SIRT6 Is Responsible for More Efficient DNA Double-Strand Break Repair in Long-Lived Species

Abstract

DNA repair has been hypothesized to be a longevity determinant, but the evidence for it is based largely on accelerated aging phenotypes of DNA repair mutants. Here, using a panel of 18 rodent species with diverse lifespans, we show that more robust DNA double-strand break (DSB) repair, but not nucleotide excision repair (NER), coevolves with longevity. Evolution of NER, unlike DSB, is shaped primarily by sunlight exposure. We further show that the capacity of the SIRT6 protein to promote DSB repair accounts for a major part of the variation in DSB repair efficacy between short- and long-lived species. We dissected the molecular differences between a weak (mouse) and a strong (beaver) SIRT6 protein and identified five amino acid residues that are fully responsible for their differential activities. Our findings demonstrate that DSB repair and SIRT6 have been optimized during the evolution of longevity, which provides new targets for anti-aging interventions.

Keywords: DNA DSB repair; DNA repair; NER; SIRT6; aging; longevity.

Figure 1. NER efficiency does not correlate with maximum lifespan in 18 rodent species.

(A) Host cell reactivation assay to measure NER efficiency. NER efficiency was calculated as the percentage of the luciferase reactivation and the formula is shown. (B, C) Luciferase reactivation of UV-damaged FFL plasmid at (B) 400 J/m² or (C) 1200 J/m² does not correlate with species MLS (also see Figure S2C, D for no correlation to body mass). Cells isolated from at least 3 different animals of each species were used and each circle represents a species. Error bars indicate s.e.m., r, Pearson correlation coefficient. p-values were determined by t-test. (D and E) Luciferase reactivation of UV-damaged FFL plasmid at (D) 400 J/m² or (E) 1200 J/m² is significantly different across the three sunlight exposure groups. The data from Figure 1B and 1C were used to plot against the level of sunlight exposure. Classification of the three sunlight exposure groups was based on the lifestyle and habitat of the species (Table S1). Statistical significance was tested using one-way ANOVA. Data was phylogenetically corrected. (F) UV LD50 is significantly different across the three sunlight exposure groups. One-way ANOVA was used to test statistical significance. Data was phylogenetically corrected. See also Table S1, Figure S1, S2.

Figure 2. The efficiency of DNA DSB repair positively correlates with maximum lifespan.

(A) NHEJ repair reporter construct. The GFP coding sequence containing an intron from the rat Pem1 gene is interrupted by an adenoviral exon (Ad). Splicing Ad into the GFP gene makes it inactive. Ad is flanked by I-SceI recognition sites in inverted orientation. Induction of a DSB by I-SceI enzyme followed by successful NHEJ reconstitutes a functional GFP gene. SD, splice donor; SA, splice acceptor. (B) HR repair reporter construct. 22 nucleotides of the GFP coding sequence was replaced with two I-SceI recognition sites in inverted orientation. This 22 nt deletion ensures that GFP cannot be reconstituted by an NHEJ repair event. An HR repair template, which lacks the ATG start codon and the second exon of GFP, is placed after the GFP coding sequence. Induction of a DSB by I-SceI enzyme followed by a gene conversion event reconstitutes a functional GFP gene. (C) DNA repair substrate. Two adjacent I-SceI cuts generate non-compatible DNA ends. (D-G) NHEJ repair efficiency (D, E) and HR repair efficiency (F, G) correlate with species maximum lifespan. NHEJ and HR reporter constructs were integrated into primary, low passage skin and lung fibroblasts. DNA DSBs repair efficiency was assayed by transfecting cells with I-SceI expression vector and a DsRed plasmid as a transfection control (Seluanov et al., 2010a). The repair efficiency was calculated as the ratio of GFP⁺/DsRed⁺ cells. Experiments were repeated at least 3 times for each cell line. Cells from at least 3 animals were assayed for each species. Error bars represent s.e.m. r, Pearson correlation coefficient. p-values were determined by t-test for the uncorrected data; the values for phylogenetically corrected data are shown in the text. See also Figure S3, S4.

Figure 3. Alignment of rodent SIRT6 proteins.

(A) Alignment of rodent SIRT6 proteins indicated by amino acid similarity and absolute complexity. The alignment of SIRT6 from 18 rodent species (Figure S1) was performed with AlignX module in Vector NTI software where the blosum62mt2 scoring matrix was used. (B) Alignment of SIRT6 protein sequences from 18 rodent species. The alignment was performed with ClustalW multiple sequence alignment program. Dots indicate amino acids identical to the mouse sequence. The SIRT6 sequences were placed in the order of increasing maximum lifespan of the species.

Figure 4. SIRT6 ability to stimulate DSB repair positively correlates with species maximum lifespan.

(A) The formula to calculate the fold of DSB repair stimulation by SIRT6. Plasmids overexpressing (OE) HPRT (control) or SIRT6 were co-transfected with the I-SceI and DsRed plasmids, and the fold stimulation for both NHEJ and HR was calculated as shown. (B) SIRT6 ability to promote NHEJ positively correlates with species maximum lifespan. SIRT6 of rodent species were expressed in mouse fibroblasts containing chromosomally integrated NHEJ reporter cassette (Figure 2A). (C) SIRT6 ability to promote HR positively correlates with species maximum lifespan. SIRT6 from 18 rodent species were expressed in rat fibroblasts containing chromosomally integrated HR reporter cassette (Figure 2B). All experiments in these figures were repeated at least 3 times and error bars show s.d. r, Pearson correlation coefficient. p-values were determined by t-test for uncorrected data; the values for phylogenetically corrected data are shown in the main text. Correlation between SIRT6-mediated stimulation of DSB repair and DNA body mass is shown in Figure S5A and S5B. (D) Western blot on the mouse cells expressing flag-tagged SIRT6 cDNAs from the corresponding species, probed with antibodies to flag tag.

Figure 5. Five amino acid substitutions are responsible for the differential activities between mouse and beaver SIRT6.

Chimeric mouse and beaver SIRT6 proteins were tested for their ability to stimulate NHEJ in mouse fibroblasts containing chromosomally integrated NHEJ reporter construct (Figure 2A). Lower panels show Western blots on the mouse cells expressing SIRT6 proteins, probed with an antibody to SIRT6 (Abcam, ab62739). The epitope that the antibody recognizes is the same between mouse and beaver SIRT6 and their mutants. The following nomenclature of the chimeric proteins is used throughout the figure: m corresponds to mouse; b corresponds to beaver. mWT corresponds to mouse WT SIRT6; bWT corresponds to beaver WT SIRT6; mNXbC is a SIRT6 protein with mouse N-terminal portion and beaver C-terminal portion starting from amino acid X; m(bX) is a mouse SIRT6 protein where aa X was replaced with a corresponding beaver amino acid; b(mX, Y) is a beaver SIRT6 protein where amino acids X and Y were replaced with the corresponding mouse amino acids. (A) Swapping C-terminal regions between mouse and beaver SIRT6. (B) Swapping amino acid 221–270 region between mouse and beaver SIRT6 is sufficient to fully swap the mouse and beaver SIRT6 activities. (C) SIRT6 sequence between amino acids 220–270. The differences between mouse and beaver are indicated. (D) Introducing any single amino acid substitution into the mouse SIRT6 corresponding to the beaver amino acid sequence in the 221–270 region is not sufficient to increase the mouse SIRT6 activity to the level of the beaver SIRT6. (E) Introducing any single amino acid substitution into the beaver SIRT6 corresponding to the mouse amino acids in 221–270 region is not sufficient to decrease the beaver SIRT6 activity to the level of the mouse SIRT6. (F) Replacing amino acids 249 and 263 together is not sufficient to swap the mouse and beaver SIRT6 activities. (G) Replacing amino acids 235, 260, and 264 is not sufficient to swap the mouse and beaver SIRT6 activities. (H) None of the quadruple amino acid replacements in the mouse SIRT6 elevate its activity to the level of the beaver SIRT6 or to the level of the quintuple replacement that involves the entire 221–270 region m(b221–270). (I) None of the quadruple amino acid replacements in the beaver SIRT6 reduce its activity to the level of the mouse SIRT6 or to the level of the quintuple replacement that involves the entire 221–270 region b(m221–270). All experiments were repeated at least three times and error bars show s.d. Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test. NS, not significant; *, p < 0.05, **, p < 0.01; ***, p < 0.001. See also Figure S5.

Figure 6. The five amino acid changes confer the beaver SIRT6 higher deacetylation and mono-ADP-ribosylation activities relative to the mouse SIRT6.

(A) Deacetylation assays with mouse and beaver SIRT6 proteins. Recombinant mouse WT, mouse WT with five beaver substitutions (mouse 5mut), beaver WT, and beaver with five mouse amino acid substitutions (beaver 5mut) SIRT6 were incubated with purified nucleosomes. The acetylation status of SIRT6 targets was assayed with acetyl-specific antibodies. The graphs on the right show quantification. The experiments were repeated 3 times and error bars show s.d. Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test. NS, not significant; *, p < 0.05, **, p < 0.01. (B) Self mono-ADP-ribosylation assays with mouse and beaver SIRT6 proteins. (C) PARP1 activation assay. SIRT6 proteins were incubated with PARP1, NAD, and linearized DNA. Poly-ADP ribose (PAR) was visualized with PAR-specific antibodies. (D-E) Structural analysis of the five amino acids in SIRT6. (D) amino acids 235 and 249 and (E) amino acids 260, 263, and 264 localize on the surface of the protein. (F) A closer view of amino acids 235 and 249. Both amino acids are located in the vicinity of H3K9 myristoyl peptide. (G) A closer view of amino acids 260, 263, and 264. Interactions of amino acid side chains between amino acids 260 and 264 in the mouse SIRT6 are altered in the beaver SIRT6. (H) Nucleosome binding by SIRT6. SIRT6 proteins were incubated with purified HeLa nucleosomes, immunoprecipitated with SIRT6 antibodies, and probed with antibodies against H3. See also Figure S6.

Figure 7. Beaver SIRT6 protects from stress-induced senescence and extends lifespan better than mouse SIRT6.

(A) Expression of the mouse, beaver, mouse 5mut, and beaver 5mut SIRT6 proteins in the SIRT6 knockout human dermal fibroblasts. (B) Clonogenic assay after serial doses of γ-irradiation in cells expressing GFP (control) or SIRT6 proteins. Linear-quadratic model was used to fit the survival data (Figure S7D) and calculate the LD50 and LD75. (C) Representative images of senescence-associated β-galactosidase (SA β-gal) staining in SIRT6 knockout dermal fibroblasts stably expressing mouse, beaver, or the corresponding mutant SIRT6 proteins after 5 Gy of γ-radiation. (D) Quantification of SA β-gal positive cells. Images from different areas of the plate were taken and quantified (n=10). Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test. NS, not significant; *, p < 0.05, **, p < 0.01; ***, p < 0.001, ****, p < 0.0001. (E-F) Beaver SIRT6 extends Drosophila lifespan better than mouse SIRT6. Drosophila transgenic lines expressing mouse WT, mouse 5mut, beaver WT, beaver 5mut SIRT6, and empty vector (EV) control were used to measure lifespan. RU486 was used to induce the expression of SIRT6 proteins. See Figure S7F–H for the female results. (G) Summary of the parameters and statistics of the lifespan measurement. Survival data was analyzed using OASIS 2. (H) Western blot showing expression of the mouse, beaver, and their mutant SIRT6 proteins in Drosophila transgenic lines. See also Figure S7.