Conversion of furfural in aerobic and anaerobic batch fermentation of glucose by Saccharomyces cerevisiae

JOURNALOF BIOSCIENCE AND BIOENGINEERING Vol. 87, No. 2, 169-174. 1999

Conversion of Furfural in Aerobic and Anaerobic Batch Fermentation of Glucose by Saccharomyces cerevisiae MOHAMMAD J. TAHERZADEH,’ LENA GUSTAFSSON,2 CLAES NIKLASSON,’ AND GUNNAR LIDl?N’* Department of Chemical Reaction Engineering, Chalmers University of Technology, S-412 96,’ and Department of General and Marine Microbiology, University of Giiteborg, Medicinaregatan SC, S-413 9O,2Gliteborg, Sweden Received 15 April 199WAccepted 22 November 1998

The effect of furfural on aerobic and anaerobic batch cultures of SaccharomycescerevisiaeCBS 8066 growing on glucose was investigated. Furfural was found to decrease both the specific growth rate and ethanol production rate after pulse additions in both anaerobic and aerobic batch cultures. The specific growth rate remained low until the furfural had been completely consumed, and then increased somewhat, but not to the initial value. The CO2 evolution rate decreased to about 35% of the value before the addition of 4 g-1-l furfural, in both aerobic and anaerobic fermentations. The decreaseof the CO2 evolution rate was rapid at first, and then a more gradual decrease was observed. The furfural was converted mainly to furfuryl alcohol, with a specific conversion rate of 0.6 ( f 0.03) g (furfural) -g-l (biomass)- h-l by exponentially growing cells. However, the conversion rate of furfural by cellsin the stationary phase was much lower. A previously unidentified compound was detected during the conversion of furfural. This compound was characterized by mass spectrometry and it is suggested that it is formed from furfural and pyruvate. [Key words: furfural,

ethanol, yeast, biotransformation]

Fuel ethanol can be produced from lignocellulose by chemical hydrolysis of lignocellulosic materials followed by fermentation. One of the major problems with hydrolyzates produced by acid hydrolysis, is the poor fermentability caused by the presence of inhibitors in the hydrolyzates. Furfural is known to be one of the most important of these inhibitors (1). It is a breakdown product from pentoses and is formed in a browning reaction during hydrolysis in the presence of strong acids (2). It therefore may be impossible to completely avoid furfural formation in a chemical hydrolysis process designed to give a high sugar yield. Since furfural is a primary breakdown product from pentoses and therefore likely to be present in hydrolyzates, the effect of furfural on fermentation by yeast has been the subject of several investigations. For concentrations on the order of 1 g. l-1, clearly negative effects on yeast viability (3), specific growth rate (4) and volumetric fermentation rate (5-8) have been shown. Several enzymes have been shown to be sensitive to furfural, and among the most sensitive are the glycolytic enzymes glyceraldehyde phosphate dehydrogenase (EC 1.2.1.12) and alcohol dehydrogenase (EC l.l.l.l), but also hexokinase (EC 2.7.1.1) is inhibited (6). The use of high-level yeast inocula has been suggested as a means of overcoming toxicity in batch cultures which ferment acid hydrolyzates (3), based on the fact that furfural can be taken up and converted by yeast cells. Conversion of furfural is, in fact, one of the first biotransformations known to occur in yeasts (9). The prime product from furfural is furfuryl alcohol (lo), accounting for more than 70% of the converted furfural. Villa et al. (11) also report formation of furoic acid, and other metabolic products reported are furoin and furil (12). With respect to the specific ethanol productivity, Azhar et al. (1) reported that in the presence of 3 g. 1-l of furfural the specific ethanol formation rate, qe, was only one third of the value without * Corresponding

furfural in batch cultures. In contrast, Fireoved and Mutharasan (13) did not find a significant change in the biomass yield on ATP due to furfural in a chemostat culture operating at low dilution rates. The specific ethanol formation rate, thus, did not change in chemostat culture, although some outliers were reported for which an increased value of qe was found. The authors suggested adaptation effects to be the explanation for these observations. In a previous work (14), we found a clear relation between the fermentation rate of dilute acid hydrolyzates from several kinds of wood, and the sum of the concentration of furfural and 5-hydroxymethyl-furfural (HMF). When the sum of furfural and HMF concentrations exceeded 20 mmol. 1-l (about 2 g . I-‘), the fermentation rate decreased strongly. Furthermore, in the hydrolyzates that fermented well, the concentration of furfural and HMF decreased during the fermentation, in agreement with what has previously been reported for acid hydrolyzates from oak (3). Presumably, the conversion of furfural and HMF is of great importance for enabling fermentation of the hydrolyzates. In the present work, the uptake of furfural and the effects on the fermentation rate of Saccharomyces cerevisiae were studied by making pulse additions of furfural into a synthetic medium. In previously reported studies, furfural has normally been present in the medium from the beginning of the batch cultivations. However, by using pulse addition the transient metabolic response can be followed. Furthermore, in comparison to chemostat cultures, the effects of a higher concentration of furfural in the broth can be studied. On-line measurements were made of the exhaust gas-composition as well as the biomass concentration to provide response times with high resolution. The uptake rate of furfural and the product distribution were determined for both growing cultures and stationary phase cells, under both aerobic and anaerobic conditions.

author. 169

170

TAHERZADEH

ET AL.

MATERIALS

J. BIOSCI.

AND METHODS

Yeast strain and medium Saccharomyces cerevkiae CBS 8066, obtained from Centraalbureau voor Schimmelcultures, Delft, the Netherlands, was used in all experiments. The strain was maintained on agar plates made from yeast extract [lo g.l-‘1, soy peptone [20 g. i-l], and agar [20 g.1.-I] with D-glucose [20 g.l-‘] as an additional carbon source. Inoculum cultures were grown in 300-ml cotton-plugged-conical flasks on a shaker with a shaking diameter of 12mm at 30°C for 24 h. The liquid volume was 100 ml and the shaker speed 170rpm. The growth medium was a defined medium as previously reported (15). Cultivation conditions Anaerobic and aerobic batch cultivations were carried out in a BioFlo III bioreactor (New Brunswick Scientific, Edison, NJ, USA) with a working volume of 2.5 1 at a temperature of 30°C and a stirring rate of 500rpm. The pH value in the medium was controlled at 5.00 (tO.01) by addition of 2 M NaOH. Nitrogen or air was continuously sparged through the reactor at a flow rate of 0.80 1 min l (at NTP) controlled by a mass flow controller (Hi-Tech, Ruurlo, the Netherlands). The nitrogen gas had a guaranteed oxygen content of less than 5 ppm (ADR class 2, l(a), AGA, Sweden). An oxygen probe was used to check the dissolved oxygen saturation in the aerobic experiments in order to avoid oxygen limitation. The experiments were carried out both aerobically and anaerobically. In the aerobic experiments 5 (or 10) g of furfural was introduced into the medium, during exponential growth in the respiro-fermentative phase. The experiments continued up to a few hours after complete uptake of ethanol in the respiratory phase. In the anaerobic experiments 5 (or 10) g of furfural was introduced into the medium, while the cells were anaerobically growing on glucose in the exponential phase. The conditions were switched to aerobic after the complete uptake of glucose (seen from measurements of the CO2 evolution rate) and were continued until the end of the ethanol uptake. In an additional experiment, 10 g of furfural was introduced into an anaerobic culture which had entered the stationary phase, i.e. after the complete uptake of glucose. This experiment ran anaerobically for 18 h after the addition, and then the conditions were changed to aerobic for 3 h. Analytical methods Gas analysis The carbon dioxide in the outlet gas was continuously measured with an acoustic gas monitor (Bruel&Kjaer 1308) (16), and gas measurement signals were averaged for 30 s. The instrument was calibrated with a gas of the following composition: 5% C02, 20% O2 with nitrogen as inert gas. Biomass concentration Cell dry weight was determined from duplicate lo-ml samples, which were centrifuged, washed with distilled water and dried for 24 h at 103°C. At least three samples were taken for each batch cultivation. An on-line flow injection analysis (FIA) system designed according to Benthin et al. (17) was used for determination of the optical density every 30min. Addition of furfural to the medium was found to have no significant effect on the accuracy of the FIA measurements. The optical density measured by FIA was calibrated against the dry weight samples. The concentrations of protein and total carbohydrates of the biomass were measured before and after furfural conversion

BIOENG.,

in all experiments. The determination of the cellular proteins were carried out by the Biuret method with modifications according to Verduyn et al. (18), and the total cellular carbohydrate content was determined according to Herbert et al. (19). Metabolite analyses Samples for HPLC analysis were withdrawn from the broth via 0.45-pm sterile filters. Glycerol and furfuryl alcohol were analyzed on an Aminex HPX-87P column (Bio-Rad, USA) as previously described (14). The amount of glucose, ethanol, acetic acid, fumaric acid, furoic acid, furfural, pyruvic acid and succinic acid was determined by an Aminex HPX-87H column (Bio-Rad) as described in (14). Mass spectrometry In order to characterize an unknown compound found by HPLC, mass spectrometry was used. Samples were obtained from preparative HPLC, and subsequently analyzed using a high resolution mass spectrometer (Zab-Spec, VG Analytical, Fisons instrument, England) operating in CI-mode, with NH3 as the ionizing gas, and operating in EI-mode. Calculations Evolved carbon dioxide was calculated from measurements of the gas composition and the gas flow rate, and other metabolite yields were obtained from HPLC data. A biomass composition of CH1.7600.56N0.17 was used in the carbon balance calculations (18). The metabolite and biomass yields were calculated from the concentrations determined at the end of the exponential growth phase. The average carbon balance determined at the end of the exponential growth phase was 96% (SD 3%). Specific growth rates and specific productivities were determined based on the biomass concentrations obtained from the FIA measurements. RESULTS Anaerobic pulse addition experiments Anaerobic conversion of furfural was studied by the injection of furfural (2 or 4 g . I-*) to exponentially growing batch cultures. The results of such an experiment are summarized in Fig. 1. When compared to the control experiment (Fig. la), the most obvious effect seen immediately after pulse addition was a decrease of the COz evolution rate, CER (Fig. lb). The decrease of CER was rapid the first few minutes after the addition, e.g. 34% (24%) within 5 min, followed by a more gradual decrease. The CER is coupled to the ethanol production rate, and measured ethanol concentrations clearly showed that the ethanol production rate also fell after addition of furfural (Fig. lb). The specific ethanol production rate fell from 1.6 (t-O.l)g.g-‘*h-l to about 0.5 (+0.2)g.g-‘.h-’ after addition of 4 g . I-* of furfural. The specific growth rate, ,u, decreased even further, from 0.4 to 0.03 (t-0.02) h.-l, and remained at this low value until the furfural had been completely consumed (cf Fig. 3a). After exhaustion of furfural in the medium p increased to somewhere between 0.13 (kO.03) h-l, i.e. to less than half of the initial value (Fig. lc). Furfural addition caused an increased pyruvate formation rate, and for some time an increased uptake of acetate from the medium (Fig. lc). However, as the added furfural was consumed, the rate of formation of pyruvate decreased, and acetate accumulation resumed. Also, compounds of the TCA-cycle were affected (Fig. Id). The formation rate of succinate decreased and the concentration of fumarate actually decreased (not shown) after addition of furfural to the medium. Fur-

VOL.

CONVERSION

87, 1999

80/(a>

OF FURFURAL

BY S. CEREVZSIAE

171

5s 47 5

3-

z B 3 0

2-

> Furfural

l-

J 0

*’ (b) --

10

20 Glucose

30 (g 1-l)

40

50

FIG. 2. Glycerol concentration vs. glucose concentration in a pulse addition experiment. Furfural (4 g.l-I) was introduced into an anaerobic batch cultivation of S. cerevisiae growing on glucose (50 g . I- *) as a carbon and energy source.

15.

9 -d 10. 5 2 5-

Q 0.4 -7 0.3

; 2

0.2

0.1

.g m -a 5 .$

protein content was 55(+-6)% and the total carbohydrate content was 25( f 6)%, with no systematic change regardless of the concentration of furfural present. The specific uptake rate of furfural, qf, was 0.6 g .g-‘. h-l and was constant during the uptake period (not shown). No appreciable lag was seen between effects on CER and uptake of furfural. Furfuryl alcohol was the main product of furfural degradation (Fig. 3a), but also furoic acid (< 1% of added furfural) was identified as a minor product. However, from a plot of the sum of furfural, furfuryl alcohol and furoic acid it can be deduced that at least one additional product is intermittently

51(a>

% 0

^ 4 s

'1125

‘100 ^ '75

;

‘50

:s

P

-25

10

20

c .s s" 5 2 u"

3

2 1 --0

30

Time (h) FIG. 1. Anaerobic batch cultivation of S. cerevisiae growing on glucose (50 g. I-r) as a carbon and energy source. (a) CER for a control batch culture in which no furfural was added. Figures (b-d) results from a pulse experiment, in which furfural(4 g. I-i) was introduced at the time indicated by the arrow. (b) CER (line) and ethanol concentration ( n ), (c) concentrations of biomass (A, note: logarithmic scale), pyruvic acid (0) and acetic acid (O), (d) concentrations of glycerol ( n ) and succinic acid ( q ). The arrows show the time of furfural addition.

fural also leads to a decreased glycerol formation rate (Fig. Id), although the formation of glycerol per consumed glucose was increased (Fig. 2). Protein and total carbohydrate levels were measured, but no major changes in biomass composition could be found. The

Time (h) FIG. 3. (a) Conversion of furfural (0) to furfuryl alcohol (0) in anaerobic batch cultivation of S. cerevisiae growing on glucose (50 g.l-I) as a carbon and energy source. (b) The sum of the concentrations of furfural, furfuryl alcohol and furoic acid ( w , left-hand scale) and the peak area of the unknown compound ( 0, right-hand scale) during a pulse addition experiment.

172

TAHERZADEH

ET AL.

Respirative phase

/ /

J. BIOSCI. BIOENO., TABLE 1. High-resolution mass spectrometry data from analysis of the unknown compound using chemical ionization (CI) with NH4+ as the ionizing gas and using electron impact ionization (EI)

Stationary phase

100.105 124.075 125.058 129.066 138.052 139.037 140.072 141.056 142.085 143.091 156.065 157.070 158.082 159.086 195.175 212.200

Relative Suggested intensitya fragment Chemical ionization (CI) 1.22 1.12 2.26 bH9021 A 1.86 2.96 1.43 WW31’ 2.20 1.29 IGWM’ 14.98 lGH&Nl+ 1.52 100.00 K+MWl+ 8.83 22.07 PXhO~N’ 1.85 2.26 [C&z10zNsl4.61 lG~z&zN~l+

27.183 28.183 29.151 34.092 36.042 39.059 41.058 43.025 43.993 44.998 81.038 95.016 97.030

Electron impact ionization (EI) 5.81 6.18 15.16 6.35 5.68 14.96 10.17 28.80 100.00 co2 9.43 6.88 C&O 26.68 CSH302 12.92 CSH502

m/z

&-----%

i $0

Time (h)

80

FIG. 4. Conversion of furfuryl alcohol ( w ) to furoic acid ( 0) by S. cerevbiue during respirative growth on ethanol, and in a stationary phase culture.

formed (Fig. 3b), accounting for at most 20% of the converted furfural. An unidentified peak could be found in the HPLC chromatograms, which matched the time profile of the missing compound well. No anaerobic conversion of furfuryl alcohol after the complete conversion of furfural was obtained with the investigated strain. The effects of aeration To be able to make a comparison between aerobic and anaerobic conditions, 2 or 4 g-1-l furfural was introduced into an aerobically growing culture in the exponential growth phase. Respirofermentative growth normally gives lower glycerol and higher biomass yields than anaerobic growth, which was also observed in the presence of furfural. The specific conversion rate of furfural was, within experimental errors, identical to the anaerobic case (0.6kO.03 g. g-l. h-l), and the presence of oxygen did not affect the product yields from furfural. The main products under aerobic conditions were also furfuryl alcohol and the previously mentioned unknown compound. The specific growth rate after furfural addition was, furthermore, the same for both aerobic and anaerobic experiments. In aerobic batch cultivations of S. cerevisiue the respiro-fermentative phase is, however, followed by respirative growth on ethanol (20). Furfuryl alcohol was converted to furoic acid during the respirative phase (Fig. 4). The conversion had already started during respiro-fermentative growth after complete conversion of furfural, but at low rate. It also continued after the complete consumption of ethanol, although, at a low rate. Shortly after the start of conversion of furfuryl alcohol to furoic acid, furfural was detected in the medium at a concentration of up to 0.06g.I-r, but later became once again undetectable. This suggests that the conversion of furfuryl alcohol to furoic acid occurs via furfural. Stationary phase cells Since furfural reduction to furfuryl alcohol requires reducing power, it is probable that the glycolytic rate influences the conversion rate. For this reason, furfural was also introduced into an anaerobic batch culture in stationary phase, immediately after the complete uptake of glucose. The experiment continued anaerobically for 18 h after the addition of furfural. The rate of furfural degradation was clearly lower (0.07 g.g-l. h-l), than that found for exponentially growing cells. The main product of furfural degradation was furfuryl alcohol. The yield of the previously mentioned unidentified compound was less than 1% of converted furfural, i.e. much smaller than in the case of exponentially growing cells.

m/z (calculated)

125.060 139.039 141.055 142.086 156.066 158.082 195.169 212.196

43.990 81.034 95.013 97.029

a Only peaks with relative intensities higher than 1% and with mass to charge ratio (m/z) higher than 100 are shown for the CI spectrum, and only peaks with relative intensities higher than 5% are shown for the EI spectrum.

Characterization by mass-spectrometry of the unknown compound The previously mentioned unknown compound (Fig. 3b) was detected by HPLC on both RI and UV detectors. It had a comparatively long residence time in the hydrogen ion exchange column, and the ratio between the UV response (at 210nm) and RI response was very high. Palmqvist et al. (21) reported formation of an unknown compound from furfural in S. cerevisiae, and previously Shvets et al. (22) also reported conversion of furfural to an unidentified compound characterized by high absorption of UV light at wavelengths less than 240nm. In order to identify the compound in the present work, the eluted peak was collected from the HPLC system and subsequently analyzed by mass spectrometry. Unfortunately, the compound was not found in available spectral libraries. However, based on identified mass fragments and known metabolic effects, a tentative candidate can be suggested. From a careful consideration of the fragments of EI and CI (Tables 1 and 2) in combination with physiological considerations, the following molecule is suggested to be the unknown compound. f:

P”

-C-y--H, HC-CH

COOH

VOL. 87, 1999

CONVERSION

This molecule has a molecular weight of 184. It can easily loose the carboxylic group as CO2 during ionization, both with CI and EI (cf. Table 1). The remaining part has a molecular weight of 140, and will give the fragments m/z= 158.082 with NH4+ and m/z= 141.055 with H+. These peaks are both found in the CI spectrum with high accuracy. A further loss of two hydrogens gives the fragments m/z=156.066 with NH4+ and 139.039 with H+ in the CI spectrum. Moreover, C4H30CO-CH3 could be the fragment giving m/z=142.086 with NH4+ and 125.060 with H+. The furfuryl group (C4H30-CO-) is, after COZ, the strongest peak in the EI spectrum. Some other hydrogenated or deoxygenated furfuryl fragments can also be identified in the EI spectrum. Another peak, m/z= 110.041, with a low intensity is also found (not shown in Table 1). It corresponds to the fragment C6H602 with the calculated m/z of 110.037. This fragment is interesting since, if combined with five or six ammonia molecules, it will give rise to the two peaks with the largest m/z in the CI spectrum (Table 1). DISCUSSION

It is clear that the glycolysis of S. cerevisiae is affected by addition of furfural, and that the effect occurs rapidly. Furthermore, as is evident from the much lower qf in stationary culture, the reduction of furfural is dependent on an active glycolysis. According to Kang and Okada (23), the aldehyde group of either acetaldehyde or furfural may be reduced by alcohol dehydrogenase (ADH) according to the reactions below. CH,-CHO + NADH ADH

+H+ c---f CH3-CH20H+NAD+ C4H30-CHO + NADH

(1)

+ H+ &y C4H30-CH20H + NAD+ (2) It was easily confirmed with an ordinary kit for enzymatic determination of ethanol (cat. nr. 176 290, Boehringer Mannheim, Germany) that furfuryl alcohol was rapidly converted by alcohol dehydrogenase. It is reasonable to believe that reduction of furfural competes with the reduction of acetaldehyde and therefore “occupies” part of the glycolytic capacity of the cells. This should lead to a lower flux of glucose to ethanol, but does not fully explain the decreased flux. The specific formation rate of ethanol prior to the addition of furfural was about 1.6g.g-‘.h-’ corresponding to 35mmol.g-1-h-1. The sum of the specific formation rate of ethanol (0.5 g.g-1. h--l) and the specific formation rate of furfuryl alcohol (0.5 g.g-‘. h-l) after the furfural addition was only about 15 mmol.gpl.h-‘. Thus, when compared to furfural-free medium, the total turnover of ADH for reduction reactions has been decreased to less than half the original value. This may be caused by direct inhibition effects of other enzymes, e.g. hexokinase (6), but may also be the result of cellular regulation in response to a decreased specific growth rate. The increased formation of pyruvate after addition of furfural (Fig. lc) may be caused by a reduced capacity for reduction of acetaldehyde “downstream” of pyruvate, but it may also be caused by a restricted capacity of pyruvate decarboxylase. The observed net consumption of acetate (Fig. lc) may occur by conversion to

OF FURFURAL

BY S. CEREVISZAE

173

acetyl-CoA (24), and may be partly explained by a decreased acetaldehyde dehydrogenase activity. Apparently, some oxidation of furfural to furoic acid takes place, although the yield of furoic acid from furfural is very low. Presumably its formation is catalyzed by aldehyde dehydrogenase, but this reaction is less favored than the reduction of furfural. Biomass synthesis is the main source of surplus NADH resulting in glycerol formation during anaerobic conditions (18, 25). One could therefore expect that the glycerol yield should be lowered in the presence of furfural, since the biomass growth yield is lowered. That was, however, not observed. On the contrary, the glycerol yield increased after introduction of furfural, which indicates another source of NADH production. The production of the new compound, discussed in the results section, could be this source. The compound may be produced from pyruvate and furfural, probably involving the action of thiamine pyrophosphate. Since formation of pyruvate produces 1 NADH, the formation of the new compound will also yield 1 NADH per formed molecule. Conclusion Furfural is a strong inhibitor of the fermentation of glucose by S. cerevisiae. However, furfural is reduced to the less inhibiting compound furfuryl alcohol by the yeast. The results therefore suggest that in situ detoxification may possibly allow fermentation of strongly inhibiting hydrolyzates. However, a complication to be kept in mind is that the degradation rate of furfural is correlated with the rate of glycolysis. NOMENCLATURE

ADH CER CI EI FIA TCA qe 4f

P

: alcohol dehydrogenase : carbon dioxide evolution rate, mmol.Z-l. h-l : chemical ionization : electron impact ionization : flow injection analysis : tricaboxylic acid cycle : specific ethanol productivity, g . g-l. h- ’ : specific furfural uptake rate, g-g-l. h-l : specific growth rate, h-’ ACKNOWLEDGMENTS

This work was financially supported by the Swedish National Board for Technical Development. The authors are grateful to Dr. Gunnar Stenhagen for help with the MS analyses and to Dr. Nils-Olof Nilvebrant for comments on the manuscript. REFERENCES 1. Azhar, A.F., Bery, M. K., Colcord, A. R., Roberts, R. S., aad Corbitt, G. V.: Factors affecting alcohol fermentation of wood acid hydrolyzate. Biotechnol. Bioeng. Symp., 11, 293-300 (1981). 2. Sjiistriim, E.: Wood chemistry. Fundamental and applications, 2nd ed. Academic Press, San Diego (1993). 3. Cbung, I. S. and Lee, Y. Y.: Ethanol fermentation of crude acid hydrolyzate of cellulose using high-level yeast inocula. Biotechnol. Bioeng., 27, 308-315 (1985). 4. Boyer, L. J., Vega, J. L., Klasson, K. T., Clausen, E. C., and Gaddy, J. L.: The effects of furfural on ethanol production by Saccharomyces cerevisiae in batch culture. Biomass Bioenergy, 3, 41-48 (1992). 5. Zauner, E., Bronn, W. K., Dellweg, H., and Tress& R.: Inhibitory activity of impurities of beet molasses and plant protectants on yeast (Saccharomyces cerevisiae) respiration and

174

TAHERZADEH

ET AL.

fermentation. Branntweinwirtschaft, 119, 154-163 (1979). 6. Banerjee, N., Bhatnagar, R., and Viswanathan, L.: Inhibition of glycolysis by furfural in Saccharomyces cerevisiae. Eur. J. Appl. Microbial. Biotechnol., 11, 226-228 (1981). 7. Banerjee, N., Bhatnagar, R., and Viswanathan, L.: Development of resistance in Saccharomyw cerevisiae against inhibitory effects of browning reaction products. Enzyme Microb. Technol., 3, 24-28 (1981). 8. Sanchez, B. and Bautista, J.: Effects of furfural and 5-hydroxymethylfurfural on the fermentation of Saccharomyces cerevisiae and biomass production from Candida guilliermondii. Enzyme Microb. Technol., 10, 315-318 (1988). 9. Windisch, W.: Uber die Bildung und den Verbleib des Furfurols, sowie dessen Bedeutung im Brauereibetriebe. Chemisches Central-Blatt, 69, 1213-1214 (1898). 10. Diaz de Villegas, M. E., Villa, P., Guerra, M., Rodriguez, E., Redondo, D., and Martinez, A.: Conversion of furfural into furfuryl alcohol by Saccharomyces cerevisiae 354. Acta Biotechnol., 12, 351-354 (1992). 11. Villa, G. P., Bartroli, R., Lopez, R., Guerra, M., Enrique, M., Penas, M., Rodriguez, E., Redondo, D., Iglesias, I., and Diaz, M.: Microbial transformation of furfural to furfuryl alcohol by Saccharomyces cerevisiae. Acta Biotechnol., 12, 509-512 (1992). 12. Morimoto, S., Hirashima, T., and Matutani, N.: Studies on fermentation product from furfural by yeast, III. Identification of furoin and furil. J. Ferment. Technol., 47, 486-490 (1969). 13. Fireoved, R.L. and Mutharasan, R.: Effect of furfural and ethanol on the growth and energetics of yeast under microaerobit conditions. Ann. N. Y. Acad. Sci., 469,433-446 (1986). 14. Taherzadeh, M. J., Ekluod, R., Gustafsson, L., Nlklasson, C., and Lid&n, G.: Characterization and fermentation of diluteacid hydrolyzates from wood. Ind. Eng. Chem. Res., 36, 46594665 (1997). 15. Taherzadeh, M. J., Liden, G., Gustafssoo, L., and Niklasson, C.: The effects of pantothenate deficiency and acetate addition on anaerobic batch fermentation of glucose by Saccharomyces

J. BIOSCI. BIOENG.. cerevisiae. Appl. Microbial. Biotechnol., 46, 176-182 (1996). 16. Christensen, L., Schulze, U., Nielsen, J., and Vllladsen, J.: Acoustic off-gas analyser for bioreactors. Precision, accuracy and dynamics of detection. Chem. Eng. Sci., 50, 2601-2610 (1995). 17. Benthin, S., Nielsen, J., and Villadsen, J.: Characterization and application of precise and robust flow-injection analysers for on-line measurement during fermentations. Anal. Chim. Acta, 247, 45-50 (1991). 18. Verduyn, C., Postma, E., Scheffers, W. A., and van-Dijken, J. P.: Physiology of Saccharomyces cerevisiae in anaerobic glucose-limited chemostat cultures. J. Gen. Microb., 136, 395403 (1990). 19. Herbert, G., Phillips, P. J., and Strange, R. E.: Chemical analysis of microbial cells, p. 209-344. In Norris, J. R. and Ribbons, D. W. (ed.), Methods in microbiology, 5B. Academic Press, London (1971). 20. Fiechter, A., Fuhrmann, G. F., and Kitppeli, 0.: Regulation of glucose metabolism in growing yeast cells. Adv. Microbial. Physiol., 22, 123-183 (1981). 21. Palmqvist, E., AImeida, J., and Hahn-Hiigerdal, B.: Influence of furfural on anaerobic glycolytic kinetics of Saccharomyces cerevisiae in batch culture. Biotechnol. Bioeng. (1999). (in press) 22. Shvets, V. N., Nlkolalchuk, T.V., Slyusarenko, T. P., and Usenko, V. A.: Effect of furfural on alcohol fermentation, alcohol quality, and bakers’ yeasts. Izv. Vyssh. Uchebn. Zaved., Pishch. Tekhnol., 2, 48-52 (1977). 23. Kang, S. S. and Okada, H.: Alcohol dehydrogenase of Cephalosporium sp. induced by furfuryl alcohol. J. Ferment. Technol., 51, 118-124 (1973). 24. Taherzadeh, M. J., Niklasson, C., and Lid&, G.: Acetic acid-friend or foe in anaerobic batch conversion of glucose to ethanol by Saccharomyces cerevisiae? Chem. Eng. Sci., 52, 2653-2659 (1997). 25. Nordstrom, K.: Yeast growth and glycerol formation. Acta Chem. Stand., 20, 1016-1025 (1966).