YeastModel UncoversDual RolesofMitochondriaintheActionofArtemisininWeiLi1,WeikeMo1,DanShen1,LiboSun1,JuanWang1,ShanLu1,JaneM.Gitschier2,BingZhou1*1TheStateKeyLaboratoryof BiomembraneandMembraneBiotechnology, Department of Biological SciencesandBiotechnology, TsinghuaUniversity, Beijing, China,2DepartmentsofMedicineandPediatricsandHowardHughesMedical Institute, UniversityofCalifornia, SanFrancisco, California, UnitedStatesofAmericaArtemisinins, derivedfromthewormwoodherbArtemisiaannua, arethemost potent antimalarial drugscurrentlyavailable. Despite extensive research, the exact mode of action of artemisinins has not been established. Here we useyeast, Saccharamyces cerevisiae, toprobe the core workingmechanismof this class of antimalarial agents. Wedemonstrate that artemisinins inhibitory effect is mediatedby disruptingthe normal functionof mitochondriathrough depolarizing their membrane potential. Moreover, in a genetic study, we identify the electron transport chainas an important player in artemisinins action: Deletion of NDE1 or NDI1, which encode mitochondrial NADHdehydrogenases, confersresistancetoartemisinin, whereasoverexpressionofNDE1orNDI1dramaticallyincreasessensitivity toartemisinin. Mutations or environmental conditions that affect electrontransport alsoalter hostssensitivity to artemisinin. Sensitivity is partially restored when the Plasmodium falciparum NDI1 ortholog is expressedin yeast ndi1 strain. Finally, we showed that artemisinins inhibitory effect is mediated by reactive oxygen species. Ourresultsdemonstratethatartemisininseffectisprimarilymediatedthroughdisruptionofmembranepotentialbyitsinteractionwiththeelectrontransport chain, resultingindysfunctional mitochondria. Weproposeadual roleofmitochondria played during the action of artemisinin: the electron transport chain stimulates artemisinins effect, mostlikelybyactivatingit,andthemitochondriaaresubsequentlydamagedbythelocallygeneratedfreeradicals.Citation: Li W, MoW, ShenD, SunL, WangJ, etal (2005)Yeastmodel uncoversdual rolesofmitochondriaintheactionofartemisinin. PLoSGenet1(3): e36.IntroductionMalaria, the most prevalent and pernicious parasitic diseaseof humans, is estimated to kill between 1 million and 2 millionpeople, mainly children, each year. Resistance has emerged toall classes of antimalarial drugs except artemisinin, anendoperoxide antimalarial drug derived froman ancientChineseherbal remedy, qinghaosu. Themechanismof theaction of artemisinin remains a mystery, although ironappears tobeinvolvedinactivatingthis endoperoxidetogeneratecytotoxicfreeradicals[1]. Several candidateshavebeen hypothesized as targets of artemisinins, including haem,a translationally controlled tumor protein and some parasitemembrane proteins [13], but none of these have beenconvincingly shown to be functionally relevant. Recently,Eckstein-Ludwiget al. [4] proposedPfATP6, asarco/endo-plasmicreticulumCa2-ATPase, as artemisinins target andinferred that artemisinin might act by mobilizing intracellularCa2stores; however, this conclusion has also been debated [5].One important reason for the lack of satisfying progress inunderstandingartemisininisthatitisdifculttocarryoutgeneticanalysis of malariaparasites. Incontrast, theyeastSaccharamyces cerevisiae is an ideal model organism to uncover avariety of molecular mechanisms that otherwise might bedifcult to address with other systems. In this study, wedevelopedayeast model andusedit toprobethefunda-mental mechanisms of artemisinins action.ResultsArtemisininInhibitsYeastGrowthinNonfermentableMediaIn a pilot experiment, we tested whether artemisinin is toxicto the growth of yeast cells. As demonstrated in Figure 1A, theinhibitoryeffect of artemisininonyeast growthinYPD(afermentable mediumwithdextrose as the carbonsource)liquidwas not detectableat 8lM, andonlystartedtobeobservable at 80 lM, a concentrationfar above the levelneeded to inhibit the growth of Plasmodium falciparum (a 50%growth inhibition concentration [IC50] of approximately a fewnM) [6,7]. On YPD plates, no signicant inhibition of growth iseven noticeable in the presence of 80 lM artemisinin.This level of insensitivity to artemisinin led us to considerwhetheruseoftheYPDmediaprecludedourassessmentofone important organelle, the mitochondrion, in the action ofartimisinin. Yeast cangroweitheranaerobicallyoraerobi-cally by utilizing fermentable or nonfermentable carbonsources. In the presence of a fermentable carbon source suchas glucose(dextrose), as inYPD, yeast preferentiallyadoptglycolysis to generate energy even under aerobic conditions,and can grow normally even when mitochondrial respirationlevelisminimal. Inordertodeterminewhetherartemisinininterfereswithmitochondrial function, yeast wasgrowninYPGorYPEmediainwhichglucosewasreplacedwiththenonfermentable carbonsource glycerol or ethanol. Amaz-ReceivedMay9, 2005; AcceptedAugust8, 2005; PublishedSeptember16, 2005DOI: 10.1371/journal.pgen.0010036Copyright: 2005Li, et al. Thisisanopen-accessarticledistributedunder thetermsof theCreativeCommons AttributionLicense, whichpermitsunrestricteduse, distribution, andreproductioninanymedium, providedtheoriginal authorandsourcearecredited.Abbreviations: BPS, bathophenanthrolinedisulfonicacid; IC50, 50%growthinhib-itionconcentration; ROS, reactiveoxygenspecies; SG, syntheticyeastmediawithglycerolasthecarbonsource;YPG(E),yeastmediawithglycerolorethanolasthecarbonsource; YPGE, yeastmediawithglycerol andethanol ascarbonsourcesEditor: SusanDutcher, WashingtonUniversity, UnitedStatesofAmerica*Towhomcorrespondenceshouldbeaddressed: zhoubing@tsinghua.edu.cnPLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0329ingly, inYPG(E) media, yeastgrowthwasseverelyinhibitedwhen artemisinin was present. Figure 1A shows the dramaticreduction of cell growth in the presence of 8 lM. Indeed, foryeastgrowninYPG, theIC50(denedhereas50%growthinhibitionin48h) wasdeterminedtobeapproximately10nM(Figure 1B), a concentration comparable to what iseffective in P. falciparum. The dramatic contrast of sensitivitybetween yeast grown on artemisinin-supplemented YPD andYPG(E) mediaindicates that artemisininaffects mitochon-drial function in YPG(E).ArtemisininDepolarizestheMitochondrial MembraneTheobservationofgrowthinhibitionofyeastonartemi-sinin-supplemented YPG(E) media prompted us to investigatewhetherthemitochondrial membranepotential of yeast isaffectedbyartemisinin, becausemaintenanceof themem-branepotentialiskeytomitochondrialfunctions, includingoxidativephosphorylationandmetabolismof aminoacids,lipids, and haem, as well as intracellular Ca2homeostasis. Asshown inFigure 1C, treatment of yeast with artemisinincausedsubstantiallossofmembranepotential, asevidencedbythedecreasedrodamineuptakeintothemitochondria.This loss of membrane potential was also observed with yeastgrown in YPD media, but required the presence of a slightlyhigherconcentrationof artemisininoralongertreatmenttime.GeneticStudiesReveal ElectronTransportChainPlaysImportantRolesintheActionofArtemisininTogainfurtherinsightsintothemechanismofactionofartemisinin, wescreenedforyeast genesthat, wheninacti-vated, can confer artemisinin resistance. A library comprisedof 4,757 homozygous deletionstrains [8,9] was platedonYPG(E)agarplateswith35lMofartemisinin. Threegeneswhose deletion resulted in artemisinin resistance wereidentied. Two of these, NDE1 and NDI1, encode yeastNADHdehydrogenasesinthemitochondrialelectrontrans-port chain [1012]. ndi1D and nde1D displayed similar level ofgrowthonartemisinin-supplemented YPGEplates (Figure2A). Overexpression of NDE1 and NDI1 in their correspond-ingdeletionstrainsorintheparental BY4742background(wild type) resulted in signicantly increased sensitivity to thedrug(Figure2B, rst threerows). Thesedatasuggest thatNADHdehydrogenaseactivityis positivelycorrelatedwithyeast sensitivity to artemisinin.InadditiontoNDE1andNDI1, athirdhomologousyeastgene, NDE2, is known to provide NADHdehydrogenaseactivity in yeast. These NADH dehydrogenases oxidize NADHat either the cytosolic side (Nde1p and Nde2p) or the matrixside (Ndi1p) of the inner mitochondrial membrane. Althoughdeletion of NDE1 leads to a major loss of NADH dehydrogen-aseactivityatthecytosolicside, deletionofbothNDE1andNDE2 results in complete loss of NADHdehydrogenaseactivity at the cytosolic side [1113]. Because of sequence andfunctionsimilaritybetweenNde1p, Ndi1p, andNde2p, weasked whether NDE2 interacts with artemisinin. Compared tothat of nde1Dor ndi1D, nde2D manifested only marginalFigure1.ArtemisininInhibitsYeastRespiratoryGrowthbyDepolarizingtheMitochondrial Membrane(A) Artemisinin (Art) inhibits yeast growth in nonfermentablemedia. InYPD the effect ofartemisinin isminimal, whereas in YPG, artemisininis highlyeffective.(B) Yeast growth is inhibited by artemisinin in YPG with an IC50 that is comparable to that required to kill cultured malaria parasites. Relative growth inthe presence of artemisinin was measured against to that of the yeast grown in the absence of artemisinin. Experiments shown were performed threetimesinliquidYPGmedia. Errorbarsrepresentstandarderrorsofthemeanforeachassay.(C) Artemisinin depolarizes mitochondrial membrane. The peak shift toward the left represents a decrease of fluorescence signal indicating the loss ofmembranepotential. CellsweregrowninYPGwithorwithoutartemisinin(Art)for2h.DOI: 10.1371/journal.pgen.0010036.g001PLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0330ModeofActionofArtemisininSynopsisMalaria kills at least 1 million people worldwide a year. Recent yearssaw the rapid emergence of drug-resistant malaria strains.Artemisinins, derivedfromtheChinesewormwoodherbArtemisiaannua, are the most potent antimalarials currently available. Despiteextensive research, the exact mode of action of artemisinins has notbeenestablished. Inthis article, Li et al. investigatedyeast as amodel to probe the core working mechanismof this class ofantimalarials. Theyshowedthatartemisinincandisruptthenormalfunctionof mitochondriabydepolarizingitsmembranepotential,andthatartemisininseffectcanbeaffectedbyitsinteractionwiththe mitochondrial electron transport chain, an apparatus thatcouplesoxygenoxidationandenergygenerationinthecell. Theyproposedadual roleof mitochondriaplayedduringtheactionofartemisinin:theelectrontransportchainlikelyactivatesartemisinin,andthe mitochondria are subsequently damagedby the locallygeneratedfree radicalsassociated withthisactivation.The researchhas provided a fine tool for the study of the mechanismofartemisinin in a model organism (yeast), and laid the framework forasetofpossiblefutureexperimentstobeconductedinyeastandmalariaparasites.resistance to artemisinin, and the double mutant nde1D nde2Dwas not substantially more resistant to artemisinin thannde1D. This is consistent withthe observationthat at thecytosolic side Nde1p is the major NADH dehydrogenase andNde2p only plays a minor role [1012].Next, weinvestigatedwhether thedoublemutant nde1Dndi1Dismoreresistanttotheactionofartemisininornot.nde1D ndi1Dyeastcangrowwell inglucose media.However,their growth on nonfermentable media is greatly affected (seeFigure2A), indicatingseverelycompromisedmitochondrialfunction of the strain. This growth defect precludes ourcomparative analysis of nde1D ndi1D resistance againstartemisinin.Although three NADHdehydrogenases are present inyeast, only one such dehydrogenase, PfNDI1 (PFI0735c),with closest homology to Ndi1p, was found in the P.falciparum genome database [13]. Similarly, Nde1p andNde2pdonothavecounterpartsinmammaliancells, whichutilizeredoxshuttlemechanismstocouplecytosolicNADHoxidation to NADH dedydrogenase (complex I) in thematrix side. To test whether PfNDI1 expression in yeastconfersartemisininssensitivity, PfNDI1wasampliedfroma cDNApool of P. falciparum. As expected, expressionofPfNDI1 in ndi1D background partially restored its sensitivityto artemisinin (Figure 2C). This indicates that PfNDI1expressedinyeast interacts withartemisininina similarfashiontoNdi1p.Thethirdartemisinin-resistantstrain is sip5D(Figure2A).Little is knownabout Sip5pexcept that it biochemicallyinteracts withSnf1 proteinkinase andReg1/Glc7 proteinphosphatase [14], both of which are involved in theregulation of alternative carbon source utilization andrespiration. Thus, it is possible that SIP5 somehow indirectlyaffects energy generation, or, morespecically, mitochon-drial respiration. To test the hypothesis that respirationcontrol is involved in artemisinin resistance, mutants ofsevenothercomponents involvedincarbon-sourceregula-tionandrespiration[15] wereexamined. Mostofthemarevariablyresistanttotheactionofartemisinin(seeMaterialsand Methods).Theepistatic relationship of SIP5 with NDI1 or NDE1 wasalsoexplored. OverexpressionofSIP5inthecontrol strain,ndi1Dor nde1Ddidnot have any effect onsensitivity orresistancetoartemisinin. Incontrast, wheneitherNDI1orNDE1 is overexpressed in sip5D background, yeast still exhibithypersensitivitytoartemisinin. Thesendings suggest thatsip5D affects artemisinin resistance indirectly by acting on theelectrontransport chain. Theidenticationof thesethreegenesthataredirectlyorindirectlyinvolvedinrespirationreinforcedourpreviousinterpretationthatartemisininactsthrough mitochondria.Because the electrontransport chainis involvedintheaction of artemisinin, we asked further whether othercomponents downstreamof NADHdehydrogenaseinteractwith artemisinin. We have checked available deletionmutants(25strainsaltogether)ofotherknowngenesintheelectrontransportchainforpossibleartemisininresistance.Although some of these strains cannot grow on nonferment-able media, of those that can, none appears to be resistant tothe action of artemisinin (see Materials and Methods).Althoughit is likelysomeof thesemutants haveimpairedelectron transport activity, the fact that all 25 deletionstrains do not manifest either artemisinin hypersensitivity orartemisinin resistance (compared to the wild-type strain)implies that not all components of the pathway interact withartemisinin.ArtemisininsActionGeneratesReactiveOxygenSpeciesGeneration of oxidative stress is believed to be involved inthe antiparasitic effects of artemisinin [16]. Figure 3A showsthat the levels of reactive oxygen species (ROS) produced inthree strains after treatment with artemisinin positivelycorrelatedwiththeirlevelsofsensitivitytoartemisinin. Wethusinvestigatedwhethertheartemisinin-resistantmutantsthat we isolated were cross-resistant to other oxidative agents.Unexpectedly, these mutants did not show resistance toperoxideor paraquat, and, if anything, manifestedhyper-Figure 2. The Genetic Screen for Artemisinin-Resistant MutationsIdentifiedGenesintheElectronTransport Chainor inthePathwayofRespiratoryControl(A) The three mutants isolated display increased resistance toartemisinin. YPGEplateswithor without 4lMartemisininwereused.nde1Dndi1Dexhibitedseveregrowthdefectinnonfermentablemedia.(B)Increased activities ofNADHdehydrogenasesexacerbate artemisininsensitivity, andSip5maybepositionedupstreamof NADHdehydro-genases. Plates areall SG-Ura (withor without 4lMartemisinin) toprevent plasmid loss. ADH1-NDE1 and ADH1-SIP5 here denote constructsthat expressNDE1andSIP5under thecontrol of ADH1promoter. Theresults of ADH1-NDI1 are similar to that of ADH1-NDE1 and are not shownonthetwoplates.(C) Expression of PfNDI1 in ndi1D restores yeast sensitivity to artemisinin.PlatesusedhereareSG-Ura(withorwithout8 lMartemisinin).Art, artemisinin; SG, syntheticyeastmediawithglycerol asthecarbonsource; WT, wildtype.DOI: 10.1371/journal.pgen.0010036.g002PLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0331ModeofActionofArtemisininsensitivity to these oxidative agents (Figure 3B). On the otherhand, althoughsod1Dandsod2Dmutants displayincreasedsensitivity to oxidative stress, neither of them exhibited moresensitivity to artemisinin than the wild type (result notshown). This result, combined with the observation that yeastcells are unaffected by artemisinininYPD, suggests thatartemisinin is not simply a promiscuous oxidant. Instead, wesuggest that artemisinin may have local effects on themitochondria.Because activation of artemisinin is known to depend uponthe cleavage of a peroxide bridge via aniron-dependentmechanism [17], we investigated the interaction between ironandartemisinininyeast. Whenbathophenanthrolinedisul-fonic acid (BPS), an iron chelator, was added into YPG mediatogether with 8 lMartemisinin, the inhibitory effect ofartemisinin on growth was attenuated (Figure 3C). In aseparateseries of experiments todeterminewhetherextraironcould aggravate the effect of artemisinin, we addedbetween1and100lMFe2 orFe3totheplates. Attheseconcentrationsiron-enhancedtoxicitywasnotobserved. Inaddition, incubationwithironbeforeartemisininexposuredoes not exacerbateyeast sensitivitytoartemisinin. Theseresults suggest iron plays an important role in the activationof artemisinin, but additional iron above a certain thresholdlevel has nofurther effect. This is inagreement withthereport that artemisinincanalsoeffectivelykill parasitesatstages that lack haemozoin [18].DiscussionInthis report we developeda sensitive yeast model toanalyze the action of artemisinin. We demonstrated thatartemisininis able todepolarize the mitochondrial mem-brane and that increased electron transport activity increasesthe sensitivity of the cell to artemisinin. Further, we showedartemisininproduces anincreaseinROSbyamechanismthat differs from that of general oxidants.Althoughlowamountsofartemisininaffectyeastgrowthonly innonfermentable media, inbothYPG(E) andYPDmedia artemisinin depolarizes the mitochondrial membrane.The fact that YPD slightly attenuates the effect of artemisininindepolarizingmitochondrial membranepotentialispossi-bly due to the partial repression of the activity of respirationenzymes (including NDI1 and NDE1) by glucose [19]. Likewise,mutations that indirectly affect respiration will possibly alsoaffect artemisininsensitivity. The positive correlationob-served between respiration activity and sensitivity to artemi-sinin suggests that the electron transport chain acts tostimulate the activityofartemisinin. As such, the mitochon-drial electron transport chain per se does not seem to be theprimary target of artemisinin; instead, the electron transportchainseems toplay a role inactivating artemisinin. Theactivatedartemisininmaythenlocallydepolarizethemito-chondrial membrane, impedingmanyfunctions dependentupon its potential. Although the disruption of mitochondrialFigure3.ArtemisininGeneratesROSinYeastWhenapplicable, 8 lMartemisininwasused.(A)Artemisinin-resistantstrainsgeneratefewerROS. Yeastuntreatedwithartemisininwasusedasthecontrol. Theexperimentwasperformedthreetimeswithsimilarresults. NDE1denotestheoverexpressorstrainofNDE1drivenbyADH1promoter.(B)Isolatedartemisinin-resistantstrainsarenotcross-resistanttoparaquatorperoxide. Shownherearethewild-type(WT)parental strain(BY4742),nde1Dandndi1DonYPDplateswithoutorwith0.02%paraquat.(C) Iron is possibly involved in artemisinin (Art) activation. Addition of BPS to the medium reduces yeasts sensitivity to artemisinin, whereas BPS alonedoes not enhance general yeast survival on drug-free plates. We did not use a higher amount of BPS to further reduce the iron level because a severereductioninirondramaticallyaffectsyeastgrowthonYPG.DOI: 10.1371/journal.pgen.0010036.g003PLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0332ModeofActionofArtemisininfunctiondoes not dramaticallyaffect yeast growthinYPDmedia, for P. falciparum, loss of membrane potential willadditionally affect pyrimidine biosynthesis [20], a keymetabolic process for the survival of the parasite. Thisproposed mode of action for artemisinin differs from that ofatovaquone, abroad-spectrumantiparasiticdrug, whichhasbeenshowntoinhibitcomplexIIIoftheelectrontransportchainand, consequently, collapsemitochondrial membranepotential and kill P. falciparum [21].Our ndings indicate that iron is involved in the action ofartemisinin. Mitochondria are well known to be a rich sourceof transition metals, including iron and copper. Conceivably,an active electron transport chain provides the electronsource to Fe-S in catalyzing artemisinin and producescarbon-centeredfreeradicals. Althoughwecannot excludethepossibilitythatothercomponentsintheelectrontrans-portingchains may interact withartemisinin, our currentndingssuggestthatlimitedcomponents, includingNADHdehydrogenases, are involved in this process.Distinctive ultrastructural changes have been noted inartemisinin-treatedP. falciparum, includingmarkedswellingof the mitochondria, followed by the appearance of electron-dense chromatin materials in the nuclei, clumping ofribosomes, nuclearmembraneblebbing, andsegregationofthe nucleoplasm [2226]. At high artemisinin concentration,even mammalian cells partially lose their mitochondriafunction[27,28]. These observations are consistent withaprimaryroleof mitochondriaintheactionof artemisinin,whereas other physiological changes are probably secondary.Not surprisingly, P. falciparum has a functional mitochondrialelectrontransport systemandanoxygen-requiringsystemthat is necessary for growth and survival [2931]. The relativeinsensitivityofmammaliancellstoartemisininsmaybeduetothefundamental structural difference of their electrontransport chains; for example, human complex I is composedofatleast43componentswhereasitscounterpartsinyeastandP. falciparumaremuchsimpler(onepolypeptidechain).Indeed, aBLASTsearchwithNde1por Ndi1pas aqueryrevealed no signicant homolog in the human genome. Evenmore, nosignicant homologs couldbefoundinavailableNational Center for Biotechnology Information (NCBI)databases across the whole Metazoankingdom, indicatingNADH dehydrogenase is not evolutionarily highly conservedbetween yeast and Metazoa.Theyeast model provides us withinsights intomodeofaction of artemisinin: It interacts with the electron transportchain, generates local ROS, and causes the depolarization ofthe mitochondrial membrane. This scheme agrees with whatweknowaboutthechemical propertiesofartemisininandndings in malaria studies. Although some physiologicaldifferences exist betweenyeast andmalarial parasites, thepathway or basic principles that we have derived inourstudies of artemisinininyeast should aid future malariaresearch. The yeast model that we have developed provides anewwaytoprobetheactionof artemisinin. Our ndingssuggest many future genetic and biochemical analyses of thisnatural, powerful, and mysterious drug.MaterialsandMethodsYeast growth and library screen. Standard yeast media and growthconditions were used. The yeast deletion library used is a collection of4,757homozygous diploidS. cerevisae strains (BY4743: MATa/MATahis3D1/his3D1 leu2D0/leu2D0 lys2D0/met15D0/ura3D0/ura3D0) inwhich each strain has a single open reading frame replaced with theKanMX4 module. We used homozygous diploid strain library for theoriginal screenbecausehaploidstrainappearstohavearelativelyhigher rate of screening background, presumably due to spontaneousmutation. An aliquot of the pooled yeast library was plated on YPG(2% glycerol as carbon source) or YPE (3% ethanol as carbon source)agar plates supplemented with 35 lM artemisinin. (Altogether about100,000total colonies wereplated). Artemisinin-resistant colonieswere isolated, and serial dilutions were made and spotted onartemisinin plates to conrm the original phenotype. About a dozenrelativelymore resistant colonieswerechosen, andthe correspond-inggeneswereidentiedbybubblePCRorinversePCRandDNAsequencing analysis. Phenotypes of these mutants were furtherconrmed in haploid background (BY4742: MATa his3D1 leu2D0lys2D0 ura3D0).Artemisinin sensitivity/resistance assay. For growth in liquid, yeastweregrownovernight inYPDmedium, spundown, washedthreetimes with YPG medium, and diluted to an A600 of 0.5. Then, 10 ll wasinoculated into YPD or YPG media with or without artemisinin, andA600was measured over time. For growth on agar plates, yeastpreviously grown in liquid YPDwas spotted with 10-fold serialdilutions.Besides SIP5, genes that may be involved in carbon-sourceutilization were individually checked for artemisinin sensitivity/resistance. These include GAL83, SIP1, 2, 3, and 4, and SNF1 and4. snf1 cannot grow on YPG. sip1D, sip3D, sip4D, and gal83D displayartemisinin resistance.Mutants of genes in the electron transport chain, including SDH1,2, 3, and4, COR1, CYT1, QCR2, 6, 7, 8, 9, and -10, RIP1, INH1,STF1and-2, andCOX4, 5A, 5B, 6, 7, 8, 9, 12, and-13wereindividually examined for artemisinin resistance. Among these, SDH1and -3, CYT1, QCR9, and COX4, 6, 7, and -13 are required for yeastgrowthonYPGmedia, precludinganalysisoftheirinvolvementinartemisinins action.Gene deletion and expression. Genereplacementwasachievedbyhomologousrecombination with URA3 or LEU2 as the replacementmarkerinstrainBY4742background. GenedeletionswereveriedbyPCR. Forexpression, SIP5, NDI1, NDE1, andPfNDI1werePCRamplied fromBY4743 genomic DNAor a cDNAlibrary of P.falciparum. AmpliedtargetDNAswerethenclonedintoanADH1-driven expression vector (pADH1-YES2, a vector modied frompYES2 fromInvitrogen (Carlsbad, California, United States), inwhichtheGAL1promoterwas replacedwithanADH1promoter).Yeast transformation was done by the lithiumacetate method.Transformed cells were selected on SD-Ura. Sequences for PCRprimers usedfor SIP5, NDI1, NDE1, andPfNDI1areavailablebyrequest.Analysis of ROS production. Flow cytometric analysis was used toassay the production of free intracellular radicals. Briey, cells wereincubated with dihydrorhodamine 123 for 2 h and then analyzed by aFACS Calibur (Becton Dickinson, San Jose, California, United States)atalowowratewithexcitationandemissionsettingsof488and525550 nm (lter FL1), respectively.Assay of the electrochemical potential. After treatment in 20 mMHEPESbuffer(pH7.4) containing50mMglucose, 1ml ofthecellsuspensionwasincubatedwith2lMRh123(rhodamine123)for30min, washed, andthenresuspendedin100ll PBS. Mitochondrialelectrochemical potential was expressed as the uorescence intensityof Rh123, which was read through a FACS Calibur (Becton Dickinson)with excitation at 480 nm and emission at 530 nm.AcknowledgmentsWe thank C. Vulpe for reading the paper. The P. falciparum cDNA wasa kindgift fromP. Rosenthals lab. JMGis aninvestigator withHoward Hughes Medical Institute. This work is supported by grantsfromTsinghua985(toBZ), NSFC(30470973and30330340toBZ),ChinaPostdoctoral ScienceFoundation(toWL), andtheExcellentYoung Teacher Program of MOE, Peoples Republic of China (to BZ).Competing interests. The authors have declared that no competinginterests exist.Author contributions. WLandBZconceivedanddesignedtheexperiments. WL, WM, DS, JW, LS, andSLperformedtheexperi-ments. BZanalyzedthedata. JMGcontributedreagents/materials/analysis tools. WL, JMG, and BZ wrote the paper. &PLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0333ModeofActionofArtemisininReferences1. MeshnickSR(2002) Artemisinin: mechanisms of action, resistance andtoxicity. Int J Parasitol 32: 16551660.2. AsawamahasakdaW, Benakis A, MeshnickSR(1994) Theinteractionofartemisinin with red cell membranes. J Lab Clin Med 123: 757762.3. BhisutthibhanJ, PanXQ, HosslerPA, WalkerDJ, YowellCA, etal. (1998)The Plasmodium falciparum translationallycontrolledtumor protein homo-loganditsreactionwiththeantimalarial drugartemisinin. JBiol Chem273: 1619216198.4. Eckstein-LudwigU., WebbRJ, VanGoethemID, East JM, LeeAG, et al.(2003) Artemisinins target the SERCA of Plasmodium falciparum. Nature 424:957961.5. Mercereau-PuijalonO, FandeurT(2003) Antimalarial activityofartemi-sinins: identication of a novel target? Lancet 362: 20352036.6. Reed MB, Saliba KJ, Caruana SR, Kirk K, Cowman AF (2000) Pgh1modulates sensitivity and resistance to multiple antimalarials in Plasmodiumfalciparum. Nature 403: 906909.7. Price RN, CassarC, BrockmanA, Duraisingh M, vanVugt M, et al.(1999)Thepfmdr1geneis associatedwithamultidrug-resistant phenotypeinPlasmodiumfalciparumfromthewesternborder of Thailand. AntimicrobAgents Chemother 43: 29432949.8. GiaeverG, ChuAM, Ni L, ConnellyC, Riles L, et al. (2002) Functionalproling of the Saccharomyces cerevisiae. Genome. Nature 418: 387391.9. Steinmetz LM, Scharfe C, Deutschbauer AM, Mokranjac D, Herman ZS, etal. (2002) Systematic screen for human disease genes in yeast. Nat Genet 31:400404.10. Luttik MA, Overkamp KM, Kotter P, de Vries S, van Dijken JP, et al. (1998)The Saccharomyces cerevisiae NDE1 and NDE2 genes encode separatemitochondrial NADH dehydrogenases catalyzing the oxidation of cytosolicNADH. J Biol Chem 273: 2452924534.11. Marres CA, de Vries S, Grivell LA (1991) Isolation and inactivation of thenucleargeneencodingtherotenone-insensitiveinternal NADH: Ubiqui-noneoxidoreductaseof mitochondriafromSaccharomyces cerevisiae. EurJBiochem 195: 857862.12. Small WC, McAlister-Henn L (1998) identication of a cytosolicallydirectedNADHdehydrogenasein mitochondria of Saccharomyces cerevisiae.J Bacteriol 180: 4051405513. Sanz P, Ludin K, Carlson M (2000) Sip5 interacts with both the Reg1/Glc7protein phosphatase and the Snf1 protein kinase of Saccharomyces cerevisiae.Genetics 154: 99107.14. GardnerMJ, HallN, FungE, WhiteO, BerrimanM, etal. (2002)Genomesequence of the human malaria parasite Plasmodium falciparum. Nature 419:498511.15. Schuller HJ (2003) Transcriptional control of nonfermentative metabolismin the yeast Saccharomyces cerevisiae. Curr Genet 43: 139160.16. KrungkraiSR,YuthavongY (1987)Theantimalarialactionon Plasmodiumfalciparumofqinghaosuandartesunateincombinationwithagentswhichmodulate oxidant stress. Trans R Soc Trop Med Hyg 81: 710714.17. MeshnickSR, YangYZ, LimaV, KuypersF, KamchonwongpaisanS, etal.(1993) Iron-dependent free radical generation from the antimalarial agentartemisinin (qinghaosu). Antimicrob Agents Chemother 37: 11081114.18. ter Kuile F, White NJ, Holloway PH, Pasvol G, Krishna S (1993) Plasmodiumfalciparum: Invitrostudies of thepharmacodynamicproperties of drugsused for the treatment of severe malaria. Exp Parasitol 76: 8595.19. DeRisi JL, Iyer VR, Brown PO (1997) Exploring the metabolic and geneticcontrol of gene expression on a genomic scale. Science 278: 680686.20. Krungkrai J (2004) The multiple roles of the mitochondrion of the malarialparasite. Parasitology 129: 511524.21. Srivastava IK, Rottenberg H, Vaidya AB (1997) Atovaquone, a broadspectrum antiparasitic drug, collapses mitochondrial membrane potentialin a malarial parasite. J Biol Chem 272: 39613966.22. Ellis DS, Li ZL, Gu HM, Peters W, Robinson BL, et al. (1985) Thechemotherapy of rodent malaria XXXIX. Ultrastructural changes followingtreatmentwithartemisinineofPlasmodiumberghei infectioninmice, withobservations of the localization of [3H]-dihydroartemisinine in P. falciparumin vitro. Ann Trop Med Parasitol 79: 367374.23. KawaiS, KanoS, Suzuki M(1993) MorphologiceffectsofartemetheronPlasmodium falciparum in Aotus trivirgatus. Am J Trop Med Hyg 49: 812818.24. Maeno Y, Toyoshima T, Fujioka H, Ito Y, Meshnick SR, et al. (1993)Morphologic effects of artemisinin in Plasmodium falciparum. Am J Trop MedHyg 49: 485491.25. Jiang JB, Jacobs G, Liang DS, Aikawa M (1985) Qinghaosu-induced changesin the morphology of Plasmodium inui. Am J Trop Med Hyg 34: 424428.26. Ye ZG, Li ZL, Li GQ, Fu XQ, Liu HP, et al. (1983) Effects of Qinghaosu andchloroquine on the ultrastructure of the erythrocytic stage of P. falciparumin continuous cultivation in vitro. J Tradit Chin Med 3: 95102.27. ReungpatthanaphongP, MankhetkornS(2002) Modulationofmultidrugresistancebyartemisinin, artesunateanddihydroartemisinininK562/adrand GLC4/adr resistant cell lines. Biol Pharm Bull 25: 15551561.28. FishwickJ, EdwardsG, WardSA, McLeanWG(1998) Morphological andimmunocytochemical effects of dihydroartemisinin on differentiatingNB2a neuroblastoma cells. Neurotoxicology 19: 393403.29. GinsburgH, DivoAA, GearyTG, BolandMT, JensenJB(1986). Effectofmitochondrial inhibitorsonintraerythrocyticPlasmodiumfalciparumininvitro cultures. J Protozool 33: 121125.30. Uyemura SA, Luo S, Moreno SN, DocampoR (2000) Oxidativephosphor-ylation, Ca2transport, and fatty acid-induced uncoupling in malariaparasites mitochondria. J Biol Chem 275: 97099715.31. Krungkrai J, Burat D, KudanS, Krungkrai S, Prapunwattana P(1999)Mitochondrialoxygenconsumptionin asexualandsexualbloodstagesofthe human malarial parasite, Plasmodium falciparum. Southeast Asian J TropMed Public Health 30: 636642.PLoSGenetics| www.plosgenetics.org September2005|Volume1| Issue3| e36 0334ModeofActionofArtemisinin