INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

A yeast strain capable of fermenting xylose and glucose to ethanol with high yields would increase the economic feasibility of fuel ethanol production from lignocellulosic biomass. Xylose fermentation by natural and recombinant yeasts has recently been reviewed (14, 17, 20). Saccharomyces cerevisiae, which is used for industrial ethanol production, cannot ferment xylose but can ferment its isomer, xylulose (44). In yeast, xylose reductase (XR) and xylitol dehydrogenase (XDH) catalyze the conversion of xylose to xylulose via the intermediate xylitol. Xylulokinase (XK), encoded by the gene XKS1 (32), phosphorylates xylulose to xylulose 5-phosphate, which is then metabolized through the pentose phosphate pathway and glycolysis. S. cerevisiae has been transformed with XYL1 and XYL2 from the xylose-fermenting yeast Pichia stipitis encoding XR and XDH, respectively (22, 37, 40, 42). Xylose fermentation by these recombinant strains of S. cerevisiae yields little ethanol, and xylitol is the major product (22, 37, 40, 42), perhaps due to limited XK activity in S. cerevisiae (6).

Saccharomyces sp. strain 1400(pLNH32), a fusion between Saccharomyces uvarum and Saccharomyces diastaticus (4), which overexpresses XYL1, XYL2, and XKS1, had an estimated ethanol yield of 0.44 carbon-millimole (c-mmol)/c-mmol in complex medium (18). Recently, S. cerevisiae CEN.PK overexpressing XYL1, XYL2, and XKS1 was quantitatively characterized under anaerobic conditions in defined media fermenting mixtures of glucose and xylose (9). This strain, TMB3001, gave an ethanol yield of 0.27 c-mmol/c-mmol. How much of the difference in ethanol yield is due to overexpression of XKS1, to media composition and to strain background is not known.

Recombinant xylose utilizing S. cerevisiae strains have been characterized in yeast extract-peptone (YP) complex medium (18, 37) and in defined medium (9, 22, 42). The use of YP limits the interpretation of the fermentation results, since YP medium contains all of the cellular components of yeast grown on hexose sugars, including some hexose sugars. Thus, components of the YP media are cofermented with xylose and enhance product yields. Furthermore, YP is too expensive for use in industrial ethanol production (46), which makes YP unsuitable for characterizing the performance of novel xylose-fermenting recombinant yeast strains.

Strains of S. cerevisiae differ in their ability to ferment xylulose (8), suggesting inherent differences in their capacities to ferment pentose sugars. Recently, a majority of yeast laboratories within the European Community agreed to use S. cerevisiae strain CEN.PK (10) as a reference strain (43). CEN.PK is a laboratory strain specifically designed for physiological and genetic research, including the development of metabolic engineering strategies (10). CEN.PK grows well on various carbon sources, sporulates efficiently, and is available with many different markers and genotypes (10). A recombinant xylose-fermenting S. cerevisiae CEN.PK strain is now available (9), but S. cerevisiae H158 (31) has been extensively used in earlier studies (15, 16, 26, 27, 28, 40, 41, 42).

In this study, we overexpressed XKS1 along with XYL1 and XYL2 in two strains of S. cerevisiae, H158 (31) and CEN.PK (10), to quantitatively determine the contribution of XKS1 overexpression and strain background, respectively, on the ethanolic fermentation of xylose. Sugar consumption and product formation in defined mineral medium, complex medium, and a birch wood hydrolysate were also monitored to quantify the contribution of YP and lignocellulose-derived fermentation inhibitors, respectively, to ethanol production in strains overexpressing XKS1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains and plasmids. All yeast and bacterial strains were maintained at 8°C on solid cloning medium (described below) and Luria-Bertani (LB) medium (1) with 100 mg of ampicillin liter¹, respectively. The XKS1 gene was cloned from S. cerevisiae CBS 8066. We obtained S. cerevisiae CEN.PK2-1C (MAT leu2-3 leu2-112 ura3-52 trp1-289 his3-1 MAL2-8^c SUC2) (10) from Echard Boles (University of Duesseldorf, Duesseldorf, Germany) and S. cerevisiae GPY55-15 (leu2-3 leu2-112 ura3-52 trp1-289 his4-519 prb1 cir⁺) from Greg Payne (University of California, Berkeley) (31), transformed them with plasmid pY6 (40), and named them CEN.PK and H158, respectively. pY6 contains XYL1 controlled by the ADH1 promoter, XYL2 controlled by the PGK1 promoter, a yeast 2µm multicopy ORI, and the URA3 marker for uracil prototrophy. The ADH1 promoter is weaker than the PGK1 promoter, resulting in lower XR than XDH activity (40). The yeast strains were also transformed with the integrative plasmid pDF1 (23), resulting in inactivation of the chromosomal FUR1 gene by gene replacement. Since a fur1 strain must have an active URA3 gene to survive, even in the presence of uracil, the FUR1 inactivation made the strains maintain the pY6 plasmid even under nonselective conditions. XKS1 was first subcloned in the vector YEp24PGK (40) and then transferred to the vector YEplac112 (11). Escherichia coli DH5 [F 80dlacZM15 (lacZYA-argF)U169 deoR recA1 endA1 hsdR17(r_K m_K⁺) supE44 thi-1 gyrA96 relA1] (Life Technologies, Rockville, Md.) was used for subcloning.

Cloning of the XKS1. XKS1 was amplified from S. cerevisiae CBS 8066 chromosomal DNA (13) with a 5' primer (primer 1, 5'-GCGGATCCTCTAGAATGGTTTGTTCAGTAATTCAG-3') and either one of two 3' primers (primer 2 [5'-AGATCTGGATCCTTAAGGGGACAATCTTGG-3'] or primer 3 [5'-AGATCTGGATCCTTAGATGAGAGTCTTTTCCAG-3']). Primer 1 was designed from published sequence information (19, 32). Primer 2 was designed to amplify a 1,776-bp open reading frame (ORF) (19), and primer 3 was designed to anneal 27 bp further downstream on the same sequence, yielding a 1,803-bp ORF (32). Both ORFs have been claimed to encode a protein with XK activity (19, 32). The complementary sequences are underlined. BamHI restriction sites used for cloning are shown in boldface. Primer 1 introduced base substitutions at positions 3 (T to A) and at 2 (T to G) to maximize translational efficiency, where +1 is the A in the start codon (7, 25). Similarly, the codon of the N-terminal amino acid was altered. Protein stability in S. cerevisiae depends partly on the N-terminal amino acid of the protein; the half-life can range from minutes to several hours (2). Thus, we changed the XKS1 N-terminal amino acid from a destabilizing (TTG, Leu) to a stabilizing (GTT, Val) one. The PCR product was ligated in the BglII site between the PGK1 promoter and terminator (29) in plasmid YEp24PGK using BamHI sites present on extra nucleotides added onto the primers, resulting in YEp24PGK/XK. The expression cassette containing promoter, gene, and terminator was cut out with BamHI and SmaI and ligated into YEplac112 (11) using the same sites, resulting in pXks. CEN.PK and H158 carrying the pY6 plasmid were transformed with pXks, resulting in CEN.PK-pXks and H158-pXks.

Transformations. S. cerevisiae was transformed using the lithium acetate method (12), and E. coli DH5 was transformed with the calcium chloride method (33).

Cloning media. Yeast strains were grown in SD medium (35) supplied with 250 mg of L-leucine, 50 mg of L-tryptophan, 50 mg of L-histidine, and 50 mg of uracil per liter. Transformants were selected by omission of the appropriate amino acids or nucleotide. Bacterial strains were grown in LB medium (1), and transformants were selected by adding 100 µg of ampicillin ml¹.

Lignocellulose hydrolysate. A birch wood lignocellulose hydrolysate was provided by Robert Eklund (Mid Sweden University, Örnsköldsvik, Sweden). It was prepared by mixing birch wood (10 kg) with water and concentrated sulfuric acid (5 g liter¹) to a total of 30 kg and hydrolyzed for 7 min at 188°C by adding steam (36). The hydrolysate was adjusted to pH 5.5 and filter sterilized (0.45 µm [pore size]). The hydrolysate contained 39 g of xylose, 5.7 g of glucose, 3.5 g of mannose, 3.1 g of galactose, 1.7 g of arabinose, 0.16 g of hydroxy methyl furfural (HMF), and 0.7 g of furfural per liter based on high-pressure liquid chromatography (HPLC) analysis (see below).

Fermentation media. Defined minimal (45) or complex media were used with either birch wood hydrolysate or a mixture of sugars with the same sugar composition as the birch wood hydrolysate (Table 1). Amino acids L-histidine and L-tryptophan, each at 50 mg liter¹, were added to complement amino acid auxotrophy in defined media. Complex medium contained 10 g of yeast extract and 20 g of peptone per liter in addition to the carbon source. Fermentation using xylose as sole carbon source was conducted with 80 g of xylose per liter. A high xylose concentration was used to overcome the absence of a specific xylose transport system in S. cerevisiae (21, 28).

Preparation of inoculum. The inoculum for batch fermentation was prepared by adding a single colony to a 500-ml shaking flask containing 200 ml of defined medium. After overnight incubation at 30°C and 120 rpm, the culture was harvested by centrifugation at 6,400 × g for 10 min at 4°C and used to inoculate up to eight 1-liter shaking flasks containing 500 ml of defined medium. The cells were grown to an optical density at 620 nm (OD₆₂₀) of 4 to 5, harvested at 4°C by centrifugation (6,400 × g for 10 min), and resuspended in ice-cold 0.9% (wt/vol) NaCl. An aliquot of this cell suspension was centrifuged at 6,400 × g for 10 min and resuspended in 25 ml of ice-cold medium of the same type to be used in fermentation.

Fermentation. Fermentation was conducted batchwise in 120-ml fermentors, with a 100-ml working volume and magnetic bar stirring (100 rpm), and 10 g (dry weight) of inoculum per liter, at 30°C by water jacket and at pH 5.5 by the addition of NaOH.

Analysis of substrates and products. Samples for quantification of substrates and products were analyzed by HPLC. Xylose, glucose, mannose, galactose, arabinose, and xylitol were separated using an HPX-87P ion-exchange column (Bio-Rad Laboratories, Hercules, Calif.) operated at 85°C using water as the mobile phase at 0.6 ml min¹ and determined with a refractive index detector (Shimadzu, Kyoto, Japan). Ethanol, glycerol, acetate, xylulose, HMF, and furfural were separated using an Aminex HPX-87H (Bio-Rad) ion-exchange column operated at 45°C, with a mobile phase of 5 mM H₂SO₄ at a flow rate of 0.6 ml min¹, and detected using a refractive index detector (Shimadzu)except for HMF and furfural which were detected using a SPD 6A UV detector (Shimadzu).

Enzyme activity analyses. Cell extracts were prepared from 50-ml batch cultures grown to an OD₆₂₀ of 4 to 5. The cells were lysed in 100 mM triethanolamine buffer (pH 7.0), 1 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol by vortexing them twice with glass beads for 5 min each time at 4°C, with cooling on ice for 5 min in between vortexing. Bovine serum albumin (5 mg ml¹) was added after the cells were lysed to increase protein stability. The protein content in cell extracts was measured by the method of Bradford (3). XR, XDH, and XK activities were measured as previously described (9). XK activity was determined in two steps. First, the XDH activity was determined in the absence of ATP, and then the sum of the XK and XDH activities in the presence of ATP was determined, the XK activity being the difference (9). All enzyme activity measurements were performed at 30°C. The absorbance change per minute (A min¹) was divided by the molar absorptivity for NADH (6.22 cm¹µmol¹) to calculate substrate consumption per minute. One unit of enzyme activity is defined as 1 µmol of substrate converted per min for all assays.

Calculations. Carbon balances were calculated using single carbon unit equivalents (i.e., c-mmol) (5) consumed and produced after 65 h to allow comparison of hexose and pentose sugar metabolism. Yields were expressed as c-mmol/c-mmol. The carbon balance was calculated assuming 1 c-mmol of CO₂ produced for every 2 c-mmol of ethanol and acetate produced, according to the metabolic stoichometry. No cell growth, as measured by dry weight determination, occurred during fermentation (data not shown). Biomass production was omitted from the carbon balance calculations for this reason. The specific xylose consumption rate was calculated as the c-mmol gram (cell dry weight)¹ hour¹, as based on the amount of xylose consumed after 65 h.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning of the xylulokinase gene (XKS1). Two sequences have been reported to encode XK activity in S. cerevisiae; one with an ORF of 1,776 bp (19) and another one that is similar but slightly longer, with an ORF of 1,803 bp, originally designated YGR194c (accession no. Z72979) (38) and now called XKS1 (32). PCR amplification of the XKS1 gene from S. cerevisiae CBS 8066 using primers 1 and 2 (see Materials and Methods) generated the 1,776-bp ORF, whereas the use of primers 1 and 3 generated the 1,803-bp ORF. Both ORFs were fused to the PGK1 promoter and terminator in the YEp24PGK vector. When transformed into either S. cerevisiae CEN.PK or H158, the 1,803-bp ORF resulted in more than 300 times higher XK activity (Table 2), whereas the 1,776-bp ORF did not cause any measurable increase in XK activity (results not shown). The 1,803-bp sequence was used throughout this work. The 1,776-bp sequence was considered incomplete and did not code for an active enzyme since several PCR products of the 1,776-bp ORF were cloned to rule out the possibility of PCR errors. When Saccharomyces sp. strain 1400(pLNH32) was constructed, a larger piece of DNA was cloned (18, 19), which yielded the complete gene, whereas the 1,776-bp ORF (19) is too short to generate an active enzyme.

XR, XDH, and XK activity. Enzyme activities were measured under conditions used for preparation of inoculum (Table 2). XR activities were 0.7 to 0.8 U mg¹ and were similar in all strains. XDH activities were higher in CEN.PK strains than in H158 strains (18.2 to 18.9 and 13.9 to 15.3 U mg¹, respectively). A low XR/XDH ratio was deliberately chosen, since xylitol production is reduced in such strains compared to strains in which the activity ratio is high (40). The XR and XDH activities were higher than previously reported (40), which may be due to the use of different growth media. XK activities increased at least 300-fold to 28 to 36 U mg¹ from overexpression of XKS1 on a multicopy plasmid under the PGK1 promoter-terminator sequences (Table 2). When XKS1 was chromosomally integrated, the specific XK activity was only 2 U mg¹ (9). The XK activity cannot be directly compared to that seen in other reports (6, 18), since an assay (34) determining the sum of XK and XDH activity was used (see Materials and Methods).

Xylose fermentation in defined and complex media. CEN.PK-pXks consumed more xylose than H158-pXks and generated slightly higher ethanol concentrations and considerably higher final xylitol concentrations (Table 3). Both strains consumed more xylose in complex medium than in the defined medium. H158-pXks had a higher ethanol yield and a lower xylitol yield than CEN.PK-pXks.

Sugar mixture fermentation in defined and complex media. A sugar mixture reflecting the sugar composition of a birch wood hydrolysate (see Materials and Methods) was fermented by H158, H158-pXks, CEN.PK, and CEN.PK-pXks in both defined and complex media (Table 4). The xylose consumption rate was initially much higher for H158 than for H158-pXks, whereas H158-pXks consumed xylose at a constant rate throughout the fermentation (Fig. 1). Ethanol, xylitol, and acetic acid were produced when xylose was consumed, whereas glycerol production was not related to xylose consumption.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Fermentation of a sugar mixture (see Materials and Methods) by S. cerevisiae H158 expressing XYL1 and XYL2 (A) and by S. cerevisiae H158-pXks also overexpressing XKS1 (B) in complex medium. The concentrations of xylose () are indicated on the left axis, and the concentrations of ethanol (), acetate (), and glycerol () are indicated on the right axis in both panels. The concentration of xylitol () is on the left axis in panel A and in the right axis in panel B. Glucose, mannose, and galactose were omitted for clarity since they were consumed within 5 h. Duplicate fermentation experiments differed less than 10%.

Hydrolysate fermentation in defined and complex media. Sugar consumption was reduced in fermentation of birch wood hydrolysate compared to a medium without hydrolysate, a finding possibly due to inhibitory components in the birch wood hydrolysate (24, 36) (Tables 4 and 5). Xylitol production was reduced in hydrolysate media, more so for CEN.PK than for H158. Complex medium promoted xylose consumption in birch wood hydrolysate. CEN.PK-pXks consumed slightly more xylose than H158-pXks in hydrolysate media but also produced more xylitol, as was the case in the sugar mixture medium. CEN.PK and CEN.PK-pXks showed higher ethanol yields on consumed sugars with hydrolysate than without hydrolysate (Tables 4 and 5), whereas the ethanol yields were lower for H158-pXks.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

XKS1 overexpression in recombinant S. cerevisiae improved ethanolic fermentation of xylose because the xylitol yield decreased in all media we examined. However, XKS1 overexpression also reduced xylose consumption considerably. The final ethanol concentration was only marginally affected, since reduced xylose consumption was balanced by reduced xylitol production. The decreased xylose utilization in H158-pXks and CEN.PK-pXks may have been caused by uncontrolled XK activity since XKS1 was overexpressed under the control of the strong PGK1 promoter, which overrules possible feedback control of xylulose phosphorylation. Teusink et al. (39) suggested that uncontrolled sugar kinase activity in the beginning of a metabolic pathway could lead to abnormal accumulation of sugar phosphates and concomitant depletion of the intracellular ATP pool. A mathematical model (39) has been developed in which glucose utilization is reduced by lack of feedback control of hexokinase and phosphofructokinase activity. This lack of control also led to glucose-6-phosphate and fructose-1,6-bisphosphate accumulation and ATP depletion. The model showed that sugar phosphate accumulation and ATP depletion could be relieved by lower hexokinase activity. By analogy with the effect of uncontrolled hexokinase activity, XKS1 overexpression may retard cellular metabolism by xylulose-5-phosphate accumulation and/or ATP depletion.

The level of XK activity may be crucial for the xylose uptake rate and the subsequent ethanolic fermentation. When XYL1, XYL2, and XKS1 were integrated into the his3 locus of a CEN.PK strain, yielding strain TMB3001 (9), the XK activity was only about 2 U mg of protein¹ compared to ca. 30 U mg of protein¹ in H158-pXks and CEN.PK-pXks. Strain TMB3001 had a maximum xylose consumption rate of 6.8 c-mmol g (cell dry weight)¹ h¹ compared to 1 c-mmol g (cell dry weight)¹ h¹ for CEN.PK-pXks (Table 4). For Saccharomyces sp. strain 1400(pLNH32), with an XK activity of 0.1 U mg¹, the xylose consumption rate has been estimated to be 14.3 c-mmol g (cell dry weight)¹ h¹ (9, 18). However, in this strain the XK activity is not directly comparable to our results, since an assay also measuring XDH activity was used (34). When XKS1 was overexpressed in S. cerevisiae FY1679 and W303, the XK activity was not reported (32). However, whereas aerobic growth on xylulose was reduced, growth on glucose was unaffected. These results suggest that it is necessary to carefully modulate the XK activity to achieve efficient xylose fermentation by recombinant S. cerevisiae.

Complex medium overcame the inhibitory effects of lignocellulose-derived inhibitors so that the specific xylose consumption and the ethanol yield increased when YP medium was added to birch wood hydrolysate. Complex medium had little effect on xylose fermentation in the absence of inhibitory hydrolysate. However, yeast extract and peptone are too expensive for industrial ethanol production from lignocellulosic hydrolysate (46). Xylitol production was lower in hydrolysate media than in sugar mixture media, possibly due to the reduction of furfural to furfuryl alcohol, which may provide the XDH reaction with reduced cofactors (30).

The host strain influenced the efficiency of the xylose fermentation by the resulting recombinant strain. The highest ethanol yield, 0.51 c-mmol/c-mmol, was obtained with H158-pXks in complex medium (Table 4). H158 is a laboratory strain used extensively as a host for genetic engineering for xylose fermentation (15, 16, 26, 27, 28, 40, 41, 42). CEN.PK is a laboratory strain, specifically designed for physiological and genetic research, including the development of metabolic engineering strategies (10). CEN.PK grows well on various carbon sources, sporulates efficiently, and is available with many different markers and genotypes (10). CEN.PK has been chosen as a standard strain for laboratories within the European Community (43). CEN.PK-pXks consumed more xylose than H158-pXks in the fermentation of xylose only and of the sugar mixture. CEN.PK-pXks also performed better in hydrolysate media than did H158-pXks, with higher sugar consumption and higher ethanol production. Furthermore, when H158-pXks fermented hydrolysate media, the ethanol yield decreased compared with the level of fermentation in other media.

Our study shows that the results of a metabolic engineering strategy aimed at introducing a new metabolic pathway is highly dependent on the choice of host strain and the modulation of overexpressed genes. In particular, our results on the deleterious effect of uncontrolled XKS1 overexpression showed that it is necessary to quantify the effect of individual genetic modifications introduced in a metabolic pathway. Medium composition influenced the results to a lower extent. Even so, any quantitative characterization of new metabolically engineered strains must be considered incomplete if not also performed in defined mineral medium.