Identification of TERRA locus unveils a telomere protection role through association to nearly all chromosomes

Abstract

Telomeric RNAs (TERRAs) are UUAGGG repeat-containing RNAs that are transcribed from the subtelomere towards the telomere. The precise genomic origin of TERRA has remained elusive. Using a whole-genome RNA-sequencing approach, we identify novel mouse transcripts arising mainly from the subtelomere of chromosome 18, and to a lesser extend chromosome 9, that resemble TERRA in several key aspects. Those transcripts contain UUAGGG-repeats and are heterogeneous in size, fluctuate in abundance in a TERRA-like manner during the cell cycle, are bound by TERRA RNA-binding proteins and are regulated in a manner similar to TERRA in response to stress and the induction of pluripotency. These transcripts are also found to associate with nearly all chromosome ends and downregulation of the transcripts that originate from chromosome 18 causes a reduction in TERRA abundance. Interestingly, downregulation of either chromosome 18 transcripts or TERRA results in increased number of telomere dysfunction-induced foci, suggesting a protective role at telomeres.

Figure 1. Identification of UUAGGG-containing transcripts that arise from the subtelomere of chromosome 18 that resembles TERRA.

(a) Graphs show the read density corresponding to (left) the TERRA biotin pull-down sample or (right) to the sum of all sample in a 30-kb region adjacent to the telomere of each chromosome. Note that chromosome 4 and Y are not sequenced until the telomere. (b) Reverse-transcribed RNA prepared with either an oligo complementary to the telomeric repeat (4xCCCTAA; ‘telomeric cDNA’) or with random hexamers (‘standard cDNA) was used for quantitative PCR detection of chromosome 18 transcripts and the non-telomeric genes Cyclin D2 and Cytochome b. Data provided are the mean values±s.e.m. from three different iPS clones. (Bottom) Reverse-transcribed RNA prepared with two different oligos complementary to the telomeric repeat (CCCTAA)₄ or (CCCTAA)₂ were used for PCR detection of chromosome 18 transcripts using primer Chr18-1. Two different concentrations of oligos were used. RT(-) reactions (performed in the absence of reverse transcriptase) are shown to exclude possible amplifications because of genomic DNA contamination (see full gel in Supplementary Note 1). (c) Northern blotting using ³²P-dCTP-labelled probes targeting either transcripts arising from the subtelomere of chromosome 18 (Chr18-RNAs) or TERRA’s telomeric track (TERRA) in three independent clones of pMEF and iPS; hybridization of 18S was included as a loading control. Both northern blots were done on the same membrane, first chromosome 18 probe and, after stripping, TERRA probe. *Unspecific band due to cross-hybridization with rRNA 18S and 28S. (Graph) Northern blot quantification. Mean values±s.e.m. from the three different clones are indicated. (d) Confocal microscopy images of double RNA-FISH preparations using probes targeting either chromosome 18-RNAs (probes 18-3-1 and 18-3-4; red) or TERRA’s telomeric track (green). (Graph) The percentages of co-localization per nuclei of chromosome 18 probes with TERRA spots are represented (mean±s.d., n=number of nuclei; three independent experiments). Total number of foci and nuclei used for the analysis are indicated. Scale bar, 10 μm. The Student’s t-test was used for all statistical analysis (*P<0.05 and **P<0.001).

Figure 2. Transcripts arising from the subtelomere of chromosome 18 show a ‘TERRA behaviour’.

(a) Immunofluorescence to detect the telomere marker Rap1 (green) followed by RNA-FISH to detect either TERRA or chromosome 18-transcripts (red). (Graph) Percentages of Rap1 foci co-localizing/associating with TERRA or chromosome 18-RNAs per nuclei (mean±s.e.m., n=number of nuclei; three different antibodies were used for telomere detection (Rap1, TPP1 and TRF1); see the results of the other two in Supplementary Fig. 8A and B). (b) Upon pMEF synchronization, TERRA and chromosome 18-RNA levels were measured by RNA dot-blot at different time points upon serum release; 18S serves as a loading control. (Top graph) Quantification of transcripts levels normalized by 18S. (Bottom graph) Percentage of cells in G0/G1 and S phase upon serum release. (c) Immunoprecipitation (IP) assay with antibodies recognizing hnRNP A1 or HuR followed by qRT–PCR for chromosome 18-RNAs detection using primers against different regions. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA was used for normalization. Data were compared with respect an IgG-IP (mean values±s.e.m. from three different iPS clones). (d) Cells were irradiated with ultraviolet C and, upon recovery, fixed for RNA-FISH. Confocal microscopy images of double RNA-FISH using probes targeting either chromosome 18-RNAs (probes 18-3-1 and 18-3-4; red) or TERRA’s telomeric track (green) are shown. (Graph) Percentages of co-localizing foci per nuclei (mean values +s.e.m., n=number of nuclei; two different probes were used to detect chromosome 18-RNAs). (e) Upon 5′Azacytidine treatment, RNA was isolated and use for (left) TERRA detection by RNA dot-blot with a probe against the telomeric track; 18S serves as loading control. (Graph) TERRA quantification normalized by 18S (mean values±s.d., n=three replicates). (f) Quantification chromosome 18 transcripts by qRT–PCR using primers targeting different regions (mean values±s.d., n=three replicates). Tmx3 is the coding gene closest to chromosome 18 telomere and Malat1 a long-noncoding RNA located in a non-subtelomeric region. Student’s t-test was used in all statistical analysis (*P<0.05, **P<0.001 and ***P<0.0001). Total number of foci and nuclei are indicated in the corresponding panels. Arrowheads and arrows indicate co-localization and association events (partial co-localizations), respectively. Untr, untreated. Scale bar, 5 μm. NS, not significant.

Figure 3. Identification of transcription initiation and termination sites of chromosome 18-RNAs as well as their promoter.

(a) UCSC snapshot depicting, from top to bottom, putative promoter regions (A2, A3, B3, C2 and D), genomic scale, genomic position, primer position, ‘genome walking’ transcripts, annotated mouse EST and the sequences cloned upon 5′RACE and 3′RACE experiments (RACE: Rapid Amplification of CDNA Ends; red lines indicate mismatches with respect to the reference genome). (b) The promoter-free vector pGL3 (firefly luciferase reporter system) containing the different putative promoter regions was transiently cotransfected into iPS or their parental pMEFs cells along with pGL4-Renilla (used to normalize for transfection efficiency); 48 h later, protein was extracted and used for the detection of firefly and renilla luciferase activities. Graph shows the relative fold increase in firefly luciferase activity seen in the pGL3-containing promoter regions relative to the empty vector after normalization to renilla activity. Mean values±s.d., n=3 technical replicates from one representative experiment (two independent transfections were performed in pMEFs and three in iPS; the activity of regions A2 and A3 was significant different compared with the other regions in all transfections). Student’s t-test was used for statistical analysis (*P<0.05 and **P<0.001). (c) Diagram showing, from top bottom, genomic scale, annotated EST, position of promoter regions A2 and A3 and examples of transcription factor-binding sites from ChIP data from the Stanford/Yale/ENCODE Project. NS, not significant.

Figure 4. Identification of the chromosomes bound by chromosome 18-TERRAs.

(a) RNA-FISH with probes targeting chromosome either 18-RNAs (red) or TERRA’s telomeric track (green) or (b) with both probes (yellow) is shown: (top-middle) RNA-FISH staining in metaphases, (top-right) SKY hybridization and (bottom) chromosome identification by SKY. Those chromosomes showing probe association have been circled and identified in the RNA-FISH preparations according to the SKY information. Zoom of these associations is shown on the left. (Graph a) Percentage of metaphases showing one o more association with either TERRA or with chromosome 18-RNAs (mean±s.e.m.). n=number of metaphases analysed; three independent experiments. Scale bar, 10 μm. (Graph b) Percentage of co-localization of TERRA and chromosome 18-RNAs per metaphases. n=number of metaphases analysed; two independent experiments.

Figure 5. Downregulation of chromosome 18-TERRAs induces telomere damage.

(a) Cells were transfected with either control Gapmer-LNA (Ctrl) or Gapmer-LNA targeting chromosome 18-RNAs and RNA collected 2 days post transfection. The graph shows the percentage of chromosome 18-RNA levels normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) levels upon Gapmer-LNA transfection (mean±s.e.m.; three independent transfections). (b) (Left) RNA dot-blot to detect TERRA using a ³²P-dCTP-labelled probe; hybridization of 18S rRNA was included as a loading control. (Right) Quantification of the RNA dot-blot signals normalized by 18S rRNA (mean±s.e.m.; three independent transfections). (c) Representative images of TRF1 (green) and γH2AX (red) fluorescence and of the merged images. Co-localization events (arrowheads) indicate telomere dysfunction-induced foci (TIF). Scale bar, 10 μm. (Graph) Percentage of cells with ≥2 or ≥3 TIFs/nuclei (mean±s.d., n=number of nuclei; three independent transfections). Student’s t-test was used for statistical analysis (*P<0.05 and **P<0.001).