Oxidative stress at telomeres triggers internal DNA loops, TRF1 dissociation, and TRF2-dependent R-loops

Abstract

Telomeres are the nucleoprotein structures at chromosome ends. Telomeres are particularly sensitive to oxidative stress, which can induce telomere damage, shortening, and premature cellular senescence. How oxidative damage influences telomere structure has not been defined. Here, we induce oxidative damage at telomeres using menadione, which damages mitochondria mimicking intrinsic oxidative stress. We find that oxidative stress induces at telomeres single-stranded DNA breaks, internal DNA loop structures, dissociation of the shelterin component TRF1, upregulation of TERRA long noncoding RNA, and increased DNA:RNA hybrid structures known as R-loops. R-loop formation is enhanced not only in cis at telomeres, which show increased TERRA transcription, but also in trans at telomeres at which TERRA transcription is not induced indicating post-transcriptional R-loop formation. Finally, we show that oxidative damage induced R-loop formation requires TRF2, whose R-loop promoting activity may be unleashed upon TRF1 dissociation from telomeres. Altogether, our findings uncover in response to oxidative stress major remodelling of telomeric DNA, RNA, and shelterin complexes, and they unravel a physiological role of TRF2's ability to stimulate TERRA R-loop formation. We propose that the identified structural changes may facilitate DNA damage signalling and repair pathways to maintain telomere integrity during development and aging.

Graphical Abstract

Figure 1.

Accumulation of DNA SSBs at telomeres upon oxidative stress induction in HEK293E cells. Detection of telomeric SSB accumulation at G-rich (A) and C-rich (B) strands by alkaline gel electrophoresis upon menadione treatment with or without prior treatment with NAC. The signal intensity curves are shown on the right. (C) Quantification of DNA damage levels in panels (A) and (B), represented by the percentage decrease in telomeric fragment length of the menadione-treated sample compared to untreated. (D) Detection of telomeric SSB accumulation at C-rich strand by alkaline gel electrophoresis upon menadione and H₂O₂ treatment. The signal intensity curves are shown on the right. (E) Quantification of DNA damage levels in panel (D) represented by the percentage decrease in telomeric fragment length of the menadione-treated sample compared untreated. (*) P < .05, ns—not significant. Data represent mean ± SD. n = 3.

Figure 2.

Accumulation of i-loop structures at telomeres upon oxidative stress induction in HEK293E cells. (A) Hypothetical models for i-loop structures induced by ssDNA gaps on telomeric strands (Adapted from [27]). This illustration was created in BioRender. Nguyen, T. (2025) https://BioRender.com/d44u613. (B) Schematic representation of the telomeric DNA enrichment procedure. (C) Dot blot analysis for quantification of the enrichment of telomeric repeats after the telomeric DNA purification procedure. The indicated amounts of DNA from each round of enrichment were blotted onto a membrane, which was then hybridized with telomere-specific probes. (D) Quantification of the telomeric DNA signal relative to the not-enriched DNA. (E) Representative EM images of telomeric DNA molecules with i-loops observed in the menadione-treated samples. (F) Percentages of molecules containing i-loops is determined based on the EM analysis in panel (E). (*) P < .05, ns—not significant. Data represent mean ± SD. n = 5.

Figure 3.

Upregulation of TERRA R-loops at damaged telomeres upon oxidative stress in HEK293E cells. Detection of total R-loops at telomeres upon menadione (A) or H₂O₂ (C) treatment by DRIP-dot blot assay using S9.6 antibody. In vitro digestion with RNaseH prior to the immunoprecipitation served as a control for R-loop specificity. Immunoprecipitated and input samples were blotted onto a membrane, followed by hybridization with a telomere-specific probe. (B, D) Quantification of telomeric DNA signal in (A) and (C), respectively, as fold change over untreated sample. (*) P < .05, ns—not significant. Data represent mean ± SD. n = 3

Figure 4.

TERRA R-loops are formed both in-cis and in-trans at damaged telomeres upon oxidative stress in HEK293E cells. (A) R-loop levels at indicated chromosome ends measured by DRIP–qPCR upon menadione treatment. qPCR was performed using primers binding to chromosome-specific subtelomeric sequences. Data were calculated as percent of input, and shown as fold change over untreated sample. (B) Detection of TERRA transcribed from indicated chromosome ends upon menadione treatment by RT-qPCR, which was performed using primers binding to chromosome-specific subtelomeric sequences. Data were normalized to GAPDH, and are shown as fold change over untreated sample. (C) Detection of TERRA transcribed from indicated chromosome ends by RT-qPCR. The cells were treated with actinomycin D for 30 min, followed by 2 h of menadione treatment. Data are shown as fold change over untreated sample. (D) Detection of total R-loop level by DRIP-dot blot assay. The cells were treated with actinomycin D for 30 min, followed by 2 h of menadione treatment. (E) Quantification of telomeric DNA signal in panel (D) as fold change over untreated sample. (*) P < .05, ns—not significant. Data represent mean ± SD. n = 3

Figure 5.

TRF1 proteins dissociate from oxidatively damaged telomeres in a ROS-dependent manner, while TRF2 proteins are unchanged. (A) Telomeric TRF1 and TRF2 protein levels detected by ChIP-dot blot following menadione treatment in HeLa cells. Immunoprecipitated DNA was blotted onto a membrane, which was then hybridized with a telomere-specific probe. (B) Quantification of telomeric DNA signal in panel (A) as percent of input. (C) Total WCL and chromatin-bound (Ch) levels of TRF1 and TRF2 proteins. Chromatin was extracted after menadione treatment using HEK293E cells containing endogenously FLAG-tagged TRF1 and TRF2 proteins. Western blot was done with an antibody against FLAG tag. Histone H3 and tubulin served as the loading control and as the fractionation control. (D) Telomeric TRF1 level detected by ChIP—dot blot following menadione treatment with or without prior NAC treatment in HeLa cells. (E) Quantification of telomeric DNA signal in panel (D) as fold change relative to untreated sample. (F) Representative figures of immunofluorescence (IF) of mCherry (green) and FISH of telomeric DNA (red), which were employed to assess the level of TRF1 protein at telomeres upon H₂O₂ treatment using HeLa cells containing endogenously mCherry-tagged TRF1. DAPI (blue) was used to stain nuclei. White squares show 15× zoom-in. Scale bar indicates 10 μm. (G) Average integrated density of mCherry foci colocalizing with telomeric DNA foci per nucleus. Each dot represents one nucleus. 150 cells were analyzed per condition, across three independent biological replicates. (H, I) The affinity of TRF1 for oligonucleotides containing non-telomeric dsDNA (non-TelDNA), telomeric dsDNA (TelDNA), and telomeric DNA:RNA hybrids (telomeric hybrids) was analyzed by electrophoretic mobility shift assay (EMSA). ³²P-labelled (asterisks) oligonucleotides were incubated with increasing concentrations of TRF1 protein. The relative dissociation constants (K_d) are shown in panel (I).

Figure 6.

The upregulation of telomeric R-loops upon oxidative stress is TRF2-dependent. (A) Western blot analysis of TRF2 protein levels in HeLa cells transfected with empty vector (EV) or shTRF2 plasmids. Tubulin was used as a loading control. (B) Detection of telomeric TRF1 and TRF2 levels by ChIP-dot blot in wild-type (EV) and TRF2-knockdown (shTRF2) HeLa cells treated with menadione. Immunoprecipitated DNA was blotted onto a membrane, which was then hybridized with a telomere-specific probe. (C) Quantification of telomeric DNA signal in panel (B) as percent of input. (D) Levels of TERRA molecules transcribed from indicated chromosome ends upon menadione treatment were determined by RT-qPCR, using primers binding to chromosome-specific subtelomeric sequences. Data were normalized to GAPDH and shown as fold change relative to untreated sample. (E) Detection of total R-loop level by DRIP-dot blot assay upon menadione treatment in HeLa cells transfected with EV or shTRF2 plasmids. (F) Quantification of telomeric DNA signal in panel (E) as fold change over untreated sample. (*) P < .05, ns—not significant. Data represent mean ± SD. n = 3

Figure 7.

(A) Oxidative stress induces ssDNA damage, which may promote the formation of i-loops at telomeres. At damaged telomeres, TERRA transcription and recruitment are upregulated, leading to increases in R-loop formation both in cis and in trans. This enhanced TERRA recruitment and R-loop formation in trans may be due to the dissociation of TRF1 proteins from damaged telomeres, which alleviates TRF1’s inhibitory effect on TRF2-mediated TERRA invasion into telomeric DNA. (B) Hypothetical positive feedback model representing the potential interplay among TRF1, TRF2, and TERRA R-loops at oxidatively damaged telomeres.