The non-oxidative pentose phosphate pathway controls the fermentation rate of xylulose but not of xylose in Saccharomyces cerevisiae TMB3001

FEMS Yeast Research 2 (2002) 277^282 www.fems-microbiology.org The non-oxidative pentose phosphate pathway controls the fermentation rate of xylulose but not of xylose in Saccharomyces cerevisiae TMB3001 Bjo«rn Johansson 1,Ba«rbel Hahn-Ha«gerdal Department of Applied Microbiology, Lund University, P.O. Box124, 221 00 Lund, Sweden Received 23 January 2002; received in revised form 4 April 2002; accepted 26 April 2002 First published online 28 May 2002 Abstract Saccharomyces cerevisiae is able to ferment xylose,when engineered with the enzymes xylose reductase (XYL1) and xylitol dehydrogenase (XYL2). However,xylose fermentation is one to two orders of magnitude slower than glucose fermentation. S. cerevisiae has been proposed to have an insufficient capacity of the non-oxidative pentose phosphate pathway (PPP) for rapid xylose fermentation. Strains overproducing the non-oxidative PPP enzymes ribulose 5-phosphate epimerase (EC 5.1.3.1),ribose 5-phosphate ketol isomerase (EC 5.3.1.6),transaldolase (EC 2.2.1.2) and transketolase (EC 2.2.1.1),as well as all four enzymes simultaneously,were compared with respect to xylose and xylulose fermentation with their xylose-fermenting predecessor S. cerevisiae TMB3001,expressing XYL1, XYL2 and only overexpressing XKS1 (xylulokinase). The level of overproduction in S. cerevisiae TMB3026,overproducing all four non-oxidative PPP enzymes,ranged between 4 and 23 times the level in TMB3001. Overproduction of the non-oxidative PPP enzymes did not influence the xylose fermentation rate in either batch cultures of 50 g l 31 xylose or chemostat cultures of 20 g l 31 glucose and 20 g l 31 xylose. The low specific growth rate on xylose was also unaffected. The results suggest that neither of the non-oxidative PPP enzymes has any significant control of the xylose fermentation rate in S. cerevisiae TMB3001. However,the specific growth rate on xylulose increased from 0.02^0.03 for TMB3001 to 0.12 for the strain overproducing only transaldolase (TAL1) and to 0.23 for TMB3026,suggesting that overproducing all four enzymes has a synergistic effect. TMB3026 consumed xylulose about two times faster than TMB30001 in batch culture of 50 g l 31 xylulose. The results indicate that growth on xylulose and the xylulose fermentation rate are partly controlled by the non-oxidative PPP,whereas control of the xylose fermentation rate is situated upstream of xylulokinase,in xylose transport,in xylose reductase,and/or in the xylitol dehydrogenase. ß 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords: Xylose; Xylulose; Saccharomyces cerevisiae; Pentose phosphate pathway; Multiple overexpression 1. Introduction Ethanol produced by fermentation of ligno-cellulosic biomass could be a renewable alternative to liquid fossil fuels. Ligno-cellulosic biomass contains the pentose sugar xylose,which the preferred ethanol production organism Saccharomyces cerevisiae cannot metabolise. S. cerevisiae * Corresponding author. Tel.: +46 (46) 222 8428; Fax: +46 (46) 222 4203. E-mail addresses: bjorn_johansson@bio.uminho.pt (B. Johansson), barbel.hahn-hagerdal@tmb.lth.se (B. Hahn-Ha«gerdal). 1 Present address: Centro de Ciencias do Ambiente,Departamento de Biologia,Universidade do Minho,Campus de Gualtar,4710-057 Braga, Portugal. can be engineered to utilise xylose by expression of the gene XYL1 encoding xylose reductase (XR) and XYL2 encoding xylitol dehydrogenase (XDH) from Pichia stipitis [1]. XR converts xylose to xylitol and XDH subsequently converts xylitol to xylulose,which S. cerevisiae can metabolise [2,3]. Xylose fermentation by such recombinant yeast strains is slow and xylitol is a major by-product [1,4,5]. The non-oxidative pentose phosphate pathway (PPP) is the catabolic pathway for xylose in naturally xylose-fermenting yeasts,and for xylulose in S. cerevisiae [6]. Xylose is metabolised at 60% of the rate of glucose by Candida shehatae [7],while S. cerevisiae ferments xylulose only at 5^15% of the rate of glucose [7,8]. The latter observation has been ascribed to a low capacity of the non-oxidative PPP in S. cerevisiae [1,9]. S. cerevisiae has been found to metabolise xylose togeth- 1567-1356 / 02 / $22.00 ß 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S1567-1356(02)00114-9

278 B. Johansson, B. Hahn-Ha«gerdal / FEMS Yeast Research 2 (2002) 277^282 er with ribose,but neither of the sugars alone [10,11]. Xylose and ribose co-metabolism would by-pass the nonoxidative PPP enzyme ribulose 5-phosphate isomerase (RPE). Flux modelling has shown that the ux through the reaction catalysed by RPE is very low in anaerobic xylose fermenting S. cerevisiae TMB3001 [12]. The nonoxidative PPP intermediate sedoheptulose 7-phosphate accumulates both in xylose-fermenting [1] and in xylulosefermenting [13] S. cerevisiae,suggesting insu cient transaldolase activity. Overexpression of transaldolase leads to faster aerobic growth on xylose [9],but not to faster xylose fermentation. Engineering of single genes to increase metabolic ux has generally not been successful. The theory of metabolic control analysis [14,15] suggests that overproduction of individual metabolic enzymes will not produce large increases in ux unless the control coe cient is larger than 0.6 [16]. Therefore simultaneous alteration of multiple genes would be necessary to achieve increased metabolic ux [16,17]. Overproduction of single enzymes in glycolysis does not improve the rate of glucose fermentation in S. cerevisiae [18]. However,simultaneous overproduction of seven glycolytic enzymes [19] increases the fermentative capacity under certain conditions [20]. Tryptophan biosynthesis may also be enhanced by multiple overproduction of ve enzymes,where overproduction of the single enzymes has little e ect [21]. The S. cerevisiae genes RPE1, RKI1, TAL1 and TKL1 encode the non-oxidative PPP enzymes ribulose 5-phosphate epimerase (RPE),ribose 5-phosphate isomerase (RKI),transaldolase (TAL) and transketolase (TKL),respectively. These enzymes carry out the conversion of xylulose 5-phosphate to glyceraldehydes 3-phosphate and fructose 6-phosphate,which are intermediates of glycolysis. In this investigation we explored the e ect of single and simultaneous overexpression of the non-oxidative PPP genes on xylose and xylulose fermentation by S. cerevisiae TMB3001 [22]. TMB3001 is a laboratory xylose-fermenting reference strain expressing P. stipitis XYL1 and XYL2 and overexpressing the endogenous xylulokinase gene (XKS1). 2. Materials and methods 2.1. Strains The prototrophic S. cerevisiae TMB3001 [22] was used as the host strain for all genetic constructions and as control strain. TMB3001 is a S. cerevisiae CEN.PK 113-7A [23] with the XYL1 (XR) and XYL2 (XDH) genes from P. stipitis and the endogenous XKS1 gene,encoding xylulokinase,stably integrated into the chromosomal HIS3 locus. S. cerevisiae strains TMB3013,TMB3014, TMB3016,and TMB3023 overexpress the genes RPE1 (4U), TAL1 (24U), TKL1 (17U),and RKI1 (14U),respectively [24] (values in parentheses represent increased enzyme activity compared with the level in TMB3001). The S. cerevisiae genes RPE1, RKI1, TAL1 and TKL1 encode the non-oxidative PPP enzymes ribulose 5-phosphate epimerase (EC 5.1.3.1),ribose 5-phosphate ketol isomerase (EC 5.3.1.6),transaldolase (EC 2.2.1.2) and transketolase (EC 2.2.1.1),respectively. TMB3026 overexpresses all four genes, RPE1 (4U), TAL1 (23U), TKL1 (17U),and RKI1 (14U),simultaneously [24]. The overexpressing strains were constructed by cloning the respective gene in the vector pb3 PGK behind the strong PGK1 promoter [25]. The constructs were integrated at the locus of each gene [24]. 2.2. Preparation of inoculum and batch fermentation De ned mineral medium [26] was supplied with 20 g l 31 glucose for growth of inoculum and 50 g l 31 xylose or xylulose for batch fermentation. Strains were initially grown in 500-ml ba ed shake asks in 100 ml glucose medium. Cells were harvested at a dry weight of ca. 3 g l 31,well within the exponential phase. The cells were washed twice in ice-cold water and re-suspended in icecold xylose or xylulose medium,respectively. Xylulose was produced in-house as previously described [27]. For each strain two 25-ml asks with rubber stoppers and 20 ml working volume were inoculated with 5 g l 31 cell dry weight. The asks were incubated at 30 C with magnetic stirring in a water bath. Samples were withdrawn through a 2-mm hypodermic needle with a syringe and fermentation gases were expelled through a 0.8-mm needle stu ed with breglass cotton. 2.3. Aerobic growth rate measurements For speci c growth rate determination de ned mineral medium [26] was supplied with 10 g l 31 glucose,xylulose or xylose. Cells were prepared as previously described for preparation of inoculum. Speci c growth rates were determined in 7-ml test tubes containing 3 ml medium. Two independent experiments including three test tubes were conducted for each strain and each carbon source. The tubes were inoculated with washed glucose-grown cells (prepared as described in Section 2.2),supplied with loosely tted caps and incubated at 30 C with vigorous shaking. Growth was monitored as optical density during 6^8 h with direct spectrophotometric measurement at 620 nm. 2.4. Continuous cultivation Yeast cells were grown for 12 h in 200 ml de ned mineral medium [26] containing 50 g l 31 glucose,10 mg l 31 ergosterol and 0.4 g l 31 Tween 80 in a 250-ml ba ed shake ask. Cells were centrifuged at 5000Ug for 5 min

B. Johansson, B. Hahn-Ha«gerdal / FEMS Yeast Research 2 (2002) 277^282 279 at 4 C,and used to inoculate 1.5 l of the same medium to OD 620 0.5 in a Bio o III fermenter (New Brunswick Scienti c,edison,nj,usa). Antifoam was added at 0.5% (v/v) (Dow Corning Antifoam RD Emulsion,BDH Laboratory Supplies,Poole,UK). Continuous cultivation was set up at dilution rates of 0.06 and 0.12 h 31 at 30 C,pH 5.5 controlled by addition of 5 M NaOH,and a stirring speed of 200 rpm. The fermenter was sparged with 0.2 l min 31 nitrogen (containing less than 5 ppm O 2 ) as measured with a gas mass ow meter (Bronkhorst,Ruurlo, The Netherlands). 2.5. Analysis of substrates and products Fermentation samples were analysed by HPLC. Glucose,xylose,xylitol,succinate,glycerol,acetate and ethanol were separated with an Aminex HPX-87H (Bio-Rad, Hercules,CA,USA) ion exchange column operated at 45 C,with a mobile phase of 5 mm H 2 SO 4 at a ow rate of 0.6 ml min 31 and detected using a refractive index detector (Shimadzu,Kyoto,Japan). Cell dry weight was determined by drying to constant weight in a microwave oven. 3. Results 3.1. Xylose batch fermentation The xylose consumption rate in oxygen-limited batch culture was calculated as the amount of xylose (g) consumed after 40^50 h,divided by time (h) and biomass (g) (Table 1). The xylose consumption rate was almost the same for all strains,ranging from 0.11 g (g biomass) 31 h 31 for TMB3014,overexpressing TAL1,to 0.14 g (g biomass) 31 h 31 for TMB3017,overexpressing RKI1. Production of xylitol,ethanol,acetate,and glycerol did not vary signi cantly between strains (results not shown). TMB3026,overexpressing all four non-oxidative PPP Table 1 Speci c xylose consumption rates (g xylose (g biomass) 31 h 31 ) calculated after 40^50 h of duplicate batch cultivation experiments per strain Strain Overexpressed gene(s) Xylose consumption rate TMB3001 0.13 TMB3013 RPE1 0.13 TMB3017 RKI1 0.14 TMB3014 TAL1 0.11 TMB3016 TKL1 0.12 TMB3026 RPE1, RKI1, TAL1, TKL1 0.12 The values are the average of two batch cultivations,which di ered by less than 10%. genes,showed an intermediate xylose consumption rate of 0.12 g (g biomass) 31 h 31 (Table 1). 3.2. Continuous cultivations with xylose and glucose Since only small di erences in xylose consumption rate were detected between strains overproducing the non-oxidative PPP enzymes in batch cultivation,tmb3026 was also compared with TMB3001 in anaerobic chemostat cultivation with 20 g l 31 glucose and 20 g l 31 xylose in the feed. Two steady states were obtained with TMB3001 and TMB3026 at dilution rates of 0.06 h 31 and 0.12 h 31 (Table 2). The two strains consumed nearly identical amounts of glucose and xylose. The product formation di ered by 2^ 10% between strains at the same dilution rate. The expected di erence between duplicates is about 5^8% in our experimental set-up,which makes the di erences between the strains insigni cant. The di erence in ethanol production is not signi cant,since nitrogen sparging of the fermenter makes this value less stable between duplicate experiments. 3.3. Xylulose batch fermentation The e ect of simultaneous overexpression of the four genes encoding enzymes of the non-oxidative PPP on xy- Table 2 Speci c consumption rates (negative values) and production rates (positive values) given as g (g biomass) 31 h 31 of substrates and products at dilution rates of 0.06 h 31 and 0.12 h 31 for TMB3001 and TMB3026 (overexpressing RKI1, RPE1, TAL1 and TKL1) TMB3001 TMB3026 D = 0.06 h 31 D = 0.12 h 31 D = 0.06 h 31 D = 0.12 h 31 Xylose 30.23 30.29 30.23 30.29 Glucose 30.61 31.24 30.63 31.19 CO 2 0.29 0.51 0.30 0.52 Ethanol 0.25 0.47 0.28 0.51 Xylitol 0.10 0.11 0.10 0.11 Glycerol 0.07 0.15 0.08 0.13 Acetate 2.4U10 33 4.2U10 33 n. d. n. d. Succinate 1.9U10 33 4.2U10 33 8.4U10 33 1.6U10 33 The presented values are the average of duplicate samples from two chemostat experiments per strain. n.d.,not detectable (below 0.5U10 33 g (g biomass) 31 h 31 ).

280 B. Johansson, B. Hahn-Ha«gerdal / FEMS Yeast Research 2 (2002) 277^282 Table 3 Xylulose concentration at di erent timepoints during 20-ml batch fermentation of 50 g l 31 xylulose and 5 g l 31 biomass with TMB3001 and TMB3026 Strain Xylulose concentration in fermenter (g l 31 ) 0 h 8 h 45 h 56 h 68 h TMB3001 57 47 37 32 24 TMB3026 54 30 5 2 0.4 The presented values are the average of two independent fermentation experiments with less than 10% di erence. lulose metabolism was also investigated. Xylulose is naturally metabolised by S. cerevisiae [2,3]. Table 3 shows the time course of xylulose consumption in oxygen-limited batch cultures of 50 g l 31 xylulose with TMB3001 and TMB3026. TMB3026 consumed xylulose considerably faster with a nal xylulose concentration of 0.4 g l 31 while TMB3001 left 24 g l 31 at the end of the fermentation (Table 3). Xylulose consumption rates in oxygen-limited batch culture were calculated after 45 h in the same way as for xylose fermentation rates. TMB3001 consumed 0.09 g (g biomass) 31 h 31 while TMB3026 consumed xylulose about two times faster (0.22 g (g biomass) 31 h 31, Table 4). The result shows that one or more of the non-oxidative PPP enzymes control the xylulose fermentation rate. TMB3026 gave slightly higher ethanol yield and nine times higher xylitol yield than TMB3001,while the glycerol yield was ve times higher in TMB3001 than in TMB3026 (Table 4). The acetate yield was similar in both strains. 3.4. Aerobic growth rates on di erent carbon sources Maximum speci c growth rates on di erent carbon sources were determined (Table 5) to identify physiological di erences resulting from non-oxidative PPP enzyme overproduction. The speci c growth rates of the various strains were similar with glucose (0.28^0.31 h 31 ) and xylose (0^0.03 h 31 ) as carbon sources (Table 5). The growth rate with xylulose was in the range of 0.01^0.03 h 31 for TMB3001 and strains individually overproducing RPE, RKI and TKL (Table 5). These values correspond well to literature data [2,28]. TMB3014,overexpressing TAL1,and TMB3026,overexpressing all four non-oxidative PPP genes,grew at a rate of 0.12 h 31 and 0.23 h 31, respectively. 4. Discussion In this investigation xylose and xylulose metabolism, respectively,were compared in two recombinant xylosefermenting strains of S. cerevisiae. TMB3001 expresses a xylose-utilising pathway comprising XR,XDH and XK, while TMB3026 in addition overexpresses the four nonoxidative PPP enzymes RPE,RKI,TAL and TKL. The aim of the investigation was to assess the control of the non-oxidative PPP on the rates of xylose utilisation and xylose fermentation. TMB3026 did not show a signi cantly higher xylose fermentation rate compared to the control strain TMB3001 in either chemostat or batch culture. Thus,the non-oxidative PPP did not have any signi cant control of the xylose fermentation rate in TMB3001. On the other hand,tmb3026 fermented xylulose twice as fast as TMB3001,suggesting that the nonoxidative PPP exerts signi cant control on xylulose metabolism. During xylose metabolism this control is probably concealed because other metabolic pathway steps constitute higher degrees of control. After transport across the plasma membrane xylose is metabolised to xylitol by XR, which is further metabolised to xylulose by XDH. Subsequent metabolic steps are shared for xylose and xylulose, indicating that other steps such as xylose transport,xr, and/or XDH are controlling the xylose consumption rate to such an extent that control by the non-oxidative PPP is not detectable during xylose fermentation. Control of the xylose fermentation rate may also be exercised by the supply of NAD(P)H for XR and/or NAD þ for XDH. A prime control point in xylose fermentation would be the oxidative PPP,since this pathway is the principal NADPH-generating pathway in the cell [29,30]. We have recently demonstrated that the xylose consumption rate drastically decreases with a non-functional oxidative PPP [31]. The increased xylitol yield in TMB3026 is probably due to NADH-dependent xylitol formation from xylulose by xylitol dehydrogenase,since the equilibrium favours xylitol production [32]. The NADH consumption rate due to combined glycerol and xylitol production in TMB3026 and TMB3001 was similar,1.24 and 1.35 mmol (g biomass) 31 h 31,respectively (calculated from Table 4). This suggests that TMB3026 maintains its NADH co-factor balance during xylulose fermentation by secreting xylitol rather than glycerol. Table 4 Xylulose consumption rate and ethanol,xylitol,acetate and glycerol yields after 45 h Strain Xylulose consumption rate (g (g biomass) 31 h 31 ) Yield (g (g consumed xylulose) 31 ) Ethanol Xylitol Acetate Glycerol TMB3001 0.09 0.33 0.02 0.03 0.10 TMB3026 0.22 0.39 0.17 0.02 0.02 The presented values are the average of two independent fermentation experiments with less than 10% di erence.

B. Johansson, B. Hahn-Ha«gerdal / FEMS Yeast Research 2 (2002) 277^282 281 Table 5 Aerobic speci c growth rates in de ned mineral medium with 10 g l 31 glucose,xylose or xylulose as carbon source Strain Overexpressed gene(s) W max (h 31 ) (S.D.,%) Glucose Xylose Xylulose TMB3001 0.32 (1) 0.01 (43) 0.03 (38) TMB3013 RPE1 0.31 (3) n.d. (^) 0.01 (79) TMB3017 RKI1 0.30 (4) 0.02 (28) 0.02 (36) TMB3014 TAL1 0.28 (2) 0.01 (20) 0.12 (6) TMB3016 TKL1 0.28 (2) n.d. (^) 0.02 (42) TMB3026 RPE1, RKI1, TAL1, TKL1 0.29 (6) 0.03 (27) 0.23 (3) Values are the average of two experiments performed in triplicate. The relative standard deviation is given as percentage of the average speci c growth rate. n.d.,not detectable (below 0.005 h 31 ). The maximum speci c growth rate on xylulose increased upon TAL overproduction,supporting the hypothesis that TAL activity limits pentose metabolism in S. cerevisiae [1,9,13]. Overproduction of TAL increased the aerobic growth on solid xylose media of a XYL1/XYL2-expressing strain [9] as measured by visual inspection of cell mass after 3 days. TAL overproduction did not improve the growth rate in liquid xylose medium (Table 5),possibly because growth was measured during a shorter time. Also, oxygen could be less available in liquid than on solid culture. Oxygen is known to increase the growth rate in xylose medium [33,34]. Overproduction of RPE had no e ect on either xylose or xylulose consumption rate,suggesting that the relatively e cient xylose and ribose co-metabolism [10,11] and the ux modelling results [12] cannot be ascribed to insu cient RPE activity. Overexpression of TKL1 has been reported to inhibit growth on fermentative carbon sources [35]. Similarly,expression of the P. stipitis TKT1 gene encoding transketolase in S. cerevisiae expressing XYL1 and XYL2 resulted in increased generation times of aerobic xylose growth [36]. None of the two strains overexpressing TKL1 reported here,tmb3016 and TMB3026,had impaired xylose metabolism (Table 1),neither did TKL1 overexpression impair growth of a XYL1/XYL2-expressing S. cerevisiae [9]. In addition to di erences in host strain properties and cultivation conditions,these discrepancies could be due to a lower speci c TKL activity in TMB3016 and TMB3026 (2.7^2.8 U (mg protein) 31 [24] and 0.92^0.97 U (mg protein) 31 [9]),compared with 6.7 U (mg protein) 31 reported by Metzger and Hollenberg [36]. The speci c growth rate on xylulose increased four times by overproducing TAL,but did not change by overproduction of RPE,RKI or TKL. Yet,the growth rate on xylulose increased up to eight times the reference level when all four genes were overproduced. This synergistic e ect suggests that overproduction of more than one enzyme is necessary in order to maximise strain performance and that the important enzymes cannot always be identi- ed by the e ect of overproducing a single enzyme [16,17,20,21]. Once the initial xylose metabolism has been improved,it is conceivable that higher xylose uxes could be obtained by additional overproduction of the non-oxidative PPP enzymes. The integrative vector system by which TMB3026 was created [24] facilitates virtually unlimited rounds of genetic engineering so any future genetic modi cations are easily carried out. Acknowledgements This work was nancially supported by The Swedish National Energy Administration and The Nordic Energy Research Programme. References [1] Ko«tter,P. and Ciriacy,M. (1993) Xylose fermentation by Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 38,776^783. [2] Wang,P.Y. and Schneider,H. (1980) Growth of yeasts on D-xylulose. Can. J. Microbiol. 26,1165^1168. [3] Wang,P.Y.,Shopsis,C. and Schneider,H. (1980) Fermentation of a pentose by yeasts. Biochem. Biophys. Res. Commun. 94,248^ 254. [4] Tantirungkij,M.,Nakashima,N.,Seki,T. and Yoshida,T. (1993) Construction of xylose-assimilating Saccharomyces cerevisiae. J. Ferment. Bioeng. 75,83^88. [5] Walfridsson,M.,Anderlund,M.,Bao,X. and Hahn-Ha«gerdal,B. (1997) Expression of di erent levels of enzymes from the Pichia stipitis XYL1 and XYL2 genes in Saccharomyces cerevisiae and its e ects on product formation during xylose utilisation. Appl. Microbiol. Biotechnol. 48,218^224. [6] Ligthelm,M.E.,Prior,B.A.,du Preez,J.C. and Brandt,V. (1988) An investigation of D-{1-13 C} xylose metabolism in Pichia stipitis under aerobic and anaerobic conditions. Appl. Microbiol. Biotechnol. 28, 293^296. [7] Yu,S.,Jeppsson,H. and Hahn-Ha«gerdal,B. (1995) Xylulose fermentation by Saccharomyces cerevisiae and xylose-fermenting yeast strains. Appl. Microbiol. Biotechnol. 44,314^320. [8] Jeppsson,H.,Yu,S. and Hahn-Ha«gerdal,B. (1996) Xylulose and glucose fermentation by Saccharomyces cerevisiae in chemostat culture. Appl. Environ. Microbiol. 62,1705^1709. [9] Walfridsson,M.,Hallborn,J.,Penttila«,M.,Kera«nen,S. and Hahn- Ha«gerdal,B. (1995) Xylose-metabolizing Saccharomyces cerevisiae strains overexpressing the TKL1 and TAL1 genes encoding the pen-

282 B. Johansson, B. Hahn-Ha«gerdal / FEMS Yeast Research 2 (2002) 277^282 tose phosphate pathway enzymes transketolase and transaldolase. Appl. Environ. Microbiol. 61,4184^4190. [10] van Zyl,C.,Prior,B.A.,Kilian,S.G. and Kock,J.L. (1989) D-Xylose utilization by Saccharomyces cerevisiae. J. Gen. Microbiol. 135,2791^ 2798. [11] van Zyl,C.,Prior,B.A.,Kilian,S.G. and Brandt,E.V. (1993) Role of D-ribose as a co-metabolite in D-xylose metabolism by Saccharomyces cerevisiae. Appl. Environ. Microbiol. 59,1487^1494. [12] Wahlbom,C.F.,Eliasson,A. and Hahn-Ha«gerdal,B. (2001) Intracellular uxes in a recombinant xylose-utilizing Saccharomyces cerevisiae cultivated anaerobically at di erent dilution rates and feed concentrations. Biotechnol. Bioeng. 72,289^296. [13] Senac,T. and Hahn-Ha«gerdal,B. (1990) Intermediary metabolite concentrations in xylulose- and glucose-fermenting Saccharomyces cerevisiae cells. Appl. Environ. Microbiol. 56,120^126. [14] Kacser,H. and Burns,J.A. (1973) The control of ux. Symp. Soc. Exp. Biol. 27,65^104. [15] Heinrich,R. and Rapoport,T.A. (1974) A linear steady-state treatment of enzymatic chains: General properties,control and e ectorstrength. Eur. J. Biochem. 42,89^95. [16] Fell,D.A. and Thomas,S. (1995) Physiological control of metabolic ux: The requirement for multi-site modulation. Biochem. J. 311,35^ 39. [17] Kacser,H. and Acerenza,L. (1993) A universal method for achieving increases in metabolite production. Eur. J. Biochem. 216,361^367. [18] Schaa,I.,Heinisch,J. and Zimmermann,F.K. (1989) Overproduction of glycolytic enzymes in yeast. Yeast 5,285^290. [19] Hauf,J.,Zimmermann,F.K. and Mu«ller,S. (2000) Simultaneous genomic overexpression of seven glycolytic enzymes in the yeast Saccharomyces cerevisiae. Enzyme Microb. Technol. 26,688^698. [20] Smits,H.P.,Hauf,J.,Mu«ller,S.,Hobley,T.J.,Zimmermann,F.K., Hahn-Ha«gerdal,B.,Nielsen,J. and Olsson,L. (2000) Simultaneous overexpression of enzymes of the lower part of glycolysis can enhance the fermentative capacity of Saccharomyces cerevisiae. Yeast 16, 1325^1334. [21] Niederberger,P.,Prasad,R.,Miozzari,G. and Kacser,H. (1992) A strategy for increasing an in-vivo ux by genetic manipulations. The tryptophan system of yeast. Biochem. J. 287,473^479. [22] Eliasson,A.,Christensson,C.,Wahlbom,C.F. and Hahn-Ha«gerdal, B. (2000) Anaerobic xylose fermentation by recombinant Saccharomyces cerevisiae carrying XYL1, XYL2,and XKS1 in mineral medium chemostat cultures. Appl. Environ. Microbiol. 66,3381^3386. [23] van Dijken,J.P.,Bauer,J.,Brambilla,L.,Duboc,P.,Francois,J.M., Gancedo,C.,Giuseppin,M.L.,Heijnen,J.J.,Hoare,M.,Lange, H.C.,Madden,E.A.,Niederberger,P.,Nielsen,J.,Parrou,J.L.,Petit, T.,Porro,D.,Reuss,M.,van Riel,N.,Rizzi,M.,Steensma,H.Y., Verrips,C.T.,Vindelov,J. and Pronk,J.T. (2000) An interlaboratory comparison of physiological and genetic properties of four Saccharomyces cerevisiae strains. Enzyme Microb. Technol. 26,706^714. [24] Johansson,B. and Hahn-Ha«gerdal,B. (2002) Overproduction of pentose phosphate pathway enzymes using a new CRE/loxP expression vector for repeated genomic integration in Saccharomyces cerevisiae. Yeast 19,225^231. [25] Mellor,J.,Dobson,M.J.,Roberts,N.A.,Tuite,M.F.,Emtage,J.S., White,S.,Lowe,P.A.,Patel,T.,Kingsman,A.J. and Kingsman, S.M. (1983) E cient synthesis of enzymatically active calf chymosin in Saccharomyces cerevisiae. Gene 24,1^14. [26] Verduyn,C.,Postma,E.,Sche ers,w.a. and Van Dijken,J.P. (1992) E ect of benzoic acid on metabolic uxes in yeasts: A continuousculture study on the regulation of respiration and alcoholic fermentation. Yeast 8,501^517. [27] Olsson,L.,Linde n,t. and Hahn-Ha«gerdal,B. (1994) A rapid chromatographic method for the production of preparative amounts of xylulose. Enzyme Microb. Technol. 16,388^394. [28] Richard,P.,Toivari,M.H. and Penttila«,M. (2000) The role of xylulokinase in Saccharomyces cerevisiae xylulose catabolism. FEMS Microbiol. Lett. 190,39^43. [29] Bruinenberg,P.M.,van Dijken,J.P. and Sche ers,w.a. (1983) An enzymic analysis of NADPH production and consumption in Candida utilis. J. Gen. Microbiol. 129,965^971. [30] Nogae,I. and Johnston,M. (1990) Isolation and characterization of the ZWF1 gene of Saccharomyces cerevisiae,encoding glucose-6- phosphate dehydrogenase. Gene 96,161^169. [31] Jeppsson,M.,Johansson,B.,Hahn-Ha«gerdal,B. and Gorwa-Grauslund,M.F. (2002) Reduced oxidative pentose phosphate pathway ux in recombinant xylose-utilizing Saccharomyces cerevisiae strains improves the ethanol yield from xylose. Appl. Environ. Microbiol. 68, 1604^1609. [32] Rizzi,M.,Harwart,K.,Erlemann,P.,Bui-Thanh,N.-A. and Dellweg,H. (1989) Puri cation and properties of the NAD þ -xylitol-dehydrogenase from the yeast Pichia stipitis. J. Ferment. Bioeng. 67, 20^24. [33] Skoog,K. and Hahn-Ha«gerdal,B. (1990) E ect of oxygenation on xylose fermentation by Picha stipitis. Appl. Environ. Microbiol. 56, 3389^3394. [34] Toivari,M.H.,Aristidou,A.,Ruohonen,L. and Penttila«,M. (2001) Conversion of xylose to ethanol by recombinant Saccharomyces cerevisiae: Importance of xylulokinase (XKS1) and oxygen availability. Metab. Eng. 3,236^249. [35] Sundstro«m,M.,Lindqvist,Y.,Schneider,G.,Hellman,U. and Ronne,H. (1993) Yeast TKL1 gene encodes a transketolase that is required for e cient glycolysis and biosynthesis of aromatic amino acids. J. Biol. Chem. 268,24346^24352. [36] Metzger,M.H. and Hollenberg,C.P. (1994) Isolation and characterization of the Pichia stipitis transketolase gene and expression in a xylose-utilising Saccharomyces cerevisiae transformant. Appl. Microbiol. Biotechnol. 42,319^325.