TRF2 Protein Interacts with Core Histones to Stabilize Chromosome Ends

Abstract

Mammalian chromosome ends are protected by a specialized nucleoprotein complex called telomeres. Both shelterin, a telomere-specific multi-protein complex, and higher order telomeric chromatin structures combine to stabilize the chromosome ends. Here, we showed that TRF2, a component of shelterin, binds to core histones to protect chromosome ends from inappropriate DNA damage response and loss of telomeric DNA. The N-terminal Gly/Arg-rich domain (GAR domain) of TRF2 directly binds to the globular domain of core histones. The conserved arginine residues in the GAR domain of TRF2 are required for this interaction. A TRF2 mutant with these arginine residues substituted by alanine lost the ability to protect telomeres and induced rapid telomere shortening caused by the cleavage of a loop structure of the telomeric chromatin. These findings showed a previously unnoticed interaction between the shelterin complex and nucleosomal histones to stabilize the chromosome ends.

Keywords: DNA damage response; GAR domain; chromatin; chromosomes; core histone; histone; telomere.

FIGURE 1.

Interaction between TRF2 and core histones. A, co-immunoprecipitation of TRF2 and core histones. Chromatin-bound and -released fractions were obtained from Myc-tagged TRF2ts-expressing cells (ts) and TRF2^F/− (F/−) parental cells. Released fractions were subjected to co-immunoprecipitation with anti-Myc antibody. Immunoprecipitates were then subjected to immunoblotting with the indicated antibodies. B, co-immunoprecipitation of ΔBasic TRF2 and core histones. The chromatin-released fraction from Myc-tagged TRF2ts-expressing cells and Myc-tagged ΔBasic TRF2ts-expressing cells were subjected to co-immunoprecipitation with anti-Myc antibody. Immunoprecipitates were then subjected to immunoblotting with anti-histone H3 and anti-Myc antibodies. C, co-immunoprecipitation of TRF2 and core histones after removal of DNA and RNA. The chromatin-released fraction from Myc-tagged TRF2ts-expressing cells was treated with DNase I and RNase A, and then subjected to co-immunoprecipitation with anti-Myc antibody. Immunoprecipitates were then subjected to immunoblotting with anti-histone H3 and anti-Myc antibodies. Immunoprecipitation using control IgG served as a negative control. D, Duolink PLA assay of TRF2 and histone H3. The Duolink assay was performed using anti-TRF2 and anti-histone H3 antibodies in HeLa cells. Duolink signals are visualized in red merged with DAPI staining (blue). Duolink using only anti-TRF2 antibody served as a negative control. E, quantification of the Duolink assay in D. Bars represent mean values from three independent experiments ± S.D. (error bars). *, p < 0.05, based on unpaired Student's t test. F, regular immunofluorescence images for TRF2 and histone H3. The merged image is shown with DNA staining using DAPI. G, Duolink PLA assay of TRF2 ΔBasic and histone H3. HeLa cells were transformed to express Myc-tagged full-length TRF2 (TRF2 FL) or TRF2 ΔBasic. Cells were then subjected to Duolink PLA assay using anti-Myc and anti-histone H3 antibodies. Duolink signals are visualized in red merged with DAPI staining (blue). H, quantification of Duolink assay in G. Bars represent mean values from three independent experiments ± S.D. (error bars). *, p < 0.05, n.s., not significant, based on one-way ANOVA with Tukey's test. I, immunoblotting to verify the equal expression of Myc-tagged FL, ΔB TRF2 in HeLa cells. GAPDH served as a loading control. J, regular immunofluorescence images using anti-Myc and anti-histone H3 antibodies to verify equal localization of Myc-TRF2 FL and Myc-TRF2 ΔB. The merged image is shown with DNA staining using DAPI.

FIGURE 2.

Direct binding of the GAR domain of TRF2 and core histones. A, in vitro binding assay for the GAR domain of TRF2 and core histones. Recombinant GST-fused TRF2 GAR domain (GST-Basic) and GST protein (GST) were treated with or without DNase I, RNase A, or both DNase I and RNase A, and then captured by glutathione-conjugated beads and incubated with core histones purified from HeLa cells. Beads were washed extensively and then subjected to SDS-PAGE. The upper panel shows CBB staining of an SDS-PAGE gel. The bottom panel shows immunoblotting using an anti-histone H3 antibody. WB, Western blot. B, complete hydrolysis of the nucleic acid contaminants in the bacterial lysate used for in vitro binding assay in A. Nucleic acid extracted from the lysate was subjected to the agarose gel electrophoresis. The gel was stained with ethidium bromide. C, in vitro binding assay for histone H2A-H2B and H3-H4. GST-Basic protein treated with DNase I and RNase A was captured by glutathione-conjugated beads. Beads were incubated with purified H2A-H2B dimer or H3-H4 tetramer and then subjected to SDS-PAGE after the extensive wash. The top panel shows CBB staining of an SDS-PAGE gel. The middle and bottom panels show immunoblotting for histone H2B and histone H3, respectively. D, in vitro binding assay for tailless core histones. DNase I- and RNase A-treated GST-Basic protein was captured by glutathione-conjugated beads. Beads were incubated with purified core histones or tailless core histones and then subjected to SDS-PAGE after the extensive wash, followed by CBB staining. E, multiple sequence alignments of the basic domain from LANA, human TRF2 (hTRF2), and mouse TRF2 (mTRF2). Amino acid residues required for LANA-histone binding are marked by the line. Conserved arginine residues are marked by asterisks. The result of an in vitro binding assay for GST-Basic deletion mutants and core histones is shown in the middle. The residues that were changed to alanine or lysine are marked in red (Basic RA and Basic RK). F, in vitro binding assay for GST-Basic mutants and core histones. GST-fused Basic, GST-Basic RA, and GST-Basic RK proteins treated with DNase I and RNase A were captured by glutathione-conjugated beads. Beads were incubated with purified core histones and then subjected to SDS-PAGE after the extensive wash. The top panel shows CBB staining of an SDS-PAGE gel, and the bottom panel shows immunoblotting for histone H3.

FIGURE 3.

Interaction between the GAR domain of TRF2 and core histones in vivo. A, chromatin bound Basic-NLS-EGFP in metaphase. NLS-EGFP control, Basic-NLS-EGFP, and Basic-RA-NLS-EGFP were introduced into NIH-3T3 cells. Images of metaphase cells were captured with a fluorescence microscope. The top panel shows the EGFP signal, the middle panel shows DAPI staining pseudo-colored in red, and the bottom panel shows the merged image. B, nuclear localization of Basic-NLS-EGFP after nucleoplasmic pre-extraction. NIH-3T3 cells expressing NLS-EGFP, Basic-NLS-EGFP, and Basic-RA-NLS-EGFP were extracted in situ with the nucleoplasm removal buffer. The images of DAPI staining and EGFP signal were captured with a fluorescence microscope. C, fluorescence microscopic images of Basic-NLS-EGFP, Basic-RA-NLS-EGFP, and NLS-EGFP to verify the equal nuclear localization in interphase NIH-3T3 cell before nucleoplasmic pre-extraction. D, co-immunoprecipitation of Basic-NLS-EGFP and core histones. NIH-3T3 cells expressing NLS-EGFP, Basic-NLS-EGFP, and Basic-RA-NLS-EGFP were extracted with radioimmunoprecipitation assay buffer. Cell extracts were subjected to co-immunoprecipitation with an anti-GFP antibody. Immunoprecipitants were then applied to SDS-PAGE. Silver staining (top) and immunoblotting for histone H3 (middle) and EGFP (bottom) are presented. WB, Western blot. E, Duolink PLA assay of TRF2 mutants and histone H3. Myc-tagged TRF2 FL-, ΔB-, and RA-expressing MEFs were subjected to Duolink PLA assay using anti-Myc and anti-histone H3 antibodies. Duolink signals are visualized in red merged with DAPI staining (blue). F, quantification of Duolink assay in E. Bars represent mean values from three independent experiments ± S.D. (error bars). **, p < 0.01, ***, p < 0.001, n.s., not significant, based on one-way ANOVA with Dunnett's test. G, IF-FISH images to verify the equal telomere localization of TRF2 FL, ΔB, and RA. Myc-tagged TRF2 FL-, ΔB-, and RA-expressing MEFs were subjected to IF-FISH. Immunofluorescence for Myc (red), FISH using telomeric TTAGGG probe (green), and merged images with DAPI (blue) are presented.

FIGURE 4.

Loss of telomere protection in TRF2-histone interaction-deficient cells. A, schematic of the experiment for TRF2 complementation in MEFs with TRF2 mutant alleles. TRF2^F/− MEFs were retrovirally introduced with a wild-type TRF2 (FL) or mutant (ΔB and RA) alleles and then were induced with Cre-mediated deletion of TRF2 flox allele. B, immunoblotting to verify the equal expression of Myc-tagged FL, ΔB, and RA mutated TRF2. GAPDH served as a loading control. C, PCR to verify Cre-mediated deletion of the endogenous allele in TRF2 FL-, ΔB-, and RA-complemented MEFs. D, TIF induction in TRF2 RA cells. TRF2 FL-, ΔB-, RA-, and RK-complemented MEFs were subjected to a TIF assay. Immunofluorescence for γH2AX (red), FISH using telomeric TTAGGG probe (green), and merged images with DAPI (blue) are presented. The white dotted boxes indicate the site of the magnified images. E, quantification of the TIF assay in D. Bars represent mean values from three independent experiments ± S.D. (error bars). ***, p < 0.001, n.s., not significant, based on one-way ANOVA with Dunnett's test. F, cell cycle-specific induction of TIF in TRF2 RA cells. TRF2 RA mutant-complemented MEFs were co-immunostained for cyclin A (magenta) and γH2AX (red), and FISH detection with telomeric TTAGGG probe (green) was performed. An arrow indicates a cyclin A-positive cell. G, quantification of the TIF assay in F. Bars represent mean values from three independent experiments ± S.D. (error bars). ***, p < 0.001, based on unpaired Student's t test. H, immunofluorescence image for BrdU (green) and 53BP1 (red) of TRF2 RA-complemented MEFs. DNA was visualized with DAPI (blue). I, proliferation of IMR90 cells expressing TRF2 FL, TRF2 RA, and TRF2 ΔB. Cells were retrovirally transfected with the indicated cDNA. The growth curves after drug selection are presented. Data represent mean values from three independent experiments ± S.D. (error bars). J, SA-β-gal assay of IMR90 cells overexpressing TRF2 FL, TRF2 RA, and TRF2 ΔB at day 10 after drug selection. K, immunoblotting to verify the equal expression of TRF2 FL, TRF2 RA, and TRF2 ΔB in IMR90 cells. GAPDH served as a loading control.

FIGURE 5.

Rapid telomere DNA loss and t-circle generation by loss of histone binding of TRF2. A, metaphase FISH analysis using telomeric TTAGGG probe (green) in TRF2 FL, TRF2 RA, and TRF2 ΔB cells. DNA was visualized with DAPI as a pseudo-colored red. Arrows and arrowheads indicate chromosomes with telomeric signal free ends. The white boxes indicate the sites of the magnified images (bottom right). B, quantification of metaphase FISH analysis in A. Bars represent mean values from three independent experiments ± S.D. (error bars). *, p < 0.05, n.s., not significant, based on one-way ANOVA with Dunnett's test. C, loss of telomere signals in TRF2 mutants. TRF2^F/− MEFs expressing TRF2 wt, ΔB, and RA mutant were harvested 14 days after Cre treatment and subjected to Southern blotting of terminal restriction fragments using a DIG-labeled telomeric TTAGGG probe (left). The numbers below the lane represent the relative telomeric signal normalized by the major satellite probe signal (right). D, rapid telomeric DNA loss in TRF2 RA-complemented MEFs. The cells were harvested at the indicated time point after Cre treatment and subjected to Southern blotting using a DIG-labeled telomeric TTAGGG probe (bottom). DNA visualized with EtBr (top) served as a loading control. E, t-circle production in TRF2 RA-complemented MEFs detected by t-circle amplification assay. Phi29-dependent t-circles (TCs) were detected 4 days after Cre treatment of TRF2^F/− MEFs expressing wt-, ΔB-, and RA-mutated TRF2 cDNA. F, t-circle production in TRF2 RA-complemented MEFs detected by 2D Southern blotting. Cells were harvested 5 days after Cre treatment. Equal amounts of MboI-digested genomic DNA were subjected neutral 2D agarose electrophoresis. Telomeric DNA was visualized by Southern blotting using a DIG-labeled TTAGGG probe. Arrows indicate t-circles. G, quantification of 2D Southern blotting in F. Bars represent mean values from three independent experiments ± S.D. (error bars). **, p < 0.01, n.s., not significant, based on one-way ANOVA with Tukey's test. H, method for quantification of circular-to-linear telomere DNA ratio in G. The signal intensity and area of 2D Southern blotting were measured using the ImageJ software. The yellow dotted area (a) was used to quantify the circular telomere DNA, and the magenta dotted area (b) was used to quantify the linear telomere DNA.

FIGURE 6.

Regain of telomere protection by forced dimerization of TRF2 and core histone, independent of the GAR domain. A, schematic of the experiment for A/C dimerizer-induced dimerization of FRB-fused histone H2B and FKBP-fused TRF2 ΔB. B, Duolink PLA assay to detect the dimerization between histone H2B-FRB and FKBP-TRF2 ΔB. FKBP-TRF2 ΔB-complemented MEFs expressing histone H2B-FRB were incubated with A/C dimerizer for 12 h and then subjected to Duolink PLA assay using antibody against HA (H2B-FRB) and Myc (FKBP-TRF2 ΔB). C, immunoblotting to verify the equal expression of Myc-tagged FKBP-TRF2 FL and FKBP-TRF2 ΔB and HA-tagged histone H2B-FRB. Nonspecific (n.s.) signal served as a loading control. D, PCR to verify Cre-mediated deletion of the endogenous allele from TRF2^F/− MEFs expressing H2B-FRB transformed to express FKBP-TRF2 FL and FKBP-TRF2 ΔB. E, IF-FISH images to verify equal localization of Myc-tagged FKBP-FL and FKBP-ΔB TRF2 at telomeres. IF-FISH co-staining results using anti-Myc (red) and anti-histone H3 (magenta) antibodies in conjunction with FISH with a telomeric TTAGGG-specific probe (green) are presented. F, TIF assay in histone H2B-FRB and FKBP-TRF2 ΔB dimerized cells. FKBP-TRF2 ΔB-complemented MEFs expressing histone H2B-FRB were incubated with A/C dimerizer for 4 h and then subjected to a TIF assay. Immunofluorescence results for γH2AX (top), FISH using telomeric TTAGGG probe (middle), and merged images with DAPI (bottom) are presented. G, quantification of TIF assay in F. Bars represent mean values from three independent experiments ± S.D. (error bars). ***, p < 0.001, n.s., not significant, based on two-way ANOVA with Sidak's test. H, inhibition of t-circle production by dimerization of histone H2B and TRF2 ΔB. FKBP-TRF2 ΔB-complemented MEFs expressing histone H2B-FRB were incubated with A/C dimerizer for 48 h and then subjected to t-circle amplification assay. The numbers below the lane represent the relative t-circle signal (TCs) to A/C dimerizer negative control. I, t-circle detection in histone H2B-FRB and FKBP-TRF2 ΔB dimerized cells by 2D Southern blotting. FKBP-TRF2 ΔB-complemented MEFs expressing histone H2B-FRB were harvested 48 h after the addition of A/C dimerizer. Equal amounts of MboI-digested genomic DNA were subjected neutral 2D agarose electrophoresis. Telomeric DNA was visualized by Southern blotting using a DIG-labeled TTAGGG probe. An arrow indicates t-circles.