Holistic Approaches in Lipid Production by Yarrowia lipolytica

2018·2.7 MB·English

by  Zbigniew Lazar

Review
Holistic Approaches in Lipid Production by

Yarrowia lipolytica

Zbigniew Lazar,1,2,3,* Nian Liu,1,3 and Gregory Stephanopoulos1,*

Concerns about climate change have driven research on the production of Highlights

lipid-derived biofuels as an alternative and renewable liquid fuel source. Yarrowia lipolytica is a model oleagi-

Using oleaginous yeasts for lipid synthesis creates the potential for cost- nousyeastfortheproductionoflipids

and lipid-derived biofuels, studies of
effective industrial-scale operations due to their ability to reach high lipid
lipidmetabolism,andthebiosynthesis
titer,yield ,andproductivi tyresulting from th eirun iqueme ta bolism .Yarr owia of various ind ustr ially important
lipoly tica is the model oleag inous ye ast, w ith th e best- studied lipid metabo- m etabolites.

lism,thegreatestnumberofgenetictools,andafullysequencedgenome.In
Multiomicsmeasurementsandinsilico

this review we highlight multiomics studies that elucidate the mechanisms metabolic modeling for Y. lipolytica
deepens our under stan din g of the
allowing this yeast to achieve lipid overaccumulation and then present sev- organism ’s m etabolism and aid s in
eral major metabolic engineering efforts that enhanced the production met- identifying the limiting steps in li pid

rics in Y. lipolytica. Recent achievements that applied novel engineering biosynthesis.

strategies are emphasized.
A variety of metabolic engineering

attemptstoenhancelipidproduction

have been conducted, exploring the
Lipid Production and Its Metabolism in Yarrowia lipolytica overe xpres sionoffatty acidand tria-

Contemporary society relies heavily on fossil fuels (petroleum, coal, and natural gas) as the cylglyceride synthesis pathways and

thedeletionofantagonisticdegrada-
sourceofenergy[1].However,concernsaboutclimatechangehavepromptedresearchinto
tionpathways.
thedevelopmentofrenewableliquidfuels[2].Thesenewtechnologiesneedtosupplyfuelsina

cost-effectiveandsustainablemannerwhilecontributingtothereductionofgreenhousegases Morerecentstrategiesinvolveaholis-
[3].Biofuelsp rodu cedfromm icrobes, prim arilybioethan ol and biodiese l,a resuchprom ising ticun derstan dingoflim itingfa ct orsin

lipidbiosynthesis.Thesestudiestarget
alternatives (Box 1). In particular, lipid-derived biodiesels are garnering much attention due to the bottlenecks id entifie d in pre vious
theirhighenergydensity,whichmakesthemsuperiorsubstitutesfordieselfuelsandjetfuels om icsstudiesan dformula tem etabolic

compared with other forms of renewable energy. The cost-effective production of biodiesel engineeringstrategiesaccordingly.

relies on several important criteria, including lipid content (see Glossary), lipid titer, lipid

productivity,andlipidyield.Therefore,oleaginousorganisms,whichexcelataccumulating

intracellular lipids, are often chosen as the industrial workhorse. Y. lipolytica, an oleaginous

yeastbelongingtotheYarrowiaclade[4,5],iswidelyregardedasthemodelorganismforthis

purpose[6].Itslipidmetabolismandsupportingpathwayshasbeenstudiedextensively,ithasa 1DepartmentofChemicalEngineering,
plethora ofg en etic engineering tool s,anditsg enomehas be enful lyseque nced[7–14 ]. Mor e Massachuset ts Instituteo f
recently, th eubiquit ousgenome editin gtec hn iqueCRIS PR –Cas9 has alsobeend emonst rated Technology, 77 Massac husetts
Avenue,Cambridge,MA02139,USA
inY.lipolytica,allowinghigh-frequencyhomologousrecombinationaswellastargetedgene 2Departm entofBiote chn ologya nd

insertion and deletion [15,16]. These characteristics enhance the potential of Y. lipolytica in FoodMicrobiology,Wroclaw
Unive rsityofEnviro nmentalandLife
achieving economic biodieselproductionatan industrialscale.
Sciences,Chelmonskiego37,51-630

Wroclaw,Poland
ToengineerY.lipolyticaefficientlyforenhancedlipidaccumulation,athoroughunderstanding 3Theseau thorscontributedequally

ofitsmetabolismmustbedevelopedfirst.Denovolipidsynthesisrequirescellstocoordinate

variousbiochemicalpathwaysthatcanproducetriacylglycerides(TAGs)fromglucoseorother
*Correspondence:
small-moleculecarbonsubstratessuchassugars,organicacids,andalcohols.Thisprocessis [email protected](Z.Lazar)

activatedwhennitrogeninthemediumbecomesscarce,causingaseriesofregulatoryevents [email protected]
(G.S tephanopoulos).
to cascade. Initially, nitrogen limitation causes a rapid decrease in intracellular AMP levels due UR L:https://stephanopouloslab.org/.

TrendsinBiotechnology,November2018,Vol.36,No.11 https://doi.org/10.1016/j.tibtech.2018.06.007 1157

©2018ElsevierLtd.Allrightsreserved.
Box1.TheImportanceofMicrobialBiofuels
Glossary

Duetotheconcernsregardingclimatechangeandenvironmentalpollutioncausedbyextensiveuseoffossilfuels,there
Acetatekinase(ACK):catalyzes
hasbeenagrowingdemandforalternativeformsofrenewableenergy.Biofuelsproducedfromengineeredmicrobes
theformationofacetylphosphate
(FigureI)representapromisingexample.
fromacetate.
Acet yl-CoAcarboxylase(ACC):
First-GenerationBiofuels carboxylates acetyl-CoAto malonyl-

CoA.
Theseareprimarilybioethanolproducedbymicrobialfermentationofsugarsobtainedfromfoodcropssuchascornand Acetyl-CoAsynthetase(ACS):

sugarcane. Commonly used organisms include Saccharomyces cerevisiae and Zymomonas mobilis [27,28]. In catalyzestheproductionofacetyl-
addition,bi odieselprod uced fromthetra nsesteri ficationofFAsfa llintothisc ateg ory.TheFAs areofte nextract ed CoAfrom ace tate.
fromveg etableoils andanima lfats [1]. Nearlyall-firstgen era tion biof uels utiliz esubstrat esth atar ede rived fromfood AMP dea minase1(AMPD1):

crops,causingamajorcompetitionwithfoodsupply. convertsAMPintoinosine

monophosphate.
Second-Generation Biofuels and Biodiesel tAhTePfo-crmitraatitoen lyoafsaec e(tAylC-CLo):A caatnadlyzes

oxaloacetatefromcitrate.
Toavoid such an issue,the industry hasnow shifted its focustowards second-generation biofuel production from nonedible DGA1,DGA 2:DA Gacyltransferases;
feedstockssuchasplantbiomass,forestryresidues,energycrops(e.g.,lignocellulosics),wasteoils,crudeglycerol,
catalyzetheterminalstepofTAG
municipalwastes,andgaseoussubstrates[77].Theendfuelproductremainsthesamecomparedwiththeprevious
formation.
generation. One major avenue of research in this area involves the synthesis of microbial biolipids as a substrate for biodiesel FAA1:long-chainFA-CoAligase1;
pdroowdnuscttrieoanm. Tihneto lipFAidsm aerteh yalcecsutemrsul(aFtAedM eEisth)e(i.re d.,eb nioodvioe soerl )euxs ninogvoe itbhye roalecaidgi-noorubs amseic-rcoaotarglyazneisdmtrsa nasneds ctearnifi cbaet icoonn.vTehretesde esterifie s long-cha in FAs in to acti ve
CoAthioesters.
FmAoMdiEficmaotiolencubleefso erexhuisbeitc[6h8e]m.Ficuartlhaenrdmpohreys,ibciaold pireospeelhrtaiessasihmigilharetnoethrgoysed eonfcsiotyn,vceonmtiopnaarladbielesweliathndotdhoernfootrrmeqsuoirfereenxetewnasbivlee FAS 1, FAS2: FA synthases;
synthesizepalmitatefromacetyl-CoA
energy,andisbiodegradable,nontoxic,andessentiallyfreeofsulfurandaromaticcomponents[69,70].
andmalonyl-CoAinthepresenceof

NADPH.
OleaginousOrganisms Fattyaciddesaturase2(FAD2):

requiredforlinoleicacidsynthesis.
Theseorganismsareclassifiedasbacteria,yeast,ormicroalgaecapableofaccumulatingmorethan20%oftheircelldry GapC:g lyc eraldehy de3 -phosphate

massaslipids.Oleaginousyeastsarethebeststudiedduetotheirfastgrowthrateandabilitytoreachhighbiomass dehydrogenase;catalyzesoxidative

densities, which is crucial for fast and abundant lipid production. Examples of oleaginous yeast species include phosphorylationofglyceraldehyde3-

Rhodosporidium,Lipomyces,Candida,andYarrowia[11,71,72].Typicallyinoleaginousyeasts,lipidsaremostlystored phosphateto1,3-

intheformoftriacylglycerolsinspecialorganellescalledLBs. bisphosphoglycerateusingthe
cofactorNADP+.
GPD1:N AD-dependentG3P

dehydrogenase;keyenzymein

glycerolsynthesis.
GSY1:g lycogensynthase;transfers

theglycosylresiduefromUDP-Glcto

alpha-1,4-glucan.
Bioethanol Biodiesel Lipidcontent:gramsoflipidper

gramCDW.

Lipidproductivity:gramsoflipid

perliterofculturevolumeperhour.
Waste oils Lignocellulose
Lipidtiter:gramsoflipidperliterof
Sucrose Starch
culturevolume.

Lipidyield:gramsoflipidpergram
First- Second-
ofcarbonsubstrateconsumed.
genera(cid:2)on genera(cid:2)on LR O1:tria cylglycero lformation

biofuels biofuels enzyme; transfers acyl groups from
thesn-2positionofaphospholipid
CO + H2 toD AG.
M AE:mitochondrialmalicenzyme;
Municipal Syngas catalyz esoxidatived ecarb oxylation
Oils
wastes ofmalatetopyruvatewhileproviding

NADPH.
MCE2:cytosolicNADP+-dependent

malicenzyme.
Figure I. A Comparison between First- and Second-Generation Biofuels. Meta bolicfluxanalysis(MFA):an
experiment alflu xomicste chnique

usedtodeterminetheproduction

1158 TrendsinBiotechnology,November2018,Vol.36,No.11
toAMPdeaminase1(AMPD1)recyclingtheaminegroup.LowAMPconcentrationcauses andconsumptionratesof

theinhibitionofisocitratedehydrogenase(IDH).Thisleadstodownregulationofthetricarbox- metabolites.
MFE2:peroxisomalmultifunctional
ylicacid(TCA)cycle,accumulationofcitrateinthemitochondria,andexportofexcesscitrate
enzymetype2;actsonthebeta-
from themitochondria to thecytosol.In the next step, citricacidiscleaved byATP-citrate
oxidationpathway.
lyase (AC L)intocytoso lic ace tyl-CoAa nd ox aloac etate. Overa ll,th isp athway, dis tinctfromthe MIG1:tra nscriptionfactorinvolvedin

pyruvate–acetaldehyde–acetate pathway used by conventional yeasts, is characteristic of glucoserepression;regulatedbythe

SNF1kinase.
oleaginousorganisms,anditallowstheformationofcytosolicacetyl-CoA,whichisthestarting
Multiomics:anewapproachin
materialfor lipidsynthesis. biologicalsys te msc ombining data

fromgenomics,transcriptomics,

Oncecytosolicacetyl-CoAisformed,itcanbeusedasatwo-carbonbuildingblockforfattyacid proteomics, lipidomics, and
metabolomics;revealstheholistic
(FA)synthesis.TheelongationofFAsbeginswiththeactionofacetyl-CoAcarboxylase(ACC),
cata lyzingthet rans formationo f acet yl-CoA into mal onyl-C oA .TheFAsyn thasecomplex (FAS1 OpicLtEu1re: Dof9 thFeA idnevessattiugraatseed; sreyqstueirmed.

andFAS2)thenactsonmalonyl-CoAtobuildC16acyl-ACPs,whicharethentransportedintothe formonounsaturatedFAsynthesis.
endo plasm ic ret iculu m ( ER) for further elo ngat ion a nd desatura tion [1 7–1 9]. De saturases in the ER Ptrea rnosxfiesroamseal( pcearrCnAitiT n2e) : atcraentsyfle-CrsoA
producepalmitoleic(16:1)andoleic(C18:1)acidsduetoOLE1activityorlinoleicacid(18:2)dueto
FAdesa turase2( FAD2 )act ivity. Theove rallst oich iom etryo fFAsy nt hesisis athceentybl egrsohuuptstl etod caacrrnoistisnem, ewmhbicrha nceasn.
n n PEX 3,P EX10,P EX11: peroxisomal
AcCoAþ ATPþðn(cid:2)2þmÞNADPH!Cðn:mÞFA; [1]
2 2 membrane proteins required for
peroxisomebiogenesis.
wherenandmarethechainlengthanddegreeofunsaturation,respectively.Thisprocessrelies POT1:3-ke toacyl-CoAthiolase;

heavily on theavailability ofATPand NADPH. cleavesacetyl-CoAfrom3-ketoacyl-

CoAduringbeta-oxidation.
POX 1–6:fa ttyacyl-CoAoxidases;
ToformTAGs,threeFAscondensewithoneglycerol3-phosphate(G3P)throughtheKennedy
involvedinthebeta-oxidation
pathway[9].Initially,G3PisacylatedbyG3Pacyltransferase(SCT1)toformlysophosphatidicacid
pathway.

(LPA).LPAissubsequentlyacylatedbyLPAacyltransferase(SLC1),whichproducesphospha- SCT1:G3P/dihydroxyacetone
tidicac id(P A) .PAisthende phosphor yla tedb yPAphosphata se(PAP )andd iacylglycer ol(DAG)is phosph atesn-1acyltransferase.
relea sed. Fina lly,T A Gsa resynthesizedeith er by DAGacyltrans ferase (DG A1orDGA2 ),whic h SLC1: 1-ac yl-sn - G3P
acyltransferase;catalyzesthe
usesacyl-CoAasthefinalacylgroupdonor,orbyphospholipidDAGacyltransferase(LRO1), acylationofLPA toformP A.
which utilizesg lyce rop hosp holip idsas theacy lgr oup donor.Thes ereac tionsoccurbetw eenthe SNF1:AM P -acti vat edS/ Tprotein
ERan d theli pid body(LB)surface , w her ethe relev ant enz ymes havebeen locat ed[20].T he kinase; acts as a regu lator y protein
forglucose-repressedtranscription,
pathwaysinvolvedinlipidsynthesisaresummarizedinFigure1.
heat shock, sporulation, and
peroxisomebiogenesis.
Inadditiontosynthesispathways,lipiddegradationpathwaysarealsorelevantinengineeringY. Stearyl-Co Adesaturase(SCD):
lip olytica fo r li pid overpr oduction a s the y often beco me targets for dele tion. Fre e F As (FFAs) ca n Dde9spatousriatitoens. fa tty acyl-CoA s at the
bbee raecletaivsaetedd thtroouagchy lt-hCeo aAcstiothnr oofu tghhe itnhteraceenlzluylmare lipaacsyel- CTGoAL4s [y1n1t,h2e1ta].s Tehe(F rAelAea1s)etdo FbFeAsfu mrthuesrt iSnUve Crt2a:s Searcecsphaornosmibylecefos rcseurcervoissiaee

processedbiologically[22].Thedegradationoftheselong-chainacyl-CoAsbybeta-oxidation hydrolysisintoglucoseandfructose.
TGL3:tria cylg lycerollip ase; in
(Figure2)occursintheperoxisome.Therearefourreactionsoccurringinacyclicmannerin
Yarrowialipolyticaservesasa
beta-oxidation where the combined effect is to release an acetyl-CoA molecule, thereby
regulatoryproteinforTGL4.
shortening the acyl chai n by two carbon u nit s. In Y. lip olytica six a cyl-CoA o xidases TGL4:lipa se4;re lea sesFAsfrom

(POX1–6),whichcatalyzethefirstreaction,wereidentified,eachhavingdifferentchain-length TAGs.
XPKA:xylulose5-phosphate
preferences [11,23,24]. The second and third reactions are catalyzed by a multifunctional
phosphoketolase;formsacetyl
enzyme (MFE2) and the last reaction is catalyzed by 3-ketoacyl-CoA thiolase (POT1) phosphatefromx ylulose 5-

[11,25,26]. Theseenzymesareoften thetargets fordeletion. phosphate.
ylYEF:ATP-NADHkinasewith

phosphorylationactivityofboth
Understanding the Biology of Lipid Synthesis through Multiomics NADHandNAD +topro du ceNADPH
Approaches andNA DP+ .

TofurtherexpandontheknowledgeofY.lipolyticalipidmetabolismandtoidentifybottleneck-

ing locations in the metabolic network, quantitative biological measurements have become

increasinglypopularintheinvestigationofthisorganism.Thesestudiesoftenguidesubsequent

TrendsinBiotechnology,November2018,Vol.36,No.11 1159
Glucose Xylose
Biomass
Xyl Xyli Xylu
Glucose
Xu5P S7P E4P
G6P 6PG Ru5P
R5P GAP F6P
F6P Biomass
FBP DHAP G3P LPA PA DAG TAG
GAP
Endoplasmic
Acyl-CoA
Glycerol re(cid:415)culum
BPG
Glycerol Lipid
TAG
3PG Biomass body
2PG Peroxisome
Acetate FA
PEP
Acetate
Acyl-CoA
Pyr OAA
Biomass MalCoA
Pyr Mal Mal
AcCoA OAA AcCoA
Cit Cit
OAA
Suc Glx
iCit
Mal
AcCoA
αKG Biomass
Fum
Suc
Acetate
Mitochondrion
Cytosol
Acetate
Figure1.OverviewofYarrowialipolyticaLipidSynthesisMetabolism.Differentlycoloredarrowsareusedto

represent different metabolic pathways: green, substrate incorporation reactions; blue, glycolysis/gluconeogenesis;

orange,tricarboxylicacid(TCA)cycleandrelatedanapleroticreactions;gray,pentosephosphatepathway;darkblue,

glyoxylateshunt; yellow, triacylglyceride(TAG)synthesis reactions;red,beta-oxidation pathway.Various intracellular

organelles (mitochondrion, peroxisome, endoplasmic reticulum, and lipid body) and their relations to the metabolic

pathwaysarealsodepicted.

1160 TrendsinBiotechnology,November2018,Vol.36,No.11
O
CoA
O
S O2
CoA
R S
CoASH HO
2 2
POT1 POX1–6
O
O O
CoA
CoA R S
R S
MFE2 MFE2 HO
2
NADH
OH O
NAD+ CoA
R S
Figure2.Beta-OxidationPathway,WhichDegradesFattyAcidsintoAcetyl-CoAinaCyclicManner.

engineeringstrategiesinachievingthebestresults.Wasylenkoandcolleaguesperformed13C

metabolic flux analysis (MFA)on Y.lipolytica cultured on glucoseand comparedthe flux

distributionbetweenawild-typestrainandanengineeredstrain[12].Theauthorsconcluded

that the availability of NADPH is limiting lipid synthesis. By comparing how fluxes change

betweenthetwostrains,theywerealsoabletodeducethatthelipogenicNADPHissynthe-

sizedalmostexclusivelyfromtheoxidativepentosephosphatepathway(PPP).Thiscorrobo-

ratedpreviousgenomicstudiesshowingthatY.lipolyticadidnothaveacytosoliccopyofmalic

enzyme [29], which makes this yeast unique compared with other oleaginous organisms. A

similarstudyanalyzedthefluxdistributionofY.lipolyticaculturedonacetate[13].Itdemon-

strated the importance of gluconeogenesis in supporting both biomass precursor synthesis

and lipogenic NADPH synthesis. A regulatory node that controls the flux through gluconeo-

genesiswas alsoidentified.

Many other analyses have also elucidated important mechanisms of lipid metabolism in Y.

lipolytica.Theapproachesappliedincludegenome-scalemodelreconstructionandanalysisas

well as transcriptomic, metabolomic, lipidomic, and proteomic measurements [29–34]. A

transcriptomicanalysisperformedbyMorinandcolleaguesdemonstratedthattheexpression

levelsofmanygenesdiffersignificantlydependingonwhetherY.lipolyticawasactivelydividing

or synthesizing lipids [29]. Furthermore, they concluded that lipid accumulation could be a

consequence of passive carbon flux rerouting to cytosolic acetyl-CoA as opposed to being

controlledatthetranscriptomiclevel.Liuandcolleagues[31]observedasimilarphenomenon

whereexcesscarbondirectedtolipidbiosynthesiscouldbeaconsequenceofglycolysisand

TCA cycle activity imbalance. However, genes related to protein synthesis, including ones

encoding ribosomal subunits and translation initiation and elongation factors, were actively

downregulated [31]. This observation was confirmed by multiomics analysis of Y. lipolytica

strains exhibiting a lipid-overproducing phenotype under carbon- and nitrogen-limiting con-

ditions [30]. In particular, lipid accumulation did not involve transcriptional regulation of lipid

TrendsinBiotechnology,November2018,Vol.36,No.11 1161
metabolismbutwasclearlyassociatedwiththeregulationofaminoacidbiosynthesis.Genes

related to amino acid metabolism were downregulated, whereas those involved in protein

turnoverandautophagy,whichprovidesalternativesuppliesofnitrogen,wereoverexpressed.

Anespeciallyimportantfactorthataffectslipidsynthesisappearedtobeleucinemetabolism

[34].Interestingly,theauthorsprovidedevidencepointingtowardsseveralspecificregulations

that are associated only with nitrogen limitation in combination with DGA1 overexpression,

demonstratinghowmetabolicengineeringofthecellscouldaltergeneregulation.Finallyusing

proteomic methods, Pomraning and colleagues [33] showed that ACL, ACC, and lecithin

cholesterolacyltransferasearephosphorylatedduringnitrogenlimitation[33].Thisobservation

suggested the importance of post-translational modification in the regulation of lipid

accumulation.
Enhancing Lipid Production Metrics through Metabolic Engineering

Wild-typestrainsofY.lipolyticaarenotthemostefficientproducersoflipids.Forinstance,

Rhodosporidium toruloides can naturally accommodate 70% lipid content in the cell,

whereasY.lipolyticatypicallyachieves20–40%natively[35–37].Nevertheless,theadvan-

tage of using Y. lipolytica for lipid production lies in the ability to alter its metabolism

[14,16,38]. Studies focused on improving lipid production in Y. lipolytica have expanded

considerably in the past several years and various research groups have applied different

engineering strategies (Box 2).

Lipid degradation, as an antagonistic pathway to lipid synthesis, hinders lipid accumulation,

and hence related genes have become the main targets for deletion [11,24]. For instance,

MFE2 deletion was performed in engineered strains and yielded a lipid-overaccumulation

phenotype[26,39,40].Similarly,genesencodingproteinsresponsibleforperoxisomebiogen-

esis,PEX3,PEX10,andPEX11,havebeendeletedtoabolishbeta-oxidationactivityentirely

[41,42].Thedeletionofcellularmetabolismregulatorscanalsobenefitlipidproduction[43,44].

As an example, Y. lipolytica strains with an SNF1 deletion accumulated FAs constitutively,

reaching amounts that are 2.6-times higher than the wild type. Similar improvements were

observed with the deletion of SNF1 in a Y. lipolytica strain engineered to produce omega-3

eicosapentaenoicacid.Thesestrainsshoweda52%increaseinEPAtiters[7.6%ofcelldry

weight (CDW)] over the control [43]. Despite this success, the mechanism of how SNF1

regulatesthelipidsynthesispathwayhasnotbeenresolved.Asimilarobservationwasfound

forstrainswithMIG1genedisruption[44].ThemutantcellsexhibitedmoreLBsthanitsparental

strainasthelipidcontentincreasedto48.7%.TheMIG1disruptionwashypothesizedtohave

downregulatedMFE2andupregulatedothergenesrelevanttolipidbiosynthesis.Finally,based

ontheassumptionthatglycogenandTAGsynthesiscompetewitheachothersincetheyare

the two major carbon storage units, Bhutada and colleagues performed null mutations on

glycogen synthase (GSY1), achieving 60% higher amounts of TAGs compared with the

wild-type strain [45].

Althoughtheworklistedabovedemonstratedthesuccessoftargetedgenedeletionforlipid

overaccumulation,performingknockoutsinY.lipolyticaischallengingasthisyeastpreferen-

tiallyusesthenonhomologousend-joining(NHEJ)mechanismforDNArepairoverhomologous

recombination(HR)[46].Bycontrast,geneoverexpressioncanbeconductedreadily.Asimple

butefficientmethodforimprovinglipidproductioninY.lipolyticainvolvesthe‘push-and-pull’

strategybasedontheoverexpressionofACC1andDGA1[47].Thecombinedoverexpression

ofthetwogenesincreasedlipidproductionsignificantly,reaching61.7%lipidcontent,withthe

overall yield and productivity from glucose being 0.195g/g and 0.143g/l/h, respectively, in

bioreactors.
1162 TrendsinBiotechnology,November2018,Vol.36,No.11
Box2.DifferentApproachestoImproveLipidBiosynthesisinYarrowialipolytica

Lipidbiosynthesiscanbeimprovedinseveraldifferentways(FigureI).Eachusesauniqueapproachbuttheyendwiththesameultimategoalofmaximizing

intracellularlipidaccumulation.

(i) Thetraditionalmethodofresearchinvolveswild-typestrainsandoptimizationofmediumcompositionaswellasprocessparameterssuchastemperature,pH,

andaeration.
(ii) Sho rtlyafterthedevelopmentofgenetictools,metabolicengineeringhasemergedasameanstoalteranorganism’sfunctionforvariousneeds,including

improvedlipidsynthesis.Thistakesintoconsiderationtheknowngenesandenzymesfromthelipidbiosynthesisandrelatedpathways,changingtheir

expressionlevelsthroughoverexpressionordeletionandmodulatingtheenzymaticproperties.

(iii) Now,theavailabilityofsequencedandannotatedgenomes(genomics)combinedwithmethodsforintra-andextracellularmetabolitedetection(metab-
olomi cs), analysisof en tiresetsofp rotei ns(proteom ics),and quantificatio nofthetra nscri ptome(tra nsc riptom ics) hasallowed scientiststo studyind etailthe

lipidmetabolismofY.lipolyticaaswellasitsregulation.Thecombinationofdifferenttechniquesthatcomplementoneanothergaverisetoanewdiscipline

calledsystemsbiology.

(iv) Insilicometabolicmodelinginvolvingthereconstructionofgenome-scalemodelswithcertainsoftwareallowedthepredictionofbottleneckingstepswithout

performinglarge-scalewet-laboratoryexperiments.Thesemethods,whenproperlyimplemented,canoftenleadtohighlyaccuratepredictionsofhowcertain
metabolicm odification scanchangea norganism’s pheno type.Henc e,res earchers oftenemploy mod elingt ogu ide engin eering.

(v) Byapplyinginsilicomodelingandsystemsbiologyalongwithmetabolicengineering,newstrainswithhigherlipidsynthesispotentialcanbeconstructed.Anotherroundof

processoptimizationwillthenneedtobeconductedtotailorthefermentationtothenewstrain.Severaliterationsmayberequiredtoachievethefullcapacityofthesystem.
(vi) CurrentlythedevelopmentofsyntheticbiologyhasgreatlyadvancedtheengineeringofY.lipolyticaandotherbiologicalsystems.Drawingprinciplesfromother
engineeri ngfi elds,researc he rscanno wbuild arti ficialan dtunable bio logicalcircu its ,w hichsign ifica ntlyfa cilitateth echanne lingofca rbonfluxe sint othe

desiredlipidandlipid-basedmoleculebiosynthesispathways.

Classical metabolic
Mul(cid:415)omics study
engineering
Δ TEF
Yarrowia lipoly(cid:415)ca
Feed Effluent
Process op(cid:415)miza(cid:415)on
Metabolic modeling
Bo(cid:425)leneck iden(cid:415)fica(cid:415)on
FigureI.SchematicRepresentationoftheDifferentWaystoBoostLipidBiosynthesisinYarrowialipolytica.

TrendsinBiotechnology,November2018,Vol.36,No.11 1163
ThecombinationofgeneoverexpressionanddeletiontoheavilymodifyY.lipolyticalipid-related

metabolismprevailedintheearlierstagesofstraindevelopment[48,49].Lazarandcolleagues

combinedthedeletionofPOX1–6andTGL4withtheoverexpressionofGDP1andDGA2[39].

The resulting strain had diminishing FA degradation flux and increased intracellular G3P

concentrationsforenhancedTAGsynthesis.Thesemodificationswerefurtherexpandedwith

the overexpression of hexokinase for improved fructose metabolism and the introduction of

invertaseforbettersucroseutilization[48].Theresultingstrainproduced9.15g/loflipids,witha

lipidcontentof26%fromsucroseasthesolecarbonsourceafter96hculturetime.Thesame

strainwasalsodemonstratedtosuccessfullyproducelipidsfrommolassesandcrudeglycerol

[49]. Thehighest lipidcontent,40%of CDW,wasobtainedwithavolumetricproductivityof

0.43g/l/hduringcontinuousculturewithglycerol.Gajdošandcolleaguesengineeredstrains

usingasimilarstrategy[50].WhenmultiplecopiesofthenativeY.lipolyticaDGA2geneunder

the constitutive TEF1 promoter were introduced into a strain where four acyltransferases

(Ddga1, Ddga2, Dlro1, and Dare1) were deleted, lipid accumulation was improved. The

modificationswerethencombinedwithMFE2deletionandSUC2overexpressionforsucrose

utilization.Bothstrainsaccumulatedmorethan40%oflipidscontentwisefromsucrose,which

was a twofold increase compared with the wild type. A more complex study performed by

Blazeckandcolleaguescoupledcombinatorialmultiplexingoflipogenesistargetswithpheno-

typicinduction[26].TheauthorsanalyzedoverexpressionofAMPD,ACL,andMAEfortheir

potential to increase acetyl-CoA and NADPH supply. These modifications were multiplexed

withoverexpressionofDGA1andDGA2anddeletionsofMFE2andPEX10.Aftercharacteri-

zation,thestrainwithDGA1overexpressionandMFE2andPEX10deletionwasdeterminedto

bemostefficient.ThisengineeredY.lipolyticastrainincreaseddenovolipidaccumulationby

>60-foldinlipidtiter(25g/l)comparedwiththeparentalstrain,withlipidcontentapproaching

90%.
HeterologousgeneexpressioninY.lipolyticahasalsobeenlargelysuccessfulinengineering

the lipid-overproducing phenotype. For instance, the DGA1-encoded enzyme from R. toru-

loides and DGA2-encoded enzyme from Claviceps purpurea were found to outperform the

nativeonesofY.lipolytica[51].Overexpressionoftheseheterologousgeneswascombined

with the deletion of TGL3, which encodes the regulatory protein for TAG lipase [21]. These

combinedmodificationsresultedinalipidcontentof77%andayieldof0.21g/gduringbatch

culture[51].Applyingfed-batchmodewithglucoseastheonlysubstrateallowedthestrainto

produce85g/loflipidsataproductivityof0.73g/l/h.Similarly,overexpressionofheterologous

ACLfromMusmusculus,withalowerK valueforitssubstrate,wasshowntoimproveTAG
M

accumulation[26,52].Theauthorsreportedanincreaseinlipidcontentfrom7.3%to23.1%

resultingfromthismodification,indicatingtheimportanceofimprovingthesupplyofcytosolic

acetyl-CoA forlipid synthesis[26,52].

Finally, there have been numerous studies focusing on expanding the substrate range of Y.

lipolytica.Sinceacomprehensivereviewonthistopichasbeencoveredpreviously[53],weonly

touch on two new studies that demonstrated lipid synthesis on inexpensive and renewable

feedstocks. The first study, by Niehus and colleagues [54], showed robust growth of highly

engineeredY.lipolyticastrainsonlignocellulosichydrolysatescontainingglucoseandxylose.

Theauthorsexpressedphosphoketolase(XPKA)andacetatekinase(ACK)alongwiththe

enzymesrequiredforxyloseassimilation(xylosereductase,xylitoldehydrogenase,andxylulose

kinase) forthe ultimate conversionof xyloseinto acetate, whichcanbe furtherconverted to

lipogenicacetyl-CoAbyacetyl-CoAsynthetase(ACS).Combiningthesegeneticmodifica-

tionswiththosementionedabove,theyreachedalipidtiterof16.5g/lwiththemaximumlipid

contentof67%fromlignocellulosicbiomasshydrolysatesalone.Thesecondstudy,performed

1164 TrendsinBiotechnology,November2018,Vol.36,No.11
byXuandcolleagues,demonstratedefficientlipidproductionfromaceticacidusingasemi-

continuous, dynamically controlled bioreactor scheme with cell recycle [55]. They achieved

simultaneousconcentrationandconversionoflow-strengthaceticacid(<30g/l)intointracel-

lular lipids. Exceptional results were reported where 115g/l of lipids was produced after

continuous feeding of 30g/l acetic acid for 144h. These results illustrate how Y. lipolytica

can play an important role in upgrading the carbon sources from dilute waste streams to

concentrated,value-addedproductsbiologically.Table1(KeyTable)showsasummaryofallof

the studies discussedso far.

Novel Strategies to Maximize the Lipid Production Potential of Y. lipolytica

Inadditiontothemodulationofgenetargetslistedabove,anumberofnovelstrategieshave

emerged in the past several years [56–58]. These studies emphasize the importance of

understandingthelimitingfactorsoflipidproductionandhavebuiltonconclusionsformulated

in previous omics studies. The lipid production metrics from these studies are among the

highest inthe field,indicatingthesuccessof these approaches.

Inthe workbyQiao andcolleagues,thepotentialallostericinhibitionofACC1bysaturated

FAswastakenintoconsideration[56,59–61].Thisissuewasovercomethroughtheconver-

sion of saturated to monounsaturated FAs by the expression of stearoyl-CoA desaturase

(SCD), which is a central metabolic regulatory enzyme that catalyzes D9-desaturation of

palmitoyl-CoAandstearoyl-CoAtopalmitoleoyl-CoAandoleoyl-CoA,respectively[62].The

authors observed significant phenotypical changes after introducing this enzyme into the

previouslymentioned lipid-overproducingstrainwithACC1andDGA1overexpression[56].

Thelevelofoleicacid(C18:1)increasedto71%andpalmitoleicacidreached8%ofthetotal

FApool.Theconcentrationofneutrallipidsattheendofabioreactorculturereached55g/l

with the overall yield of 0.234g/g from glucose and productivity of 0.707g/l/h. The yield

achievedwas84.7%ofthetheoreticalmaximum.Theauthorshypothesizedthattheactionof

SCD released theinhibitory effect on ACC1 caused bypalmitoyl-CoA and stearoyl-CoAby

pushing these acids to monounsaturated versions. The monounsaturated acyl-CoAs were

thenrapidlyincorporatedintoTAGthroughtheactionofDGA1withoutfurtherinhibitionofthe

desaturase itself.

Ledesma-AmaroandcolleaguesfocusedonthesecretionofintracellularFAsintotheculture

mediumfollowedbyinsituextractionusinganorganiclayer[63].Theyengineeredtwounique

strategies to achieve this. The first involved the deletion of several genes related to free FA

activation and degradation, which were FAA1 and MFE2, respectively. As a result, the cells

couldnolongerconverttheimmobilizedFFAsintoacyl-CoAsorutilizetheminbeta-oxidation,

leadingtoFFAaccumulationandsubsequentsecretion.Thesecondapproachreportedinthe

study aimed at abolishing LB formation entirely and redirecting FA synthesis to the cytosol,

therebymimickingthebacterialpathway.Theauthorsconstructedastrainwithdeletionsinfour

genes(Dare1,Ddga1,Ddga2,andDlro1)thatexhibitedthisphenotypeand,combinedwith

deletionsinFAA1andMFE2,thisresultedinsignificantlymoreFFAsecretion.Usingtheirbest-

performingstrainandoptimizedreactorconditions,anextracellularFAconcentrationof10.4g/

lwasachieved,whichamountedtoatotalof1.2goflipidsynthesizedpergramCDW,well

exceedingthe storagelimitof individualcells.

Inaseparatestudy,Xuandcolleaguesfocusedonliftingthebottleneckoflipogenicacetyl-CoA

availability[57].TheynotedthattheactionofACLprovideslargeamountsofcytosolicacetyl-

CoA only when nitrogen becomes limited. However, a nitrogen source is required at the

beginning of the culture to sustain high biomass production before lipid accumulation. As a

TrendsinBiotechnology,November2018,Vol.36,No.11 1165
KeyTable

Table 1. Summary of the Lipid Production Metrics and Strain Genotypes in Various Studies

Genotype Lipidtiter(g/l) Lipidproductivity(g/l/h) Lipidyield(g/g) Lipidcontent(%) Substrate Refs

FlaskCultures

Y4184(ATCC20362):Dsnf1 1.10 – – 30.0 Glucose [43]

PO1h:mmACL 1.70 – – 23.1 Glycerol [52]

ACA-DC50109:Dmig1 2.44 – – 48.7 Glucose [44]

PO1d:Dtgl4,Dgsy1,DGA2,GPD1 2.62 – 52.4 Glycerol [45]

PO1d:Dpox1–6,Dtgl4,DGA2, 2.84 – 0.130 – Starch [73]

GPD1,(riceAlphaAmylase,

glucoamylase)x2
PO1d:Ddga1Dlro1Dare1Ddga2, 6.70 – – 49.0 Sucrose [50]
Dmfe2 ,DGA2x3,scSUC2

BioreactorCultures

PO1d:GAL1,7,10E,10M 3.22 – 0.056 16.6 Galactose [74]

PO1d:Dpox1–6,Dtgl4,DGA2, 9.15 – 0.063 26.0 Sucrose [48]

GPD1,HXK1,scSUC2

PO1d:Dfaa1Dmfe1DGA2TLG4 10.40 – 0.200 120.4 Glucose [63]

klTGL3
E26(Dpex10,Dmfe1,DGA1, 15.00 0.190 – – Xylose [75]

evolved):ssXYL1,ssXYL2,starved

PO1d:Dpox1–6,Dtgl4,DGA2, 16.50 0.185 0.344 67.0 Glucose/xylose [54]

GPD1,XPKA,ACK,XK,XDH,XR

PO1g:ACC1,DGA1 17.60 0.143 0.195 61.7 Glucose [47]

PO1d:Dpox1–6,Dtgl4,DGA2, 23.82 0.158 0.160 43.0 Inulin [40]

GPD1,HXK1,GAL1,7,10E,10M,

scSUC2,kmINU1

PO1d:Dpox1–6,Dtgl4,DGA2, 24.20 0.430 0.100 40.0 Molasses/glycerol [49]

GPD1,HXK1,scSUC2

PO1f:Dpex10,Dmfe2,DGA1 25.00 0.210 – 71.0 Glucose [26]

E26(Dpex10,Dmfe1,DGA1, 25.00 0.145 0.213 – Glucose [31]
evol ved):Dmg a2

PO1d:Dpox1–6,Dtgl4,GDP1, 50.50 – 0.120 42.0 Xylose/glycerol [76]

DGA2,ssXR,ssXDH,ylXK

PO1g:ACC1,DGA1,SCD 55.00 0.707 0.234 67.0 Glucose [56]

PO1g:ACC1,DGA1,scCAT2 66.40 0.565 0.229 70.0 Glucose [57]

NS18:rtDGA1,cpDGA2,Dtgl3 85.00 0.730 0.200 73.0 Glucose [51]

PO1g:ACC1,DGA1,caGapC, 99.00 1.200 0.270 66.8 Glucose [58]

mcMCE2
PO1g:ACC1,DGA1 115.00 0.800 0.160 59.0 Aceticacid [55]

result,Y.lipolyticacultivationisgenerallybiphasic,requiringagrowthphase(nitrogenpresent)

andalipidproductionphase(nitrogendepleted),whichprolongsthefermentation.Todecouple

the onset of lipid synthesis from nitrogen availability, the engineering of other pathways that

providecytosolicacetyl-CoAevenwhennitrogenispresentisessential[57].Themostefficient

1166 TrendsinBiotechnology,November2018,Vol.36,No.11