Survival of starving yeast is correlated with oxidativestress response and nonrespiratory mitochondrialfunctionAllegra A. Pettia,b, Christopher A. Crutchfielda,c, Joshua D. Rabinowitza,c, and David Botsteina,b,1

aLewis-Sigler Institute for Integrative Genomics and Departments of bMolecular Biology and cChemistry, Princeton University, Princeton, NJ 08544

Edited by Jasper Rine, University of California, Berkeley, CA, and approved June 8, 2011 (received for review January 31, 2011)

Survival of yeast during starvation has been shown to depend onthe nature of the missing nutrient(s). In general, starvation for“natural” nutrients such as sources of carbon, phosphate, nitro-gen, or sulfate results in low death rates, whereas starvation foramino acids or other metabolites in auxotrophic mutants results inrapid loss of viability. Here we characterized phenotype, gene ex-pression, and metabolite abundance during starvation for methi-onine. Some methionine auxotrophs (those with blocks in thebiosynthetic pathway) respond to methionine starvation like yeaststarving for natural nutrients such as phosphate or sulfate: theyundergo a uniform cell cycle arrest, conserve glucose, and survive.In contrast, methionine auxotrophs with defects in the transcrip-tion factors Met31p and Met32p respond poorly, like other auxo-trophs. We combined physiological and gene expression data froma variety of nutrient starvations (in both respiratory competentand incompetent cells) to show that successful starvation responseis correlated with expression of genes encoding oxidative stressresponse and nonrespiratory mitochondrial functions, but not res-piration per se.

mitochondrion | longevity | Warburg effect

Understanding global coordination of subcellular processesduring adaptation to environmental change is a central chal-

lenge in systems biology. The ability of free-living organisms toadapt to changes in their nutritional environment is clearly oneof the driving forces of their evolution. In natural environments,yeast are exposed to extreme variations in “natural nutrient”availability, particularly in their sources of carbon (and energy),phosphorus, sulfur, and nitrogen. Unlike wild type strains,auxotrophic mutant yeast strains unable to make an essentialmetabolite (e.g., leucine or uracil) can also be starved for themissing metabolite, but adaptation to this kind of “nonnatural”starvation has not been subject to evolutionary selection. Star-vation of Saccharomyces cerevisiae for a single, growth-limitingnutrient offers the opportunity to study the coordination of nu-trient sensing, metabolism, growth, and cell division. Propercoordination results in prolonged survival, concerted cell cyclearrest, and glucose conservation during starvation, and dependsstrongly on the specific nutrient being depleted. For instance, thesurvival of auxotrophic yeast starved for leucine, histidine, oruracil is substantially impaired (exhibiting a roughly 10-folddifference in half-life) relative to the same strain starved for the“natural” nutrients sulfate or phosphate (1). Starvation for sul-fate or phosphate elicits rapid, nearly uniform G0/G1 cell cyclearrest and slows glucose consumption, whereas starvation forleucine, histidine, or uracil results in incomplete cell cycle arrestand markedly higher rates of glucose consumption.We are interested in understanding what, if any, general

principles determine starvation phenotype. Early work on nu-trient starvation posited the existence of a starvation “signal”that promotes concerted cell cycle arrest and survival in responseto nutritional scarcity (2). However, fundamental questionsabout the identity and effects of the starvation signal have neverbeen addressed. The answers to these questions have implica-

tions for other areas of biology, because connections among thesame processes that are influenced by nutrient starvation havebeen documented in cancer, aging, and the yeast metabolic cycle.As a first step toward understanding general determinants of

starvation phenotype, we sought to identify transcriptional cor-relates of “successful” or “survivable” starvation—that is, starva-tion that leads to concerted cell cycle arrest, glucose conservation,and, most importantly, survival. To this end, we compared thephysiology and gene expression of S. cerevisiae in response tostarvation for a variety of previously characterized nutrients, andfor one additional nutrient, methionine. We chose methioninebecause previous studies hint that it coordinates diverse cellularfunctions, thereby acting as a regulatory hub. For example, me-thionine starvation of a methionine auxotroph causes rapid anduniform G0/G1 cell cycle arrest (2), unlike the result obtainedwith other auxotrophies (3). In addition, genome-wide analysesof periodic gene expression during the cell cycle showed thatmethionine is the only amino acid whose biosynthetic genes ex-hibit periodic expression (4). In the yeast metabolic cycle, themethionine-regulated transcription factors exhibit the most highlyperiodic transcriptional activity, implying a connection betweenthe methionine regulon and the respiro-fermentative balance (5–7). This implied connection is strengthened by a study of glucoserepletion in yeast, which shows that methionine biosyntheticgenes are highly induced upon relief from glucose starvation(8). The methionine-regulated transcription factors are also wellknown to regulate the response to various toxins, including heavymetals and oxidizing agents (9, 10).We began our comparison of multiple nutrient starvations by

characterizing physiology, gene expression, and metabolite abun-dance during methionine starvation. We first created a panel ofisogenic methionine auxotrophs. Two of these deletion mutants(met6Δ and met13Δ) lack genes encoding steps in methioninebiosynthesis, creating a “metabolic” defect. The others lackgenes encoding one or more methionine transcription factors,thereby creating “regulatory” defects that result in a methioninerequirement. Phenotypic characterization of these strains con-firmed that methionine starvation is indeed different from thepreviously studied auxotrophic starvations and results in a pheno-type intermediate between that of the natural nutrient starvationsand the other auxotrophic starvations. Methionine starvation of

Author contributions: A.A.P., J.D.R., and D.B. designed research; A.A.P. and C.A.C. per-formed research; A.A.P. contributed new reagents/analytic tools; A.A.P. analyzed data;and A.A.P. and D.B. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

Data deposition: The microarray expression data reported in this paper have been de-posited at http://genomics-pubs.princeton.edu/StarvationSurvival/.1To whom correspondence should be addressed. E-mail: botstein@princeton.edu.

See Author Summary on page 18217.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101494108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1101494108 PNAS | November 8, 2011 | vol. 108 | no. 45 | E1089–E1098

SYST

EMSBIOLO

GY

PNASPL

US

met6Δ or met13Δ mutants, like sulfate or phosphate starvation ofa wild-type strain, results in uniform G0/G1 arrest, substantiallyincreased survival relative to leucine or uracil starvation (assessedin mutants with analogous metabolic defects), and in conservationof residual glucose.By comparing mRNA levels during methionine starvation in

met6Δ and met13Δ with the previously published results of similarexperiments during starvation for leucine, uracil, phosphate, andsulfate, we found a group of genes whose expression is stronglycorrelated with successful response to starvation. These nuclearlyencoded genes are enriched for assorted mitochondrial functions,response to heat and oxidative stress, and regulation of therespiro-fermentative balance. We then narrowed down the role ofthese functions in survival: during starvation for various nutrients,we measured survival of ρ0 mutants and hydrogen peroxide-treated (ρ+) strains, which showed that oxidative phosphorylationis not required for survival but that survival half-life is correlatedwith ability to detoxify reactive oxygen species (ROS). This resultis fortified by our observation that the short-lived regulatorymethionine auxotroph,met31Δmet32Δ, differs from the long-livedmetabolic methionine auxotrophs primarily in its reduced abilityto combat oxidative stress during methionine starvation.Our results speak against the intuitive and commonly held

view that oxidative metabolism is deleterious per se. Instead, thedata suggest that induction of the transcriptional programs as-sociated with, but not dependent on, respiration exerts a pro-tective effect on the cell during nutritional shortages, possiblybecause the many protective mechanisms against oxidativedamage and related stresses become increasingly engaged as thecell shifts from fermentative to oxidative metabolism.

ResultsStarvation for Methionine Resembles Starvation for Natural Nutrients.The response of S. cerevisiae to starvation for phosphate, sulfate,ammonium, leucine, and uracil has been previously described (1,3, 11). To determine where methionine starvation lies on thespectrum of starvation phenotypes, we measured survival, glucoseconsumption, and cell cycle arrest during methionine starvationof the metabolic methionine auxotrophs met6Δ and met13Δ. Tocalibrate our measurements against those in the literature, wealso measured these parameters during starvation of a previouslystudied leucine auxotroph, leu2Δ. We compared all measure-ments with those made during phosphate starvation.To measure the starvation survival ofmet6Δ,met13Δ, and leu2Δ,

a single colony of each strain was grown in medium limited fora single nutrient, either phosphate or the required amino acid. Eachstrain was then diluted into medium lacking only that nutrient, andviability over time was measured (Materials and Methods). Thesemethionine auxotrophs, when starved for methionine, survivedlonger than leucine or uracil auxotrophs starved for their respectiverequirements but not as long as the same strains starved for phos-phate or sulfate (Fig. 1A and Table S1) (1). Fig. 1A also containsdata for the regulatory methionine auxotroph met31Δmet32Δ, inwhich the methionine requirement is caused by the deletion of twogenes,MET31 andMET32, which encode transcription factors; weaddress its properties below. We classified the starvations as long-lived [phosphate, sulfate, and methionine (metabolic; met6Δ ormet13Δ)] or short-lived [(leucine, uracil, and methionine (regula-tory; met31Δmet32Δ)] according to the data in Fig. 1A.Survival during starvation correlates strongly with uniformity

of cell cycle arrest in stationary phase cultures (1, 3, 11). Wemeasured bud index for three cultures each of met6Δ and met13Δundergoing methionine starvation, and for one culture each ofstrains undergoing leucine (leu2Δ), uracil (ura3-52), and phos-phate starvation (FY4) (Fig. 1B and Table S1). Combining ourresults with those in ref. 3, we found that methionine starvation isan outlier with respect to the other studied auxotrophic starva-tions: methionine starvation induces cell cycle arrest in G0/G1 to

nearly the same extent as phosphate starvation (P = 0.07 by two-sided Student t test) but to a significantly greater extent thanleucine and uracil starvation (P = 9.9 × 10−5). This, as expected,recapitulates the cell cycle arrest during methionine starvationthat was initially observed decades ago (2).In previously studied starvations, survival time and arrest

phenotype correlated with the rate of glucose consumption after

A

B

C

20 40 60 800

120

80

40

Hours

0 1612840.01

0.1

1

10

100

Days

Per

cent

Via

bilit

y

5 15 25 3530

50

70

90

Hours

Per

cent

Arr

este

dP

erce

nt R

esid

ual G

luco

se

Pho (met6 (panel A))

Pho (met13 )

Pho (met31Δmet32 )Pho (leu2 )

Ura (ura3-52)

Met (met6 )Met (met13 )

Leu (leu2 )Met (met31Δmet32 )

Pho (FY4 (panel B))

Fig. 1. Physiological responses to nutrient starvation. (A) Survival over timeduring starvation for methionine (Met), phosphate (Pho), or Leucine (Leu)using the indicated strains. Survival was measured as the ability to forma colony on rich medium plates. (B) Cell cycle arrest, as measured by per-centage of cells with no bud (assessed by light microscopy), during starvationfor Met, Pho, Leu, or Uracil (Ura) using the indicated strains. Representa-tive time courses are shown, with more comprehensive data provided inTable S1. (C) Glucose consumption (measured as described in Materials andMethods) after biomass (measured by optical density) has reached its pla-teau during starvation for Met or Leu. Error bars show SD of technicalreplicates.

E1090 | www.pnas.org/cgi/doi/10.1073/pnas.1101494108 Petti et al.

biomass accumulation stopped increasing (1, 12). We measuredthe concentration of residual, extracellular glucose in batch cul-tures ofmet6Δ,met13Δ, leu2Δ, and the exceptionalmet31Δmet32Δundergoing starvation for the required amino acid. Like proto-trophs starving for a natural nutrient, metabolic methionine aux-otrophs starving for methionine stopped consuming glucose oncebiomass accumulation reached a plateau. In contrast, the leucineauxotroph continued consuming glucose, as reported previously(Fig. 1C) (1).A diverse set of survival assays has been used to show that re-

pression of glucose-activated signaling pathways, such as PKAand Tor, prolongs survival (1, 13, 14). To test the influence ofglucose-activated signaling on survival during methionine star-vation, we repeated themet6Δ andmet13Δ survival measurementsusing a combination of glycerol and ethanol as the carbon source,which increased survival during methionine starvation almosttwofold (Table S1).

Methionine Starvation Activates a Characteristic Gene ExpressionProgram. We followed gene expression during methionine star-vation of the metabolic methionine auxotrophs met6Δ andmet13Δ using a filter-based cell growth protocol that permitsinstantaneous removal of external methionine (Materials andMethods). This facilitates comparison of multiple strains becauseit allows the time courses to be aligned to the instant of methi-onine removal. As described in SI Materials and Methods, linearregression was used to select genes exhibiting (i) a statisticallysignificant dependence on time and (ii) at least a twofold changein one or both time courses. In this regression analysis, theP value of the F statistic for the regression model represents thesignificance of the time dependence, and Q-VALUE software(15) was used to identify P values corresponding to a false dis-covery rate (FDR) of at most 0.01 (SI Materials and Methods).The resulting gene set shows that a wide variety of biologicalprocesses are perturbed during methionine starvation. We fo-cused on identifying processes that are perturbed specificallyduring methionine starvation and might therefore provide insightinto the unique role of methionine in the cell cycle, the metaboliccycle, and carbohydrate metabolism. To isolate such processes,we compared our methionine starvation gene expression datawith previously published gene expression data collected duringphosphate, sulfate, leucine, and uracil starvations, the “Mega-Cluster” in ref. 3.We processed, combined, and filtered these data sets as de-

scribed in SI Materials and Methods. To control for potentialdifferences between our instantaneous starvation protocol andthe gradual starvation protocols used for leucine, phosphate,sulfate, and uracil, we also collected expression data for met13Δaccording to the protocol used for these other nutrients. Acomparison of met13Δ expression data using the two alternateprotocols showed essentially identical patterns of gene expression(Fig. S1 and Dataset S1). To identify genes whose differentialexpression is specific to methionine starvation, we clustered thecombined expression data set into 40 distinct clusters usingk-means clustering (SI Materials and Methods, Fig. S2A, andDataset S1). We then focused on gene clusters displaying expres-sion profiles specific to methionine starvation or to both methio-nine and sulfate starvation.The k-means clustering yielded a cluster of genes (cluster 25)

that are induced only in methionine or sulfate starvation (Fig. 2,Table S2, and Dataset S1). As expected, a significant number ofthese genes function in sulfur metabolism and methionine bio-synthesis, including the entire methionine biosynthetic pathway.Remarkably, this cluster also contains a number of the NAD andNADP-binding enzymes on which methionine biosynthesis in-directly depends. Several other biological processes that dependheavily onmethionine abundance are also enriched here, includingthe overlapping categories “oxidation reduction,” “oxidative stress

response,” and “response to drug” (e.g., YAP1, GSH1, OPT1,GRX8,MXR1, andZWF1) and “phosphatidylcholine biosynthesis”(OPI3, CKI1, SER33).The substantial overlap between the transcriptional responses to

sulfate and methionine starvation suggests that methionine mightserve as an “indicator metabolite” for sulfate starvation. Despite

Cluster 31: Genes specifi-

cally induced by methio-

nine starvation

Process: No GO enrich-ment; MIPS enrichment for DNA topology; stress response; aerobic respirationComponent: Mitochodrial inner membrane*, integral to membrane*, respiratory chain*

Cluster 12: Genes specifi-

cally repressed by

methionine starvation

Process: Cell cycle; nucleotide metabolism; cell aging; chromatid segrega-tion; nuclear division; DNA replication; cell wall organiza-tionComponent: Chromosomal part, mitotic cohesin complex, intrinsic to plasma membrane

Cluster 25: Genes

regulated by methionine

and sulfur starvation

Process: Sulfur amino acid biosynthesis; methionine metabolism; serine family metabolism; aspartate family metabolism; sulfate assimilation; response to drug; oxidation reduction; siroheme metabolism; responses to metal ion and hydrogen peroxide; phosphatidylcholine biosynthesisComponent: Sulfite reductase complex (NADPH)

M(6)

L P S UM(13)

M(TF)

-3 30

Fig. 2. Expression profiles specific to methionine and sulfate starvation.K-means clusters characterized by gene expression specific to methioninestarvation (cluster 31 and cluster 12) or methionine and sulfate starvation(cluster 25). Each heatmap shows microarray expression data for sevenstarvation time courses, separated by gray columns. The starvations are,from left to right, methionine using met6Δ [M(6)], methionine using met13Δ[M(13)], methionine using the Transcription factor mutant met31Δmet32Δ[M(T)], leucine using leu2Δ (L), phosphate using FY4 (P), sulfate using FY4 (S),and uracil using ura3-52 (U). The grayscale triangles above each time courseshow how the time courses were grouped for analysis and interpretation.Here, for example, methionine starvation in metabolic methionine auxo-trophs (black triangles) was compared with phosphate, sulfur, leucine, anduracil starvation (gray triangles). met31Δmet32Δ expression data (white tri-angle) is displayed but was not used in the analysis. The functional enrich-ment (FDR ≤ 0.1, or, where indicated with an asterisk, between 0.1 and 0.2)of each gene set is summarized next to the corresponding heat map.Detailed enrichment data are provided in Table S2.

Petti et al. PNAS | November 8, 2011 | vol. 108 | no. 45 | E1091

SYST

EMSBIOLO

GY

PNASPL

US

this overlap, some genes are purely methionine starvation-specific.Of these, the repressed subset (cluster 12) is enriched for cell cycleregulation (e.g., TOR2, NDD1, PCL1, YOX1, RAD9, RAD53) andrelated processes, such as chromatid segregation,DNA replication,and nucleotide metabolism. It is also enriched for genes implicatedin response to aging (SCH9, HDA2, LAC1, SGS1, and ACS2) andcell wall organization through the SHO1 osmosensing pathway.Genes that are repressed in all starvations but methionine

(cluster 27) are enriched for assorted mitochondrial functions,including the folate cycle (GCV1, GCV2, GCV3) and translation(e.g., AIM10, MST1, MSK1, MSD1, QCR6, MRPS17, PPA2,MRPL22, ALG6). This reflects the fact that the methioninebiosynthetic pathway interacts directly with the folate pathway atits penultimate step, in which the methylenetetrahydrofolate re-ductase encoded by MET13 catalyzes the reduction of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate.Only a few terms are significantly overrepresented among the

genes that are most strongly and specifically induced duringmethionine depletion (cluster 31) [e.g., the Munich InformationCenter for Protein Sequences (MIPS) categories related to DNAtopology and aerobic respiration]. At a more lenient FDR, genesinvolved in ATP biosynthesis (QCR2, QCR8, and COX6) andmetabolism of membrane phospholipids (e.g., IPT1, PBN1,PSD1, TGL2, and TLG2) are overrepresented.

Transcriptional Regulation of Metabolic Pathways and IsozymesDuring Methionine Starvation. To understand metabolic changesduring methionine starvation from a transcriptional perspective,we mapped the most differentially regulated enzyme-encodinggenes (≥32-fold change in met6Δ and/or met13Δ) onto the met-abolic network (Fig. S3). Notable are the high induction of an-tioxidant biosynthetic genes (CTT1, GSH1, GTT1, GTT2, andGPX1); differential regulation of most pentose phosphatepathway genes; induction of the NAD metabolic genes BNA3and PNC1; and repression of nucleotide biosynthesis and thefolate cycle.The set of differentially expressed genes contains several iso-

zyme pairs that are particularly informative because they exhibitdifferent specificities for anaerobic and aerobic conditions. Inevery pair, the isozyme specific to respiration is induced, whereasthat specific to fermentation is repressed (cf. GLK1 and HXK1vs. HXK2, PDC5 vs. PDC6, ALD4 vs. ALD6, GDH1 vs. GDH3,GND1 vs. GND2, and TKL1 vs. TKL2). This suggests that me-thionine starvation may cause the cell to rely increasingly onrespiration for energy generation.

Direct Assessment of Metabolism During Methionine Starvation. Wemeasured intracellular metabolites for one methionine auxo-troph, met13Δ, during starvation (Materials and Methods). Nor-malization and processing of these data are described in SIMaterials and Methods. Hierarchical clustering was used to iden-tify clusters of metabolites that exhibit coordinated changes inabundance (Fig. 3A and Dataset S1). As expected, the metabo-lites whose abundance decreases earliest and most dramatically(Fig. 3A, turquoise bar) are predominantly associated with me-thionine biosynthesis or salvage (e.g., methionine, 4-methyl-thio-2-oxobutanoic acid, S-adenosyl methionine, homocysteine,5′-methylthioadensoine, and cysteine), reflecting the cell’s needto replenish methionine. However, the set of rapidly depletedmetabolites also includes fructose-1,6-bisphosphate and NADH,reflecting the broader metabolic reconfiguration that occurs dur-ing methionine starvation. Fructose-1,6-bisphosphate (F1,6BP) isproduced from fructose-6-phosphate (F6P) by phosphofructoki-nase (PFK), the main regulator of glycolytic flux (16). Twoobservations suggest that PFK activity is reduced. First, F6Pabundance increases concomitantly with the decrease in F1,6BP,implying inhibition of PFK. Second, citrate, a major inhibitor ofPFK, is the metabolite whose abundance increases most. The de-

B

Hours

Arb

itrar

y U

nits

0 2 4 602468

10

L-methionine4-methylthio-2-oxobutanoic acidS-adenosyl-L-methioninehomocysteineallantoated-fructose 1,6-bisphosphate5’-methylthioadenosinecystationineNADH2-propylmalic acidL-2-aminoadipic acidD-lactatekynurenic acidascorbic acid2-dehydro-D-gluconateD-gluconateL-methylhistidinemyo-inositolmaleic acidcitraconic acidmethylmalonic acido-acetylcarnitine2-aminoisobutyric acidglycerophosphocholinedihydroxyacetone phosphateD-glyceraldehyde 3-phosphateD-ribose 5-phosphatecellobiosephosphatidylglycerophosphatepyruvatenicotinatepyridoxineL-alanineL-threoninepantothenatesucroseindole-3-carboxylic acidD-glucono-1,5-lactone 6-phosphatep-aminobenzoateL-glutamineL-dihydroorotateadenosineUDPcytidinecytosinen-acetylornithineL-prolineL-alanineL-threoninecis-aconitateD-glucose 1-phosphateL-histidinolD-fructose 6-phosphateL-malateNAD+ (1)ATPphenylpropiolic acid3-hydroxy-3-methylglutarate (1)3-hydroxy-3-methylglutarate (2)controltrehalosecysteineisocitrate

NADP+hexosesedoheptulose 7-phosphateglutathione (1)controlalpha-D-glucose 6-phosphatethiamineshikimate 3-phosphateglutathione (2)NAD+ (2)4-pyridoxic acidinosineguanosineADPphenylpyruvateL-ornithineadenineAMPn-acetyl-glutaminen6-acetyl-L-lysineglutamateL-glutamineL-pipecolateL-arginineL-asparagineL-leucineudp-D-glucoseL-phenylalanineL-histidineL-lysineO-acetylhomoserineudp-n-acetyl-D-glucosamineL-aspartateserineUTPCTPGTPATPcitrullinetrehalose 6-phosphate

L-valineL-isoleucinetyrosinetryptophann-acetylputrescinecitrate (1)citrate (2) -7 0 7

A

Fig. 3. Metabolite abundance during methionine starvation of met13Δ. (A)Metabolite abundance data in arbitrary units of fold-change relative to thezero time point. Colored bars indicate the metabolites exhibiting thegreatest changes in abundance (range is ±128-fold). (B) NAD:NADH ratioduring the time course.

E1092 | www.pnas.org/cgi/doi/10.1073/pnas.1101494108 Petti et al.

crease inNADHabundance causes an increase in theNAD:NADHratio (Fig. 3B), which is also consistent with an increase in respi-ration. TheNAD:NADHratio is a potentially important clue to themechanism by which methionine starvation prolongs survival. Thisratio is widely thought to reflect the overall cellular energy balance,and an increase in NAD:NADH due to falling NADH has beenexplicitly implicated in longevity extension by dietary restriction(17). Taken together, the metabolite measurements support theinference made from gene expression patterns that methioninestarvation tips the metabolic balance toward respiration.There are two broad classes of metabolites whose abundance

changes in a consistent and correlated fashion: amino acid levelsincrease throughout the time course, and the nucleotide tri-phosphates ATP, GTP, UTP, and CTP increase transiently. Theincrease in amino acid abundance may result from a combinationof factors, including decreased amino acid consumption anddegradation of ribosomes and other proteins.

Genes at the Intersection of Methionine Metabolism, Oxidative StressResponse, and Mitochondrial Function Are Required for MaximalSurvival Under Multiple Starvations. Methionine-specific gene ex-pression depends largely on the activity of several transcriptionfactors, Met4p, Met31p, Met32p, Met28p, and Cbf1p (18). Wedeleted these transcription factors individually and in pairwisecombinations and found that the double mutantmet31Δmet32Δdies much faster than the met6Δ and met13Δ mutants during me-thionine starvation (Fig. 1A and Table S1). (met4Δ also dies muchmore quickly but was not further analyzed in this work.) Theportion of methionine-specific gene expression regulated byMet31p and Met32p is therefore required for survival duringmethionine starvation. met31Δmet32Δ arrests in G0/G1 withslightly lower efficiency than the other (metabolic) methionineauxotrophs (Fig. 1B), but it does conserve glucose (Fig. 1C).To test whether the MET31/MET32-dependent expression

program is relevant to survival more generally, we measured thesurvival of the met31Δmet32Δ double mutant during phosphatestarvation, which differs more frommethionine starvation than theother starvations studied. Deleting MET31 and MET32 signifi-cantly decreases survival during phosphate starvation (Fig. 1A, P=1× 10−4), but the decrease is significantly smaller than the decreaseobserved under methionine starvation (P = 3 × 10−7) (SI Materialsand Methods). The Met31p/Met32p regulon is therefore requiredfor full survival during methionine and phosphate starvation.To better understand how Met31p and Met32p contribute to

survival, we measured gene expression during methionine star-vation of met31Δmet32Δ. These data were combined with thatfrom met6Δ and met13Δ, and multiple regression was used toidentify genes whose expression is significantly different inmet31Δmet32Δ compared with met6Δ and met13Δ (SI Materialsand Methods). We then selected genes that rank higher than the95th or 90th percentile with respect to the F statistic P value (P ≤9.19 × 10−6 or 2.07 × 10−4; FDR ≤ 0.01), depend statisticallysignificantly on time (FDR ≤ 0.01), and change twofold or morein at least one time course. The resulting genes are stronglyenriched for (i) ribosomal protein-encoding genes and (ii) genesthat function at the intersection of methionine biosynthesis,regulation of cell division, and mitochondrial function, such aspurine biosynthesis, folate metabolism, serine and glycine me-tabolism, iron metabolism, inositol metabolism, and aspartatemetabolism (Table S3 and Dataset S1).This gene set contains two particularly informative subsets (Fig.

4). The first, which depends on Met31p and Met32p for induction,is enriched for sulfur and methionine metabolism, oxidation-reduction, siroheme biosynthesis, glutathione biosynthesis, and re-sponse to oxidative stress (including key oxidative stress responsegenes such asGSH1,ZWF1,GTO1,OPT1, andMXR1). The secondsubset, which is repressed bymet31Δmet32Δ, is strongly enriched forgenes associated with purine, nucleotide, or glycine biosynthesis, or

with iron homeostasis. The most striking examples are the purinebiosynthetic genes SHM2, FCY2, ADE1, ADE2, ADE5,7, ADE13,and ADE17; the folate metabolic genes MTD1, GCV2, and GCV3;and several genes—FET3,FTR1,SIT1, andTIS11—that are inducedwhen mitochondrial iron-sulfur cluster biogenesis is disrupted (19).

Gene Expression Patterns Correlated with Starvation PhenotypeAcross Multiple Nutrient Starvations. As a first step toward identi-fying and characterizing the starvation signal proposed in previouswork (2), we compared gene expression data for methionine,phosphate, sulfate, leucine, and uracil starvations to find expres-sion signatures that correlate with various starvation phenotypes.We analyzed the combined expression data for these nutrientswith the hope of identifying groups of genes whose expressionpatterns are correlated with survival, arrest, and glucose conser-vation. To this end, we first filtered out expression patterns thatare common to all starvations and expression patterns specific toa single starvation. We then asked three related questions: (i)Which genes are expressed differently in the “survivable” (i.e.,glucose-conserving, cell cycle arresting, long-lived) starvationscompared with the “unsurvivable” (i.e., glucose-wasting, shortest-lived) starvations? (ii)Which genes are expressed differently in allof the auxotrophs (including methionine) compared with thelongest-lived, natural nutrient starvations? (iii) Which expressionpatterns best correlate with survival half-life?

Genes That Are Expressed Differently Between Survivable andUnsurvivable Starvations. To identify genes that are expresseddifferently during the survivable starvations (methionine, phos-phate, sulfate) than during the unsurvivable starvations (leucine,uracil), we used two independent methods, multiple regression(Fig. 5A) and k-means clustering (SI Materials and Methods andFig. S2B), which yielded similar results. Multiple regression wasused to extract a set of genes with a statistically significant de-pendence on the survivability of the starvation (SI Materials andMethods). As before, we selected genes that rank higher thanthe 95th or 90th percentile with respect to the F statistic P value(P ≤ 1.6 × 10−14 or 9.03 × 10−11; FDR ≤ 5.7 × 10−14 or 1.6 × 10−10),depend statistically significantly on time (FDR ≤ 0.01), andchange twofold or more in at least one time course. Hierarchicalclustering of these genes revealed two subclusters, one charac-terized by stronger induction in the successful starvations, theother by stronger repression (Fig. 5A and Dataset S1).Overall, the functional enrichment of these clusters implies

that successful starvation is correlated with expression of genesthat support stress response and several functions executedlargely within the mitochondria, such as aerobic metabolism,oxidative stress response, and redox homeostasis (Fig. 5A andTable S4). More specifically, genes expressed more highly in thesuccessful starvations are enriched for heat stress, mitochondrialtranslation, and energy generation through respiration, including

M(6)

M(13)

M(TF)

Methionine biosynthesis; sulfur assimilation; DNA repair; antioxidant biosynthesis

Iron homeostasis; folate cycle; purines

Fig. 4. MET31/MET32-dependent genes and oxidative stress response.Genes whose expression depends most strongly on MET31 and MET32 asdetermined by multiple regression (Materials and Methods). Functional en-richment is indicated as in Fig. 2 and further described in Table S3.

Petti et al. PNAS | November 8, 2011 | vol. 108 | no. 45 | E1093

SYST

EMSBIOLO

GY

PNASPL

US

the Gene Ontology (GO) categories “oxidation reduction” and“generation of precursor metabolites and energy.” (In GO, thedescendents of the “oxidation reduction” term are involved inmitochondrial respiration and its constituent functions, such aselectron transport and ATP synthesis. “Generation of precursormetabolites and energy” contains TCA cycle components andenzymes that transfer electrons from metabolites to the electrontransport chain.) Specifically, this gene set contains severalindicators of aerobic metabolism (e.g., HXK1, ENO1, PGK1, andGPM1), which we also identified using k-means. With respect tooxidative stress response, it contains key regulators of oxidativestress, including the superoxide dismutase SOD1, the thioredoxinTRX3, and the glutathione synthetase GTT1, as well as TDH1,BNA2, RCK2, PST2 (a Yap1p target), OYE3, and the pentosephosphate pathway components TAL1 and GND2.This gene set is also enriched for the mitochondrial folate

cycle [including glycine and serine metabolism (e.g., SER33)](Table S4) and for several documented starvation-responsegenes (SVF1, SSA4, GIS1, SIP2, and PNC1) that were alsoidentified using k-means. Although not statistically enriched asa GO class, we also identified in this cluster a number of phos-pholipid biosynthetic genes, including OPI1 (a PKA-activatedtranscriptional repressor of phospholipid biosynthetic genes),OPI3, CHO1, CHO2,GPI6, and SER33. This gene set is enrichedfor genes whose products are localized to the mitochondrion,cytoplasm, and cell membrane.

Genes That Are Expressed Differently During Starvation for NaturalNutrients vs. Starvation for Auxotrophic Requirements. Presumably,the auxotrophic starvations—including methionine—share somephysiological features that limit survival relative to the naturalnutrient starvations. We used multiple regression to identifygenes with a statistically significant dependence on nutrient class(either auxotrophic or “natural”), selecting genes that rankhigher than the 95th or 90th percentile as above (P ≤ 1.2 × 10−13

or 1.5 × 10−9; FDR ≤ 5.4 × 10−13 or 3.3 × 10−9). Hierarchicalclustering of the significant genes yielded two main subclusters,one containing genes that are more highly induced duringphosphate or sulfate starvation than in methionine, leucine, oruracil starvation, and a second with the opposite expressionpattern (Fig. 5B, Table S5, and Dataset S1). [Results obtainedusing k-means clustering (Fig. S2C) are discussed in SI Results.]The first cluster, containing genes that are more highly

expressed in the natural nutrient starvations, is enriched foroxidative stress response (e.g., BNA2, GTO3, GRX2, TRX2,OYE3, and HYR1). Using FunSpec (20) to identify enrichmentfor MIPS functional categories (21), we found enrichment forgenes involved in glutamate biosynthesis and genes involved ingenerating input to the citric acid cycle and thence to the elec-tron transport chain (e.g., IDH1, IDH2, ACO1, and MEU1).[This cluster is also enriched for genes annotated to the GOglucose metabolism category (YOR283W, PGK1, HXK2, TDH3,ENO2, and TDH1); because most of those genes function in bothglycolysis and gluconeogenesis, the meaning of this enrichmentis ambiguous.]

Genes with positive

dependence on successful

starvation

Response to heat; oxidation-reduction; mitochondrial translation; glycine & serine metabolism; TCA cycle; glycolysis; gluconeogenesis; membrane fraction; assorted mitochondrial; large ribosomal subunit; AMP-activated protein kinase

Genes with negative

dependence on success-

ful starvation

Assorted amino acid biosynthesis; nucleotide transport; nucleocytoplas-mic transport; response to glucose stimulus; vitamin biosynthetic process

-3 30

M(6)

L P S UM(13)

M(T)

M(6)

L P S UM(13)

M(T)

A

B

Higher expression in

auxotrophic starvations

Assorted amino acid metabo-lism; autophagy; triglyceride & glycerol catabolism; cardiolipin metabolism; phosphatidylglyc-erol metabolism; sulfur metabolism; vitamin biosyn-thesis

Higher expression in

natural nutrient starva-

tions

Glycolysis; glutamate biosynthesis; Red genes: oxidation-reduction; TCA cycle enzyme complex, cytoplasm, assorted mitochondrion, etc.

M(6)

L P S UM(13)

M(T)

M(6)

L P S UM(13)

M(T)

Positive correlation

Response to heat; oxidation reduction; electron transport chain; TCA cycle; glucose catabolism; redox homeostasis; response to oxidative stress; membrane; cytoplasm; assorted mitochondrial

M(6)

L P S UM(13)

M(T)

Negative correlation

Assorted amino acid metabolism; tRNA aminoacylation; pyrimidine transport; folic acid biosynthesis; cytosolic ribosome; mitochondrion

C M(6)

L P S UM(13)

M(6)

L P S UM(13)

M(T)

Fig. 5. Genes that differ by starvation phenotype or nutrient class, as iden-tified using multiple regression. Each panel shows genes ranking above the90th percentile with respect to statistical significance for the given compari-son. Functional enrichment and color coding are as specified in Fig. 2. Theseare independent analyses that address related questions (see text) and giveconsistent results; thus, the resulting gene sets (and their functional enrich-ments) necessarily overlap. (A) Expression profiles shared by successful star-vations. Genes that are expressed differently in the survivable starvations(Met, Pho, and Sul) than in the unsurvivable starvations (Leu and Ura).

Functional enrichment is further described in Table S4. (B) Expression profilesdependent on nutrient class. Genes that are expressed differently in thenatural nutrient starvations (Pho and Sul) than in the auxtrophic starvations(Met, Leu, and Ura). Genes with particularly strong induction in Pho and Sul(red boxes) are strongly enriched for functions related to aerobic metabolism.Functional enrichment is further described in Table S5. (C) Expression profilescorrelated with survival time. The gene expression template used for thePavlidis template matching (SI Materials and Methods) is shown above geneswhose expression profiles match the template mean with a Pearson correla-tion coefficient of at least 0.7 (“Positive correlation,” Left) or −0.7 (“Negativecorrelation,” Right). Functional enrichment is further described in Table S6.

E1094 | www.pnas.org/cgi/doi/10.1073/pnas.1101494108 Petti et al.

The second cluster, characterized by greater induction duringthe auxotrophic starvations, is strongly enriched for amino acidbiosynthesis (as expected), autophagy, metabolism of glycer-olipids, cardiolipin, phosphatidylglycerol, and related compounds.This cluster is also enriched for the MIPS categories related tomorphogenesis and cell wall integrity (e.g., MKK2, SLT2, RLM1,SOG2, and SOG2) and contains PPM1, whose deletion dramati-cally increases the survival during leucine starvation (1, 22, 23).

Gene Expression Program Correlated with Survival Time over AllTypes of Starvations. The preceding analyses began with classi-fications based on either (i) the observation that met6Δ andmet13Δ auxotrophs survive longer during starvation than leu2Δ orura3Δ auxotrophs, or (ii) the fact that methionine-requiringmutants are auxotrophs. To avoid potential bias introduced byprior classification, we used Pavlidis template matching (24) toidentify genes that are highly correlated with survival time per se(91 genes with correlation at least ±0.7; Fig. 5C, Table S6, andDataset S1). These genes are significantly enriched for the sameheat and oxidative stress, redox, citric acid cycle, and respiration-related processes previously identified, with many genes havingannotations in several of these categories.

Overlap Between Met31p/Met32p-Dependent Gene Expression andSurvival-Correlated Gene Expression. Deletion of MET31 andMET32 significantly decreases survival during methionineand phosphate starvation. To determine whether Met31p andMet32p promote survival by the same means as the survival-correlated gene expression signature discussed above, we ex-amined the intersection between Met31p/Met32p-dependentgenes and survival-correlated genes. Nine genes are shared be-tween these gene sets. They are involved in methylmethionineimport and metabolism (MMP1 and MHT1), methionine im-port (MUP1), glutathione transport (OPT1), purine bio-synthesis (SHM2 and ADE5,7), the folate cycle (MTD1 andGCV2), and glycine/serine metabolism (SER33). From this weconclude that, through Met31p and Met32p, these methionine-regulated processes must be properly regulated during starva-tion for a variety of nutrients, not just methionine.

Cells Undergoing Survivable Starvations Consume More Oxygen, butan Intact Electron Chain Is Not Required for Survival. The analysis ofgene expression patterns directly correlated with survival impli-cates processes associated with mitochondria and processesassociated with oxidative stress response. We carried out twoexperiments to better characterize the involvement of theseprocesses. In one, we studied oxygen consumption during star-vation, and in the other, we studied survival in strains lackingmitochondrial DNA (and therefore lacking an electron transportchain). Fig. S4 shows that there are indeed differences in oxygenconsumption for different starvation regimes and that the cul-tures with poorer survival consume less oxygen. By the secondday of starvation, methionine-starved (met6Δ) cells consumeslightly more oxygen than leucine-starved or methionine-starved (met31Δmet32Δ) cells, but the difference is not statis-tically significant. Strikingly, cells starved for phosphate (FY4or met6Δ) consume by far the most oxygen (P = 4.3 × 10−7),paralleling their extraordinarily prolonged survival.These results led us to test whether respiration is required

for starvation survival. We used ethidium bromide to createρ0 (cytoplasmic petite) derivatives of the strains tested above(Materials and Methods) and measured starvation survival as de-scribed above (Fig. 6A). The ρ0 derivatives survive nearly per-fectly in phosphate starvation, almost as well in methioninestarvation of the metabolic methionine auxotroph, and verypoorly in leucine starvation and methionine starvation of theregulatory methionine auxotroph. Loss of the electron transportchain (and all of the other functions encoded in the mitochon-

drial DNA) results in only marginal impairment of survival of themethionine regulatory mutant and the leucine mutant. Becausethe presence of an electron transport chain had no consistenteffect on survival, we conclude that survival does not depend onrespiration per se (although respiration may contribute to sur-vival in certain cases). Of course, this experiment does not ruleout the possibility that the efficiency of respiration in yeast withan intact electron transport chain also influences survival.

Survival Is Correlated with Ability to Detoxify ROS. If survival is notaffected by the respiration genes we identified in the expressionanalysis, it may be affected by the other respiration-associated

Pho (FY4) Leu (leu2 )

met32 )Met (met6 )

A

B

0.001

1

100

0 7 14

0.001

1

100

0 7 14

0 7 14

0 7 14%

via

bilit

y%

via

bilit

y

days days

days days

ρ0 strainsρ+ strains

0

5

10

15

20

25

30

35

40

45

Met(met31

met32 )

Metmet6 )

Pho(FY4)

Leu(( leu2 )

% v

iabi

lity

replicate 1 replicate 2

Met (met31

Fig. 6. (A) Survival of ρ0 derivatives during starvation for phosphate, me-thionine, and leucine. Survival of ρ0 derivatives of FY4, met6Δ, leu2Δ, andmet31Δmet32Δ (black lines) was measured as the ability to form colonies onrich medium plates (Materials and Methods). Survival of the ρ+ parent strainsis shown for comparison (gray lines). (B) Viability after hydrogen peroxidetreatment. For duplicate cultures limited for phosphate (FY4), methionine(met6Δ or met31Δmet32Δ), or leucine (leu2Δ), colony-forming units on richmedium plates after treatment with 20 mM hydrogen peroxide, a source ofoxidative stress, is reported as a percentage of colony-forming units aftertreatment with water. Black and gray bars indicate biological replicates.

Petti et al. PNAS | November 8, 2011 | vol. 108 | no. 45 | E1095

SYST

EMSBIOLO

GY

PNASPL

US

genes we identified. In our data, oxidative stress response is oneof the most prominent respiration-associated functions. More-over, the data in Fig. 4 suggest that impairing the oxidative stressresponse, such as by deleting MET31 and MET32, impairs sur-vival in multiple nutrient starvations. To test the idea that star-vation survival correlates more generally with the ability tosurvive oxidative stress, we measured the survival of hydrogenperoxide-treated cultures limited for methionine (met6Δ andmet31Δmet32Δ), leucine (leu2Δ), or phosphate (FY4) (Materialsand Methods). Fig. 6B shows that starvation survival is indeedhighly correlated (Pearson correlation, 0.997) with the robust-ness of the oxidative stress response.

DiscussionPrevious work suggested the existence of a starvation signal-and-response network that is differently activated in different nutri-tional environments, thereby leading to different phenotypes (2).Under this interpretation, the signal-and-response network istriggered during starvation for naturally required nutrients (e.g.,sulfate, phosphate, nitrogen, and carbon), resulting in cell cyclearrest, glucose conservation, and survival, but is not triggeredduring starvation for auxotrophic requirements. We sought tobetter characterize the function of this signal by studying starva-tion for diverse nutrients. We have shown that cells undergoingsurvivable starvations (i.e., those in which cells survive for manydays or weeks) induce expression of genes involved in mitochon-drial function and/or stress response. They also survive externallyapplied oxidative stress far better than those undergoing unsur-vivable starvations. In addition, we have shown that cells un-dergoing survivable starvations consume more oxygen than cellsundergoing unsurvivable starvations, but that respiration per se isnot required for prolonged starvation survival. We conclude thatthe differences between the survivable and unsurvivable starva-tions represent differential activation of the signal-and-responsenetwork, and that this network controls functions that supportstress response and nonrespiratory mitochondrial functions.

Role of Mitochondrial Function and Stress Response in StarvationSurvival. According to our gene expression analyses, “successful”response to starvation is associated with induction of genes thatcan be broadly characterized as supporting stress response and/ormitochondrial function. When we identified genes whose ex-pression is correlated with survival time during starvation, wefound that many of them are involved in three basic cellularfunctions: (i) electron transport, aerobic respiration, and generalmitochondrial activity (such as mitochondrial translation), (ii)acetyl-CoA and storage carbohydrate metabolism, and (iii) re-sponse to heat and oxidative stress. Some of the genes involved inthe latter are mitochondrial and help maintain mitochondrialintegrity and respiratory efficiency—for example, through syn-thesis of antioxidants such as superoxide dismutase, thioredoxin,peroxiredoxin, glutaredoxin, and glutathione, which neutralizeROS generated by electron transport (25). Additionally, aerobic-specific isozymes are preferentially induced relative to theirfermentation-specific counterparts, particularly in methioninestarvation. The fact that many genes typically associated with thecanonical MSN2/MSN4-mediated stress response cocluster withgenes involved in aerobic metabolism suggests that stress re-sponse and aerobic metabolism are transcriptionally coordinated.Physiological measurements suggest that the relationship be-

tween survival and mitochondrial function may well be causativerather than merely correlative: we and others have shown thatforcing the cell to respire during starvation (for example, byproviding only a nonfermentable carbon source), increases sur-vival time. The case for causality is supported by a recent studythat measured survival of the prototrophic yeast deletion collec-tion during phosphate and leucine starvation (26): genes whosedeletion decreased survival in both starvations were more highly

enriched for localization in the mitochondrion than any othercellular component.Determining more precisely whether, how, and why mitochon-

dria influence starvation survival is a complex task. The variousprocesses that occur in the mitochondrion might influence star-vation survival in different, even opposing, ways. For instance,electron transport generates ROS that might shorten survival, butcompensatory stress response mechanisms might ultimatelylengthen survival. Indeed, the literature is rife with conflictingreports on the role of mitochondrial function in chronologicalaging (e.g., refs. 14 and 27). Our results concerning survival ofmutants lacking mitochondrial DNA show that respiration per seis not the major determinant of survival.If survival does not depend on respiration, then it may depend

on the stress response pathways that seem to be coupled to re-spiratory growth. This dependency is supported by our resultswith the short-lived met31Δmet32Δ mutant, which cannot mounta response to oxidative stress, and our results showing that abilityto detoxify hydrogen peroxide is correlated with starvation sur-vival across a variety of starvations. Previous work suggested thatglucose signaling during leucine and uracil starvation activatesthe Tor pathway, leading to an inappropriate progrowth signalthat results in cell death (1). Our data suggest that the survival-correlated stress response pathways are the same ones that arerepressed by the Tor pathway [many of which are regulated byMSN2/MSN4 (23)]. Thus, survivable starvation conditions mayrepress the Tor pathway, thereby derepressing the Tor-regulatedstress response pathways.Other recent systems-level gene expression analyses have also

found connections between mitochondrial functions and stressresponse pathways. It may be that over evolutionary time, induc-tion of oxidative metabolism became associated with inductionof protective mechanisms (e.g., thioredoxins, glutathione, super-oxide dismutases, etc.) that provide recourse against respiration-generated ROS. Consequently, inducing oxidative metabolismmay increase the abundance of antioxidants, whose protectiveeffects may benefit the entire cell. This interpretation suggests thatoxidative metabolism indirectly exerts a protective effect on thecell, in direct contrast with the standard view that increased oxi-dative metabolism is necessarily associated with increased abun-dance of damaging ROS.

Methionine Regulates Cellular Redox Homeostasis and OxidativeStress Response, Which Contributes to Longevity During MethionineStarvation. Our results beg the question of why starvation formethionine (in met6Δ or met13Δ mutants) is more like starvationfor a natural nutrient. Our interpretation, based on the similarityof gene expression during methionine starvation and sulfurstarvation, is that methionine is the “indicator metabolite” forsulfate availability. Our data show that regulation of methioninemetabolism, iron and sulfur metabolism, oxidative metabolism,antioxidant biosynthesis, and one-carbon metabolism is highlyinterdependent. Thus, methionine auxotrophs ultimately up-regulate some or all of the protective functions normally asso-ciated with increased oxidative metabolism. It is therefore notsurprising that deleting MET31 and MET32, which curtails sur-vival during methionine starvation, phosphate starvation, andhydrogen peroxide exposure, severely affects the expression ofoxidative stress response genes.We conclude that proper regulation of these interdependent

processes, particularly induction of the oxidative stress responsegenes, contributes to longevity in multiple starvations. This sup-ports previous evidence for the role of oxidative stress in aging (28–32) and contributes to a growing body of evidence suggesting thatmethionine abundance regulates the pathways that protect the cellfrom the ravages of oxidative metabolism. In higher organisms,methionine restriction has been found to increase longevity and

E1096 | www.pnas.org/cgi/doi/10.1073/pnas.1101494108 Petti et al.

reduce oxidative damage to proteins and mitochondrial DNA andto decrease ROS generated in the mitochondria (32).

Nutrient Starvation in Yeast as a Model for the Warburg Effect inCancer Cells. A major metabolic hallmark of cancer is rapid glu-cose consumption even in the presence of oxygen, known as theWarburg effect. Although the uncontrolled glucose consumptionof yeast starving for leucine or uracil is superficially analogous,there was no previous evidence that the two phenomena havea common cause. The Warburg effect was long thought ancillaryto more fundamental molecular changes in cancer, but recentwork suggests that it results directly from oncogene activation.Mounting evidence suggests that oncogene activation may causemitochondrial dysfunction, which increases the cell’s dependenceon glycolysis (33) and the abundance of ROS (25, 34).Our gene expression data suggest that glucose wasting in yeast

may also be caused by failure to induce a variety of mitochon-drial functions. In our data, a major difference between glucose-conserving and glucose-wasting starvations is greater inductionof the mitochondrial genes in the glucose-conserving starvations.This suggests that glucose wasting during certain yeast starva-tions may be a useful model for the Warburg effect.

Nutrient Starvation in Yeast as a Model for Aging. A variety ofstudies on chronological and replicative aging show that a met-abolic regime favoring respiration instead of glycolysis increasesyeast lifespan (1, 14, 35–38), but there is conflicting evidence forthe role of respiration itself (27, 39, 40). Our results suggest thatthe situation is more nuanced: survival during starvation, oftencalled “longevity,” is correlated with the induction of a non-respiratory subset of mitochondrial functions, particularly thepathways that alleviate heat and oxidative stress. Protectivemechanisms may be coinduced with oxidative metabolism, suchthat metabolic perturbations that favor respiration over fer-mentation may indeed increase longevity. We believe that theessential connections between longevity, metabolism, and stressresponse may well be conserved in whole or in part betweenyeast and higher organisms, and that data from yeast may welllead to a better understanding of the interplay between metab-olism, stress response, cell growth control, and longevity.

Materials and MethodsMedia compositions are listed in Table S7. Strain descriptions are listed inTable S8.

Establishing the Range of Linear Dependence on Methionine Concentration.The range of linear dependence was determined as in ref. 3, using 0, 5, 10, 20,40 or 200 mg/L [i.e., 0, 33.5, 67, 134, 268.1, or 1340.4 μM (μmoles/L)] me-thionine. A methionine concentration of 7.5 mg/L (i.e., 50.3 μM) lies in thecenter of the linear range.

Methionine Dependence and Growth on Different Carbon Sources. A singlecolony of each strain was grown to steady state in richmedium and spotted inthreefold serial dilutions on YNB agar plates containing 20 g/L glucose or12.6 g/L glycerol and 8.1 g/L ethanol as the carbon source, with or without20 mg/L methionine.

Cell Cycle Arrest. A single colony of each strain was grown for 24 h in 5–7 mLof methionine-limited (7.5 mg/L) minimal medium, then diluted to a densityaround 1 × 105 to 5 × 105 cells/mL in fresh methionine-limited medium.Beginning at the end of lag phase (12–15 h after dilution), turbidity (Klettunits), cell density (Coulter Count), and bud index were measured periodi-cally. Bud index, the fraction of unbudded cells in a stationary phase culture(Table S1 lists strain-specific stationary phase Klett values), was measured bymanual scoring of a sonicated culture under a microscope.

Starvation Survival Assays. Survival assays on various carbon sources wereperformed as described in ref. 1. Here, nutrient-limited medium contained7.5 mg/L methionine, 40 mg/L leucine, or 13.3 mg/L phosphate plus 200 mg/L

methionine or 240 mg/L leucine. ρ0 mutants were generated using ethidiumbromide as described in ref. 41.

Hydrogen Peroxide Survival Assays. A single colony was inoculated intonutrient-limited medium and grown for 24 h with shaking at 30 °C. Theculture was diluted into fresh nutrient-limited medium and grown withshaking at 30 °C until midexponential phase (the length of time required isstrain-specific). An aliquot containing 1 × 108 cells was resuspended in 10 mLof fresh nutrient-limited medium, treated with 0 mM or 20 mM hydrogenperoxide for 20 min, washed twice with water, and plated for viable cells.Viability in 20 mM hydrogen peroxide is reported as a percentage of theviable cells counted after treatment in 0 mM hydrogen peroxide.

Residual Glucose Measurements. A single colony was inoculated into nutrient-limited medium (7.5 mg/L methionine or 40 mg/L leucine) and grown for 24 hwith shaking at 30 °C. Each culture was diluted into fresh nutrient-limitedmedium to a density between 1 × 105 and 4 × 105 cells/mL and grown withshaking at 30 °C. Starting 12–15 h after dilution, turbidity was measured,and 1-mL samples were frozen periodically over the course of 3 d. Theglucose concentration in each thawed sample was measured in triplicateusing the Boehringer-Mannheim D-glucose UV test.

Methionine-Starvation Filter-Switching Time Courses for met6Δ, met13Δ,and met31Δmet32Δ. A single colony of each strain was inoculated intomethionine-limited minimal medium and grown with shaking at 30 °C. After24 h, each culture was diluted to low density (1 × 105 to 1 × 106 cells/mL) in250–300 mL of fresh methionine-limited minimal medium and grown withshaking for 12–15 h to a density of 20–25 Klett units (approximately 5 × 106

cells/mL). As described in ref. 42, the culture was then filtered in 5-mL ali-quots onto 0.45-μm, 82-mm-diameter Nylon filters. Each filter was placed ona Petri dish containing YNB5 (minimal medium made with agarose) and7.5 mg/L methionine. The plates were incubated at 30 °C for 4 h, until theculture reached a density of 80–90 Klett units (approximately 2 × 107 cells/mL). At this time, each filter was transferred to a fresh plate containingYNB5 but no methionine. At specified time points thereafter (0 min, 10 min,30 min, 1 h, 1.5 h, 2 h, 2.5 h, 3 h, 3.5 h, 4 h, and 6 h), filters from two plateswere placed in 50-mL screw-capped tubes and immediately immersed inliquid nitrogen. A third filter was washed and used for Klett readings,Coulter counts, and bud index measurements.

Metabolite Measurements in met13Δ. During the filter-switching time coursefor met13Δ described above, samples were simultaneously collected for me-tabolite extraction. Metabolites were extracted using methanol quenching,and their abundance was measured using LC-MS/MS as described in ref. 43.

Oxygen Consumption Measurements. A single colony was inoculated into 5 mLof nutrient-limited medium (7.5 mg/L methionine, 40 mg/L leucine, or 13.3mg/L postassium phosphate) and grown overnight with shaking at 30 °C. Theculture was diluted to a density between 1 × 105 and 4 × 105 cells/mL in 300mL of fresh nutrient-limited medium and grown overnight with shaking at30 °C. At Klett 20 (≈5 × 106 cells/mL), 250 mL of the culture was transferredto a chemostat vessel equipped with two dissolved-oxygen probes andgrown to stationary phase in batch mode (400 rpm, 30 °C) with an airflowrate of 40. The oxygen probes were calibrated in water at 25 °C, so that fullprobe saturation corresponded to a known dissolved oxygen concentration(0.27 mMol O2/L). Three days after inoculation of the chemostat vessel,cellular oxygen consumption was measured by periodically turning off thechemostat air supply for 20 min and recording the change in dissolved ox-ygen per unit time during the first 10 min.

Methionine-Depletion Batch-Growth Time Course for met13Δ. A single colonywas inoculated into 100 mL methionine-limited minimal medium and grownwith shaking at 30 °C for 32 h. The culture was then diluted to 5 × 105 cells/mL in 500 mL of fresh methionine-limited minimal medium and grown inbatch with shaking. Beginning 5 h after dilution and every hour thereafteruntil the culture reached steady state with respect to density and bud index,samples were removed for bud index, culture density, and cell countmeasurements, and cell samples for RNA extraction were obtained by fil-tering and freezing in liquid nitrogen.

Preparation of Cells for Microarray Reference Samples. In a chemostat, FY4was grown to steady state in chemostat medium containing limiting phos-phate (10 mg/L) at a dilution rate of 0.17 h−1. Cells were harvested by filteringonto a nitrocellulose filter and frozen in liquid nitrogen.

Petti et al. PNAS | November 8, 2011 | vol. 108 | no. 45 | E1097

SYST

EMSBIOLO

GY

PNASPL

US

RNA Isolation, Labeling, and Hybridization. RNA was collected from thefrozen filters using phenol-chloroform extraction, purified using an theQiagen RNeasy kit, labeled with Cy5 CTP (samples) or Cy3 CTP (phosphate-limited chemostat reference) using the Agilent Low-input Linear Ampli-fication kit, and hybridized to 2 × 105k, 4 × 44k, or 8 × 15k Agilent YeastOligo V2 microarrays. Arrays were washed and scanned using extendeddynamic range according to Agilent protocols. Agilent Feature ExtractionSoftware was used with default settings and linear and lowess dyebias correction.

Expression and Metabolite Data Analysis. Expression and metabolite dataanalysis are described in SI Materials and Methods.

ACKNOWLEDGMENTS. We thank Sanford Silverman and Viktor Boer forhelpful discussions of this work and Olivia Ho-Shing for technical assistancewith oxygen consumption measurements. A.A.P. was supported by RuthKirschstein Cancer Training Grant T32 CA-009528. Research was supportedby National Institute of General Medical Sciences Center for QuantitativeBiology Grant P50 GM-071508 and National Institutes of Health Grant GM-046406 (to D.B.).

1. Boer VM, Amini S, Botstein D (2008) Influence of genotype and nutrition on survivaland metabolism of starving yeast. Proc Natl Acad Sci USA 105:6930–6935.

2. Unger MW, Hartwell LH (1976) Control of cell division in Saccharomyces cerevisiae bymethionyl-tRNA. Proc Natl Acad Sci USA 73:1664–1668.

3. Saldanha AJ, Brauer MJ, Botstein D (2004) Nutritional homeostasis in batch andsteady-state culture of yeast. Mol Biol Cell 15:4089–4104.

4. SpellmanPT, etal. (1998)Comprehensive identificationofcell cycle-regulatedgenesof theyeast Saccharomyces cerevisiae bymicroarray hybridization.Mol Biol Cell 9:3273–3297.

5. Klevecz RR, Bolen J, Forrest G, Murray DB (2004) A genomewide oscillation intranscription gatesDNA replication and cell cycle. ProcNatl Acad Sci USA 101:1200–1205.

6. Murray DB, Beckmann M, Kitano H (2007) Regulation of yeast oscillatory dynamics.Proc Natl Acad Sci USA 104:2241–2246.

7. Tu BP, Kudlicki A, Rowicka M, McKnight SL (2005) Logic of the yeast metabolic cycle:Temporal compartmentalization of cellular processes. Science 310:1152–1158.

8. Kresnowati MTAP, et al. (2006) When transcriptome meets metabolome: Fast cellularresponses of yeast to sudden relief of glucose limitation. Mol Syst Biol 2:49.

9. Wheeler GL, Trotter EW, Dawes IW, Grant CM (2003) Coupling of the transcriptionalregulation of glutathione biosynthesis to the availability of glutathione andmethioninevia the Met4 and Yap1 transcription factors. J Biol Chem 278:49920–49928.

10. Yen JL, Su N-Y, Kaiser P (2005) The yeast ubiquitin ligase SCFMet30 regulates heavymetal response. Mol Biol Cell 16:1872–1882.

11. Brauer MJ, Saldanha AJ, Dolinski K, Botstein D (2005) Homeostatic adjustment andmetabolic remodeling in glucose-limited yeast cultures. Mol Biol Cell 16:2503–2517.

12. Brauer MJ, et al. (2008) Coordination of growth rate, cell cycle, stress response, andmetabolic activity in yeast. Mol Biol Cell 19:352–367.

13. Kaeberlein M, et al. (2005) Regulation of yeast replicative life span by TOR and Sch9 inresponse to nutrients. Science 310:1193–1196.

14. Bonawitz ND, Chatenay-Lapointe M, Pan Y, Shadel GS (2007) Reduced TOR signalingextends chronological life span via increased respiration and upregulation of mito-chondrial gene expression. Cell Metab 5:265–277.

15. Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. ProcNatl Acad Sci USA 100:9440–9445.

16. Voet D, Voet JG, eds (1995) Biochemistry (John Wiley & Sons, New York), 2nd Ed.17. Lin SJ, Ford E, Haigis M, Liszt G, Guarente L (2004) Calorie restriction extends yeast life

span by lowering the level of NADH. Genes Dev 18:12–16.18. Lee T, et al. (2010) Dissection of combinatorial control by the Met4 transcriptional

complex. Mol Biol Cell 21:456–469.19. ChenOS, et al. (2004) Transcriptionof the yeast iron regulondoes not responddirectly to

iron but rather to iron-sulfur cluster biosynthesis. J Biol Chem 279:29513–29518.20. Robinson MD, Grigull J, Mohammad N, Hughes TR (2002) FunSpec: A web-based

cluster interpreter for yeast. BMC Bioinformatics 3:35.21. Ruepp A, et al. (2004) The FunCat, a functional annotation scheme for systematic

classification of proteins from whole genomes. Nucleic Acids Res 32:5539–5545.22. Wu WS (2006) The signaling mechanism of ROS in tumor progression. Cancer

Metastasis Rev 25:695–705.23. Zaman S, Lippman SI, Zhao X, Broach JR (2008) How Saccharomyces responds to

nutrients. Annu Rev Genet 42:27–81.

24. Saeed AI, et al. (2006) TM4 microarray software suite.Methods Enzymol 411:134–193.25. Murphy MP (2009) How mitochondria produce reactive oxygen species. Biochem J

417:1–13.26. Gresham D, et al. (2011) System-level analysis of genes and functions affecting

survival during nutrient starvation in Saccharomyces cerevisiae. Genetics 187:299–317.27. Kaeberlein M, et al. (2005) Increased life span due to calorie restriction in respiratory-

deficient yeast. PLoS Genet 1:e69.28. Finkel T, Holbrook NJ (2000) Oxidants, oxidative stress and the biology of ageing.

Nature 408:239–247.29. Richie JP, Jr., et al. (1994) Methionine restriction increases blood glutathione and

longevity in F344 rats. FASEB J 8:1302–1307.30. Miller RA, et al. (2005) Methionine-deficient diet extends mouse lifespan, slows

immune and lens aging, alters glucose, T4, IGF-I and insulin levels, and increases

hepatocyte MIF levels and stress resistance. Aging Cell 4:119–125.31. Pamplona R, Barja G (2006) Mitochondrial oxidative stress, aging and caloric

restriction: The protein and methionine connection. Biochim Biophys Acta 1757:

496–508.32. Sanz A, et al. (2006) Methionine restriction decreases mitochondrial oxygen radical

generation and leak as well as oxidative damage to mitochondrial DNA and proteins.

FASEB J 20:1064–1073.33. Matoba S, et al. (2006) p53 regulates mitochondrial respiration. Science 312:

1650–1653.34. Sung HJ, et al. (2010) Mitochondrial respiration protects against oxygen-associated

DNA damage. Nat Commun 1:5.35. Lin SJ, Defossez PA, Guarente L (2000) Requirement of NAD and SIR2 for life-span

extension by calorie restriction in Saccharomyces cerevisiae. Science 289:2126–

2128.36. Lin S-J, et al. (2002) Calorie restriction extends Saccharomyces cerevisiae lifespan by

increasing respiration. Nature 418:344–348.37. Easlon E, Tsang F, Skinner C, Wang C, Lin SJ (2008) The malate-aspartate NADH

shuttle components are novel metabolic longevity regulators required for calorie

restriction-mediated life span extension in yeast. Genes Dev 22:931–944.38. Kaeberlein M, Burtner CR, Kennedy BK (2007) Recent developments in yeast aging.

PLoS Genet 3:e84.39. Masoro EJ (2005) Overview of caloric restriction and ageing. Mech Ageing Dev 126:

913–922.40. Guarente L (2007) Sirtuins in aging and disease. Cold Spring Harb Symp Quant Biol 72:

483–488.41. Fox TD, et al. (1991) Analysis and manipulation of yeast mitochondrial genes.

Methods Enzymol 194:149–165.42. Brauer MJ, et al. (2006) Conservation of the metabolomic response to starvation

across two divergent microbes. Proc Natl Acad Sci USA 103:19302–19307.43. Boer VM, Crutchfield CA, Bradley PH, Botstein D, Rabinowitz JD (2010) Growth-

limiting intracellular metabolites in yeast growing under diverse nutrient limitations.

Mol Biol Cell 21:198–211.

E1098 | www.pnas.org/cgi/doi/10.1073/pnas.1101494108 Petti et al.