SIRT6 recruits SNF2H to DNA break sites, preventing genomic instability through chromatin remodeling

Abstract

DNA damage is linked to multiple human diseases, such as cancer, neurodegeneration, and aging. Little is known about the role of chromatin accessibility in DNA repair. Here, we find that the deacetylase sirtuin 6 (SIRT6) is one of the earliest factors recruited to double-strand breaks (DSBs). SIRT6 recruits the chromatin remodeler SNF2H to DSBs and focally deacetylates histone H3K56. Lack of SIRT6 and SNF2H impairs chromatin remodeling, increasing sensitivity to genotoxic damage and recruitment of downstream factors such as 53BP1 and breast cancer 1 (BRCA1). Remarkably, SIRT6-deficient mice exhibit lower levels of chromatin-associated SNF2H in specific tissues, a phenotype accompanied by DNA damage. We demonstrate that SIRT6 is critical for recruitment of a chromatin remodeler as an early step in the DNA damage response, indicating that proper unfolding of chromatin plays a rate-limiting role. We present a unique crosstalk between a histone modifier and a chromatin remodeler, regulating a coordinated response to prevent DNA damage.

Figure 1. SIRT6 interacts with SNF2H and recruits it to chromatin

A) Flag-IP of protein extracts from Flag-SIRT6- or empty vector-expressing cells was used for mass spectrometry analysis where SNF2H was identified as SIRT6 interactor. Silver-staining of extracts from Flag-control (Vector) or SIRT6-Flag transfected cells is shown. B–C) Exogenous (B) and endogenous (C) co-IPs were performed, and western blots were developed with the indicated antibodies. Flag-vector (Ctrl), Flag-WT-SIRT6 (WT) and Flag-HY-SIRT6 (HY) transfected cells. D) IP for SIRT6 Flag (WT), SIRT6-HY (HY), SIRT6 fragments lacking N (ΔN), C (ΔC) or both termini (Core). E) GST-Full Length SIRT6 and SIRT6 C-terminus, N-terminus or Core domain were tested in vitro for interaction with baculovirus purified SNF2H. F) Glycerol gradient fractionation for WT cells showing SIRT6, SNF2H and H3 bands. G) Fractions 7–9 from WT and 7–9 from SIRT6-KO cells. H) Nucleosome shift assay was performed with 75nM nucleosomes, SIRT6 (μM range) and SNF2H (nM range). Marked with the arrows are the different complexes formed with the nucleosomes under SIRT6/SNF2H incubation, or both proteins. See also Fig. S1.

Figure 2. SIRT6 and SNF2H act together in DNA repair

A) Whole cell extracts (WCE) and chromatin fractions from WT and SIRT6 KO cells, showing impaired SNF2H recruitment to chromatin. Quantification of SNF2H levels within the chromatin fraction in three independent experiments is shown, B) SIRT6-WT and SIRT6-HY mutant were expressed in 293T cells and the presence of SNF2H in WCE and chromatin was analyzed by western blot. Quantification of SNF2H levels within the chromatin fraction in three independent experiments is shown. C) Western blot with the indicated antibodies was performed using chromatin fractions of NAM-treated or untreated cells. D) Cells were irradiated and Flag-SIRT6 immunoprecipitated 15 or 30 minutes after IR exposure. Cohesin protein SMC1 is shown as a SIRT6 interactor that was not affected by damage as a control. E) SNF2H expression was silenced in Sirt6^+/+ and Sirt6^−/− ES cells using two different shRNA sequences. F) Survival assays upon IR of ES cells of the indicated genotypes. Data is represented as mean +/− SEM. (right panel, * p< 0.01). See also Fig. S2.

Figure 3. Kinetics of SIRT6 and SNF2H recruitment to DSBs

A) Laser-induced DNA damage was performed in cells transfected with GFP or GFP-SIRT6, fixed at indicated time-points, and immunostained with anti-γH2AX and anti-GFP antibodies. B) Laser induced DNA damage in U2OS shCtrl or shSIRT6 cells immunostained with γH2AX and SNF2H at different time points. Graph shows time-course analysis. SNF2H-positive cells among γH2AX positive cells were quantified. C) RFP-SIRT6 and SNF2H-GFP recruitment was measured on sites of damage, and followed over time with live cell imaging. A representative series is shown. D) Quantitative analysis of the recruitment experiments described in (C). E) Analysis of SNF2H recruitment in shCtrl versus shSIRT6 cells. Fluorescence intensity is normalized as 1=maximum intensity reached in each case after 30 minutes. F) Graphic representation of the early- and late-recruitment phases of the graphs shown in (E). Data is represented as mean +/− SEM.. *p< 0.05, **p<0.001. See also Fig. S3.

Figure 4. SIRT6 modulates SNF2H-dependent DSBs chromatin opening and repair

A) Schematic representation of the DR-GFP/ISce-I system. B) Left panel: ChIP of cells transfected with Flag-cmv empty vector, or Flag-S6, with or without I-SceI transfection. Middle panel: ChIP of SNF2H in shCtrl or shSIRT6 cells with or without Isce-I transfection. Right panel: Sequential ChIP from Flag-SIRT6 eluted chromatin (ctrl or I-sceI treated), where SNF2H or IgG were used for the second ChIP. C-F) HR efficiency measured by GFP positive cells in (C) shCtrl vs shSIRT6, (D) NAM treated cells (10mg/ml, 12hrs). E) SIRT6-KD cells transfected with either WT-SIRT6, the catalytic mutant SIRT6-HY (catalytically inactive) or the C-terminus deleted (non-SNF2H interacting) SIRT6 fragments, and (F) shCtrl and shSIRT6 cells pre-treated for two hours with chloroquine. G) Chromatin accessibility at DNA breaks. Scheme of the experiment: DSBs were induced with the I-SceI endonuclease, nuclei were isolated and digested with MNase. Different nucleosomal fractions (mono, di and upper) were separated on an agarose gel, and the abundance of the I-SceI site in the isolated DNA of each fraction was quantified using specific primers adjacent to the breaks. (H) qPCR of the isolated DNA from the nucleosomes with primers adjacent to the site of damage in shCtrl cells vs shSIRT6 cells. Data is represented as mean +/− SEM. *p< 0.05, **p<0.01, ***p<0.001. See also Fig. S4.

Figure 5. SIRT6 modulates H3K56 deacetylation and recruitment of repair factors at DSBs

A) Immunofluorescence showing H3K56ac at sites of damage after laser induced damage in shCtrl and shSIRT6 U2OS cells. H3K56Ac levels were measured at and besides damage sites, data is represented as mean +/− SEM.. B) Quantification of GFP positive U2OS-DR-GFP cells transfected with H3-WT or H3K56Q mutant. C–E) High-throughput analysis of foci number showing (C) 53BP1, (D) RPA and (E) γH2AX foci number per cell at different time points after IR in shCtrl vs shSIRT6 U2OS cells. F) Comet tail length for shCtrl and shSIRT6 is quantified at the indicated time-points. Representative pictures are shown (15 min. time-point). Data is represented as mean +/− SEM. p-values are abbreviated as in Figure 4. See also Fig. S5.

Figure 6. Decreased recruitment of repair factors to laser-induced breaks in the absence of SIRT6

Following laser-induced damage, recruitment of (A) ATM-P, (B) RPA, (C) 53BP1 and (D) BRCA1 was quantified in the indicated genotypes, data is represented as mean +/− SEM. See also Fig. S6.

Figure 7. SIRT6 modulates SNF2H recruitment and DNA repair in vivo

A–B) Chromatin fractions or whole cell extracts from brains of 18 or 22 days old mice from WT or SIRT6 KO animals with the noted antibodies C–E) Primary brain cells cultured for 14 days before damage and collected at different time points (shown is 30min post IR) for RPA and 53BP1. F) Proposed model: SIRT6 is mobilized very early to the sites of damage, recruiting SNF2H and deacetylating H3K56, allowing opening of chromatin and recruitment of downstream repair factors, such as 53BP1, RPA and BRCA1. Data is represented as mean +/− SEM. *p< 0.05, **p<0.001. See also Fig. S7.