Telomere length-dependent transcription and epigenetic modifications in promoters remote from telomere ends

Abstract

Telomere-binding proteins constituting the shelterin complex have been studied primarily for telomeric functions. However, mounting evidence shows non-telomeric binding and gene regulation by shelterin factors. This raises a key question-do telomeres impact binding of shelterin proteins at distal non-telomeric sites? Here we show that binding of the telomere-repeat-binding-factor-2 (TRF2) at promoters ~60 Mb from telomeres depends on telomere length in human cells. Promoter TRF2 occupancy was depleted in cells with elongated telomeres resulting in altered TRF2-mediated transcription of distal genes. In addition, histone modifications-activation (H3K4me1 and H3K4me3) as well as silencing marks (H3K27me3)-at distal promoters were telomere length-dependent. These demonstrate that transcription, and the epigenetic state, of telomere-distal promoters can be influenced by telomere length. Molecular links between telomeres and the extra-telomeric genome, emerging from findings here, might have important implications in telomere-related physiology, particularly ageing and cancer.

Fig 1. Significantly altered TRF2 occupancy at non-telomeric sites in cells with short versus long telomeres.

A. Distance of selected promoters from nearest telomere end. Gene promoters with TRF2 binding sites (within 1.5 kb of transcription start sites (TSS)) selected from replicate ChIP-seq studies (raw data publicly available-SRA 304653) and published reports of extra-telomeric occupancy of TRF2. B-C. Promoter occupancy of TRF2. TRF2 occupancy at gene promoters was checked by ChIP-qPCR in HT1080 (B) and MRC5 cells (C). CTCF and GAPDH promoters with no TRF2 binding within +/- 5 kb of TSS were used as negative controls. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. D-E. TRF2 silencing transcriptionally affects gene expression. Effect of TRF2 silencing on gene promoters was tested in HT1080 (D) and MRC5 cells (E). CTCF was used as a negative control gene; normalized with respect to GAPDH expression. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. F-G. HT1080-LT cells with long telomeres have more telomeric TRF2 occupancy in comparison to HT1080 cells. ChIP with TRF2 antibody (or isotypic control) was followed by PCR using telomere-specific primer (TEL-PCR) in HT1080 and HT1080-LT cells (F). Input samples and TEL-PCR products were blotted on membrane and hybridized with telomere-specific probes. Quantification of three independent dot blot assays (G). Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. H. Significantly reduced TRF2 occupancy at gene promoters in cells with long telomeres. TRF2 occupancy at many promoters was lower in HT1080-LT cells compared to HT1080 cells. CTCF and GAPDH promoters with no TRF2 binding within +/- 5 kb of TSS were used as negative controls. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01.

Fig 2. Altered promoter TRF2 occupancy in normal fibroblast cells with long or short telomeres.

A-B. Telomere elongation. Telomere elongation in MRC5 cells was confirmed by flow-FISH using FITC-tagged telomere-specific probes following treatment with G-rich terminal repeat (GTR) oligonucleotides for either 7 (MRC5-OF7), 14 (MRC5-OF14) cycles or without treatment (MRC5) (A); also see Supplementary S2A Fig. Relative telomere length was quantified by three independent flow-FISH experiments (B). Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. C-D. MRC5 cells with long telomeres have more telomeric TRF2. ChIP with TRF2 antibody (or isotypic control) was followed by PCR using telomere-specific primers in MRC5 cells. Input samples and the TEL-PCR products were blotted on a membrane and hybridized with telomere-specific probes—dot blot assay showing cells with long telomeres have enriched TRF2 occupancy at telomeres in MRC5-OF7 and MRC5-OF14 cells (C). Quantification of three independent dot blot assays (D). Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. E. Significantly lower TRF2 occupancy at promoters in cells with long telomeres. TRF2 occupancy at multiple promoter sites was reduced in MRC5-OF7 and OF14 cells in comparison to untreated cells. CTCF and GAPDH promoters with no TRF2 binding within +/- 5 kb of TSS were used as negative controls. Error bars indicate ± SD from three independent experiments. Significance was tested by paired T-test -* <0.05; **<0.01.

Fig 3. Transcription activity of p21 is altered in cells with short vis-à-vis long telomeres.

A-B. p21 expression in HT1080-LT and MRC5 cells. p21 promoter activity, mRNA and protein levels increased in HT1080-LT relative to HT1080 cells (A) and MRC5-OF7/14 compared to untreated MRC5 cells (B). Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. C. Altered transcription of p21 through loss of REST and LSD1 from the p21 gene promoter. Loss of REST and LSD1 occupancy from p21 promoter in HT1080-LT as compared to HT1080 cells. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01. D. Histone modifications at the p21 promoter in cells with short/long telomeres. HT1080-LT cells with elongated telomeres had enrichment of activating histone marks, H3K4me1 and H3K4me3, and reduction in the suppression histone mark H3K27me3. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01.

Fig 4. Transcription of many genes altered in cells with short compared to long telomeres.

A-B. Differential mRNA expression of genes with altered promoter occupancy of TRF2 in cells with short/long telomeres. mRNA expression was checked by qPCR for genes in which TRF2 promoter occupancy was reduced in HT1080-LT cells compared to HT1080 cells (A), and MRC5-OF7 or OF14 cells compared to untreated MRC5 cells (B). CTCF was used as a negative control. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01.

Fig 5. Promoter epigenetic modifications altered in cells with short versus long telomeres.

A-C. Promoter occupancy H3K4me1 (A), H3K4me3 (B) and H3K27me3 (C) in HT1080-LT relative to HT1080 cells. Error bars indicate ± SD from three independent experiments; significance was tested by paired T-test -* <0.05; **<0.01.

Fig 6. Telomere sequestration-partition model–non-telomeric versus telomeric TRF2 binding in cells with long vis-à-vis short telomeres.

In cells with long telomeres increase in telomere-bound TRF2 restricts non-telomeric TRF2 occupancy. On the other hand, in cells with short telomeres, more TRF2 is available for binding at non-telomeric promoter sites. Increased TRF2 binding at promoters in cells with short vis-à-vis elongated telomeres result in altered chromatin histone modifications and influence transcription.