SEARCH

SEARCH BY CITATION

Keywords:

  • xylose fermentation;
  • Saccharomyces cerevisiae;
  • glucose-6-phosphate dehydrogenase;
  • xylitol;
  • redox balance;
  • inhibitors;
  • synthetic promoter library;
  • NADPH

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Disruption of the ZWF1 gene encoding glucose-6-phosphate dehydrogenase (G6PDH) has been shown to reduce the xylitol yield and the xylose consumption in the xylose-utilizing recombinant Saccharomyces cerevisiae strain TMB3255. In the present investigation we have studied the influence of different production levels of G6PDH on xylose fermentation. We used a synthetic promoter library and the copper-regulated CUP1 promoter to generate G6PDH-activities between 0% and 179% of the wild-type level. G6PDH-activities of 1% and 6% of the wild-type level resulted in 2.8- and 5.1-fold increase in specific xylose consumption, respectively, compared with the ZWF1-disrupted strain. Both strains exhibited decreased xylitol yields (0.13 and 0.19 g/g xylose) and enhanced ethanol yields (0.36 and 0.34 g/g xylose) compared with the control strain TMB3001 (0.29 g xylitol/g xylose, 0.31 g ethanol/g xylose). Cytoplasmic transhydrogenase (TH) from Azotobacter vinelandii has previously been shown to transfer NADPH and NAD+ into NADP+ and NADH, and TH-overproduction resulted in lower xylitol yield and enhanced glycerol yield during xylose utilization. Strains with low G6PDH-activity grew slower in a lignocellulose hydrolysate than the strain with wild-type G6PDH-activity, which suggested that the availability of intracellular NADPH correlated with tolerance towards lignocellulose-derived inhibitors. Low G6PDH-activity strains were also more sensitive to H2O2 than the control strain TMB3001. Copyright © 2003 John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Saccharomyces cerevisiae efficiently converts the hexose sugars present in lignocellulosic hydrolysates to ethanol. However, it cannot utilize pentoses such as xylose. The xylose pathway has therefore been integrated into S. cerevisiae. The XYL1 and XYL2 genes from Pichia stipitis, encoding xylose reductase (XR) and xylitol dehydrogenase (XDH), respectively, as well as endogeneous XKS1 encoding xylulokinase (XK), have been chromosomally integrated into S. cerevisiae, resulting in the stable xylose-fermenting strain TMB3001 (Eliasson et al., 2000). Even though TMB3001 produces ethanol from xylose (0.21 g/g xylose), considerable amounts of xylitol are formed as well (0.2–0.4 g/g xylose). Xylitol formation has been attributed to the co-factor imbalance generated by the NAD(P)H-dependent XR and NAD+-dependent XDH reactions (Bruinenberg et al., 1983a).

The oxidative pentose phosphate pathway (PPP) is a major source of NADPH in yeast, but NADPH is also produced via the NADP+-linked isocitrate dehydrogenase reaction (Bruinenberg et al., 1983b). The ZWF1 gene encoding glucose-6-phosphate dehydrogenase has been disrupted in strain TMB3255 to decrease the NADPH level (Jeppsson et al., 2002). TMB3255 showed an enhanced ethanol yield (0.41 g/g xylose) and a decreased xylitol yield (0.05 g/g xylose), however, at the expense of a six-fold lower xylose consumption rate. The xylose consumption rate in TMB3255 has been enhanced by increasing the production of xylose reductase (Jeppsson et al., 2003). Another strategy to maintain a high ethanol yield without losing the xylose flux would be to allow a low flux through the oxidative PPP.

In the present investigation we constructed strains with a range of production levels of G6PDH to generate a range of intracellular NADPH levels. The copper promoter, CUP1 (Labbe and Thiele, 1999) was used to facilitate the regulation of the ZWF1 gene expression by changing the Cu2+ concentration of the medium. Additionally, a synthetic promoter library was constructed for expression of the ZWF1 gene at a wider span of activities. Furthermore, a transhydrogenase-producing strain was included to investigate the effects of the reaction converting NADPH and NAD+ into NADP+ and NADH (Nissen et al., 2001) on xylose fermentation. Ethanol and xylitol yields, as well as the xylose consumption rate were monitored.

Strains with low G6PDH-activity are likely to have reduced levels of NADPH. This may influence their tolerance towards inhibitors present in lignocellulose hydrolysates. For example, 5-hydroxymethyl furfural (5-HMF) requires NADPH for reduction (Wahlbom and Hahn-Hägerdal, 2002) and this may result in reduced growth rate. In this investigation, growth rates were determined in a lignocellulose hydrolysate. Strains disrupted in the oxidative pentose phosphate pathway have also shown enhanced sensitivity to oxidative stress (Nogae and Johnston, 1990). The tolerance towards hydrogen peroxide was therefore also evaluated for strains with a range of different G6PDH activities.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Strains

Strains of S. cerevisiae used in this investigation are summarized in Table 1. Escherichia coli DH5α (Life Technologies, Rockville, MD) was used for sub-cloning. All strains were stored in 20% glycerol at −80 °C. Yeast cells from freshly streaked YPD plates (Ausubel et al., 1995) were used for inoculation.

Table 1. Strains used in this study
StrainGenotypeReference
TMB3001CEN.PK 113-7A (MATahis3-Δ1 MAL2-8c SUC2) his3::YIp XR/XDH/XK(Eliasson et al., 2000)
TMB3255TMB3001 zwf1::KanMX(Jeppsson et al., 2002)
TMB3256TMB3001 promoterless ZWF1This work
TMB3037TMB3001 YRP13 ZWF1This work
TMB3035TMB3001 YRP25 ZWF1This work
TMB3034TMB3001 YRP34 ZWF1This work
TMB3030TMB3001 CUP1 ZWF1This work
TMB3253TMB3001 YEplac112 KanRThis work
TMB3254TMB3001 YEplac112 PGK CTH KanRThis work

Nucleic acid manipulation

Plasmid DNA was prepared with BioRad Plasmid Miniprep kit (Hercules, CA). Restriction and modification enzymes were obtained from Roche (Roche Diagnostics AB, Bromma, Sweden), Fermentas (Vilnius, Lithuania) and Life Technologies (Rockville, MD). DNA extractions from agarose gel were made by QIAGEN Gel Extraction Kit (QIAGEN GmbH, Hilden, Germany).

Transformation

Competent cells of E. coli DH5α were prepared and transformed as described elsewhere (Inoue et al., 1990) and yeast transformations were made using a lithium acetate method (Gietz et al., 1992), which was slightly modified (Güldener et al., 1996). E. coli transformants were selected on Luria–Bertani (LB) medium plates (Ausubel et al., 1995) with 100 µg/ml ampicillin (IBI Shelton Scientific Inc., Shelton, CT). S. cerevisiae transformants were selected on Yeast Nitrogen Base without amino acids (Difco, Sparks, MD) or on YPD plates with 100 µg/ml zeocin (Invitrogen, Groningen, The Netherlands) or 150 µg/ml geneticin (Life Technologies, Rockville, MD).

Construction of the pB3 ZWF1 and pB3 CUP1 ZWF1 vectors

Approximately 75% of the 5′ end of ZWF1 was cloned by PCR from S. cerevisiae CBS 8066 chromosomal DNA using the primers 5ZWF1clon 5′-GAGGATCCAGAATGAGTGAAGGCCCCGTCAAATTC-3′ and 3ZWF1clon 5′-GAGGATCCCTGCACTCTGATGACCAGTTCG-3′, adding BamHI sites at both ends (bold face). Only the 5′ part was cloned to allow simultaneous integration and disruption of the wild-type copy of the ZWF1 gene. The flanking BamHI sites were cleaved at the 1150 bp PCR product. The partial ZWF1 ORF was ligated to the pB3 PGK, cut with BglII, resulting in pB3 PGK ZWF1. The PGK1 promoter was removed by restriction cleavage with SacI and XbaI, resulting in the linear fragment pB3 ZWF1′.

The cohesive ends of the remaining promoterless plasmid were blunted by Klenow DNA polymerase. The resulting vector was closed by ligation with T4 DNA ligase, generating pB3 ZWF1 vector.

The CUP1 promoter was released from the plasmid pCu413 CUP1 (Labbe and Thiele 1999) by restriction cleavage with SacI and SpeI. The CUP1 promoter was ligated to pB3 ZWF1′, resulting in pB3 CUP1 ZWF1.

Oligonucleotides for synthetic promoter design

Two synthetic oligonucleotides (DNA Technology A/S, Aarhus, Denmark) were used to generate synthetic promoters: oligonucleotide 1: 5′ATCAGAATTCTCGAGNNNNNCTTCCNNNNNACCCATACANNNNNNNNACCCATACANNNNNCTTCCNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNTATAAANNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNCTTCTTTCTTGTAACATC3′(159-mer) and oligonucleotide 2: 5′ATCGGGATCCATTTTGATTTAGTGTTTGTGTGTTGATAAGCAGTTGCTTGGTTTTTTA TGAAAAATAGCTAGAAGGAATAAGGAATTACAAGAGAGATGTTACAAGAAAGAAG3′ (113-mer). The underlined sequences are complementary, facilitating annealing of the downstream part of oligonucleotides 1 and 2.

Synthesis of synthetic promoters

To create double-stranded DNA, 0.6 nmol oligonucleotide 1 and 1.04 nmol oligonucleotide 2 were mixed with 100 nmol of each dNTP and incubated for 2 h at 37 °C with 23.6 units Klenow DNA polymerase (Pharmacia Biotech, Uppsala, Sweden) in a total volume of 200 µl. The double-stranded DNA was purified by agarose gel electrophoresis and cleaved with EcoRI and BamHI.

Construction of the pYLZ-2 YRP vectors

The pYLZ-2 vector (Hermann et al., 1992) containing the E. coli lacZ gene was cut with EcoRI and BamHI and the synthetic double-stranded promoters were ligated to the vector, resulting in the pYLZ-2 YRP vectors. Positive clones were identified by PCR using the primers 5YRPclon and 3YRPclon, as described below, and E. coli colonies as template. Plasmid DNA was prepared from 37 positive clones, designated pYLZ-2 YRP 1–37, and used for transformation of S. cerevisiae CEN.PK 113-5D (Entian and Kötter, 1998). Transformants were selected by uracil prototrophy.

Construction of the pB3 YRP ZWF1 vectors

Three clones of the synthetic promoter library (YRP13, YRP25 and YRP34) were used for PCR-amplification with the primers 5YRPclon 5′-GATCGAGCTCTGGCCGATTCATTAATCCAGCTGAA-3′ and 3YRPclon 5′-GATCTCTAGATTTGATTTAGTGTTTGTGTGTTGAT-3′. The synthetic promoters were inserted into the promoterless pB3 ZWF1′ vector after cleavage with SacI and XbaI (bold face), generating pB3 YRP13 ZWF1, pB3 YRP25 ZWF1 and pB3 YRP34 ZWF1.

Construction of TMB3256, TMB3030, TMB3037, TMB3035 and TMB3034

The vectors pB3 ZWF1, pB3 CUP1 ZWF1, pB3 YRP13 ZWF1, pB3 YRP25 ZWF1 and pB3 YRP34 ZWF1 were cleaved with BglII within the ZWF1 gene. The cleavage products were used for transformation of TMB3001, generating TMB3256, TMB3030, TMB3037, TMB3035 and TMB3034 after selection for zeocin resistance. Integration was confirmed by PCR amplification with the pB3 PGK-specific primer 5′-GAAGTTATTAGGTGATATCAGATCC-3′ and the ZWF1-specific primer 3ZWF1clon.

Construction of TMB3253 and TMB3254

The cth gene encoding the Azotobacter vinelandii cytoplasmic transhydrogenase was removed from YEp24-pPGK-CTH (Nissen et al., 2001) using HindIII and inserted into YEplac112 (Gietz and Sugino, 1988) using the same site, resulting in YEplac112 PGK CTH. The KanMX gene was removed from pFA6–KanMX4 (Wach et al., 1994) using BamHI and SacI, and thereafter inserted in YEplac112 and YEplac112 PGK CTH using the same sites, resulting in YEplac112 KanR and YEplac112 PGK CTH KanR. TMB3001 was transformed with YEplac112 KanR and YEplac112 PGK CTH KanR, resulting in TMB3253 and TMB3254, respectively, after selection for geneticin resistance.

Batch fermentation

Batch fermentation under oxygen-limited conditions in a defined mineral medium (Verduyn et al., 1992) with xylose as sole carbon source was performed as previously described (Jeppsson et al., 2002). Sampling occurred before xylose was totally consumed. Geneticin (150 µg/ml) was added in the batch-fermentation with TMB3253 and TMB3254.

Maximal specific growth rates in the presence of 2% lignocellulose hydrolysate (Nilvebrant et al., 2003), 20 g/l glucose, and 0.1 mM methionine were determined using a defined mineral medium (Verduyn et al., 1992) and in-house-built fermentors. Cultivation was conducted anaerobically by sparging with 0.6 l/min N2 gas at 30 °C, 250 rpm stirring, pH 5.5 controlled by autoaddition of 3 M NaOH and a total volume of 750 ml. For TMB3030, 10 µM Cu2+ was added to generate a G6PDH-activity above the one of the wild-type.

Enzymatic measurements

Crude extracts for G6PDH-activity determination were made using the Y-PER reagent (Pierce, Rockford, IL). Crude extracts for TH-activity determination were made with 0.5 mm glass beads in disintegration buffer (100 mM triethanolamine, 0.5 mM EDTA, 0.5 mM DTT and 1 mM PMSF) through 3 × 5 min of vigorous vortexing. Protein concentration was determined using Coomassie Protein Assay Reagent (Pierce), with bovine serum albumin as standard. The glucose-6-phosphate dehydrogenase activity (G6PDH; EC 1.1.1.49) was measured according to Bergmeyer (1974). β-Galactosidase activities were measured in whole cells (Miller, 1972) and the transhydrogenase activity was measured according Voordouw et al. (1979).

Enzyme activities for strains used in batch fermentation were measured in precultures, since the cells are in a better physiological condition before than after fermentation with xylose as sole carbon source.

Analysis of substrates and products

Xylose, xylitol, glycerol, acetate and ethanol concentrations were determined by column liquid chromatography (CLC) as previously described (Jeppsson et al., 2002). Cell dry weight was determined as previously described (Jeppsson et al., 2002).

Sensitivity towards H2O2

Minimal inhibitory concentrations (MIC) of H2O2 were determined in 2 ml YPD medium kept overnight in test tubes at 30 °C and 140 rpm. MIC was defined as the lowest H2O2 concentration for which no growth was found.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Effect of modulation of G6PDH activity using the CUP1 promoter

The connection between the availability of NADPH and xylitol formation during xylose fermentation was investigated by expressing the ZWF1 gene under control of the Cu2+-regulated CUP1 promoter (Labbe and Thiele, 1999). Integration of pB3 CUP1 ZWF1 into TMB3001 resulted in TMB3030. The addition of 0 and 10 µM Cu2+ gave G6PDH activities of 1.040 and 1.567 U/mg protein, respectively. In the control strain TMB3001 the activity was 1.015 U/mg protein, which is about the same as the lowest activity feasible with the copper promoter.

A batch fermentation of xylose was performed under oxygen-limited conditions with the control strain TMB3001 and with TMB3030 at 10 µM Cu2+. The cells did not grow under these conditions, and the xylose was therefore converted only to xylitol, glycerol, acetate, ethanol and carbon dioxide. The higher G6PDH-activity in TMB3030 resulted in a higher xylitol yield (0.33 g/g) than in TMB3001 (0.29 g/g) (Table 2). The xylose consumption rate was the same for the two strains, indicating that the xylose consumption rate can not be enhanced by increasing the G6PDH activity above the wild-type level. Addition of 10 µM Cu2+ to TMB3001 did not affect xylose consumption and product yields (data not shown). Since the CUP1 promoter turned out to be unsuitable for downregulation of ZWF1 expression, a synthetic promoter library was constructed.

Table 2. Effect of different expression levels of ZWF1 in recombinant S. cerevisiae in oxygen-limited batch fermentation with 50 g/l xylose as sole carbon source. Specific xylose consumption rate (g/g biomass/h) and ethanol, xylitol, acetate and glycerol yields (g/g consumed xylose) and carbon recoveries (cmoles out/cmoles in) were determined after 60–70 h. The presented values are the mean of two independent fermentation experiments plus/minus the deviation from the mean. In calculation of carbon recovery, 0.5 moles CO2 are assumed to be formed for each cmole of ethanol or acetate produced. Maximal specific growth rates, µmax (h−1) were measured under anaerobic conditions with 2% lignocellulose hydrolysate and 20 g/l glucose. For minimal inhibitory concentration, MIC, of H2O2 (%) steps of 0.75% H2O2 were investigated and double samples were made. NM, not measured. G6PDH activities were measured in precultures
StrainsG6PDH activityXylose consumption rateY(Ethanol)Y(Xylitol)Y(Acetate)Y(Glycerol)C- recoveryµmax (h−1)MIC (% H2O2)
  • *

    Fermentation data from Jeppsson et al. (2002).

TMB3255*<0.010.022 ± 0.0020.41 ± 0.020.05 ± 0.010.084 ± 0.0050.054 ± 0.0080.950.196.75
TMB32560.0140.062 ± 0.0020.36 ± 0.020.13 ± 0.020.054 ± 0.0030.044 ± 0.0010.960.21NM
TMB30370.0560.112 ± 0.0030.34 ± 0.010.19 ± 0.010.039 ± 0.0010.037 ± 0.0010.990.266.75
TMB3001*1.0150.145 ± 0.0020.31 ± 0.010.29 ± 0.010.025 ± 0.0010.052 ± 0.0040.980.289.75
TMB3030 (0 µM Cu2+)1.040NMNMNMNMNMNMNMNM
TMB3030 (10 µM Cu2+)1.5670.150 ± 0.0010.28 ± 0.010.33 ± 0.010.025 ± 0.0010.051 ± 0.0030.970.21NM

Construction of a promoter library for downregulation of ZWF1

In order to modulate the ZWF1 expression, a library of synthetic promoters was designed. The promoter (Figure 1), called YRP, contains two regulatory structures, RPG boxes (Nieuwint et al., 1989) and CT boxes (Baker, 1986, 1991). These elements promote transcription of ribosomal protein genes and glycolytic genes (Rotenberg and Woolford, 1986; Baker, 1991). The CT boxes and RPG boxes were designed to be degenerated; at each position, 1% each of the three other bases was incorporated during synthesis (Figure 1). Completely degenerated intervening sequences combined with less degenerated transcription factor binding sites results in a wide range of different promoter strengths (Jensen, 1997; Jensen and Hammer, 1998a, 1998b).

thumbnail image

Figure 1. Base sequence of the synthetic promoter YRP. n = 25% of each base, transcription factor binding sites (RPG box and CT box) have 1% each of the bases other than indicated at each position

Download figure to PowerPoint

In Lactococcus lactis, nucleotides at a number of positions are well conserved between different promoters, and upon these sequences the synthetic promoter can be modelled (Jensen and Hammer, 1998a, 1998b). However, S. cerevisiae does not appear to have a strict sequence for promoters; instead the synthetic promoter was pieced together with a combination of structures from several S. cerevisiae promoters. The distance between the RPG box and the CT box was the same as in the glycolytic PYK1 promoter (Drazinic et al., 1996), since it had been shown to be the best for interaction between the transcription factors GCR1p and RAP1p (Drazinic et al., 1996), which bind to the RPG box and CT box, respectively. The distance between the RPG boxes was chosen to be the same as in the ribosomal protein promoter RP39A (Rotenberg and Woolford, 1986). The constant sequence between the TATA box and the 3′ end of the promoters was the same as in the ENO1 promoter (Uemura et al., 1997). The synthetic double-stranded promoters were ligated to the vector pYLZ-2 (Hermann et al., 1992) containing the E. coli lacZ reporter gene, resulting in pYLZ-2 YRP vectors. Thirty-seven different clones of the promoter were used to control the lacZ reporter gene in vectors pYLZ-2 YRP 1–37. The resulting β-galactosidase activities ranged from 0.007 to 37 Miller units (Figure 2), covering about three orders of magnitude between the lowest and the highest activity.

thumbnail image

Figure 2. Specific β-galactosidase activity (Miller units) from YRP1 to YRP37 controlling lacZ in pYLZ-2 in S. cerevisiae. Shaded clones (YRP13, YRP25, YRP34) were used to regulate the expression of ZWF1

Download figure to PowerPoint

Three selected promoter clones YRP13 [1 Miller unit (MU)], YRP25 (4 MU), and YRP34 (22 MU) (shaded in Figure 2), were used to control the expression of the ZWF1 gene in TMB3001, resulting in TMB3037, TMB3035 and TMB3034, respectively. When harvested in exponential phase on YPD medium TMB3037, TMB3035 and TMB3034 had G6PDH-activities of 0.11, 0.77 and 1.82 U/mg protein, respectively, which corresponded with the expected order of the strength of the promoters.

Xylose fermentation with strains with downregulated ZWF1

The synthetic promoter YRP13 produced the lowest G6PDH activity in TMB3037 (0.056 U/mg protein in defined medium), so this strain was selected for batch fermentation with xylose. TMB3037 consumed 0.112 g xylose g biomass/h, which is five times faster than in TMB3255 (zwf1Δ), and represents 77% of the specific consumption observed in the control strain TMB3001 (Table 2). As expected, TMB3037 had a higher ethanol yield (0.34 g/g) than TMB3001 (0.31 g/g), accompanied by a lower xylitol yield (0.19 g/g vs. 0.29 g/g).

In an attempt to obtain even lower expression levels of ZWF1, a strain with a promoterless ZWF1 gene was constructed by integration of pB3 ZWF1 into TMB3001. The resulting strain, TMB3256, had a G6PDH activity of 0.014 U/mg protein, which is about 1% of the wild-type activity (Table 2). Batch fermentation conducted with 50 g/l xylose showed that TMB3256 had a specific xylose consumption of 0.062 g/g biomass/h, which corresponds to a 2.8-fold increase compared with TMB3255 (zwf1Δ). At the same time, TMB3256 showed an ethanol yield of 0.36 g/g and a xylitol yield of 0.13 g/g, which are close to the values of TMB3255 (zwf1Δ) (Table 2).

The connection between G6PDH activity, xylitol yields, ethanol yields and xylose consumption are illustrated in strains with G6PDH activities spanning from 0 U/mg protein to 1.6 U/mg protein (Table 2). The ZWF1-disrupted strain TMB3255 exhibited the highest ethanol yield (0.41 g/g) and lowest xylitol yield (0.05 g/g), whereas the TMB3030 strain at 10 µM Cu2+ had the lowest ethanol yield (0.28 g/g) and the highest xylitol yield (0.33 g/g) of the strains in Table 2. The xylose consumption rate was not further enhanced at G6PDH-activity above wild-type level.

Overproduction of transhydrogenase

If the fermentation pattern seen in low G6PDH-activity strains is a result of low intracellular levels of NADPH, a similar phenotype should be observed at overproduction of the cytoplasmic transhydrogenase (TH) from Azotobacter vinelandii. In S. cerevisiae, this enzyme has been shown to catalyse the conversion of NADPH and NAD+ to NADP+ and NADH, so that the ratio NADPH : NADP+ shifted from 5.1 to 3.0 at TH-overproduction (Nissen et al., 2001).

TH was overproduced at 1.0 U/mg in TMB3254 (Table 3). No TH activity was detected in the control strain. The xylose consumption was similar in the TH-overexpressing strain compared to its control. TH-overproduction resulted in 12% lower xylitol yield and 29% higher glycerol yield. Ethanol and acetate yields were similar with and without TH overproduction.

Table 3. Effect of transhydrogenase overproduction in recombinant S. cerevisiae in oxygen-limited batch fermentation with 50 g/l xylose as sole carbon source. Specific xylose consumption rate (g/g biomass/h) and ethanol, xylitol, acetate and glycerol yields (g/g consumed xylose) and carbon recoveries (cmoles out/cmoles in) were determined after 40 h. The presented values are the mean of two independent fermentation experiments ± the deviation from the mean. In calculation of carbon recovery, 0.5 moles CO2 are assumed to be formed for each cmole ethanol or acetate produced. Specific TH activities were measured in cells harvested in exponential phase after growth on YPD medium
StrainsTH activityXylose consumption rateY(Ethanol)Y(Xylitol)Y(Acetate)Y(Glycerol)C-recovery
TMB3253<0.010.157 ± 0.0040.28 ± 0.0070.34 ± 0.0010.036 ± 0.0010.059 ± 0.0020.98
TMB32541.00.162 ± 0.0020.28 ± 0.0030.30 ± 0.0020.036 ± 0.0010.076 ± 0.0010.97

Maximal specific growth rates in lignocellulose hydrolysate

Strains in which the G6PDH activity is reduced are impaired in their ability to produce NADPH. This may influence the strains' tolerance towards inhibitors present in lignocellulosic hydrolysates, since for example, 5-HMF is reduced with NADPH (Wahlbom and Hahn-Hägerdal, 2002).

In previous work, we showed that the disrupted strain TMB3255 (zwf1Δ) and TMB3001 had similar anaerobic maximal specific growth rates on glucose in defined mineral medium (Verduyn et al., 1992) supplemented with 0.1 mM methionine (Jeppsson et al., 2003). TMB3255 (zwf1Δ) grows 54% slower than TMB3001 when methionine is left out (Jeppsson et al., 2003). Under aerobic conditions, the growth rate of TMB3255 was not fully restored by methionine-addition (Jeppsson et al., 2003). Anaerobic growth rates were measured for TMB3030, TMB3001, TMB3037, TMB3256 and TMB3255 in 2% lignocellulosic hydrolysate and 20 g/l glucose, supplemented with 0.1 mM methionine. Since we supplemented the medium with methionine, and the cultures were run anaerobically, differences in growth rate between strains result solely from the presence of hydrolysate. For TMB3030, 10 µM Cu2+ was also added.

Maximal specific growth rates decreased with decreasing G6PDH-activity (Table 2). TMB3001, TMB3037, TMB3256 and TMB3255 had maximal specific growth rates of 0.28, 0.26, 0.21 and 0.19/h, respectively. TMB3030 had the highest G6PDH activity but it only grew at 0.21/h, which is probably due to inhibition by Cu2+.

Minimal inhibitory concentrations (MIC) of H2O2

Strains disrupted in the oxidative pentose phosphate pathway display enhanced sensitivity to oxidative stress (Nogae and Johnston, 1990), which has been ascribed to reduced levels of NADPH. This was also observed in our hands, where TMB3001 had a MIC of 9.75 vol% H2O2, whereas TMB3037 and TMB3255 both had a MIC of 6.75 vol% H2O2 (Table 2).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In a previous investigation (Jeppsson et al., 2002), we demonstrated the connection between a low flux through the oxidative PPP and a low xylitol yield. However, the specific xylose consumption in the ZWF1-disrupted strain (TMB3255) was six-fold lower than for the control strain TMB3001. In the present investigation we looked into the effect of different expression levels of ZWF1 on xylose fermentation. We showed that xylose consumption in a G6PDH-deficient strain could be enhanced by allowing a low G6PDH activity, although at the expense of a higher xylitol yield. TMB3256 and TMB3037 had 1% and 6% of the wild-type level of G6PDH activity, respectively. The xylose consumption rate was 2.8 times higher for TMB3256 and 5.1 times higher for TMB3037 than in the ZWF1-disrupted strain TMB3255 (zwf1Δ). The xylitol yields increased 2.6 times for TMB3256 and 3.8 times for TMB3037. Thus, there is a trade-off between high xylose consumption rate and low xylitol yield when ZWF1 expression level is modulated.

The synthetic promoter library was useful for downregulation of ZWF1. The β-galactosidase activities covered a range of about three orders of magnitude between the lowest and the highest activity, and since the maximal strength appears to be in the same range as the CUP1 promoter, this library should be useful for both up- and downregulation of many genes in S. cerevisiae. Synthetic promoter libraries have earlier been described for Lactococcus lactis (Jensen and Hammer, 1998a), E. coli (Jensen and Hammer, 1998b) and mammalian cells (Tornoe et al., 2002). The synthetic promoters did not require an inducing agent, in contrast to, for example, the CUP1 promoter which requires supplementation of Cu2+.

The cytoplasmic transhydrogenase from A. vinelandii has been shown to transfer NAD+ and NADPH into NADH and NADP+ (Nissen et al., 2001). The overproduction of transhydrogenase in S. cerevisiae resulted in a lower xylitol yield. Since this was also observed in strains with decreased G6PDH activity, it further strengthens the hypothesis that the decreased xylitol yield is a result of lower NADPH level. We observed enhanced glycerol yield from xylose at TH overproduction. This has also been observed on glucose (Nissen et al., 2001) and was then ascribed to the necessary reoxidation of NADH formed in the TH reaction.

If the transhydrogenase had been working in the opposite direction we might have been able to remove excess NADP+ and NADH formed in the XR and XDH reactions, respectively. A strategy to overcome the redox imbalance between the XR and XDH reactions has recently been initiated by Verho et al. (2002). The Kluyveromyces lactis gene encoding a NADP+-dependent glyceraldehyde 3-phosphate dehydrogenase has been introduced in xylose-fermenting S. cerevisiae to reduce accumulation of NADP+ and NADH, and thereby enhance the ethanol yield.

Another strategy for creating a co-factor balanced breakdown of xylose was to subject xylose reductase to site-specific mutagenesis to shift the co-factor preference towards NADH (Zhang and Lee, 1997; Kostrzynska et al., 1998). However, when the Km of NADPH was enhanced, Km for xylose increased concomitantly.

The growth rates in lignocellulose hydrolysate were significantly lower in the strains with 0% and 1% G6PDH activity, whereas no significant decrease was observed for TMB3037 with 6% G6PDH activity compared to the control strain. Lignocellulose-derived inhibitors fall in three groups: low-molecular-weight aliphatic acids (predominantly acetic acid, formic acid and levulinic acid), furan derivatives (furfural and 5-HMF) and phenolic compounds (Larsson et al., 1999). 5-HMF requires NADPH for reduction (Wahlbom and Hahn-Hägerdal, 2002). This co-factor is also needed for biomass formation, which may explain the lower growth rates found for the low G6PDH activity strains.

Another mechanism of growth inhibition is the interference of aliphatic acids with ATP requirements. Acetic acid has been shown to lower the maximal specific growth rate of S. cerevisiae (Pampulha and Loureiro-Dias, 2000) and at high external acetic acid concentrations the intracellular pH is decreased (Pampulha and Loureiro-Dias, 1989). When the plasma membrane ATPase pumps out protons it uses ATP, which results in less ATP being available for biomass formation. We showed in the present investigation that the acetate yield is enhanced in strains with low G6PDH activity. These strains may therefore encounter difficulties with providing the ATP required for maintaining the intracellular pH.

In this investigation we have illustrated the effect of modulated G6PDH activities on xylose fermentation. The best ethanol yield was found in the ZWF1-disrupted strain TMB3255, but this strain also had the lowest xylose consumption rate and a significantly reduced growth rate in lignocellulose hydrolysate. TMB3037 had only 6% of wild-type G6PDH activity but the xylose consumption rate was 77% of the consumption rate found in the control strain. This strain also had almost the same maximal specific growth rate in a lignocellulose hydrolysate as TMB3001.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Professor Morten C. Kielland-Brandt, Carlsberg Laboratories, and Professor Jens Nielsen, DTU, for providing us with the YEp24-pPGK-CTH vector. Åsa Ekman and Christer Larsson are gratefully acknowledged for technical assistance. This work was financially supported by The Swedish National Energy Administration and the Nordic Energy Research Programme. M.F.G.-G. was supported by Marie Curie fellowship QLK3-CT-1999-51355 from The European Community.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References