Uncovering Indirect Genetic Modulators of Longevity Beyond SIR2 in Saccharomyces cerevisiae

Sulekha Osman, CAS '25

Bachelor of Science: Global Public Health/Biology

Advised by Dr. Mark Siegal

Abstract

Aging is a highly regulated biological process that impacts health and longevity. Yet, the genetic factors driving cellular aging and its variability across individuals remain poorly understood. Studies in the model organism Saccharomyces cerevisiae have provided key insights into the mechanisms of aging. One important discovery was SIR2, a gene that extends lifespan by promoting genomic stability through chromatin silencing. In humans, its homolog SIRT1 is involved in similar processes related to DNA repair and cellular stress. While SIR2’s molecular function is well studied, less is known about how natural genetic variation shapes lifespan through this pathway. My study goes beyond SIR2 by using quantitative trait locus (QTL) analysis to map broader genomic regions associated with lifespan, and focuses on chronological lifespan and the survival of non-dividing yeast cells in the stationary phase. I crossed two genetically distinct S. cerevisiae strains and performed survival-based selection followed by QTL mapping. This approach allows for the detection of modifiers and genetic interactions that may act upstream of the SIR pathway or that influence lifespan through mechanisms related to, but distinct from, SIR2 activity. I identified multiple longevity-associated regions, including candidates that regulate chromatin, stress response, and mitochondrial function. As the strongest associations do not localize within the SIR family itself, I conclude that the genes identified will likely be part of regulatory networks that interact with the SIR family or modulate its activity. The SIR family functions within a complex regulatory mechanism, and my findings support the idea that lifespan variation may arise from indirect genetic effects rather than direct genetic variation in the SIR family. Identifying these genetic contributors will provide insight into how natural variation influences aging and reveal new potential targets for aging-related interventions in humans.

Introduction

Aging is a fundamental biological process with extensive impacts on health and longevity. The genetic mechanisms underlying cellular aging have been explored in model systems such as S. cerevisiae. One of the most significant breakthroughs in the study of aging was the discovery of SIR2 in S. cerevisiae¹. Originally identified for its role in transcriptional silencing², SIR2 was later found to regulate lifespan in yeast. Past experiments have established that its overexpression can extend lifespan, whereas its deletion results in premature aging³. This discovery generated interest in the function of SIR2 across different organisms, especially humans, where its closest homolog is SIRT1. It laid the groundwork in aging biology, paving the way for therapies such as SIRT1-activating compounds (STACs)⁴, which target SIR2-related pathways.

Although SIR2 has been widely studied, significant gaps remain in our understanding of how natural genetic variation within this and related genes influences molecular aging processes. SIR2 belongs to a larger family of genes—the SIR family—which includes SIR1, SIR2, SIR3, and SIR4, and has homologs across various species⁵. My research examines how variation in genes, including but not limited to those in the SIR family, influences cellular aging and longevity using S. cerevisiae as a model organism. By characterizing this natural variation, my study aims to uncover new regulatory mechanisms of lifespan and to identify genetic variants that improve both adaptation and longevity, especially under nutrient stress conditions like uracil deprivation. These findings may provide broader insight into how conserved regulatory pathways shape aging across species.

SIR2’s role in repressing recombination and extending lifespan

Prior research has demonstrated that SIR2 is an NAD+-dependent histone deacetylase⁶ that regulates transcriptional silencing and genome stability in S. cerevisiae. SIR2 also functions as part of the Sir2/3/4 protein complex, repressing transcription at telomeres, ribosomal DNA (rDNA) regions, and the haploid mating (HM) loci². This repression is essential for maintaining genomic integrity by preventing harmful recombination events, particularly within rDNA, where excessive homologous recombination leads to extrachromosomal rDNA circles (ERCs)⁷. The accumulation of ERCs is a major contributor to yeast aging, creating a metabolic strain that eventually reduces lifespan⁷. SIR2 helps suppress ERC formation by silencing transcription at the rDNA locus, reducing the chances of replication fork stalling and recombination⁶. In addition, its deacetylase activity contributes to tighter chromatin packaging at rDNA, further limiting the accessibility of recombination machinery².

Identifying sources of natural variation in aging

This study focuses on chronological lifespan (CLS), which measures how long non-dividing cells survive⁸. CLS tracks survival in a quiescent state and relies heavily on a cell’s ability to adapt to nutrient stress, oxidative damage, and metabolic shifts that occur during the stationary phase. Chronological lifespan can be contrasted with replicative lifespan (RLS), which measures how often a mother cell divides⁸, which has contributed to some of the literature in the field. These two models highlight different aging processes, and genes like SIR2 may affect each one differently. Previous research has established that SIR2 influences lifespan in yeast; however, the impact of natural genetic variation at other loci on how the SIR family alters their functions remains unexplored. Identifying existing modulators of SIR2 or uncovering additional contributors could provide deeper insights into how natural variation impacts aging, including their potential effects in humans.

To investigate this, we capture natural variation by performing genetic crosses between distinct wild S. cerevisiae strains cultured from diverse environments, which have been exposed to different stressors⁹ over the course of their recent evolutionary history. Sporulating crosses between two genetically diverse yeast strains allow for the generation of large numbers of recombinant offspring that inherit different combinations of parental alleles. These recombinant offspring can then be selected for the desired phenotype and sequenced. By comparing allele frequencies in the selected versus unselected bulks, it becomes possible to pinpoint where one allele is favored over another, allowing for identification of quantitative trait loci (QTL) for the trait of interest on a genome-wide scale. I aim to find the specific genes that contribute to extended lifespan and compare whether they are associated with the SIR family and its regulatory network, or whether other genetic factors outside this pathway play a significant role in modulating aging.

In summary, despite extensive research on SIR2 and its role in lifespan regulation, there is a significant gap in understanding how natural genetic variation influences its function and that of related genes in aging. Previous studies have primarily focused on targeted genetic modifications and on replicative lifespan, rather than naturally occurring variation and its effects on chronological lifespan, leaving unanswered questions about how SIR family genes interact within diverse genetic backgrounds. However, by leveraging natural variation through QTL mapping and BSA, my study aims to fill this gap by identifying key genetic contributors to longevity. My work not only helps us understand how lifespan is regulated in yeast but could also point toward new genetic pathways relevant for the development of treatments for age-related decline in humans.

Materials and Methods

Strain selection and genetic requirements for mating

To determine how natural variation in yeast contributes to longevity, two divergent haploid yeast strains were selected, one isolated from an oak tree (MLS.CB.007, MATα) and the other from a wine barrel (NCYC 3591, MATa). To ensure proper crossing and selection, the strains had to meet several requirements: they needed to be of opposite mating types (MATα and MATa), and each had to carry a unique drug resistance marker to allow for diploid selection after mating. Additionally, the α strain carries HSV-TK (FuDR sensitivity), which is integrated into the MATα locus¹⁰ and acts as a negative selection marker after sporulation; 5-fluorodeoxyuridine (FuDR) eliminates MATα haploids, ensuring only MATa haploids remain to maintain a single mating type in the population (Fig. 1). The wine strain carries Hygromycin (HygMX) resistance, while the oak strain has Nourseothricin (Nat) resistance, allowing for selection of diploids that inherit both resistance markers. Both strains are Ura3 knockouts, which was required for selection in later nutrient deprivation experiments. URA3 encodes orotidine-5'-phosphate decarboxylase, an enzyme required for de novo synthesis of uracil¹¹. As a result, these strains cannot grow on synthetic complete media without uracil (SC-Ura) unless uracil is supplemented to the media.

Crossing genetically distinct strains and isolating haploid progeny

Following strain selection, both the oak strain (MLS.CB.007, MATα) and the wine strain (NCYC 3591, MATa) were cultured from frozen stock, each streaked out on a YPD plate to ensure there was no contamination, and then tested on plates under different conditions. Four types of plates were used: a standard YPD plate as a positive control (both strains grew), a +Nat plate where only the oak strain grew, a +Hyg plate where only the wine strain grew, and an SC-Ura plate, which neither strain grew on due to the ura3Δ knockout (negative control). To begin the mating procedure, 10 µL of saturated culture from each strain was mixed in an Eppendorf tube, pipetted vigorously, and plated in a small circle on a YPD plate. The plate was incubated at 30°C for 3 hours. Once mating had 24 hours to occur, potential diploid colonies were picked and streaked onto three plates: SC-Ura (to confirm lack of uracil synthesis), YPD (positive control), and a +Hyg/+Nat double-selection plate to confirm successful mating. Only diploids that inherited both resistance markers grew on the double-selection plate. Two diploid colonies were selected and grown in 3 mL of YPD supplemented with Hygromycin and Nourseothricin at 30°C overnight. The following day, cultures were back-diluted at a 1:50 ratio and incubated further to ensure sufficient growth before transfer to sporulation media. Sporulation was carried out in 2% potassium acetate (KAc) at room temperature (25 °C) for 11 days. To remove unsporulated diploids, cultures were spun down (1 min, 6000 RPM), resuspended in 50% diethyl ether (500 µL sterile water + 500 µL diethyl ether), and vortexed in a fume hood for 10 minutes. Cultures were spun down again (2 min, 15,000 RPM), the ether was removed, and the cells were rinsed twice with 1 mL sterile water. The resulting pellet was resuspended in ~100 µL zymolyase and incubated at 37°C for exactly 12 minutes, using the same volume of enzyme for all samples, to dissolve the ascus. Following zymolyase treatment, cells were spun down (3 min at 6000 RPM) and resuspended in 1 mL sterile water. Finally, spores were transferred to experimental flasks and plates containing FuDR-supplemented media, allowing for selective elimination of MATα haploids via the HSV-TK negative selection marker (Fig. 1). This ensured that only MATa haploids remained and prevented the formation of diploids in subsequent experiments.

Dosage testing to determine survival thresholds

To determine appropriate survival thresholds for QTL mapping, sporulated cells were plated on SC-Ura media supplemented with 25 μL of FuDR to eliminate MATα haploids. 100 μL of sporulated cells were added to each plate. Each plate received 200 μL of uracil-containing amino acid mix at different time points to allow surviving cells to resume growth. Plates were prepared for 10 different time points, Day 0 through Day 9, with Day 0 serving as the control condition. Two replicate plates were prepared for each time point, resulting in 20 plates. This dosage testing assay allowed us to monitor survival over time and identify Day 3 and Day 7 as corresponding to approximately 10 percent and 1 percent survival, respectively, which were then used for selection in downstream QTL analysis. Plates were incubated at room temperature (25 °C) for three days before colonies were counted and data was visualized using a line graph in R.

Survival assay

Following sporulation, cells were subjected to uracil deprivation selection, so that only cells surviving until uracil was added at a later time point would be harvested for sequencing. Survival was measured at two time points, day 3 and day 7, which were determined in the dosage tests to correspond to approximately 10% and 1% survival. The experiment was carried out in liquid culture flasks. On day 0, a media stock was prepared in a total volume of 250 mL, consisting of 50% synthetic defined base (SD), 2% dextrose, 10% amino acid dropout mix lacking uracil, in sterile water. Since the amino acid dropout mix did not include adenine, histidine, or arginine, a separate stock containing those components was added to 5% in the total mixture. Each flask was also supplemented with 25 μL of 1000x FuDR per 25 mL. Three flasks were designated as controls and received 10 μL of sporulated culture, while the remaining flasks received 100 μL. 1.25 mL of uracil (from a 400 mg/L stock) was then added to each control flask, resulting in a final uracil concentration of 178 µM. Flasks were then placed in a shaker and incubated at 30°C. Uracil was added to 3 additional flasks on day 3 and to an additional 3 flasks on day 7 (Fig. 1). After uracil was added, the flasks were incubated in a 30 °C shaker before harvesting. After reaching saturation in the 3-day incubation period, cells were spun down and collected for DNA extraction. To confirm that survival at day 3 and day 7 remained at approximately 10% and 1%, respectively, colony counts were performed as follows: after adding uracil, the flasks were mixed well by swirling, and 100 μL from each flask was plated onto SC-Ura plates with 25 μL of FuDR. Plates were incubated at room temperature (25 °C) for three days before colonies were counted to ensure the survival matched the dosage tests.

Library preparation and sequencing

Genomic DNA was extracted using the Qiagen DNeasy Blood and Tissue Kit (69506). Library preparation followed the protocol described by Baym et al., using the Illumina Nextera kit (FC-131-2001) and KAPA HiFi DNA polymerase (KK2612). DNA fragments ranging from 200–600 bp were size-selected using AMPure XP beads, and concentrations were measured using a Qubit 3.0 fluorometer and Roche 480 LightCycler qPCR¹⁰. Prepared libraries were pooled and sequenced on an Illumina NovaSeq 6000 at an average ∼380X per site per bulk.

Quantitative trait loci (QTL) analysis of data

QTL locus identification was performed as described in Buzby et al. 2025¹⁰. Briefly, after DNA sequencing, a quantitative trait locus (QTL) analysis was conducted on the data. First, the sequences were aligned to the genome of the lab strain S288C to provide a baseline for identifying genetic differences and mapping SNPs accurately (Fig. 2). Sequence alignment and processing were conducted using NYU High Performance Computing resources and the following software: GATK, Picard, BWA, SAMtools, and BAMtools^{12,13,14,15,16}. After the sequences were aligned and variants identified, the remainder of the analysis was conducted using R¹⁷. Variants from the reference genome were called to produce a VCF table representing alleles belonging to either the oak or wine parent. The variants were then analyzed separately based on selected bulk populations: unselected (control) vs. day 3 and unselected (control) vs. day 7, to track bulk differences and changes over time. Since the oak and wine strains differ at 1 in 200 bases on average, there are roughly 30,000 SNPs between them. For each bulk, the number of reads of each oak or wine allele per position was smoothed in a 300-SNP window using a sliding window to reduce noise from sequencing. To quantify differences in allele frequencies between selected and control populations, I used logistic regression at each smoothed genomic position using the formula: log(allele odds) ~ Selection (Fig. 2). A false discovery rate (FDR) of 5% was determined by permuting the unselected (control) bulk to generate a distribution of z-scores of allele frequency differences occurring by chance. The significance of the z-score at each position was assessed by comparing real data against this cutoff. Lastly, the cybrBSA¹⁸ peak finder was used to identify regions where the linear regression slope across positions shifted from positive to negative, pinpointing candidate QTL peaks.

Results

Growth of yeast strains under nutrient (uracil) deprivation to identify genetic variants that improve adaptation and longevity

To determine appropriate selection timepoints for QTL mapping, it was important to define when survival began to decline sharply in the absence of uracil. This part of the experiment establishes a survival window that allows isolation of long-lived versus short-lived cells for downstream genomic analysis.

I tracked yeast survival over 9 days on media lacking uracil (SC-Ura) using crosses of strains MLS.CB.007 (oak) and NCYC 3591 (wine), both Ura3 knockouts that lack the ability to synthesize their own uracil. Yeast requires uracil for proper metabolic function, and I measured how survival was affected without supplementation of uracil to determine the 1% and 10% survival rates, which optimizes QTL identification¹⁹. Yeast survival was at its highest on day 0, followed by a gradual decline by day 1 and day 2 (Fig. 3). By day 3, survival had dropped to 10%, marking the upper limit of the target range. After this point, survival continued to decrease sharply, with day 7 reaching 1%, defining the lower survival limit. The population remained consistently low at later time points, with minimal variation in survival (Fig. 3). I expected to see a gradual decline in survival over time with a sharp drop between day 3 and day 7, which agrees with our observed data. The steep decline after day 3 indicates that the bulk of the population could not survive in the absence of uracil, while a small subset of the population persisted. These findings define a clear selection window for BSA, during which very resistant and moderately resistant populations (1% and 10%, respectively) can be compared to identify genetic variants that confer survival in the absence of uracil.

Mapping survival loci reveals SIR4 as a key target under strong selection

After establishing day 3 and day 7 as key survival timepoints, the next step was to map which genes become important as survival becomes limiting. With the thresholds for the BSA established, I then proceeded to the survival assay to measure how yeast populations declined over time without uracil. By comparing populations that survived moderate selection (day 3, ~10% survival) and strong selection (day 7, ~1% survival) to unselected controls, I was able to map specific loci that contribute to chronological lifespan. This comparison allows me to see which genes are consistently linked to survival, and whether stronger selection reveals additional, stronger regulators of aging. Looking at how selection strength changes the associated genes helps show which genes matter most for survival under stress. On Day 3, 10 peaks passed the 5% FDR threshold of 3.85 (Fig. 4A). The graph shows a narrower distribution of peaks that are generally less significant and closer to the cutoff compared to Day 7 (Fig. 4B), suggesting fewer loci with strong associations. In contrast, Day 7 displayed more widespread and more significant peaks, with 13 loci surpassing the 5% FDR threshold of 5.51, indicating that stronger selection reveals additional variants linked to survival. This outcome is consistent with previous BSA studies, which have shown that applying stronger selection often leads to more distinct QTL peaks due to the increased contrast between selected and unselected populations²⁰. Based on this, I expected Day 7 to yield both more peaks and stronger signals than Day 3, which agrees with our observed data. These results suggest that surviving longer without uracil likely depends on more specific or stress-responsive genes that aren’t as important early on. The additional peaks I see at Day 7 could represent factors that only start to have an effect after longer uracil deprivation, showing how stronger selection helps uncover a wider range of survival-related loci.

Focusing on the Day 7 peaks on Chromosome IV (Fig. 4C), we see a sharp peak that overlaps with the SIR4 locus. SIR2 does not show a corresponding signal, despite our expectation to see a peak at SIR2 based on previous data highlighting its role in aging². I also expected to see peaks for SIR3 and SIR4, since their proteins function together with Sir2 protein to suppress the accumulation of extrachromosomal rDNA circles (ERCs), one of the major drivers of aging in yeast². However, our observed data only shows a clear peak for SIR4, with no significant signal at SIR2 or SIR3. This suggests that natural genetic variation affecting survival under uracil deprivation may act more strongly through SIR4 or its regulatory network, rather than through direct changes at SIR2, despite its well-established role in aging².

Functional categorization reveals distinct gene classes under moderate and strong selection

To understand the types of genes associated with longevity at each selection point, I categorized the QTL peaks from Day 3 and Day 7 based on the predicted biological function of the closest gene. On Day 3, most of the peaks were linked to protein degradation and epigenetic control, which are involved in general maintenance pathways supporting genome stability and cellular homeostasis. I expected these categories to be most prominent under moderate selection, as cells likely rely on baseline quality control systems to persist through early nutrient stress. This outcome is consistent with previous studies showing that early-stage survival often depends on general stress buffering mechanisms²¹. By comparison, Day 7 revealed a more diverse set of functional categories, including mitochondrial function, stress response, and translational regulation. In addition, a strong peak overlapped with SIR4, a SIR family gene involved in chromatin silencing and lifespan regulation. These categories suggest that prolonged survival in the absence of uracil requires more specialized mechanisms that become essential under stronger stress conditions²⁰. The increase in peak-adjacent genes’ diversity supports the idea that stronger selection highlights additional regulators of survival that are not evident under milder stress²⁰.

Discussion

Understanding how gene variation of epigenetic regulation influences cellular lifespan and disease progression is important for understanding and possible treatments of aging. In this study, I explored the role of SIR2 in genomic stability and aging using S. cerevisiae as a model organism. SIR2 functions primarily through chromatin silencing as a histone deacetylase, reducing genomic instability, which is a key driver of cellular aging⁶. In humans, SIRT1 closely resembles the yeast SIR2, performing similar functions in DNA repair regulation, histone deacetylation, and the management of stress responses and metabolic processes associated with aging²². The activity of SIRT1 has direct implications for delaying age-related decline in humans, making it a promising target for therapeutic intervention in neurodegenerative diseases, cardiovascular conditions, and degenerative joint disorders like osteoarthritis. This study’s open-ended approach ultimately identified SIR family genes, offering insight into how natural genetic variation in chromatin regulators may inform future strategies for aging and disease prevention. Although SIR2 was not identified as a major contributor in my study, other members of the SIR family were, along with additional genes outside the family that likely interact with the SIR2 pathway.

In the first part of my study, I conducted preliminary dose experiments in two genetically distinct yeast strains, MLS.CB.007 (oak) and NCYC 3591 (wine). By exposing progeny to uracil-deprived media, I determined that Day 3 (10%) and Day 7 (1%) survival rates provided the ideal selection thresholds for maintaining genetic diversity while still applying selective pressure. These results identified Day 3 and Day 7 as optimal timepoints for QTL identification in the next phase of the study. In the second part of my study, I used a bulk segregant analysis (BSA) to identify genetic loci linked to survival under nutrient deprivation by sequencing progeny that survived the 1% and 10% selection conditions. A quantitative trait locus (QTL) analysis then revealed genomic regions associated with increased survival. Peaks were found on chromosomes IV, VII, and XV, with a strong signal overlapping SIR4. Although I expected to find a peak at SIR2 based on its established role in aging, no significant signal appeared at that locus. However, the identification of SIR4 and other stress-related loci suggests that survival may be driven more by regulators acting alongside SIR2 rather than by direct variation in SIR2 itself.

Absence of SIR2 signal reveals limits of natural variation

Although I expected to find a peak at SIR2 based on its well-established role in aging⁶, no significant signal appeared at that locus. This raises two interesting possibilities. First, the genetic variants influencing aging in natural populations may differ from those identified through gene knockouts in laboratory settings. While deleting SIR2 shortens lifespan², natural variants may be too subtle to produce a detectable QTL signal, or they may not be present at all between these two strains. Second, there may not be any biologically relevant variation in SIR2 between the oak and wine strains, meaning the locus is not polymorphic in a way that influences survival under these conditions. Although the two strains differ on average at 1 in every 200 base pairs, this level of variation may still not affect SIR2 in a way that impacts aging. The absence of a peak at SIR2 suggests that, although it plays a central role in controlled experimental systems, other components of the same pathway may exhibit more functional variability across natural populations. In this study, SIR4, where a strong signal was detected, may represent one such component.

Model-dependent differences in aging-associated QTLs

In this study, I identified several QTLs associated with variation in chronological lifespan (CLS) in yeast. Notably, these included SIR4 from the SIR family, along with other genes involved in mitochondrial function, stress response, and protein degradation. However, a previous study⁹ identified different QTLs—most notably SIR2—likely due to their use of replicative lifespan (RLS) as a model. Since RLS measures the number of daughter cells produced by a mother cell, while CLS reflects the survival of non-dividing cells in the stationary phase, the underlying genetic drivers of aging can differ between the two.

While both CLS and RLS are used to study aging in yeast, they capture different aspects of cellular aging. Replicative lifespan centers on the aging of actively dividing mother cells and is especially sensitive to genes involved in chromatin silencing and genomic stability⁸. SIR2 regulates chromatin structure by maintaining silencing at rDNA loci, telomeres, and mating-type regions, which are functions particularly important during replication and mitotic division². Chronological lifespan, on the other hand, tracks how long non-dividing cells survive in a quiescent state⁸. This mode of aging relies more heavily on a cell’s ability to manage nutrient stress, oxidative damage, and metabolic shifts during the stationary phase. As a result, genes like ENA5 and SER1, which regulate stress responses and central carbon metabolism, may play a larger role in CLS than in RLS. This distinction likely explains why SIR2 did not appear as a peak in my analysis, since its role is more relevant to dividing cells. In contrast, SIR4, which also participates in chromatin regulation², may influence lifespan through mechanisms that remain active or have lingering effects even in non-dividing, quiescent cells, making it more relevant to CLS.

SER1 emerges as a QTL peak across conditions but through distinct mechanisms

A previous study using chronological lifespan as a model also identified SER1 as a gene associated with lifespan variation through QTL mapping²¹. SER1 was also identified as a QTL in my analysis, suggesting a potentially conserved role for this gene in regulating survival during aging. However, there are important experimental differences that may explain why other genes identified in the previous study, such as RIM15, a protein kinase involved in cell proliferation, and MIL1, a putative lipase located on chromosome VI²², did not appear in my study. This previous experiment allowed yeast to grow until they naturally entered the stationary phase as nutrients became depleted, triggering quiescence-associated stress responses²². In contrast, my experiment limited growth earlier by removing a specific nucleotide from the media, which prevented the cells from undergoing full metabolic adaptation to the stationary phase. Due to these differences, SER1 may be influencing lifespan through distinct mechanisms in each context, potentially through nutrient stress signaling in their setup and nucleotide starvation in mine²⁴.

Additional pathways supporting lifespan under uracil deprivation

In addition to SER1, my study identified several other genes with potential roles in longevity regulation under uracil starvation. COX20, a mitochondrial membrane protein involved in cytochrome c oxidase assembly, appeared as a significant peak on day 7. Given the established link between mitochondrial function and aging²⁵, COX20 may influence survival through respiration-dependent pathways. Another gene, RRM3, is a helicase known to resolve replication stress at telomeric and rDNA regions²⁶. Its identification aligns with prior studies showing the importance of genome maintenance in extending lifespan⁷. I also identified YPK9, a vacuolar protein with oxidative stress response functions and homology to human ATP13A2, which is associated with Parkinson's disease²⁷. These results suggest that, in addition to the SIR family, multiple pathways—mitochondrial regulation, DNA replication fidelity, and stress response—converge in support of long-term survival under nutrient limitation.

Future directions: uncovering gene interactions through chromosome IV replacement

To uncover how genetic variation across the genome interacts with Chromosome IV (where I saw the strongest signal overlapping SIR4, as well as the region where SIR2 is located), I plan to detect epistatic interactions by testing how the rest of the genome responds when the Chromosome IV background is held constant in an oak or wine background. By fixing Chromosome IV, I can use bulk segregant analysis (BSA) to identify genes whose effects depend on the specific allele present at this locus¹⁰. This approach could help reveal new regulators of aging that are not detectable through single-gene analysis alone but instead act in combination with specific alleles at key loci like SIR4 and SIR2. Identifying these interactions will help us better understand how natural variation shapes longevity across genetic backgrounds and may provide insight into the types of gene networks that influence aging in humans²⁸.

In addition to mapping epistatic interactions, the loci I identified may also provide insight into the broader regulatory network surrounding the SIR genes. Several of the candidate loci identified in my analysis are not directly involved in chromatin silencing but may influence lifespan by modulating the expression or activity of SIR family members. Exploring how these loci connect to known aging regulators could help reconstruct the signaling and transcriptional networks that shape aging phenotypes across genetic backgrounds. This type of network-level analysis could enhance the allele-specific effects uncovered through epistasis and point toward more integrated models of lifespan regulation.

Conclusion

Understanding how natural genetic variation shapes longevity is essential for uncovering the broader regulatory networks that influence aging. In this study, I used S. cerevisiae as a model to investigate the genetic basis of chronological lifespan (CLS) by mapping survival-associated loci through QTL analysis. My approach involved crossing two genetically distinct yeast strains and selecting for long-lived progeny under conditions of nucleotide deprivation. This allowed me to isolate genomic regions linked to enhanced survival in the absence of uracil.

I identified several QTL peaks associated with longevity, including a strong signal overlapping SIR4, a member of the SIR family known for its role in chromatin silencing. Although SIR2 has been widely studied as a key aging regulator, it did not appear as a peak in my analysis, suggesting that survival under nutrient stress may depend more heavily on other epigenetic regulators like SIR4, especially in non-dividing, stationary-phase cells. Functional categorization of the QTL peaks revealed a shift in genetic drivers depending on the level of selection pressure, with early survival relying on protein degradation and chromatin maintenance, and later survival involving more diverse pathways, including mitochondrial function and stress response.

Notably, SER1 emerged as a significant QTL peak in my study, consistent with findings from a previous CLS study that identified it under nutrient depletion conditions. Despite the shared peak, the experimental designs differed: the prior study allowed yeast to progress naturally into the stationary phase, while my selection prevented full quiescence by removing a key nucleotide. These differences suggest that SER1 may act through distinct mechanisms depending on the metabolic and environmental context, highlighting its versatility as a modulator of lifespan. Overall, my findings showcase the importance of looking beyond canonical regulators like SIR2 and considering how indirect genetic factors shape lifespan through stress response, metabolism, and chromatin dynamics. This work not only adds to the growing understanding of how aging is regulated at the systems level but also offers new points for identifying conserved longevity pathways relevant to humans.

Acknowledgements

I would like to express my gratitude to Dr. Duncan Smith for being on my thesis committee. I am especially thankful to Dr. Mark Siegal for mentoring and supporting me throughout the project and over the past year during my undergraduate study. I also wish to acknowledge the contribution of Carina Terry for their contribution to my survival assay and DNA sequencing. I also acknowledge Dr. Cassandra Buzby's contribution to the DNA sequencing analysis. To Dr. Eugene Plavskin, thank you for answering all my questions and giving me advice on figures and my code. I am also grateful to the NYU Dean’s Undergraduate Research Funds Committee for awarding me the grant that I used to fund my DNA sequencing. I also thank the members of the Siegal lab for their input on the paper.

References

1. Ha, C. W., & Huh, W.-K. (2011). The implication of Sir2 in replicative aging and senescence in Saccharomyces cerevisiae. Aging (Albany, NY), 3(3), 319–324.
2. Kaeberlein, M., McVey, M., & Guarente, L. (1999). The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes & Development, 13(19), 2570–2580. https://doi.org/10.1101/gad.13.19.2570
3. Kaeberlein, M., Kirkland, K. T., Fields, S., & Kennedy, B. K. (2004). Sir2-Independent Life Span Extension by Calorie Restriction in Yeast. PLOS Biology, 2(9), e296. https://doi.org/10.1371/journal.pbio.0020296
4. Giblin, W., & Lombard, D. B. (2016). Chapter 3—Sirtuins, Healthspan, and Longevity in Mammals. In M. R. Kaeberlein & G. M. Martin (Eds.), Handbook of the Biology of Aging (Eighth Edition) (pp. 83–132). Academic Press. https://doi.org/10.1016/B978-0-12-411596-5.00003-4
5. Hickman, M. A., Froyd, C. A., & Rusche, L. N. (2011). Reinventing Heterochromatin in Budding Yeasts: Sir2 and the Origin Recognition Complex Take Center Stage. Eukaryotic Cell, 10(9), 1183–1192. https://doi.org/10.1128/EC.05123-11
6. Imai, S., Armstrong, C. M., Kaeberlein, M., & Guarente, L. (2000). Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature, 403(6771), 795–800. https://doi.org/10.1038/35001622
7. Sinclair, D. A., & Guarente, L. (1997). Extrachromosomal rDNA Circles—A Cause of Aging in Yeast. Cell, 91(7), 1033–1042. https://doi.org/10.1016/S0092-8674(00)80493-6
8. Longo, V. D., Shadel, G. S., Kaeberlein, M., & Kennedy, B. (2012). Replicative and Chronological Aging in Saccharomyces cerevisiae. Cell Metabolism, 16(1), 18–31. https://doi.org/10.1016/j.cmet.2012.06.002
9. Stumpferl, S. W., Brand, S. E., Jiang, J. C., Korona, B., Tiwari, A., Dai, J., Seo, J.-G., & Jazwinski, S. M. (2012). Natural genetic variation in yeast longevity. Genome Research, 22(10), 1963–1973. https://doi.org/10.1101/gr.136549.111
10. Buzby, C., Plavskin, Y., Sartori, F. M. O., Tong, Q., Vail, J. K., & Siegal, M. L. (2025). Epistasis and cryptic QTL identified using modified bulk segregant analysis of copper resistance in budding yeast. Genetics, 229(4), iyaf026. https://doi.org/10.1093/genetics/iyaf026
11. URA3 orotidine-5’-phosphate decarboxylase [Saccharomyces cerevisiae S288C]—Gene—NCBI. (n.d.). Retrieved April 23, 2025, from https://www.ncbi.nlm.nih.gov/gene/856692
12. Van der Auwera GA & O'Connor BD. (2020). Genomics in the Cloud: Using Docker, GATK, and WDL in Terra (1st Edition). O'Reilly Media.
13. Li, H., & Durbin, R. (2009). Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics, 25(14), 1754–1760. https://doi.org/10.1093/bioinformatics/btp324
14. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R., & 1000 Genome Project Data Processing Subgroup. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16), 2078–2079. https://doi.org/10.1093/bioinformatics/btp352
15. Barnett, D. W., Garrison, E. K., Quinlan, A. R., Strömberg, M. P., & Marth, G. T. (2011). BamTools: A C++ API and toolkit for analyzing and managing BAM files. Bioinformatics, 27(12), 1691–1692. https://doi.org/10.1093/bioinformatics/btr174
16. Broad Institute, GitHub Repository. Picard Toolkit. Broad Institute; 2019. https://broadinstitute.github.io/picard/
17. R Core Team (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.
18. Buzby Cassandra, GitHub. cybrBSA. 2024 https://github.com/cbuzby/cybrBSA
19. Magwene, P. M., Willis, J. H., & Kelly, J. K. (2011). The Statistics of Bulk Segregant Analysis Using Next Generation Sequencing. PLOS Computational Biology, 7(11), e1002255. https://doi.org/10.1371/journal.pcbi.1002255
20. Ehrenreich, I. M., Torabi, N., Jia, Y., Kent, J., Martis, S., Shapiro, J. A., Gresham, D., Caudy, A. A., & Kruglyak, L. (2010). Dissection of genetically complex traits with extremely large pools of yeast segregants. Nature, 464(7291), 1039–1042. https://doi.org/10.1038/nature08923
21. Gasch, A. P., Spellman, P. T., Kao, C. M., Carmel-Harel, O., Eisen, M. B., Storz, G., Botstein, D., & Brown, P. O. (2000). Genomic Expression Programs in the Response of Yeast Cells to Environmental Changes. Molecular Biology of the Cell, 11(12), 4241–4257. https://doi.org/10.1091/mbc.11.12.4241
22. Haigis, M. C., & Sinclair, D. A. (2010). Mammalian Sirtuins: Biological Insights and Disease Relevance. Annual Review of Pathology: Mechanisms of Disease, 5(Volume 5, 2010), 253–295. https://doi.org/10.1146/annurev.pathol.4.110807.092250
23. Jung, P. P., Zhang, Z., Paczia, N., Jaeger, C., Ignac, T., May, P., & Linster, C. L. (2018). Natural variation of chronological aging in the Saccharomyces cerevisiae species reveals diet-dependent mechanisms of life span control. Npj Aging and Mechanisms of Disease, 4(1), 1–11. https://doi.org/10.1038/s41514-018-0022-6
24. Petti, A. A., Crutchfield, C. A., Rabinowitz, J. D., & Botstein, D. (2011). Survival of starving yeast is correlated with oxidative stress response and nonrespiratory mitochondrial function. Proceedings of the National Academy of Sciences, 108(45), E1089–E1098. https://doi.org/10.1073/pnas.1101494108
25. Sun, N., Youle, R. J., & Finkel, T. (2016). The Mitochondrial Basis of Aging. Molecular Cell, 61(5), 654–666. https://doi.org/10.1016/j.molcel.2016.01.028
26. Ivessa, A. S., Lenzmeier, B. A., Bessler, J. B., Goudsouzian, L. K., Schnakenberg, S. L., & Zakian, V. A. (2003). The Saccharomyces cerevisiae Helicase Rrm3p Facilitates Replication Past Nonhistone Protein-DNA Complexes. Molecular Cell, 12(6), 1525–1536. https://doi.org/10.1016/S1097-2765(03)00456-8
27. Joseph, F., Miller, D., Evgrafov, O. V., & Chirico, W. J. (2021). YPK9 and WHI2 Negatively Interact during Oxidative Stress. Microorganisms, 9(12), Article 12. https://doi.org/10.3390/microorganisms9122584
28. Podolskiy, D., Molodtcov, I., Zenin, A., Kogan, V., Menshikov, L. I., Gladyshev, V. N., Reis, R. J. S., & Fedichev, P. O. (2016). Critical dynamics of gene networks is a mechanism behind aging and Gompertz law (No. arXiv:1502.04307). arXiv. https://doi.org/10.48550/arXiv.1502.04307