Conceived and designed the experiments: MM GB. Performed the experiments: GB ES LB XH GS CE CM SC SS KJ JX ME. Analyzed the data: MH GB ES LB. Contributed reagents/materials/analysis tools: MH ES LB. Wrote the paper: MM GB.
The authors have declared that no competing interests exist.
The yeast sir2 gene and its orthologues in
Silent information regulator 2 (sir2) is a yeast gene encoding a nicotinamide adenine dinucleotide (NAD+)-dependent histone deacetylase
Mammals have 7 homologues of the Sir2 protein, Sirtuins 1-7 (SirT1-7)
We
Several lines of evidence suggest that SirT1 plays a role in energy metabolism. The dependence on NAD+ as a cofactor for catalysis is thought to link SirT1 activity to the energetic state of the cell
Calorie restriction (CR) has been known for decades to extend lifespan of virtually all organisms from yeast to mammals. Sir2 is required for CR to mediate lifespan extension in yeast and flies
We have previously created mice that do not synthesize the SirT1 protein
A. Body weights of sibling animals of normal (SirT1+/+ and SirT1+/−) and SirT1(−/−)-null genotypes at 2–4, 5–8 and 13–20 months of age. B. Daily food intake of mice as described in A. C. Daily food intake normalized to body weight. D. Caloric content of feces from 2–4 months old normal and SirT1-null mice. Means and standard errors are represented along with the number of animals used for each determination. Unpaired T-tests were performed to assess statistical significance.
The low body mass and hyperphagia of the SirT1-null mice might be a consequence of enhanced activity but we found that young SirT1-null mice are much less active than their normal littermates, particularly during the dark period (
Activity during 60-minute intervals of individually caged SirT1-null and normal littermates 3–6 months (A) and 9–12 months (B) of age. Insets show histograms of total activity in a 24 hour period. Grey and black lines above the X-axis indicate light and dark periods, respectively. Means and standard errors are represented. Unpaired T-tests were performed to assess statistical significance.
Systemic energy metabolism is under the control of hormones. Thyroxine (T4) levels were slightly lower in the SirT1-null serum (
A. T4 levels were measured from serum samples of SirT1-null and normal mice using an EIA kit. B. Corticosterone levels were measured from plasma samples of SirT1-null and normal mice at 4 hours intervals during the diurnal cycle using an EIA kit. The peak values are plotted. Unpaired T-tests were performed to assess statistical significance. Blood glucose (C) and insulin (D) levels of 2–3 months old mutant and wild type mice were measured after 24 hour fasting and again 3 hours after refeeding using a standard glucometer (C) and an ELISA kit (D). Means and standard errors are represented. Two-way ANOVAs was performed to assess statistical significance of interaction between genotype and dietary condition (p*-values on top of panel), and unpaired T-test to assess difference between genotypes within dietary condition groups (p-value over bars).
To estimate metabolic rate and substrate utilization, we measured oxygen consumption and carbon dioxide production from whole animals. When normalized to body weight, oxygen consumption of SirT1-null mice was higher than that of controls, indicating that they are hypermetabolic (
Whole animal indirect calorimetry (IC) was used to assess oxygen consumption normalized to body weight (VO2/BWT) plotted (panel A) at 30 minute intervals during a 24 hour period or (panel B) as the percent relative cumulative frequency (PRCF) of VO2/BWT. The respiratory exchange ratio (RER = VO2/VCO2) was calculated from VO2 and VCO2 data and plotted at 30 minute intervals during a 24 hour period (panel C) or as PRCF (panel D). In a fasting-refeeding experiment (panel E), food was either removed or added at the indicated times (arrows) and the interval RER plotted. Grey and black lines above the X-axis (panels A, C, and E) indicate light and dark periods, respectively. An RER of 1.0 is expected for glucose oxidation and an RER of 0.7 occurs during lipid oxidation. Means and standard errors are represented. An unpaired T-test using medians was performed to assess statistical significance (B).
It has been proposed that SirT1 is required for the induction and maintenance of fatty acid oxidation in response to low glucose concentration
The observation that SirT1-null mice have higher rates of lipid oxidation than normal is consistent with previous observations that in older SirT1-null mice the lipid content of cells in white adipose tissue (WAT) are smaller than those of normal tissues
Organs from 3–5 month (A–C) and 13–26 month old mice (D–F) were weighed and plotted. Inguinal fat pad (A, D), interscapular BAT (B, E) and brain (C, F) are plotted as means and standard errors. Unpaired T-tests were performed to assess statistical significance.
SirT1-null mice are hypermetabolic but lethargic, suggesting that their energy generation system might be defective. We investigated the state 3 (maximal phosphorylating) respiration rate as well as proton conductance in isolated mitochondria from liver and skeletal muscle. We detected no significant differences in the mitochondria from skeletal muscle (data not shown). However, the rate of respiration under state 3 conditions was lower in SirT1-null liver mitochondria suggesting that these mitochondria would produce less ATP at full capacity than those from normal mice (
A. State 3 oxygen consumption rate of liver mitochondria from SirT1-null and normal mice was determined using succinate, in the presence of rotenone. Proton motive force (PMF) was determined in liver mitochondria (B) using succinate, in the presence of rotenone and a saturating amount of oligomycin. The farthest point on the right represents the maximal state 4 oxygen consumption rates. The kinetic response of PMF was determined by inhibiting respiration targeting complex II by incremental additions of malonate (up to 5 mM). C. Western blot of UCP2 from liver mitochondria. Thirty µg of mitochondria-enriched proteins from normal spleen (Spl., positive control) or liver from 3 different SirT1-null and normal mice were loaded separately in each lane, electrophoresed, blotted and probed with an antibody to UCP2. Note that UCP2 protein expression is not de-repressed in mutants. D–E. ROS production from liver mitochondria was measured using the p-hydroxyphenylacetate (PHPA) assay using the substrates and respiration inhibitors indicated at the top of each column. F. Western blot of phospho-AMPKα (pAMPK), AMPα (AMPK), and α−tubulin. Hundred-fifty µg of liver proteins from 2–3 different SirT1-null and normal mice under the indicated dietary condition was loaded separately in each lane, electrophoresed, blotted and probed with an antibody to pAMPK. Membrane was stripped and reprobed sequentially with antibodies to AMPK and α−tubulin. G. Densitometry of western blot signal in F. Bands in F were quantified by densitometry using the ImageJ software. Means and standard errors are represented. Unpaired (A, G) or paired (D–E) T-tests were performed to assess statistical significance (see
Uncoupling proteins such as UCP2 can diminish proton motive force and SirT1 is known to inhibit the expression of ucp2 in pancreatic beta-cells
In addition to ATP synthesis, mitochondria are the major site for the generation of reactive oxygen species (ROS). We measured the capacity of liver mitochondria to produce H2O2 under several conditions. In the presence of succinate or palmitoylcarnitine, liver mitochondria from SirT1-null animals produced less H2O2 than normal (
We also took advantage of this H2O2 production assay as an indirect method to probe the capacity of pathways upstream of the electron transport chain (ETC) to provide reducing equivalents (RE) to the ETC. Mitochondria were subjected to three different substrates (succinate, pyruvate+malate or palmitoylcarnitine) in the presence of antimycin so that H2O2 would be produced predominantly from complex III in all conditions. Succinate feeds into complex II directly, so the H2O2 produced using this substrate is not affected by the capacity of any pathways upstream of the ETC to provide RE. Mitochondria from both normal and SirT1-null liver produced similar levels of H2O2 under this condition (
AMP-activated protein kinase (AMPK) is a sensor for the availability of energy in cells. Our observations with SirT1-null liver mitochondria suggest that they are less efficient than normal in producing ATP. We measured the levels of the active form of AMPK, phopho(Thr172)-AMPKα (pAMPK), in liver from mice fed AL or fasted for 24 hours. Similar levels of activated pAMPK were observed in liver from AL-fed mice, regardless of the genotype (
Sir2 is required for lifespan extension in
CR resulted in a reduced body weight that was similar in both normal and SirT1-null animals (
The weights of whole body (A), brain (B), inguinal fat pad (C), and interscapular BAT (D) were obtained from 13–15 months old mice, AL-fed or CR for 25 to 28 weeks. Percentages above bars represent the percent reduction in weight compared to AL-fed mice.
The daily cumulative physical activity of normal animals increased with CR, as previously reported
Total activity of normal (A) and SirT1-null (B) mice was monitored using a Micromax system for 24-hour periods at the indicated time after the start of CR. Animals were 5–7 months old at the beginning of CR. Means and standard errors are represented. The activity of CR normal mice increased with time whereas that of AL controls decreased. CR did not alter the activity of SirT1-null mice.
Despite a 40% reduction in caloric intake, body weight-normalized oxygen consumption of normal mice subjected to CR was the same or slightly higher than AL fed animals (
Oxygen consumption normalized to whole body weight (VO2/BWT) for normal (A, D) and SirT1-null (B, E) mice was measured after 2–4 weeks (A–C) and 21–26 weeks (D–F) of CR or AL diet as indicated. RER from the same animals at 2–4 weeks (C) or 21–26 weeks (F) of CR were also obtained. Data were plotted as PRCF as in
The SirT1-null animals are less fit than normal and many fail to survive the first year following birth. As CR is a well-established means of prolonging the lifespan of mice, we looked at the survival of SirT1-null and normal mice during our CR experiments. These experiments involved relatively few animals but the data seem to indicate that the decreased viability of SirT1-null mice was exacerbated during CR (
Survival curve for normal (A) and SirT1-null (B) mice that were fed the CR of AL diets. Mutant mice that were euthanized because of eyelid inflammation were excluded from this graph. Grey rectangles represent the period when CR started (5 to 7 months old).
Since its catalytic activity is dependent on NAD+, SirT1 deacetylase activity has been postulated to be controlled by the metabolic state of the cell
SirT1-null mice are smaller and lethargic compared to their normal littermates but, per gram of body weight, they consume more food and oxygen. Thus, the SirT1-null animal is metabolically inefficient compared to normal. Liver mitochondria appear to have a lower capacity to produce ATP because of lower state 3 respiration capacity and higher proton leak through the inner mitochondrial membrane. Perhaps to compensate for their reduced oxidative phosphorylation capacity, mitochondria from SirT1-null liver have increased capacity for TCA cycle and beta-oxidation. This compensation seems to be successful in maintaining the level of ATP but overnight fasting resulted in much higher levels of phospho-AMPK suggesting that ATP levels are not sustained in SirT1-null mice in the face of food deprivation.
SirT1 is known to modulate the activities of several regulators of metabolism but an examination of the characteristics of the SirT1-null mouse and the expectations based on published studies yields a number of unexpected contradictions. As a negative regulator of PPARγ, SirT1 decreases transcription of genes involved in fat storage
The means by which CR extends lifespan is not yet clear. In yeast and
Lifespan is thought to be determined by the accumulation of cellular damage arising from ROS although recent evidence from C. elegans suggests that ROS might in fact be responsible for extending lifespan (Schulz et al., 2007). It is perhaps relevant that SirT1-null mice have liver mitochondria that produce less ROS than normal. In regards to mitochondrial uncoupling, this observation is consistent with increased proton leak in these mitochondria; however, the increased leak is not due to a derepression of ucp2, as one could infer from studies in pancreatic beta-cells
In conclusion, our study indicates that the absence of SIRT1 results in a metabolically inefficient animal that fails to adapt to CR conditions. Given that there are over 30 SirT1 substrates to date
All animal experiments were performed according to the Guidelines for the Care and Use of Animals established by the Canadian Council on Animal Care. The mutant mice used in this study carried the
Two to four months old mice were caged individually in metabolic chambers and feces were collected after 48 to 72 hours. Feces were dehydrated in a speedvac at 50°C overnight and grounded to powder. Gross energy (GE) of feces was determined using an automatic bomb calorimeter (Parr, 1271, Parr Instruments, Moline, IL, USA).
Mice were caged individually and activity was recorded for 24-hour periods in a MicroMax activity monitoring system with 16 infrared beams per cage (AccuScan Instruments, Columbus, OH, USA). Total activity data were used for analyses. Lighting was on a normal 12 h light/dark cycle.
Mice were caged individually and oxygen consumption (VO2) and carbon dioxide production (VCO2) were measured using a four-chamber Oxymax system with automatic temperature and light controls (Columbus Instruments, Columbus, OH). Temperature was maintained at 24°C, and lighting was on a normal 12 h light/dark cycle. System settings included a flow rate of 0.5 L/min, a sample line-purge time of 2 min, and a measurement period of 60 s every 12 minutes. The respiratory exchange ratio (RER) was calculated as the ratio of VCO2 produced/VO2 consumed.
Four to 5 month old mice were euthanized by decapitation for isolation of liver and skeletal muscle mitochondria. All media were ice-cold, and the procedures done on ice or at 4°C. Isolation of skeletal muscle mitochondria was performed using a modified method of Chappell and Perry
Oxygen consumption was measured in isolated mitochondria (0.5 mg/ml) at 37°C using a Clark-type oxygen electrode (Hansatech, Norfolk, UK) and incubated in standard incubation medium (IM: 120 mM KCl, 1 mM EGTA, 5 mM KH2PO4, 5 mM MgCl2 and 5 mM HEPES; pH 7.4) containing 0.3% defatted BSA and assumed to contain 406 nmol O/ml at 37°C
A methyl-triphenyl-phosphonium (TPMP+)-sensitive electrode was used to assess mitochondrial protonmotive force (Δp)
Mitochondrial H2O2 production rate was determined in freshly isolated mitochondria from liver using the
For the evaluation of UCP2 levels, liver mitochondria from SirT1 mutant and WT mice, as well as spleen mitochondria from WT mice, were isolated (see “Isolation of mitochondria” section above) and 30 µg of protein was loaded on a NuPAGE Bis-Tris 4–12% gradient precast polyacylamide gel (Invitrogen, Burlington, ON, Canada), electrophoresed, blotted on a nitrocellulose membrane and probed with an antibody to UCP2 (C20) (cat. #: sc-6525, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Spleen was used as a positive control for UCP2 expression. Loading was verified by Ponceau Red staining.
For evaluation of phospho-AMPKα and AMPKα levels in liver tissue, SirT1 mutant and WT mice were euthanized by cervical dislocation, and pieces of liver were harvested and flash frozen within 1 min. Tissues were homogenized in ice-cold RIPA-phosphatase inhibitor buffer (Tris 20 mM pH 8.0, NP-40 1%, sodium deoxycholate 0.25%, NaCl 150 mM, EDTA 1 mM, Complete protease inhibitor cocktail 1× (Roche, Laval, QC, Canada), Na3VO4 1 mM and NaF 1 mM) using a polytron homogenizer. Samples were centrifuged in a microfuge at 4°C, at full speed for 10 min and supernatants were collected and frozen at −80°C. A hundred fifty µg of proteins of each sample were loaded on a NuPAGE Bis-Tris 4–12% gradient precast polyacylamide gel (Invitrogen, Burlington, ON, Canada), electrophoresed, blotted on a nitrocellulose membrane and probed overnight at 4°C with an antibody to phospho-AMPKα (Thr172) (cat. #: 2531, Cell Signalling Technology, Danvers, MA, USA). Membrane was stripped in stripping buffer (SDS 2%, Tris 62.5mM pH 6.8, β-mercaptoethanol 100 mM) at 55°C for 30 min, washed, blocked and reprobed overnight at 4°C with an antibody to AMPKα (23A3) (cat. #: 2603, Cell Signalling Technology, Danvers, MA, USA). The day after, the membrane was reprobed for an hour at room temperature with an antibody to α-tubulin (cat. #: 2125, Cell Signalling Technology, Danvers, MA, USA), and AMPK and α-tubulin signals were revealed together. Signals were quantitated by densitometry using the ImageJ software (NIH, Bethesda, MD, USA).
Two to three months old mice were fasted for 24 hours and blood was collected between 9am and 11am from the saphenous vein. For the refeeding experiments, mice were left at least one week to recover and were fasted again for 24 hours, refed between 9am and 11am and then, blood was collected 3 hours later. Blood glucose concentration was measured immediately when blood was collected using a One Touch II glucometer (LifeScan Canada Ltd., Burnaby, BC, Canada) and serum was frozen at −80°C. Serum insulin was measured using an ELISA kit (cat. #: EZRMI-13K, Linco Research Inc., MO, USA), according to the manufacturer's instructions.
For T4 measurement, blood was collected from the saphenous vein in the afternoon and plasma was frozen at −80°C. Plasma T4 levels were measured using an EIA kit (cat. #: 07BC-1007, MP Biomedicals, Orangeburg, NY, USA) according to the manufacturer's instructions. For corticosterone measurements, 20 µl of blood was collected every 4 hours during a day, starting at 6am. In order to let the mice recover between harvests, blood was taken every 28 hours for 6 days. Serum from every sample was frozen at −80°C. Serum corticosterone levels were measured using an EIA kit (cat. #: 900-097, Assay Design, Ann Harbor, MI, USA) according to the manufacturer's instructions.
Pre-weighed food pellets from Bioserv (Frenchtown, NJ, USA) were used for CR studies. Daily food intake of mice was determined for several weeks and averaged by genotype. Mice were adapted to the food for at least 1 month before CR started. SirT1 mutant mice and controls were assigned to AL or CR. Cohorts were matched for sex and litter when possible. Mice from CR groups were given 60% of average daily food intake of their corresponding genotype and mice from AL were given food unrestricted or 95% of average to prevent obesity. Age of mice was between 5 and 7 months when CR started and the diet was sustained for up to 44 weeks.
The authors would like to thank the following people for allowing us to use materials: Alexander McKenzie for his Micromax activity monitoring system and Christopher Kennedy for his metabolic chambers. We are also grateful to Manfred Hansel and the Department of Animal and Poultry Science of the University of Guelph for performing bomb calorimetry analyses.