Tier 3 — preclinical

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

Tian X, Firsanov D, Zhang Z, Cheng Y, Luo L, Tombline G, Tan R, Simon M, Henderson S, Steffan J, Goldfarb A, Tam J, Cornwell A, Johnson A, Yang JN, Mao Z, Manta B, Dang W, Zhang Z, Vijg J, Wolfe A, Moody K, Kennedy BK, Bohmann D, Gladyshev VN, Seluanov A, Gorbunova V

Cell 2019 Volume 177, issue 3, pages 622–638.e22

Bibliography

DOI: 10.1016/j.cell.2019.03.043
PubMed: PMID 31002797
PubMed Central: PMC6499390
Funding: US National Institutes of Health grants to V.G., A.S., V.N.G., Z.Z., and J.V., and Life Extension Foundation grants to V.G. and A.S.
Competing interests: The authors declared no competing interests.

Study snapshot

Design	Comparative interspecies analysis across 18 rodent species using cell-based DNA repair reporter assays; mouse–beaver SIRT6 chimeric mutagenesis; in vitro enzymatic characterisation; Drosophila lifespan validation of species-specific SIRT6.
Model	Primary skin and lung fibroblasts from 18 rodent species with maximum lifespans of 3–32 years (including mouse, rat, gerbil, hamster, deer mouse, chinchilla, guinea pig, chipmunk, red squirrel, gray squirrel, woodchuck, porcupine, paca, capybara, blind mole rat, naked mole rat, and beaver); SIRT6-knockout mouse embryonic fibroblasts; SIRT6-knockout human dermal fibroblasts; UAS-Sirt6 transgenic Drosophila melanogaster on the Actin-GeneSwitch-Gal4 driver.
Sample	DSB repair assays run in at least 3 independent experiments per species per cell line, pooled from ≥100 stable clones per line to control for integration site effects. NHEJ efficiency and HR efficiency reporters integrated into skin and lung fibroblasts from all species. Drosophila lifespan cohorts of approximately 340–366 flies per genotype per sex, 6 bottles of 50–60 flies each.
Intervention	Chromosomal integration of GFP-based NHEJ and HR reporter constructs; overexpression of species-specific SIRT6 cDNAs in mouse and rat cells; site-directed mutagenesis of 5 identified amino acid residues (235, 249, 260, 263, 264) between mouse and beaver SIRT6; shRNA knockdown of SIRT6 in beaver fibroblasts; lentiviral expression of mouse WT, mouse-5mut, beaver WT, and beaver-5mut SIRT6 in SIRT6-knockout human dermal fibroblasts; RU486-induced transgenic Drosophila expression of species-specific SIRT6.
Duration	3-day WST-1 cell survival readouts after UV irradiation; 12-day senescence-associated β-galactosidase assay after 5 Gy γ-radiation; Drosophila lifespan measurements to natural death (approximately 70–80 days).
Endpoints	Non-homologous end joining efficiency (GFP+/DsRed+ ratio); homologous recombination efficiency; nucleotide excision repair measured by host cell reactivation of UV-damaged luciferase plasmid; γ-H2AX and 53BP1 co-localised foci resolution at 1 h and 24 h; UV LD50 and γ-radiation LD50 by clonogenic assay; SIRT6-mediated stimulation of DSB repair across species; in vitro SIRT6 deacetylation of H3K9, H3K18, and H3K56; SIRT6 mono-ADP-ribosyltransferase activity; PARP1 activation by SIRT6; nucleosome binding by SIRT6; senescence-associated β-galactosidase positivity after γ-radiation; Drosophila median and maximum lifespan

What the study showed, in plain terms

Different mammal species have wildly different maximum lifespans — a mouse lives about 3 years, a beaver about 32. The question this paper asks is whether long-lived species have measurably better DNA repair, and if so, whether SIRT6 is part of the reason.

The authors collected primary skin and lung cells from 18 rodent species spanning that lifespan range and put each cell line through a series of standardised DNA repair assays. Two kinds of repair were tested: nucleotide excision repair, which handles the bulky DNA damage caused by ultraviolet light, and double-strand break repair, which handles the more catastrophic breaks that come from radiation, replication errors, and endogenous damage.

Nucleotide excision repair did not track with lifespan. What it tracked with was sunlight exposure — species that live in daylight (diurnal) had better nucleotide excision repair than nocturnal or underground species, regardless of how long they lived. Double-strand break repair told a completely different story: both of its major sub-pathways, non-homologous end joining and homologous recombination, correlated very strongly with maximum lifespan across the 18 species. Longer-lived species had substantially more efficient repair.

The authors then narrowed in on SIRT6. Each species' SIRT6 gene was cloned and its ability to stimulate DSB repair was tested in mouse cells. Longer-lived species had SIRT6 proteins that were substantially better at boosting DSB repair. By making chimeric SIRT6 proteins that swapped regions between mouse (short-lived) and beaver (long-lived), the authors found the sequence between amino acids 221 and 270 accounts for the difference. Within that stretch, five specific amino acid substitutions were fully responsible.

Beaver SIRT6 was a stronger deacetylase, a stronger mono-ADP-ribosyltransferase, and a stronger stimulator of PARP1 than mouse SIRT6. Introducing the five beaver amino acids into mouse SIRT6 conferred beaver-level enzymatic activity; introducing the five mouse amino acids into beaver SIRT6 reduced it to mouse-level activity.

To ask whether these enzymatic differences mattered for whole-organism lifespan, the authors expressed each SIRT6 variant in fruit flies. Beaver SIRT6 extended fruit fly lifespan roughly 14%, while mouse SIRT6 extended it about 5–7%. Introducing beaver's five residues into mouse SIRT6 pushed the mouse protein toward beaver-level lifespan extension, and vice versa.

Key findings

Non-homologous end joining efficiency in skin fibroblasts across 18 rodent species correlated strongly with maximum lifespan (r² = 0.76, p < 0.0001), and the same held in lung fibroblasts (r² = 0.59, p = 0.0008). The correlation held after phylogenetic correction.
Homologous recombination efficiency also correlated with maximum lifespan in both skin (r² = 0.76, p < 0.0001) and lung fibroblasts (r² = 0.65, p = 0.0002). Neither DSB repair pathway correlated with body mass.
Nucleotide excision repair did not correlate with maximum lifespan (r² = 0.03–0.04 across UV doses). It correlated instead with a species' sunlight exposure classification (low, medium, or high UV), with high-UV species substantially more resistant.
SIRT6's ability to stimulate NHEJ, tested by expressing each species' SIRT6 in mouse cells with an integrated repair reporter, correlated strongly with species maximum lifespan (r² = 0.54, p = 0.0008). SIRT6-stimulated HR correlated similarly (r² = 0.64, p = 0.0001).
Statistical decomposition showed that species-level differences in SIRT6 activity account for a major portion of the correlation between DSB repair and maximum lifespan.
Between mouse and beaver SIRT6, the sequence between amino acids 221 and 270 was fully responsible for the difference in DSB repair stimulation. Within that stretch, five specific residues (positions 235, 249, 260, 263, and 264) together account for the entire difference. No subset of four was sufficient.
Beaver SIRT6 was a substantially stronger deacetylase than mouse SIRT6 at H3K9, H3K18, and H3K56; a stronger mono-ADP-ribosyltransferase; and a stronger stimulator of PARP1 poly-ADP-ribose activity. Introducing the five beaver residues into mouse SIRT6 conferred beaver-level enzymatic activity.
Beaver SIRT6 bound HeLa nucleosomes more tightly than mouse SIRT6, consistent with the crystal-structure prediction that residues 235 and 249 sit near the substrate-binding site.
SIRT6-knockout human dermal fibroblasts expressing beaver WT or mouse-5mut (mouse with beaver residues) SIRT6 showed significantly fewer senescence-associated β-galactosidase-positive cells after 5 Gy γ-radiation than cells expressing mouse WT or beaver-5mut SIRT6.
In Drosophila, RU486-induced ubiquitous expression of beaver WT SIRT6 extended median lifespan by 14.3% and maximum lifespan by 14.7% in males; mouse WT SIRT6 extended median lifespan by 5.4% and maximum lifespan by 7.4%. Introducing the five beaver residues into mouse SIRT6 pushed lifespan extension toward beaver-level (10.7% median), and introducing the five mouse residues into beaver SIRT6 reduced it toward mouse-level (8.9% median). Similar pattern was seen in female flies.

What this study can and cannot tell us

The comparative correlations across 18 rodent species are strong, but correlation between DSB repair efficiency and maximum lifespan does not on its own prove that more efficient DSB repair causes longer lifespan. The paper's causal inference is strengthened by the mouse-beaver swap experiments and Drosophila lifespan validation, but a controlled intervention in a long-lived mammal has not been performed.

The Drosophila lifespan comparison expresses beaver and mouse SIRT6 in flies. This is a heterologous system: fly biology is not mammalian biology, and the extent to which the lifespan extension observed in Drosophila would translate to a mammal receiving a "better" SIRT6 is not established here.

The five critical amino acid residues identified are specific to the mouse–beaver comparison. When SIRT6 sequences are aligned across all 18 species, only two of the five (residues 249 and 263) show a consistent pattern with lifespan. The paper explicitly notes that different species pairs likely have partially overlapping but not identical sets of critical residues, meaning the mouse–beaver five are not a universal signature.

The nucleotide excision repair finding is worth qualifying. The authors classify species into three sunlight-exposure groups based on lifestyle and habitat, which is necessarily approximate. Quantitative measurement of actual sunlight exposure per species was not possible, so the correlation between NER and sunlight is measured against a coarse classification.

SIRT6 activity in stimulating DSB repair was measured by overexpressing each species' SIRT6 in mouse cells. This tests intrinsic protein activity in a foreign cellular context but does not directly measure how SIRT6 behaves in its native cellular environment in each species. Species-specific chromatin, co-factor availability, and interacting partners are not controlled for.

The paper does not test whether pharmacological SIRT6 activation to "beaver-like" activity levels is achievable, and no supplement, dosing, or human clinical outcome can be inferred from this work.