Blank-2005-Large-scale 13C-flux

GenomeBiology2005,6:R49commentreviewsreportsdepositedresearchrefereedresearchinteractionsinformationOpenAccess2005Blanketal.Volume6,Issue6,ArticleR49ResearchLarge-scale13C-fluxanalysisrevealsmechanisticprinciplesofmetabolicnetworkrobustnesstonullmutationsinyeastLarsMBlank,LarsKuepferandUweSauerAddress:InstituteofBiotechnology,ETHZürich,8093Zürich,Switzerland.Correspondence:UweSauer.E-mail:[e-mail]©2005Blanketal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.Large-scale13C-fluxanalysisinyeastpGenome-scale13supC/sup-fluxanalysisinSaccharomycescerevisiaerevealedthattheapparentdispensabilityofknockoutmutantswithmetabolicfunctioncanbeexplainedbygeneinactivityunderaparticularcondition,bynetworkredundancythroughdupli-catedgenesorbyalternativepathways./pAbstractBackground:Quantificationofintracellularmetabolitefluxesby13C-tracerexperimentsismaturingintoaroutinehigher-throughputanalysis.Thequestionnowarisesastowhichmutantsshouldbeanalyzed.Hereweidentifykeyexperimentsinasystemsbiologyapproachwithagenome-scalemodelofSaccharomycescerevisiaemetabolism,therebyreducingtheworkloadforexperimentalnetworkanalysesandfunctionalgenomics.Results:Genome-scale13Cfluxanalysisrevealedthatabouthalfofthe745biochemicalreactionswereactiveduringgrowthonglucose,butthatalternativepathwaysexistforonly51gene-encodedreactionswithsignificantflux.Theseflexiblereactionsidentifiedinsilicoarekeytargetsforexperimentalfluxanalysis,andwepresentthefirstlarge-scalemetabolicfluxdataforyeast,coveringhalfofthesemutantsduringgrowthonglucose.Themetaboliclesionswereoftencounteractedbyfluxrerouting,butknockoutofcofactor-dependentreactions,asintheadh1,ald6,cox5A,fum1,mdh1,pda1,andzwf1mutations,causedfluxresponsesinmoredistantpartsofthenetwork.Byintegratingcomputationalanalyses,fluxdata,andphysiologicalphenotypesofallmutantsinactivereactions,wequantifiedtherelativeimportanceof'geneticbuffering'throughalternativepathwaysandnetworkredundancythroughduplicategenesforgeneticrobustnessofthenetwork.Conclusions:Theapparentdispensabilityofknockoutmutantswithmetabolicfunctionisexplainedbygeneinactivityunderaparticularconditioninabouthalfofthecases.Fortheremaining207viablemutantsofactivereactions,networkredundancythroughduplicategeneswasthemajor(75%)andalternativepathwaystheminor(25%)molecularmechanismofgeneticnetworkrobustnessinS.cerevisiae.BackgroundTheavailabilityofannotatedgenomesandaccumulatedbio-chemicalevidenceforindividualenzymestriggeredthereconstructionofstoichiometricreactionmodelsfornet-work-basedpathwayanalysis[1,2].Formanymicrobes,suchnetworkmodelsareavailableatthegenomescale,providingalargelycomprehensivemetabolicskeletonbyinterconnect-ingallknownreactionsinagivenorganism[3,4].Thus,net-workpropertiessuchasoptimalperformance,flexibilitytocopewithever-changingenvironmentalconditions,andPublished:17May2005GenomeBiology2005,6:R49(doi:10.1186/gb-2005-6-6-r49)Received:1February2005Revised:8March2005Accepted:6April2005Theelectronicversionofthisarticleisthecompleteoneandcanbefoundonlineat://genomebiology.com/2005/6/6/R49GenomeBiology2005,6:R49enzymedispensability(alsoreferredtoasrobustnessorgeneticrobustness[5,6])becomemathematicallytractable.Thesecomputationaladvancesarematchedwithpost-genomicadvancesinexperimentalmethodsthatassessthecell'smolecularmake-upatthelevelofmRNA,protein,ormetaboliteconcentrations.Asthefunctionalcomplementtothesecompositionaldata,quantificationofintracellularinvivoreactionratesormolecularfluxeshasbeenafocalpointofmethoddevelopmentintherealmofmetabolism[7-9].Recentprogressinincreasingthethroughputofstable-iso-tope-basedfluxanalyses[8,10,11]hasallowedthequantifica-tionoffluxresponsestomorethanjustafewintuitivelychosengeneticorenvironmentalperturbations[12-14].Nowthatfluxquantificationinhundredsofnullmutantsunderaparticularconditionisfeasibleinprinciple,thequestionarisesofwhichmutantsshouldbeanalyzed.Asperhapsthemostwidelyusedmodeleukaryote,theyeastSaccharomycescerevisiaefeaturesametabolicnetworkofabout1,200reactionsthatrepresentabout750biochemicallydistinctreactions[3,15].Isitnecessarytoquantifyfluxresponsestonullmutationsinallreactionsforacomprehen-siveviewofthemetaboliccapabilitiesunderagivencondi-tion?Toaddressthisquestion,weusedarecentlymodifiedversion(iLL672;LKuepfer,USauerandLMBlank,unpub-lishedwork)oftheoriginaliFF708genome-scalemodelpub-lishedbyFörsteretal.[3].Onthebasisofthismodel,weestimatedthegenome-scalefluxdistributioninwild-typeS.cerevisiaefrom13C-tracerexperiments,toidentifythe339biochemicalreactionsthatwereactiveduringgrowthonglu-cose.Yeastmetabolismhasthepotentialflexibilitytousealternativepathwaysfor105oftheseactivereactions.Foramajorfractionofthepotentiallyflexiblereactionsthatcata-lyze