Appl Microbiol Biotechnol (2002) 59:436–442
DOI 10.1007/s00253-002-1056-y
ORIGINAL PAPER
J.Zaldivar· A.Borges· B.Johansson· H.P.Smits
S.G.Villas-Bôas· J.Nielsen· L.Olsson
Fermentation performance andintracellular metabolite patterns inlaboratory andindustrial xylose-fermentingSaccharomyces cerevisiae
Received: 4 March 2002 / Revised: 21 May 2002 / Accepted: 26 May 2002 / Published online: 3 July 2002
©Springer-Verlag 2002
Abstract Heterologousgenesforxyloseutilizationwere introducedintoanindustrial Saccharomycescerevisiae, strain A, with the aim of producing fuel ethanol from lig- nocellulosicfeedstocks.Twotransformants,A4andA6, wereevaluatedbycomparingtheperformancein4-lan- aerobicbatchcultivationstoboththeparentstrainanda laboratory xylose-utilizing strain: S. cerevisiae TMB
3001.Duringgrowthinaminimalmediumcontaininga mixtureofglucoseandxylose(50g/leach),glucosewas preferentiallyconsumed.Duringthefirstgrowthphase onglucose,thespecificgrowthrateswere0.26,0.32,
0.27and0.30 h–1 forstrainsTMB3001,A(parental strain),A4,andA6,respectively.Thespecificethanol productivities were 0.04, 0.13, 0.04 and 0.03 g/g.per hour,forTMB3001,A,A4andA6,respectively.The specificxyloseconsumptionrateswere0.06,0.21and
0.14g/g.perhour,respectivelyforstrainsTMB3001,A4 andA6.Xyloseconsumptionresultedmainlyinthefor- mationofxylitol,withbiomassandethanolbeingminor products.Themetaboliteprofileofintermediatesinthe pentosephosphatepathwayandkeyglycolyticinterme- diatesweredeterminedduringgrowthonglucoseand xylose,respectively.Themetabolitepatterndifferedde- pendingonwhetherglucoseorxylosewasutilized.The levelsofintracellularmetaboliteswerehigherinthein- dustrial strains than in the laboratory strain during growth on xylose.
J.Zaldivar· A.Borges· H.P.Smits· S.G.Villas-Bôas· J.Nielsen
L.Olsson (✉)
CenterforProcessBiotechnology, Biocentrum-DTU, Building223, TechnicalUniversityofDenmark,
2800Kgs.Lyngby, Denmark
e-mail:
Tel.: +45-45-252677, Fax: +45-45-884148
B.Johansson
AppliedMicrobiology, LundUniversity, P.O.Box124,
22100Lund, Sweden
Presentaddress:
B.Johansson, CentrodeCienciasdoAmbiente-Departamento de Biologia, UniversidadedoMinho, CampusdeGualtar, Braga, Portugal
Introduction
Manycountriesintheworldarecurrentlycommittedto reduceatmosphericCO2 levels.Theviewthatuseofbio- ethanolasanadditiveinfuelsfortransportationcanhelp inreducingexhaustemissionsisnowadaysgenerallyac- cepted(Wyman1996).Cheapfeedstockssuchasligno- cellulosicwaste(sugarcanebagasse,cornstalks,wheat straw)canpotentiallybeusedforcompetitiveethanol production(Hahn-Hägerdaletal.2001;Mielenz2001). However,theutilizationofsugarspresentinlignocellu- loserequiresefficienthydrolyticmethodsandefficient fermentationmicroorganisms,capableoffermentingthe pentosesaswellasthehexosesthatoriginatefromligno- cellulosicmaterial(Ingrametal.1999;Zaldivaretal.
2001).
Saccharomycescerevisiae hasevolvedintoaneffi- cient fermentation microorganism that has acquired qual- itiessuchashighethanolproductivity,tolerancetopro- cess hardiness, tolerance to fermentation by-products andis,therefore,preferredforethanolproductionfrom crops.Moreover,thepresenceofextrasetsofchromo- somesintheindustrialpolyploidstrain,andwiththatthe concomitantoverexpressionofgenes(Pretorius2000), couldbeanadditionaladvantage.Along-heldbeliefalso attributes higher genetic stability to polyploid strains, sincemultiplemutationaleventswillberequiredinor- dertobringaboutanychanges,butgeneticvariabilityin industrial strains is currently accepted to occur under strongselectivepressure,althoughatmuchlowerfre- quency than in laboratory strains (Hammond 1995; Pretorius2000).However,applyingS.cerevisiaeforfer- mentationoflignocellulosichydrolysateshasthedraw- back that it cannot naturally utilize pentoses.
Thevastaccumulatedknowledgeregardingthephysi- ologyandgeneticsof S.cerevisiae basicallyoriginated frominvestigationoflaboratorystrains.Unfortunately, afteryearsofhandlinginlaboratorysurroundings,labo- ratorystrainsdonotdisplaysomeofthetraitsthatchar- acterizeS.cerevisiaestrainsusedinindustry(Whealset al.1999).However,whenmetabolicengineeringgoalsin
437
Fig.1 Overviewofmetabolicpathwaysresultinginethanolpro- ductionfromglucose(A)andxylose(B),adaptedfromSchaaff- GerstenschlägerandMiosga(1997);Walker(1998).PPPPentose phosphate pathway, EMP Embden-Meyerhof-Parnas pathway
(glycolysis),Ru5pribulose-5-phopshate,Ri5Pribose-5-phosphate, Xu5Pxylulose-5-phosphate,S7Psedoheptulose-7-phosphate,F6P fructose-6-phosphate, Glyceral3P glyceraldehyde-3-phosphate, Acetalacetaldehyde,G6Pglucose-6-phosphate,XRxylosereduc- tase, XDH xylitol dehydrogenase, XK xylulokinase. NADP+, NADPH,NAD+,NADHandATPcofactorsareshown.Thesizeof thearrowshasbeenscaledaccordingtofluxdistributionbasedon Nissen et al. (1997)
S.cerevisiae aretobeproven,laboratorystrainshave primarilybeenused,asgeneticmanipulationsareeasier. Nonetheless,topursueanindustrialapplication,thecon- cepthastobeproveninanindustrialstrain,whichmight
grating plasmid harboring the endogenous gene encoding xylulokinase (XKS1) and genes for xylose reductase
(XYL1)andxylitoldehydrogenase(XYL2)from Pichia stipitis,whichenabledtheutilizationofxylose(Eliasson etal.2000;Fig1B).Thefermentationcapabilitiesofthe strainsconstructedwereevaluatedinminimalmediaun- deranaerobicconditionsintermsofgrowth,substrate consumption,productandby-productformationduring batch growth. The metabolite levels for the pentose phosphate pathway (PPP) were measured since this is the primarymetabolicpathwayforxylose(Ligthelm1988). Furthermore,theconnectionbetweenthemetabolicpro- fileandfactorssuchasgenotype,background(laborato- ry or industrial) of the strains, and sugars consumed
(glucose or xylose) was investigated.
havevaluableadditionalproperties.Alongtheselines,
usefulhostsforxylose-utilizing S.cerevisiae couldbe theacid-tolerantS.cerevisiaestrainsisolatedfromharsh environments(Lindénetal.1992)orbredstrainswith industrial background, as demonstrated by Ho et al.
(1998).
Inmetabolicengineeringitisimportanttoassessthe physiologicaloutcomeofthemodifiedmetabolicpath- way(StaffordandStephanopoulus2001).Forthisanaly- sisitishighlydesirabletoobtainaphysiologicalsnap- shotofthecellthatfaithfullyrepresentsitsmetabolic stateattheverymomentofharvesting.Suchasnapshot canbeobtainedbyanalysisofintracellularmetabolites
(deKoningandvanDam1992;Theobaldetal.1993; Gonzalezetal.1997;Smitsetal.1998;Groussacetal.
2000).
Inthiswork,anS.cerevisiaestrainusedintheindus- trialproductionofethanolwastransformedwithaninte-
Materials andmethods
Media forbatch cultivations
ThestrainsofS.cerevisiaewerecultivatedinminimalmediaac- cordingtoVerduynetal.(1992).Thecompositionofthemedium
(in g/l) was: (NH4)2SO4 5.0, KH2PO4 3.0, MgSO4·7H2O 0.5. Themediumwassupplementedwith1mloftracemetalsolution,
1 ml of vitamins solution and 1 ml ergosterol/Tween 80. The trace metal solution had the following composition (in g/l): EDTA15,ZnSO4·7H2O4.5,MnCl2·2H2O0.84,CoCl2·6H2O0.30, CuSO4·5H2O 0.30, Na2MoO4·2H2O 0.40, CaCl2·2H2O 4.5, FeSO4·7H2O3.0,H3BO3 1.0,KI0.10.Thecompositionofthevi- taminsolution(ing/l)was: D-biotin0.05,calciumpantothenate
1.0,nicotinicacid1.0,myoinositol25.0,thiaminehydrochloride
1.0, pyridoxal hydrochloride 1.0, p-aminobenzoic acid 0.20. A mixtureof50 g/lglucoseand50 g/lxylosewasusedascarbon source.Ergosterol,aprecursorofcellularmembranenotsynthe- sizedbyS.cerevisiaeinanaerobiccultures(AndreasenandStier
1953, 1954) was added as a solution containing ergosterol
(15mg/l) and Tween 80 (660mg/l) (Verduyn et al. 1992).
438
S. cerevisiaestrains
CEN.PK113-7A MATa his3-1MAL2-8c SUC2 wasusedastherecipient for the integrating plasmid YIpXR/XDH/XK (Entian andKötter1998).ThisplasmidharborstheXYL1andXYL2genes fromP.stipitisandanendogenousXKS1geneunderthecontrol ofthePGK1promoter.Theconstructedstrain,TMB3001(Elias- sonetal.2000),wasusedasaxylose-utilizingreferencestrainin thisstudy.IndustrialstrainAwastransformedwiththeplasmid YIpLoxZEO(Jeppssonetal.2002).Thisplasmidcontainsaho- mologousxylulokinaseXKS1,andheterologousXYL1andXYL2. Inadditionitcontainstwoselectionmarkers:zeocinandampicil- lin(forselectioninEscherichiacoli),andafragmentofanS.cer- evisiae gene to promote recombination (HIS3). Plasmid DNA
(10µg)waslinearizedwithNdeI.Fromthispreparation,5µgwas usedtotransformS.cerevisiaeAappropriatelytreatedwithlithi- umacetate(SchiestlandGietz1989).Screeningwasmadeon YPDplatescontaining50,150or300µg/mlzeocin,respectively. Plateswereincubatedat30°Candafter24–48 hcolonieswere transferredtofreshplateshavingasimilarformulation.Colonies thatgrewinthissecondselectionwereevaluatedin50mlYPX
(xylose20g/l)inbaffled300-mlErlenmeyerflasksandincubated inashakerat150rpm,at30°C.Potentiallygoodcandidateswere growninYPD(20 g/lglucose),thenwashedtwiceindistilled waterandfinallytransferredtoYPX(20g/lxylose).Toenablea faircomparison,inoculumwasnormalized,i.e.,initialbiomass wassimilarinallcases,equivalentto20mgDW/l.Growthwas evaluatedat600nminaspectrophotometer(HitachiU-1100,To- kyo, Japan).
Cultivation conditions
Batch fermentations were performed in in-house manufactured fermenterswithatotalvolumeof5l,equippedwithtwoRushton turbines,containing4lminimalmedium.Technicalqualitynitro- gen(AGA,Copenhagen,Denmark)containinglessthan5ppmO2 wasflushedthroughthevessels(0.4l/min)toobtainananaerobic environment,andtheexhaustgaspassedthrougharefluxcooler maintainedat2°Ctominimizeethanolevaporation.ThepHwas maintainedat5.0with2MNaOH.Thefermentationswererunat
30°Catastirringspeedof600 rpm.Carbondioxidewasmoni- toredduringfermentationwithanacousticgas-analyzer(Brüel& Kjaertype1308,Nærum,Denmark).Reactorswereinoculatedto aninitialbiomassconcentrationof2 mgDW/lwithprecultures grown in unbaffled flasks at 30°C and 150rpm for 15h.
Analytical methods
Cell growth was monitored by absorbance measurements at
600 nminaspectrophotometer(HitachiU-1100,Tokyo,Japan) andgravimetricaldeterminationbydryweightasdescribedby OlssonandNielsen(1997).Samplesfordeterminationofsugars andextracellularmetaboliteswerefilteredthrough0.45-µm-pore- sizeacetatefilters(Osmonics,Minnetonka,Minn.)andthefil- trateswerefrozenat–20°Candlaterusedforanalyses.Glucose, arabinose, galactose, mannose and xylose were determined by high-performanceanionexchangechromatography(Dionex,Sun- nyvale,Calif.),usingpulsedamperometricdetection.Sugarswere separatedinaCarboPacPA1columnat30°C.Twoeluents,Aand B,200mMNaOHand1mMNaOH/0.03mMNaAc,respectively, wereusedasthemobilephases,ataflowof1ml/min.Thegradi- entwasestablishedasfollows:from0–40mineluentAwasused at100%oftheflowrate;from40–45min,theflowofAwasde- creasedtozero,whereastheflowofBwasincreasedto100%; from 45–50 min the flow of B was kept constant, and from
50–55mintheflowofBwasdecreasedto0%,whereastheflow of A was increased to 100% to finish the cycle.
Sugaralcohols,xylitolandarabitol,wereseparatedinaCarbo- PacMA1columnat30°C,utilizing612mMNaOHasthemobile phase, at a flow rate of 1ml/min.
Extracellularmetabolites,ethanol,succinate,pyruvate,acetate, andglycerolweremeasuredbyhigh-performanceliquidchroma- tography(Waters,Milford,Mass.),usinganAminexion-exclusion HPX-87Hcationexchangecolumn(Bio-Rad,Calif.)at65°Cand
5 mMH2SO4 ataflowrateof0.6 ml/minasthemobilephase. Succinate,glycerolandethanolweredeterminedbyRI-detection and pyruvate and acetate by UV-detection.
Fortheanalysisofintracellularmetabolites,5 mlbrothwas harvestedinduplicatefromthereactors,beforeglucoseexhaus- tion(at22and26 hofcultivation)andafterglucoseexhaustion
(42,79and131hofcultivation).Proceduresformetabolicarrest, solid-phaseextractionofmetabolitesandanalysishavebeende- scribedindetailbySmitsetal.(1998).However,theanalysisby high-pressure ion exchange chromatography coupled to pulsed amperometricdetectionusedtoanalyzecellextracts,wasslightly modified.SolutionsusedwereeluentA,75mMNaOH,andeluent B,500mMNaAc.Topreventcontaminationofcarbonateinthe eluentsolutions,a50%NaOHsolutionwithlowcarbonatecon- centration(BakerAnalysed,Deventer,TheNetherlands)wasused insteadofNaOHpellets.TheeluentsweredegassedwithHefor
30minandthenkeptunderaHeatmosphere.Thegradientpump wasprogrammedtogeneratethefollowinggradients:100%Aand
0%B(0min),alineardecreaseofAto70%andalinearincrease ofBto30%(0–30min),alineardecreaseofAto30%andalin- earincreaseofBto70%(30–70min),alineardecreaseofAto
0%andalinearincreaseofBto100%(70–75 min),0%Aand
100%B(75–85min),alinearincreaseofAto100%andalinear decreaseofBto0%(85–95min).Themobilephasewasrunata flowrateof1ml/min.OtherconditionswereaccordingtoSmitset al. (1998).
Calculations
Ethanol evaporation
Duetoitslowboilingpoint,ethanolevaporateswhenthereactor isspargedwithnitrogen.Tocompensateforethanollossinthe calculations,evaporationratewasdeterminedexperimentally.The reactorwassetupasusedinthebatchcultivationsandethanol wasadded.Theethanolevaporationwasestimatedbymeasuring theethanolconcentrationintheliquidphaseovertime.Theevap- orationatethanolconcentrationsobtainedinthisworkfollowed theequationdCethanol/dt=-k Cethanol,wheretherateconstantwas k=0.1134 h–1.
Yields ofbiomass andethanol
Yieldsofbiomassandethanolonsugarswerecalculatedbasedon thetimeofglucoseexhaustionfromthereactor(36h)fortheref- erencestrainA,sinceitisunabletoconsumexylose.Forthexy- lose-utilizingstrains,theendofthefermentation(191h)wasused for the calculations.
Specific ethanol productivity
Specificethanolproductivitywasbasedonthevolumetricproduc- tivity(g/l.perhour)dividedbythebiomassatthetimeofglucose exhaustion(forthereferencestrain)andattheendofthecultiva- tionfortherecombinantstrains.Specificethanolproductivitywas expressed in g/g DW per hour.
Sugar uptake rate
Sugar(glucoseorxylose)uptakeratewascalculatedbasedonthe equation:uptakerate=dCsugar/dt 1/DW,whereDWisthebio- massconcentration.Sugaruptakeratewasexpresseding/gDW per hour.
Fig.2A–C FermentationperformanceofSaccharomycescerevisi- aestrainsTMB3001( , ),A( , ),A4( , )andA6( , ). TMB3001hasalaboratorybackground,whereasA,A4andA6 haveindustrialbackgrounds;TMB3001,A4andA6arexylose- utilizingstrains.Timecourseofglucose(hollowsymbols)andxy- lose(filledsymbols)consumption(A),biomassconcentration(B), andethanolproduction(C)inanaerobicbatchculturesemploying minimal medium
439
Table1 SummaryofthegrowthcharacteristicsofstrainsofSac- charomycescerevisiae duringanaerobicbatchfermentationofa mixture of 50g/l glucose and 50g/l xylose
TMB 3001aA A4A6
µMaxb0.260.320.270.30
Biomassmaxc3.54.44.04.6
Ethanolmaxd23.320.825.225.2
Glycerolmaxe8.76.58.38.3
Xylitolmaxf4.10.413.615.9
Ysxg0.0340.0900.0400.046
Yseh0.230.420.270.27 rEthanoli0.040.130.040.03 rXylosej0.060 0.21 0.14
aTMB 3001, A4 and A6 are xylose-metabolizing strains
bMaximum specific growth rate (h–1)
c,d,e,fMaximalconcentrationsofbiomass,ethanol,glycerolandxy- litol (g/l), respectively
gBiomassyield(gbiomass/gsugarconsumed).Forthecalculation ofyieldsofbiomassandethanol,coefficientswerecalculatedon glucose(50 g/l)onlyforstrainA,whereasfortherecombinant strains calculations were based on glucose + xylose (100g/l)
hEthanol yield (g ethanol/g sugar consumed)
iSpecificethanolproductivity(gethanol/gbiomassperhour).The volumetricproductivity,i.e.,ethanolproducedperunittime(g/l perhour),dividedbythebiomassconcentrationatthetimeofglu- coseexhaustionforstrainA,andtheend-pointfortherecombi- nant strains
jSpecificxyloseconsumptionrate(gxylose/gbiomassperhour), calculatedbasedontheequation:uptakerate=dCxylose/dt1/DW, where DW is the biomass concentration
Growth(andethanolproduction)occurredpredomi- nantlyduringtheinitial36h,inwhichintervalglucose waspredominantlyconsumed.Aminimalgrowthofthe recombinant strains on xylose occurred afterwards
(Fig. 2B). Thus, the values presented in Table 1 for strainsTMB3001,A4andA6representincreasesof3%,
15%and13%,respectively,overgrowthonglucose.No growthonxyloseoccurredinstrainA.Theyieldofbio- massonsugarswas0.09 g/gontheparentalstrainA, whichisclosetothetypicallyfoundvalueof0.1g/gfor anaerobicS.cerevisiaecultures.ThehighestYsx among
recombinantstrainswas0.046 g/g(Table 1),i.e.,50%
Results
Sugar consumption, growth, andformation ofextracellular metabolites
Therecombinantstrainswereabletoutilizexyloseand onlythosereactorscontainingstrainswithanindustrial backgroundexhaustedxyloseduringthe191hofculti- vation(Fig. 2A).Sugarconsumptionoccurredsequen- tially:firstglucoseandthenxylose.Glucoseexhaustion occurredat36 handthespecificglucoseconsumption rateswere2.7,2.8,2.9and2.9g/gperhour,forstrains TMB3001,A,A4andA6,respectively.Thespecificxy- loseconsumptionrateswere0.06,0.21and0.14g/gper hour,respectively,forstrainsTMB3001,A4andA6
(Table1),i.e.,3.5and2.4-foldhigherintherecombinant strains with industrial background compared to TMB
3001 with a CEN.PK laboratory strain background.
lowerthanintheparentalstrain,causedbytheminimal growthduringthexyloseconsumptionphase.Astheref- erencestrainwasunabletoutilizexylose,yieldcoeffi- cientswerebasedonglucose(50g/l),whereasforthere- combinantstrainsglucoseandxylose(100g/l)werecon- sidered for the calculations (Table1).
Thefinalethanolconcentrationwas23.3,20.8,25.2 and25.2 g/l,instrainsTMB3001,A,A4andA6,re- spectively (Fig. 2C, Table 1). For the recombinant strains,thesevaluesrepresentincreasesof7%,15%and
12%, respectively, compared with production on glu- cose.
Peakconcentrationsofglycerolwere8.7,6.5,8.3and
8.3g/linstrainsTMB3001,A,A4andA6,respectively, whereastheproductionofxylitolreached4.1,0.4,13.6 and 15.9g/l, respectively (Table1).
440
Fig.3 Intracellular metabolite profiles determined during growth on glucose + xylose
(open bars) and xylose (filled bars) inS. cerevisiaestrains TMB 3001, A, A4 and A6.
Bars represent peak heights and not absolute concentrations. Samples were harvested at
26 and 79h, respectively. G6PGlucose-6-phosphate, E4Perythrose-4-phosphate, F6Pfructose-6-phosphate, Ri5Pribose-5-phosphate, Xu5Pxylulose-5-phosphate, S7Psedoheptulose-7-phos- phate,F1,6fructose-1,6- diphosphate
Intracellular metabolites formation
Toevaluatethephenotypeoftherecombinantstrainfur- ther, intracellular metabolites were analyzed. Samples werecollectedfromthereactorduringtheglucosecon- sumptionphase(at22and26 h)andduringthexylose consumptionphase(samples42,79and131 h).After quenchingandadequatesamplepreparation,theywere analyzedformetabolitesfromthePPPandtheEmbden- Meyerhof-Parnas(EMP)pathway.Erythrose-4-P,ribose-
5-P,xylulose-5-P,andsedoheptulose-7-P,areexclusive tothePPP,fructose-6-Pandglucose-6-Pcanbepresent inbothpathways,andfructose-1,6-diPisexclusiveto theEMPpathway(Fig. 1A,B).Itshouldbenotedthat themetaboliteprofilesarenotbasedonabsoluteconcen- trations,butonlypeakheightsarerepresented(Fig. 3). However,thesamesamplevolumewasappliedduring analysis,makingitpossibletocomparetheheightasa representationoftheamountofthecomponentinthe samples. To facilitate the discussion, samples corre- sponding to 26h and 79h were chosen as representatives oftheeffectofglucoseandxylose,respectively,onthe metabolism of the strains.
Duringtheglucoseconsumptionphase,thelaboratory
ashighasinstrainA,butPPPintermediatesribose-5-P and xylulose-5-P were one order of magnitude lower thanfructose-1,6-diP.Theleveloffructose-6-P,thepre- cursoroffructose-1,6-diPintheEMPpathway,was40- foldlowerthanthelatter.InstrainA4,metabolitelevels
(includingfructose1,6-diP)were10-foldlowerthanin strains A and A6.
Duringthexyloseconsumptionphase(afterglucose exhaustion) the levels of PPP intermediates in strain TMB3001didnotincreaseandlevelsoftheglycolytic intermediatefructose-1,6-diPdidnotchangedramatical- ly.InstrainA,therewasaminimalleveloffructose-1,6- diP,65-foldlowerthanduringgrowthonglucose,possi- bly due to gluconeogenesis (Gancedo and Gancedo
1997).InstrainA6theleveloffructose-1,6-diPwasre- ducedbyoneorderofmagnitudecomparedtotheglu- cosephase.Furthermore,intherecombinantstrainswith industrialbackground,erythrose-4-Plevelsineachstrain wereinthesamerange.Similarly,thelevelsofribose-
5-Pineachstrain,orxylulose-5-Pineachstrain,werein thesamerange.Noaccumulationofseptulose-7-phos- phatecouldbeshown,challengingearlierstudies(Kötter and Ciriacy 1993).
strain(TMB3001)hadlowlevelsofPPPmetabolites,
andthelevelsofglucose-6-Pandfructose-1,6-diPwere alsolowcomparedtothelevelsinstrainsAandA6.In theindustrialstrainA,unabletometabolizexylose,in- termediatessuchasglucose-6-P,erythrose-4-Pandri- bose-5-Pappearedasexpectedonlywhenglucosewas availableinthebroth(PPPintermediateserythrose-4-P andribose-5-Pareusedinbiosyntheticreactions).The leveloffructose-1,6-diPwas20-foldhigherthanthatof erythrose-4-P.InstrainA6,theleveloffructose-6-Pwas
Discussion
Sugar consumption, growth, andformation ofextracellular metabolites
This work verified that the recombinant strains con- sumed sugars sequentially, xylose being utilized after glucoseexhaustion.TherecombinantstrainsA4andA6 withindustrialbackgroundweresuperiortoTMB3001
(withlaboratorybackground),asindicatedbytheen- hancedxyloseconsumption,biomassandethanolpro- duction(Fig.2A–C).Ithasbeenshownpreviouslythat overexpressionofXRledtoanincreasedxylosecon- sumptionrateand,concomitantly,ahigherxylitolyield
(Johansson2001).Theyieldsofxylitolonconsumedxy- losewere0.27and0.31 g/gforstrainsA4andA6,re- spectively,whichwere1.5to2-foldhigherthantheyield of0.16g/gforTMB3001.Thehigherxylitolyieldsin the industrial recombinant strains might be a conse- quenceoftheincreasedxyloseconsumptionrates.This mightbecausedbyincreasedXRactivityintherecom- binantindustrialstrainssince,duetotheirpolyploidna- ture,severalintegrationsoftheplasmidmighthaveoc- curred.
Asaresultoftheincreasedxyloseconsumption,bio- massandethanolconcentrationinstrainsA4andA6in- creased.InspiteofthebetterperformanceofstrainsA4 andA6comparedtoTMB3001,theslowconsumption ofxyloseincomparisonwithglucoseconsumptionisa problem in direct implementation. A specific xylose transporterisabsentin S.cerevisiae;uptakeiscarried outbythehexosetransporter(Walshetal.1994;Özcan andJohnston1999),whichhasanaffinityforxylosein therange98–137mM(SinghandMishra1995)oreven lower(170 mMisindicatedbyKotyk1967).Theslow metabolismofxyloseresultsinlongfermentationtimes, lengtheningtheexposureofthecellstostressingcondi- tionsinthereactorsuchasincreasedlevelsofethanol andby-products,affectingcellviability.Viabilitywould befurtherreducedbythelimitedbiosyntheticcapability afterglucoseexhaustion,asdiscussedbelow.Further- more,asignificantpercentageofconsumedxyloseisdi- rectedtowardsxylitolformation,duetocofactorimbal- anceinxylosemetabolism(Bruinenbergetal.1983)as well as to unfavorable thermodynamics (Rizzi et al.
1988,1989).Inthisregard,approximately30%ofthe xylose consumed was directed towards xylitol formation.
Intracellular metabolite levels
Togainfurtherinsightintothemetabolismoflaboratory andindustrialS.cerevisiaestrains,keyintermediatesin thePPPandEMPpathwaywereanalyzedinthiswork. Theresultsverifiedthattheprofilesofmetabolitesare relatedto:(1)thegeneticbackgroundofthestrains;the laboratorystrainshadlevelsofmetaboliteslowerthan thestrainswithindustrialbackground,(2)straingeno- type;onlystrainstransformedwiththegenesforxylose utilizationpossess,toalower(TMB3001)orhigher(A4 andA6)extent,aprofileofPPPmetabolitesafterglu- coseexhaustion,(3)thetypeofcarbonsourcepresentin thereactor;thepresence(andutilization)ofxylosere- sultedindifferencesinmetaboliteprofilesbetweenre- combinant and non-transformed strains.
Asverifiedinthiswork,itisacriticalissuethatin batchcultivationsemployingminimalmedia,xyloseis predominantlyconsumedafterglucoseexhaustionand
441
that the non-oxidative PPP gains a catabolic role
(Fig.1B).Sincereducingpowerresultsfromtheoxida- tivePPP(NogaeandJohnston1990)andATPinanaero- bic cultures results from glycolytic activity, cellular growth is minimized when these pathways are by- passed.BecauseATP(andethanol)formationreliesona highglycolyticflux,aloweffluxoftheintermediates glyceraldehyde-3-Pandfructose-6-PfromPPPtoEMP pathway results in both minimal growth and minimal ethanolformation.Thus,withtheaimbeingahighpro- ductionofethanol,metabolicengineeringdemandsan acceleratedfluxthroughthesequenceof“pipelines”in thePPPtogeneratehighlevelsoftheEMPpathwayin- termediates glyceraldehyde-3-P and fructose-6-P. It is expectedthatsuchinfluxwillfuelahighfluxthrough theEMPpathway,yieldingethanollevelscomparableto thoseachievedwithglucose.Inthisstudy,weshowed thatbyintroductionofxylosemetabolizinggenesintoan industrialstrain,increasedxyloseconsumptionrateand elevated levels of PPP metabolites were achieved in comparisonwiththosedeterminedforxyloseutilization inalaboratorystrain.TheincreasedlevelsofPPPme- tabolitescouldbeanindicationofhigherfluxthrough thispathwayortolimitationsin,ordownstreamof,the PPP. However, further studies would be required for clarification.
Acknowledgements We thank Bruno Jarry, Orsan-Amylum, France,forkindlyprovidingthestrainSaccharomycescerevisiae A.TheworkonxylosefermentationattheCenterforProcessBio- technologyattheTechnicalUniversityofDenmarkissupported under the European Commission Framework V, contract no. QLK3-CT-1999–00080.
References
AndreasenAA,StierTJB(1953)AnaerobicnutritionofSaccharo- mycescerevisiae.I.Ergosterolrequirementforgrowthinade- fined medium. JCell Comp Physiol 41:23–36
AndreasenAA,StierTJB(1954)AnaerobicnutritionofSaccharo- mycescerevisiae. II.Unsaturatedfattyacidrequirementfor growth in a defined medium. JCell Comp Physiol 43:271–281
Bruinenberg PM, de Bot PHM, van Dijken JP, Scheffers WA
(1983)Theroleofredoxbalancesintheanaerobicfermenta- tionbyxyloseutilizingyeasts.EurJ ApplMicrobiolBiot- echnol 18:287–292
Eliasson A, Christensson C, Wahlbom CF, Hahn-Hägerdal B
(2000)Anaerobicxylosefermentationbyrecombinant Sac- charomycescerevisiae harbouring XYL1, XYL2 and XKS1 in mineralmediachemostatcultivations.ApplEnvironMicrobiol
66:3381–3386
EntianK-D,Kötter P(1998)Yeastmutantandplasmidcollec- tions.In:BrownAJP,TuiteMF(eds)Methodsinmicrobiolo- gy. Academic Press, London, pp431–449
Gancedo JM,Gancedo C(1997)Gluconeogenesisandcatabolite inactivation. In: Zimmermann FK, Entian K-D (eds) Yeast sugarmetabolism.Biochemistry,genetics,biotechnology,and applications. Technomic, Basel, pp359–378
Gonzalez B,Francois J,Renaud M(1997)Arapidandreliable methodformetaboliteextractioninyeastusingboilingbuf- fered ethanol. Yeast 13:1347–1355
Groussac E,Ortiz M,Francois J(2000)Improvedprotocolsfor quantitativedeterminationofmetabolitesfrombiologicalsam- plesusinghighperformanceionic-exchangechromatography
442
withconductimetricandpulsedamperometricdetection.En- zyme Microb Technol 26:715–723
PretoriusIS(2000)Tailoringwineyeastforthenewmillennium:
novel approaches to the ancient art of winemaking. Yeast
Hahn-HägerdalB,WahlbomCF,GardonyiM,vanZylWH,Cor- deroOtero RR,Jönsson LJ(2001)Metabolicengineeringof Saccharomycescerevisiaeforxyloseutilization.AdvBiochem Eng Biotechnol 73:53–84
HammondJRM(1995)Geneticallymodifiedbrewingyeastsfor the 21st century. Progress to date. Yeast 11:1613–1627
HoNWY,Chen Z,Brainard AP(1998)Geneticallyengineered Saccharomyces yeastcapableofeffectivecofermentationof glucose and xylose. Appl Environ Microbiol 64:1852–1859
Ingram LO,Aldrich HC,BorgesACC,Causey TB,Martinez A, Morales F,Saleh A,Underwood SA,Yomano LP,York SW, Zaldivar J,Zhou S(1999)Entericbacterialcatalystsforfuel ethanol production. Biotechnol Prog 15:855–866
Jeppsson M, Johansson B, Hahn-Hägerdal B, Gorwa-Graus- lund MF(2002)Reducedoxidativepentosephosphatepath- wayfluxinrecombinantxylose-utilizingSaccharomycescere- visiae strainsimprovestheethanolyieldfromxylose.Appl Environ Microbiol 68:1604–1609
Johansson B(2001)Metabolicengineeringofthepentosephos- phatepathwayofxylosefermenting Saccharomycescerevisi- ae. PhD thesis, Lund University, Sweden
KoningWde,vanDamK(1992)Amethodforthedetermination ofchangesofglycolyticmetabolitesinyeastonasubsecond time scale using extraction at neutral pH. Anal Biochem
201:118–123
KötterP,CiriacyM(1993)XylosefermentationbySaccharomy- ces cerevisiae. Appl Microbiol Biotechnol 38:776–783
KotykA(1967)Propertiesofthesugarcarrierinbaker’syeast.2. Specificity of transport. Folia Microbiol 12:121–131
LigthelmME,PriorBA,duPreezJC,BrandtV(1988)Aninvesti- gationofD-{1–13C}xylosemetabolisminPichiastipitisun- deraerobicandanaerobicconditions.ApplMicrobiolBiot- echnol 28:293–296
LindénT,PeetreJ,Hahn-HägerdalB(1992)Isolationandcharac- terizationofaceticacidtolerantgalactose-fermentingstrains of Saccharomycescerevisiae fromaspentsulfiteliquorfer- mentation plant. Appl Environ Microbiol 58:1661–1669
MielenzJR(2001)Ethanolproductionfrombiomass:technology and commercialization status. Curr Opin Microbiol 4:324–329
NissenTL,SchulzeU,NielsenJ,VilladsenJ(1997)Fluxdistribu- tioninanaerobic,glucose-limitedcontinuousculturesofSac- charomyces cerevisiae. Microbiology 143:203–218
NogaeI,JohnstonM(1990)Isolationandcharacterizationofthe ZWF1geneofSaccharomycescerevisiae,encodingglucose-6- phosphate dehydrogenase. Gene 96:161–169
OlssonL,NielsenJ(1997)On-lineandin-situmonitoringofbio- massinsubmergedcultivations.TrendsBiotechnol15:517–
522
ÖzcanS,JohnstonM(1999)Functionandregulationoftheyeast hexose transporters. Microbiol Mol Biol Rev 63:554–569
16:675–729
RizziM,ErlemannP,Bui-ThanhN-A,DellwegH(1988)Xylose fermentationbyyeast.4.Purificationandkineticstudiesof xylosereductasefrom Pichiastipitis.ApplMicrobiolBiot- echnol 29:148–154
RizziM,KleinC,SchulzeC,Bui-ThanhN-A,DellwegH(1989) Xylose fermentation by yeast. 5. Use of ATP balances for modelingoxygen-limitedgrowthandfermentationofyeastPi- chiastipitiswithxyloseascarbonsource.BiotechnolBioeng
34:509–514
Schaaff-Gerstenschläger I,Miosga T(1997)Thepentosephos- phatepathway.In:ZimmermannFK,EntianK-D(eds)Yeast sugarmetabolism.Biochemistry,genetics,biotechnology,and applications. Technomic, Basel, pp-271–283
Schiestl RH,Gietz RD(1989)Highefficiencytransformationof intactyeastcellsusingsinglestrandednucleicacidsascarrier. Curr Genet 16:339–346
SinghA,MishraP(1995)Microbialpentoseutilization.Elsevier, Amsterdam
SmitsHP,CohenA,ButtlerT,NielsenJ,OlssonL(1998)Cleanup andanalysisofsugarphosphatesinbiologicalextractsbyus- ingsolid-phaseextractionandanion-exchangechromatogra- phy with pulse amperometric detection. Anal Biochem
261:36–42
Stafford DE,Stephanopoulus G(2001)Metabolicengineeringas anintegratingplatformforstraindevelopment.CurrOpinMi- crobiol 4:336–340
TheobaldU,MailingerW,ReussM,RizziM(1993)Invivoanal- ysisofglucose-inducedfastglycolyticchangesinyeastade- nine nucleotide pool applying a rapid sampling technique. Anal Biochem 214:31–37
VerduynC,PostmaE,ScheffersWA,vanDijkenJP(1992)Effect ofbenzoicacidonmetabolicfluxesinyeasts:acontinuous- culturestudyontheregulationofrespirationandalcoholicfer- mentation. Yeast 8:501–517
WalkerG(1998)Yeastgrowth.In:WalkerG(ed)Yeast:physiolo- gy and biotechnology. Wiley, New York, pp101–202
WalshMC,SmitsHP,ScholteM,vanDamK(1994)Affinityof glucosetransportin Saccharomycescerevisiae ismodulated during growth on glucose. JBacteriol 176:953–958
WhealsAE,BassoLC,AlvesDMG,AmorimHV(1999)Fueleth- anol after 25 years. Trends Biotechnol 17:482–487
Wyman CE(1996)Ethanolproductionfromlignocellulosicbio- mass:overview.In:WymanCE(ed)Handbookonbioethanol: productionandutilization.TaylorandFrancis,Levittown,Pa. pp1–18
Zaldivar J,Nielsen J,Olsson L(2001)Fuelethanolproduction fromlignocellulose:achallengeformetabolicengineeringand process integration. Appl Microbiol Biotechnol 56:17–34