Introduction

Lignocellulosic feedstocks from agriculture and forestry provide cheap and sustainable resources for production of transportation biofuels and biochemicals. A significant proportion of the hemicellulose fraction of lignocellulose is pentose sugars and therefore the ability to metabolize them along hexoses is essential for the production hosts to be used in plant material-based processes. The most abundant pentose sugar is d-xylose that is the primary constituent of xylans making up the bulk of hemicellulose in plant cell walls [55].

S. cerevisiae is one of the most prominent organisms for the industrial bioethanol production due to its very efficient hexose fermentation and good inhibitor tolerance [55]. Consequently, the extension of substrate range for fermentation of d-xylose is one of the most active fields in metabolic engineering of S. cerevisiae [20, 27]. d-Xylose-fermenting S. cerevisiae strains have been constructed by over-expression of the genes of Scheffersomyces stipitis (former Pichia stipitis) encoding NAD(P)H-dependent xylose reductase (XR) and NAD+-dependent xylitol dehydrogenase (XDH) [21, 22]. d-Xylose utilization and ethanol production was further improved by over-expression of the endogenous gene encoding xylulokinase [9, 14, 47]. The xylose pathway with the XR and XDH from S. stipitis, however, results in redox cofactor imbalance due to different cofactor requirements of XR and XDH enzymes [8]. In recombinant S. cerevisiae, this has been attributed to one of the major reasons for low ethanol production rates from d-xylose, substantial production of xylitol as a side product, and dependence of oxygen for growth on d-xylose [8, 9, 16]. d-Xylose fermentation by S. cerevisiae has also been enabled by expression of xylose isomerase (XI) encoding gene from bacteria or fungi that directly isomerizes d-xylose to d-xylulose with no involvement of redox cofactors [6, 13]. Expression of the XI pathway in S. cerevisiae has resulted in higher ethanol yields but lower growth rates and productivities on d-xylose compared with the strains with the XR-XDH pathway [6, 19, 24, 25, 34].

One of the first approaches to relieve the redox imbalance of recombinant d-xylose-fermenting S. cerevisiae was the expression of the bacterial transhydrogenase that catalyses the conversion of NADH and NADP+ to NAD+ and NADPH, however, when expressed in S. cerevisiae it operated in the reverse direction converting NADPH to NADH [1]. The other strategies for relieving the redox imbalance include among others expression of a NADP+-dependent d-glyceraldehyde 3-phosphate dehydrogenase [7, 51], modification of the cofactor preference of the ammonium assimilation pathway from NADPH to NADH [40], alteration of the cofactor affinity of XR and XDH [18, 35, 41, 53, 54], or disruption of the oxidative pentose phosphate pathway [17]. Typically, these approaches resulted in a 15–50 % increase in ethanol yields on d-xylose with concomitant decrease in xylitol production but the fermentation rates were still lower than that of d-glucose.

Recently, the endogenous NADH-kinase Pos5p of S. cerevisiae was overexpressed in mitochondria and cytosol of S. cerevisiae with the oxidoreductive d-xylose pathway aiming at reduced NADH levels and increased NADPH availability for the XR reaction and subsequent enhancement in d-xylose fermentation [15]. The mitochondrial NADH-kinase Pos5p phosphorylates NADH to NADPH using ATP as a phosphate donor. Pos5p has kinase activity also towards NAD+, but with an affinity 50-fold lower compared with NADH [5, 31, 44]. Pos5p is the major source of mitochondrial NADPH and important for example iron-sulphur cluster biogenesis and for full respiratory activity [32, 44, 45]. The first 17 amino acids of the Pos5p constitute the signal peptide, which targets the enzyme into mitochondria and removal of these amino acids led to cytosolic expression of the enzyme [46]. When S. cerevisiae expressing the oxidoreductive d-xylose pathway and POS5Δ17 was grown anaerobically on a mixture of d-glucose and d-xylose, increased xylitol and decreased ethanol yields from d-xylose were observed, contrary to the set hypothesis [15]. The result suggested increased NADPH availability for the XR reaction but insufficient NAD+ supply for efficient conversion of xylitol to d-xylulose. Under anaerobic conditions, NADH produced in glycolysis and in the synthesis of biomass can be oxidized to NAD+ only by the reactions catalyzed by alcohol dehydrogenase and glycerol-3-phosphate dehydrogenase producing ethanol and glycerol, respectively [39, 48]. Thus, as increasing the glycerol production would reduce the carbon flux to ethanol, the results suggested that yet another reaction producing NAD+ elsewhere than the central carbon pathways would be required to increase the carbon flux from xylitol onwards (Fig. 1).

Fig. 1
figure 1

Schematic picture of d-glucose and d-xylose metabolic pathways in recombinant S. cerevisiae, also showing the relevant cofactors and overexpressed fumarate reductase (FRD) and NADH–kinase reactions

Full size image

Fumarate reductases catalyze the reduction of fumarate to succinate in the citric acid cycle. Most of the FRDs belong to a multimeric complex associated with the respiratory chain and transfer electrons from quinol to fumarate [49]. However, also soluble monomeric FRDs, which transfer electrons from non-covalently bound FADH2/FMNH2 or NADH to fumarate, have been identified from some species. These include FRDs from several Shewanella species that use quinol as electron donor and are involved in fumarate respiration [11]. S. cerevisiae has two soluble FRDs (cytosolic and mitochondrial) being unique using FADH2/FMNH2 to reduce fumarate, and they are neither linked to the electron transfer chain nor involved in the oxidative phosphorylation [2]. Most interestingly, the African trypanosome Trypanosoma brucei expresses a soluble NADH-dependent FRD located in the peroxisome-like organelle glycosome. This 120-kDa enzyme plays an important role in maintaining the glycosomal NAD+/NADH balance in T. brucei [4]. In addition to fumarate reductase activity, it has domains with homology to AbpE enzymes involved in thiamine biosynthesis and cytochrome b5 reductases/nitrate reductases involved in biosynthesis of unsaturated fatty acids and nitrate metabolism, suggesting that the enzyme may be multifunctional, possessing several different metabolic roles [4].

The aim of the present study was to increase both cytosolic NADPH and NAD+ by combining in cells the cytosolic NADH–kinase and fumarate reductase reactions that specifically produce NADPH and NAD+ cofactors. Fumarate has been shown to accumulate during d-xylose fermentation with the XR-XDH strain [3], suggesting that this metabolite is abundant and present for the reduction. NADH-dependent fumarate reductase of T. brucei was expressed in S. cerevisiae with the oxidoreductive d-xylose pathway either alone or together with the truncated cytosol-directed form of the endogenous NADH-kinase. The strains generated were studied in aerobic and anaerobic batch fermentations using d-glucose and d-xylose as carbon sources.

Materials and methods

Plasmid construction

The plasmids and yeast strains used in this study are listed in Table 1. The FRD from T. brucei (GenBank: AF457132) was codon optimized for S. cerevisiae and synthesized by GENEART (Regensburg, Germany). The synthesized gene was ligated into BglII site between the PGK1 promoter and terminator of vector B1181, which was constructed from YEplac195 [10] as described earlier [42]. The POS5Δ17 lacking the first 17 amino acids (mitochondrial targeting sequence) was amplified by PCR from the chromosomal DNA of S. cerevisiae using the following primers: 5′CGGGATCCCGAAGCTTAAAAATGAGTACGTTGGATTCACATTCC3′ and 5′CGGGATCCCGAAGCTTTTAATCATTATCAGTCTGTCTCTTG3′, and cloned between the TPI1 promoter and ADH1 terminator of plasmid pMLV12 (BluescribeM13 with TPI1-promoter and ADH1-terminator). The pTPI1-POS5Δ17-tADH1 cassette was released by BamHI digestion and cloned into BamHI site of the vector pKK27-1.

Table 1 Strains and plasmids used in the study
Full size table

PKK27-1 was constructed by releasing the KanMX4 cassette from the plasmid pFA6a-KanMX4 [52] by BamHI + EcoRI digestion. The KanMX4 cassette was further ligated into YEplac195 [10] digested with BamHI + EcoRI resulting in pKK27-1.

Strain construction

Yeast strain used in the study was VTT C-10880, i.e., [CEN.PK113-1A (URA3. HIS3. LEU2. TRP1. MAL2-8 c . SUC2), ura3::XYL1 XYL2, xks1::XKS1]. XYL1 and XYL2 genes of S. stipitis were chromosomally integrated into URA3 locus under PGK1 and ADH1 promoters, respectively. The integration cassette was constructed as described earlier [47]. XKS1 of S. cerevisiae was integrated into XKS1 locus as described below.

The XKS1 expression cassette pADH1 m- XKS1- tADH1 was released as a BamHI-fragment from plasmid pMLV14 (BluescribeM13) and ligated into the bacterial plasmid pMLV24. The plasmid pMLV24 contains loxP-S. cerevisiae MEL5 (encoding α-galactosidase)-loxP marker cassette with 60-bp flanking regions for targeting to the XKS1 locus. The XKS1 flanking regions were from nucleotides-250 to -192 and from nucleotides 1,801–1,861, where numbers are relative to the nucleotide A in the XKS1 ATG start codon. The BamHI cloning site was included in the XKS1 5′ flanking sequence. The resulting plasmid, containing XKS1, was named pMLV46. The expression cassette for XKS1, together with the loxP-MEL5-loxP marker, was released from pMLV46 with PstI and SpeI and introduced into yeast cells by transformation. Blue colored, MEL5 (α-galactosidase) expressing yeast colonies were collected from agar-solidified YP containing 2 % w/v d-galactose, supplemented with 5-Bromo-4-chloro-3-indolyl-α-d-galactopyranoside (X-α-Gal, 40 μg/ml). To remove the MEL5 marker cassette from the yeast chromosome, the transformant was retransformed with a plasmid pSH47 [12], expressing the Cre recombinase. Integration of the expression cassette pADH1 m-XKS1 -tADH1 into the XKS1 locus in the genome was verified with PCR. The yeast strain generated was named VTT C-10880.

Control plasmids (B1181 and pKK27-1) and plasmids expressing FRD (pPGK1-FRD) and POS5Δ17 (pTPI1-POS5Δ17) were transformed into VTT C-10880 strain resulting in control (Ctrl), FRD and FRD + POS5Δ17 strains, respectively.

Batch bioreactor cultivations

FRD and FRD + POS5Δ17 and their control strain were cultivated in the 15-vessel parallel bioreactor prototype system Medicel Explorer Cultivation Unit (Medicel, Espoo, Finland) [36]. Cultivation conditions were: temperature 30 °C, 800 rpm stirrer speed with a 26-mm-diameter Rushton-type impeller, cultivation volume of 200 ml, and pH 5, maintained by adding 1 M NaOH. For aerobic conditions, 100 ml/min (0.5 vvm) air was sparged into the reactors. For anaerobic conditions, the head space of the reactors was flushed with 50 ml/min nitrogen (99.999 %). Cultivation medium contained 1.7 g/l yeast nitrogen base without amino acids and ammonium sulphate (Difco), 1 g/l glutamate (Fluka), 0.0135 g/l adenine, 0.348 g/l arginine, 0.266 g/l aspartic acid, 0.058 g/l histidine, 0.036 g/l myo-inositol, 0.525 g/l isoleucine, 0.262 g/l leucine, 0.091 g/l lysine, 0.149 g/l methionine, 0.083 g/l phenylalanine, 0.105 g/l serine, 0.119 g/l threonine, 0.082 g/l tryptophan, 0.03 g/l tyrosine, 0.117 valine, 1.6 mg/l Geneticin sulphate G-418 (Gibco) and either 20 g/l d-glucose (AnalaR NORMAPUR) or 50 g/l d-xylose (Sigma-Aldrich, St. Louis, MO, USA).

Inocula were cultivated in two phases on the above medium with 10 g/l d-glucose and 20 g/l d-xylose in shake flasks at 30 °C with 150 rpm shaking. For the first phase, 25-ml cultivations in 100-ml Erlenmeyer flasks were inoculated from single colonies on plates and incubated overnight. For the second phase, 75-ml cultivations in 250-ml Erlenmeyer flasks were inoculated by adding the total contents of the first phase and incubated for 5 h. An adequate amount of the cell suspension from the second phase was centrifuged and re-suspended in the cultivation medium so that the initial optical density (A600 nm) in the bioreactor cultivations was 1 in aerobic and 3 in anaerobic conditions. Three replicate cultivations were carried out with each of the strains.

Prior to the bioreactor cultivations, FRD and FRD + POS5Δ17 and their control strain were also cultivated in aerobic and anaerobic shake cultures as described in Supplemental Material Figs. 8 and 9.

Extracellular metabolite analysis

Samples were taken with the automated sampling device of the reactor system to empty tubes in a cold bath at −30 °C and let to freeze. Samples of 4 ml for extracellular metabolite and cell-density analyses were taken every 3 h from the glucose cultivations and every 12 h from the xylose cultivations. Samples of 10 ml for enzyme analyses were taken after 24 h cultivation. Concentrations of d-xylose, d-glucose, glycerol, xylitol, ethanol, and acetate were measured with HPLC by using a Waters 2690 Separation Module and Waters System Interphase Module liquid chromatography coupled with a Waters 2414 differential refractometer and a Waters 2487 dual λ absorbance detector (Waters Co., Milford, MA, USA). A fast Acid Analysis Column (100 × 7.8 mm, Bio-Rad, Hercules, CA, USA) and an Aminex HPX-87H Organic Acid Analysis Column (300 × 7.8 mm, Bio-Rad) were equilibrated with 5 mM H2SO4 (Titrisol, Merck, Germany) in water at 55 °C and samples were eluted with 5 mM H2SO4 in water at a 0.6 ml/min flow rate. Data were acquired with Waters Empower software. Cell dry weight was measured gravimetrically by drying the washed cells at 105 °C in Eppendorf tubes. Optical density was measured with spectrophotometer at 600-nm wavelength. In addition, CO2 and O2 levels in the exhaust gas were measured from the exhaust gas at 1.25-h intervals with BlueSens probes (BlueSens, Herten, Germany).

Enzyme activity measurement

Samples for the enzyme activity assays were harvested from the batch cultures from the time point 24 h. Cells were broken with glass beads in 100 mM Tris–HCl, pH 8, supplemented with the Complete Mini, EDTA-free Protease Inhibitor Cocktail (Roche) using a Precellys homogenizator. Protein concentrations were measured by Bradford reagent (Bio-Rad). NAD- and NADH-kinase activities were determined according to the previously described methods [44]. One unit (U) of enzyme activity was defined as 1 μmol of NADH or NADPH produced in 1 min at 30 °C in 1 ml of assay mixture. Specific activity was expressed in units per mg of total soluble proteins. Fumarate reductase activity assay mixture consisted of 100 mM fumarate and 0.4 mM NADH in 100 mM Tris–HCl, pH 8, in a total volume of 220 μl. The reaction was started by addition of 10 μl of the cell lysate prepared as described above and the oxidation of NADH was followed at 340 nm. One unit (U) of enzyme activity was defined as 1 μmol of NADH oxidized in 1 min at 30 °C in an assay mixture. Both assays were performed using a Konelab Arena photometric analyzer (Thermo Electron Oy, Vantaa, Finland).

Results

The NADH-dependent fumarate reductase FRD of T. brucei was overexpressed alone and together with the endogenous cytosol-directed NADH-kinase (POS5Δ17) in d-xylose-utilizing S. cerevisiae to study their effect on the aerobic and anaerobic metabolism of d-glucose and fermentation of d-xylose.

Enzyme activities

The functional overexpression of POS5Δ17 and FRD was confirmed by in vitro enzymatic assays. Since Pos5p functions both as NADH and NAD+ kinase the activity was measured with both of these cofactors. The NADH-FRD and Pos5p activities were measured from the 24-h time point of three replicate aerobic and anaerobic d-glucose and d-xylose cultures. NADH-FRD activity was not detected in the control strain not expressing FRD. In FRD strain fumarate reductase activity was 53 ± 10 mU/mg protein in the d-glucose cultivations and 38 ± 12 mU/mg protein in the d-xylose cultivations. This was substantially more than 7 mU/mg protein measured from T. brucei cell extracts [30] that confirms the functional overexpression of FRD in S. cerevisiae. Fumarate reductase activity was not significantly different between the aerobic and anaerobic conditions (data not shown). NADH-kinase activity of the control strain with the empty B1181 and pKK27-1 plasmids was 0.2 ± 0.1 U/mg protein while its cytosolic overexpression increased the activity to 1.1 ± 0.3 U/mg protein. NAD+-kinase activity was 0.05 ± 0.03 U/mg protein in the control strain and 0.08 ± 0.02 U/mg protein in the strains with POS5Δ17 overexpression. There were no significant differences in the NAD+ and NADH kinase activities between the aerobic and anaerobic d-xylose and d-glucose cultures (data not shown).

Aerobic and anaerobic cultivations on d-glucose

The effect of overexpression of FRD alone or combination of FRD and POS5Δ17 on d-glucose metabolism of VTT C-10880 was studied in aerobic and anaerobic batch fermentations on 20 g/l d-glucose. The product yields on carbon utilized and the specific rates of d-glucose utilization and metabolite production are presented in Fig. 2. The g/l titers over the time course of the fermentations are shown the Fig. 1 and 2 in Supplemental Material. Overall, the differences between the strains on d-glucose were rather small. The expression of FRD and FRD + POS5Δ17 slightly increased the d-glucose consumption rate but decreased the ethanol yield both under aerobic and anaerobic conditions. The CO2 yield of FRD + POS5Δ17 was lower compared with the control under anaerobic conditions.

Fig. 2
figure 2

The yields (g/g) of ethanol, glycerol, carbon dioxide (CO2) and biomass over total d-glucose consumed in a aerobic and b anaerobic batch cultures on 20 g/l d-glucose. The specific consumption and production rates (g g−1 h−1) of d-glucose, ethanol, glycerol, carbon dioxide (CO2) and biomass in c aerobic and d anaerobic batch cultures on 20 g/l d-glucose. The yields and rates were calculated from the logarithmic phase (4–6 h) and (0–6 h) of the aerobic and anaerobic cultivations, respectively. (Ctrl black bars, FRD grey bars, and FRD + POS5Δ17 light grey bars)

Full size image

Aerobic and anaerobic cultivations on d-xylose

The effect of overexpression of FRD alone or FRD + POS5Δ17 on xylose fermentation was studied in aerobic and anaerobic batch fermentations on 50 g/l d-xylose. The product yields on d-xylose utilized and the specific rates of d-xylose utilization and metabolite production are presented in Fig. 3. The g/l titers over the time course of the fermentations are shown the Figs. 2–7 in Supplemental Material. The cultivations were also carried out in shake flasks on 20 g/l d-xylose. Aerobic growth of strains measured as OD600 and fermentation measured as weight loss of glycerol-lock shake flasks due to CO2 escape are shown in Supplemental Material Figs. 8 and 9.

Fig. 3
figure 3

The yields (g/g) of xylitol, ethanol, glycerol, acetate, carbon dioxide (CO2) and biomass over total d-xylose consumed in a aerobic and b anaerobic batch cultures on 50 g/l d-xylose. The specific consumption and production rates (g g−1 h−1) of d-xylose, ethanol, glycerol, acetate, carbon dioxide (CO2), and biomass in c aerobic and d anaerobic batch cultures on 50 g/l d-xylose. The yields and rates were calculated from the d-xylose consumption phase (0–90 h) of the aerobic and anaerobic cultivations. (Ctrl black bars, FRD grey bars, and FRD + POS5Δ17 light grey bars)

Full size image

Under aerobic conditions, FRD expression decreased the d-xylose consumption rate (31 %) and also xylitol, glycerol, and CO2 production rates and yields were decreased while acetate yield was increased. FRD expression decreased the d-xylose consumption rate also under anaerobic conditions. However, ethanol yield increased 60 % and its production rate 40 %. At the same time, xylitol yield and production rate decreased but the specific glycerol production rate increased.

Under aerobic conditions, the expression of FRD + POS5Δ17 had no effect on the d-xylose consumption rate but the specific xylitol production rate was the lowest of the strains studied (73 % lower compared with the control). On the other hand, there was an increase in glycerol production while acetate production was decreased. Under anaerobic conditions, the expression of FRD + POS5Δ17 decreased the d-xylose consumption rate significantly (73 %). Ethanol yield was slightly higher compared with the control (23 %) but its production rate was lower as well as the rates of xylitol, glycerol, acetate, and CO2 production.

Discussion

In the present study, fumarate reductase (FRD) and NADH-kinase (POS5Δ17) reactions were introduced into the cytosol of recombinant S. cerevisiae with the oxidoreductive XR-XDH d-xylose pathway. The aim was to increase the cytosolic NAD+ and NADPH pools in order to alleviate the redox imbalance caused by the oxidoreductive d-xylose utilization pathway, and thus enhance d-xylose fermentation. Alternative solutions to redox imbalance have been previously provided for instance by the expression of the d-xylose isomerase pathway in S. cerevisiae that has resulted in ethanol yields ranging from 0.13 to 0.48 g/g of d-xylose consumed depending on the strain background and fermentation conditions [26]. The other interesting option is alteration of the cofactor affinity of XR or XDH that has resulted in yeast strains able to produce 0.12–0.37 g ethanol/g consumed d-xylose as summarized by Krahulec et al. [23].

Expression of FRD and FRD + POS5Δ17 resulted in 60 and 23 % increase in ethanol yield on d-xylose under anaerobic conditions. The maximum ethanol yield was 0.19 g/g with the strain expressing FRD, which was lower compared with the best ethanol yields reported in the literature [26]. However, as the aim of the present study was to demonstrate the effect of FRD and FRD + POS5Δ17 expression on d-xylose fermentation, we used a very basic laboratory yeast strain with the XR-XDH pathway and small replicate cultures in a 15-vessel parallel bioreactor system that were not expected to result in the highest possible ethanol yields. The results suggest that in a strain with the XR-XDH pathway provision of NAD+ may be at least as important as the provision of NADPH, as the expression of FRD alone was adequate to increase the ethanol yield and to decrease the xylitol yield from d-xylose. Similar results, supporting the present data, were recently obtained by overexpression of the water-forming NADH oxidase from Lactococcus lactis in S. cerevisiae with the XR-XDH pathway [56]. The reason may be that under anaerobic conditions there are only two major ways to oxidize NADH and close the redox balance; alcoholic fermentation and glycerol production [38, 50]. Indeed, FRD expression resulted also in increased yield and production rate of glycerol suggesting that the NADH produced in glycolysis and oxidation of xylitol was not completely oxidized by production of ethanol.

On the other hand, for the regeneration of NADPH more options exist. The cytosolic NADPH can be regenerated via the reactions catalyzed by glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase in the oxidative pentose phosphate pathway, and by isocitrate dehydrogenase or acetaldehyde dehydrogenase [29]. Our previous results from proteome and metabolic flux analyses of d-xylose-fermenting S. cerevisiae suggest that on d-xylose NADPH regeneration mainly takes place via the reaction catalyzed by acetaldehyde dehydrogenase, and via the reactions in the oxidative pentose phosphate pathway that, however, simultaneously result in loss of carbon as CO2 [37, 43].

Compared with the expression of FRD only, the expression of FRD + POS5Δ17 significantly decreased the d-xylose consumption and ethanol, xylitol, glycerol and acetate production rates, suggesting that the phosphorylation of NADH to NADPH by NADH-kinase was energetically too costly under anaerobic conditions on d-xylose. This was observed also in the previous study where expression of POS5Δ17 in xylose-utilizing S. cerevisiae resulted in decreased rates of d-glucose and d-xylose consumption and growth, and biomass yield compared with the control strain when cells were grown on d-glucose or on a mixture of d-glucose and d-xylose under both aerobic and anaerobic conditions [15]. However, in the present study FRD + POS5Δ17 rather increased the d-glucose consumption rate and had no effect on biomass yield or production rate in both the aerobic and anaerobic d-glucose cultures. These results are more in line with earlier results from overexpression of the endogenous NADH-kinase in A. nidulans that resulted in increased specific growth rate and biomass yield, suggesting that the overexpression increased NADPH availability for biomass formation [33]. The present results additionally indicate that the extra NAD+ provided by FRD may have a positive effect on the d-glucose consumption rate counteracting the negative effect of NADH-kinase expression reported before. The FRD alone slightly increased the aerobic and anaerobic d-glucose consumption rates and aerobic biomass production. The yield and fermentation rate of ethanol were decreased under both aerobic and anaerobic conditions proposing that the FRD reaction served as an alternative redox sink replacing the need for regeneration of NAD+ by fermentation.

FRD reduced the d-xylose uptake rate under aerobic conditions on xylose, resulting in decreased xylitol, glycerol and CO2 production rates while the acetate production rate was increased. This suggests that FRD expression resulted in surplus of oxidized cofactors which were reduced to NADH e.g. by acetaldehyde dehydrogenase under these conditions when also the oxidative phosphorylation was active. On the other hand, FRD + POS5Δ17 restored the d-xylose uptake rate to the level of the control strain and the yield and production rate of xylitol were further decreased as compared to the FRD strain. This in turn indicates that under aerobic conditions the ATP supply was adequate and the redox cycle created by FRD and NADH-kinase worked most optimally, providing NADPH and NAD+ for xylose reductase and xylitol dehydrogenase reactions. However, the carbon was directed rather to glycerol than to acetate, as was the case when FRD was expressed alone. Apparently there was no need to oxidize a surplus of reduced cofactors, which could have been formed by the xylitol dehydrogenase reaction if the NADH-kinase and FRD reactions were not completely in balance. Moreover, hardly any ethanol was observed in aerobic xylose cultures, indicating that the ethanol produced was consumed as a co-substrate of d-xylose [28, 43] that may have further increased the NADH pool to be oxidized via the glycerol synthesis.

Overall, the results show in particular, the beneficial effect of provision of NAD+ on d-xylose fermentation by the oxidoreductive xylose pathway. Expression of the NADH-using FRD provides a simple and still a novel way to enhance d-xylose fermentation of S. cerevisiae strains with XR-XDH pathway without extensive metabolic engineering. Moreover, its expression takes advantage of the higher concentration of fumarate observed in d-xylose fermenting S. cerevisiae with XR-XDH pathway [3]. Simultaneously increasing the anaerobic pool of NADPH by overexpression of NADH-kinase turned out to be more challenging, but appeared beneficial under aerobic conditions, as seen by decreased xylitol production. Combination of FRD with an ATP independent production of NADPH, such as the expression of NADP+-dependent glyceraldehyde 3-phosphate dehydrogenase [51] would be an interesting option to test.