optimalcjblaze
cjblaze 时间:2021-01-14 阅读:()
RESEARCHOpenAccessGenome-scalemodel-drivenstraindesignfordicarboxylicacidproductioninYarrowialipolyticaPranjulMishra1,Na-RaeLee1,MeiyappanLakshmanan2,MinsukKim3,Byung-GeeKim3andDong-YupLee1,2,4*FromThe28thInternationalConferenceonGenomeInformaticsSeoul,Korea.
31October-3November2017AbstractBackground:Recently,therehavebeenseveralattemptstoproducelong-chaindicarboxylicacids(DCAs)invariousmicrobialhosts.
Ofthese,Yarrowialipolyticahasgreatpotentialduetoitsoleaginouscharacteristicsanduniqueabilitytoutilizehydrophobicsubstrates.
However,Y.
lipolyticashouldbefurtherengineeredtomakeitmorecompetitive:thecurrentapproachesaremostlyintuitiveandcumbersome,thuslimitingitsindustrialapplication.
Results:Inthisstudy,weproposedmodel-guidedmetabolicengineeringstrategiesforenhancedproductionofDCAsinY.
lipolytica.
Attheoutset,wereconstructedgenome-scalemetabolicmodel(GSMM)ofY.
lipolytica(iYLI647)bysubstantiallyexpandingthepreviousmodels.
Subsequently,themodelwasvalidatedusingthreesetsofpublishedcultureexperimentdata.
Itwasfinallyexploitedtoidentifygeneticengineeringtargetsforoverexpression,knockout,andcofactormodificationbyapplyingseveralinsilicostraindesignmethods,whichpotentiallygiverisetohighyieldproductionoftheindustriallyrelevantlong-chainDCAs,e.
g.
,dodecanedioicacid(DDDA).
Theresultanttargetsinclude(1)malatedehydrogenaseandmalicenzymegenesand(2)glutamatedehydrogenasegene,insilicooverexpressionofwhichgeneratedadditionalNADPHrequiredforfattyacidsynthesis,leadingtotheincreasedDDDAfluxesby48%and22%higher,respectively,comparedtowild-type.
Wefurtherinvestigatedtheeffectofsupplyingbranched-chainaminoacidsontheacetyl-CoAturn-overratewhichiskeymetaboliteforfattyacidsynthesis,suggestingtheirsignificanceforproductionofDDDAinY.
lipolytica.
Conclusion:Insilicomodel-basedstraindesignstrategiesallowedustoidentifyseveralmetabolicengineeringtargetsforoverproducingDCAsinlipidaccumulatingyeast,Y.
lipolytica.
Thus,thecurrentstudycanprovideamethodologicalframeworkthatisapplicabletootheroleaginousyeastsforvalue-addedbiochemicalproduction.
Keywords:Yarrowialipolytica,Dicarboxylicacid,Genome-scalemetabolicmodels,Straindesign,Metabolicengineering*Correspondence:dongyuplee@skku.
eduEqualcontributors1NUSSyntheticBiologyforClinicalandTechnologicalInnovation(SynCTI),LifeSciencesInstitute,NationalUniversityofSingapore,28MedicalDrive,Singapore117456,Singapore2BioprocessingTechnologyInstitute,AgencyforScience,TechnologyandResearch(A*STAR),20BiopolisWay,#06-01,Centros,Singapore138668,SingaporeFulllistofauthorinformationisavailableattheendofthearticleTheAuthor(s).
2018OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.
0InternationalLicense(http://creativecommons.
org/licenses/by/4.
0/),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.
TheCreativeCommonsPublicDomainDedicationwaiver(http://creativecommons.
org/publicdomain/zero/1.
0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12https://doi.
org/10.
1186/s12918-018-0542-5BackgroundLong-chaindicarboxylicacids(DCAs)arewidelyusedinthemanufacturingofpolyamidesandpolyestersasthemonomericintermediates[1].
ThemostcommonlyemployedchemicalprocesstoproduceDCAsisthering-openingoxidationofcycliccompounds.
However,itrequiresexpensivestartingmaterialaswellastheenvir-onmentallyhazardousprocedures[2].
Alternatively,DCAscanbesynthesizedthroughbio-routesinvariousmicrobialhosts,suchasCandidasp.
[3,4],Yarrowialipolytica[5],Pseudomonasaeruginosa[6],Cyptococcusneoformans[7],andEscherichiacoli[8].
Ofthese,re-cently,Y.
lipolyticahasattractedgreatattentionasacellfactorytomanufactureDCAs[9]sincethisoleaginousyeastiscapableofaccumulatinglargeamountsoflipidsandpossessauniqueω-oxidationpathwaytocatalyzethehydrophobicsubstratessuchasn-alkaneandfattyacids[10–13].
Inω-oxidationpathway,aftertheoxida-tionofω-terminalofthefattyacid,fattyacidaldehydeisproducedfromω-hydroxyfattyacidbyfattyalcoholoxidase,followedbyitsoxidizationtoDCAsbyNAD-dependentfattyaldehydedehydrogenase[14].
Indeed,Y.
lipolyticaasoleaginousyeasthashugepotentialtobecomeamodelorganismtoproducefatty-acidderivedproductsincludingDCAs.
Inordertomakeitindustriallycompetitive,itsproductivityshouldbefur-therenhanced,whichcanbeachievedbymodificationofrelevanttargetgenes.
However,metabolicengineeringofY.
lipolyticahasbeenmainlyfocusedonover-producinglipids,e.
g.
,triacylglycerols(TAGs)withonlyahandfulofstudiesfortheincreasedDCAsproduction[15].
Inaddition,mostofgeneticengineeringtargetshavebeenidentifiedinanintuitiveoradhocmanner,whichlimitedourdesignscopeforstrainimprovement.
There-fore,itisnowimperativetoadoptmorerationalsystemsapproaches[16].
Inthisregard,variousstraindesignstrategiesguidedbyinsilicogenome-scalemetabolicmodel(GSMM)havebeendevelopedandsuccessfullyappliedtoseveralindustrialhostsincludingE.
coli[17,18]andS.
cerevisiae[19,20].
Similarly,inthiswork,weexploitedsuchinsilicomethodsusingGSMMofY.
lipolyticawhichwasnewlyreconstructedbysubstantiallyexpandingthepreviousmodels,thusallowingustoidentifyvariousgeneticengineeringtargetsforoverexpres-sion,knockout,andcofactormodificationtowardsDCAoverproduction.
ResultsanddiscussionGenome-scalemetabolicreconstructionofY.
lipolyticaWehavereconstructedagenome-scalemetabolicmodelofY.
lipolyticafordesigningDCAoverproducingstrains.
Initially,fourexistingY.
lipolyticaGSMMs(iNL895[21],iYL619[22],iMK735[23]andiYALI4[24])werecomparedonthebasisoftheirgeneannotations(seeMethodsandFig.
1a).
WhileiNL895wasdevelopedfromthephylogeneticallydistantyeastS.
cerevisiaemodel,iYL619wasreconstructedbasedonbiochemicalinformationdatabasessuchasKEGGandBRENDA.
iMK735wascompletelyderivedfromS.
cerevisiaemodeliND750[25]byaddingafewreactionsrelatedtoalkaneuptakeandlipidmetabolismwhereasiYALI4standsabitdifferentfromotherthreemodelssinceitwasbuiltautomaticallyusingRAVENtoolboxfromyeastconsen-susmodelastemplate.
Afterqualitativeandquantitativeassessment,iMK735wasselectedassuitablescaffoldmodelsinceithaswidercoverageofcellularphysiologyofY.
lipolyticawithbetterqualityintermsofrepresenta-tionandpredictionquality.
Withtheaimofbuildinganobjective-orientedmodelforsimulatingDCAsproduction,wefirstadded9biosyntheticreactionswhicharesequentiallycatalyzedbyhydroxylase,oxidaseanddehydrogenaseenzymeswithintheω-oxidationpathway,thuscapturingtheoxi-dationoffattyacidstoDCAs.
The27DCAsdegradationreactionswerealsoincludedtorepresentβ-oxidationpathwaytogetherwith6dicarboxylicacidstransportandanexchangereactionpertainingtoDDDAsecretion.
Basedonthescaffoldmodel,Y.
lipolyticaappearedtonotdegradeleucineandotherBCAAintoacetyl-CoA,buttherearereportsimplicatingthebiosyntheticcap-acityofleucineasaneffectoroflipogeniccapacityinoleaginousorganisms[26];aputativeacetyl-CoAprodu-cingleucinedegradationpathwaywasrecentlyidentified[27].
Therefore,wehaveadded4relevantenzymaticreactionsforleucinedegradationinthemodel.
Intotal,weadded50newreactionstothescaffoldmodelmainlyrelatedtoω-oxidationandBCAAdegradation.
Duringthemanualcuration,weremoved55reactionswhichareresponsiblefordead-endsorfutilecycleswithoutliteratureevidence.
Forexample,lactaldehydedehydrogenasewasdeletedsincethereisnotenoughexperimentalevidencetosupporttheexistenceofsuchreactioninY.
lipolytica.
Additionally,wecorrectedtheelementalbalanceanddirectionalityof45reactions.
Forexample,thedirectionalityofTHRArepresentingthethreoninealdolasemediatedbreakdownofthreoninetoglycineandacetaldehyde,waschangedtooppositedirec-tionfavoringacetaldehydeformationbasedonthedirec-tionreportedinS.
cerevisiae[28].
AllthechangesmadefromthescaffoldareprovidedinAdditionalfile1andsummarizedinFig.
1b.
Finally,theresultinginsilicoY.
lipolyticamodel(iYLI647)consistsof1347reactionsand1152metabolitesencodedby647genes(Fig.
1c).
TheiYLI647isavailableasSystemsBiologyMarkupLanguage(SBML)file(Additionalfile2).
Themodelpredictabilityofcellgrowthrelieshighlyontheaccuracyofbiomassequation.
However,wefoundthatthebiggestshortcomingofthepreviousmodelsisMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page10of130theinaccuratebiomasscompositionsincetheymainlyusedtheinformationofS.
cerevisiaebiomass.
Hence,thewholebiomasscompositionsofY.
lipolyticawerederivedbasedonvariousY.
lipolyticaexperimentaldatapublishedundercarbonlimitedandnitrogenlimitedconditionswhichcansignificantlyaltertheaminoacidandlipidcomposition.
ThecalculatedcompositionswerethenassimilatedintoiYLI647astwoseparatebiomasssynthesisequationspertainingtoC-limitedandN-limitedconditions(Additionalfile3).
Growthandnon-growthassociatedATPmaintenance(GAMandNGAM)requirementsforcellularprocesseswerealsoderivedusingrelevantliteraturedata[29].
TheNGAMofY.
lipolyticawasestimatedtobe5.
03mmolATP/gDCW,andGAMwas23.
09mmolATP/gDCW.
ComparativevalidationofiYLI647withOtherY.
lipolyticaGSMMsTheiYLI647wasevaluatedforitspredictionscorrelatedwithexperimentalphenotypes.
ThegrowthpredictionsunderdifferentcultureconditionsofiYLI647werecomparedwiththoseforotherY.
lipolyticaGSMMs.
Todoso,theinsilicobiomassyieldsofallthemodelsonglucoseandglycerolminimalmediumundersteady-stateconditionswerepredictedusingFBA.
Foreachsetofculturedatatakenfromindependentstudies,thecarbonsourceuptakerateswereconstrainedaccord-ingly,whilemaximizingbiomass.
Additionally,theCO2evolutionrate(CER)wasalsoconstrainedaccordingtotheexperimentaldatawhereverprovided.
ForthebiomassmaximizationiniMK735,biomassequationcorrespond-ingto5%lipidwasused.
Themaximumspecificgrowthratewasdeterminedforallmodelssimulatedundersameconstraintsandtheresultswerecompared(Fig.
2).
Datasets1and2weretakenfromWorkmanetal.
,[30]inwhichY.
lipolyticawasgrowninbatchcultureonglycerolandglucoseasasolecarbonsource,respectively.
Duringthebatchculturewithglucose,onlybiomassandCO2wereproducedasnometaboliteproductionwasob-served.
Inthecaseofglycerol,apartfrombiomassandCO2,smallamountofpolyolsintheformofmannitolandarabitol,werealsoproduced.
Since,theindividualbreak-downofpolyolswasnotmentionedintheliteratureandamountsweresmallenoughtodrasticallyaffecttheresult,wedidn'ttakeitintoconsiderationforsimulation.
Dataset3wastakenfromDulermoetal.
,[31]whichendeavoredtoanalyzetheY.
lipolyticamutantsforfattyacidproduc-tion.
ThecarbonsourceuptakeandCERwereconstrainedasperthevaluesreportedintheliteratureandbiomasswasmaximized.
Fig.
1ReconstructionprocessandcharacteristicofinsilicomodelsaComparisonofpreviousY.
lipolyticamodelsbSchematicdiagramhighlightingtheoverallreconstructionprocessofiYLI647,followedbyapplicationforstraindesigningcGeneralfeaturesofiYLI647incomparisonwithpreviousfourmodelsMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page11of130ItcanbeseenfromcomparativevalidationthatiYLI647canpredicttheevaluatedmacroscopicgrowthparametersmoreaccurately,i.
e.
smallerdeviationsfromexperimentaldatashowninblackdottedlines,ascom-paredtoothermodels.
Webelievethegrowthpredic-tionsaredirectlyinfluencedbytheaccuracyofthebiomasscomposition,andespecially,incaseofY.
lipoly-ticathebiomasscompositioncandrasticallychangedependingontheculturecondition.
AnotherimportantfactorthataffectsthepredictionsistheGAMandNGAMvaluesusedinthemodel.
Although,incompara-tivepredictiondoneinthecurrentstudy,wehaveusedsameNGAMinallthemodelsbutGAMvaluewasunchanged,andsomemodelsreportedveryhighGAMvalue,resultingindiscrepanciesinthemodelprediction.
MetabolicengineeringstrategiesforDCAsproductionIntheoleaginousorganism,denovoaccumulationoflipidsstartswiththeformationofanabolicacetyl-CoAviaglycolysis.
Thefattyacidsformedbytheseacetyl-CoAgetesterifiedtoformTAGs[32].
Ahighnumberofcarbonsourcesexcludingcelluloseandmethanol,havebeenconsideredassubstratesforthedenovoDCAsbiosynthesisinoleaginousmicroorganisms.
Amongthevariouscarbonsourcesavailable,weusedglucosetodesignthemetabolicengineeringstrategiesfordenovooverproductionofDDDA,arepresentativeoflong-chainDCAs.
Herein,weappliedthreemodel-guideddesignstrategiestooverproduceDDDAusingfourtools.
First,weemployedgeneticdesignbylocalsearch(GDLS)[33]tofindthegrowth-coupledsolutiontooverproduceDCAsbyknockingoutasetofreactions.
Then,weusedfluxactivityanalysis[34]toidentifythebottlenecksinthemetabolicnetworkwhichcanbeconsideredaspotentialoverexpressiontargets.
Inaddition,weimple-mentedtranscriptomic-basedstrainoptimizationtool(tSOT)[35]whichfirstgeneratesactivatedreactionsasareferencestatetoidentifythedeactivatedreactions,additionoftheeachdeactivatedreactionscanthenleadtotheincreaseintheproductyield.
Finally,tosupple-mentthemetabolicengineeringtargetswithcofactoravailability,weperformedcofactormodificationanalysis(CMA)[36]tofindthecofactorspecificityengineeringtargetsthatcanincreasethepoolofcofactorsrequiredforcatalyzingthereactions.
Figure3illustratesthecentralmetabolicnetworkofY.
lipolyticaandidentifiedgeneticengineeringtargetsforenhancingDCAsproduction.
OverexpressiontargetsExpressionlevelofgenesandtheactivityoftheirenzymeproductsarehighlyoptimizedtomeettheperformancedemandofabiologicalsystem[37].
How-ever,theseenzymeexpressionsaltertoadaptagainstthechangingbiologicalconditions.
Thisrangeoffluctua-tionsingeneexpressionlevelcanbeexploitedtodesignanoverexpressionsystem[38].
Basedonthisprinciple,weperformedfluxactivityanalysisunderglucoseminimalmediumbyfixinganoptimalbiomassandsys-tematicallyincreasingfluxactivityofeachreactionfrom0%to100%andmaximizingDDDA.
SimulationresultsshowthatwhenY.
lipolyticagrowingin10mmol/gDCW-hrofglucose,theincreaseinfluxactivityofsomereactionshasproportionaleffectonthemaximumachievableyieldofDDDA(Fig.
4).
Thesedirectionallycoupledreactionscanbegeneticallyoverexpressedtocomplementtheω-oxidationpathwayenzymestopro-duceDDDAatitstheoreticalmaximum.
Someofthebottleneckreactions,e.
g.
,DDCAH,acytochromeP450Fig.
2ComparativevalidationofiYLI647withall4Y.
lipolyticamodelsavailableunder3differentdatasets.
ThedottedlinesrepresentexperimentalvalueMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page12of130hydroxylase(CYP52)andDDCAFAO,afattyalcoholoxidaseidentifiedbyinsilicoanalysishavebeenverifiedasoverexpressiontargetsforDDDAproduction[39].
Theotherhypothesizedreactionsincludeacetyl-CoAcarboxylase(ACCOAC),overexpressionofwhichmayincreasethemalonyl-CoApool.
ItcanbefurtherutilizedbyFAScomplextogeneratemorefattyacids,whichcanthenbechannelizedtoω-oxidationpathway.
Consideringthatfattyacidbiosyntheticpathwayisverytightlyregulated,transcriptomicdatacanprovidesomeusefulinsightintotheON/OFFstateofthereactionsinaparticularcondition.
ToidentifytheoverexpressiontargetsFig.
4Simulationresultbyfluxactivityanalysis.
Overexpressiongenesandproductionratechangesdependonthealterationoffluxactivitiesofrespectivegenes.
GAPD(Glyceraldehyde-3-phosphatedehydrogenase),PGK(Phosphoglyceratekinase),TPI(Triosephosphatekinase),ACCOAC(Acetyl-CoAcarboxylase),ALDDHDD(Aldehydedehydrogenase),DDCAFAO(Fattyacidoxidase),DDCAH(Fattyacidhydroxylase)Fig.
3ThecentralmetabolicnetworkofY.
lipolyticadepictingmetabolicengineeringtargetstoproduceDDDAMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page13of130forincreasingtheDCAsproduction,weimplementedtSOTbyresortingtotime-coursetranscriptomicprofileoftheY.
lipolytica,duringacontrolledfed-batchusingglu-coseasthesolecarbonsourceafter27htime-pointwhichcorrespondstoearlystationaryphase[40].
Anitrogenlimitationwasappliedduringthefed-batchtoinitiatedenovolipidsynthesiswhichcanrepresentthepreconditionforhighDCAsproduction.
ThebasicprincipleoftSOTistoascertainthegeneoverexpressiontargetsbyrestoringthereactionswhichareremovedfromGSMMbydata-integrationalgorithmswhiledevelopingacontext-specificmodel.
Asaresult,tSOTidentifiedMDH,bothcytosolicandmitochondrial,asanoverexpressiontarget.
Inaddition,italsofoundmitochondrialNAD-dependentmalicenzyme(ME1m)andglutamatedehydrogenase(GLUDy)tobetheoverexpressiontargets(Table1).
Interestingly,owingtothefactthatω-oxidationisanoxidativeprocesswiththehighdemandofredoxcofactors,alltheidentifiedreactionsareinvolvedincofactorregeneration.
MEishypothesizedtobethesupplierofNADPHduringlipidbiosynthesisinmostoleaginousyeaststhroughtheintracellularsubstratecyclesinvolvingMDH,pyruvatecarboxylase(PC)andME,alsocalled"transhydrogenasecycle"[41].
AlthoughY.
lipolyticalacksacytosoliccopyofMErequiredtocompensateforNADPHdemand,itcouldbeinterestingtoinvestigatethecompoundeffectofoverexpressingMDHandMEbecauseapartfromNADPH,MEinconjunctionwithmitochondrialpyruvatedehydrogenase(PDH)alsoprovidemitochondrialacetyl-CoA.
ThemostinterestingfindingfromtSOTisGLUDy,whichapartfromregenerationofNADPH;alsoplayskeyroleinmaintainingthebalanceofcarbonandnitrogen.
ThereisconsiderableevidenceofthepresenceofGLUDyshuntinplants,whichreturnsthecarboninaminoacidsbiosynthesisbackintoreactionsofcarbonmetabolismandTCAcycle[42]whichisthecaseinnitrogenstarvingcondition[24],makingitparticularlyaninterestingtargettoexploreasitlinksaminoacidbiosynthesistofattyacidmetabolism.
Downregulation/knockouttargetsGDLSalgorithmwasusedtosearchforgrowth-coupledsolutionsforDCAproduction,identifyingupto5reactiondeletioncandidates.
Basically,GDLSproposethepathwaydesignwhichcouplestheproductformationwiththecellgrowth,makingitsproductionnecessarytoreachoptimalgrowth.
ThestraindesignstrategydecipheredbyGDLScombinedthesimultaneousknockoutofDESAT16(Stear-oyl-CoAdesaturase)andMI1PS(Myo-inositol-1-phos-phatesynthase).
DESAT16istheenzymethatcatalyzestheconversionofsaturatedfattyacidtomonounsaturatedfattyacids.
TheoverexpressionofthisenzymeisshowntoincreaselipidaccumulationinY.
lipolyticawhichappar-entlytakesthefluxawayfromDCAsproductionwhichrequiresfreefattyacid[43],deletingwhichcanresultinincreasedpooloffreefattyacids.
SinceDCAsisanon-growthcoupledproduct,deletionofDESAT16andMI1PSassuggestedbyGDLSmaynotgiverisetoimprovedproductyield.
However,inthesimplenetworkperspective,DESAT16andMI1PSreactionsbranchthecarbonfluxawayfromDCAsformationwhichmakestheminterestingknockouttargetstotestexperimentally.
ItisworthnoticingherethatMI1PScatalyzesthefirstreactionintheinositolpathwaywhichproducesmembraneformingmetabolites,soknockingoutMI1PScouldshowdeleteri-ouseffectsoncellgrowth.
Nonetheless,sinceDCAsisnon-growthassociatedproduct,aninducibleknockoutstraincouldbeusedwhereinMI1PSissuppressedwhentheculturereachesstationaryphase.
Furthermore,formedDCAscangetdegradedviaβ-oxidationpathway.
Thereforeblockingitbydeletionofacyl-CoAoxidaseencodedbyPOX1–6genescanfurtherenhancethetiterofDCAs.
Y.
lipolyticawhichlacksacyl-CoAoxidasescanmoreefficientlyconvertn-alkanesandfattyacidsortheirderivativestotheircorrespondingDCAs[5].
CofactorspecificityengineeringtargetsSimilartootherbiosyntheticpathways,theDCAsbiosynthesisviaω-oxidationpathwayinY.
lipolyticainvolvesseveraluniquereactionsandiscommonlycontrolledbythesupplyofprecursorsandcofactors.
Earlierpartofthisstudyhasfocusedonoverexpressionanddownregulationofsomekeyenzymesregulatingtheω-oxidationpathway.
However,cofactorsareveryimportanttoachieveimprovementinproductivity.
NADPH,asareducingequivalent,usuallyplaysanim-portantroleincouplingcatabolismwithanabolismandenergygenerationduringmetabolism.
Severalmetabolicengineeringapproacheshavebeenimplementedtomanipulatethecofactorsleveltoincreasetheproductyieldinothermicroorganisms[44,45]Increasingtheα-santaleneproductionbymodifyingtheammoniumTable1OverexpressiontargetssimulatedbytSOTtoincreaseDCAsproductionTargetsReactionNameReactionDefinitionYieldImprovement(%)GLUDyGlutamateDehydrogenase(NADPH-forming)glu_L[c]+h2o[c]+nadp[c]akg[c]+h[c]+nadph[c]+nh4[c]22.
2MDHMalateDehydrogenase(cytosol)mal_L[c]+nad[c]h[c]+nadh[c]+oaa[c]47.
8MDHmMalateDehydrogenase(mitochondrial)mal_L[m]+nad[m]h[m]+nadh[m]+oaa[m]47.
8ME1mMalicenzyme(NAD-dependent)mal_L[m]+nad[m]->co2[m]+nadh[m]+pyr[m]47.
8Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page14of130assimilationfrombeingNADPHtoNADHdependentbythedeletionofGDH1andtheoverexpressionofGDH2;overexpressionofStreptococcusmutantsgapNgenewhichencodesaGAPDHtoincreasetheL-lysineproduction[46];engineeringNADPHregenerationforimprovingpentosefermentationbyoverexpressingtheGDP1,aNADP-dependentGAPDHfromKluyveromyceslactis[47];overexpressionoftranshydrogenaseandNADkinasetoimproveisobutanolproduction[48],areafewexamplesshowingtheemergenceofNADPHlevelengineeringaspotentandfeasiblestrategytoincreasetheproductioninmicrobialhosts.
Fattyacidbiosynthesisandω-oxidationpathwayareNADPHdemandingoxidativepathwayinY.
lipolytica.
Thisexcessivecofactordemandismainlysatisfiedthroughthepentosephosphatepathway(PPP)reactions.
But,theincreasedusageofthePPPthroughoverexpres-sionissuboptimalasonemoleofcarbonislostasCO2foreverytwomolesofNADPHproduced[49].
Tocir-cumventthis,weperformedCMAinordertoidentifythetargetsforcofactorspecificityengineeringtoim-provetheNADPHpoolwhichcanenhanceDDDAyield.
FromtheCMAresults,itwasobservedthattheDDDAyieldcanbeimprovedbyincreasingNADPHregener-ationthroughmodificationofcofactorspecificityfromNADtoNADP.
Amongthetargetsfound,changingthecofactorspecificityofGAPDandMDHfromNADtoNADPwasfoundtogivethebestimprovementinDDDAyield(Table2).
Toachieveahigh-levelproductionofDCAinyeast,weproposetoenhancethesupplyofNADPHtotheω-oxidationpathway.
Hence,amulti-stepmetabolicengin-eeringstrategycanbedevisedtosimultaneouslymodifytheglycolysisstep,optimizetheNADPHsupplyandtoenhancetheactivityofsomeNADPHproducingenzymes.
Effectofbranched-chainaminoacidssupplementationonDCAproductionInmostrecombinantDCAsproductionstudies,thehydrophobicsubstratessuchasalkaneandfattyacidmethylester(FAME)wereusedforthebiotransform-ation[5].
Assuch,glucoseisusedasthecarbonsourceforthegrowthbeforeinducingtheω-oxidationpathwayusingalkane.
However,ithasbeenshowninY.
lipolyticathatlipidaccumulationcanoccurdenovowithoutexogenoussupplyofhydrophobicsubstrate,layingthegroundsforapossibilitytobeusedfordenovoDCAsproductionaswell,fromprimarycarbonsources,i.
e.
,glucoseorglycerol[50].
Inadditiontothis,theabilitytoefficientlyproduceacetyl-CoAbymetabolizingothercarbonsourcesinthemediumwilldrivetheimportantprecursortowardsDCAsproduction.
Presently,thereareevidencesimplicatingastrongcorrelationbetweenlipidaccumulationandleucinemetabolisminS.
cerevi-siae[51]andY.
lipolytica[52].
TakingthecluefromlipidaccumulationstudiesonY.
lipolytica,wesoughttoexploretheeffectofBCAAsupplementationonacetyl-CoApool,whichcanbeproducedviadegradationpathwayofBCAA.
Inordertotheoreticallyanalyzethefluxdistributionandflux-sumchangesthatoccurwhenBCAAissupplemented,wemaximizedtheDDDAproductionusingglucoseastheprimarycarbonsourcesupplementedwith10C-mmol/gDCW-hrofleucine,isoleucine,orvaline.
Toelucidatethedifferenceinpath-wayutilizationandmetaboliteturn-overindifferentaminoacidsduringDDDAproduction,wepreparedtheheatmapoffluxdistributionandflux-sumasshowninFig.
5.
Itcanbeseenthatacetyl-CoAbeingcriticalde-terminantofDCAssynthesiswasproducedatthehigh-estlevelonvalinesupplementationfollowedbyleucinewhichcorrelatedwithDDDAturnovertrend.
ApartfromtheroleofBCAAsinacetyl-CoAproduction,supplementingaminoacidsdecreasestheprimarycarbondemandinaminoacidbiosynthesisforbiomassformation.
ThisextracarboncanthenbedivertedtoDCAsbiosynthesis.
Thisobservationisincloseresem-blancewithnitrogenstarvationconditionforlipidaccu-mulationbecauseinnitrogenlimitingcondition,biosynthesisofaminoacidseizesandcarbonpresentthencanbeutilizedinotherbiosyntheticpathways.
Fromoursimulation,wehaveidentifiedthatglucosesupplementedwithvalineorleucinecanincreasethegrowthaswellasdenovoaccumulationofDCAsinY.
lipolytica.
Thisisthefirstreportimplicatingaprominentroleofvalinedegradationasanalternaterouteofacetyl-CoAbiosynthesisinY.
lipolytica,experimentalvalidationofwhichcangiveinsighttointricatecorrelationbetweenfattyacidbiosynthesisandaminoaciddegradation.
ConclusionTheGSMMofY.
lipolytica,iYLI647,wasreconstructedusingiMK735asascaffoldandtherelevantinformationfromotheravailableY.
lipolyticamodels.
iYLI647con-sistsof1347reactionsand647genes.
ThebiomassTable2CofactorspecificityengineeringtargetsbyCMAtoincreaseDCAsproductionReactionNameReactionDefinitionYieldofDDDA(mmol/gDCW-hr)NADNADPMDHmal_L[c]+nad[c]h[c]+nadh[c]+oaa[c]2.
6012.
781GAPDg3p[c]+nad[c]+pi[c]13dpg[c]+h[c]+nadh[c]2.
6012.
778Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page15of130equationswerecarefullyformulatedwithvariousexperi-mentalinformationofY.
lipolytica,whichisperhapsthereasonforaccuratemodelprediction.
ThepotentialofdenovoDCAsproductioninY.
lipolyticacombinedwithamodel–drivenstraindesignformetabolicengin-eeringandmediaoptimizationstrategieswerethenevaluatedusingthereconstructediYLI647.
ThefluxtowardsDDDAproductionwasincreasedfollowingtheoverexpressionanddeletionoffewreactionsandthemodel-basedstraindesigngivesusagoodstartingpointtoexplorethemetaboliccapabilitiesofY.
lipolyticatoproducefattyacidderivedproducts.
Moreover,theworkflowandprocedurepresentedinthisanalysiscanbeutilizedasaplatformtoperformsimilaranalyseswithdifferentorganisms.
MethodsMetabolicnetworkreconstructionWechoseallfourseparatelydevelopedpubliclyavailablemetabolicnetworkreconstructions,iNL895,iYL619,iMK735,andiYali4,tocompareandselectthebestsuitedasscaffoldmodel.
Followingourqualitativeandquantitativecomparison,wechoseiMK735asascaffoldmodelandproceededtomanualcurationtoexpandthemodelcoverageandcharacteristics.
Todothis,usingliteraturedata,weverifiedthepresenceofreactionsandrelevanceinY.
lipolyticametabolism.
Then,weaddedω-oxidationpathwaytoconvertfattyacidstoDCAs,andalsothesubsequentdegradingβ-oxidationpathway.
Followingtheliteratureevidenceanditsestablishedim-portanceinfattyacidmetabolism,wealsoincludedBCAAdegradationpathways.
Furthermore,wecheckedthemassbalanceofreactionsandmadeappropriatechangestomakestoichiometricallybalancedreactions.
Inaddition,wederivedthebiomassequationincarbonandnitrogenlimitedconditionsusingupdatedandrele-vantliteraturesources.
Usingrefinedbiomassequation,wesimulatedthemodeltoidentifytheloops,andmissinglinkreactionsusingGapFinder[53].
LoopswereremovedbychangingthedirectionalityofcoupledFig.
5Fluxdistributionandflux-sum.
Heatmapshowingfluxdistributionandflux-sumofimportantreactionsandmetabolites,respectively,duringDDDAproductionMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page16of130reaction,orremovingthenon-metaboliclumpedreactions,whereas,gapswerefilledwithreactionsfromorthologsorintroducingtransport/sinkreaction.
Constraints-basedfluxanalysisThecellularmetabolismofY.
lipolyticawassimulatedundervaryingenvironmentalconditionsusingconstraints-basedfluxanalysis.
Thebiomassreactionwasmaximizedtosimulatethegrowthundervariouscultureconditionsasdescribedelsewhere[54–56].
Themaximizationofbiomassissubjectedtostoichiometricandcapacityconstraints,whichcanmathematicallybeformulatedas:maxZ1Xjcjvj1s:t:XjSijvj0metaboliteivminj≤vj≤vmaxjreactionjwhereSijreferstothestoichiometriccoefficientofmetaboliteiinvolvedinreactionj,vjdenotestothefluxorspecificrateofmetabolicreactionj,vminjandvmaxjrep-resentthelowerandupperlimitsonthefluxofreactionj,respectively;andZ1correspondstothecellularobject-iveasalinearfunctionofallthemetabolicreactionswheretherelativeweightsaredeterminedbythecoeffi-cientcj.
Inthisstudy,theconstraints-basedfluxanalysisproblemsweresolvedusingCOBRAtoolbox[57].
Insilicomodel-basedstraindesignFourinsilicostraindesignapproacheshavebeenemployedtoidentifymetabolicengineeringtargetsforoverproducingDCAfromglucose.
Geneticdesignbylocalsearch(GDLS)GDLSstraindesignalgorithm[33]wasimplementedinCOBRAtoolboxinMatlab,searchinguptomaximumof5knockoutreactionswiththeouterobjectiveofmaximizingDDDAexchangeflux.
FluxactivityanalysisToidentifytheupregulationanddownregulationgene/reactiontargetswhichleadtotheenhancedproductionofdesiredcompound,wefirstneededtoquantifythefluxactivityofallreactionsinthewild-typestrain.
Fluxactivity,fj,isdefinedastheabsolutevalueofreactionflux,vj[34].
Thiscanbedeterminedbyfirstsolvingtheconstraints-basedfluxanalysisproblem(Eq.
1)withbio-massmaximizationasobjective,andthenobtainingtheabsolutevaluesofindividualreactionfluxes.
Next,wesolvedthebelow-mentionedmixed-integerlinearpro-gramming(MILP)problemtoidentifythemaximumandminimumfluxactivitiestodeterminethefeasiblerangesofindividualreactionssuchthattheycanbeup-regulatedanddownregulatedwithinthislimit:max=minfj2Subjectto:XjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1Where,αjandβjareupperandlowerboundsoffluxes,respectively.
Thefj+andfjarethepositivetwovariablesintowhichtheflux,vj,isdecomposed.
Itisobservedthatfjfjjfjfjj,ifandonlyifeitherfj+orfjisequaltozero.
ThisconditionwasintroducedbynewbinaryvariablesIj+andIjwhicharewhenmultipliedwithalargeintegerwhichshouldbelargerthanthefluxofex-perimentallymeasuredvalue,M,willbealwayslesserthanfj+andfj,respectively.
Additionally,theconstraintIjIj1isalsointroducedtoensurethateitherfj+orfjareequaltozero.
Oncethereferencefluxactivitiesareestablished,i.
e.
wild-type,maximumandminimumvalues,wethensolvethebelowmentionedMILPproblemtoanalyzetheeffectsofperturbingaparticularfluxactivityoncellulargrowth.
maxvbiomass3Subjectto:XjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1C1:fjfj≤fminjkattfWTjfminjMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page17of130ORC2:fjfj≥fWTjkintfmaxjfWTjWhere,constraints(C1)and(C2)areapplicableforupregulationanddownregulationproblems,respectively.
Parameterskattandkintaregraduallyvariedbetween0and1instepsof0.
1toanalyzetheeffectofreactionupregulationbetweenminimalandwild-typevalues,andreactiondownregulationbetweenthewild-typeandmaximalvalues,respectively.
Finally,theobjectivevalueobtainedfromthesolutionof(Eq.
3)isusedasthelowerlimitforcellgrowthinthefourthstepwhereby(Eq.
3)issolvedagainwiththetargetedproductastheobjectivefunction.
Thecorre-spondingmathematicalformulationisasfollows:minvEXsuccSubjectto:vbiomass≥Bj;kXjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1C1:fjfj≤fminjkattfWTjfminjORC2:fjfj≥fWTjkintfmaxjfWTjwhereBj,kisthemaximumbiomassobtainablewhilesolvingproblem(Eq.
3)forjthreactionatkthupregula-tion/downregulationlevels.
Alltheoptimizationprob-lemsweresolvedusingtheGAMSIDEsoftwareversion22.
4withIBMILOGCPLEXsolver.
Transcriptomics-basedstrainoptimizationtool(tSOT)SinceoptimalityassumptionbasedFBAalgorithmsignorestheregulatoryconsiderationwhilestraindesigning,weimplementedtranscriptomics-basedstrainoptimizationtool(tSOT)[35]toidentifymetabolicengineeringtargetsbasedontranscriptomicdataintegratedinthemodel.
Acomprehensivetime-coursetranscriptomicprofilefromthecultureofY.
lipolytica,duringacontrolledfed-batchonglucose,wasusedastranscriptomicdataobtainedfromGeneExpressionOmnibus(GEO)databaseusingaccessionnumberGSE29046[40].
Cofactormodificationanalysis(CMA)Sincetheω-oxidationpathwayisanoxidativeprocessandrequirescofactoroptimizationtomaximizetheoreticalyield,weusedCMAtoidentifythecofactorspecificityengineeringtargetwhichcanincreasetheyieldofDDDA.
CMAwasimplementedasdescribedinourpreviouswork[36].
Mathematically,thebi-levelmixed-integernonlinearprogramming(MINLP)optimizationproblemspecifictotheCMAcanberepresentedasfollows:maxφproduct0:5XjSproduct;jvjs:t:maxvbiomasss:t:XjSijvjScModijvcModj0metaboliteivbiomass≥vminbiomass1ycModjvminj≤vj≤1ycModjvmaxjycModjvminj≤vcModj≤ycModjvmaxjycModj0;1fgreactionj266666666664377777777775XjycModj≤kWhere,ScModijisthecofactormodifiedstoichiometricmatrixwherethecoefficientsaresameasSij,exceptthere-actionswhichinvolveeitherNAD(H)orNADP(H).
ThesereactionsareswappedforcofactorsintheScModijmatrixsuchthatSNADH;jScModNADPH;jandSNADPH;jScModNADH;j.
vcModjisthefluxthroughthecofactormodifiedreactionandvminbiomassistheminimumamountofbiomassthatneedstobeproduced.
ThebinaryvariableycModjensuresthatthecofactorassociatedreactionsareallowedtocarryfluxeitherwithitsoriginalorswappedcofactorbutnotboth.
Thenumberofcofactorswitchesallowedinaparticularsimula-tioniscontrolledbythenumber,kandwhichisfixedat1forallsimulationsinthiswork.
Thebi-levelMINLPproblemwasreformulatedasasingle-levelMINLPprob-lemusingtheprimaldualtransformationasimplementedearlier[58].
TheMILPoptimizationproblemwassolvedusingtheGAMSIDEsoftwareversion22.
4withIBMILOGCPLEXsolver.
Flux-sumIntheconstraints-basedfluxanalysis,thereisnoaccu-mulationofintermediatemetabolitesduetothesteady-statecondition.
However,theturnoverrate,whichisalsoequivalenttothetotalconsumptionorproductionrate,oftheintermediatescanbenonzero,whichisdefinedasMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page18of130theirflux-sum[59].
Sincetheoverallconsumptionandgenerationratesareequalunderthesteady-stateas-sumption,theflux-sumofmetaboliteicanbeformu-latedasΦi=0.
5∑|Sijvj|.
Eachterminthissummationseriesgivesustheabsoluterateofconsumption/gener-ationofmetaboliteiduetoreactionjandthusbyhalv-ingthesumoftheseterms,wecanobtaintheoverallturnoverrateformetabolitei.
AdditionalfilesAdditionalfile1:Changed,addedanddeletedreactionsiniYLI647modelincomparisonwithiMK735scaffoldmodel.
(XLSX22kb)Additionalfile2:SMBLfileofiYLI647.
(XML2312kb)Additionalfile3:BiomasscompositionofY.
lipolyticainC-andN-limitedconditionsandGAMandNGAMcalculations.
(DOCX52kb)AcknowledgementsWethankDr.
Jung-OhAhn(KoreaResearchInstituteofBioscienceandBio-technology)forhelpanddiscussion.
FundingThisworkwassupportedbytheAcademicResearchFund(R-279-000-476-112)oftheNationalUniversityofSingapore,BiomedicalResearchCouncilofA*STAR(AgencyforScience,TechnologyandResearch),Singapore,theGlobalR&Dprojectprogram(N011500017),MinistryofTrade,IndustryandEnergy(MOTIE),RepublicofKoreaandtheNext-GenerationBioGreen21Pro-gramoftheRuralDevelopmentAdministration,RepublicofKorea(SystemsandSyntheticAgrobiotechCenter;grantno.
PJ01334605).
PublicationcostwascoveredbyNUSSynCTIprogram.
AvailabilityofdataandmaterialsAlldatageneratedoranalyzedduringthisstudyareincludedinthispublishedarticleanditssupplementaryinformationfiles.
AboutthissupplementThisarticlehasbeenpublishedaspartofBMCSystemsBiologyVolume12Supplement2,2018:Proceedingsofthe28thInternationalConferenceonGenomeInformatics:systemsbiology.
Thefullcontentsofthesupplementareavailableonlineathttps://bmcsystbiol.
biomedcentral.
com/articles/supplements/volume-12-supplement-2.
Authors'contributionsPM,NRLandDYLconceivedanddesignedtheinsilicostudy.
PM,NRL,MLandMSKperformedthecomputationalsimulationsanddraftedthemanuscript.
PM,NRL,BGKandDYLrevisedthemanuscript.
Allauthorsreadandapprovedthefinalmanuscript.
EthicsapprovalandconsenttoparticipateNotapplicable.
ConsentforpublicationNotapplicable.
CompetinginterestsTheauthorsdeclarethattheyhavenocompetinginterests.
Publisher'sNoteSpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
Authordetails1NUSSyntheticBiologyforClinicalandTechnologicalInnovation(SynCTI),LifeSciencesInstitute,NationalUniversityofSingapore,28MedicalDrive,Singapore117456,Singapore.
2BioprocessingTechnologyInstitute,AgencyforScience,TechnologyandResearch(A*STAR),20BiopolisWay,#06-01,Centros,Singapore138668,Singapore.
3SchoolofChemicalandBiologicalEngineering,InstituteofMolecularBiologyandGenetics,andBioengineeringInstitute,SeoulNationalUniversity,1Gwanak-ro,Gwanak-gu,Seoul151-742,RepublicofKorea.
4SchoolofChemicalEngineering,SungkyunkwanUniversity,2066Seobu-ro,Jangan-gu,Suwon,Gyeonggi-do16419,RepublicofKorea.
Published:19March2018References1.
CornilsB,LappeP.
Dicarboxylicacids,Aliphatic.
Ullmann'sEncyclIndChem.
2012;11:287–304.
https://doi.
org/10.
1002/14356007.
a08.
2.
PolenT,SpelbergM,BottM.
Towardbiotechnologicalproductionofadipicacidandprecursorsfrombiorenewables.
JBiotechnol.
2013:75–84.
https://doi.
org/10.
1016/j.
jbiotec.
2012.
07.
008.
3.
GreenKD,TurnerMK,WoodleyJM.
CandidaCloacaeoxidationoflong-chainfattyacidstodioicacids.
EnzymMicrobTechnol.
2000;27:205–11.
https://doi.
org/10.
1016/S0141-0229(00)00217-9.
4.
JiaoP,MaS,HuaY,HuangY,CaoZ.
IsolationandenzymedeterminationofCandidaTropicalismutantsforDCAproduction.
JGenApplMicrobiol.
2000;46:245–9.
https://doi.
org/10.
2323/jgam.
46.
245.
5.
SmitMS,MokgoroMM,SetatiE,NicaudJ-M.
Alpha,omega-dicarboxylicacidaccumulationbyacyl-CoAoxidasedeficientmutantsofYarrowialipolytica.
BiotechnolLett.
2005;27:859–64.
https://doi.
org/10.
1007/s10529-005-6719-1.
6.
ChanEC,ChengCS,HsuYH.
ContinuousproductionofdicarboxylicacidbyimmobilizedPseudomonasAeruginosacells.
JFermentBioeng.
1997;83:157–60.
https://doi.
org/10.
1016/S0922-338X(97)83575-1.
7.
ChanEC,KuoJ.
BiotransformationofdicarboxylicacidbyimmobilizedCryptococcuscells.
EnzymMicrobTechnol.
1997;20:585–9.
https://doi.
org/10.
1016/S0141-0229(96)00198-6.
8.
Sathesh-PrabuC,LeeSK.
Productionoflong-chainalpha,omega-dicarboxylicacidsbyengineeredEscherichiaColifromrenewablefattyacidsandplantoils.
JAgricFoodChem.
2015;63:8199–208.
https://doi.
org/10.
1021/acs.
jafc.
5b03833.
9.
Ledesma-AmaroR,NicaudJM.
Yarrowialipolyticaasabiotechnologicalchassistoproduceusualandunusualfattyacids.
ProgLipidRes.
2016;61:40–50.
https://doi.
org/10.
1016/j.
plipres.
2015.
12.
001.
10.
EschenfeldtWH,ZhangY,SamahaH,StolsL,EirichLD,WilsonCR,etal.
TransformationoffattyacidscatalyzedbycytochromeP450monooxygenaseenzymesofCandidaTropicalis.
ApplEnvironMicrobiol.
2003;69:5992–9.
https://doi.
org/10.
1128/AEM.
69.
10.
5992-5999.
2003.
11.
GatterM,FrsterA,BrK,WinterM,OttoC,PetzschP,etal.
Anewlyidentifiedfattyalcoholoxidasegeneismainlyresponsiblefortheoxidationoflong-chainω-hydroxyfattyacidsinYarrowialipolytica.
FEMSYeastRes.
2014;14:858–72.
https://doi.
org/10.
1111/1567-1364.
12176.
12.
EirichLD,CraftDL,SteinbergL,AsifA,EschenfeldtWU,StolsL,etal.
CloningandcharacterizationofthreefattyalcoholoxidasegenesfromCandidaTropicalisstrainATCC20336.
ApplEnvironMicrobiol.
2004;70:4872–9.
https://doi.
org/10.
1128/AEM.
70.
8.
4872-4879.
2004.
13.
IwamaR,KobayashiS,IshimaruC,OhtaA,HoriuchiH,FukudaR.
FunctionalrolesandsubstratespecificitiesoftwelvecytochromesP450belongingtoCYP52familyinn-alkaneassimilatingyeastYarrowialipolytica.
FungalGenetBiol.
2016;91:43–54.
https://doi.
org/10.
1016/j.
fgb.
2016.
03.
007.
14.
BroadwayNM,DickinsonFM,RatledgeC.
Theenzymologyofdicarboxylic-acidformationbyCorynebacteriumSpstrain7E1CgrownonN-alkanes.
JGenMicrobiol.
1993;139:1337–44.
https://doi.
org/10.
1099/00221287-139-6-1337.
15.
NicaudJ,ThevenieauF,LeDallM,MarchalR.
Productionofdicarboxylicacidsbyimprovedmutantstrainsofyarrowialipolytica.
USpatapp11/;2005.
p.
1.
Available:http://www.
google.
com/patents/US2010004111516.
SimeonidisE,PriceND.
Genome-scalemodelingformetabolicengineering.
JIndMicrobiolBiotechnol.
2015:327–38.
https://doi.
org/10.
1007/s10295-0e14-1576-3.
17.
AlperH,JinYS,MoxleyJF,StephanopoulosG.
IdentifyinggenetargetsforthemetabolicengineeringoflycopenebiosynthesisinEscherichiaColi.
MetabEng.
2005;7:155–64.
https://doi.
org/10.
1016/j.
ymben.
2004.
12.
003.
18.
AlperH,MiyaokuK,StephanopoulosG.
Constructionoflycopene-overproducingE.
Colistrainsbycombiningsystematicandcombinatorialgeneknockouttargets.
NatBiotechnol.
2005;23:612–6.
https://doi.
org/10.
1038/nbt1083.
19.
BroC,RegenbergB,FrsterJ,NielsenJ.
InsilicoaidedmetabolicengineeringofSaccharomycesCerevisiaeforimprovedbioethanolproduction.
MetabEng.
2006;8:102–11.
https://doi.
org/10.
1016/j.
ymben.
2005.
09.
007.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page19of13020.
SunZ,MengH,LiJ,WangJ,LiQ,WangY,etal.
IdentificationofnovelknockouttargetsforimprovingterpenoidsbiosynthesisinSaccharomycesCerevisiae.
PLoSOne.
2014;9https://doi.
org/10.
1371/journal.
pone.
0112615.
21.
LoiraN,DulermoT,NicaudJ-M,ShermanDJ.
Agenome-scalemetabolicmodelofthelipid-accumulatingyeastYarrowialipolytica.
BMCSystBiol.
BioMedCentralLtd.
2012;6:35.
https://doi.
org/10.
1186/1752-0509-6-35.
22.
PanP,HuaQ.
Reconstructionandinsilicoanalysisofmetabolicnetworkforanoleaginousyeast,Yarrowialipolytica.
PLoSOne.
2012;7:e51535.
https://doi.
org/10.
1371/journal.
pone.
0051535.
23.
KavekM,BhutadaG,MadlT,NatterK,KrainerF,DietzschC,etal.
Optimizationoflipidproductionwithagenome-scalemodelofYarrowialipolytica.
BMCSystBiol.
2015;9:72.
https://doi.
org/10.
1186/s12918-015-0217-4.
24.
KerkhovenEJ,PomraningKR,BakerSE,NielsenJ.
Regulationofamino-acidmetabolismcontrolsfluxtolipidaccumulationinYarrowialipolytica.
Npj.
SystBiolAppl.
2016;2:16005.
https://doi.
org/10.
1038/npjsba.
2016.
5.
25.
DuarteNC,HerrgrdMJ,PalssonB.
ReconstructionandvalidationofSaccharomycesCerevisiaeiND750,afullycompartmentalizedgenome-scalemetabolicmodel.
GenomeRes.
2004;14:1298–309.
https://doi.
org/10.
1101/gr.
2250904.
26.
BlazeckJ,LiuL,KnightR,AlperHS.
HeterologousproductionofpentaneintheoleaginousyeastYarrowialipolytica.
JBiotechnol.
2013;165:184–94.
https://doi.
org/10.
1016/j.
jbiotec.
2013.
04.
003.
27.
VorapreedaT,ThammarongthamC,CheevadhanarakS,LaotengK.
Alternativeroutesofacetyl-CoAsynthesisidentifiedbycomparativegenomicanalysis:involvementinthelipidproductionofoleaginousyeastandfungi.
Microbiology.
2012;158:217–28.
https://doi.
org/10.
1099/mic.
0.
051946-0.
28.
VanMarisAJA,LuttikMAH,WinklerAA,VanDijkenJP,PronkJT.
Overproductionofthreoninealdolasecircumventsthebiosyntheticroleofpyruvatedecarboxylaseinglucose-limitedchemostatculturesofSaccharomycesCerevisiae.
ApplEnvironMicrobiol.
2003;69:2094–9.
https://doi.
org/10.
1128/AEM.
69.
4.
2094-2099.
2003.
29.
PapanikolaouS,AggelisG.
LipidproductionbyYarrowialipolyticagrowingonindustrialglycerolinasinglestagecontinuousculture.
BioresourTechnol.
2002;82:43–9.
https://doi.
org/10.
1186/s13068-015-0286-z.
30.
WorkmanM,HoltP,ThykaerJ.
ComparingcellularperformanceofYarrowialipolyticaduringgrowthonglucoseandglycerolinsubmergedcultivations.
AMBExpress.
2013;3:58.
https://doi.
org/10.
1186/2191-0855-3-58.
31.
DulermoT,LazarZ,DulermoR,RakickaM,HaddoucheR,NicaudJM.
AnalysisofATP-citratelyaseandmalicenzymemutantsofYarrowialipolyticapointsouttheimportanceofmannitolmetabolisminfattyacidsynthesis.
BiochimBiophysActa-MolCellBiolLipids.
ElsevierB.
V.
2015;1851:1107–17.
https://doi.
org/10.
1016/j.
bbalip.
2015.
04.
007.
32.
MostafaNa,MaherA,AbdelmoezW.
Productionofmono-,di-,andtriglyceridesfromwastefattyacidsthroughesterificationwithglycerol.
AdvBiosciBiotechnol.
2013;4:900–7.
http://dx.
doi.
org/10.
4236/abb.
2013.
4911833.
LunDS,RockwellG,GuidoNJ,BaymM,KelnerJA,BergerB,etal.
Large-scaleidentificationofgeneticdesignstrategiesusinglocalsearch.
MolSystBiol.
2009;5:296.
https://doi.
org/10.
1038/msb.
2009.
57.
34.
ChungBK,SelvarasuS,AndreaC,RyuJ,LeeH,AhnJ,etal.
Genome-scalemetabolicreconstructionandinsilicoanalysisofmethylotrophicyeastPichiaPastorisforstrainimprovement.
MicrobCellFactories.
2010;9:50.
https://doi.
org/10.
1186/1475-2859-9-50.
35.
KimM,YiJS,LakshmananM,LeeDY,KimBG.
Transcriptomics-basedstrainoptimizationtoolfordesigningsecondarymetaboliteoverproducingstrainsofStreptomycesCoelicolor.
BiotechnolBioeng.
2016;113:651–60.
https://doi.
org/10.
1002/bit.
25830.
36.
LakshmananM,ChungBK-S,LiuC,KimS-W,LeeD-Y.
Cofactormodificationanalysis:acomputationalframeworktoidentifycofactorspecificityengineeringtargetsforstrainimprovement.
JBioinformaComputBiol.
2013;11:1343006.
https://doi.
org/10.
1142/S0219720013430063.
37.
ZaslaverA,MayoAE,RosenbergR,BashkinP,SberroH,TsalyukM,etal.
Just-in-timetranscriptionprograminmetabolicpathways.
NatGenet.
2004;36:486–91.
https://doi.
org/10.
1038/ng1348.
38.
LittleJW,ShepleyDP,WertDW.
Robustnessofageneregulatorycircuit.
EMBOJ.
1999;18:4299–307.
https://doi.
org/10.
1093/emboj/18.
15.
4299.
39.
IwamaR,KobayashiS,OhtaA,HoriuchiH,FukudaR.
AlcoholdehydrogenasesandanalcoholoxidaseinvolvedintheassimilationofexogenousfattyalcoholsinYarrowialipolytica.
FEMSYeastRes.
2015;15:1–12.
https://doi.
org/10.
1093/femsyr/fov014.
40.
MorinN,CescutJ,BeopoulosA,LelandaisG,LeBerreV,UribelarreaJ-L,etal.
TranscriptomicanalysesduringthetransitionfrombiomassproductiontolipidaccumulationintheoleaginousyeastYarrowialipolytica.
PLoSOne.
2011;6:e27966.
https://doi.
org/10.
1371/journal.
pone.
0027966.
41.
RatledgeC.
TheroleofmalicenzymeastheproviderofNADPHinoleaginousmicroorganisms:areappraisalandunsolvedproblems.
BiotechnolLett.
2014;36:1557–68.
https://doi.
org/10.
1007/s10529-014-1532-3.
42.
MiflinBJ,HabashDZ.
Theroleofglutaminesynthetaseandglutamatedehydrogenaseinnitrogenassimilationandpossibilitiesforimprovementinthenitrogenutilizationofcrops.
JExpBot.
2002;53:979–87.
https://doi.
org/10.
1093/jexbot/53.
370.
979.
43.
QiaoK,ImamAbidiSH,LiuH,ZhangH,ChakrabortyS,WatsonN,etal.
EngineeringlipidoverproductionintheoleaginousyeastYarrowialipolytica.
MetabEng.
2015;29:56–65.
https://doi.
org/10.
1016/j.
ymben.
2015.
02.
005.
44.
GaoC,ZhangL,XieY,HuC,ZhangY,LiL,etal.
Productionof(3S)-acetoinfromdiacetylbyusingstereoselectiveNADPH-dependentcarbonylreductaseandglucosedehydrogenase.
BioresourTechnol.
2013;137:111–5.
https://doi.
org/10.
1016/j.
biortech.
2013.
02.
115.
45.
SunJA,ZhangLY,RaoB,ShenYL,WeiDZ.
EnhancedacetoinproductionbySerratiaMarcescensH32withexpressionofawater-formingNADHoxidase.
BioresourTechnol.
2012;119:94–8.
https://doi.
org/10.
1016/j.
biortech.
2012.
05.
108.
46.
TakenoS,MurataR,KobayashiR,MitsuhashiS,IkedaM.
EngineeringofCorynebacteriumglutamicumwithanNADPH-generatingglycolyticpathwayforL-lysineproduction.
ApplEnvironMicrobiol.
2010;76:7154–60.
https://doi.
org/10.
1128/AEM.
01464-10.
47.
VerhoR,LondesboroughJ,PenttilM,RichardP.
EngineeringredoxcofactorregenerationforimprovedpentosefermentationinSaccharomycesCerevisiae.
ApplEnvironMicrobiol.
2003;69:5892–7.
https://doi.
org/10.
1128/AEM.
69.
10.
5892-5897.
2003.
48.
ShiA,ZhuX,LuJ,ZhangX,MaY.
ActivatingtranshydrogenaseandNADkinaseincombinationforimprovingisobutanolproduction.
MetabEng.
2013;16:1–10.
https://doi.
org/10.
1016/j.
ymben.
2012.
11.
008.
49.
WasylenkoTM,AhnWS,StephanopoulosG.
TheoxidativepentosephosphatepathwayistheprimarysourceofNADPHforlipidoverproductionfromglucoseinYarrowialipolytica.
MetabEng.
2015;30:27–39.
https://doi.
org/10.
1016/j.
ymben.
2015.
02.
007.
50.
SeipJ,JacksonR,HeH,ZhuQ,HongSP.
Snf1isaregulatoroflipidaccumulationinYarrowialipolytica.
ApplEnvironMicrobiol.
2013;79:7360–70.
https://doi.
org/10.
1128/AEM.
02079-13.
51.
KamisakaY,TomitaN,KimuraK,KainouK,UemuraH.
DGA1(diacylglycerolacyltransferasegene)overexpressionandleucinebiosynthesissignificantlyincreaselipidaccumulationintheΔsnf2disruptantofSaccharomycesCerevisiae.
BiochemJ.
2007;408:61.
https://doi.
org/10.
1042/bj20070449.
52.
BlazeckJ,HillA,LiuL,KnightR,MillerJ,PanA,etal.
HarnessingYarrowialipolyticalipogenesistocreateaplatformforlipidandbiofuelproduction.
NatCommun.
NaturePublishingGroup.
2014;5:3131.
https://doi.
org/10.
1038/ncomms4131.
53.
SatishKumarV,DasikaMS,MaranasCD.
Optimizationbasedautomatedcurationofmetabolicreconstructions.
BMCBioinformatics.
2007;8:212.
https://doi.
org/10.
1186/1471-2105-8-212.
54.
OrthJD,ThieleI,PalssonB.
WhatisfluxbalanceanalysisNatBiotechnol.
2010;28:245–8.
https://doi.
org/10.
1038/nbt.
1614.
55.
OberhardtMA,PalssonB,PapinJA.
Applicationsofgenome-scalemetabolicreconstructions.
MolSystBiol.
2009;5:320.
https://doi.
org/10.
1038/msb.
2009.
77.
56.
LeeJM,GianchandaniEP,PapinJA.
Fluxbalanceanalysisintheeraofmetabolomics.
BriefBioinform.
2006:140–50.
https://doi.
org/10.
1093/bib/bbl007.
57.
SchellenbergerJ,QueR,FlemingRMT,ThieleI,OrthJD,FeistAM,etal.
Quantitativepredictionofcellularmetabolismwithconstraint-basedmodels:theCOBRAtoolboxv2.
0.
NatProtoc.
2011;6:1290–307.
https://doi.
org/10.
1038/nprot.
2011.
308.
58.
LakshmananM,YuK,KoduruL,LeeD-Y.
Insilicomodel-drivencofactorengineeringstrategiesforimprovingtheoverallNADP(H)turnoverinmicrobialcellfactories:JIndMicrobiolBiotechnol.
SpringerBerlinHeidelberg;2015.
https://doi.
org/10.
1007/s10295-015-1663-0.
59.
KimP-J,LeeD-Y,KimTY,LeeKH,JeongH,LeeSY,etal.
MetaboliteessentialityelucidatesrobustnessofEscherichiaColimetabolism.
ProcNatlAcadSciUSA.
2007;104:13638–42.
https://doi.
org/10.
1073/pnas.
0703262104.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page20of130
31October-3November2017AbstractBackground:Recently,therehavebeenseveralattemptstoproducelong-chaindicarboxylicacids(DCAs)invariousmicrobialhosts.
Ofthese,Yarrowialipolyticahasgreatpotentialduetoitsoleaginouscharacteristicsanduniqueabilitytoutilizehydrophobicsubstrates.
However,Y.
lipolyticashouldbefurtherengineeredtomakeitmorecompetitive:thecurrentapproachesaremostlyintuitiveandcumbersome,thuslimitingitsindustrialapplication.
Results:Inthisstudy,weproposedmodel-guidedmetabolicengineeringstrategiesforenhancedproductionofDCAsinY.
lipolytica.
Attheoutset,wereconstructedgenome-scalemetabolicmodel(GSMM)ofY.
lipolytica(iYLI647)bysubstantiallyexpandingthepreviousmodels.
Subsequently,themodelwasvalidatedusingthreesetsofpublishedcultureexperimentdata.
Itwasfinallyexploitedtoidentifygeneticengineeringtargetsforoverexpression,knockout,andcofactormodificationbyapplyingseveralinsilicostraindesignmethods,whichpotentiallygiverisetohighyieldproductionoftheindustriallyrelevantlong-chainDCAs,e.
g.
,dodecanedioicacid(DDDA).
Theresultanttargetsinclude(1)malatedehydrogenaseandmalicenzymegenesand(2)glutamatedehydrogenasegene,insilicooverexpressionofwhichgeneratedadditionalNADPHrequiredforfattyacidsynthesis,leadingtotheincreasedDDDAfluxesby48%and22%higher,respectively,comparedtowild-type.
Wefurtherinvestigatedtheeffectofsupplyingbranched-chainaminoacidsontheacetyl-CoAturn-overratewhichiskeymetaboliteforfattyacidsynthesis,suggestingtheirsignificanceforproductionofDDDAinY.
lipolytica.
Conclusion:Insilicomodel-basedstraindesignstrategiesallowedustoidentifyseveralmetabolicengineeringtargetsforoverproducingDCAsinlipidaccumulatingyeast,Y.
lipolytica.
Thus,thecurrentstudycanprovideamethodologicalframeworkthatisapplicabletootheroleaginousyeastsforvalue-addedbiochemicalproduction.
Keywords:Yarrowialipolytica,Dicarboxylicacid,Genome-scalemetabolicmodels,Straindesign,Metabolicengineering*Correspondence:dongyuplee@skku.
eduEqualcontributors1NUSSyntheticBiologyforClinicalandTechnologicalInnovation(SynCTI),LifeSciencesInstitute,NationalUniversityofSingapore,28MedicalDrive,Singapore117456,Singapore2BioprocessingTechnologyInstitute,AgencyforScience,TechnologyandResearch(A*STAR),20BiopolisWay,#06-01,Centros,Singapore138668,SingaporeFulllistofauthorinformationisavailableattheendofthearticleTheAuthor(s).
2018OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.
0InternationalLicense(http://creativecommons.
org/licenses/by/4.
0/),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.
TheCreativeCommonsPublicDomainDedicationwaiver(http://creativecommons.
org/publicdomain/zero/1.
0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12https://doi.
org/10.
1186/s12918-018-0542-5BackgroundLong-chaindicarboxylicacids(DCAs)arewidelyusedinthemanufacturingofpolyamidesandpolyestersasthemonomericintermediates[1].
ThemostcommonlyemployedchemicalprocesstoproduceDCAsisthering-openingoxidationofcycliccompounds.
However,itrequiresexpensivestartingmaterialaswellastheenvir-onmentallyhazardousprocedures[2].
Alternatively,DCAscanbesynthesizedthroughbio-routesinvariousmicrobialhosts,suchasCandidasp.
[3,4],Yarrowialipolytica[5],Pseudomonasaeruginosa[6],Cyptococcusneoformans[7],andEscherichiacoli[8].
Ofthese,re-cently,Y.
lipolyticahasattractedgreatattentionasacellfactorytomanufactureDCAs[9]sincethisoleaginousyeastiscapableofaccumulatinglargeamountsoflipidsandpossessauniqueω-oxidationpathwaytocatalyzethehydrophobicsubstratessuchasn-alkaneandfattyacids[10–13].
Inω-oxidationpathway,aftertheoxida-tionofω-terminalofthefattyacid,fattyacidaldehydeisproducedfromω-hydroxyfattyacidbyfattyalcoholoxidase,followedbyitsoxidizationtoDCAsbyNAD-dependentfattyaldehydedehydrogenase[14].
Indeed,Y.
lipolyticaasoleaginousyeasthashugepotentialtobecomeamodelorganismtoproducefatty-acidderivedproductsincludingDCAs.
Inordertomakeitindustriallycompetitive,itsproductivityshouldbefur-therenhanced,whichcanbeachievedbymodificationofrelevanttargetgenes.
However,metabolicengineeringofY.
lipolyticahasbeenmainlyfocusedonover-producinglipids,e.
g.
,triacylglycerols(TAGs)withonlyahandfulofstudiesfortheincreasedDCAsproduction[15].
Inaddition,mostofgeneticengineeringtargetshavebeenidentifiedinanintuitiveoradhocmanner,whichlimitedourdesignscopeforstrainimprovement.
There-fore,itisnowimperativetoadoptmorerationalsystemsapproaches[16].
Inthisregard,variousstraindesignstrategiesguidedbyinsilicogenome-scalemetabolicmodel(GSMM)havebeendevelopedandsuccessfullyappliedtoseveralindustrialhostsincludingE.
coli[17,18]andS.
cerevisiae[19,20].
Similarly,inthiswork,weexploitedsuchinsilicomethodsusingGSMMofY.
lipolyticawhichwasnewlyreconstructedbysubstantiallyexpandingthepreviousmodels,thusallowingustoidentifyvariousgeneticengineeringtargetsforoverexpres-sion,knockout,andcofactormodificationtowardsDCAoverproduction.
ResultsanddiscussionGenome-scalemetabolicreconstructionofY.
lipolyticaWehavereconstructedagenome-scalemetabolicmodelofY.
lipolyticafordesigningDCAoverproducingstrains.
Initially,fourexistingY.
lipolyticaGSMMs(iNL895[21],iYL619[22],iMK735[23]andiYALI4[24])werecomparedonthebasisoftheirgeneannotations(seeMethodsandFig.
1a).
WhileiNL895wasdevelopedfromthephylogeneticallydistantyeastS.
cerevisiaemodel,iYL619wasreconstructedbasedonbiochemicalinformationdatabasessuchasKEGGandBRENDA.
iMK735wascompletelyderivedfromS.
cerevisiaemodeliND750[25]byaddingafewreactionsrelatedtoalkaneuptakeandlipidmetabolismwhereasiYALI4standsabitdifferentfromotherthreemodelssinceitwasbuiltautomaticallyusingRAVENtoolboxfromyeastconsen-susmodelastemplate.
Afterqualitativeandquantitativeassessment,iMK735wasselectedassuitablescaffoldmodelsinceithaswidercoverageofcellularphysiologyofY.
lipolyticawithbetterqualityintermsofrepresenta-tionandpredictionquality.
Withtheaimofbuildinganobjective-orientedmodelforsimulatingDCAsproduction,wefirstadded9biosyntheticreactionswhicharesequentiallycatalyzedbyhydroxylase,oxidaseanddehydrogenaseenzymeswithintheω-oxidationpathway,thuscapturingtheoxi-dationoffattyacidstoDCAs.
The27DCAsdegradationreactionswerealsoincludedtorepresentβ-oxidationpathwaytogetherwith6dicarboxylicacidstransportandanexchangereactionpertainingtoDDDAsecretion.
Basedonthescaffoldmodel,Y.
lipolyticaappearedtonotdegradeleucineandotherBCAAintoacetyl-CoA,buttherearereportsimplicatingthebiosyntheticcap-acityofleucineasaneffectoroflipogeniccapacityinoleaginousorganisms[26];aputativeacetyl-CoAprodu-cingleucinedegradationpathwaywasrecentlyidentified[27].
Therefore,wehaveadded4relevantenzymaticreactionsforleucinedegradationinthemodel.
Intotal,weadded50newreactionstothescaffoldmodelmainlyrelatedtoω-oxidationandBCAAdegradation.
Duringthemanualcuration,weremoved55reactionswhichareresponsiblefordead-endsorfutilecycleswithoutliteratureevidence.
Forexample,lactaldehydedehydrogenasewasdeletedsincethereisnotenoughexperimentalevidencetosupporttheexistenceofsuchreactioninY.
lipolytica.
Additionally,wecorrectedtheelementalbalanceanddirectionalityof45reactions.
Forexample,thedirectionalityofTHRArepresentingthethreoninealdolasemediatedbreakdownofthreoninetoglycineandacetaldehyde,waschangedtooppositedirec-tionfavoringacetaldehydeformationbasedonthedirec-tionreportedinS.
cerevisiae[28].
AllthechangesmadefromthescaffoldareprovidedinAdditionalfile1andsummarizedinFig.
1b.
Finally,theresultinginsilicoY.
lipolyticamodel(iYLI647)consistsof1347reactionsand1152metabolitesencodedby647genes(Fig.
1c).
TheiYLI647isavailableasSystemsBiologyMarkupLanguage(SBML)file(Additionalfile2).
Themodelpredictabilityofcellgrowthrelieshighlyontheaccuracyofbiomassequation.
However,wefoundthatthebiggestshortcomingofthepreviousmodelsisMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page10of130theinaccuratebiomasscompositionsincetheymainlyusedtheinformationofS.
cerevisiaebiomass.
Hence,thewholebiomasscompositionsofY.
lipolyticawerederivedbasedonvariousY.
lipolyticaexperimentaldatapublishedundercarbonlimitedandnitrogenlimitedconditionswhichcansignificantlyaltertheaminoacidandlipidcomposition.
ThecalculatedcompositionswerethenassimilatedintoiYLI647astwoseparatebiomasssynthesisequationspertainingtoC-limitedandN-limitedconditions(Additionalfile3).
Growthandnon-growthassociatedATPmaintenance(GAMandNGAM)requirementsforcellularprocesseswerealsoderivedusingrelevantliteraturedata[29].
TheNGAMofY.
lipolyticawasestimatedtobe5.
03mmolATP/gDCW,andGAMwas23.
09mmolATP/gDCW.
ComparativevalidationofiYLI647withOtherY.
lipolyticaGSMMsTheiYLI647wasevaluatedforitspredictionscorrelatedwithexperimentalphenotypes.
ThegrowthpredictionsunderdifferentcultureconditionsofiYLI647werecomparedwiththoseforotherY.
lipolyticaGSMMs.
Todoso,theinsilicobiomassyieldsofallthemodelsonglucoseandglycerolminimalmediumundersteady-stateconditionswerepredictedusingFBA.
Foreachsetofculturedatatakenfromindependentstudies,thecarbonsourceuptakerateswereconstrainedaccord-ingly,whilemaximizingbiomass.
Additionally,theCO2evolutionrate(CER)wasalsoconstrainedaccordingtotheexperimentaldatawhereverprovided.
ForthebiomassmaximizationiniMK735,biomassequationcorrespond-ingto5%lipidwasused.
Themaximumspecificgrowthratewasdeterminedforallmodelssimulatedundersameconstraintsandtheresultswerecompared(Fig.
2).
Datasets1and2weretakenfromWorkmanetal.
,[30]inwhichY.
lipolyticawasgrowninbatchcultureonglycerolandglucoseasasolecarbonsource,respectively.
Duringthebatchculturewithglucose,onlybiomassandCO2wereproducedasnometaboliteproductionwasob-served.
Inthecaseofglycerol,apartfrombiomassandCO2,smallamountofpolyolsintheformofmannitolandarabitol,werealsoproduced.
Since,theindividualbreak-downofpolyolswasnotmentionedintheliteratureandamountsweresmallenoughtodrasticallyaffecttheresult,wedidn'ttakeitintoconsiderationforsimulation.
Dataset3wastakenfromDulermoetal.
,[31]whichendeavoredtoanalyzetheY.
lipolyticamutantsforfattyacidproduc-tion.
ThecarbonsourceuptakeandCERwereconstrainedasperthevaluesreportedintheliteratureandbiomasswasmaximized.
Fig.
1ReconstructionprocessandcharacteristicofinsilicomodelsaComparisonofpreviousY.
lipolyticamodelsbSchematicdiagramhighlightingtheoverallreconstructionprocessofiYLI647,followedbyapplicationforstraindesigningcGeneralfeaturesofiYLI647incomparisonwithpreviousfourmodelsMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page11of130ItcanbeseenfromcomparativevalidationthatiYLI647canpredicttheevaluatedmacroscopicgrowthparametersmoreaccurately,i.
e.
smallerdeviationsfromexperimentaldatashowninblackdottedlines,ascom-paredtoothermodels.
Webelievethegrowthpredic-tionsaredirectlyinfluencedbytheaccuracyofthebiomasscomposition,andespecially,incaseofY.
lipoly-ticathebiomasscompositioncandrasticallychangedependingontheculturecondition.
AnotherimportantfactorthataffectsthepredictionsistheGAMandNGAMvaluesusedinthemodel.
Although,incompara-tivepredictiondoneinthecurrentstudy,wehaveusedsameNGAMinallthemodelsbutGAMvaluewasunchanged,andsomemodelsreportedveryhighGAMvalue,resultingindiscrepanciesinthemodelprediction.
MetabolicengineeringstrategiesforDCAsproductionIntheoleaginousorganism,denovoaccumulationoflipidsstartswiththeformationofanabolicacetyl-CoAviaglycolysis.
Thefattyacidsformedbytheseacetyl-CoAgetesterifiedtoformTAGs[32].
Ahighnumberofcarbonsourcesexcludingcelluloseandmethanol,havebeenconsideredassubstratesforthedenovoDCAsbiosynthesisinoleaginousmicroorganisms.
Amongthevariouscarbonsourcesavailable,weusedglucosetodesignthemetabolicengineeringstrategiesfordenovooverproductionofDDDA,arepresentativeoflong-chainDCAs.
Herein,weappliedthreemodel-guideddesignstrategiestooverproduceDDDAusingfourtools.
First,weemployedgeneticdesignbylocalsearch(GDLS)[33]tofindthegrowth-coupledsolutiontooverproduceDCAsbyknockingoutasetofreactions.
Then,weusedfluxactivityanalysis[34]toidentifythebottlenecksinthemetabolicnetworkwhichcanbeconsideredaspotentialoverexpressiontargets.
Inaddition,weimple-mentedtranscriptomic-basedstrainoptimizationtool(tSOT)[35]whichfirstgeneratesactivatedreactionsasareferencestatetoidentifythedeactivatedreactions,additionoftheeachdeactivatedreactionscanthenleadtotheincreaseintheproductyield.
Finally,tosupple-mentthemetabolicengineeringtargetswithcofactoravailability,weperformedcofactormodificationanalysis(CMA)[36]tofindthecofactorspecificityengineeringtargetsthatcanincreasethepoolofcofactorsrequiredforcatalyzingthereactions.
Figure3illustratesthecentralmetabolicnetworkofY.
lipolyticaandidentifiedgeneticengineeringtargetsforenhancingDCAsproduction.
OverexpressiontargetsExpressionlevelofgenesandtheactivityoftheirenzymeproductsarehighlyoptimizedtomeettheperformancedemandofabiologicalsystem[37].
How-ever,theseenzymeexpressionsaltertoadaptagainstthechangingbiologicalconditions.
Thisrangeoffluctua-tionsingeneexpressionlevelcanbeexploitedtodesignanoverexpressionsystem[38].
Basedonthisprinciple,weperformedfluxactivityanalysisunderglucoseminimalmediumbyfixinganoptimalbiomassandsys-tematicallyincreasingfluxactivityofeachreactionfrom0%to100%andmaximizingDDDA.
SimulationresultsshowthatwhenY.
lipolyticagrowingin10mmol/gDCW-hrofglucose,theincreaseinfluxactivityofsomereactionshasproportionaleffectonthemaximumachievableyieldofDDDA(Fig.
4).
Thesedirectionallycoupledreactionscanbegeneticallyoverexpressedtocomplementtheω-oxidationpathwayenzymestopro-duceDDDAatitstheoreticalmaximum.
Someofthebottleneckreactions,e.
g.
,DDCAH,acytochromeP450Fig.
2ComparativevalidationofiYLI647withall4Y.
lipolyticamodelsavailableunder3differentdatasets.
ThedottedlinesrepresentexperimentalvalueMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page12of130hydroxylase(CYP52)andDDCAFAO,afattyalcoholoxidaseidentifiedbyinsilicoanalysishavebeenverifiedasoverexpressiontargetsforDDDAproduction[39].
Theotherhypothesizedreactionsincludeacetyl-CoAcarboxylase(ACCOAC),overexpressionofwhichmayincreasethemalonyl-CoApool.
ItcanbefurtherutilizedbyFAScomplextogeneratemorefattyacids,whichcanthenbechannelizedtoω-oxidationpathway.
Consideringthatfattyacidbiosyntheticpathwayisverytightlyregulated,transcriptomicdatacanprovidesomeusefulinsightintotheON/OFFstateofthereactionsinaparticularcondition.
ToidentifytheoverexpressiontargetsFig.
4Simulationresultbyfluxactivityanalysis.
Overexpressiongenesandproductionratechangesdependonthealterationoffluxactivitiesofrespectivegenes.
GAPD(Glyceraldehyde-3-phosphatedehydrogenase),PGK(Phosphoglyceratekinase),TPI(Triosephosphatekinase),ACCOAC(Acetyl-CoAcarboxylase),ALDDHDD(Aldehydedehydrogenase),DDCAFAO(Fattyacidoxidase),DDCAH(Fattyacidhydroxylase)Fig.
3ThecentralmetabolicnetworkofY.
lipolyticadepictingmetabolicengineeringtargetstoproduceDDDAMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page13of130forincreasingtheDCAsproduction,weimplementedtSOTbyresortingtotime-coursetranscriptomicprofileoftheY.
lipolytica,duringacontrolledfed-batchusingglu-coseasthesolecarbonsourceafter27htime-pointwhichcorrespondstoearlystationaryphase[40].
Anitrogenlimitationwasappliedduringthefed-batchtoinitiatedenovolipidsynthesiswhichcanrepresentthepreconditionforhighDCAsproduction.
ThebasicprincipleoftSOTistoascertainthegeneoverexpressiontargetsbyrestoringthereactionswhichareremovedfromGSMMbydata-integrationalgorithmswhiledevelopingacontext-specificmodel.
Asaresult,tSOTidentifiedMDH,bothcytosolicandmitochondrial,asanoverexpressiontarget.
Inaddition,italsofoundmitochondrialNAD-dependentmalicenzyme(ME1m)andglutamatedehydrogenase(GLUDy)tobetheoverexpressiontargets(Table1).
Interestingly,owingtothefactthatω-oxidationisanoxidativeprocesswiththehighdemandofredoxcofactors,alltheidentifiedreactionsareinvolvedincofactorregeneration.
MEishypothesizedtobethesupplierofNADPHduringlipidbiosynthesisinmostoleaginousyeaststhroughtheintracellularsubstratecyclesinvolvingMDH,pyruvatecarboxylase(PC)andME,alsocalled"transhydrogenasecycle"[41].
AlthoughY.
lipolyticalacksacytosoliccopyofMErequiredtocompensateforNADPHdemand,itcouldbeinterestingtoinvestigatethecompoundeffectofoverexpressingMDHandMEbecauseapartfromNADPH,MEinconjunctionwithmitochondrialpyruvatedehydrogenase(PDH)alsoprovidemitochondrialacetyl-CoA.
ThemostinterestingfindingfromtSOTisGLUDy,whichapartfromregenerationofNADPH;alsoplayskeyroleinmaintainingthebalanceofcarbonandnitrogen.
ThereisconsiderableevidenceofthepresenceofGLUDyshuntinplants,whichreturnsthecarboninaminoacidsbiosynthesisbackintoreactionsofcarbonmetabolismandTCAcycle[42]whichisthecaseinnitrogenstarvingcondition[24],makingitparticularlyaninterestingtargettoexploreasitlinksaminoacidbiosynthesistofattyacidmetabolism.
Downregulation/knockouttargetsGDLSalgorithmwasusedtosearchforgrowth-coupledsolutionsforDCAproduction,identifyingupto5reactiondeletioncandidates.
Basically,GDLSproposethepathwaydesignwhichcouplestheproductformationwiththecellgrowth,makingitsproductionnecessarytoreachoptimalgrowth.
ThestraindesignstrategydecipheredbyGDLScombinedthesimultaneousknockoutofDESAT16(Stear-oyl-CoAdesaturase)andMI1PS(Myo-inositol-1-phos-phatesynthase).
DESAT16istheenzymethatcatalyzestheconversionofsaturatedfattyacidtomonounsaturatedfattyacids.
TheoverexpressionofthisenzymeisshowntoincreaselipidaccumulationinY.
lipolyticawhichappar-entlytakesthefluxawayfromDCAsproductionwhichrequiresfreefattyacid[43],deletingwhichcanresultinincreasedpooloffreefattyacids.
SinceDCAsisanon-growthcoupledproduct,deletionofDESAT16andMI1PSassuggestedbyGDLSmaynotgiverisetoimprovedproductyield.
However,inthesimplenetworkperspective,DESAT16andMI1PSreactionsbranchthecarbonfluxawayfromDCAsformationwhichmakestheminterestingknockouttargetstotestexperimentally.
ItisworthnoticingherethatMI1PScatalyzesthefirstreactionintheinositolpathwaywhichproducesmembraneformingmetabolites,soknockingoutMI1PScouldshowdeleteri-ouseffectsoncellgrowth.
Nonetheless,sinceDCAsisnon-growthassociatedproduct,aninducibleknockoutstraincouldbeusedwhereinMI1PSissuppressedwhentheculturereachesstationaryphase.
Furthermore,formedDCAscangetdegradedviaβ-oxidationpathway.
Thereforeblockingitbydeletionofacyl-CoAoxidaseencodedbyPOX1–6genescanfurtherenhancethetiterofDCAs.
Y.
lipolyticawhichlacksacyl-CoAoxidasescanmoreefficientlyconvertn-alkanesandfattyacidsortheirderivativestotheircorrespondingDCAs[5].
CofactorspecificityengineeringtargetsSimilartootherbiosyntheticpathways,theDCAsbiosynthesisviaω-oxidationpathwayinY.
lipolyticainvolvesseveraluniquereactionsandiscommonlycontrolledbythesupplyofprecursorsandcofactors.
Earlierpartofthisstudyhasfocusedonoverexpressionanddownregulationofsomekeyenzymesregulatingtheω-oxidationpathway.
However,cofactorsareveryimportanttoachieveimprovementinproductivity.
NADPH,asareducingequivalent,usuallyplaysanim-portantroleincouplingcatabolismwithanabolismandenergygenerationduringmetabolism.
Severalmetabolicengineeringapproacheshavebeenimplementedtomanipulatethecofactorsleveltoincreasetheproductyieldinothermicroorganisms[44,45]Increasingtheα-santaleneproductionbymodifyingtheammoniumTable1OverexpressiontargetssimulatedbytSOTtoincreaseDCAsproductionTargetsReactionNameReactionDefinitionYieldImprovement(%)GLUDyGlutamateDehydrogenase(NADPH-forming)glu_L[c]+h2o[c]+nadp[c]akg[c]+h[c]+nadph[c]+nh4[c]22.
2MDHMalateDehydrogenase(cytosol)mal_L[c]+nad[c]h[c]+nadh[c]+oaa[c]47.
8MDHmMalateDehydrogenase(mitochondrial)mal_L[m]+nad[m]h[m]+nadh[m]+oaa[m]47.
8ME1mMalicenzyme(NAD-dependent)mal_L[m]+nad[m]->co2[m]+nadh[m]+pyr[m]47.
8Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page14of130assimilationfrombeingNADPHtoNADHdependentbythedeletionofGDH1andtheoverexpressionofGDH2;overexpressionofStreptococcusmutantsgapNgenewhichencodesaGAPDHtoincreasetheL-lysineproduction[46];engineeringNADPHregenerationforimprovingpentosefermentationbyoverexpressingtheGDP1,aNADP-dependentGAPDHfromKluyveromyceslactis[47];overexpressionoftranshydrogenaseandNADkinasetoimproveisobutanolproduction[48],areafewexamplesshowingtheemergenceofNADPHlevelengineeringaspotentandfeasiblestrategytoincreasetheproductioninmicrobialhosts.
Fattyacidbiosynthesisandω-oxidationpathwayareNADPHdemandingoxidativepathwayinY.
lipolytica.
Thisexcessivecofactordemandismainlysatisfiedthroughthepentosephosphatepathway(PPP)reactions.
But,theincreasedusageofthePPPthroughoverexpres-sionissuboptimalasonemoleofcarbonislostasCO2foreverytwomolesofNADPHproduced[49].
Tocir-cumventthis,weperformedCMAinordertoidentifythetargetsforcofactorspecificityengineeringtoim-provetheNADPHpoolwhichcanenhanceDDDAyield.
FromtheCMAresults,itwasobservedthattheDDDAyieldcanbeimprovedbyincreasingNADPHregener-ationthroughmodificationofcofactorspecificityfromNADtoNADP.
Amongthetargetsfound,changingthecofactorspecificityofGAPDandMDHfromNADtoNADPwasfoundtogivethebestimprovementinDDDAyield(Table2).
Toachieveahigh-levelproductionofDCAinyeast,weproposetoenhancethesupplyofNADPHtotheω-oxidationpathway.
Hence,amulti-stepmetabolicengin-eeringstrategycanbedevisedtosimultaneouslymodifytheglycolysisstep,optimizetheNADPHsupplyandtoenhancetheactivityofsomeNADPHproducingenzymes.
Effectofbranched-chainaminoacidssupplementationonDCAproductionInmostrecombinantDCAsproductionstudies,thehydrophobicsubstratessuchasalkaneandfattyacidmethylester(FAME)wereusedforthebiotransform-ation[5].
Assuch,glucoseisusedasthecarbonsourceforthegrowthbeforeinducingtheω-oxidationpathwayusingalkane.
However,ithasbeenshowninY.
lipolyticathatlipidaccumulationcanoccurdenovowithoutexogenoussupplyofhydrophobicsubstrate,layingthegroundsforapossibilitytobeusedfordenovoDCAsproductionaswell,fromprimarycarbonsources,i.
e.
,glucoseorglycerol[50].
Inadditiontothis,theabilitytoefficientlyproduceacetyl-CoAbymetabolizingothercarbonsourcesinthemediumwilldrivetheimportantprecursortowardsDCAsproduction.
Presently,thereareevidencesimplicatingastrongcorrelationbetweenlipidaccumulationandleucinemetabolisminS.
cerevi-siae[51]andY.
lipolytica[52].
TakingthecluefromlipidaccumulationstudiesonY.
lipolytica,wesoughttoexploretheeffectofBCAAsupplementationonacetyl-CoApool,whichcanbeproducedviadegradationpathwayofBCAA.
Inordertotheoreticallyanalyzethefluxdistributionandflux-sumchangesthatoccurwhenBCAAissupplemented,wemaximizedtheDDDAproductionusingglucoseastheprimarycarbonsourcesupplementedwith10C-mmol/gDCW-hrofleucine,isoleucine,orvaline.
Toelucidatethedifferenceinpath-wayutilizationandmetaboliteturn-overindifferentaminoacidsduringDDDAproduction,wepreparedtheheatmapoffluxdistributionandflux-sumasshowninFig.
5.
Itcanbeseenthatacetyl-CoAbeingcriticalde-terminantofDCAssynthesiswasproducedatthehigh-estlevelonvalinesupplementationfollowedbyleucinewhichcorrelatedwithDDDAturnovertrend.
ApartfromtheroleofBCAAsinacetyl-CoAproduction,supplementingaminoacidsdecreasestheprimarycarbondemandinaminoacidbiosynthesisforbiomassformation.
ThisextracarboncanthenbedivertedtoDCAsbiosynthesis.
Thisobservationisincloseresem-blancewithnitrogenstarvationconditionforlipidaccu-mulationbecauseinnitrogenlimitingcondition,biosynthesisofaminoacidseizesandcarbonpresentthencanbeutilizedinotherbiosyntheticpathways.
Fromoursimulation,wehaveidentifiedthatglucosesupplementedwithvalineorleucinecanincreasethegrowthaswellasdenovoaccumulationofDCAsinY.
lipolytica.
Thisisthefirstreportimplicatingaprominentroleofvalinedegradationasanalternaterouteofacetyl-CoAbiosynthesisinY.
lipolytica,experimentalvalidationofwhichcangiveinsighttointricatecorrelationbetweenfattyacidbiosynthesisandaminoaciddegradation.
ConclusionTheGSMMofY.
lipolytica,iYLI647,wasreconstructedusingiMK735asascaffoldandtherelevantinformationfromotheravailableY.
lipolyticamodels.
iYLI647con-sistsof1347reactionsand647genes.
ThebiomassTable2CofactorspecificityengineeringtargetsbyCMAtoincreaseDCAsproductionReactionNameReactionDefinitionYieldofDDDA(mmol/gDCW-hr)NADNADPMDHmal_L[c]+nad[c]h[c]+nadh[c]+oaa[c]2.
6012.
781GAPDg3p[c]+nad[c]+pi[c]13dpg[c]+h[c]+nadh[c]2.
6012.
778Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page15of130equationswerecarefullyformulatedwithvariousexperi-mentalinformationofY.
lipolytica,whichisperhapsthereasonforaccuratemodelprediction.
ThepotentialofdenovoDCAsproductioninY.
lipolyticacombinedwithamodel–drivenstraindesignformetabolicengin-eeringandmediaoptimizationstrategieswerethenevaluatedusingthereconstructediYLI647.
ThefluxtowardsDDDAproductionwasincreasedfollowingtheoverexpressionanddeletionoffewreactionsandthemodel-basedstraindesigngivesusagoodstartingpointtoexplorethemetaboliccapabilitiesofY.
lipolyticatoproducefattyacidderivedproducts.
Moreover,theworkflowandprocedurepresentedinthisanalysiscanbeutilizedasaplatformtoperformsimilaranalyseswithdifferentorganisms.
MethodsMetabolicnetworkreconstructionWechoseallfourseparatelydevelopedpubliclyavailablemetabolicnetworkreconstructions,iNL895,iYL619,iMK735,andiYali4,tocompareandselectthebestsuitedasscaffoldmodel.
Followingourqualitativeandquantitativecomparison,wechoseiMK735asascaffoldmodelandproceededtomanualcurationtoexpandthemodelcoverageandcharacteristics.
Todothis,usingliteraturedata,weverifiedthepresenceofreactionsandrelevanceinY.
lipolyticametabolism.
Then,weaddedω-oxidationpathwaytoconvertfattyacidstoDCAs,andalsothesubsequentdegradingβ-oxidationpathway.
Followingtheliteratureevidenceanditsestablishedim-portanceinfattyacidmetabolism,wealsoincludedBCAAdegradationpathways.
Furthermore,wecheckedthemassbalanceofreactionsandmadeappropriatechangestomakestoichiometricallybalancedreactions.
Inaddition,wederivedthebiomassequationincarbonandnitrogenlimitedconditionsusingupdatedandrele-vantliteraturesources.
Usingrefinedbiomassequation,wesimulatedthemodeltoidentifytheloops,andmissinglinkreactionsusingGapFinder[53].
LoopswereremovedbychangingthedirectionalityofcoupledFig.
5Fluxdistributionandflux-sum.
Heatmapshowingfluxdistributionandflux-sumofimportantreactionsandmetabolites,respectively,duringDDDAproductionMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page16of130reaction,orremovingthenon-metaboliclumpedreactions,whereas,gapswerefilledwithreactionsfromorthologsorintroducingtransport/sinkreaction.
Constraints-basedfluxanalysisThecellularmetabolismofY.
lipolyticawassimulatedundervaryingenvironmentalconditionsusingconstraints-basedfluxanalysis.
Thebiomassreactionwasmaximizedtosimulatethegrowthundervariouscultureconditionsasdescribedelsewhere[54–56].
Themaximizationofbiomassissubjectedtostoichiometricandcapacityconstraints,whichcanmathematicallybeformulatedas:maxZ1Xjcjvj1s:t:XjSijvj0metaboliteivminj≤vj≤vmaxjreactionjwhereSijreferstothestoichiometriccoefficientofmetaboliteiinvolvedinreactionj,vjdenotestothefluxorspecificrateofmetabolicreactionj,vminjandvmaxjrep-resentthelowerandupperlimitsonthefluxofreactionj,respectively;andZ1correspondstothecellularobject-iveasalinearfunctionofallthemetabolicreactionswheretherelativeweightsaredeterminedbythecoeffi-cientcj.
Inthisstudy,theconstraints-basedfluxanalysisproblemsweresolvedusingCOBRAtoolbox[57].
Insilicomodel-basedstraindesignFourinsilicostraindesignapproacheshavebeenemployedtoidentifymetabolicengineeringtargetsforoverproducingDCAfromglucose.
Geneticdesignbylocalsearch(GDLS)GDLSstraindesignalgorithm[33]wasimplementedinCOBRAtoolboxinMatlab,searchinguptomaximumof5knockoutreactionswiththeouterobjectiveofmaximizingDDDAexchangeflux.
FluxactivityanalysisToidentifytheupregulationanddownregulationgene/reactiontargetswhichleadtotheenhancedproductionofdesiredcompound,wefirstneededtoquantifythefluxactivityofallreactionsinthewild-typestrain.
Fluxactivity,fj,isdefinedastheabsolutevalueofreactionflux,vj[34].
Thiscanbedeterminedbyfirstsolvingtheconstraints-basedfluxanalysisproblem(Eq.
1)withbio-massmaximizationasobjective,andthenobtainingtheabsolutevaluesofindividualreactionfluxes.
Next,wesolvedthebelow-mentionedmixed-integerlinearpro-gramming(MILP)problemtoidentifythemaximumandminimumfluxactivitiestodeterminethefeasiblerangesofindividualreactionssuchthattheycanbeup-regulatedanddownregulatedwithinthislimit:max=minfj2Subjectto:XjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1Where,αjandβjareupperandlowerboundsoffluxes,respectively.
Thefj+andfjarethepositivetwovariablesintowhichtheflux,vj,isdecomposed.
Itisobservedthatfjfjjfjfjj,ifandonlyifeitherfj+orfjisequaltozero.
ThisconditionwasintroducedbynewbinaryvariablesIj+andIjwhicharewhenmultipliedwithalargeintegerwhichshouldbelargerthanthefluxofex-perimentallymeasuredvalue,M,willbealwayslesserthanfj+andfj,respectively.
Additionally,theconstraintIjIj1isalsointroducedtoensurethateitherfj+orfjareequaltozero.
Oncethereferencefluxactivitiesareestablished,i.
e.
wild-type,maximumandminimumvalues,wethensolvethebelowmentionedMILPproblemtoanalyzetheeffectsofperturbingaparticularfluxactivityoncellulargrowth.
maxvbiomass3Subjectto:XjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1C1:fjfj≤fminjkattfWTjfminjMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page17of130ORC2:fjfj≥fWTjkintfmaxjfWTjWhere,constraints(C1)and(C2)areapplicableforupregulationanddownregulationproblems,respectively.
Parameterskattandkintaregraduallyvariedbetween0and1instepsof0.
1toanalyzetheeffectofreactionupregulationbetweenminimalandwild-typevalues,andreactiondownregulationbetweenthewild-typeandmaximalvalues,respectively.
Finally,theobjectivevalueobtainedfromthesolutionof(Eq.
3)isusedasthelowerlimitforcellgrowthinthefourthstepwhereby(Eq.
3)issolvedagainwiththetargetedproductastheobjectivefunction.
Thecorre-spondingmathematicalformulationisasfollows:minvEXsuccSubjectto:vbiomass≥Bj;kXjSijvj0αj≤vj≤βjvjfjfjfj≥0;fj≥0fj≤IjM;fj≤IjMIj∈0;1fg;Ij∈0;1fgIjIj1C1:fjfj≤fminjkattfWTjfminjORC2:fjfj≥fWTjkintfmaxjfWTjwhereBj,kisthemaximumbiomassobtainablewhilesolvingproblem(Eq.
3)forjthreactionatkthupregula-tion/downregulationlevels.
Alltheoptimizationprob-lemsweresolvedusingtheGAMSIDEsoftwareversion22.
4withIBMILOGCPLEXsolver.
Transcriptomics-basedstrainoptimizationtool(tSOT)SinceoptimalityassumptionbasedFBAalgorithmsignorestheregulatoryconsiderationwhilestraindesigning,weimplementedtranscriptomics-basedstrainoptimizationtool(tSOT)[35]toidentifymetabolicengineeringtargetsbasedontranscriptomicdataintegratedinthemodel.
Acomprehensivetime-coursetranscriptomicprofilefromthecultureofY.
lipolytica,duringacontrolledfed-batchonglucose,wasusedastranscriptomicdataobtainedfromGeneExpressionOmnibus(GEO)databaseusingaccessionnumberGSE29046[40].
Cofactormodificationanalysis(CMA)Sincetheω-oxidationpathwayisanoxidativeprocessandrequirescofactoroptimizationtomaximizetheoreticalyield,weusedCMAtoidentifythecofactorspecificityengineeringtargetwhichcanincreasetheyieldofDDDA.
CMAwasimplementedasdescribedinourpreviouswork[36].
Mathematically,thebi-levelmixed-integernonlinearprogramming(MINLP)optimizationproblemspecifictotheCMAcanberepresentedasfollows:maxφproduct0:5XjSproduct;jvjs:t:maxvbiomasss:t:XjSijvjScModijvcModj0metaboliteivbiomass≥vminbiomass1ycModjvminj≤vj≤1ycModjvmaxjycModjvminj≤vcModj≤ycModjvmaxjycModj0;1fgreactionj266666666664377777777775XjycModj≤kWhere,ScModijisthecofactormodifiedstoichiometricmatrixwherethecoefficientsaresameasSij,exceptthere-actionswhichinvolveeitherNAD(H)orNADP(H).
ThesereactionsareswappedforcofactorsintheScModijmatrixsuchthatSNADH;jScModNADPH;jandSNADPH;jScModNADH;j.
vcModjisthefluxthroughthecofactormodifiedreactionandvminbiomassistheminimumamountofbiomassthatneedstobeproduced.
ThebinaryvariableycModjensuresthatthecofactorassociatedreactionsareallowedtocarryfluxeitherwithitsoriginalorswappedcofactorbutnotboth.
Thenumberofcofactorswitchesallowedinaparticularsimula-tioniscontrolledbythenumber,kandwhichisfixedat1forallsimulationsinthiswork.
Thebi-levelMINLPproblemwasreformulatedasasingle-levelMINLPprob-lemusingtheprimaldualtransformationasimplementedearlier[58].
TheMILPoptimizationproblemwassolvedusingtheGAMSIDEsoftwareversion22.
4withIBMILOGCPLEXsolver.
Flux-sumIntheconstraints-basedfluxanalysis,thereisnoaccu-mulationofintermediatemetabolitesduetothesteady-statecondition.
However,theturnoverrate,whichisalsoequivalenttothetotalconsumptionorproductionrate,oftheintermediatescanbenonzero,whichisdefinedasMishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page18of130theirflux-sum[59].
Sincetheoverallconsumptionandgenerationratesareequalunderthesteady-stateas-sumption,theflux-sumofmetaboliteicanbeformu-latedasΦi=0.
5∑|Sijvj|.
Eachterminthissummationseriesgivesustheabsoluterateofconsumption/gener-ationofmetaboliteiduetoreactionjandthusbyhalv-ingthesumoftheseterms,wecanobtaintheoverallturnoverrateformetabolitei.
AdditionalfilesAdditionalfile1:Changed,addedanddeletedreactionsiniYLI647modelincomparisonwithiMK735scaffoldmodel.
(XLSX22kb)Additionalfile2:SMBLfileofiYLI647.
(XML2312kb)Additionalfile3:BiomasscompositionofY.
lipolyticainC-andN-limitedconditionsandGAMandNGAMcalculations.
(DOCX52kb)AcknowledgementsWethankDr.
Jung-OhAhn(KoreaResearchInstituteofBioscienceandBio-technology)forhelpanddiscussion.
FundingThisworkwassupportedbytheAcademicResearchFund(R-279-000-476-112)oftheNationalUniversityofSingapore,BiomedicalResearchCouncilofA*STAR(AgencyforScience,TechnologyandResearch),Singapore,theGlobalR&Dprojectprogram(N011500017),MinistryofTrade,IndustryandEnergy(MOTIE),RepublicofKoreaandtheNext-GenerationBioGreen21Pro-gramoftheRuralDevelopmentAdministration,RepublicofKorea(SystemsandSyntheticAgrobiotechCenter;grantno.
PJ01334605).
PublicationcostwascoveredbyNUSSynCTIprogram.
AvailabilityofdataandmaterialsAlldatageneratedoranalyzedduringthisstudyareincludedinthispublishedarticleanditssupplementaryinformationfiles.
AboutthissupplementThisarticlehasbeenpublishedaspartofBMCSystemsBiologyVolume12Supplement2,2018:Proceedingsofthe28thInternationalConferenceonGenomeInformatics:systemsbiology.
Thefullcontentsofthesupplementareavailableonlineathttps://bmcsystbiol.
biomedcentral.
com/articles/supplements/volume-12-supplement-2.
Authors'contributionsPM,NRLandDYLconceivedanddesignedtheinsilicostudy.
PM,NRL,MLandMSKperformedthecomputationalsimulationsanddraftedthemanuscript.
PM,NRL,BGKandDYLrevisedthemanuscript.
Allauthorsreadandapprovedthefinalmanuscript.
EthicsapprovalandconsenttoparticipateNotapplicable.
ConsentforpublicationNotapplicable.
CompetinginterestsTheauthorsdeclarethattheyhavenocompetinginterests.
Publisher'sNoteSpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
Authordetails1NUSSyntheticBiologyforClinicalandTechnologicalInnovation(SynCTI),LifeSciencesInstitute,NationalUniversityofSingapore,28MedicalDrive,Singapore117456,Singapore.
2BioprocessingTechnologyInstitute,AgencyforScience,TechnologyandResearch(A*STAR),20BiopolisWay,#06-01,Centros,Singapore138668,Singapore.
3SchoolofChemicalandBiologicalEngineering,InstituteofMolecularBiologyandGenetics,andBioengineeringInstitute,SeoulNationalUniversity,1Gwanak-ro,Gwanak-gu,Seoul151-742,RepublicofKorea.
4SchoolofChemicalEngineering,SungkyunkwanUniversity,2066Seobu-ro,Jangan-gu,Suwon,Gyeonggi-do16419,RepublicofKorea.
Published:19March2018References1.
CornilsB,LappeP.
Dicarboxylicacids,Aliphatic.
Ullmann'sEncyclIndChem.
2012;11:287–304.
https://doi.
org/10.
1002/14356007.
a08.
2.
PolenT,SpelbergM,BottM.
Towardbiotechnologicalproductionofadipicacidandprecursorsfrombiorenewables.
JBiotechnol.
2013:75–84.
https://doi.
org/10.
1016/j.
jbiotec.
2012.
07.
008.
3.
GreenKD,TurnerMK,WoodleyJM.
CandidaCloacaeoxidationoflong-chainfattyacidstodioicacids.
EnzymMicrobTechnol.
2000;27:205–11.
https://doi.
org/10.
1016/S0141-0229(00)00217-9.
4.
JiaoP,MaS,HuaY,HuangY,CaoZ.
IsolationandenzymedeterminationofCandidaTropicalismutantsforDCAproduction.
JGenApplMicrobiol.
2000;46:245–9.
https://doi.
org/10.
2323/jgam.
46.
245.
5.
SmitMS,MokgoroMM,SetatiE,NicaudJ-M.
Alpha,omega-dicarboxylicacidaccumulationbyacyl-CoAoxidasedeficientmutantsofYarrowialipolytica.
BiotechnolLett.
2005;27:859–64.
https://doi.
org/10.
1007/s10529-005-6719-1.
6.
ChanEC,ChengCS,HsuYH.
ContinuousproductionofdicarboxylicacidbyimmobilizedPseudomonasAeruginosacells.
JFermentBioeng.
1997;83:157–60.
https://doi.
org/10.
1016/S0922-338X(97)83575-1.
7.
ChanEC,KuoJ.
BiotransformationofdicarboxylicacidbyimmobilizedCryptococcuscells.
EnzymMicrobTechnol.
1997;20:585–9.
https://doi.
org/10.
1016/S0141-0229(96)00198-6.
8.
Sathesh-PrabuC,LeeSK.
Productionoflong-chainalpha,omega-dicarboxylicacidsbyengineeredEscherichiaColifromrenewablefattyacidsandplantoils.
JAgricFoodChem.
2015;63:8199–208.
https://doi.
org/10.
1021/acs.
jafc.
5b03833.
9.
Ledesma-AmaroR,NicaudJM.
Yarrowialipolyticaasabiotechnologicalchassistoproduceusualandunusualfattyacids.
ProgLipidRes.
2016;61:40–50.
https://doi.
org/10.
1016/j.
plipres.
2015.
12.
001.
10.
EschenfeldtWH,ZhangY,SamahaH,StolsL,EirichLD,WilsonCR,etal.
TransformationoffattyacidscatalyzedbycytochromeP450monooxygenaseenzymesofCandidaTropicalis.
ApplEnvironMicrobiol.
2003;69:5992–9.
https://doi.
org/10.
1128/AEM.
69.
10.
5992-5999.
2003.
11.
GatterM,FrsterA,BrK,WinterM,OttoC,PetzschP,etal.
Anewlyidentifiedfattyalcoholoxidasegeneismainlyresponsiblefortheoxidationoflong-chainω-hydroxyfattyacidsinYarrowialipolytica.
FEMSYeastRes.
2014;14:858–72.
https://doi.
org/10.
1111/1567-1364.
12176.
12.
EirichLD,CraftDL,SteinbergL,AsifA,EschenfeldtWU,StolsL,etal.
CloningandcharacterizationofthreefattyalcoholoxidasegenesfromCandidaTropicalisstrainATCC20336.
ApplEnvironMicrobiol.
2004;70:4872–9.
https://doi.
org/10.
1128/AEM.
70.
8.
4872-4879.
2004.
13.
IwamaR,KobayashiS,IshimaruC,OhtaA,HoriuchiH,FukudaR.
FunctionalrolesandsubstratespecificitiesoftwelvecytochromesP450belongingtoCYP52familyinn-alkaneassimilatingyeastYarrowialipolytica.
FungalGenetBiol.
2016;91:43–54.
https://doi.
org/10.
1016/j.
fgb.
2016.
03.
007.
14.
BroadwayNM,DickinsonFM,RatledgeC.
Theenzymologyofdicarboxylic-acidformationbyCorynebacteriumSpstrain7E1CgrownonN-alkanes.
JGenMicrobiol.
1993;139:1337–44.
https://doi.
org/10.
1099/00221287-139-6-1337.
15.
NicaudJ,ThevenieauF,LeDallM,MarchalR.
Productionofdicarboxylicacidsbyimprovedmutantstrainsofyarrowialipolytica.
USpatapp11/;2005.
p.
1.
Available:http://www.
google.
com/patents/US2010004111516.
SimeonidisE,PriceND.
Genome-scalemodelingformetabolicengineering.
JIndMicrobiolBiotechnol.
2015:327–38.
https://doi.
org/10.
1007/s10295-0e14-1576-3.
17.
AlperH,JinYS,MoxleyJF,StephanopoulosG.
IdentifyinggenetargetsforthemetabolicengineeringoflycopenebiosynthesisinEscherichiaColi.
MetabEng.
2005;7:155–64.
https://doi.
org/10.
1016/j.
ymben.
2004.
12.
003.
18.
AlperH,MiyaokuK,StephanopoulosG.
Constructionoflycopene-overproducingE.
Colistrainsbycombiningsystematicandcombinatorialgeneknockouttargets.
NatBiotechnol.
2005;23:612–6.
https://doi.
org/10.
1038/nbt1083.
19.
BroC,RegenbergB,FrsterJ,NielsenJ.
InsilicoaidedmetabolicengineeringofSaccharomycesCerevisiaeforimprovedbioethanolproduction.
MetabEng.
2006;8:102–11.
https://doi.
org/10.
1016/j.
ymben.
2005.
09.
007.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page19of13020.
SunZ,MengH,LiJ,WangJ,LiQ,WangY,etal.
IdentificationofnovelknockouttargetsforimprovingterpenoidsbiosynthesisinSaccharomycesCerevisiae.
PLoSOne.
2014;9https://doi.
org/10.
1371/journal.
pone.
0112615.
21.
LoiraN,DulermoT,NicaudJ-M,ShermanDJ.
Agenome-scalemetabolicmodelofthelipid-accumulatingyeastYarrowialipolytica.
BMCSystBiol.
BioMedCentralLtd.
2012;6:35.
https://doi.
org/10.
1186/1752-0509-6-35.
22.
PanP,HuaQ.
Reconstructionandinsilicoanalysisofmetabolicnetworkforanoleaginousyeast,Yarrowialipolytica.
PLoSOne.
2012;7:e51535.
https://doi.
org/10.
1371/journal.
pone.
0051535.
23.
KavekM,BhutadaG,MadlT,NatterK,KrainerF,DietzschC,etal.
Optimizationoflipidproductionwithagenome-scalemodelofYarrowialipolytica.
BMCSystBiol.
2015;9:72.
https://doi.
org/10.
1186/s12918-015-0217-4.
24.
KerkhovenEJ,PomraningKR,BakerSE,NielsenJ.
Regulationofamino-acidmetabolismcontrolsfluxtolipidaccumulationinYarrowialipolytica.
Npj.
SystBiolAppl.
2016;2:16005.
https://doi.
org/10.
1038/npjsba.
2016.
5.
25.
DuarteNC,HerrgrdMJ,PalssonB.
ReconstructionandvalidationofSaccharomycesCerevisiaeiND750,afullycompartmentalizedgenome-scalemetabolicmodel.
GenomeRes.
2004;14:1298–309.
https://doi.
org/10.
1101/gr.
2250904.
26.
BlazeckJ,LiuL,KnightR,AlperHS.
HeterologousproductionofpentaneintheoleaginousyeastYarrowialipolytica.
JBiotechnol.
2013;165:184–94.
https://doi.
org/10.
1016/j.
jbiotec.
2013.
04.
003.
27.
VorapreedaT,ThammarongthamC,CheevadhanarakS,LaotengK.
Alternativeroutesofacetyl-CoAsynthesisidentifiedbycomparativegenomicanalysis:involvementinthelipidproductionofoleaginousyeastandfungi.
Microbiology.
2012;158:217–28.
https://doi.
org/10.
1099/mic.
0.
051946-0.
28.
VanMarisAJA,LuttikMAH,WinklerAA,VanDijkenJP,PronkJT.
Overproductionofthreoninealdolasecircumventsthebiosyntheticroleofpyruvatedecarboxylaseinglucose-limitedchemostatculturesofSaccharomycesCerevisiae.
ApplEnvironMicrobiol.
2003;69:2094–9.
https://doi.
org/10.
1128/AEM.
69.
4.
2094-2099.
2003.
29.
PapanikolaouS,AggelisG.
LipidproductionbyYarrowialipolyticagrowingonindustrialglycerolinasinglestagecontinuousculture.
BioresourTechnol.
2002;82:43–9.
https://doi.
org/10.
1186/s13068-015-0286-z.
30.
WorkmanM,HoltP,ThykaerJ.
ComparingcellularperformanceofYarrowialipolyticaduringgrowthonglucoseandglycerolinsubmergedcultivations.
AMBExpress.
2013;3:58.
https://doi.
org/10.
1186/2191-0855-3-58.
31.
DulermoT,LazarZ,DulermoR,RakickaM,HaddoucheR,NicaudJM.
AnalysisofATP-citratelyaseandmalicenzymemutantsofYarrowialipolyticapointsouttheimportanceofmannitolmetabolisminfattyacidsynthesis.
BiochimBiophysActa-MolCellBiolLipids.
ElsevierB.
V.
2015;1851:1107–17.
https://doi.
org/10.
1016/j.
bbalip.
2015.
04.
007.
32.
MostafaNa,MaherA,AbdelmoezW.
Productionofmono-,di-,andtriglyceridesfromwastefattyacidsthroughesterificationwithglycerol.
AdvBiosciBiotechnol.
2013;4:900–7.
http://dx.
doi.
org/10.
4236/abb.
2013.
4911833.
LunDS,RockwellG,GuidoNJ,BaymM,KelnerJA,BergerB,etal.
Large-scaleidentificationofgeneticdesignstrategiesusinglocalsearch.
MolSystBiol.
2009;5:296.
https://doi.
org/10.
1038/msb.
2009.
57.
34.
ChungBK,SelvarasuS,AndreaC,RyuJ,LeeH,AhnJ,etal.
Genome-scalemetabolicreconstructionandinsilicoanalysisofmethylotrophicyeastPichiaPastorisforstrainimprovement.
MicrobCellFactories.
2010;9:50.
https://doi.
org/10.
1186/1475-2859-9-50.
35.
KimM,YiJS,LakshmananM,LeeDY,KimBG.
Transcriptomics-basedstrainoptimizationtoolfordesigningsecondarymetaboliteoverproducingstrainsofStreptomycesCoelicolor.
BiotechnolBioeng.
2016;113:651–60.
https://doi.
org/10.
1002/bit.
25830.
36.
LakshmananM,ChungBK-S,LiuC,KimS-W,LeeD-Y.
Cofactormodificationanalysis:acomputationalframeworktoidentifycofactorspecificityengineeringtargetsforstrainimprovement.
JBioinformaComputBiol.
2013;11:1343006.
https://doi.
org/10.
1142/S0219720013430063.
37.
ZaslaverA,MayoAE,RosenbergR,BashkinP,SberroH,TsalyukM,etal.
Just-in-timetranscriptionprograminmetabolicpathways.
NatGenet.
2004;36:486–91.
https://doi.
org/10.
1038/ng1348.
38.
LittleJW,ShepleyDP,WertDW.
Robustnessofageneregulatorycircuit.
EMBOJ.
1999;18:4299–307.
https://doi.
org/10.
1093/emboj/18.
15.
4299.
39.
IwamaR,KobayashiS,OhtaA,HoriuchiH,FukudaR.
AlcoholdehydrogenasesandanalcoholoxidaseinvolvedintheassimilationofexogenousfattyalcoholsinYarrowialipolytica.
FEMSYeastRes.
2015;15:1–12.
https://doi.
org/10.
1093/femsyr/fov014.
40.
MorinN,CescutJ,BeopoulosA,LelandaisG,LeBerreV,UribelarreaJ-L,etal.
TranscriptomicanalysesduringthetransitionfrombiomassproductiontolipidaccumulationintheoleaginousyeastYarrowialipolytica.
PLoSOne.
2011;6:e27966.
https://doi.
org/10.
1371/journal.
pone.
0027966.
41.
RatledgeC.
TheroleofmalicenzymeastheproviderofNADPHinoleaginousmicroorganisms:areappraisalandunsolvedproblems.
BiotechnolLett.
2014;36:1557–68.
https://doi.
org/10.
1007/s10529-014-1532-3.
42.
MiflinBJ,HabashDZ.
Theroleofglutaminesynthetaseandglutamatedehydrogenaseinnitrogenassimilationandpossibilitiesforimprovementinthenitrogenutilizationofcrops.
JExpBot.
2002;53:979–87.
https://doi.
org/10.
1093/jexbot/53.
370.
979.
43.
QiaoK,ImamAbidiSH,LiuH,ZhangH,ChakrabortyS,WatsonN,etal.
EngineeringlipidoverproductionintheoleaginousyeastYarrowialipolytica.
MetabEng.
2015;29:56–65.
https://doi.
org/10.
1016/j.
ymben.
2015.
02.
005.
44.
GaoC,ZhangL,XieY,HuC,ZhangY,LiL,etal.
Productionof(3S)-acetoinfromdiacetylbyusingstereoselectiveNADPH-dependentcarbonylreductaseandglucosedehydrogenase.
BioresourTechnol.
2013;137:111–5.
https://doi.
org/10.
1016/j.
biortech.
2013.
02.
115.
45.
SunJA,ZhangLY,RaoB,ShenYL,WeiDZ.
EnhancedacetoinproductionbySerratiaMarcescensH32withexpressionofawater-formingNADHoxidase.
BioresourTechnol.
2012;119:94–8.
https://doi.
org/10.
1016/j.
biortech.
2012.
05.
108.
46.
TakenoS,MurataR,KobayashiR,MitsuhashiS,IkedaM.
EngineeringofCorynebacteriumglutamicumwithanNADPH-generatingglycolyticpathwayforL-lysineproduction.
ApplEnvironMicrobiol.
2010;76:7154–60.
https://doi.
org/10.
1128/AEM.
01464-10.
47.
VerhoR,LondesboroughJ,PenttilM,RichardP.
EngineeringredoxcofactorregenerationforimprovedpentosefermentationinSaccharomycesCerevisiae.
ApplEnvironMicrobiol.
2003;69:5892–7.
https://doi.
org/10.
1128/AEM.
69.
10.
5892-5897.
2003.
48.
ShiA,ZhuX,LuJ,ZhangX,MaY.
ActivatingtranshydrogenaseandNADkinaseincombinationforimprovingisobutanolproduction.
MetabEng.
2013;16:1–10.
https://doi.
org/10.
1016/j.
ymben.
2012.
11.
008.
49.
WasylenkoTM,AhnWS,StephanopoulosG.
TheoxidativepentosephosphatepathwayistheprimarysourceofNADPHforlipidoverproductionfromglucoseinYarrowialipolytica.
MetabEng.
2015;30:27–39.
https://doi.
org/10.
1016/j.
ymben.
2015.
02.
007.
50.
SeipJ,JacksonR,HeH,ZhuQ,HongSP.
Snf1isaregulatoroflipidaccumulationinYarrowialipolytica.
ApplEnvironMicrobiol.
2013;79:7360–70.
https://doi.
org/10.
1128/AEM.
02079-13.
51.
KamisakaY,TomitaN,KimuraK,KainouK,UemuraH.
DGA1(diacylglycerolacyltransferasegene)overexpressionandleucinebiosynthesissignificantlyincreaselipidaccumulationintheΔsnf2disruptantofSaccharomycesCerevisiae.
BiochemJ.
2007;408:61.
https://doi.
org/10.
1042/bj20070449.
52.
BlazeckJ,HillA,LiuL,KnightR,MillerJ,PanA,etal.
HarnessingYarrowialipolyticalipogenesistocreateaplatformforlipidandbiofuelproduction.
NatCommun.
NaturePublishingGroup.
2014;5:3131.
https://doi.
org/10.
1038/ncomms4131.
53.
SatishKumarV,DasikaMS,MaranasCD.
Optimizationbasedautomatedcurationofmetabolicreconstructions.
BMCBioinformatics.
2007;8:212.
https://doi.
org/10.
1186/1471-2105-8-212.
54.
OrthJD,ThieleI,PalssonB.
WhatisfluxbalanceanalysisNatBiotechnol.
2010;28:245–8.
https://doi.
org/10.
1038/nbt.
1614.
55.
OberhardtMA,PalssonB,PapinJA.
Applicationsofgenome-scalemetabolicreconstructions.
MolSystBiol.
2009;5:320.
https://doi.
org/10.
1038/msb.
2009.
77.
56.
LeeJM,GianchandaniEP,PapinJA.
Fluxbalanceanalysisintheeraofmetabolomics.
BriefBioinform.
2006:140–50.
https://doi.
org/10.
1093/bib/bbl007.
57.
SchellenbergerJ,QueR,FlemingRMT,ThieleI,OrthJD,FeistAM,etal.
Quantitativepredictionofcellularmetabolismwithconstraint-basedmodels:theCOBRAtoolboxv2.
0.
NatProtoc.
2011;6:1290–307.
https://doi.
org/10.
1038/nprot.
2011.
308.
58.
LakshmananM,YuK,KoduruL,LeeD-Y.
Insilicomodel-drivencofactorengineeringstrategiesforimprovingtheoverallNADP(H)turnoverinmicrobialcellfactories:JIndMicrobiolBiotechnol.
SpringerBerlinHeidelberg;2015.
https://doi.
org/10.
1007/s10295-015-1663-0.
59.
KimP-J,LeeD-Y,KimTY,LeeKH,JeongH,LeeSY,etal.
MetaboliteessentialityelucidatesrobustnessofEscherichiaColimetabolism.
ProcNatlAcadSciUSA.
2007;104:13638–42.
https://doi.
org/10.
1073/pnas.
0703262104.
Mishraetal.
BMCSystemsBiology2018,12(Suppl2):12Page20of130
- optimalcjblaze相关文档
- Sterncjblaze
- tioncjblaze
- contentcjblaze
- deterministiccjblaze
- activationcjblaze
- RESULTScjblaze
香港云服务器最便宜价格是多少钱一个月、一年?
香港云服务器最便宜价格是多少钱一个月/一年?无论香港云服务器推出什么类型的配置和活动，价格都会一直吸引我们，那么就来说说香港最便宜的云服务器类型和香港最低的云服务器价格吧。香港云服务器最便宜最低价的价格是多少?香港云服务器只是服务器中最受欢迎的产品。香港云服务器有多种配置类型，如1核1G、2核2G、2核4G、8到16核32G等。这些配置可以满足大多数用户的需求，无论是电商站、视频还是游戏、小说等。...
妮妮云（119元/季）日本CN2 2核2G 30M 119元/季
妮妮云的知名度应该也不用多介绍了，妮妮云旗下的云产品提供商，相比起他家其他的产品，云产品还是非常良心的，经常出了一些优惠活动，前段时间的八折活动推出了很多优质产品，近期商家秒杀活动又上线了，秒杀产品比较全面，除了ECS和轻量云，还有一些免费空间、增值代购、云数据库等，如果你是刚入行安稳做站的朋友，可以先入手一个119/元季付的ECS来起步，非常稳定。官网地址：www.niniyun.com活动专区...
云雀云（larkyun）低至368元/月，广州移动1Gbps带宽VDS（带100G防御），常州联通1Gbps带宽VDS
云雀云（larkyun）当前主要运作国内线路的机器，最大提供1Gbps服务器，有云服务器（VDS）、也有独立服务器，对接国内、国外的效果都是相当靠谱的。此外，还有台湾hinet线路的动态云服务器和静态云服务器。当前，larkyun对广州移动二期正在搞优惠促销！官方网站：https://larkyun.top付款方式：支付宝、微信、USDT广移二期开售8折折扣码：56NZVE0YZN （试用于常州联...
cjblaze为你推荐
-
手机网站空间手机登陆qq空间网址是什么?韩国虚拟主机韩国虚拟主机好还是香港的好虚拟主机评测网怎么选一台好的虚拟主机重庆虚拟主机重庆市邮政速递物流公司渝北分公司双龙揽投部客服电话windows虚拟主机在windows上怎么安装虚拟机shopex虚拟主机西部数码虚拟主机，适合做独立shopex或者echsop网店吗，我想开网店，推荐一下哪个型号的好www二级域名顶级域名，二级域名，网站域名邮箱域名是干什么的？域名邮箱和自己注册的邮箱有什么不一样吗？网络域名网络域名是指什么？ 买个域名需要多少钱啊?万网域名查询万网域名证书查询去那里啊？/