Functional characterization of the TERRA transcriptome at damaged telomeres

Abstract

Telomere deprotection occurs during tumorigenesis and aging upon telomere shortening or loss of the telomeric shelterin component TRF2. Deprotected telomeres undergo changes in chromatin structure and elicit a DNA damage response (DDR) that leads to cellular senescence. The telomeric long noncoding RNA TERRA has been implicated in modulating the structure and processing of deprotected telomeres. Here, we characterize the human TERRA transcriptome at normal and TRF2-depleted telomeres and demonstrate that TERRA upregulation is occurring upon depletion of TRF2 at all transcribed telomeres. TRF2 represses TERRA transcription through its homodimerization domain, which was previously shown to induce chromatin compaction and to prevent the early steps of DDR activation. We show that TERRA associates with SUV39H1 H3K9 histone methyltransferase, which promotes accumulation of H3K9me3 at damaged telomeres and end-to-end fusions. Altogether our data elucidate the TERRA landscape and defines critical roles for this RNA in the telomeric DNA damage response.

Figure 1. TERRA expression is upregulated upon persistent DDR and TRF2 depletion

(a) DNA-ChIP of telomeric and centromeric DNA with anti-TRF2 antibody performed in HeLa cells treated with zeocin for the indicated times. (b) Quantification of three independent ChIP experiments represented in a (mean±s.d., n=3). (c) DNA-ChIP of telomeric and centromeric DNA with anti-γH2AX antibody performed in HeLa cells treated with zeocin for the indicated times. (d) Quantification of three independent ChIP experiments represented in c (mean±s.d., n = 3). (e) Northern blot analysis of total RNA from HeLa cells treated with zeocin. TERRA was detected using a telomeric DNA probe complementary to the UUAGGG repeats. Filters were then stripped and reprobed for 18S rRNA. (f) DNA-ChIP of telomeric and centromeric DNA with anti-γH2AX antibody performed in HeLa cells depleted of POT1 and TRF2, respectively. (g) Quantification of three independent ChIP experiments represented in f (mean±s.d., n=3). (h) Northern blot analysis of total RNA from HeLa cells depleted of POT1 and TRF2.

Figure 2. Characterization of TERRA transcriptome at TRF2-depleted telomeres

(a) Enrichment profiles of RNA-seq data of UUAGGG-containing RNAs versus total nuclear RNA prepared from HeLa cells depleted of TRF2 at 10 different chromosome ends showing transcription within subtelomeric regions. Red and green labelling indicates the p- and the q-arms, respectively. Yellow rectangles indicate TTAGGG repeats interspersed in subtelomeric regions. The blue line is a moving average over a window of 10 nucleotides. Enrichment profiles of ChIP-seq analysis of CTCF are shown below the RNA-seq profile data for each chromosome end. (b) Comparison of the GC content of TERRA transcribed versus non-transcribed subtelomeres. The plots were generated using the Wistar GC content track and are an average of the two subtelomere subsets. (c) Average GC content calculated in the first 5 kb of each subtelomere. The boxes represent the GC content distribution of the two subtelomere subsets (TERRA transcribed versus non-transcribed subtelomeres). The middle line is the median, whereas the lower and upper bounds are the mininimal and maximal values. (d) Reads per kilobase per million reads (RPKM, times 10² for practicality) as a measure for chromosome-specific TERRA levels in wild-type (shEV) and TRF2-depleted cells (shTRF2).

Figure 3. TERRA transcriptional rate but not half-life increases upon TRF2 depletion

(a) TERRA half-life. Wild-type and TRF2-depleted HeLa cells were treated with 5 µg ml⁻¹ actinomycin D for the indicated times. Total RNA was prepared and subjected to Northern blot analysis using a telomeric DNA probe detecting UUAGGG repeats. Filters were then stripped and reprobed successively for c-MYC, β-actin and 18S rRNA. (b) Quantification of TERRA half-life in wild-type and TRF2-depleted HeLa cells based on northern blots shown in a. Relative TERRA levels were calculated using a scatter plot analysis after normalizing TERRA to 18S rRNA. The best-fit exponential regression curve of the data points was determined for both conditions. R² values are indicated. (c) DNA-ChIP of telomeric and centromeric DNA with anti-RNAPII antibodies performed in HeLa cells depleted of TRF2. (d) Quantification of three independent ChIP experiments represented in c (mean±s.d., n=3).

Figure 4. DDR activation is not required for TERRA upregulation at uncapped telomeres

(a) Western blot analysis to evaluate knockdown efficiency of endogenous MRE11, NBS1 and TRF2. Tubulin was used as loading control. (b) Wild-type (shEV) and TRF2-depleted (shTRF2) HeLa cells transfected with control vector (shEV) or additionally depleted for MRE11 (shMRE11) and NBS1 (shNBS1). Total RNA was prepared and subjected to northern blot analysis using a telomeric DNA probe detecting UUAGGG repeats. Filters were then stripped and reprobed subsequently for 18S rRNA. (c) Western blot analysis to evaluate knockdown efficiency of endogenous ATM, ATR and TRF2. Tubulin was used as loading control. (d) Wild-type (shEV) and TRF2-deficient (shTRF2) HeLa cells transfected with control vector (shEV) or depleted for ATM (shATM) and ATR (shATR). Total RNA was prepared and subjected to northern blot analysis using a telomeric DNA probe detecting UUAGGG repeats. Filters were then stripped and reprobed subsequently for 18S rRNA. (e) Northern blot analysis of total RNA prepared wild-type (p53 + ) and p53 knockout (p53 –) HCT116 cells. The blot was probed using a telomeric DNA probe detecting UUAGGG repeats. The filter was stripped and reprobed subsequently for 18S rRNA. Uncropped western blots are shown in Supplementary Fig. 9a,b.

Figure 5. TRF2 negatively regulates TERRA transcription through its homodimerization TRFH domain

(a) Schematic representation of TRF2/TRF1 chimeric constructs. (b) Western blot analysis to test expression of TRF2/TRF1 chimeric proteins in HeLa cells depleted of endogenous TRF2. Uncropped blots are shown in Supplementary Fig. 9c. (c) Immunofluorescence fluorescence in situ hybridization (IF-FISH) analysis to assess the localization of the TRF2/TRF1 chimeric proteins (red signals) at telomeres (green signals) in HeLa cells depleted of endogenous TRF2. (d) Northern blot analysis. Total RNA was prepared from wild-type HeLa cells or HeLa cells depleted of TRF2 that had been transfected with the indicated constructs. Northern blot analysis was then performed using a telomeric DNA probe detecting UUAGGG repeats. Filters were then stripped and reprobed subsequently for 18S rRNA.

Figure 6. SUV39H1 associates with TERRA and is recruited to uncapped telomeres

(a) ChIP of telomeric and centromeric DNA with H3 antibodies recognizing histone marks in HeLa cells transfected with control vector (shEV) or depleted of TRF2 (shTRF2). (b) Quantification of three independent ChIP experiments represented in a (mean±s.d., n=3). Statistical analysis was done using a two-tailed Student’s t-test (*P<0.05). (c) ChIP of telomeric and centromeric DNA with anti-SUV39H1 antibodies performed in HeLa cells transfected with control vector (shEV) or depleted of TRF2 (shTRF2). (d) Quantification of three different ChIP experiments represented in c (mean±s.d., n=3). Statistical analysis was done using a two-tailed Student’s t-test (*P<0.05). (e) RNA-ChIP assays with two antibodies against endogenous SUV39H1 performed in HeLa cells transfected with control vector (shEV) or depleted of TRF2 (shTRF2). IP-recovered RNA was detected with probes annealing with TERRA (left panel) or 18S rRNA (right panel).

Figure 7. The N-terminal domain of SUV39H1 directly binds UUAGGG-TERRA repeats

(a) Coomassie stained gel of GST-SUV39H1 affinity purified from E. coli. Bands labelled with asterisks are SUV39H1 degradation products. (b) Agarose EMSA used to measure SUV39H1 RNA binding activity in vitro. GST-SUV39H1 was assayed for binding to ³²P-(UUAGGG)₁₀ RNA or ³²P-(TTAGGG)₁₀ DNA oligonucleotides, respectively. Single asterisks indicate free oligonucleotides and double asterisks protein/oligonucleotide complexes. (c) Agarose supershift assay performed with GST-SUV39H1 incubated with ³²P-(UUAGGG)₁₀ RNA oligonucleotide and an antibody recognizing SUV39H1 polypeptide. Single asterisks indicate free oligonucleotides and double asterisks protein/oligonucleotide complexes. (d) EMSA competition experiments of GST-SUV39H1 binding to ³²P-(UUAGGG)₇. Increasing amounts of unlabeled (UUAGGG)₇, (AAAGGG)₇, (UUUCGG)₇ and (UUAGCC)₇ RNA oligonucleotides were added as competitors. Numbers indicate fold excess of competitor over ³²P-labelled probe. Single asterisks indicate free oligonucleotides and double asterisks protein/oligonucleotide complexes. (e) Scheme of SUV39H1 domains and deletion constructs. (f) Coomassie stained gel of GST-SUV39H1 deletion constructs affinity purified from E. coli. (g) Agarose EMSA used to identify domains of SUV39H1 with binding activity for ³²P-(UUAGGG)₁₀ RNA oligonucleotides. Single asterisks indicate free oligonucleotides and double asterisks protein/oligonucleotide complexes.

Figure 8. SUV39H1 sustains telomere end fusions occurring upon TRF2 depletion

(a) Timeline of the experiment in b and c). (b) Western blot of total protein extracts from HeLa cells with or without shTRF2 expression and siRNAs transfection. Uncropped blots are shown in Supplementary Fig. 9d. (c) Metaphase chromosome spreads and telomere-FISH performed on HeLa cells depleted (shTRF2 on) or not depleted (shTRF2 off) for TRF2 and concomitantly transfected with siRNAs against luciferase (siLuc) or SUV39H1 (siSUV39H1a and siSUV39H1b). (d) Model for the regulation and here-described roles of TERRA at dysfunctional telomeres.