency of less than 0.3% per generation. Thus, we estimate that the contribution of recombination to telomere maintenance is 20‑30‑fold lower than that of telomerase. In summary, our analysis of a single round of telomere replication demonstrates that telomere length homeostasis is achieved via a switch between telomerase extendible and non-extendible states. We are now exploiting our experimental system to study the function of other telomere replication factors for telomerase recruitment and processivity.

In human cells, the telomeric repeat binding factor TRF1 inhibits telomerase at telomeres in cis in a length-dependent manner to achieve telomere length homeostasis. We found through increasing telomerase activity in human cells over a wide range that telomerase association with telomeres as measured by chromatin immunoprecipitation depends on telomerase concentration. Overexpression of telomerase increased its association with telomeric DNA and this was sufficient to elongate telomeres in primary or cancer cells in a length-independent manner far beyond physiological size. Thus, even long telomeres are extendible, indicating that the non-extendible state is not adopted permanently. In addition we predict existence of a distinct third telomeric state with extendible telomeres switching to an extending state, upon productive association with telomerase (Fig. 2). In this model, the transition between the non-extendible and extendible states is length dependent, whereas the transition between extendible and extending states depends on the concentration of telomerase. Thus, a low cellular concentration of telomerase is critical to achieve preferential elongation of short telomeres and telomere length homeostasis. Our ability to detect human telomerase at chromosome ends by chromatin immunoprecipitation will allow us to identify the factors, which regulate this interaction.

Figure 2: A three-state model for telomere length homeostasis.
Telomeres are proposed to switch between non-extendible (curly end, left), extendible (straight, middle) and extending states (straight end associated with telomerase, right). The equilibrium between non-extendible and extendible states (1) is a function of telomere length, while the equilibrium between extendible and extending states (2) is a function of telomerase concentration. The molecular natures of extendible and non-extendible states remain speculative.

Telomeric states: function of human POT1 and of EST1-homologs

The presumed major single-stranded telomere binding protein in humans, POT1, has been proposed to act as the terminal transducer of telomere length control from TRF1. We have purified recombinant human POT1 to determine its function in telomerase control. Association of recombinant POT1 with telomeric DNA oligonucleotide ends inhibits binding of telomerase. This is consistent with the notion that POT1 negatively affects telomere length. Thus, POT1 bound to the telomeric 3’ end may correspond to the non-extendible telomeric state discussed above. It is currently unclear if for telomere elongation, POT1 dissociates from the telomeric 3’ end or if POT1 looses its telomerase inhibitory effects when associated with other factors. The latter would be reminiscent of the situation in S  cerevisiae, where the putative POT1 ortholog Cdc13p, can either function as a negative, or, when associated with ScEst1p, as a positive regulator of telomerase.

In S. cerevisiae, the Ever Shorter Telomeres 1 (EST1) gene product recruits telomerase to 3’ ends of telomeres allowing telomere elongation in S‑phase. It is unclear if similar mechanisms of telomerase regulation have been conserved in vertebrates during evolution. By iterative profile searches we and others identified a human homolog of yeast Est1p (hEST1A) that associates with at least 70% of active telomerase in HeLa cells. Over-expression of EST1A in a human fibrosarcoma-derived cell line (HT1080) induces telomere-telomere associations causing chromosome bridges during anaphase, followed by a rapid apoptotic response. Telomeric fusions in hEST1A‑overexpressing cells may stem from telomere-uncapping followed by DNA break repair, reminiscent of the loss of function phenotype of the double strand telomere binding protein TRF2. Thus, we hypothesize that endogenous hEST1A is involved in modulating the telomere structure. To further elucidate hEST1A function at telomeres, we are characterizing the domains of hEST1A that mediate its interaction with telomerase and are analyzing the function of mutant hEST1A‑alleles.

Dual roles of hEST1A and UPF1 in DNA and RNA surveillance pathways

Recent reports by several laboratories uncovered dual functions of a few polypeptides in DNA and RNA surveillance pathways. The nonsense-mediated RNA surveillance pathway (NMD) triggers rapid decay of mRNAs that bear premature stop codons. The above discussed telomerase-associated hEST1A is identical to hSMG6 and also involved in NMD. Similarly, the human phosphoinositide 3‑kinase‑related protein kinase (PIKK) SMG1 is activated in response to DNA damage and required for optimal activation of p53. However, SMG1 is also required for NMD. We now find that the RNA‑ and DNA‑dependent 5’‑3’ NMD‑helicase UPF1 is also required for genome stability. ShRNA‑mediated depletion of UPF1 causes human cells to arrest early in S‑phase inducing an ATR-dependent DNA damage response. A fraction of hyperphosphorylated UPF1 associates with chromatin in S‑phase and upon g‑irradiation. The DNA damage checkpoint PIKK ATR phosphorylates UPF1 both in vitro and in vivo, and shRNA‑mediated downregulation of ATR strongly diminishes the DNA damage-induced association of UPF1 with chromatin while it did not affect NMD. In contrast downregulation of the NMD factor UPF2 does not interfere with cell cycle progression and DNA stability, supporting the notion that only some NMD‑factors play a role in genome stability. Dual functions of polypeptides in RNA and DNA/telomere surveillance pathways may facilitate integration of cellular responses to multiple macromolecular damage, which is induced by genotoxic agents.

Keywords

Telomeres, telomerase, cellular senescence, DNA damage response, genome stability