The wild-type (WT) strain S. cerevisiae BY4742 (MATα his3Δ leu2Δ0 lys2Δ ura3Δ0) from Thermo Scientific/Open Biosystems and long-lived mutant species 3, 5, and 12 derived from this strain were grown in liquid YP medium (1% yeast extract, 2% peptone) initially containing 0.2% glucose as a carbon source, with or without LCA, as detailed in Section “Results.” Cells were cultured at 30°C with rotational shaking at 200 rpm in Erlenmeyer flasks at a “flask volume/medium volume” ratio of 5:1.

A sample of cells was taken from a culture at a certain day following cell inoculation into the medium. A fraction of the sample was diluted in order to determine the total number of cells using a hemocytometer. Another fraction of the cell sample was diluted and serial dilutions of cells were plated in duplicate onto YEP [1% (w/v) yeast extract, 2% (w/v) peptone] plates containing 2% (w/v) glucose as carbon source. After 2 days of incubation at 30°C, the number of colony forming units (CFU) per plate was counted. The number of CFU was defined as the number of viable cells in a sample. For each culture, the percentage of viable cells was calculated as follows: (number of viable cells per ml/total number of cells per ml) × 100. The percentage of viable cells in mid-logarithmic growth phase was set at 100%.

A small patch of cells of mating type α [i.e., the haploid WT strain BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0)] was applied to the surface of a master YEPD (1% yeast extract, 2% peptone, 2% glucose, 2% agar) plate. 10⁶ cells of mating type α [i.e., the haploid WT strain BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) or the selected long-lived haploid mutant strains 3, 5, or 12] were spread on the surface of a separate crossing plate with YEPD medium. The master plate was replica plated onto a lawn of cells on each of the four crossing plates; different velvet was used for each crossing plate. The crossing plates were incubated overnight at 30°C. Each of the four crossing plates was then replica plated onto a synthetic minimal YNB medium plate (0.67% yeast nitrogen base without amino acids, 2% glucose, 2% agar) supplemented with 20 mg/l L-histidine, 30 mg/l L-leucine and 20 mg/l uracil. These plates were incubated overnight at 30°C. A positive mating reaction between cells of the haploid WT strain BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and cells of the haploid WT strain BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) or cells of each of the selected long-lived haploid mutant strains 3, 5, or 12 resulted in confluent growth of diploid cells on a YNB plate (supplemented with L-histidine, L-leucine, and uracil) at the position of a patch of haploid BY4741 cells.

Cells of the diploid strains recovered in the mating assay were spread on a separate SPO (0.1% yeast extract, 1% potassium acetate, 0.05% glucose, 2% agar) plate supplemented with 20 mg/l L-histidine, 30 mg/l L-leucine, and 20 mg/l uracil. The plates were incubated at 30°C for 5–6 days, until numerous asci were observed microscopically.

An inoculating loop was used to resuspend spores from the lawn on SPO plates in 50 μl of 0.2 mg/ml Zymolyase 20T solution in sterile water. The suspension of spores was incubated for 10 min at room temperature. A total of 450 μl of sterile water was gently added to the suspension of spores. The mix was then incubated for 30–40 min at room temperature. An inoculating loop was used to spread an aliquot of digested asci onto a tetrad dissecting YEPDD (1% yeast extract, 2% peptone, 2% glucose, 4% agar) plate. Dissection of asci, separation of ascospores, and tetrad analysis were performed according to established procedures (Sherman, 2002; Amberg et al., 2005; Boone et al., 2016).

A sample of cells was taken from a culture at a certain time-point. Cells were pelleted by centrifugation and resuspended in 1 ml of fresh YP medium containing 0.2% glucose. Oxygen uptake by cells was measured continuously in a 2-ml stirred chamber using a custom-designed biological oxygen monitor (Science Technical Center of Concordia University) equipped with a Clark-type oxygen electrode.

For the analysis of hydrogen peroxide (oxidative stress) resistance, serial dilutions (1:10 to 1:10⁵) of WT and mutant cells removed from each culture at various time-points were spotted onto two sets of plates. One set of plates contained YP medium with 2% glucose alone, whereas the other set contained YP medium with 2% glucose supplemented with 5 mM hydrogen peroxide. Pictures were taken after a 3-day incubation at 30°C.

For the analysis of thermal stress resistance, serial dilutions (1:10 to 1:10⁵) of WT and mutant cells removed from each culture at various time-points were spotted onto two sets of plates containing YP medium with 2% glucose. One set of plates was incubated at 30°C. The other set of plates was initially incubated at 55°C for 30 min, and was then transferred to 30°C. Pictures were taken after a 3-day incubation at 30°C.

For the analysis of osmotic stress resistance, serial dilutions (1:10 to 1:10⁵) of WT and mutant cells removed from each culture at various time-points were spotted onto two sets of plates. One set of plates contained YP medium with 2% glucose alone, whereas the other set contained YP medium with 2% glucose supplemented with 1 M sorbitol (S6021; Sigma). Pictures were taken after a 3-day incubation at 30°C.

Statistical analysis was performed using Microsoft Excel’s (2010) Analysis ToolPak-VBA. All data on cell survival are presented as mean ± SEM. The p values for comparing the means of two groups using an unpaired two-tailed t-test were calculated with the help of the GraphPad Prism 7 statistics software. The log-rank test for comparing each pair of survival curves was performed with GraphPad Prism 7. Two survival curves were considered statistically different [and two strains were concluded to exhibit a significant difference in chronological lifespan (CLS)] if the p value was less than 0.05.

To empirically verify our hypothesis suggesting that hormetic selective forces can drive the evolution of longevity regulation mechanisms within ecosystems, we carried out a three-step selection of long-lived yeast species by a lasting exposure to LCA under laboratory conditions.

For the first step of such selection, yeast cells from an overnight culture were inoculated into a fresh, nutrient-rich YP medium (with or without LCA) containing 0.2% glucose to the initial cell titer of 1 × 10⁵ cells/ml (Figure 1). We chose 0.2% as the initial concentration of glucose in liquid medium for the LCA-driven selection of long-lived yeast species because the longevity-extending efficiency of LCA under caloric restriction (CR) on 0.2% glucose exceeds that under non-CR conditions on 2% glucose, a concentration typically used to culture yeast under laboratory conditions (Goldberg et al., 2009, 2010b; Arlia-Ciommo et al., 2014a). After 1 week of the incubation in this medium and following approximately 45 cell generations, yeast entered into a non-proliferative state of quiescence by reaching stationary (ST) growth phase at the cell titer of 2 × 10⁸ cells/ml (Figure 1). Following their entry into a non-proliferative state, yeast cells were cultured for additional 5 weeks. By the end of this cultivation, only ∼0.01% of cells in the medium without LCA remained viable. In the cell cultures that were supplemented with LCA, ∼1% of cells remained viable—due to the ability of LCA to increase the CLS of non-dividing (quiescent) yeast cells. Thus, the enrichment factor for cells surviving each selection step in LCA-treated samples was 10² (Figure 1). According to our hypothesis of the hormetic selective forces driving the evolution of longevity regulation mechanisms, the genomes of some of the cells surviving a selection step may contain mutations that extend yeast CLS. Noteworthy, it is conceivable that the genomes of other cells surviving a selection step may not undergo any changes affecting their CLS; the survival of these cells may have been caused only by the ability of LCA to extend yeast CLS. The following five cultures were used to carry out the selection of long-lived yeast species: the first culture lacked LCA; the second culture contained 5 μM LCA added at the moment of cell inoculation; the third culture contained 50 μM LCA added at the moment of cell inoculation; the fourth culture contained 250 μM LCA added at the moment of cell inoculation; and the fifth culture was supplemented with 10 doses of 5 μM LCA each, by adding a 5-μM dose of LCA every 3 or 4 days (Figure 1). Of note, LCA exhibited the greatest longevity-extending effect in chronologically aging yeast if used at a final concentration of 50 μM (Goldberg et al., 2010b). At the end of the first selection step, aliquots of each of the five cultures were plated onto plates with solid YP medium containing 2% glucose. After 2 days, each of the 10 randomly chosen colonies was inoculated into a liquid medium without LCA to carry out a viability testing step (Figure 1). Every week, an aliquot of each culture was used for a spot assay of cell viability to identify long-lived mutants (if any) induced due to a lasting exposure of yeast to LCA during the selection step (Figure 1). If such mutants were detected, an aliquot of their culture in medium lacking LCA was frozen at -80°C.

To conduct the second selection step, an aliquot of the culture recovered at the end of the first step of selection was diluted 100-folds in a fresh YP medium containing 0.2% glucose, with or without LCA; if added to a culture for the second selection step, LCA was present at the same final concentration as that in the medium used for the first step of selection (Figure 1). From that point, the second selection step was carried out exactly as the first one and was followed by a spot-assay viability step described above (Figure 1). The second selection step was followed by the third selection step and then by a spot-assay viability step, both conducted as described previously (Figure 1).

The three-step selection of long-lived yeast species depicted in Figure 1 was expected to increase the percentage of such mutant species within a population by the end of each selection step (Figure 2). Moreover, each next selection step was anticipated to yield the higher percentage of such mutants than the previous one (Figure 2; Table 1).

By carrying out this three-step selection of long-lived yeast species under laboratory conditions, we found no such mutant species at the end of the first selection step, four long-lived mutants at the end of the second selection step, and 16 long-lived mutants at the end of the third selection step (Figures 3A–C, respectively). We therefore concluded that a long-term exposure of WT yeast to LCA during three consecutive selection steps yielded yeast species that live longer in the absence of LCA than their ancestor.

Our findings revealed the following order of different LCA concentrations ranked by the efficiency with which they cause the appearance of long-lived species (the frequencies of such appearances are shown in parentheses; they were calculated as detailed in Table 1): 5 μM LCA (≈4 × 10^-8 per generation) > 10 doses × 5 μM LCA (≈3 × 10^-8 per generation) > 50 μM LCA (≈1 × 10^-8 per generation) > 250 μM LCA (no long-lived species found). Because the lowest used concentration of LCA resulted in the highest frequency of long-lived species appearance, it is unlikely that the longevity-extending mutations they carry are due to mutagenic action of LCA. Thus, these mutations are likely to arise spontaneously in yeast populations subjected to a lasting exposure to LCA.

The three-step experimental evolution of long-lived yeast species by a prolonged exposure of WT yeast to LCA resulted in selection of 20 species that in a spot-assay lived longer than their ancestor in medium lacking LCA (Figure 3). An aliquot of the culture of each of these long-lived yeast species was frozen at -80°C immediately after being recovered during the second or third selection step. To test the abilities of these species to maintain their considerably increased lifespans during the first passage in medium without LCA, each aliquot was thawed and then inoculated into liquid YP medium lacking this bile acid and initially containing 0.2% glucose. Our comparative analysis of the CLS of WT strain and all of these selected long-lived yeast species revealed that each of the 20 species maintains its ability to live longer than WT during the first passage in liquid medium (Supplementary Figure S1). Yeast species 3, 5, and 12 exhibited the highest extent of longevity extension during the first passage in medium without LCA (Supplementary Figure S1). These three yeast species underwent four more passages (each carried out as described for the first passage) in liquid medium without LCA; they then were tested again for their abilities to maintain greatly extended lifespans, now following five successive passages in medium without LCA. As we found, each of the three extremely long-lived yeast species evolved under laboratory conditions sustains its significantly prolonged CLS after five passages in medium without LCA (Figure 4).

Each of the three selected long-lived yeast mutants (each of mating type α and in the BY4742 genetic background, i.e., MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) was subjected to the backcross mating with the WT strain BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) of opposite mating type. A WT × WT diploid strain was created by mating between the haploid WT strains BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0). We compared the CLS of each, the WT × 3, WT × 5, and WT × 12 diploid strain, to those of the parental haploid mutant strain, parental haploid WT strain and WT × WT diploid strain; yeast cells in these experiments were cultured in YP medium lacking LCA and initially containing 0.2% glucose. We found that chronologically aging cells of the WT × 3, WT × 5, and WT × 12 diploid strains live (1) almost as long as chronologically aging cells of the parental haploid mutant strain; and (2) significantly longer than chronologically aging cells of the WT × WT diploid strain and the parental haploid WT strain (Figure 5). Both the mean and maximum CLS of each, the WT × 3, WT × 5, and WT × 12 diploid strain, exceeded those of the WT × WT diploid strain to a similar degree as the mean and maximum CLS of the parental haploid mutant strains exceeded those of the parental haploid WT strain (Figure 6). Based on these observations, we concluded that the extended longevity of each of the three long-lived yeast mutants is a dominant genetic trait.

To elucidate if the extended longevity of each of the three long-lived yeast mutants is a monogenic or polygenic genetic trait, we first subjected the WT × 3, WT × 5, and WT × 12 diploid strains to sporulation. We then removed the outer cell wall surrounding the four ascospores within individual asci by treatment with Zymolyase, and used micromanipulation to separate these ascospores. For each of the three sporulated diploids, we conducted tetrad analysis of ascospores from six randomly chosen asci (i.e., the four ascospores recovered within each ascus). We compared the CLS of individual meiotic segregants originated from the WT × 3, WT × 5, and WT × 12 diploid strains to the CLS of the parental mutant and WT strains. This comparative analysis revealed the following: (1) chronologically aging cells of all four ascospores within each of the six tested tetrads exhibit an extended CLS characteristic of the parental mutant strain and live significantly longer than cells of the parental WT strain; and (2) this relationship between the CLS of individual meiotic segregants, the parental mutant strain and the parental WT strain is observed for each of the three diploid strains, namely WT × 3, WT × 5, and WT × 12 (Supplementary Figures S2–S10). The observed relationship between the CLS of individual meiotic segregants, the parental mutant strain and the parental WT strain differed from a relationship that was expected if the extended longevity of a long-lived yeast mutant evolved under laboratory conditions was due to a single-gene mutation or mutations in two nuclear genes. Indeed, if the trait of extended longevity was monogenic, it was anticipated that two of the four ascospores within each of the six randomly chosen tetrads do not exhibit an extended CLS characteristic of the parental mutant strain and live almost as short life as cells of the parental WT strain (Sherman, 2002; Amberg et al., 2005; Boone et al., 2016). Furthermore, if the trait of extended longevity was caused by dominant mutations in two nuclear genes, it was projected that one or two of the four ascospores within three, four, or five out of the six randomly chosen tetrads do not exhibit an extended CLS characteristic of the parental mutant strain and live almost as short life as cells of the parental WT strain (Sherman, 2002; Amberg et al., 2005; Boone et al., 2016). Based on these observations, we concluded that the extended longevity of each of the three selected long-lived yeast mutants is a polygenic genetic trait caused by mutations in more than two nuclear genes.

One of the key aspects of our hypothesis of the hormetic selective forces driving the evolution of longevity regulation mechanisms is an assumption that all yeast species within an ecosystem can respond to LCA and other bile acids released into the ecosystem by animals (Goldberg et al., 2010a,b; Burstein et al., 2012a). This response consists in the ability of yeast to develop mechanisms of protection against cellular damage caused by the mildly toxic molecules of bile acids (Goldberg et al., 2010a,b; Burstein et al., 2012a). Of note, our analysis of how different concentrations of LCA impact yeast longevity has revealed that this bile acid delays yeast chronological aging by eliciting a hormetic stress response (Goldberg et al., 2010b; Burstein et al., 2012a), which is characterized by a non-linear and biphasic dose–response curve (Goldberg et al., 2010a; Calabrese and Mattson, 2011; Burstein et al., 2012a; Calabrese et al., 2012; Leonov et al., 2015; Lutchman et al., 2016). Our recent studies also demonstrated that the ability of LCA to elicit such longevity-extending stress response is due in part to alterations in mitochondrial functionality caused by LCA, including changes in the age-related chronology of mitochondrial respiration (Goldberg et al., 2010b; Burstein et al., 2012b; Beach et al., 2013, 2015a,b; Arlia-Ciommo et al., 2014a,b; Burstein and Titorenko, 2014; Medkour and Titorenko, 2016). These LCA-driven changes in mitochondrial respiration allow mitochondria to establish and sustain an aging-delaying pattern of the entire cell (Goldberg et al., 2010b; Burstein et al., 2012b; Beach et al., 2013, 2015a,b; Arlia-Ciommo et al., 2014a,b; Burstein and Titorenko, 2014; Medkour and Titorenko, 2016). Enhanced resistance of chronologically aging yeast to chronic oxidative, thermal, and osmotic stresses is part of such LCA-driven cellular pattern (Goldberg et al., 2010b; Burstein et al., 2012b; Beach et al., 2015b; Medkour and Titorenko, 2016); this kind of mitochondria-dependent hormetic stress response is called “mitohormesis” (Tapia, 2006; Schulz et al., 2007; Ristow and Zarse, 2010; Ristow, 2014; Yun and Finkel, 2014). We thought, therefore, that some of the three long-lived yeast mutants selected under laboratory conditions may have evolved an altered pattern of mitochondrial respiration and/or stress susceptibility. We found that in WT yeast grown in YP medium with 0.2% glucose in the absence of LCA, the rate of oxygen consumption by mitochondria was (1) greatly amplified when yeast entered diauxic (D) growth phase that begins on day 2 of culturing; (2) sharply declined during the subsequent post-diauxic (PD; occurs between days 3 and 7 of culturing) phase of growth; and (3) gradually reduced during ST (begins after day 7 of culturing) growth phase that follows PD phase (Figure 7). In contrast, in all three selected long-lived yeast mutants grown in this medium, the rate of oxygen consumption by mitochondria was (1) amplified to a significantly lesser extent during D phase than it was in WT yeast; (2) not declined as sharply during the subsequent PD phase of growth as it was in WT yeast; and (3) elevated again for a long period of time during ST phase (to the level similar to that seen in these yeast species during PD phase) and began to decline slowly only deep into ST phase—unlike the situation seen in WT yeast (Figure 7). Moreover, we found that each of the three selected long-lived yeast mutants grown in this medium displays enhanced cell resistance to chronic oxidative, thermal, and osmotic stresses, especially during D, PD, and ST growth phases (Figure 8).