Conversion of glucose‐xylose mixtures to pyruvate using a consortium of metabolically engineered Escherichia coli

Two strains of Escherichia coli were engineered to accumulate pyruvic acid from two sugars found in lignocellulosic hydrolysates by knockouts in the aceE, ppsA, poxB, and ldhA genes. Additionally, since glucose and xylose are typically consumed sequentially due to carbon catabolite repression in E. coli, one strain (MEC590) was engineered to grow only on glucose while a second strain (MEC589) grew only on xylose. On a single substrate, each strain generated pyruvate at a yield of about 0.60 g/g in both continuous culture and batch culture. In a glucose‐xylose mixture under continuous culture, a consortium of both strains maintained a pyruvate yield greater than 0.60 g/g when three different concentrations of glucose and xylose were sequentially fed into the system. In a fed‐batch process, both sugars in a glucose‐xylose mixture were consumed simultaneously to accumulate 39 g/L pyruvate in less than 24 h at a yield of 0.59 g/g.     

Succinate production from xylose‐glucose mixtures using a consortium of engineered Escherichia coli

The conversion of variable sugar mixtures into biochemicals poses a challenge for a single microorganism. For example, succinate has not been effectively generated from mixtures of glucose and xylose. In this work, a consortium of two Escherichia coli strains converted xylose and glucose to succinate in a dual phase aerobic/anaerobic process. First, the optimal pathway from xylose or glucose to succinate was determined by expressing either heterologous pyruvate carboxylase or heterologous adenosine triphosphate‐forming phosphoenol pyruvate (PEP) carboxykinase. Expression of PEP carboxykinase (pck) resulted in higher yield (0.86 g/g) and specific productivity (155 mg/gh) for xylose conversion, while expression of pyruvate carboxylase (pyc) resulted in higher productivity (76 mg/gh) for glucose conversion. Then, processes using consortia of the two optimal xylose‐selective and glucose‐selective strains were designed for two different feed ratios of glucose/xylose. In each case the consortia generated over 40 g/L succinate efficiently with yields greater than 0.90 g succinate/g total sugar. This study demonstrates two advantages of microbial consortia for the conversion of sugar mixtures: each sugar‐to‐product pathway can be optimized independently, and the volumetric consumption rate for each sugar can be controlled independently, for example, by altering the biomass concentration of each consortium member strain.     

Accumulation of d-Glucose from Pentoses by Metabolically Engineered Escherichia coli

Escherichia coli that is unable to metabolize d-glucose (with knockouts in ptsG, manZ, and glk) accumulates a small amount of d-glucose (yield of about 0.01 g/g) during growth on the pentoses d-xylose or l-arabinose as a sole carbon source. Additional knockouts in the zwf and pfkA genes, encoding, respectively, d-glucose-6-phosphate 1-dehydrogenase and 6-phosphofructokinase I (E. coli MEC143), increased accumulation to greater than 1 g/liter d-glucose and 100 mg/liter d-mannose from 5 g/liter d-xylose or l-arabinose. Knockouts of other genes associated with interconversions of d-glucose-phosphates demonstrate that d-glucose is formed primarily by the dephosphorylation of d-glucose-6-phosphate. Under controlled batch conditions with 20 g/liter d-xylose, MEC143 generated 4.4 g/liter d-glucose and 0.6 g/liter d-mannose. The results establish a direct link between pentoses and hexoses and provide a novel strategy to increase carbon backbone length from five to six carbons by directing flux through the pentose phosphate pathway.     

Amino acid production from rice straw and wheat bran hydrolysates by recombinant pentose-utilizing <Emphasis Type="Italic">Corynebacterium glutamicum</Emphasis>

Corynebacterium glutamicum wild type lacks the ability to utilize the pentose fractions of lignocellulosic hydrolysates, but it is known that recombinants expressing the araBAD operon and/or the xylA gene from Escherichia coli are able to grow with the pentoses xylose and arabinose as sole carbon sources. Recombinant pentose-utilizing strains derived from C. glutamicum wild type or from the l-lysine-producing C. glutamicum strain DM1729 utilized arabinose and/or xylose when these were added as pure chemicals to glucose-based minimal medium or when they were present in acid hydrolysates of rice straw or wheat bran. The recombinants grew to higher biomass concentrations and produced more l-glutamate and l-lysine, respectively, than the empty vector control strains, which utilized the glucose fraction. Typically, arabinose and xylose were co-utilized by the recombinant strains along with glucose either when acid rice straw and wheat bran hydrolysates were used or when blends of pure arabinose, xylose, and glucose were used. With acid hydrolysates growth, amino acid production and sugar consumption were delayed and slower as compared to media with blends of pure arabinose, xylose, and glucose. The ethambutol-triggered production of up to 93 ± 4 mM l-glutamate by the wild type-derived pentose-utilizing recombinant and the production of up to 42 ± 2 mM l-lysine by the recombinant pentose-utilizing lysine producer on media containing acid rice straw or wheat bran hydrolysate as carbon and energy source revealed that acid hydrolysates of agricultural waste materials may provide an alternative feedstock for large-scale amino acid production.     

Improved strains of recombinant Escherichia coli for ethanol production from sugar mixtures

Hemicellulose hydrolysates of agricultural residues often contain mixtures of hexose and pentose sugars. Ethanologenic Escherichia coli that have been previously investigated preferentially ferment hexose sugars. In some cases, xylose fermentation was slow or incomplete. The purpose of this study was to develop improved ethanologenic E. coli strains for the fermentation of pentoses in sugar mixtures. Using fosfomycin as a selective agent, glucose-negative mutants of E. coli KO11 (containing chromosomally integrated genes encoding the ethanol pathway from Zymomonas mobilis) were isolated that were unable to ferment sugars transported by the phosphoenolpyruvate-dependent phosphotransferase system. These strains (SL31 and SL142) retained the ability to ferment sugars with independent transport systems such as arabinose and xylose and were used to ferment pentose sugars to ethanol selectively in the presence of high concentrations of glucose. Additional fosfomycin-resistant mutants were isolated that were superior to strain KO11 for ethanol production from hexose and pentose sugars. These hyperproductive strains (SL28 and SL40) retained the ability to metabolize all sugars tested, completed fermentations more rapidly, and achieved higher ethanol yields than the parent. Both SL28 and SL40 produced 60 gl^–1 ethanol from 120 gl^–1 xylose in 60 h, 20% more ethanol than KO11 under identical conditions. Further studies illustrated the feasibility of sequential fermentation. A mixture of hexose and pentose sugars was fermented with near theoretical yield by SL40 in the first step followed by a second fermentation in which yeast and glucose were added. Such a two-step approach can combine the attributes of ethanologenic E. coli for pentoses with the high ethanol tolerance of conventional yeasts in a single vessel.     

Ethanol production from pentoses and hexoses by recombinantEscherichia coli

Summary The fermentation of glucose (5g/L), xylose (80g/L) and arabinose (5g/L) produced 42.5g/L of ethanol in 96 hours, yielding 0.49g of alcohol per g of sugar using recombinantEscherichia coli. At these concentrations, the first sugar to be consumed was glucose, followed by arabinose and xylose last.     

Systems Metabolic Engineering of Escherichia coli Improves Coconversion of Lignocellulose‐Derived Sugars

Currently, microbial conversion of lignocellulose‐derived glucose and xylose to biofuels is hindered by the fact that most microbes (including Escherichia coli [E. coli], Saccharomyces cerevisiae, and Zymomonas mobilis) preferentially consume glucose first and consume xylose slowly after glucose is depleted in lignocellulosic hydrolysates. In this study, E. coli strains are developed that simultaneously utilize glucose and xylose in lignocellulosic biomass hydrolysate using genome‐scale models and adaptive laboratory evolution. E. coli strains are designed and constructed that coutilize glucose and xylose and adaptively evolve them to improve glucose and xylose utilization. Whole‐genome resequencing of the evolved strains find relevant mutations in metabolic and regulatory genes and the mutations’ involvement in sugar coutilization is investigated. The developed strains show significantly improved coconversion of sugars in lignocellulosic biomass hydrolysates and provide a promising platform for producing next‐generation biofuels.     

Growth of engineered Pseudomonas putida KT2440 on glucose,xylose, and arabinose: Hemicellulose hydrolysates and their major sugars as sustainable carbon sources

Lignocellulosic biomass is the most abundant bioresource on earth containing polymers mainly consisting of d ‐glucose, d ‐xylose, l ‐arabinose, and further sugars. In order to establish this alternative feedstock apart from applications in food, we engineered Pseudomonas putida KT2440 as microbial biocatalyst for the utilization of xylose and arabinose in addition to glucose as sole carbon sources. The d ‐xylose‐metabolizing strain P. putida KT2440_xylAB and l ‐arabinose‐metabolizing strain P. putida KT2440_araBAD were constructed by introducing respective operons from Escherichia coli. Surprisingly, we found out that both recombinant strains were able to grow on xylose as well as arabinose with high cell densities and growth rates comparable to glucose. In addition, the growth characteristics on various mixtures of glucose, xylose, and arabinose were investigated, which demonstrated the efficient co‐utilization of hexose and pentose sugars. Finally, the possibility of using lignocellulose hydrolysate as substrate for the two recombinant strains was verified. The recombinant P. putida KT2440 strains presented here as flexible microbial biocatalysts to convert lignocellulosic sugars will undoubtedly contribute to the economic feasibility of the production of valuable compounds derived from renewable feedstock.     

A novel strategy for production of ethanol and recovery of xylose from simulated corncob hydrolysate

Objectives

To develop a xylose-nonutilizing Escherichia coli strain for ethanol production and xylose recovery.

Results

Xylose-nonutilizing E. coli CICIM B0013-2012 was successfully constructed from E. coli B0013-1030 (pta-ack, ldhA, pflB, xylH) by deletion of frdA, xylA and xylE. It exhibited robust growth on plates containing glucose, arabinose or galactose, but failed to grow on xylose. The ethanol synthesis pathway was then introduced into B0013-2012 to create an ethanologenic strain B0013-2012PA. In shaking flask fermentation, B0013-2012PA fermented glucose to ethanol with the yield of 48.4 g/100 g sugar while xylose remained in the broth. In a 7-l bioreactor, B0013-2012PA fermented glucose, galactose and arabinose in the simulated corncob hydrolysate to 53.4 g/l ethanol with the yield of 48.9 g/100 g sugars and left 69.6 g/l xylose in the broth, representing 98.6% of the total xylose in the simulated corncob hydrolysate.

Conclusions

By using newly constructed strain B0013-2012PA, we successfully developed an efficient bioprocess for ethanol production and xylose recovery from the simulated corncob hydrolysate.

     

Improved <Emphasis Type="Italic">p</Emphasis>-hydroxybenzoate production by engineered <Emphasis Type="Italic">Pseudomonas putida</Emphasis> S12 by using a mixed-substrate feeding strategy

The key precursors for p-hydroxybenzoate production by engineered Pseudomonas putida S12 are phosphoenolpyruvate (PEP) and erythrose-4-phosphate (E4P), for which the pentose phosphate (PP) pathway is an important source. Since PP pathway fluxes are typically low in pseudomonads, E4P and PEP availability is a likely bottleneck for aromatics production which may be alleviated by stimulating PP pathway fluxes via co-feeding of pentoses in addition to glucose or glycerol. As P. putida S12 lacks the natural ability to utilize xylose, the xylose isomerase pathway from E. coli was introduced into the p-hydroxybenzoate producing strain P. putida S12palB2. The initially inefficient xylose utilization was improved by evolutionary selection after which the p-hydroxybenzoate production was evaluated. Even without xylose-co-feeding, p-hydroxybenzoate production was improved in the evolved xylose-utilizing strain, which may indicate an intrinsically elevated PP pathway activity. Xylose co-feeding further improved the p-hydroxybenzoate yield when co-fed with either glucose or glycerol, up to 16.3 Cmol% (0.1 g p-hydroxybenzoate/g substrate). The yield improvements were most pronounced with glycerol, which probably related to the availability of the PEP precursor glyceraldehyde-3-phosphate (GAP). Thus, it was demonstrated that the production of aromatics such as p-hydroxybenzoate can be improved by co-feeding different carbon sources via different and partially artificial pathways. Moreover, this approach opens new perspectives for the efficient production of (fine) chemicals from renewable feedstocks such as lignocellulose that typically has a high content of both glucose and xylose and (crude) glycerol.     

Engineered Escherichia coli capable of co-utilization of cellobiose and xylose

Natural ability to ferment the major sugars (glucose and xylose) of plant biomass is an advantageous feature of Escherichia coli in biofuel production. However, excess glucose completely inhibits xylose utilization in E. coli and decreases yield and productivity of fermentation due to sequential utilization of xylose after glucose. As an approach to overcome this drawback, E. coli MG1655 was engineered for simultaneous glucose (in the form of cellobiose) and xylose utilization by a combination of genetic and evolutionary engineering strategies. The recombinant E. coli was capable of utilizing approximately 6 g/L of cellobiose and 2 g/L of xylose in approximately 36 h, whereas wild-type E. coli was unable to utilize xylose completely in the presence of 6 g/L of glucose even after 75 hours. The engineered strain also co-utilized cellobiose with mannose or galactose; however, it was unable to metabolize cellobiose in the presence of arabinose and glucose. Successful cellobiose and xylose co-fermentation is a vital step for simultaneous saccharification and co-fermentation process and a promising step towards consolidated bioprocessing.     

Fermentation of sugar cane bagasse hemicellulosic hydrolysate and sugar mixtures to ethanol by recombinant Escherichia coli KO11

Escherichia coli KO11, carrying the ethanol pathway genes pdc (pyruvate decarboxylase) and adh (alcohol dehydrogenase) from Zymomonas mobilis integrated into its chromosome, has the ability to metabolize pentoses and hexoses to ethanol, both in synthetic medium and in hemicellulosic hydrolysates. In the fermentation of sugar mixtures simulating hemicellulose hydrolysate sugar composition (10.0 g of glucose/l and 40.0 g of xylose/l) and supplemented with tryptone and yeast extract, recombinant bacteria produced 24.58 g of ethanol/l, equivalent to 96.4% of the maximum theoretical yield. Corn steep powder (CSP), a byproduct of the corn starch-processing industry, was used to replace tryptone and yeast extract. At a concentration of 12.5 g/l, it was able to support the fermentation of glucose (80.0 g/l) to ethanol, with both ethanol yield and volumetric productivity comparable to those obtained with fermentation media containing tryptone and yeast extract. Hemicellulose hydrolysate of sugar cane bagasse supplemented with tryptone and yeast extract was also readily fermented to ethanol within 48 h, and ethanol yield achieved 91.5% of the theoretical maximum conversion efficiency. However, fermentation of bagasse hydrolysate supplemented with 12.5 g of CSP/l took twice as long to complete. This revised version was published online in November 2006 with corrections to the Cover Date.     

Comparing the xylose reductase/xylitol dehydrogenase and xylose isomerase pathways in arabinose and xylose fermenting Saccharomyces cerevisiae strains

Background

Ethanolic fermentation of lignocellulosic biomass is a sustainable option for the production of bioethanol. This process would greatly benefit from recombinant Saccharomyces cerevisiae strains also able to ferment, besides the hexose sugar fraction, the pentose sugars, arabinose and xylose. Different pathways can be introduced in S. cerevisiae to provide arabinose and xylose utilisation. In this study, the bacterial arabinose isomerase pathway was combined with two different xylose utilisation pathways: the xylose reductase/xylitol dehydrogenase and xylose isomerase pathways, respectively, in genetically identical strains. The strains were compared with respect to aerobic growth in arabinose and xylose batch culture and in anaerobic batch fermentation of a mixture of glucose, arabinose and xylose.

Results

The specific aerobic arabinose growth rate was identical, 0.03 h^-1, for the xylose reductase/xylitol dehydrogenase and xylose isomerase strain. The xylose reductase/xylitol dehydrogenase strain displayed higher aerobic growth rate on xylose, 0.14 h^-1, and higher specific xylose consumption rate in anaerobic batch fermentation, 0.09 g (g cells)^-1 h^-1 than the xylose isomerase strain, which only reached 0.03 h^-1 and 0.02 g (g cells)^-1h^-1, respectively. Whereas the xylose reductase/xylitol dehydrogenase strain produced higher ethanol yield on total sugars, 0.23 g g^-1 compared with 0.18 g g^-1 for the xylose isomerase strain, the xylose isomerase strain achieved higher ethanol yield on consumed sugars, 0.41 g g^-1 compared with 0.32 g g^-1 for the xylose reductase/xylitol dehydrogenase strain. Anaerobic fermentation of a mixture of glucose, arabinose and xylose resulted in higher final ethanol concentration, 14.7 g l^-1 for the xylose reductase/xylitol dehydrogenase strain compared with 11.8 g l^-1 for the xylose isomerase strain, and in higher specific ethanol productivity, 0.024 g (g cells)^-1 h^-1 compared with 0.01 g (g cells)^-1 h^-1 for the xylose reductase/xylitol dehydrogenase strain and the xylose isomerase strain, respectively.

Conclusion

The combination of the xylose reductase/xylitol dehydrogenase pathway and the bacterial arabinose isomerase pathway resulted in both higher pentose sugar uptake and higher overall ethanol production than the combination of the xylose isomerase pathway and the bacterial arabinose isomerase pathway. Moreover, the flux through the bacterial arabinose pathway did not increase when combined with the xylose isomerase pathway. This suggests that the low activity of the bacterial arabinose pathway cannot be ascribed to arabitol formation via the xylose reductase enzyme.     

Regulation of Arabinose and Xylose Metabolism in Escherichia coli

Bacteria such as Escherichia coli will often consume one sugar at a time when fed multiple sugars, in a process known as carbon catabolite repression. The classic example involves glucose and lactose, where E. coli will first consume glucose, and only when it has consumed all of the glucose will it begin to consume lactose. In addition to that of lactose, glucose also represses the consumption of many other sugars, including arabinose and xylose. In this work, we characterized a second hierarchy in E. coli, that between arabinose and xylose. We show that, when grown in a mixture of the two pentoses, E. coli will consume arabinose before it consumes xylose. Consistent with a mechanism involving catabolite repression, the expression of the xylose metabolic genes is repressed in the presence of arabinose. We found that this repression is AraC dependent and involves a mechanism where arabinose-bound AraC binds to the xylose promoters and represses gene expression. Collectively, these results demonstrate that sugar utilization in E. coli involves multiple layers of regulation, where cells will consume first glucose, then arabinose, and finally xylose. These results may be pertinent in the metabolic engineering of E. coli strains capable of producing chemical and biofuels from mixtures of hexose and pentose sugars derived from plant biomass.The transporters and enzymes in many sugar metabolic pathways are conditionally expressed in response to their cognate sugar or a downstream pathway intermediate. While the induction of these pathways in response to a single sugar has been studied extensively (28), far less is known about how these pathways are induced in response to multiple sugars. One notable exception is the phenomenon observed when bacteria are grown in the presence of glucose and another sugar (10, 15). In such mixtures, the bacteria will often consume glucose first before consuming the other sugar, a process known as carbon catabolite repression (27). The classic example of carbon catabolite repression is the diauxic shift seen in the growth of Escherichia coli on mixtures of glucose and lactose, where the cells first consume glucose before consuming lactose. When the cells are consuming glucose, the genes in the lactose metabolic pathway are not induced, thus preventing the sugar from being consumed. A number of molecules participate in this regulation, including the cyclic AMP receptor protein (CRP), adenylate cyclase, cyclic AMP (cAMP), and EIIA from the phosphoenolpyruvate:glucose phosphotransferase system (PTS) (33). In addition to lactose, the metabolic genes for many other sugars are subject to catabolite repression by glucose in E. coli (27). While the preferential utilization of glucose is well known, it is an open question whether additional hierarchies exist among other sugars.Recently, substantial effort has been directed toward developing microorganisms capable of producing chemicals and biofuels from plant biomass (1, 34, 42). After glucose, l-arabinose and d-xylose are the next most abundant sugars found in plant biomass. Therefore, a key step in producing various chemicals and fuels from plant biomass will be the engineering of strains capable of efficiently fermenting these three sugars. However, one challenge concerns catabolite repression, which prevents microorganisms from fermenting these three sugars simultaneously and, as a consequence, may decrease the efficiency of the fermentation process. E. coli cells will first consume glucose before consuming either arabinose or xylose. As in the case of lactose, the genes in the arabinose and xylose metabolic pathways are not expressed when glucose is being consumed. In addition to glucose catabolite repression, a second hierarchy, between arabinose and xylose, appears to exist. Kang and coworkers have observed that the genes in the xylose metabolic pathway were repressed when cells were grown in a mixture of arabinose and xylose (21). Hernandez-Montalvo and coworkers also observed that E. coli utilizes arabinose before xylose (19). While a number of strategies exist for breaking the glucose-mediated repression of arabinose and xylose metabolism (8, 16, 19, 31), none exist for breaking the arabinose-mediated repression of xylose metabolism. Moreover, little is known about this repression beyond the observations made by these researchers.In this work, we investigate how the arabinose and xylose metabolic pathways are jointly regulated. We demonstrate that E. coli will consume arabinose before consuming xylose when it is grown in a mixture of the two sugars. Consistent with a mechanism involving catabolite repression, the genes in the xylose metabolic pathway are repressed in the presence of arabinose. We found that this repression is AraC dependent and is most likely due to binding by arabinose-bound AraC to the xylose promoters, with consequent inhibition of gene expression.     

Lactic acid production from hemicellulosic hydrolyzate by cells of <Emphasis Type="Italic">Lactobacillus bifermentans</Emphasis> immobilized in Ca-alginate using response surface methodology

Consumption of hexoses/pentoses and production of lactic acid by Lactobacillus bifermentans were investigated in optimized culture medium and hemicellulosic hydrolyzates. The hydrolyzate used had the following composition (expressed in gL⁻¹): xylose 50 ± 5 gL⁻¹; glucose 18 ± 3 gL⁻¹; arabinose 29 ± 5 gL⁻¹. The immobilization experiments were conducted with microbial cells entrapped in calcium alginate beads. The results indicate that maximum concentrations of lactic acid were produced after 54 h of fermentation. All glucose and arabinose in wheat bran hydrolyzate were consumed during fermentation. Only xylose was not completely consumed. The substrate consumption rate was 3.2 gh⁻¹, 1.9 gh⁻¹, 1.6 gh⁻¹ respectively for glucose, arabinose, and xylose. The optimized culture condition gave a lactic acid concentration and metabolic yield of 62.77 gL⁻¹ and 0.83 gg⁻¹. These parameters improved to 41.3 gL⁻¹ and 0.47 gg⁻¹ respectively, when cell free was used.     

Fermentation performance of Candida guilliermondii for xylitol production on single and mixed substrate media

Semidefined media fermentation simulating the sugar composition of hemicellulosic hydrolysates (around 85 g l^-1 xylose, 17 g l^-1 glucose, and 9 g l^-1 arabinose) was investigated to evaluate the glucose and arabinose influence on xylose-to-xylitol bioconversion by Candida guilliermondii. The results revealed that glucose reduced the xylose consumption rate by 30%. Arabinose did not affect the xylose consumption but its utilization by the yeast was fully repressed by both glucose and xylose sugars. Arabinose was only consumed when it was used as a single carbon source. Xylitol production was best when glucose was not present in the fermentation medium. On the other hand, the arabinose favored the xylitol yield (which attained 0.74 g g^-1 xylose consumed) and it did not interfere with xylitol volumetric productivity (Q _P=0.85 g g^-1), the value of which was similar to that obtained with xylose alone.     

A co-fermentation strategy to consume sugar mixtures effectively

We report a new approach for the simultaneous conversion of xylose and glucose sugar mixtures into products by fermentation. The process simultaneously uses two substrate-selective strains of Escherichia coli, one which is unable to consume glucose and one which is unable to consume xylose. The xylose-selective (glucose deficient) strain E. coli ZSC113 has mutations in the glk, ptsG and manZ genes while the glucose-selective (xylose deficient) strain E. coli ALS1008 has a mutation in the xylA gene. By combining these two strains in a single process, xylose and glucose are consumed more quickly than by a single-organism approach. Moreover, we demonstrate that the process is able to adapt to changing concentrations of these two sugars, and therefore holds promise for the conversion of variable sugar feed streams, such as lignocellulosic hydrolysates.     

Production of arabitol from enzymatic hydrolysate of soybean flour by Debaryomyces hansenii fermentation

Arabitol is a low-calorie sugar alcohol with anti-cariogenic properties. Enzymatic hydrolysate of soybean flour is a new renewable biorefinery feedstock containing hexose, pentose, and organic nitrogen sources. Arabitol production by Debaryomyces hansenii using soybean flour hydrolysate was investigated. Effects of medium composition, operating conditions, and culture stage (growing or stationary phase) were studied. Production was also compared at different culture volumes to understand the effect of dissolved oxygen concentration (DO). Main factors examined for medium composition effects were the carbon to nitrogen concentration ratio (C/N), inorganic (ammonium) to organic nitrogen ratio (I/O-N), and sugar composition. Arabitol yield increased with increasing C/N ratio and a high I/O-N (0.8–1.0), suggesting higher yield at stationary phase of low pH (3.5–4.5). Catabolite repression was observed, with the following order of consumption: glucose > fructose > galactose > xylose > arabinose. Arabitol production also favored hexoses and, among hexoses, glucose. DO condition was of critical importance to arabitol production and cell metabolism. The yeast consumed pentoses (xylose and arabinose) only at more favorable DO conditions. Finally, arabitol was produced in fermentors using mixed hydrolysates of soy flour and hulls. The process gave an arabitol yield of 54%, volumetric productivity of 0.90 g/L-h, and specific productivity of 0.031 g/g-h.