Production of ethanol from wood hydrolyzate by yeasts

a ELSEVIER Bioresource Technology 72 (2000) 253-260 BIORESOURCE TECHNOLOGY b Production of ethanol from wood hydrolyzate by yeasts H.K. Sreenath a, T.W. Jeffries b,* a Department of Bioiogical Systems Engineering. University of Wisconsin, Madison, Wisconsin, USA Institute of Microbial Biochemical Technology, USDA Forest Service, Forest Products Laboratory, One Gifford Pinchot Drive, Madison, Wisconsin, WI 53705-2398, USA Received 18 March 1999; received in revised form 19 July 1999; accepted 29 July 1999 Abstract A total of 43 Forest Products Laboratory (FPL) strains of Pichia stipitis and Candida shehatae were tested for their ability to ferment a 1:1 mixture of glucose and xylose to-ethanol prior to fermentation of partially deacidified wood hydrolyzates. The starting sugar composition, ph, and concentrations of inhibitors such as acetic acid, furfural, and hydroxy methyl furfural varied from one batch to another. The delay observed in growth and fermentation depended on the amounts of inhibitors present and on the capacity of the strain to resist them. The ethanol production and yields obtained with C. shehatae were higher than those with P. stipitis. C. shehatae strain FPL-Y-049 produced up to 34 g/l ethanol from batch VII of wood hydrolyzates. Allthe glucose, mannose, galactose and xylose in the woodhydrolyzate were consumedduring fermentation. Only arabinose was unused. Addition of 10 mg/l zinc toacidhydrolyzate did not affectpeak ethanol production, but it did increase rates of sugar utilization and ethanol production. The fermentation of hydrolyzate withrecycled cells of C. shehatae Y-049 reduced the fermentation lag and increased the final ethanol concentration.the ethanol production rate was optimum in the ph range of 5.5-6.0, and an ethanol yield of 0.41-0.46 g/g was obtained in these fermentations. Because of heterogeneity between wood hydrolyzate batches, ethanol production was found to be influenced by hydrolyzate composition, ph, acetate concentration, amount of cells, and recycling of cells. Published by Elsevier Science Ltd. Keywords: Wood hydrolyzate; Pichia stipitis; Cadida shehatae;fply-049; Fermentation; Optimization; Ethanol production; Cell recycling 1. Introduction Xylose is the major sugar in the hemicellulosic fraction of agricultural and hardwood hydrolyzates (Grohmann et al., 1986). Dilute acid hydrolysis solubilizes hemicellulosic sugars and increases porosity of plant cell walls. Secondary hydrolysis releases glucose' and leaves residual lignin (Wayman and Tsuyuki, 1985). However, acid pretreatment forms inhibitors such as acetic acid, furfural, and compounds derived from lignin (Tran and Chambers, 1986; van Zyl et al., 1991; Schneider, 1996; Wilson et al., 1989; Wang et al., 1994; Parekh et al., 1986). The Forest Products Laboratory is maintained in cooperation with the University of Wisconsin. This article was written and prepared by US Government employees on official time, and it is therefore in the public domain and not subject to copyright. The use of trade or from names in this publication is for reader information and does not imply endorsement by the US Department of Agriculture of any product or service. *Corresponding author. Several naturally occurring or genetically engineered yeasts can convert hexoses and pentoses to ethanol (Parekh et al., 1986; Delgenes et al., 1988; Jeffries, 1982; Zhang et al., 1995; Sreenath and Jeffries, 1996; Hahn- Hagerdal et al., 1994; Ligthelm et al., 1988; Laplace et al.,1992). However, their capacity to ferment these sugars from acid hydrolyzates of wood using yeasts is hindered by the presence of inhibitors (Tran and Chambers, 1986; van Zyl et al., 1991; Schneider, 1996; Wilson et al., 1989; Wang et al., 1994; Parekh et al., 1986). Acetic acid is one of themost prevalent. At the ph optimum for fermentation (5.5-6.0). acetic acid is largely undissociated. This permits diffusion into the cell cytoplasm, where it dissociates and decreases the intracellular ph. As a result, the proton gradient across the membrane cannot be maintained and the transport of various nutrients is impaired (Ko and Edwards, 1975; Herrero et al., 1985), Hence, in the presence of acetate, yeast fermentation of wood hydrolyzates is poor. Previous researchers have attempted to lower acetate by continuous cellrecycling arid have developed strains with improved tolerance (Tran and Chambers, 1986; 0960-8524/00/$ - see front matter. Published by Elsevier Science Ltd. PII:S0960-8524(99)00113-3

254 H.K Sreenath. T. W. Jeffries / Bioresource Technology 72 (2000) 253-260, van Zyl et al., 1991; Schneider, 1996; Wilson et al., 1989; Parekh et al., 1986; Tran and Chambers, 1985, Jeffries and Sreenath, 1988). In the research reported here, we lowered acetate from acid hyd olzates of wood ("deacidified wood hydrolyzate") using electrodialysis (Datta, 1989). The objective of our work was to investigate selected yeast strains for growth and ethanol fermentation on wood hydrolyzates and to optimize ethanol production with the best yeast strain available. 2. Methods 2.1. Wood hydrolyzate tained 1.7 g/l filter-sterilized yeast nitrogen base (YNB) (without amino acid and ammonium sulfate) (Difco, Detroit, Michigan) and 8.83 g/1 total nitrogen (2.27 g/l urea and 6.56 g/1 peptone); ph was 5.0. During initial screening of yeasts, mixtures of pure sugars were used as the carbon source for the fermentations. Otherwise, in most experiments, samples of wood hydrolyzate containing hexoses and pentoses in the medium were the main carbon source for the organisms employed. The hydrolyzate samples were always filter-sterilized, whereas nitrogenous nutrients and supplements were autoclaved and added to the medium at room temperature (25-27 C). The final ph was adjusted to 5.5-6 using sterile 1M sodium hydroxide. Samples of eight batches of wood (mixture of hardwood and softwood) hydrolyzates ranging from 1.6 to 4.9 kg were supplied by Michigan Biotechnology Institute, Lansing, MI. Wood was.hydrolyzed by dilute acid at various liquid to solid ratios and partially deacidified by electrodialysis (Datta, 1989). The chemical composition and characteristics of various batches of partially deacidified wood hydrolyzate samples are shown in Table 1. The wood hydrolyzate samples contained sugars such as glucose, xylose, arabinose, galactose, and mannose. 2.2. Yeast strains Test strains of Pichia stipitis and Candida shehatae (Jeffries and Livingston, 1992) were from the FPL yeast collection, which is stored in 15% glycerol at -80 C. These strains were grown and maintained on fresh plates of yeast extract peptone xylose (YEPX)agar (see Media) for 48 h at 30-32 C. 2.3. Media Yeasts were grown on YEPX agar plates consisting of 10 g/l yeast extract, 20 g/l peptone, 20 g/l xylose, and 20 g/l agar. The undiluted fernentation medium con- 2.4. Shake flask fermentation 2.4.1. Mixed sugars Shake flask fermentation was conducted in triplicate with mixed sugars (80 g/l final, 40 g/l each xylose and glucose). Erlenmeyer flasks (125 ml) containing 50 ml of the fermentation medium were inoculated with 5 ml of various cell suspensions. Inocula were prepared by washing cells from the surfaces of 2-day-old YEPX agar plate cultures. Cell inocula were adjusted to 2.5 g/l. Fermentation flaskswere shaken at 100 rpm at25-27 C. Fermentation was monitored for 3-5 days by removing 1.3 ml samples or sugar and ethanol analyses. 2.4.2. Wood hydrolyzate (batches I-VIII) Batches of wood hydrolyzates were partially deacidified using an electrodialysis concentration process to decrease the inhibitors and concentrate the wood sugars (Datta, 1989). The shake flask fermentation for the wood hydrolyzate was conducted in triplicate similar to fermentation for mixed sugars. Erlenmeyer flasks (125 ml) containing 50 ml of fermentation medium, which contained 40 ml filter-sterilized partially deacidified wood hydrolyzates (batches I-VIII), 1.7 g/l filter-sterilized YNB (Difco, Michigan), and 8.83 g/l total nitrogen nutrients (2.27 g/l urea and 6.56 g/l peptone). Flasks Table 1 Chemical composition and characteristics of partially deacidified wood hydrolyzates a Acid hydrolyzate ph Total fermentable sugars b (g/l) Acetate concentration (g/l) Hydroxymethl furfural (g/l) Batch I 4.55 (0.05) A 10.0 (0.12) A 0.43 (0.01) A 0.03 (0.000) A Batch II 7.30 (0.04) B 42.3 (1.30) B 0.63 (0.03) AB 0.14 (0.005) B Batch III 3.00 (0.00) C 62.5 (0.29) C 4.31 (0.02) C 1.33 (0.005) C Batch IV 3.20 (0.01) D 70.4 (0.58) D 6.15 (0.06) D 2.20 (0.005) D Batch V 2.32 (0.01) E 77.8 (0.58) E 5.80 (0.10) E ND Batch VI 5.10 (0.05) F 57.8 (0.12) F 5.80 (0.00) E ND Batch VII 5.30 (0.02) G 121.7 (0.40) G 0.71 (0.01) B 0.20 (0.02) E Batch VIII 7.74 (0.05) H 69.3 (0.23) D ND ND a ND is not determined. Numbers in parentheses are standard errors. Means in the samecoiumn with the same alphabetic label are not significantly different at the 0.05 level. b Sugar concentration was determined using HPLC after diluting original wood hydrolyzate by 20% with other fermentation media components.

H.K. Sreenath, T. W. Jeffries / Bioresource Technology 72 (2000) 253-260 255 were inoculated with an initial 1.5-3.0 g/l fresh or re- YNB and 4.41 g/l total nitrogen nutrients (1.13 g/l urea cycled C. sheliatae FPL-Y-049 and shaken at 100 rpm at and 3.28 g/l peptone). Seventy milliliters of high-density 25 C. The wood hydrolyzate was initially diluted 20% recycled cells (second recycle) were inoculated into the by adding nitrogenous nutrients and cell inoculum. The reactor so that a final OD of 15 was achieved. The ph of ph of acid hydrolyzates of various batches was adjusted reactor operation was initially adjusted in the range of to 5.5-6 using 1 M sodium'hydroxide. The shake flask 5.5-5.7 at 25-27 C. The reactor vessel was agitated at fermentation was optimized with various reduced levels 250 rpm and sparged with air at 1.5 l/min. The ph was of nitrogen nutrients with fresh or recycled cells in the increased from 5.5 to 7.0 during fermentation run by presence or absence of 10 mg/l zinc sulfate. addition of 1 M sodium hydroxide. Samples (2.5 ml) were taken periodically for cell density measurements 2.4.3. Analytical methods and other analyses. Cell densities were measured at 525 nm after diluting cells from 0.05 to 0.5 optical density (OD). An OD of 1.0 2.5. Statistical analyses was equivalent to 0.21 mg dry weight of cells/ml. Ethanol was separated by gas chromatography (Jeffries, Statistical analyses primariiy consisted of summary 1982). Sugars and other fermentation byproducts were statistics, including means and standard errors, where determined by high performance liquid chromatography appropriate; simple t-tests for mean differences; and (HPLC) (Verhaar and Kuster, 1981). The amounts of one-way analysis of variance followed by mean comhexoses and pentoses in the wood.hydrolyzate samples, parisons using Tukey's procedure conducted at the 0.05 such as glucose, galactose, mannose, and xylose, were level (Montgomery, 1997). When appropriate, the wood determined by HPLC and expressed as total fermentable hydrolyzate batch was included as a blocking factor. sugars. The amount of arabinose quantified in the wood Most analyses were performed with SAS version 6.12 hydrolyzate was expressed separately. software (SAS Institute, 1989). The figures do not include standard error bars since relative errors were 2.4.4. Optimization of nutrients dwarfed by graphical symbols. For medium optimization, nutrients were diluted to one-half, one-fourth, one-tenth, and one-twentieth. Corn steep liquor and molasses, 56 g/l (final), was sep- 3. Resultsanddiscussion arately added to partially deacidified wood hydrolyzate fermentation without nitrogenous supplements. For 3.1. Fermentation of mixed sugars by FPL yeast strains other experiments, the fermentation medium was supplemented with biotin (0.025 mg/l) and thiamine (5 mg/l) More than 40 FPL yeast strains of P. stipitis and C. in the presence or absence of other nitrogen sources. In shehatae were screened to determine their fermentation other experiments, 4 g/l solids of fermentation stillage rates on mixed sugars. All of the tested FPL strains concentrate of wood hydrolyzate was added to a fresh fermented both glucose and xylose. The average ethanol batch of wood hydrolyzate in the presence or absence of concentration attained was 34.8 ± 2.42 g/l with nitrogenous nutrients. Addition of ZnSO 4 (10 mg/l) to P. stipitis and 34.0 ± 1.67 g/l with C. shehatae. The avwood hydrolyzate fermentation medium was also at- erage ethanol yield was 0.44 ± 0.02 and 0.43± 0.02 g/g tempted. for P. stipitis and C. shehatue, respectively. Some C. shehatae strains exhibited 50-60% less cell growth than 2.4.5. Cell recycling did P. stipitis strains for the sameamount of sugar The cells of C. shehatae Y-049 were recovered asep- consumed. Consequently, C. shehatae had higher spetically from each round of fermentation of wood cific ethanol production rates. Strains of C. shehatae hydrolyzates. Recycled cells were inoculated into a fresh generally assimilated glucose and xylose faster than did batch of wood hydrolyzate for fermentation, The effect P. stipiris strains. Production of polyols was lower than of cell recycling and number of recycling cycles on eth- that of other sugarsin both yeast strains under these anol production was studied. The effectof cell recycling during optimization of nutrient substitution on fer mentation of wood hydrolyzate was also studied. Batches of wood hydrolyzate containing 10-121 g/l total fermentable sugars were used (Table 1). The pri mary sugars were xylose, glucose, mannose, galactose, and arabinose. The amounts of xylose, glucose, man nose, and galactose were totaled and expressed as total 2.4.6. ph-controlled batch reactor fermentation of wood hydrolyzate The ph of wood hydrolyzate (batch VI) was adjusted to 5.5-7.0 using 1 M sodium hydroxide and ethanol fermentation was studied. The 1.3-1 batch reactor contained 1.2 1 wood hydrolyzate (batch IV) with 0.85 g/l fermentation conditions. 3.2. Composition of partially deacidified wood hydrolyz ates

256 H.K Sreenath. T.W.Jeffries I Bioresource Technology 72 (2000) 253-260 fermentable sugars. The amount of arabinose was expressed separately. The ph of these batches of partially deacidified wood hydrolyzates ranged from 2.3 to 7.7 and acetic acid ranged from 0.4 to 6.15 g/l. The hydroxy methyl furfural (HMF) values ranged from 0.03 to 2.2 g/l, and there were trace amounts of furfural. Some batches of hydrolyzates required ph adjustment to 5.5-6.0 prior to fermentation. For many acid hydrolyzate batches, such as batches I, II, VI, VII, and VIII, no delay in fermentation by C. shehatae FPLY-049 was observed since acetate content was low. In batches III, IV, and V, acetate content was high and the ph was therefore adjusted to 6.0. In these batches of wood hydrolyzates, growth of C. shehatae FPL-Y-049 was delayed and fermentation initiated after 20-40 h. The initial delayed growth observed in fermentation of these hydrolyzates was possibly due to cellular adjustment in response to increased acetate content (Parekh et al., 1986;Jeffries and Sreenath, 1988). The ethanol yield and production and fermentation time depended on characteristics of hydrolyzates. 3.3. Fermentation of wood hydrolyzate by C. shehatae FPL-Y-049 Compared to P. stipitis strains and other strains of C. shehatae, the FPL-Y-049 C. shehatae strain produced the most ethanol from the wood hydrolyzate under shake flask conditions. C. shehatae produced more than 20 g/l ethanol from wood hydrolyzate in shake flask fermentation in 60 h (Fig. l(a)) whereas P. stipitis produced 15.7 g/l ethanol (Fig. 1(b)). All sugars except araginose were assimilated completely (Fig. 1). Traces of arabitol were produced in the late fag phase of fermentation (data not shown). Ethanol production by C. shehatae FPL-Y-049 was not affected by batch-to-batch feeding of wood hydrolyzate sugars during fermentation (data not shown). Ethanol production rates and yields on various batches of acid hydrolyzates are summarized in Tables 2. Ethanolic fermentation was enhanced in batches of acid hydrolyzates in which sugars were concentrated with lower acetic acid levels. 3.4. Optimizationof nutrients during C. shehatae FPL-Y- 049 fermentation 3.4.1. Effect of reduced nutrients Reducing the total nutrients (nitrogen supplements and YNB) from 10.53 to 5.26 or 2.63 g/l did not change ethanol production significantly ( P>0.05) However, further reduction of nutrients to 1.05 or 0.53 g/l reduced the rate of ethanol production considerably (P = 0.01). Ethanol production and yield obtained during optimization of nitrogen supplementation during fermentation Fig. 1. Fermentation of wood hydrolyzate (batch II) during ethanol production: (a) C. shehatae FPL-Y-049, (b) P. stipitis FPL-Y-606. of acid hydrolyzates by C. shehatae FPL-Y-049 are summarized in Table 3). To produce 26-28 g/l ethanol during fermentation of various batches of wood hydrolyzates under experimental conditions, the optimum level of nutrients used was 2.63 g/l. Similar obsevations on effect of nutrients in ethanol fermentation from xylose in C. shehatae have been reported previously (Jeffries, 1985). 3.4.2. Effect of ZnSO 4 The presence of 10 mg/l zinc along with other nutrient supplements in the fermentation medium of wood hydrolyzate increased the rate of sugar utilization when compared to the absence of zinc (Fig. 2(a) and (b)) (P = 0.02). The increase inethanol production caused by the addition of zinc was probably due to high activity of zinc-dependent alcohol dehydrogenase (ADH)in the fermentative pathway. ADH has been extensively studied in Saccharomyces cerevisiae, where it occurs as four isozymes and has been found to be zinc dependent (Hahn-Hagerdal et al., 1994; Jornvall et al., 1987). 3.4.3. Effect of vitamin and stillage concentrate supplements Reducing nutrients to 2.63 g/l+ 10 mg/l zinc in the fermentation medium yielded 26-30 g/l ethanol in various batches of ph-adjusted (6.0) wood hydrolyzates

H.K. Sreenath, T.W. Jeffries / Bioresource Technology 72 (2000) 253-260 257 Table 2 Ethanol production by C. shehatae FPL-Y-049 during fermentation of wood hydrolyzates a Acid hydrolyzate b Fermentation time (h) Ethanol concentration (g/l) Ethanol yield (g/g) Residual sugars (g/l) Batch I 48 3.47 (0.00) A 0.35 (0.004) A 0.0 (0.00) A Batch II 50 21.20 (0.35) BC 0.51 (0.004) B 0.7 (0.65) A Batch III 118 26.41 (0.55) D 0.48 (0.007) CD 7.3 (0.32) B Batch IV 118 25.56 (0.58) D 0.48 (0.005) C 16.8 (0.44) C Batch V 146 30.21 (0.58) E 0.46 (0.007) DE 12.0 (0.58) D Batch VI 46 23.00 (0.12) B 0.44 (0.002) E 5.0 (0.06) E Batch VII 76 34.50 (0.52) F 0.32 (0.004) F 13.6 (0.06) D Batch VIII 76 19.73 (0.30) C 0.33 (0.004) AF 9.4 (0.26) F a Fermentations were performed in 125-ml Erlenmeyer flasksin 50-ml medium consisting of wood hydrolyzate with 2.63 g/l nutrient supplements and 10 mg/lzinc with 2.5 g/l C.shehatae and shaken at 100 rpm at 25 C. Numbers in parentheses are standard errors. Means in the same column with the same alphabetic label are not significantly different at the 0.05 level. b The ph of batches III, IV, and V rangedfrom 2.3 to 3.2; hence, ph was adjusted to 5.5-6.0 by adding 1 M NaOH. Table 3 Effect of nutrient supplement dilutions on ethanol production during fermentation of hydrolyzates by C. shehataefpl-y-049 a Acid hydrolyzate Type of cell inoculum b Nutrient supplement c (g/l) Max ethanol concentration (gl) Ethanol yield (g/g) Batch V d Fresh 10.53 29.8 (1.17) 0.45 (0.018) Recycled lx 5.26 2.63 1.05 0.53 0 30.9 (0.35) 32.2 (0.75) 23.1 (1.10) 22.3 (1.15) 16.1 (1.21) 0.42 (0.006) 0.44 (0.003) 0.42 (0.003) 0.40 (0.006) 0.39 (0.006) Batch VI e Fresh 10.53 20.3 (0.40) 5.26 21.9 (0.40) Recycled lx 2.63 24.1 (0.59) 1.05 17.5 (0.72) 0.53 13.8 (2.46) 0 10.4 (1.27) 0.44 (0.006) 0.45 (0.003) 0.48 (0.001) 0.36 (0.003) 0.35 (0.002) 0.40 (0.005) a Fermentation was performed in 125-ml Erlenmeyer flasks with 50 ml medium consisting of acid hydrolyzate with various amounts of nutrient supplements and 10 mg/l zinc. Flasks were shaken at 100 rpm at 25 C. Numbers in parentheses are standard errors. b With 2.5 g/l cells. c The basic nutrient supplement (10.53 g/l)consisted of 1.7 g/l YNB, 2.27 g/l urea, and 656 g/l peptone. Other supplements werediluted 1:2,1:4,1:10, and 1:20. d The wood hydrolyzate had an acidic ph of 2.3, and hence the ph was adjusted to 5.5-6.0 using 1 M NaOH; fermentation period was 5-6 days. e Original ph of wood hydrolyzate was 5.1 and fermentation period was 2-3 days. (batches III, IV, and V). Ethanol production and yield obtained during fermentation of batch VI with or without nutrients or other vitamins and stillage concentrate supplementation are summarized in Table 4. When the hydrolyzate was fermented without nutrient supplements, 10 g/l ethanol was obtained. Supplementing the medium with only biotin and thiamine slightly enhanced ethanol concentration to 12 g/l (Table 4). However, yeast growth was nearly identical under all of these conditions. Similarly, supplementing with 0.53 g/l nutrients (YNB +total nitrogen supplements) with thiamine and biotin caused only a slight increase in ethanol production. Recycling of stillage concentrate (prepared by concentrating post-fermentation medium as a source of residual sugars and other nutrients) during fermentation of hydrolyzate (batch VI, ph 5.1)with recycled cells and 10 mg A zinc was found to be useful since it supported yeast growth and produced 20 g/l ethanol. 3.4.4. Effect of corn steep liquor and molasses supplements Inexpensive supplements such as corn steep liquor (CSL) have been substituted for expensive nitrogen supplements in pentose fermentation mediated by strains of P. stipitis and C shehatae (Sreenath andjeffries, 1996; Amartey and Jeffries, 1994). In the work reported here, both CSL and molasses were employed in the fermentation of wood hydrolyzates. Addition of corn steep liquor alone to the hydrolyzate did not support yeast growth and fermentation, whereas addition of molasses supported slow growth and delayed fermentation (data not shown). At the end of fermentation (205 h), ethanol concentration was 26.5 g/l. 3.4.5. Effect of cell recycling Fresh cells of C. shehatae FPL-Y-049 exhibited delayed growth and fermentation in some of these batches of partially deacidified wood hydrolyzate up to 46 h.

258 H.K Sreenuth, T.W. Jeffries / Bioresource Technology 72 (2000) 253-260 Fig. 2. Effect of zinc on fermentation of wood hydrolyzate (batch V) by C. shehatae FPL-Y-049: (a) no zinc, (b) 10 mg/l zinc. Ethanol production was reduced (Fig. 3(a)). In contrast, recycled cells of C. shehatae FPL-Y-049 grew after 20 h and produced 32 g/l ethanol at the end of fermentation (Fig. 3(b)). The fermentation rate of recycled cells was slightly faster than that of fresh cell inoculum since the recycled cells were readily exposed to various inhibitor components of acid hydrolyzates (Table 1). Similar increased resistance to hydrolyzate inhibitors by C. shehatae and P. stipitis was reported after repeated recycling in wood hydrolyzates, resulting in production of much higher ethanol concentrations (Parekh et al., 1986; du Preez, 1994). Fig. 3. Effect of cell recycling on fermentation of wood hydrolyzate (batch V) by C. shehatae FTCY-049: (a) fresh cells, (b) recycled cells. The effects of number of recycling cycles and cell density on ethanol production during hydrolyzate fermentation are summarized in Table 5. Ethanol production was further enhanced when a third round of recycled cells was used in hydrolyzate fermentation (batch VIII). Ethanol production may have also been infuenced by the density of recycled cells. In batch VI of wood hydrolyzate, ethanol yield was higher with high density cells (Table 5, P = 0.02), whereas in batch V, ethanol yield was similar for both cell densities Table 4 Effect of other nutrient supplements on ethanol production during fermentation of wood hydrolyzate (batch VI) with C. shehatae FPL-Y-049 a Type of supplement Recycled cell in- Maximum ethanol Ethanol yield (g/g) Fermentation oculum (no. cycles) concentration (g/l) time (h) Hydrolyzate alone lx 10.43 (1.26) A 0.40 (0.005) A 92 Hydrolyzate +biotin +thiamine 1x 12.48 (0.65) AB 0.34 (0.003) B 116.5 Hydrolyzate +recycled hydrolyzate lx 20.04 (0.01) C 0.48 (0.002) C 92 conantrate Hydrolyzate +0.53 g/l nitrogenous 2x 13.82 (2.46) AB 0.35 (0.005) B 84 nutrient supplements Hydrolyzate +0.53 g/l nitrogenous 2x 16.91 (0.38) BC 0.47 (0.002) CD 84 nutrient supplements +recycled hydrolyzate concentrate Hydrolyzate + 0.53 g/l nitrogenous 2x 15.02 (0.39) ABC 0.46 (0.002) D 84 nutrient supplements +biotin + thiamine a Fermentation conducted in 50 ml of medium consisting of wood hydrolyzate with 10 mg/l zinc and 2.5 g/l cell inoculum with specifiedsupplements. Flasks were shaken at 100 rpm, ph 5.1 at 25 C. Numbers in parentheses are standard errors. Means in the same column with the same alphabetic label are not significantly different at the 0.05 level.

H.K Sreenath, T.W. Jeffries/ Bioresource Technology 72 (2000) 253-260 259 Table 5 Effectof cell recycling on fimnentation of wood hydrolyzates by C. shehatae FPLY-049 a Acid Type of cell Maximum ethanol Ethanol yield Fermentation Cell density b hydrolyzate inoculum concentration (g/l) (g/g) time (h) Batch V Fresh 29.80 (1.17) A 0.45 (0.003) A 119 High Recycled 1x 32.16 (0.72) A 0.44 (0.005) A 167 Medium Batch VI Recycled 1x 23.28 (0.56) B 0.49 (0.003) B 46 High Recycled 2x 20.72 (0.35) C 0.39 (0.001) C 84 Low Batch VIII Recycled lx 19.55 (0.47) D 0.33 (0.003) D 76.5 High Recycled 3x 21.43 (0.37) E 0.33 (0.002) D 76.5 High a Fermentations performed in 50 ml of medium consisting of wood hydrolyzate with 2.63 g/1nutrient supplementsand 10 mg/l zinc, ph 5.1-6.0, 100 rpm, 25 C, Numbers in parentheses are standard errors. Within each batch, means in the same column with the same alphabetic label are not significantly different at the 0.05 level. b Cell density (g/l): high 2.5. medium 1.5-2.5, low 1.5. (Table 5, P = 0.14). Because of heterogeneity between wood hydrolyzate batches, ethanol production was found to be influenced by hydrolyzate composition, ph, acetate concentration, amount of cells, and recycling of cells. 3.5. Effect of ph The rate of ethanol production from wood hydrolyzates was not significantly affected by ph in the range of 5.1-7.0 (data not shown, P > 0.05).Ethanol production was maximum at ph 6.0. The fermentation of mixed sugars in this batch of hydrolyzate (batch VI) was completed rapidly (in less than 48 h) because of the low acetate content. Acid hydrolyzate batch VIII, which had a ph of 7.7, produced 30% less ethanol by fermentation and hence suffered lower yields. For acid hydrolyzate batches with a ph of less than 4, growth and fermentation did not occur unless the ph was readjusted to ph 5.5-6. 3.6. ph-controlled batch reactor fermentation In the batch reactor, fermentation of wood hydrolyzate (batch IV) was studied in ph-controlled conditions. The ph range of 5.6-6.4 was obtained by addition of potassium hydroxide. Recycled cells and reduced nutrient supplementationwith zinc were used in the fermentation medium. The growth of C. shehatae FPL-Y-049 did not start until 24 h had passed as a result of adaptation to high acetate content of the hydrolyzate (Parekh et al., 1986). At ph 5.7, ethanol concentration was 20 g/l but ethanol yield was 0.38 (Fig. 4). The lower ethanol concentration could be attributed to either ethanol respiration or low levels of inhibition as a consequence of potassium acetate formation in the medium during buffering to maintain ph of fermentation. Xylitol (5-7 g/l) accumulated in the reactor as well as low levels of glycerol. Fig. 4. Effect of ph on ethanol production in fermentation bioreactor with C. shehatae FPL-Y-049 using wood hydrolyzate batch IV. 3.7. Sugar utilization The C. shehatae FPL-Y-049 strain completely utilized xylose, glucose, mannose, and galactose for growth and ethanol fermentation of the wood hydrolyzate. However, this strain did not ferment arabinose. When the yeast was given only arabinose, growth started very slowly and arabinose was subsequently converted into arabitol (data not shown). Ethanol fermentation was efficient, in the range of 40-80 g/l hydrolyzate sugar concentrations. In the hydrolyzate batch with 120 g/1 fermentable sugar (batch VII), only 35 g/l ethanol was obtained after about 76 h fermentation, as a result of interference 'by inhibitors still present in the hydrolyzate batch. 4. Conclusions A higher ethanol yield was obtained with Candida shehatae compared with Pichia stipitis during wood hydrolyzate fermentation. The best fermentation strain was C. shehatae FPL-Y-049. Ethanol production in various batches of hydrolyzates ranged from 20 to 35 g/ l, and ethanol yield ranged from 0.33 to 0.49 g/g. The

260 H.K. Sreenath, T.W. Jeffries I Bioresource Technology 72 (2000) 253-260 hydrolyzate batches differed in sugar composition, ph, acetate content, and other components. The higher acetate content of some batches delayed growth for 20-40 h before fermentation. When acetate content was less than 0.5-0.7 g/l, C. shehatae FPL-Y-049 grew well and fermented within 48-72 h. All fermentable sugars except arabinose were used efficiently during fermentation. Nitrogen supplementation was required for fermentation. The optimal amount of nutrients, such as YNB, urea, and peptone, required for fermentation was 0.43, 0.57, and 1.64 g/l, respectively. Addition of 10 mg/l zinc, to fermentation of acid hydrolyzate enhanced ethanol production. The optimum ph of acid hydrolyzate fermentation for ethanol production was 5.5-6.0. Recycling of cells and the amount of cells were very important in enhancing ethanol production. Recycling of stillage (as a source of residual unfermented sugars) was also helpful in supporting growth and increasing fermentation. Acknowledgements The authors acknowledge technical help by Brian Manske and financial support from American Ethanol, Inc. References Amartey, S., Jeffries, T.W., 1994. Comparisonof corn steep liquor with other nutrients in the fermentation of D-xylose by Pichia stipitis CBS 6054. Biotechnol. Lett. 16,211-214. Datta, R., 1989. Recovery and purification of lactate salts from whole fermentation broth by electrodialysis. US patent 4 885 247. Delgenes, J.P., Moletta, R., Navarro, J.M., 1988. Fermentation of D xylose, D-glucose and L-arabinose mixture by Pichia stipitis Y 7124 Sugar tolerance. Appl. Microbiol. Biotechnol. 29, 155-161. dupreez, J.C., 1994. Process parameters and environmental factors affecting D-xylose fermentation by yeasts. Enzyme Microb. Technol. 16, 944-956. Grohmann, K., Torget. R., Himmel, M., 1986. Dilute acid pretreatment of biomass at high solids concentration Biotechnol. Bioeng. Symp. 17, 135-151. Hahn-Hagerdal, B., Jeppson. H., Skoog, K., Prior, B,A., 1994. Biochemistry and physiology of xylose fermentation by yeasts. Enzyme Microb. Technol. 16,933-943. Herrero, A.A., Gomez, R.F., Snedecor, B., Tolman, C.J., Roberts, M.F., 1985. Growth inhibition of Clostridium thermocellum by carboxylicacids. A mechanism based on uncoupling by weak acids. Appl. Microbiol. Biotechnol. 22, 53-62. Jeffries, T.W., 1982. A comparison of Candida tropicalis and Pachysolen tannophilus for conversion of xylose to ethanol. Biotechnol. Bioeng. Symp. 12, 103-110. Jeffnes,T.W., 1985. Effect of culture conditions on the fermentation of xylose to ethanol by Canidida shehatae. Biotechnol.Bioeng. Symp. 15, 149-166. Jeffries, T.W., Livingston, P.L., 1992. Xylose fermenting yeast mutants. US patent 5, 126,266. Jeffries, T.W., Sreenath, H.K., 1988. Fermentation of hemicellulosic sugars and sugar mixtures by Candida shehatae. Biotechnol. Bioeng. 31, 502-506. Jornvall, W., Jeffrey, J., Persson, B., 1987. Characteristicsof alcohol polyol dehydrogenases - The zinc containing long-chain alcohol dehydrogenases. Eur. J. Biochem. 167, 195-201. Ko, R.Y.C., Edwards, V.H., 1975. Transient kinetics of yeast growth in batch and continuous culture with an inhibitory carbon and energy source. Biotechnol. Bioeng. 17, 965-983. Laplace, J.M., Delgenes, J.P.. Molleta, R., Navarro, J.M., 1992. Fermentation of lignocellulosic sugars to ethanol - selection of mutants of Pichia stipitis affected for D-glucose utilization. Enzyme Microb. Tcchnol. 14, 644-648. Ligthelm, M.E., Prior, B.A., dupreez, J.C, 1988. The oxygen requirements of yeasts for the fermentation of D-xylose and D-glucose to ethanol. Appl. Microbiol. Biotechnol. 28.43-68. Montgomery, D.C., 1997. Design and Analysis of Experiments. Wiley, New York, NY. Parekh, S.R., Yu. S., Waymann. M., 1986. Adaptation of Candida shehatae and Pichia stipitis to wood hydrolyzatcs for increased ethanol production. Appl. Microbiol. Biotechnol. 25, 300-304. SAS Institute Inc., 1989. SAS/STAT User's guide. Version 6, 4th ed., vol. 1. SAS Institute Inc., Cary, NC. Schneider, H., 1996. Selective removal of acetic acid from hardwood - spent sulfite liquor using a mutant yeast. Enzyme Microb. Technol. 19, 94-98. Sreenath, H.K, Jeffries, T.W., 1996. Effect of corn steep liquor on fermentation of mixed sugarsby Candida shehatae FPL-702. Appl. Biochem. Biotechnol. 57/58, 551-561. Tran, A.V., Chambers, R.P., 1985. Red oak wood derived inhibitors in the ethanol fermentation of xylose by Pichia stipitis CBS 5776. Biotechnol. Lett. 7, 841-846. Tran, A.V.. Chambers, R.P., 1986. Ethanol fermentation of red oak acid hydrolyzate by the yeast Pichia stipitis CBS 5776. Enzyme Microb. Technol. 8, 439-444. van Zyl, C., Prior, B.A., dupreez, J.C., 1991. Acetic acid inhibition of D-xylose fermentation by Pichia stipitis. Enzyme Microb. Technol. 13, 82-86. Verhaar, L.A.Th., Kuster, B.F.M.. 1981. Improved column efficiency in chromatographic analysis of sugars on cation-exchange resins by use of water-triethylamine eluents. J. Chromatogr. 210,279-290. Wang, P., Brenchley, J.E., Humphrey, A.E., 1994. Screening microorganism and utilization of furfural, and possible intermediatesin its degradative pathway. Biotechnol. Lett. 16, 977-982. Wayman, M., Tsuyuki, S., 1985. Fermentation of xylose to ethanol by Candida shehatae. Biotechnol. Bioeng. Symp. 15, 167-177. Wilson, J.J., Deschatelets L.. Nishikawa, N.K., 1989. Comparative fermentability of enzymatic and acid hydrolyzatcs of steampretreated aspen wood hemicellulose by Pichia stipitis CBS 5776. Appl. Microbiol. Biotechnol. 31, 592-5%. Zhang, M., Eddy, C., Deanda, K., Finkelstein, M., Picataggio, S., 1995. Metabolic engineering of a pentose metabolism pathway in ethanologenic Zymomonas mobilis. Science 267, 240-243.