Functional analysis of the telomeric long non-coding RNA TERRA

Telomeres are nucleoprotein structures present in the end of linear chromosomes. In humans, a core complex of 6 proteins cap and protect the chromosome ends from being recognized as double-strand breaks and eliciting unwanted DNA damage and repair responses, which lead to chromosomal end-to-end fusions and genomic instability. In addition to the shelterins, another 200 proteins are detected at telomeres by mass spectrometry. Their functions at telomere are not yet completely understood. Telomeres are transcribed into a long non-coding RNA called TERRA. It is believed that subtelomeric CpG islands act as TERRA promoters. TERRA is suggested to regulate telomerase activity, to participate in the processing of uncapped telomeres, in telomeric heterochromatin formation and in promoting recombination-mediated telomere elongation. TERRA levels are increased in patients affected by the ICF (immunodeficiency, centromeric instability, facial anomalies) syndrome. Since these patients also present very short telomeres, it has been suggested that TERRA regulates telomere length, either by inhibiting telomerase or by promoting access to end-attacking nucleases. We used HCT116 DNMT1 (deletion of exons3-5/exons3-5) DNMT3B (-/-) (DKO) cells as a model for the ICF syndrome to try to comprehend the relationship between TERRA expression and telomere length. In this thesis, we try to understand (1) if there is a direct correlation between TERRA expression and telomere length, (2) what drives TERRA transcription, (3) what are the consequences of TERRA overexpression, and (4) what could cause the short telomere phenotype in DKO cells. We show that not all telomeres overexpress TERRA in DKO cells, but even telomeres with normal TERRA expression are short. By analyzing subtelomeric sequences, we found that the presence and density of subtelomeric CpG islands determine if its transcript will be upregulated in DKO cells or not. We used absolute quantification methods to demonstrate that TERRA is expressed from multiple chromosome ends, disproving a recent proposal that TERRA stems only from one locus. We also discovered that CpG-negative subtelomeres express TERRA in levels similar to CpG-positive ones, suggesting that the presence of subtelomeric CpG-islands is not a requirement for high TERRA expression. We next evaluated the telomere length dynamics in DKO cells. We found that DKO cells show lower telomere elongation rates than the WT. Moreover, DKO cells fail to downregulate TERRA in S-phase, in contrast to WT HCT116 and HeLa cells. The downregulation of TERRA in late S-phase has been proposed to alleviate TERRA-dependent telomerase inhibition and thereby allow the enzyme to extend short telomeres. To better understand TERRA transcription, we performed a siRNA screen to discover new TERRA transcription regulators. We found ZNF148, ZFX, PLAG1 and EGR1 to be new repressors of TERRA transcription. Finally, we established a CRISPR activation system to induce overexpression of endogenous TERRA. Overexpression of 10q and 13q TERRA in HCT116 did not lead to telomere shortening in cis. In HeLa cells, overexpression of multiple TERRA species led to deposition of the heterochromatin marks H3K9me3 and H4K20me3 in the telomeres. These marks were also enriched in subtelomeric sequences in cis. Since TERRA has been shown to interact with SUV39H1 to ensure H3K9me3 deposition, we speculate that TERRA also recruits SUV420H2 to induce H4K20me3 accumulation.