Evolutionary engineering of mixed-sugar utilization by a xylose-fermenting Saccharomyces cerevisiae strain


  • Marko Kuyper,

    1. Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
    Search for more papers by this author
  • Maurice J. Toirkens,

    1. Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
    Search for more papers by this author
  • Jasper A. Diderich,

    1. Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
    Search for more papers by this author
  • Aaron A. Winkler,

    1. Bird Engineering B.V., Westfrankelandseweg 1, 3115 HG Schiedam, The Netherlands
    Search for more papers by this author
  • Johannes P. van Dijken,

    1. Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
    2. Bird Engineering B.V., Westfrankelandseweg 1, 3115 HG Schiedam, The Netherlands
    Search for more papers by this author
  • Jack T. Pronk

    Corresponding author
    1. Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
      *Corresponding author. Fax: +31 15 2782355., E-mail address: j.t.pronk@tnw.tudelft.nl
    Search for more papers by this author

*Corresponding author. Fax: +31 15 2782355., E-mail address: j.t.pronk@tnw.tudelft.nl


We have recently reported about a Saccharomyces cerevisiae strain that, in addition to the Piromyces XylA xylose isomerase gene, overexpresses the native genes for the conversion of xylulose to glycolytic intermediates. This engineered strain (RWB 217) exhibited unprecedentedly high specific growth rates and ethanol production rates under anaerobic conditions with xylose as the sole carbon source. However, when RWB 217 was grown on glucose–xylose mixtures, a diauxic growth pattern was observed with a relatively slow consumption of xylose in the second growth phase. After prolonged cultivation in an anaerobic, xylose-limited chemostat, a culture with improved xylose uptake kinetics was obtained. This culture also exhibited improved xylose consumption in glucose–xylose mixtures. A further improvement in mixed-sugar utilization was obtained by prolonged anaerobic cultivation in automated sequencing-batch reactors on glucose–xylose mixtures. A final single-strain isolate (RWB 218) rapidly consumed glucose–xylose mixtures anaerobically, in synthetic medium, with a specific rate of xylose consumption exceeding 0.9 g g−1 h−1. When the kinetics of zero trans-influx of glucose and xylose of RWB 218 were compared to that of the initial strain, a twofold higher capacity (Vmax) as well as an improved Km for xylose was apparent in the selected strain. It is concluded that the kinetics of xylose fermentation are no longer a bottleneck in the industrial production of bioethanol with yeast.


The concept of using, as an alternative fuel for combustion engines, ethanol from renewable agricultural carbohydrate feedstocks is by no means new. A.J. Kluyver already remarked in his inaugural address in 1922: “Their gradual depletion and the resulting higher price of fossil raw materials will consequently entail that the chemical industry will gravitate more and more towards procuring its starting materials immediately from the present-day plant worldFor a considerable time, a search for petrol substitutes has already been instituted, and universally the conviction has been gained that industrial alcohol offers the best prospects[1]. As a result of concerns about the limited reserves of fossil feedstocks, geopolitical issues, and the potential climate change consequences of large-scale carbon dioxide release, this process has met with renewed interest.

The yeast Saccharomyces cerevisiae has always been the organism of choice for the large-scale industrial production of ethanol from carbohydrate feedstocks. However, S. cerevisiae can only ferment a narrow range of sugars (essentially hexoses and disaccharides) to ethanol. Agricultural biomass, and in particular its hemicellulose fraction, contains substantial amounts of the pentose sugars d-xylose and l-arabinose. The conversion of S. cerevisiae into an efficient xylose-fermenting yeast has therefore long been one of the major challenges in yeast metabolic engineering. Several non-Saccharomyces yeasts are capable of fermenting xylose to ethanol. In such yeasts, xylose metabolism is initiated by two redox reactions that convert the aldose sugar xylose into xylulose, its keto isomer. The first of these reactions is catalysed by a (generally NADPH-dependent) xylose reductase (EC that reduces xylose to xylitol. In a second step, xylitol is oxidized to xylulose via a (generally NAD+-dependent) xylitol dehydrogenase (EC Many elegant metabolic-engineering studies have been based on the combined expression of the Pichia stipitis xylose reductase and xylitol dehydrogenase genes in S. cerevisiae[2–5]. In combination with additional genetic modifications (e.g., the overexpression of xylulokinase and pentose-phosphate pathway enzymes) this has resulted in significant rates of xylose fermentation, but these were invariably accompanied by the formation of substantial amounts of xylitol and/or glycerol. The formation of these byproducts is an inevitable consequence of an imperfect match of the cofactor specificities of xylose reductase and xylitol dehydrogenase [6,7].

In principle, the intrinsic redox problem associated with the xylose reductase/xylitol dehydrogenase approach can be avoided by expression of a heterologous gene encoding xylose isomerase (EC The viability of this concept has recently been proven by the functional expression of the Piromyces xylose isomerase gene in S. cerevisiae[8]. In contrast to bacterial and Archaeal xylose isomerase genes that have been tested previously, the fungal gene [9] yields high xylose-isomerase activities when introduced in S. cerevisiae[8].

By itself, expression of a xylose isomerase in S. cerevisiae does not enable fast fermentation of xylose. Prolonged selection of a xylose isomerase-expressing strain in aerobic, oxygen-limited and anaerobic batch cultures was required to obtain a strain that could slowly grow anaerobically on xylose with a specific growth rate of 0.03 h−1 but with a high ethanol yield [7]. In a subsequent study, knowledge-based metabolic engineering has been used to improve the kinetics of xylose fermentation [10]. Expression of the Piromyces XylA gene was combined with the overexpression of the native S. cerevisiae genes encoding xylulokinase (EC, ribose 5-phosphate isomerase (EC, ribulose 5-phosphate epimerase (EC, transketolase (EC and transaldolase (EC Moreover, to prevent the formation of even small amounts of xylitol, the S. cerevisiae GRE3 gene, which encodes an aldose reductase, was deleted. The engineered strain RWB 217 exhibited an anaerobic specific growth rate on xylose of 0.09 h−1 and efficiently produced ethanol from xylose. However, a diauxic growth pattern was observed on glucose–xylose mixtures [10]. The low rate and non-exponential kinetics of xylose consumption in these mixed-substrate batch cultures suggested that the affinity of xylose transport for its substrate might be a key factor in the slow xylose consumption.

When industrially relevant phenotypes of a microorganism can be coupled to the specific growth rate and/or the affinity for a nutrient, a rational natural selection approach may offer an attractive alternative to targeted metabolic engineering. This approach, which has been termed ‘Evolutionary Engineering’, has been applied successfully to a number of industrial microorganisms and processes [11–14]. The goal of the present study was to evaluate the applicability of evolutionary engineering for improvement of mixed-sugar utilization kinetics of a xylose-fermenting S. cerevisiae strain.

2Materials and methods

2.1Strains and maintenance

The S. cerevisiae strain used in this study was RWB 217 (MATa ura3-52 leu2-112 loxP-PTPI::(-266, -1)TAL1 gre3Δ::hphMX pUGPTPI-TKL1 pUGPTPI-RPE1 KanloxP-PTPI::(-40, -1)RKI1 {pAKX002, p415ADHXKS}). It was derived from CEN.PK102-3A (MATA ura3-52 leu2-112) [15] and has been described previously [10]. This strain overexpresses the structural genes for all enzymes involved in the intracellular conversion of xylose to glycolytic intermediates. The overexpressed enzymes are xylose isomerase (EC, xylulokinase (EC, ribose 5-phosphate isomerase (EC, ribulose 5-phosphate epimerase (EC, transketolase (EC and transaldolase (EC In addition, the GRE3 gene encoding aldose reductase (EC was deleted to further minimize xylitol production.

Stock cultures were grown at 30 °C in shake flasks on synthetic medium [16] supplemented with 20 g of glucose l−1. When stationary phase was reached, sterile glycerol was added to 30% (v/v), and 2 ml aliquots were stored in sterile vials at −80 °C.

2.2Cultivation and media

Shake-flask cultivation was performed at 30 °C in a synthetic medium [16]. The pH of the medium was adjusted to 6.0 with 2 M KOH prior to sterilization. Precultures were prepared by inoculating 100 ml medium containing 20 g l−1 glucose or xylose in a 500-ml shake-flask with a frozen stock culture. After 24–48 h incubation at 30 °C in an orbital shaker (200 rpm), this culture was used to inoculate either shake-flask cultures or fermentor cultures.

Anaerobic chemostat cultivation was carried out at 30 °C in 2-l laboratory fermentors (Applikon, Schiedam, The Netherlands) with a working volume of 1 l. The culture pH was kept at pH 5.0 by automatic addition of 2 M KOH. Cultures were stirred at 800 rpm and sparged with 0.5 l min−1 nitrogen (<10 ppm oxygen). Dissolved oxygen was monitored with an autoclavable oxygen electrode (Applisens, Schiedam, The Netherlands). Synthetic medium [16] containing 30 g l−1 xylose as the carbon source was added at a dilution rate of 0.06 h−1. The medium was supplemented with 100 μl l−1 of silicone antifoam (BDH, Poole, UK) as well as with the anaerobic growth factors ergosterol (Sigma, St. Louis, MO, USA) (0.01 g l−1) and Tween 80 (Merck, Darmstadt, Germany) (0.42 g l−1) dissolved in ethanol [17,18], this resulted in 11–13 mM ethanol in the medium. To minimize diffusion of oxygen, fermentors were equipped with Norprene tubing (Cole Palmer Instrument Company, Vernon Hills, USA) and the medium vessel was sparged with nitrogen gas.

Anaerobic sequencing-batch cultivation was carried out in fermentors with an identical setup as for chemostat cultivation on a synthetic medium containing 20 g l−1 glucose and xylose each. New cycles of batch cultivation were initiated every 20 h by timer-controlled pumps that removed 95% of the culture and refilled the fermentor to 1 l with fresh medium. In the second stage of selection the cycle time was increased to 30 h with the addition of extra xylose at 20 h (approximately 150 ml of a 50% (w/v) solution).

For the determination of the stoichiometry and kinetics of sugar fermentation anaerobic batch cultures were carried out in the same fermentor setup as mentioned above, but with a working volume of 1.5 l and sparged with 0.5 l min−1 of high-grade nitrogen (<5 ppm oxygen), on 20 g l−1 xylose, 20 g l−1 glucose and xylose each, or 100 g l−1 glucose and 25 g l−1 xylose as the carbon source, as indicated in the text. All fermentations were carried out in duplicate.

2.3Determination of culture dry weight

Culture samples (10.0 ml) were filtered over preweighed nitrocellulose filters (pore size 0.45 μm; Gelman laboratory, Ann Arbor, USA). After removal of medium the filters were washed with demineralized water and dried in a microwave oven (Bosch, Stuttgart, Germany) for 20 min at 360 W and weighed. Duplicate determinations varied by less than 1%.

2.4Gas analysis

Exhaust gas was cooled in a condensor (2 °C) and dried with a Permapure dryer type MD-110-48P-4 (Permapure, Toms River, USA). O2 and CO2 concentrations were determined with an NGA 2000 analyser (Rosemount Analytical, Orrville, USA). Exhaust gas flow rate, specific oxygen consumption and carbon dioxide production rates were determined as described previously [19,20]. In calculating biomass-specific rates, a correction was made for volume changes caused by withdrawing culture samples.

2.5Metabolite analysis

Glucose, xylose, xylitol, organic acids, glycerol and ethanol were detected by HPLC analysis on a Waters Alliance 2690 HPLC (Waters, Milford, USA) containing a Biorad HPX 87H column (Biorad, Hercules, USA). The column was eluted at 60 °C with 0.5 g l−1 H2SO4 at a flow rate of 0.6 ml min−1. Detection was by means of a Waters 2410 refractive-index detector and a Waters 2487 UV detector. Xylulose was determined enzymically with a reaction mixture consisting of 100 mM Tris–HCl buffer (pH 7.5) with 10 mM MgCl2, 0.30 mM NADH and an adequate amount of sample (1 ml total volume); the assay was started by the addition of 0.2 U sorbitol dehydrogenase (Sigma, St. Louis, USA). The xylulose concentration was calculated using an absorption coefficient of 6.3 mM−1 cm−1 at 340 nm for NADH. The enantiomeric identity of the small amounts of lactate determined by HPLC was analysed with an enzyme kit specific for l-lactate (no. 10139 084 035, R-biopharm, Darmstadt, Germany). The lactate produced appeared to be d-lactate. Acetate was determined with kit no. 10148 261 035 (R-biopharm).

2.6Carbon recoveries and ethanol evaporation

Carbon recoveries were calculated as carbon in products formed divided by the total amount of sugar carbon consumed, and based on a carbon content of biomass of 48%. To correct for ethanol evaporation during the fermentations, the amount of ethanol produced was assumed to be equal to the measured cumulative production of CO2 minus the CO2 production due to biomass synthesis (5.85 mmol CO2 per gram biomass [21]) and the CO2 production associated with acetate formation as described previously [7].

2.7Selection of strain RWB 218

Strain RWB 217 was subjected to selection by anaerobic chemostat cultivation described above, at a dilution rate of 0.06 h−1 with 30 g l−1d-xylose as the carbon source, for 1000 h. This culture was then used to inoculate an anaerobic sequencing-batch reactor (SBR) which operated, as described in Section 2.2, for 25 cycles of 20 h on a synthetic medium with 20 g l−1 glucose and xylose each. Thereafter the cycle time was extended by 10 h. At the beginning of this 10-h extension approximately 70 g of xylose was added to the culture. This 30-h cycle was repeated 20 times. From the last batch, samples were streaked on solid synthetic medium plates with glucose. Six randomly picked colonies were tested on medium with 20 g l−1 glucose and xylose each in anaerobic batch fermentors. Of these six, the best-performing isolate was designated RWB 218.

2.8Sugar transport assays

For both strains cells were harvested from duplicate fermentations by centrifugation at 4 °C (5 min at 4500g), washed twice in ice-cold 0.1 M potassium phosphate buffer (pH 6.5), resuspended in 0.1 M potassium phosphate buffer (pH 6.5) to a concentration of approximately 15 g l−1 dry weight and kept on ice until further use. Zero trans-influx of glucose and xylose was determined according to Walsh et al. [22] at 30 °C in 0.1 M potassium phosphate buffer (pH 6.5).

After 3 min aeration at 30 °C, 50 μl of cells was mixed with 12.5 μl of fivefold concentrated glucose or xylose solutions, labelled with d-[U-14C]glucose or d-[U-14C]xylose (Amersham Biosciences, Little Chalfont, UK), respectively, and incubated for 5 s (accurately timed). Then, uptake was quenched by transfer of 50 μl of the resulting mixture to 10 ml of cold quench buffer (0.1 M potassium phosphate (pH 6.5)/0.5 M d-glucose, −5 °C). Cells were rapidly harvested by vacuum filtration onto a glass fibre filter (GF/C, Whatman, Brentford, UK). The filters were washed twice with cold quench buffer and rapidly transferred to scintillation vials containing 10 ml of Emulsifier Scintillator Plus (Packard Instruments, Downers Grove, USA). Disintegrations per min were determined with an LS6500 scintillation counter (Beckman Coulter, Fullerton, USA). For each batch culture triplicate determinations were performed at each of 10 glucose concentrations (from 250 mM (9 MBq mol−1) to 0.25 mM (1600 MBq mol−1)), eight xylose concentrations (from 500 mM (9.5 MBq mol−1) to 2.5 mM (300 MBq mol−1)). Additionally four glucose concentrations, from 100 mM (21 MBq mol−1) to 1 mM (460 MBq mol−1), were tested in the presence of 50 mM unlabelled xylose to determine the inhibition of glucose uptake in the presence of xylose. The inhibition constant (Ki) is defined as the concentration of xylose at which the rate of glucose transport is halved. Glucose and xylose concentrations were determined as described above.

The parameters of sugar transport were derived according to single-component Michaelis–Menten kinetics using FigSys software (Biosoft, Cambridge, UK).


3.1Prolonged cultivation of strain RWB 217 in anaerobic xylose-limited chemostat cultures

As described previously [10], the engineered strain RWB 217 exhibits a typical diauxic growth pattern when grown anaerobically on a glucose–xylose mixture. During a rapid, exponential first growth phase on glucose, only a small fraction of the xylose was consumed. Most of the xylose was consumed in a second phase. In contrast to the glucose consumption phase, the xylose consumption phase was characterized by non-exponential (almost linear) xylose consumption (Fig. 1). As the patterns of sugar consumption and alcohol production were mirrored by the carbon dioxide production profile, we used on-line analysis of carbon dioxide production for the initial characterization of selected strains.

Figure 1.

Typical graph of anaerobic growth of strain RWB 217 in fermentors on synthetic medium with 20 g l−1 glucose and xylose each as the carbon source; duplicate experiments differed by less than 5%. Glucose (•), xylose (◯), ethanol (▪) and % CO2 measured in the off-gas per litre culture (—). Initial biomass concentration was 0.2 g l−1.

In an attempt to select for spontaneous mutants with an improved affinity for xylose, strain RWB 217 was grown in anaerobic, xylose-limited chemostat cultures at a dilution rate of 0.06 h−1. This is well below the specific growth rate of 0.09 h−1 observed in anaerobic batch cultures grown on 20 g l−1 xylose as the sole carbon source [10]. In the initial steady state following inoculation, the residual xylose concentration in the cultures was 4.5 g l−1. During prolonged cultivation, this residual concentration decreased progressively until, after ca. 85 generations, an apparent plateau was reached at a threefold lower residual xylose concentration (Fig. 2).

Figure 2.

Residual xylose concentration measured in supernatants of an anaerobic chemostat culture of RWB 217 grown on 30 g l−1 xylose as the carbon source at a dilution rate of 0.06 h−1.

A sample of this long-term chemostat culture was used to inoculate an anaerobic batch culture on a glucose–xylose mixture. The carbon dioxide release pattern was clearly different from that of the original RWB 217 strain (Fig. 3, compare lines a and b). The – probably heterogeneous – inoculum obtained from this long-term chemostat culture showed a distinct acceleration of carbon dioxide production during the second phase of the batch culture. As a result, the total length of the anaerobic batch fermentation process had decreased from 45 to 30 h. At this stage, a culture sample was diluted and plated on xylose synthetic medium. Six colonies were restreaked and analysed for their carbon dioxide production kinetics in anaerobic batch cultures grown on glucose–xylose mixtures. The varying carbon dioxide profiles (data not shown) of these single-cell isolates obtained after prolonged xylose-limited cultivation suggested that, after 85 generations, the cultures were still heterogeneous or that the spontaneous mutations were not stable. Therefore, instead of selecting a single-cell line for further research, a sample from the chemostat culture was subjected to further selection experiments in automated sequencing-batch reactors. The chemostat culture was maintained for a further 1000 h. Its residual xylose concentration at that point was only 20% lower than a thousand hours before. When a sample from this culture was tested in an anaerobic mixed-substrate batch culture (data not shown), the results were near identical to the sample that had not undergone the additional 1000 h of selection (Fig. 3, line b).

Figure 3.

CO2 production profiles, per litre culture, as measured in off-gas of anaerobic chemostat batches with 20 g l−1 glucose and xylose each. Profiles have been aligned on the glucose consumption peak to eliminate variations in initial biomass. (a) RWB 217, (b) culture after chemostat selection, (c) RWB 218. Initial biomass concentrations were 0.20 ± 0.05 g l−1.

3.2Selection for improved mixed-sugar utilization in sequencing-batch reactors

As prolonged chemostat selection did not further improve anaerobic mixed-substrate consumption, further selection was performed in automated sequencing-batch reactors (see Section 2). Initially, these cultures were grown, via an automated fill-and-draw strategy, using a medium containing 20 g l−1 glucose and 20 g l−1 xylose (Fig. 4(a)). This resulted in a very slow improvement of the xylose fermentation during mixed-sugar utilization (data not shown). A potential drawback of this experimental design is that, after the glucose has been consumed, a maximum of only one additional biomass doubling is possible on the xylose present in the medium (assuming identical biomass yields on glucose and xylose). As selection should be based on the rapid consumption of xylose, we therefore altered the automated feed regime by prolonging the cycle time with 10 h at the beginning of which extra xylose was added (Fig. 4(b)). This resulted in a progressive improvement of the xylose utilization during mixed-sugar fermentation until, after 20 extended cycles in the sequencing-batch reactor, the initial glucose–xylose mixture was completely consumed within 20 h.

Figure 4.

CO2 production profile of a single cycle of the automated sequencing-batch reactor: (a) 20-h cycle with 20 g l−1 glucose and xylose each, (b) 30-h cycle starting with 20 g l−1 glucose and xylose each; and at 20 h approximately 70 g l−1 xylose is added to the culture.

At this stage, six single-cell lines were isolated from the sequencing-batch reactor and analysed for their carbon dioxide production kinetics on a glucose–xylose mixture of 20 g l−1 each (data not shown). One culture exhibited a CO2 production profile similar to that of the heterogeneous chemostat selection (Fig. 3, line b), the other five showed a significant improvement of xylose consumption. The best of these cell lines (Fig. 3, line c) was designated RWB 218 and used for further characterization.

3.3Fermentation characteristics of a strain selected for efficient utilization of glucose–xylose mixtures

The selected strain, RWB 218, was tested like its parental strain in anaerobic batch fermentors with 20 g l−1 xylose, as well as with 20 g l−1 xylose and glucose each as the carbon source. On xylose alone the selected strain exhibited an anaerobic growth rate of 0.12 h−1, which is 30% higher than that of RWB 217 (Table 1). The maximum specific xylose consumption rate in these batches was correspondingly increased to 1.4 g xylose (g dry weight)−1 h−1. Apart from the increased rate of growth and sugar consumption this strain behaved the same as RWB 217.

Table 1.  Growth parameters, sugar consumption and product formation by the engineered strain RWB 217 and its selected derivative RWB 218 during anaerobic batch cultivation in fermentors at pH 5.0 in synthetic medium
 RWB 217RWB 217RWB 218RWB 218RWB 218
  1. Values are presented as the average and experimental deviation of two independent batch cultivations.

  2. aDetermined from the glucose consumption phase.

  3. bDetermined from the total of consumed sugar.

  4. cCalculated based on the ethanol concentrations deduced from the CO2 production, see Section 2.

  5. dDeduced from the CO2 production, see Section 2.

Carbon source (w/v)Xylose (2%)Glucose (2%) + xylose (2%)Xylose (2%)Glucose (2%) + xylose (2%)Glucose (10%) + xylose (2.5%)
Specific growth rate (h−1)0.09 ± 0.000.25 ± 0.00a0.12 ± 0.000.25 ± 0.01a0.22 ± 0.00a
Biomass yieldb (g g−1)0.085 ± 0.0020.074 ± 0.0010.100 ± 0.0030.084 ± 0.0030.084 ± 0.001
Ethanol yieldc (g g−1)0.43 ± 0.000.43 ± 0.000.41 ± 0.000.40 ± 0.030.38 ± 0.01
Carbon recoveryc (%)105.9 ± 0.9103.2 ± 0.1101.9 ± 0.698.4 ± 1.795.3 ± 0.7
Glucose consumed (mM)108.5 ± 0.2107.3 ± 5.0545.7 ± 0.2
Xylose consumed (mM)133.9 ± 0.1136.0 ± 0.3123.5 ± 0.6132.1 ± 5.8171.2 ± 0.7
Biomass (g l−1)1.70 ± 0.042.97 ± 0.041.85 ± 0.053.29 ± 0.1210.37 ± 0.18
CO2 (mmol l−1)199.9 ± 1.5391.6 ± 0.6175.2 ± 0.9362.7 ± 23.91089 ± 2
Ethanold (mM)188.5 ± 1.3370.7 ± 0.4163.5 ± 8.6342 ± 23.21024 ± 3
Xylitol (mM)0.38 ± 0.040.78 ± 0.000.18 ± 0.030.45 ± 0.050.97 ± 0.01
Xylulose (mM)<0.01<0.01<0.01<0.01<0.01
Glycerol (mM)17.8 ± 0.232.7 ± 0.315.0 ± 0.529.8 ± 1.1111.7 ± 0.2
Acetate (mM)1.40 ± 0.073.54 ± 0.020.78 ± 0.150.74 ± 0.063.68 ± 0.22
Succinate (mM)0.39 ± 0.020.96 ± 0.000.23 ± 0.000.68 ± 0.022.79 ± 0.02
d-Lactate (mM)1.46 ± 0.012.78 ± 0.031.48 ± 0.012.68 ± 0.099.30 ± 0.00

Under mixed-substrate conditions RWB 218 performed nearly as well as on xylose alone. With an initial dry weight of 0.2 g l−1, glucose and xylose (20 g l−1 each) were consumed in 24 h (Fig. 5), with ethanol, carbon dioxide, glycerol and biomass as the major products (Table 1). The xylose consumption rate in these experiments was 0.9 g xylose (g dry weight)−1 h−1. Although this rate is lower than on xylose alone, it is 1.5 times higher than the rate of xylose consumption by RWB 217 under such conditions.

Figure 5.

Typical graph of anaerobic growth of strain RWB 218 in fermentors on synthetic medium with 20 g l−1 glucose and xylose each as the carbon source; duplicate experiments differed by less than 5%. (a) Glucose (•), xylose (◯), ethanol (▪), glycerol (□) and % CO2 measured in off-gas per litre culture (−). (b) Dry weight (•), acetate (◯), xylitol (▪), d-lactate (□), and succinate (▴). Initial biomass concentration was 0.17 g l−1.

The sugar concentrations applied in our laboratory experiments do not reflect the actual concentrations applied in industry. Sugar loads as high as 20% are routinely used. In order to investigate the performance of our strain with higher sugar concentrations, we tested our selected strain in a mixture of 100 g l−1 glucose and 25 g l−1 xylose. Together with the increased sugar concentrations, double-concentrated synthetic medium [16] was used and the flow of nitrogen was increased to 1.0 l min−1 to keep the level of exhaust CO2 within the analyser range. When this fermentor was inoculated with an initial 1.1 g l−1 dry weight, all sugars were consumed within 24 h (Fig. 6) with ethanol, CO2, glycerol and biomass as the main products (Table 1).

Figure 6.

Typical graph of anaerobic growth of strain RWB 218 in fermentors on synthetic medium with 100 g l−1 glucose and 25 g l−1 xylose as the carbon source; duplicate experiments differed by less than 5%. Glucose (•), xylose (◯), ethanol (▪), glycerol (□) and % CO2 measured in off-gas per litre culture (—). Initial biomass concentration was 1.1 g l−1.

3.4Sugar transport kinetics

To investigate whether the improved fermentation characteristics of strain RWB 218 were due to improved sugar transport, zero trans-influx assays were performed on samples of both strains grown in anaerobic batches with mixed substrate. Cells were harvested in the xylose consumption phase, after glucose had been depleted (at 21 h for RWB 217 and 16 h for RWB 218), subsequently the kinetics of glucose and xylose transport were determined in 5-s uptake assays of the 14C-labelled sugars.

The sugar transport kinetics found for RWB 217 (Fig. 7) corresponded well with those reported in other studies. For glucose a high-affinity system was determined with a Km of 1.1 mM and a maximum transport capacity (Vmax) of 3.7 mmol (g dry weight)−1 h−1 (Fig. 7) [23–25]. The values for xylose transport of RWB 217 (Km 132 mM, Vmax 15.8 mmol (g dry weight)−1 h−1) also agree with those of other studies [26,27]. The capacity for xylose uptake for strain RWB 217 and RWB 218 was nearly four times higher than that for glucose uptake (Fig. 7). Although this has not previously been reported for S. cerevisiae, Gárdonyi et al. [28] have measured a similar phenomenon for Candida intermedia PYCC 4715.

Figure 7.

Zero trans-influx kinetics of strain RWB 217 (•) and its selected derivative RWB 218 (◯) for glucose (a) and xylose (b). Cells were harvested from mixed-substrate anaerobic batch cultivations, after depletion of glucose (at 21 h for RWB 217 and 16 h for RWB 218). Data result from single-component Michaelis–Menten fits of the averaged triplicates of two independent cultures, with Km in mM and Vmax in mmol (g dry weight)−1 h−1; the ± value and dotted lines indicate the 95% confidence interval.

The transport kinetics of RWB 218 were significantly different from those of RWB 217. The Km for glucose increased to 5 mM, while the transport capacity (Vmax) for glucose had nearly doubled in the selected strain as compared to the parental strain. In the selected strain the Km for xylose had decreased from 132 to 99 mM where the Vmax had increased to 32 mmol (g dry weight)−1 h−1, two times higher than that of the parental strain.

Since glucose and xylose are thought to be transported by the same family of hexose transporters, in a mixture, glucose and xylose will have to compete for these transporters. Uptake experiments of 14C-labelled glucose in the presence of 50 mM xylose showed that xylose is a competitive inhibitor of glucose transport (data not shown), with an inhibition constant (Ki) of roughly 35 mM for both strains.


4.1Evolutionary engineering of mixed-sugar utilization

Already at the time of invention of chemostat cultivation, it was realized that the constant nutrient-limited conditions represent a strong selective pressure for improved uptake kinetics of the growth-limiting substrate [29]. Indeed, microbial populations in chemostat cultures, which are by definition limited for one or more nutrients, tend to evolve towards a higher affinity for uptake of the growth-limiting substrate. In such cases, an increasing substrate affinity (which may be caused by an increased Vmax and/or a decreased Km for the growth-limiting nutrient) is reflected by a progressive decrease in the residual concentration of the growth-limiting nutrient [29–33]. In practice, this selection pressure is so strong that cultures older than 20 generations are considered to be substantially evolved from the ‘parental’ culture and are therefore unfit to study the kinetic parameters of the original culture [31,34]. The progressive decrease of the residual xylose concentrations observed during long-term cultivation of the xylose-fermenting strain RWB 217 in xylose-limited chemostat cultures was entirely consistent with this pattern. The chemostat-evolved population exhibited a strongly improved performance in mixed-substrate batch cultures (Fig. 3, line b). This is consistent with our hypothesis that the slow xylose fermentation kinetics of the parental strain in mixed-substrate cultures was due to a poor affinity for xylose.

A further improvement of mixed-sugar utilization was obtained by prolonged cultivation in anaerobic sequencing-batch reactors on a glucose–xylose medium. In terms of experimental design, the prolongation of the xylose phase was instrumental in accelerating the selection of relevant mutants. The rapid improvement of fermentation performance in this system demonstrates its power in applied research for improving mixed-substrate utilization.

4.2Kinetics of xylose transport

Although is has been demonstrated that several members of the HXT family of hexose transporter proteins in S. cerevisiae are capable of transporting xylose, their affinity for this substrate is invariably low [26,27,35]. We therefore hypothesized that the slow utilization of xylose by the parental strain RWB 217, especially in mixed-sugar cultures (Fig. 1), was indicative of poor xylose uptake kinetics. Indeed, transport kinetics for glucose as well as xylose differed strongly for the parental strain RWB 217 and the evolved strain RWB 218. For both sugars, a doubling of Vmax was observed, and in the case of xylose a ca. 25% decrease of Km further contributed to improved uptake kinetics.

There are several possible explanations for the observed kinetics in RWB 218. One explanation is an increased expression of a gene encoding a low- or intermediate-Km glucose transporter with activity towards glucose. Alternatively or additionally, the Vmax and Km of hexose transporters for both sugars may have been affected by point mutations in the corresponding genes. We are currently performing a detailed analysis of the molecular basis for improved xylose fermentation.

In a previous study on xylose-metabolizing S. cerevisiae[35], no correlation was found between xylose transport capacities and actual xylose fermentation rates. Apparently, xylose transport only controls the rate of xylose conversion when all reactions downstream of xylose uptake have been sufficiently optimized. This conclusion is corroborated by experiments in which we showed that a xylose isomerase-expressing strain of S. cerevisiae that exhibited low rates of xylose fermentation [7] could be improved threefold by overexpressing the enzymes in the pathway from xylose to glycolysis intermediates [10].

4.3Fermentation performance of the evolved strain RWB 218

In the past three years, the successful expression of a fungal xylose isomerase in S. cerevisiae has enabled a major leap in the performance of xylose-fermenting strains (Fig. 8). Strain RWB 218 represents the best strain available to date, and exhibiting high fermentation rates, even during anaerobic growth at high sugar concentrations on glucose–xylose mixtures (Fig. 6, Table 1). In our opinion, the fermentation performance of strain RWB 218 is such that, in principle, the kinetics of anaerobic xylose fermentation no longer present a true bottleneck in the fermentation of hemicellulose hydrolysates. However, this does not imply that further improvement is either impossible or undesirable.

Figure 8.

Specific ethanol production from xylose by yeast over the last two decades. Values originate from cultures grown on xylose as the sole carbon source, irrespective of varying fermentation conditions. Open symbols represent non-Saccharomyces cerevisiae yeasts, closed symbols represent S. cerevisiae strains engineered for xylose fermentation. 1 Bruinenberg et al. [6], 2 Shi et al. [37], 3 Shi et al. [38], 4 Ho et al. [39], 5 Krishnan et al. [40], 6 Toivari et al. [41], 7 Jeppson et al. [2], 8 Zaldivar et al. [42], 9 Sonderegger et al. [43], 10 Kuyper et al. [7], 11 Sonderegger et al. [44], 12 Kuyper et al. [10], 13 Kuyper et al. (this work).

Even though mixed-sugar utilization by the evolved strain RWB 218 was substantially faster than that of RWB 217, the consumption of glucose and xylose remained sequential, with glucose as the preferred carbon source. Furthermore, the anaerobic specific growth rate on xylose is still only one third of the specific growth rate on glucose. When it is assumed that the anaerobic growth rate on xylose can become as high as that of the parental strain on glucose, i.e., 0.34 h−1[8], with a biomass yield of 0.1 g g−1 and an ethanol yield of 0.45 g g−1, a maximum specific ethanol production rate of 1.5 g (g biomass)−1 h−1 would be possible. The maximum rate observed in our studies is 0.49 g ethanol (g biomass)−1 h−1 (Fig. 8). We expect that a combination of knowledge-based metabolic engineering and evolutionary engineering will enable further improvement of fermentation kinetics.

Surprisingly, the evolved strain RWB 218 exhibited a slightly but significantly higher biomass yield on xylose than RWB 217. This may be related to the higher specific growth rate of this strain, which could lead to a smaller impact of maintenance energy requirements. Furthermore, at pH 5.0, even 1 mM acetate may have a significant uncoupling effect [21]. Therefore, the differences in the (very small amounts) of acetate produced by the two cultures may also have contributed to the small difference in biomass yield (Table 1).

The present study was performed in synthetic media under controlled conditions. Although we have verified that the evolved strain performs well at elevated sugar concentrations (Fig. 6), many other parameters relevant for industrial ethanol production [36] remain to be systematically investigated. Therefore, successful transfer of fermentation performance to industrial strains, media and process conditions should be central issues in future research on xylose fermentation.


We thank Tjerko Kamminga for his contribution to the chemostat selection studies, Stefan de Kok for his efforts in the analysis of batch cultures and Wim de Laat (Royal Nedalco, The Netherlands) for a critical reading of the manuscript.

The research group of J.T. Pronk is part of the Kluyver Centre for Genomics of Industrial Fermentation, which is supported by the Netherlands Genomics Initiative.