The S. cerevisiae BY4741 strain background (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) was used for all experiments. Deletion mutants were obtained from the EUROSCARF deletion collection. The deletion mutant pim1∆ was harboring second site mutations, therefore the strain was recreated for this study. The cells were cultivated at 28°C in complex medium (YPD/YPGal (1% (w/v) yeast extract, 2% (w/v) peptone and 2% (w/v) D-glucose/galactose) or synthetic complete glucose/galactose medium (SC-glucose/galactose (2% (w/v) D-glucose/galactose, 0.17% (w/v) yeast nitrogen base without amino acids, 0.5% ammonium sulphate and 10 ml of complete dropout mixture (0.2% Arg, 0.1% His, 0.6% Ile, 0.6% Leu, 0.4% Lys, 0.1% Met, 0.6% Phe, 0.5% Thr, 0.4% Trp, 0.1% Ade, 0.4% Ura, 0.5% Tyr per liter) under constant shaking. Galactose media were prepared for strains with vectors, which are harboring a galactose-promotor (pESC-His). For oleate and olive oil experiments (0.05% oleate (v/v) with Tween-80) synthetic complete media with glucose as carbon source were prepared. Solid media were made by adding 2% (w/v) agar. Selection for plasmids was ensured via leaving out the respective amino acid(s). In the Supplementary Table S1 all used and created yeast strains are listed.

BY4741 genomic DNA was isolated from an overnight grown (3 ml YPD) culture. After washing with H₂O, the cell pellet was resuspended in 500 µL SCE (1 M sorbitol, 20 mM EDTA, 10 mM Na-citrate, pH 7) and 40 µL zymolyase (10 mg/ml). The suspension was incubated for 60 min at 37°C under constant shaking. Cell lysis was executed via addition of 60 µL 10% SDS for 30 min at 65°C without shaking. By using 200 µL 5 M potassium acetate pH 5–5.5 on ice for 60 min protein precipitation was performed. After centrifugation at 14,000 rpm for 10 min, the supernatant was mixed with 700 µL isopropanol and the DNA was precipitated at −20°C. The suspension was centrifuged at 14,000 rpm for 5 min. After washing with 70% EtOH, the pellet was resuspended in H₂O. PCR was performed using Phusion High-Fidelity DNA Polymerase (NEB, Ipswich, MA, United States). For ARE2 the primers fwd1 (gtc aag gag aaa aaa ccc cgg atc cAT GGA CAA GAA GAA GGA TC) and rev1 (aaa tca act tct gtt cca tgt cga cTT AGA ATG TCA AGT ACA ACG TAC) were used, for LOR1 the primers fwd2 (ctca cta aag ggc ggc cgc aAT GGG CAC ACT GTT TCG AAG) and rev2 (atc ctt gta atc cat cga taT TAC ATT GGG AAG GGC ATC) (capital letters: complementary regions to the gene of interest; lowercase letters: complementary regions to the vector). Gel elution and clean-up was performed via Wizard^®SV Gel and PCR Clean-Up System (Promega, Mannheim, Germany). The vector pESC-His was linearized using the restriction enzymes BamHI-HF and SalI-HF (NEB; Ipswich, United States). Using Gibson Assembly^® Master Mix (NEB; Ipswich; United States) according to the manufacturer’s protocol ARE2 was cloned into pESC-His ARE1 (Bischof et al., 2017). Respectively LRO1 was cloned into the vector pESC-His DGA1 (Bischof et al., 2017). Constructs were sequenced by Eurofins-MWG-OPERON (Ebersberg, Germany).

Via homologous recombination, the gene PIM1 was replaced by a nourseothricin resistance cassette that was amplified from the vector pSDS4 (Lettner et al., 2010) using GoTaq DNA Polymerase (Promega, Mannheim, Germany) and the primers pim1Δ fwd (TTT TCT TTT GGT TTT CGA GGT GCT TGA ACG AAA AGA TTT GCA AAT AGA) and pim1Δ rev (ATA TTT ACA GAA TGT TTA AAC AGG TAT TTA ATC CAT TTA GAT GAA AAG CTG CAG AGG TAA ACC CAG A). After gel elution and clean-up via the Wizard^®SV Gel and PCR Clean-Up System (Promega, Mannheim, Germany), the strain BY4741 was transformed with the deletion cassette (Yeast transformation and Genomic Integration Section) and the genomic integration was selected by growth on YPD plates containing 100 µg/ml nourseothricin. A correct integration was controlled by PCR using the primers Pim1 A (GAG AAG ACA AAA CCA GGT GGT AGA T) and Pim1 D (CTT CTT AGA AAA GAG GCA AAG AGG T).

The strain BY4741 was grown to an optical density (OD₆₀₀) of 0.6–0.8 prior to harvesting. After centrifugation at 3,500 g for 3 min the cells were washed with LiAc/TE (100 mM Tris, 10 mM Tris, 1 mM EDTA, pH 8.0) and resuspended in 200 µL LiAc/TE. 50 µL of this cell suspension was mixed with 5 µg plasmid DNA, 10 µg/ml single-stranded salmon sperm DNA and 300 μL LPT (100 mM LiAc, 10 mM Tris, 1 mM EDTA, pH 8.0, 50% PEG 3350). This mixture was incubated at 28°C for 30 min under constant shaking. 40 μL DMSO was added and the cells were heat-shocked for 15 min at 42°C. After mixing with 1 ml sterile H₂O, the suspension was plated on the respective selective media plates. In case of genomic integration the cells were recovered in YPD medium for 2 h at 28°C under constant shaking prior to plating.

The elutriation centrifugation was performed as described in Klinger et al. (2010) to obtain a separation of young (is defined as either one or two generations) and old (which is defined as at least fifteen generations) yeast cells according to their replicative age. A workflow of the whole protocol is presented in Figure 1. A 10 ml overnight culture was diluted to an OD₆₀₀ = 0.1 in 200 ml either complex or synthetic medium. After 24 h of growth the first elutriation round was executed. The elutriation was performed with the Beckman elutriation system (Beckman Coulter Inc., Brea CA, United States) and the rotor JE-6B with a standard elutriation chamber (flow rate 10 ml/min). Prior to this process, cells were washed twice with 1xPBS. After resuspension in 10 ml 1xPBS, yeast mother and daughter cells were separated via sonification. Cells were loaded into the elutriation chamber with a rotor speed of 3,500 rpm and a flow rate of 10 ml/min yielding fraction I (virgin cells). Reduction in rotor speed (2,700 rpm) yields fraction II (young cells). Further reduction of the rotor speed to 2,400 rpm yields fraction III and 2000 rpm fraction IV (middle aged cells). Fraction V (old cells) was obtained at a rotor speed of 1,350 rpm. Fraction III, IV and V were reinoculated in 300 ml complex or synthetic medium containing 100 mg/L ampicillin. After 2 days growth at 28°C the second elutriation was conducted to obtain young and old cells. For the FACS analysis main cultures were prepared in SC, SC-Gal, YPD or YP-Gal (OD₆₀₀ 0.1). After incubation for 24 h a first elutriation harboring fraction III-V was conducted. Care was taken that the cells reached stationary phase before elutriation or FACS analysis. Then the cells (OD₆₀₀ 0.1) were further cultivated in 10 ml BES-buffered (100 mM BES at pH 7.5) media containing 100 mg/L ampicillin.

2.5 × 10⁷ cells were washed two times with 1x PBS. After centrifugation for 3 min at 3,500 rpm the pellet was resolved in 500 µL PBS-DHE (1:1,000 dilution of a 5 mg/ml DHE stock, Sigma Aldrich—37,291). The samples were incubated for 30 min in the dark without shaking. 200 µL of the solution were pipetted into each well and the fluorescence was measured with Anthos Zenyth 3,100 (Anthos Labtec Instruments GmbH, Salzburg, Austria). Excitation was set at 535 nm and emission was detected at 625 nm for 4 s.

After washing two times with 1x PBS, 1 × 10⁷ cells in a volume of 225 µL were pipetted into a 96-well plate. The cells suspension was mixed with 25 µL formaldehyde (37%). After addition of 1 µL Nile red (0.001 mg/ml in acetone, Thermo Fisher Scientific, N-1142), the plate was incubated for 20 min in the dark on a shaker and the fluorescence was measured using the Anthos Zenyth 3,100 (Anthos Labtec Instruments GmbH, Salzburg, Austria). Excitation was set at 485 nm and emission was detected at 595 nm for 0.4 s. Because elutriation yields were low cell numbers, a modified protocol was used after elutriation: 0.5 × 10⁷ cells and 0.002 mg/ml Nile red in acetone without formaldehyde fixation were used.

Yeast cells harboring the aging reporter (Supplementary Table S1) were elutriated as indicated in Elutriation Section and Figure 1.

After either one or two elutriation rounds the cells were analyzed using the FACS CytoFLEX S (Beckman Coulter, United States) equipped with a laser (excitation wavelength 488 nm) and a GFP filter (emission wavelength 510 nm; 20 nm width) with 100,000 events and a medium flow rate of 30 µL/min. Gating was performed to obtain GFP fluorescence over a certain threshold (above autofluorescence).

Cells were cultured in selective citrate phosphate buffered media [either SC-glucose or SC-galactose, buffered with 64.2 mM Na₂HPO₄, 17.9 mM citric acid, pH 6.0 (Wu et al., 2013)]. A typical overnight culture was inoculated in either YPD or SC medium. These cultures were diluted in 100 ml buffered media to an OD₆₀₀ of 0.1. The strains were cultivated over 4 weeks under constant shaking at 28°C. Water loss, after weighing the cultures, was constantly compensated. Survival plating was performed every day. Every week DHE and Nile red stainings were performed. The survival integral (SI), meaning the area beyond the lifespan curves were calculated with online available tools (https://www.desmos.com/calculator/be5ne9vwi8).

Microscopical analysis was carried out with a Nikon (Tokyo, Japan) Eclipse Ni-U equipped with a DS-Fi2 digital camera, a Nikon Eclipse Ti2 (Tokyo, Japan) and a Leica DMi8 microscope (Wetzlar, Germany).

Overnight cultures (BY4741 pESC-His, BY4741 pESC-ARE1/ARE2 and BY4741 pESC-LRO1/DGA1) were diluted in 2% YPGal (OD₆₀₀ = 0.1) and grown for 48 h at 28°C under constant shaking (600 rpm). Oxygen consumption of 10⁸ cells was measured by using an Oxygraph 2k at 28°C (Oroboros Innsbruck, Austria).

Data are presented as standard deviations ±SD. Data were tested using one-way ANOVA followed by a TUKEY post hoc test or unpaired two-tailed Student’s t-test, and results with p < 0.05 were considered statistically significant.

Recently, we were able to demonstrate that LDs fulfil an important role in the cellular stress management of S. cerevisiae (Bischof et al., 2017). Cells devoid of LDs become sensitive to the application of stressors such as acetic acid or hydrogen peroxide, whereas yeast cells harboring a surplus of LDs show the opposite phenotype. Furthermore, LDs are perfect biomarkers for cellular stress. Immediately after stress induction the LD content increases (Bischof et al., 2017). Based on these findings, we wanted to test the role of LDs in the aging process of yeast cells. In a first approach, replicatively aged cells were isolated via elutriation. This special counterflow centrifugation technique takes advantage of the fact that during aging the cell size as well as the sedimentation coefficient increases (Klinger et al., 2010; Geltinger et al., 2020b). Usually a 300–500 ml cell culture in stationary phase is separated into four fractions (fraction II, III, IV and V). Fraction II represents young and fraction V old cells. Fraction II and fraction V cells were stained with 0.002 mg/ml Nile red. This dye shows a high affinity for neutral lipids and is used as a selective “yellow-gold” or red fluorescent probe for LDs (Greenspan et al., 1985). Fluorescence microscopy clearly revealed a strong accumulation of LDs in aged cells (Figure 2A). This finding was confirmed by fluorometric measurements which showed a two-fold increase in LD content in fraction V compared to fraction II (Figure 2B).

The replicative lifespan is based on the observation that yeast cells can perform a limited number of cell divisions. These cell divisions are asymmetrically producing a larger mother cell that ages and a smaller daughter, which rejuvenates itself. The classical way of measuring the replicative lifespan of yeast cells is very tiresome and is dependent on the usage of a micromanipulator to remove and count the constantly produced daughter cells (Mortimer and Johnston, 1959). Therefore we recently established a new method based on an “aging reporter”, which is less labor intensive, faster and allows the screening of several yeast strains in parallel (Streubel et al., 2018). This aging reporter is composed of a cell cycle specific promoter (HO promoter) and the green fluorescent protein GFP. Each time the mother cell divides GFP is expressed. Hence, aged cells show a much brighter fluorescence than younger cells. In this initial publication we performed two subsequent elutriations. The first elutriation is necessary to increase the amount of replicatively aged yeast cells. In this first elutriation fraction II was discarded and fraction III-V were reinoculated and grown for further 2 days (Figure 1). After the second elutriation round aged cells can be isolated in a sufficient amount to perform further studies. To measure the replicative lifespan the increase of fluorescence between fraction II (young cells) and fraction V (old cells) was analyzed using a fluorometer. We showed that both methods, the “classical micromanipulation” and the “aging reporter“, delivered absolutely comparable results concerning some life prolonging interventions [for further details (Streubel et al., 2018)]. In our current study, we further tried to simplify the workflow. In Supplementary Figure S1A FACS analysis data of yeast cells transformed with the vector YCplac111 are presented, whereas in Supplementary Figure S1B cells harboring the aging reporter (YCPlac111-HOprom.-GFP) are shown. As expected cells expressing GFP show a much stronger fluorescence. Afterwards the strain BY4741 YCPlac111-HOprom.-GFP was elutriated two times and fraction II (young cells), fraction III (middle-aged cells) and fraction V (old cells) were analyzed separately using the FACS CytoFLEX S (Beckman Coulter, United States) (Figure 3). A constant age-dependent increase in GFP fluorescence was observed, confirming the functionality of the aging reporter.

Furthermore, it was tested if there is a necessity for elutriation at all. Therefore, cells were elutriated once and fraction III-V were reinoculated for 48 h. These cells were then analyzed directly by FACs without a second elutriation round and were compared to cells that have undergone no elutriation at all (Supplementary Figure S2). It is quite evident that the number of GFP fluorescent and thus aged cells is enormously increased after one elutriation round. All further experiments were then performed with one elutriation (Figure 1).

To test the role of LDs during mother cell specific aging, we chose a mutant strain that is deficient for the gene SEI1 encoding the yeast seipin. Seipin is responsible for controlling three parameters of LDs in yeast cells: number of LDs, morphology and size. It is published that in a sei1∆ mutant strain either supersized or small clustered LDs can be observed (Fei et al., 2008; Wang et al., 2014). Quantification of the LD content (Figure 4A) revealed no significant difference between BY4741 and BY4741 sei1∆, but fluorescence microscopy demonstrated differences in LD size and distribution (small and clustered LDs, Figure 4B). After transformation with the aging reporter, elutriation and growth for further 2 days the control strain (BY4741 YCPlac111-HOprom.-GFP) and the deletion mutant (BY4741 sei1∆ YCPlac111-HOprom.-GFP) were analyzed using the FACS CytoFLEX S. In the deletion mutant strain a clear reduction of cells with a high GFP fluorescence signal was observed (a 1.85-fold decrease of aged cells), indicating a reduction of the replicative lifespan (Table 1; Figures 4C,D). As a second candidate the acyltransferase Lro1p was chosen. This enzyme catalyzes the reaction of diacylglycerols to triacylglycerols. Upon deletion, the LD content decreased and the amount of replicatively aged cells as shown in Table 1 is nearly halved. Because we showed that during stress response a close interaction between mitochondria and LDs exists (Bischof et al., 2017; Geltinger et al., 2020b), this organelle was further analyzed.

A hallmark of aging is a reduced efficiency of mitochondrial respiration, resulting in a premature leakage of electrons, which are transferred to O₂ (Lopez-Otin et al., 2013). As a consequence superoxide is produced which can be measured by specific dyes such as dihydroethidium (DHE) (Chen et al., 2013). After elutriation, we compared the DHE levels of fraction II and V. Confirming Harman’s famous observations (Harman, 1956), we saw a clear increase (1.36-fold) in superoxide levels in aged cells. The observed ROS levels further increased in the SEI1 deletion mutant (1.53-fold), even if this difference was not statistically significant (Figure 4E).

To confirm the previously mentioned findings, ldb16∆ cells were also analyzed. The LDB16 gene encodes a Sei1p interacting protein that is responsible for targeting Seip1p to ER-LD contact sites (Wang et al., 2014). Identical to the BY4741 sei1∆ strain, the ldb16∆ mutant strain shows a strong reduction in replicative lifespan (Table 1).

The findings so far clearly indicate that LDs fulfill a supportive role and the decline in LD numbers or the change in morphology reduces the yeast replicative lifespan. Therefore, we wanted to test, if LDs have the capacity to prolong the lifespan in yeast cells. In a first approach several methods were analyzed which could promote the cellular LD content: Recently we demonstrated that a simultaneously overexpression of DGA1 and LRO1 increases LD numbers. Both genes encode diacylglycerol acyltransferases leading to a surplus of triacylglycerols that are stored in LDs (Bischof et al., 2017). Besides triacylglycerols, LDs also contain sterol esters, which are produced by two Acyl-CoA:sterol acyltransferases (Are1p and Are2p) in yeast cells (Yang et al., 1996). A third way to stimulate LD formation is the addition of the mono-unsaturated fatty acid oleate, the main component of olive oil (Wilfling et al., 2013; Bischof et al., 2017). After Nile red staining all these interventions were analyzed either fluorometrically (Figure 5A) or via fluorescence microscopy (Figures 5B–E). With the exception of an Are1p/Are2p overexpression (BY4741 p416GPD-ARE1 pESC-ARE2) each intervention leads to a significant increase in LD content (Figure 5A). In case of oleate the effect is concentration dependent with a peak at 0.05% oleate. A change in the morphology of LDs was also observed. Addition of 0.05% oleate leads to supersized LDs that completely fill the cell (Figure 5C). In contrast to this finding, a Lro1p/Dga1p overexpression leads to a modest reduction in LD size and to an exploding LD number (Figure 5D). An Are1p/Are2p overexpression had no obvious effect on either LD size or LD number (Figure 5E).

All these strains and interventions were tested using our aging reporter. After yeast cell transformation with the vector YCPlac111-HOprom.-GFP and a one-time elutriation, the GFP signal was measured via FACS analysis. The strongest GFP signal was obtained for the strain BY4741 p416GPD-LRO1 pESC-DGA1. Compared to the strain BY4741 p416GPD pESC-HIS a more than 2-fold increase in aged cells was observed (Table 1; Figure 6). This result was confirmed by fluorometric measurements. The strains BY4741 p416GPD-LRO1 pESC-DGA1 and BY4741 p416GPD pESC-HIS were elutriated twice. The second elutriation yielded young as well as old cells and the fluorescence of these two fractions (II and V) was analyzed separately. Comparing these two fractions the increase in fluorescence (and thus age) is more obvious in the strain BY4741 p416GPD-LRO1 pESC-DGA1 than in the control strain (Figure 6A). A further control experiment was performed by using a one-vector system instead of a two-vector system. Both LR O 1 and DGA1 were cloned into the vector pESC and both genes were co-expressed from the bidirectional promoter GAL1/10. FACS analysis (Table 1) yielded a more than 3-fold enrichment in aged cells. Both LRO1 and DGA1 were also analyzed separately. Dga1p and Lro1p single-overexpression led to a ∼1.5-fold increase in aged cells (Table 1). This result is clearly indicating that a co-overexpression of Dga1p/Lro1p is life prolonging.

As a co-overexpression of Are1p/Are2p leads to an only modest increase in LD numbers (Figure 5A), the life prolonging effect of this genetic intervention is quite low (Table 1). Only a 1.1-fold non-significant increase in aged-cells was observed. Surprisingly, either an Are1p or Are2p overexpression results in a 1.7-fold or 2.1-fold increase in lifespan, respectively. Non-genetic interventions such as treatment with 0.05% oleate or 0.05% olive oil had no advantageous, but on the contrary detrimental effects (Table 1). This is most probably attributed to the strong morphological changes that LDs have gone through after treatment with these two compounds (Figure 5). Pursuing the life prolonging effect upon LD stimulation, we wanted to test if mitochondria are involved in this process. In healthy cells, mitochondria form a tubular network, whereas in stressed, sick and aged cells, the mitochondrial network starts to fragment (Klinger et al., 2010). In our view, mitochondria are a perfect marker for cellular health. For the visualization of the mitochondrial network cells were transformed with the vector pYX142 harboring GFP fused to a mitochondrial targeting sequence (pYX142 mtGFP) (Westermann and Neupert, 2000). In fact, the mitochondrial network completely collapsed in aged yeast cells that were isolated via elutriation (Figure 7A, Supplementary Movie S1). In cells with boosted LD levels (achieved by either a Lro1p/Dga1p co-overexpression or Are1p/Are2p co-overexpression) no excessive mitochondrial fission during aging was observed [Figure 7B and Figure 7C, Supplementary Movie S2 (Lro1p/Dga1p co-overexpression) and Supplementary Movie S3 (Are1p/Are2p co-overexpression)]. As a second mitochondrial marker, ROS production was monitored. DHE measurements revealed an age dependent increase in superoxide levels in the wildtype strain (BY4741 pESC-His) (Figure 7D). LD stimulation by a Lro1p/Dga1p co-overexpression (BY4741 pESC-LRO1/DGA1) showed already a significant effect in young cells. In the latter strain the superoxide levels are more than 5-fold decreased (Figure 7D). This finding is even more astonishing when the mitochondrial respiration is taken into consideration. Oxygraph measurements with cells that were grown for 48 h in 2% YPGal were performed. BY4741 pESC-His cells showed a respiration of 28 +/− 4 pmol/(sec*10⁷ cells), BY4741 pESC-LRO1/DGA1 cells a respiration of 61 +/− 13 pmol/(sec*10⁷ cells) and BY4741 pESC-ARE1/ARE2 cells a respiration of 42 +/− 13 pmol/(sec*10⁷cells) (One Way ANOVA; p < 0.05).

There is increasing evidence that in times of stress mitochondria have a central role in protein homeostasis. Protein aggregates that are formed in the cytosol are transported to mitochondria. After import into the mitochondrial matrix, these misfolded proteins are degraded by the LON protease Pim1p. This process was termed MAGIC (mitochondria as guardian in cytosol) (Ruan et al., 2016). Additionally, we demonstrated that upon stress application and aging LDs get in close contact with mitochondria and a shuttling of proteins from mitochondria to LDs occurs (Bischof et al., 2017; Geltinger et al., 2020b). Most probably, this protein transfer supports MAGIC. Therefore, we wanted to test, if the interplay of mitochondrial protein degradation and “LD shuttling” modulates the aging process in yeast cells. In a PIM1 deletion mutant several mitochondrial parameters are impaired. It has to be stated that the pim1∆ strain shows no growth on non-fermentable carbon sources (glycerol)/respiratory media, indicative for a petite-like phenotype (Supplementary Figure S3). Consequently, the doubling time in the pim1∆ strain is increased. The strain BY4741 showed a doubling time of ∼100 min, whereas BY4741 pim1∆ cells divided every ∼180 min (Figure 8A). After two elutriation rounds, the mitochondrial superoxide production was measured by a DHE-assay comparing fraction II and V. Compared to the wild-type, an enormous age-specific increase in O₂ ⁻ levels was observed (Figure 8B). This defect in the pim1∆ background also manifests on mitochondrial morphology. BY4741 and BY4741 pim1∆ cells were transformed with the vector pYX142 mtGFP. As visualized by fluorescence microscopy mitochondria form long tubules that fill the mother cells as well as daughter cells (Figure 8C, Supplementary Movie S4). After PIM1 deletion the mitochondrial network completely collapses and only small, roundish mitochondrial “blobs” remain (Figure 8D, Supplementary Movie S5). This extreme phenotype was observed in 100% of the cells. Compared to the wildtype the GFP signal was very faint in the deletion mutant, indicating a loss of the mitochondrial membrane potential, which is a prerequisite for the mtGFP import. Microscopical investigations revealed another obvious phenotype: The LDs in the pim1∆ strain were unusually enlarged (Figure 8E). This finding was confirmed by a Nile red staining and fluorometric measurements that showed a 4-fold increase in LDs in the strain defective for the LON protease (8F). Utilizing the vector-based aging reporter, we observed that the GFP signal in the pim1∆ strain decreases (Figures 8G,H). In dependence of the strain background (either BY4741 pim1∆ p416GPD YCplac111 HO-Prom.-GFP or BY4741 pim1∆ pESC-HIS YCplac111 HO-Prom.-GFP), the amount of aged cells compared to the respective control strains is reduced in the range from 1.8 to 3.8-fold (Table 1).

In further experiments, we tried to reverse some of these premature aging phenotypes by stimulating the LD content. An increase of the LD content by a LR O 1 overexpression (transformation with the vector p416GPD-LRO1) clearly had an enhancing effect on the growth rate of a pim1∆ strain. The doubling time decreased from ∼180 min in the deletion mutant (BY4741 pim1∆ p416GPD) to 104 min in the strain BY4741 pim1∆ p416GPD-LRO1 (Figure 9A).

In addition, the mitochondrial morphology was recovered by boosted LD levels. Transformation of cells with the before-mentioned vector induced a complete shift of the fragmented towards a tubular mitochondrial network in the pim1∆ strain (observed in ∼80% of all cells) (Figure 8D and Figure 9B). As can be seen in Supplementary Movie S6, the recovery was not effective in each single cell, in ∼20% of all cells a partial recovery was observed. Despite the fact that an Lro1p overexpression in the pim1∆ background further boosted the amount of neutral lipids, a decrease in LD size was observed (Figures 9C,D). In case of replicative aging this phenotype is even more distinct. As can be seen in Figures 9E–G, a LR O 1 overexpression is not only compensating the aging defect in the pim1∆ mutant, but is also leading to a prolonged lifespan. The strain BY4741 pim1∆ p416GPD-LRO1 showed a 1.6-fold increase in aged cells compared to the corresponding wildtype and a 2.8-fold increase compared to the pim1∆ strain (Table 1).

In contrast to the replicative lifespan, chronological aging is focusing on postmitotic cells. This aging process is measured by a loss of viability during stationary phase. In a meta-analysis, we compared the chronological and replicative aging process as well as the underlying pathways and found no significant overlap between these two aging mechanisms (Laun et al., 2006). Nonetheless, the role of LDs during chronological aging was studied. The following yeast strains were cultured for 25 days in buffered SC medium and cell survival was studied using a survival plating assay: BY4741; BY4741 pESC p416GPD; BY4741 pESC p416GPD- DGA1; BY4741 pESC p416GPD- LR O 1; BY4741 pESC-DGA1 p416GPD-LRO1; and BY4741 are1∆ are2∆ lro1∆ dga1∆. Survival curves for selected yeast strains (BY4741 pESC p416GPD; BY4741 pESC-DGA1 p416GPD-LRO1; and BY4741 are1∆ are2∆ lro1∆ dga1∆) are presented in Figure 10A, the survival integrals can be found in Supplementary Table S2. The wildtype showed a reduced survival rate after 5 days, the decrease was slowed down and stayed at 20% after 10 days. A strain completely devoid of LDs [BY4741 are1∆ are2∆ lro1∆ dga1∆ (Bischof et al., 2017)] started to show a sharp decrease in cell survival immediately after the first day, leading to no surviving cells after 15 days in culture. A strain with increased LD numbers (BY4741 pESC-DGA1 p416GPD-LRO1) showed similar death kinetics as the wildtype strain but the survival rate reached a constant level at ∼50% cell survival. From the survival curves quantifiable “survival integrals” (SI) were calculated (Murakami and Kaeberlein, 2009). These SIs confirm prolonged chronological lifespans for all LD stimulation interventions (LRO1 overexpression; DGA1 overexpression; and DGA1/LR O 1 co-overexpression) and clearly demonstrate a loss of viability during stationary phase for the strain BY4741 are1∆ are2∆ lro1∆ dga1∆ (Supplementary Table S2). In addition to survival plating DHE measurements were performed every week. In case of the wildtype a constant increase in superoxide levels was observed. The same increase was monitored for the strain with the co-overexpression of Lro1p/Dga1p (BY4741 pESC-DGA1 p416GPD-LRO1), although on clearly reduced levels (Figure 10B). A strain devoid of LDs (BY4741 are1∆ are2∆ lro1∆ dga1∆) showed a burst of ROS already during the first week. A similar high DHE level was observed in the strain BY4741 pESC-DGA1 p416GPD-LRO1 after 4 weeks of cultivation.