Mitochondrial sirtuins sir-2.2 and sir-2.3 regulate lifespan in C. elegans

Abstract

Mitochondrial sirtuins regulate metabolism and are emerging drug targets for metabolic and age-related diseases such as cancer, diabetes, and neurodegeneration. Yet, the extent of their functions remain unclear. Here, we uncover a physiological role for the C. elegans mitochondrial sirtuins, sir-2.2 and sir-2.3, in lifespan regulation. Using genetic alleles with deletions that destroy catalytic activity, we demonstrate that sir-2.2 and sir-2.3 mutants live an average of 25% longer than controls when fed the normal lab diet of live E. coli OP50. While decreased consumption of food is a known mechanism for lifespan extension, we did not find evidence of reduced pharyngeal pumping. Interestingly, lifespan extension effected by loss of sir-2.2 or sir-2.3 is sensitive to the diet. The lifespan extension of the sir-2.2 mutants is eliminated and that of sir-2.3 mutants is attenuated when the animals are fed the E. coli strain HT115, which is typically used for RNAi experiments. We used growth ability of the food source and a virulent pathogenic strain to ask if differences in pathogenicity are related to the mechanisms for lifespan extension. sir-2.3 deletion results in lifespan extension in all conditions. However, removing the ability of the food source to grow eliminated the sir-2-mediated effect. We also examine the response of the mutants to oxidative stress, and our results suggest that a hormetic response contributes to lifespan extension in both mutants. Our data suggest that sir-2.2 and sir-2.3 use overlapping yet distinct mechanisms for regulating lifespan.

Introduction

Sirtuins are a highly conserved family of NAD⁺-dependent enzymes that use NAD⁺ as a co-substrate to execute mono-ADP ribosylation, deacetylation and other deacylation reactions (Blander and Guarente 2004; Houtkooper et al. 2012). Distinct sirtuin family members are active in the cytoplasm, nucleus, and mitochondria, and act as molecular sensors and regulators of metabolism (Haigis and Sinclair 2010; Guarente 2011; He et al. 2012; Guarente 2013). Sirtuins are named for the Saccharomyces cerevisiae protein Sir2 (silent information regulator 2), which increases yeast replicative lifespan when overexpressed (Kaeberlein et al. 1999). Sir2 orthologs have been extensively explored in many models. For example, over-expression of the C. elegans nuclear sirtuin SIR-2.1 was originally shown to share lifespan enhancing functions with Sir2 (Tissenbaum and Guarente 2001). However, the robustness of the lifespan extension of C. elegans SIR-2.1 has been questioned (Burnett et al. 2011). Nonetheless, extensive research on SIR-2.1 and orthologous sirtuins supports the model that sirtuins are key players in modulating progression of aging and health span via regulation of metabolism and oxidative stress responses (Satoh et al. 2011; Guarente 2013; Chang and Guarente 2014).

Three of the seven sirtuins found in mammals, SIRT3–5, reside in the mitochondria (Houtkooper et al. 2012). While the activities and functions of the mitochondrial sirtuins are not fully elucidated, they appear to have distinct activities and roles. SIRT3 is the major deacetylase in the mitochondria (Lombard et al. 2007), and it regulates metabolism (Zhang et al. 2020) and the oxidative stress response (Schwer et al. 2006; Qiu et al. 2010; Tao et al. 2010). SIRT4 regulates insulin secretion and lipid homeostasis (Haigis et al. 2006; Laurent et al. 2013; Anderson et al. 2017) via ADP-ribosyltransferase (Ahuja et al. 2007), deacetylase (Laurent et al. 2013), and lipoamidase (Mathias et al. 2014) activities. SIRT5 possesses deglutarylase (Tan et al. 2014; Zhang et al. 2024), demalonylase (Du et al. 2011) and succinylase (Du et al. 2011; Sadhukhan et al. 2016) activities and regulates the urea cycle (Nakagawa et al. 2009; Du et al. 2011).

In C. elegans, there are four sirtuins, SIR-2.1 to SIR-2.4. SIR-2.2 and SIR-2.3 localize in the mitochondria and are orthologs of mammalian SIRT4 (Wirth et al. 2013; Jedrusik-Bode 2014). SIR-2.2, SIR-2.3 and SIRT4 physically interact with the mitochondrial enzymes pyruvate carboxylase, propionyl-CoA carboxylase, and methylcrotonyl-CoA carboxylase (Wirth et al. 2013). While SIR-2.1 is well-studied, much less is known about the biological roles of SIR-2.2 and SIR-2.3. We have investigated functions of C. elegans mitochondrial sirtuins using a genetic approach. We reveal novel roles of C. elegans mitochondrial sirtuins in lifespan regulation through non-redundant processes despite the high degree (~75%) of primary amino acid sequence conservation.

Materials and Methods

C. elegans strains and maintenance

Strains were maintained using standard methods (Brenner, 1974). Strains and alleles: N2 Bristol (wild-type control); RB654 sir-2.3(ok444) (provided by CGC); sir-2.2(tm2673) (provided by Mitani lab). sir-2.2(tm2673) and sir-2.3(ok444) were each outcrossed three times into the N2 control strain.

Lifespan Analysis

Lifespan assays were conducted at 20°C on NGM plates with 400 μl of live or UV-killed E. coli strains OP50 or HT115 or Pseudomonas aeruginosa PAO1. For UV-killed experiments, plates were seeded with live E. coli, allowed to dry for 2 days, and then treated with UV for 30m. Elimination of culturability was verified via subculture on LB plates. Animals were synchronized using a timed egg lay or an egg preparation (Sulston & Hodgkin, 1988). L4 animals were placed on individual plates and moved to new plates daily during assay. Worms were prodded with a platinum wire daily and scored as dead if non-responsive. Internal egg hatching or “exploded” phenotypes were censored. For anti- and pro-oxidant lifespans, 200 μl of paraquat (methyl viologen dichloride hydrate 98%, Sigma) or L-ascorbic acid (Fisher Chemical^™ A61–25) stock solution diluted in water was added to NGM plates spotted with 400 μl OP50 to achieve a final concentration of 0.1 mM or 3.75 mM for paraquat and 10 mM for ascorbic acid.

DCFDA fluorescence analysis

DCFDA (2′,7′-Dichlorofluorescin diacetate ≥97%, Sigma-Aldrich) assays were conducted on age-synchronized animals by isolating eggs with a bleaching protocol (Sulston & Hodgkin, 1988) and then allowing animals to develop to L4. L4 animals were washed with M9 buffer, and then diluted with M9 to roughly the same visual density. 50 μL of worms was added to each well of a black 96-well plate. 100 μL of M9 with 50 μM DCFDA was added to each well. Samples were analyzed using a Spectramax i3X 96-well plate scanner (abs: 455 nm ext: 530) with a reading taken every 5m. The ROS profile was normalized to total protein concentration determined using BCA assay following homogenization of samples.

Pharyngeal Pumping

Pharyngeal pumping was measured in day one adults fed ad libitum or 5m post 6h fasting on OP50 (Lemieux et al., 2015). Pumping rates were recorded and then counted in ten second intervals using a Nikon SMZ1500 stereoscope equipped with Roper Scientific Photometrics CoolSnap EZ camera.

Results

Mitochondrial sirtuin mutants live longer than N2 when fed E. coli OP50 ad libitum

Mitochondrial sirtuin mutants sir-2.2(tm2673) and sir-2.3(ok444), referred to as sir-2.2(Δ) and sir-2.3(Δ), are deletion alleles that each remove catalytically important residues, likely eliminating catalytic function (Supplementary Figure 1) (Wirth et al. 2013). We noted that these mutants were very robust as they aged. Thus, we compared the lifespan of sir-2.2(Δ) and sir-2.3(Δ) relative to the N2 strain to which they were backcrossed three times. Loss of sir-2.2 or sir-2.3 function resulted in an average 25% increase in lifespan when fed E. coli OP50 ad libitum (Fig. 1A,B, Supplementary Table 1). sir-2.2(Δ) and sir-2.3(Δ) lived an average of 3.6 and 3.7 days longer, respectively. Because an increase in sirtuin activity has been associated with lifespan extension, we wondered if the increased lifespan observed in single mitochondrial sirtuin mutants was due to upregulation of the remaining mitochondrial sirtuin. We detected no compensatory upregulation between the mitochondrial sirtuins (Fig. 1C). Sangaletti et al. previously demonstrated that the mRNA level of no other sirtuin was upregulated in sir-2.3(ok444) (Sangaletti et al. 2017).

Figure 1. Lifespan of sir-2.2(Δ) and sir-2.3(Δ) fed ad libitum on OP50.

(A) Survival curve and (B) mean lifespan of N2 control, sir-2.2(Δ) and sir-2.3(Δ) derived from seven independent experiments (See Supplementary Table 1). Statistical analysis: survival curves via log rank test; mean via one-way ANOVA followed by Dunnett’s post hoc test, ****p<0.0001. (C) Relative levels of sir-2.2 mRNA in sir-2.3(Δ) (red) and sir-2.3 mRNA in sir-2.2(Δ) (blue) measured with qRT-PCR. Values are the average of three independent experiments ±SEM. ns: p>0.05, one sample t-test.

Because dietary restriction increases lifespan (Klass 1977; Hosono et al. 1989; Sohal and Weindruch 1996), we investigated whether the lifespan extension of sir-2.2(Δ) and sir-2.3(Δ) was due to decreased food intake. The contraction or pumping rate of the C. elegans pharynx, which moves food into the intestine, is a standard proxy for assessment of short-term food intake (e.g., Avery & Shtonda, 2003; Avery & You, 2012; Lemieux et al., 2015; Paek et al., 2012; Seymour et al., 1983). We measured pumping rate and found no difference between the N2 control and the sirtuin mutants providing no support for the hypothesis that decreased food intake contributes to the increase in lifespan of sir-2.2(Δ) or sir-2.3(Δ) (Supplementary Figure 2).

Diet influences sir-2.2 and sir-2.3-related lifespan extension

In preparation for RNAi experiments, we noted that food source influenced the lifespan extension. sir-2.2(Δ) has no lifespan extension when grown on HT115 (Fig. 2A,B, Supplementary Table 2). While an increase in lifespan is still evident when sir-2.3(Δ) is grown on HT115 (Fig. 2A,B, Supplementary Table 2), the magnitude of the extension (12% or 1.5 days) is strongly reduced relative to the 25% extension observed on OP50. Thus, lifespan phenotypes of sir-2.2(Δ) and sir-2.3(Δ) can be modulated by an aspect of the diet, and the diet-dependent effect is more pronounced in sir-2.2(Δ) than in sir-2.3(Δ).

Figure 2. Lifespan of sir-2.2(Δ) and sir-2.3(Δ) fed ad libitum on HT115.

(A) Survival curve and (B) mean lifespan of N2 control, sir-2.2(Δ) and sir-2.3(Δ) fed ad libitum on HT115 across 4 experiments (See Supplementary Table 2). Statistical analysis: survival curves via log rank test; mean via one-way ANOVA followed by Dunnett’s post hoc test, ****p<0.0001, ***0.0001<p<0.001; **0.001<p<0.01; *0.01<p<0.05; ns p>0.05.

Pathogenicity influences sir-2.2-related lifespan extension

Colonization of the intestinal tract of aged C. elegans by E. coli may contribute to mortality (Garigan et al. 2002). Thus we considered the hypothesis that the lifespan extension in the sirtuin mutants might be a result of enhanced resistance to pathogenic effects of the food source. To investigate this hypothesis, we manipulated food pathogenicity. We examined lifespan on less pathogenic food by providing food sources that were unable to grow or colonize the gut, UV-killed OP50 and HT115, as well as on a highly pathogenic food source, Pseudomonas aeruginosa PAO1, a virulent isolate known to colonize the gut and lead to animal’s death within a few days (Mahajan-Miklos et al. 1999; Tan et al. 1999). Unlike when cultured on live OP50, sir-2.2(Δ) did not extend lifespan when cultured on UV-killed OP50 (Fig. 3A,B, Supplementary Table 3) or UV killed HT115 (Fig. 3C,D, Supplementary Table 4). Strikingly, the short lifespan of animals grown on Pseudomonas PAO1 was robustly extended by 2.4 days or 35% in sir-2.2(Δ) (Fig. 3E,F, Supplementary Table 5). In contrast to sir-2–2(Δ), sir-2.3(Δ) animals have an extended lifespan in all conditions tested (Fig. 3A–F). The magnitude of the sir-2.3(Δ) lifespan extension observed on UV-killed OP50 (14%) was reduced relative to the live OP50 (25%), but the effect of sir-2.3(Δ) on live and UV killed HT115 were similar, 11 and 14% respectively. sir-2.3(Δ) also robustly extended lifespan by almost 2 days or 27% when cultured on Pseudomonas, similar to the 25% extension noted on OP50.

Figure 3. Lifespan of sir-2.2(Δ) and sir-2.3(Δ)under hypo- and hyper-pathogenic conditions.

(A) Survival curve and (B) mean lifespan of N2 control, sir-2.2(Δ), and sir-2.3(Δ) animals fed ad libitum on UV-killed OP50 across three experiments (See Supplementary Table 3), (C) Survival curve and (D) mean lifespan of N2 control, sir-2.2(Δ), and sir-2.3(Δ) animals fed ad libitum on UV-killed HT115 across three experiments (See Supplementary Table 4). (E) Survival curve and (F) mean lifespan of N2 control, sir-2.2(Δ), and sir-2.3(Δ) animals fed ad libitum on Pseudomonas aeruginosa PAO1 across three experiments (See Supplementary Table 5). Statistical analysis: survival curves via log rank test; mean via one-way ANOVA followed by Dunnett’s post hoc test, ****p<0.0001, ***0.0001<p<0.001; **0.001<p<0.01; *0.01<p<0.05; ns p>0.05.

A hormetic response may underlie sir-2.2 and sir-2.3-related lifespan extension

Both sir-2.2(Δ) and sir-2.3(Δ) are hypersensitive to oxidative stress (Wirth et al. 2013), suggesting that the proteins might help ameliorate stress and that the mutants may experience an elevated constitutive level of oxidative stress. We hypothesized that a mounted response to a mild level of constitutive oxidative stress, a hormetic response, could mediate the extended lifespan. To investigate this hypothesis, we first measured intracellular reactive oxygen species (ROS) using the probe 2′,7′-Dichlorofluorescin diacetate (DCFDA), which becomes fluorescent after oxidization. sir-2.2(Δ) and sir-2.3(Δ) showed increased relative DCFDA fluorescence indicating higher levels of ROS production compared to N2 controls (Fig. 4A). We also measured ROS production when animals were grown on a UV-killed diet. Interestingly, the response of control N2 animals and the sirtuin mutants to the UV-killed food differs. A UV killed OP50 diet results in an increase in ROS levels for the N2 strain but a decrease In ROS levels for both sir-2.2(Δ) and sir-2.3(Δ) (Fig. 4A). These results are consistent with an altered oxidative stress response in the mutants.

Figure 4. Impacts of pro and antioxidant conditions on sir-2.2(Δ) and sir-2.3(Δ) lifespan.

(A) Measurement of intracellular ROS levels of N2, sir-2.2(Δ) and sir-2.3(Δ) fed ad libitum on live or UV-killed OP50. Relative DCFDA fluorescence values are averages from three biological replicates performed with duplicate technical replicates per sample. Statistical analysis via two-way ANOVA. Designations for statistical significance on top solid lines are comparisons to N2 on live food. Designations for statistical significance on bottom dotted lines are comparisons to same genotype on live food. ****p<0.0001; ***p<0.001; **p<0.01; *p<0.05. (B) Average survival curves and (C) mean lifespan comparing N2 and each sirtuin mutant cultured with 0.1 mM paraquat (see Supplementary Table 6). (D) Average survival curves comparing each sirtuin mutant cultured with and without 0.1 mM paraquat. (E) Average survival curves comparing N2 and each sirtuin mutant cultured with 10 mM vitamin C (ascorbic acid) across 3 experiments (see Supplementary Table 7). Statistical analysis for survival curves: log rank test. Statistical analysis for mean: via one-way ANOVA followed by Dunnett’s post hoc test. ****p<0.0001, ***0.0001<p<0.001; **0.001<p<0.01; *0.01<p<0.05; ns p>0.05.

To further test the hypothesis that a hormetic response contributes to the extended lifespan of the mitochondrial sirtuin mutants, we experimented with pro- and anti-oxidant conditions. Low levels of superoxide generators, such as 0.1 mM paraquat, extend the lifespan of wild-type worms but have no effect on mutants that already have elevated production of ROS (Yang and Hekimi 2010). We recapitulated the lifespan extension expected when control N2 animals are treated with a low concentration (0.1 mM) of paraquat (Fig. 4B,C, Supplementary Table 6). However, this lifespan extension associated with mutation of the sirtuins was greatly reduced in the presence of paraquat. The lifespan curves of N2, sir-2.2(Δ) and sir-2.3(Δ) treated with 0.1 mM paraquat were not statistically different, suggesting no extension (Fig. 4B). The mean lifespans did indicate some extension by sir-2.2(Δ) and sir-2.3(Δ) relative to N2 in the presence of 0.1 mM paraquat (Fig. 4C), but the magnitude of the average lifespan extension decreased from an average of 40% in this experiment without paraquat to only 12% in the presence of paraquat. In fact, this concentration of paraquat had a negative effect on the lifespan of sir-2.2(Δ) and sir-2.3(Δ) (Fig. 4D) in contrast to the extension caused by 0.1 mM paraquat on N2 (Fig. 4BC). As previously published (Wirth et al. 2013), sir-2.2 and sir-2.3 mutants were more sensitive than controls to the high concentration of 3.75 mM paraquat (Supplementary Table 6). Treatment with the antioxidant vitamin C extended the lifespan of control animals but eliminated the difference in lifespan between control and sirtuin mutants (Fig. 4E, Supplementary Table 7). Our results are consistent with the hypothesis that the sirtuin mutants have mildly elevated levels of ROS that are important for the lifespan extension.

Discussion

We have uncovered a role for the C. elegans mitochondrial sirtuins in lifespan regulation; removal of either sir-2.2 or sir-2.3 activity increases lifespan by an average of 25%. Interestingly, sir-2.1 mutants also have an increased lifespan (Moroz et al. 2014), and the lifespan extension of sir-2.3(Δ) has been corroborated (Zhang et al. 2024 Jan 2). sir-2.2(Δ) and sir-2.3(Δ) lifespan-extension is sensitive to diet or environment; they each have an extended lifespan when cultured on OP50 but the extension is either absent (sir-2.2(Δ)) or dampened (sir-2.3(Δ)) when cultured on HT115. The diet-dependent effect is not surprising as diet has significant influence on lifespan (Pang and Curran 2014). However, the distinct phenotypes observed in culture with each of the two most common food sources used in C. elegans laboratories as well as between live and dead food is likely relevant to the biology of these proteins and is also relevant to experimental design; recognition that phenotypes on OP50 might not be recapitulated on HT115 is important when designing RNAi experiments in C. elegans.

Lifespan is often positively correlated with pathogen resistance and is influenced by oxidative stress responses. Thus we investigated the impact of pathogenicity and oxidative stress in the lifespan extension. First, we manipulated the pathogenicity of the food. sir-2.2(Δ) manifested no lifespan extension when cultured on a bacterial strain incapable of colonizing the gut (UV-killed OP50) and extended lifespan more robustly (by 35%) when fed a highly pathogenic Pseudomonas strain. We conclude that lack of SIR-2.2 activity appears to enhance pathogen resistance and this may be an important function for helping the mutants live longer. sir-2.3(Δ) retained at least some ability to extend lifespan even on a food source that does not colonize the gut. Thus, at least some aspects of lifespan extension mechanisms differ between sir-2.2(Δ) and sir-2.3(Δ), with sir-2.3(Δ) mechanisms less related to pathogen resistance.

When we manipulated oxidative stress, both sir-2.2(Δ) and sir-2.3(Δ) behaved as predicted if a hormetic effect was at play. The increase in constitutive ROS levels in mutants is also consistent with a proposed hormetic effect. Notably, we see poor correlation between actual ROS levels and lifespan. For example, two conditions that raise ROS levels to similar absolute levels – mutation of sir-2.3 relative to N2 and growth of N2 on UV-killed relative to live OP50– have dissimilar effects on lifespan (extended in the first case and either unaffected or even diminished in the second). Nevertheless, the distinct direction of the changes in ROS levels in N2 compared to sir-2.2(Δ) or sir-2.3(Δ) when grown on UV-killed OP50 highlights that sir-2.2(Δ) and sir-2.3(Δ) mount a distinct response to oxidative stress relative to N2. This altered response was corroborated in another study focused on a model for aging-related neural degeneration in C. elegans. Sangaletti et al. demonstrated an altered level of ROS elevation in sir-2.3(Δ) under low nutrient conditions and a protective effect of SIR-2.3 on neural degeneration (Sangaletti et al. 2017)

The reasons for the altered oxidative stress responses and outcomes in sir-2.2(Δ) or sir-2.3(Δ) are unknown. SIR-2.2, SIR-2.3, and their mammalian ortholog Sirt4 may have the deacetylase activity that was classically associated with sirtuins (Rauh et al. 2013). However, Sirt4 is more potent lipoamidase and has deacylase activity for a variety of moieties (Mathias et al. 2014; Pannek et al. 2017). SIR-2.2, SIR2.3 and SIRT4 interact with enzymes involved in fatty acid oxidation (Haigis et al. 2006; Laurent et al. 2013; Wirth et al. 2013) and knockdown of SIRT4 impacts lipid metabolism and gluconeogenesis (Haigis et al. 2006; Haigis and Sinclair 2010; Nasrin et al. 2010; Laurent et al. 2013). Thus, SIR-2.2 and SIR-2.3 are likely poised to act as central regulator of metabolism in the mitochondria effecting changes in response to stress and nutrient conditions. Their loss is predicted to have cascading effects on mitochondrial homeostasis, and high constitutive ROS as well as altered ability to respond to additional stress might arise from a disrupted electron transport chain or from changes in levels of oxidants or oxygen scavengers.

We considered whether differences in ability to respond to oxidative stress or differences in pathogenicity account for the differences in phenotype between animals cultured on OP50 and HT115. Animals fed an OP50 diet, but not an HT115 diet, have a mild vitamin B12 deficiency that negatively affects mitochondrial function (Revtovich et al. 2019). Given the likely role of SIR-2.2 and SIR-2.3 in regulation of metabolism in the mitochondria, we consider it most likely that the difference in mitochondrial health between OP50 and HT115 accounts for the differing outcomes.

While the lifespan extending mechanism in sir-2.2(Δ) and sir-2.3(Δ) still requires investigation, our work supports a lack of redundancy between SIR-2.2 and SIR2.3, which are highly homologous, and emphasizes the importance of the sirtuin family as modulators of oxidative stress response and lifespan. These results are supportive of strategies to target this class of proteins for new therapies, potentially in ways that might counter aging.

Data availability

Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures and tables.