Calorie restriction reduces rDNA recombination independently of rDNA silencing

Abstract

Calorie restriction (CR) extends lifespan in yeast, worms, flies and mammals, suggesting that it acts via a conserved mechanism. In yeast, activation of the NAD-dependent histone deacetylase, Sir2, by CR is thought to increase silencing at the ribosomal DNA, thereby reducing the recombination-induced generation of extrachromosomal rDNA circles, hence increasing replicative lifespan. Although accumulation of extrachromosomal rDNA circles is specific to yeast aging, it is thought that Sirtuin activation represents a conserved longevity mechanism through which the beneficial effects of CR are mediated in various species. We show here that growing yeast on 0.05 or 0.5% glucose (severe and moderate CR, respectively) does not increase silencing at either sub-telomeric or rDNA loci compared with standard (2% glucose) media. Furthermore, rDNA silencing was unaffected in the hxk2Delta, sch9Delta and tor1Delta genetic mimics of CR, but inhibited by FOB1 deletion. All these interventions extend lifespan in multiple yeast backgrounds, revealing a poor correlation between rDNA silencing and longevity. In contrast, CR and deletion of the FOB1, HXK2, SCH9 and TOR1 genes, all significantly reduced rDNA recombination. This silencing-independent mechanism for suppressing rDNA recombination may therefore contribute to CR-mediated lifespan extension.

Fig. 1

Proposed mechanisms of lifespan extension by calorie restriction. Models 1 and 2 suggest that CR up-regulates Sir2 activity, which leads to increased gene silencing and decreased rDNA recombination. Consequently, less ERCs are formed, which extends the yeast replicative lifespan. Both models differ only in the events upstream of Sir2 activation. Model 1 suggests that the increase in respiration in response to nutrient deprivation modulates the NAD⁺/NADH ratio in favour of NAD⁺, whereas model 2 explains the increase in Sir2-activity via a rise in Pnc1 levels and consequently the degradation of the endogenous Sir2 inhibitor, nicotinamide. Model 3 bypasses Sir2-dependent gene silencing altogether and puts forward the Tor1/Sch9 kinase pathways as links between calorie restriction and lifespan extension; although the downstream mechanism has not been fully determined, this is postulated to involve reduced ribosomal protein synthesis.

Fig. 2

Calorie restriction does not affect subtelomeric gene silencing. (A) Qualitative silencing assay. The telomeric silencing reporter parent strain, AEY1017, and its isogenic sir2Δ strain were grown up in YPD, then diluted in H₂O and serially spotted out on synthetic complete media (SC) and SC + FOA media (FOA). Spots correspond to OD₆₀₀ = 1 (left) then serial 1 in 5 dilutions to the right. Plates were imaged after 2 (0.5% and 2% glucose) and 3 days (0.05% glucose) at 30 °C. Reduced growth on FOA compared with SC indicates reduced silencing. (B) Quantitative silencing assay. Serial dilutions of overnight cultures were spread on plates and colonies were counted after 2 (0.5% and 2% glucose) or 3 days (0.05% glucose). Silencing is proportional to viability on FOA media, which was calculated by dividing the numbers of colonies on FOA plates by the total numbers of colonies on −ura and FOA plates combined. 62 000 colonies were counted in total. The graph shows pooled data normalized to the 2% glucose control value, expressed as mean + standard error of the mean. No statistically significant difference between the conditions was found.

Fig. 3

Calorie restriction does not affect gene silencing in the rDNA. (A) Qualitative silencing assay. The rDNA silencing reporter parent strain, JS128, and its isogenic sir2Δ strain were grown up in YPD, then diluted in H₂O and serially spotted out on synthetic complete media (SC) and SC − ura media (−ura). Spots correspond to OD₆₀₀ = 1 (left) then serial 1 in 5 dilutions to the right. Plates were imaged after 2 (0.5% and 2% glucose) and 3 days (0.05% glucose) at 30 °C. Reduced growth on −ura compared with SC indicates increased silencing in this assay. (B) Quantitative silencing assay. Serial dilutions of overnight cultures were spread on plates and colonies were counted after 2 (0.5% and 2% glucose) or 3 days (0.05% glucose). Silencing is inversely proportional to viability on −ura media, which was calculated by dividing the numbers of colonies on −ura plates by the numbers of colonies on SC plates (black bars). Viability on FOA was assessed in parallel by dividing the numbers of colonies on FOA plates by numbers of colonies on SC plates (grey bars). 63 000 colonies were counted in total. The graph shows pooled data normalized to the 2% glucose control value, expressed as mean + standard error of the mean. Statistical analysis was performed using Student’s t-tests (*P<0.05).

Fig. 4

Effects of lifespan-extending mutations on rDNA silencing. (A) The rDNA silencing reporter parent strain, JS128, and isogenic deletion strains hxk2Δ and fob1Δ, were grown up in YPD. These were then diluted in H₂O and serially spotted out on synthetic complete media (SC), SC − ura media (−ura) and SC + FOA media (FOA) containing 2% glucose. Spots correspond to OD₆₀₀ = 1 (left) then serial 1 in 5 dilutions to the right. Plates were imaged after 2 days at 30 °C. Reduced growth on −ura compared with SC indicates increased silencing in this assay. (B) The JS122 control strain, where URA3 is integrated in a nonsilenced locus, and isogenic deletion strains comprising hxk2Δ and fob1Δ, were grown up in YPD and plated exactly as above. (C) The telomeric silencing reporter parent strain, AEY1017, and its isogenic deletion strains, hxk2Δ and fob1Δ, were grown up in YPD, plated as above and imaged after 3 days. Reduced growth on FOA is a readout of reduced silencing in this assay.

Fig. 5

Calorie restriction reduces rDNA recombination. The rDNA silencing reporter parent strain, JS128, the control JS122 strain, and isogenic fob1Δ and sir2Δ strains were serially cultured in YP media containing the indicated glucose concentrations for approximately 120 generations. These were then spread onto YPD plates and then replica plated onto both SC and −ura plates. Marker loss per generation was calculated by dividing the number of colonies that did not grow on −ura plates by the total number of colonies on replica SC plates and dividing this value by the total number of generations. In total, 73 000 colonies were counted for (A) and 69 000 colonies for (B). Each graph represents pooled data normalized to the appropriate JS128 parent 2% glucose control value, expressed as mean + standard error of the mean. Statistical analysis was performed using Student’s t-tests (*P<0.05).

Fig. 6

Genetic mimics of calorie restriction reduce rDNA recombination independently of rDNA silencing. (A) Qualitative silencing assay. The rDNA silencing reporter parent strain, JS128, and its isogenic sch9Δ, tor1Δ and sir2Δ strains were grown up in YPD, then diluted in H₂O and serially spotted out on synthetic complete media (SC) and SC − ura media (−ura). Spots correspond to OD₆₀₀ = 1 (left) then serial 1 in 5 dilutions to the right. Plates were imaged after 2 (0.5% and 2% glucose) and 3 days (0.05% glucose) at 30 °C. Reduced growth on −ura compared with SC indicates increased silencing in this assay. (B) Quantitative silencing assay. Threefold serial dilutions of overnight cultures of JS128 and its isogenic hxk2Δ, sch9Δ and tor1Δ strains were spread on plates and colonies were counted after 2 days. Silencing is expressed as a function of viable clones on −ura by dividing the numbers of colonies on −ura by numbers of colonies on SC. 29 000 colonies were counted in total. The graph shows pooled data normalized to the parent 2% glucose control value, expressed as mean + standard error of the mean. Statistical analysis was performed using Student’s t-tests (*P<0.05). (C) Recombination assay. JS128 and its isogenic hxk2Δ, sch9Δ and tor1Δ strains were serially cultured in YPD for approximately 120 generations. These were then spread onto YPD plates and then replica plated onto both SC and −ura plates. Marker loss per generation was calculated by dividing the number of colonies that did not grow on −ura plates by the total number of colonies on replica SC plates and dividing this value by the total number of generations. 43 000 colonies were counted in total. The graph shows pooled data normalized to the parent 2% glucose control value, expressed as mean + standard error of the mean. Statistical analysis was performed using Student’s t-tests (*P<0.05).