Seed train development for the fermentation of bagasse from sweet sorghum and sugarcane using a simplified fermentation process.

A process was developed for seed culture expansion (3.6 million-fold) using 5% of the hemicellulose hydrolysate from dilute acid pretreatment as the sole organic nutrient and source of sugar. Hydrolysate used for seed growth was neutralized with ammonia and combined with 1.0mM sodium metabisulfite immediately before inoculation. This seed protocol was tested with phosphoric acid pretreated sugarcane and sweet sorghum bagasse using a simplified process with co-fermentation of fiber, pentoses, and hexoses in a single vessel (SScF). A 6h liquefaction (L) step improved mixing prior to inoculation. Fermentations (L+SScF process) were completed in 72 h with high yields (>80 gal/US ton). Ethanol titers for this L+SScF process ranged from 24 g/L to 32 g/L, and were limited by the bagasse concentration (10% dry matter).

Simplified process for ethanol production from sugarcane bagasse using hydrolysate-resistant Escherichia coli strain MM160.

Hexose and pentose sugars from phosphoric acid pretreated sugarcane bagasse were co-fermented to ethanol in a single vessel (SScF), eliminating process steps for solid-liquid separation and sugar cleanup. An initial liquefaction step (L) with cellulase was included to improve mixing and saccharification (L+SScF), analogous to a corn ethanol process. Fermentation was enabled by the development of a hydrolysate-resistant mutant of Escherichia coli LY180, designated MM160. Strain MM160 was more resistant than the parent to inhibitors (furfural, 5-hydroxymethylfurfural, and acetate) formed during pretreatment. Bagasse slurries containing 10% and 14% dry weight (fiber plus solubles) were tested using pretreatment temperatures of 160-190°C (1% phosphoric acid, 10 min). Enzymatic saccharification and inhibitor production both increased with pretreatment temperature. The highest titer (30 g/L ethanol) and yield (0.21 g ethanol/g bagasse dry weight) were obtained after incubation for 122 h using 14% dry weight slurries of pretreated bagasse (180°C).

Deletion of methylglyoxal synthase gene (mgsA) increased sugar co-metabolism in ethanol-producing Escherichia coli.

The use of lignocellulose as a source of sugars for bioproducts requires the development of biocatalysts that maximize product yields by fermenting mixtures of hexose and pentose sugars to completion. In this study, we implicate mgsA encoding methylglyoxal synthase (and methylglyoxal) in the modulation of sugar metabolism. Deletion of this gene (strain LY168) resulted in the co-metabolism of glucose and xylose, and accelerated the metabolism of a 5-sugar mixture (mannose, glucose, arabinose, xylose and galactose) to ethanol.

Silencing of NADPH-dependent oxidoreductase genes (yqhD and dkgA) in furfural-resistant ethanologenic Escherichia coli.

Low concentrations of furfural are formed as a side product during the dilute acid hydrolysis of hemicellulose. Growth is inhibited by exposure to furfural but resumes after the complete reduction of furfural to the less toxic furfuryl alcohol. Growth-based selection was used to isolate a furfural-resistant mutant of ethanologenic Escherichia coli LY180, designated strain EMFR9. Based on mRNA expression levels in the parent and mutant in response to furfural challenge, genes encoding 12 oxidoreductases were found to vary by more than twofold (eight were higher in EMFR9; four were higher in the parent). All 12 genes were cloned. When expressed from plasmids, none of the eight genes in the first group increased furfural tolerance in the parent (LY180). Expression of three of the silenced genes (yqhD, dkgA, and yqfA) in EMFR9 was found to decrease furfural tolerance compared to that in the parent. Purified enzymes encoded by yqhD and dkgA were shown to have NADPH-dependent furfural reductase activity. Both exhibited low K(m) values for NADPH (8 microM and 23 microM, respectively), similar to those of biosynthetic reactions. Furfural reductase activity was not associated with yqfA. Deleting yqhD and dkgA in the parent (LY180) increased furfural tolerance, but not to the same extent observed in the mutant EMFR9. Together, these results suggest that the process of reducing furfural by using an enzyme with a low K(m) for NADPH rather than a direct inhibitory action is the primary cause for growth inhibition by low concentrations of furfural.

Re-engineering Escherichia coli for ethanol production.

A lactate producing derivative of Escherichia coli KO11, strain SZ110, was re-engineered for ethanol production by deleting genes encoding all fermentative routes for NADH and randomly inserting a promoterless mini-Tn5 cassette (transpososome) containing the complete Zymomonas mobilis ethanol pathway (pdc, adhA, and adhB) into the chromosome. By selecting for fermentative growth in mineral salts medium containing xylose, a highly productive strain was isolated in which the ethanol cassette had been integrated behind the rrlE promoter, designated strain LY160(KO11, Deltafrd::celY(Ec) DeltaadhE DeltaldhA, DeltaackA lacA::casAB(Ko) rrlE::(pdc( Zm)-adhA(Zm)-adhB(Zm)-FRT-rrlE)pflB(+)). This strain fermented 9% (w/v) xylose to 4% (w/v) ethanol in 48 h in mineral salts medium, nearly equal to the performance of KO11 with Luria broth.

Low salt medium for lactate and ethanol production by recombinant Escherichia coli B.

Individual nutrient salts were experimentally varied to determine the minimum requirements for efficient L (+)-lactate production by recombinant strains of Escherichia coli B. Based on these results, AM1 medium was formulated with low levels of alkali metals (4.5 mM and total salts (4.2 g l(-1)). This medium was equally effective for ethanol production from xylose and lactate production from glucose with average productivities of 18-19 mmol l(-1) h(-1) for both (initial 48 h of fermentation).

Development of industrial-medium-required elimination of the 2,3-butanediol fermentation pathway to maintain ethanol yield in an ethanologenic strain of Klebsiella oxytoca.

Fermentation efficiency and nutrient costs are both significant factors in process economics for the microbial conversion of cellulosic biomass to commodity chemicals such as ethanol. In this study, we have developed a more industrial medium (OUM1) composed of 0.5% corn steep liquor (dry weight basis) supplemented with mineral salts (0.2%), urea (0.06%), and glucose (9%). Although the growth of strain P2 was vigorous in this medium, approximately 14% of substrate carbon was diverted into 2,3-butanediol and acetoin under the low pH conditions needed for optimal cellulase activity during simultaneous saccharification. Deleting the central region of the budAB genes encoding alpha-acetolactate synthase and alpha-acetolactate decarboxylase eliminated the butanediol and acetoin coproducts and increased ethanol yields by 12%. In OUM1 medium at pH 5.2, strain BW21 produced over 4% ethanol in 48 h (0.47 g ethanol per g glucose). Average productivity (48 h), ethanol titer, and ethanol yield for BW21 in OUM1 medium (pH 5.2) exceeded that of the parent (strain P2) in rich laboratory medium (Luria broth).

Gene array-based identification of changes that contribute to ethanol tolerance in ethanologenic Escherichia coli: comparison of KO11 (parent) to LY01 (resistant mutant).

Escherichia coli KO11 (parent) and LY01 (mutant) have been engineered for the production of ethanol. Gene arrays were used to identify expression changes that occurred in the mutant, LY01, during directed evolution to improve ethanol tolerance (defined as extent of growth in the presence of added ethanol). Expression levels for 205 (5%) of the ORFs were found to differ significantly (p < 0.10) between KO11 and LY01 under each of six different growth conditions (p < 0.000001). Statistical evaluation of differentially expressed genes according to various classification schemes identified physiological areas of importance. A large fraction of differentially expressed ORFs were globally regulated, leading to the discovery of a nonfunctional fnr gene in strain LY01. In agreement with a putative role for FNR in alcohol tolerance, increasing the copy number of fnr(+) in KO11(pGS196) decreased ethanol tolerance but had no effect on growth in the absence of ethanol. Other differences in gene expression provided additional clues that permitted experimentation. Tolerance appears to involve increased metabolism of glycine (higher expression of gcv genes) and increased production of betaine (higher expression of betIBA and betT encoding betaine synthesis from choline and choline uptake, respectively). Addition of glycine (10 mM) increased ethanol tolerance in KO11 but had no effect in the absence of ethanol. Addition of betaine (10 mM) increased ethanol tolerance by over 2-fold in both LY01 and KO11 but had no effect on growth in the absence of ethanol. Both glycine and betaine can serve as protective osmolytes, and this may be the basis of their beneficial action. In addition, the marAB genes encoding multiple antibiotic resistance proteins were expressed at higher levels in LY01 as compared to KO11. Interestingly, overexpression of marAB in KO11 made this strain more ethanol-sensitive. Overexpression of marAB in LY01 had no effect on ethanol tolerance. Increased expression of genes encoding serine uptake (sdaC) and serine deamination (sdaB) also appear beneficial for LY01. Addition of serine increased the growth of LY01 in the presence and absence of ethanol but had no effect on KO11. Changes in the expression of several genes concerned with the synthesis of the cell envelope components were also noted, which may contribute to increased ethanol tolerance.

Decreasing the level of ethyl acetate in ethanolic fermentation broths of Escherichia coli KO11 by expression of Pseudomonas putida estZ esterase.

During the fermentation of sugars to ethanol relatively high levels of an undesirable coproduct, ethyl acetate, are also produced. With ethanologenic Escherichia coli strain KO11 as the biocatalyst, the level of ethyl acetate in beer containing 4.8% ethanol was 192 mg liter(-1). Although the E. coli genome encodes several proteins with esterase activity, neither wild-type strains nor KO11 contained significant ethyl acetate esterase activity. A simple method was developed to rapidly screen bacterial colonies for the presence of esterases which hydrolyze ethyl acetate based on pH change. This method allowed identification of Pseudomonas putida NRRL B-18435 as a source of this activity and the cloning of a new esterase gene, estZ. Recombinant EstZ esterase was purified to near homogeneity and characterized. It belongs to family IV of lipolytic enzymes and contains the conserved catalytic triad of serine, aspartic acid, and histidine. As expected, this serine esterase was inhibited by phenylmethylsulfonyl fluoride and the histidine reagent diethylpyrocarbonate. The native and subunit molecular weights of the recombinant protein were 36,000, indicating that the enzyme exists as a monomer. By using alpha-naphthyl acetate as a model substrate, optimal activity was observed at pH 7.5 and 40 degrees C. The Km and Vmax for alpha-naphthyl acetate were 18 microM and 48.1 micromol. min(-1). mg of protein(-1), respectively. Among the aliphatic esters tested, the highest activity was obtained with propyl acetate (96 micromol. min(-1). mg of protein(-1)), followed by ethyl acetate (66 micromol. min(-1). mg of protein(-1)). Expression of estZ in E. coli KO11 reduced the concentration of ethyl acetate in fermentation broth (4.8% ethanol) to less than 20 mg liter(-1).

Global gene expression differences associated with changes in glycolytic flux and growth rate in Escherichia coli during the fermentation of glucose and xylose.

The simplicity of the fermentation process (anaerobic with pH, temperature, and agitation control) in ethanologenic Escherichia coli KO11 and LY01 makes this an attractive system to investigate the utility of gene arrays for biotechnology applications. By using this system, gene expression, glycolytic flux, and growth rate have been compared in glucose-grown and xylose-grown cells. Although the initial metabolic steps differ, ethanol yields from both sugars were essentially identical on a weight basis, and little carbon was diverted to biosynthesis. Expression of only 27 genes changed by more than 2-fold in both strains. These included induction of xylose-specific operons (xylE, xylFGHR, and xylAB) regulated by XylR and the cyclic AMP-CRP system and repression of Mlc-regulated genes encoding glucose uptake (ptsHIcrr, ptsG) and mannose uptake (manXYZ) during growth on xylose. However, expression of genes encoding central carbon metabolism and biosynthesis differed by less than 2-fold. Simple statistical methods were used to investigate these more subtle changes. The reproducibility (coefficient of variation of 12%) of expression measurements (mRNA as cDNA) was found to be similar to that typically observed for in vitro measurements of enzyme activities. Using Student's t test, many smaller but significant sugar-dependent changes were identified (p < 0.05 in both strains). A total of 276 genes were more highly expressed during growth on xylose; 307 genes were more highly expressed with glucose. Slower growth (lower ATP yield) on xylose was accompanied by decreased expression of 62 genes concerned with the biosynthesis of small molecules (amino acids, nucleotides, cofactors, and lipids), transcription, and translation; 5 such genes were expressed at a higher level. In xylose-grown cells, 90 genes associated with the transport, catabolism, and regulation of pathways for alternative carbon sources were expressed at higher levels than in glucose-grown cells, consistent with a relaxation of control by the cyclic AMP-CRP regulatory system. Changes in expression of genes encoding the Embden-Meyerhof-Parnas (EMP) pathway were in excellent agreement with calculated changes in flux for individual metabolites. Flux through all but one step, pyruvate kinase, was predicted to be higher during glucose fermentation. Expression levels (glucose/xylose) were higher in glucose-grown cells for all EMP genes except the isoenzymes encoding pyruvate kinase (pykA and pykF). Expression of both isoenzymes was generally higher during xylose fermentation but statistically higher in both strains only for pykF encoding the isoenzyme activated by fructose-6-phosphate, a key metabolite connecting pentose metabolism to the EMP pathway. The coordinated changes in expression of genes encoding the EMP pathway suggest the presence of a common regulatory system and that flux control within the EMP pathway may be broadly distributed. In contrast, expression levels for genes encoding the Pentose-Phosphate pathway did not differ significantly between glucose-grown and xylose-grown cells.

Detoxification of dilute acid hydrolysates of lignocellulose with lime.

The hydrolysis of hemicellulose to monomeric sugars by dilute acid hydrolysis is accompanied by the production of inhibitors that retard microbial fermentation. Treatment of hot hydrolysate with Ca(OH)(2) (overliming) is an effective method for detoxification. Using ethanologenic Escherichia coli LY01 as the biocatalyst, our results indicate that the optimal lime addition for detoxification varies and depends on the concentration of mineral acids and organic acids in each hydrolysate. This optimum was shown to be readily predicted on the basis of the titration of hydrolysate with 2 N NaOH at ambient temperature to either pH 7.0 or pH 11.0. The average composition of 15 hydrolysates prior to treatment was as follows (per L): 95.24 +/- 7.29 g sugar, 5.3 +/- 2.99 g acetic acid, 1.305 +/- 0.288 g total furans (furfural and hydroxymethylfurfural), and 2.86 +/- 0.34 g phenolic compounds. Optimal overliming resulted in a 51 +/- 9% reduction of total furans, a 41 +/- 6% reduction in phenolic compounds, and a 8.7 +/- 4.5% decline in sugar. Acetic acid levels were unchanged. Considering the similarity of microorganisms, it is possible that the titration method described here may also prove useful for detoxification and fermentation processes using other microbial biocatalysts.

Use of UV absorbance To monitor furans in dilute acid hydrolysates of biomass.

A simple method based on UV spectra was developed for the estimation of total furans (furfural and hydroxymethylfurfural) in hemicellulose hydrolysates. UV spectra of hemicellulose hydrolysate contained a single dominant peak at around 278 nm. Approximately two-thirds of this peak can be attributed to furan absorbance (furfural and hydroxymethylfurfural). At 284 nm, both furfural and hydroxymethylfurfural have equal absorbance on a weight basis. A comparison of HPLC determinations for different samples of hydrolysate was used to develop a simple equation that allows the accurate prediction of total furans based on the difference in absorbance at 284 and 320 nm. This method may prove useful for quality control applications during the production of biomass syrups using a dilute acid hydrolysis process and during treatments for the amelioration of toxins. Although furans represent only a portion of the toxins present in hemicellulose hydrolysates, the abundance of furans appears to serve as a useful marker to predict relative toxicity.

Effects of Ca(OH)(2) treatments ("overliming") on the composition and toxicity of bagasse hemicellulose hydrolysates.

Hemicellulose syrups from dilute sulfuric acid hydrolysates of hemicellulose contain inhibitors that prevent efficient fermentation by yeast or bacteria. It is well known that the toxicity of these hydrolysate syrups can be ameliorated by optimized "overliming" with Ca(OH)(2). We have investigated the optimization of overliming treatments for sugar cane bagasse hydrolysates (primarily pentose sugars) using recombinant Escherichia coli LY01 as the biocatalyst. A comparison of composition before and after optimal overliming revealed a substantial reduction in furfural, hydroxymethylfurfural, and three unidentified high-performance liquid chromatography (HPLC) peaks. Organic acids (acetic, formic, levulinic) were not affected. Similar changes have been reported after overliming of spruce hemicellulose hydrolysates (Larsson et al., 1999). Our studies further demonstrated that the extent of furan reduction correlated with increasing fermentability. However, furan reduction was not the sole cause for reduced toxicity. After optimal overliming, bagasse hydrolysate was rapidly and efficiently fermented (>90% yield) by LY01. During these studies, titration, and conductivity were found to be in excellent agreement as methods to estimate sulfuric acid content. Titration was also found to provide an estimate of total organic acids in hydrolysate, which agreed well with the sum of acetic, levulinic, and formic acids obtained by HPLC. Titration of acids, measurement of pH before and after treatment, and furan analyses are proposed as relatively simple methods to monitor the reproducibility of hydrolysate preparations and the effectiveness of overliming treatments.

Enteric bacterial catalysts for fuel ethanol production.

The technology is available to produce fuel ethanol from renewable lignocellulosic biomass. The current challenge is to assemble the various process options into a commercial venture and begin the task of incremental improvement. Current process designs for lignocellulose are far more complex than grain to ethanol processes. This complexity results in part from the complexity of the substrate and the biological limitations of the catalyst. Our work at the University of Florida has focused primarily on the genetic engineering of Enteric bacteria using genes encoding Zymomonas mobilis pyruvate decarboxylase and alcohol dehydrogenase. These two genes have been assembled into a portable ethanol production cassette, the PET operon, and integrated into the chromosome of Escherichia coli B for use with hemicellulose-derived syrups. The resulting strain, KO11, produces ethanol efficiently from all hexose and pentose sugars present in the polymers of hemicellulose. By using the same approach, we integrated the PET operon into the chromosome of Klebsiella oxytoca to produce strain P2 for use in the simultaneous saccharification and fermentation (SSF) process for cellulose. Strain P2 has the native ability to ferment cellobiose and cellotriose, eliminating the need for one class of cellulase enzymes. Recently, the ability to produce and secrete high levels of endoglucanase has also been added to strain P2, further reducing the requirement for fungal cellulase. The general approach for the genetic engineering of new biocatalysts using the PET operon has been most successful with Enteric bacteria but was also extended to Gram positive bacteria, which have other useful traits for lignocellulose conversion. Many opportunities remain for further improvements in these biocatalysts as we proceed toward the development of single organisms that can be used for the efficient fermentation of both hemicellulosic and cellulosic substrates.

Biosynthetic burden and plasmid burden limit expression of chromosomally integrated heterologous genes (pdc, adhB) in Escherichia coli.

Previous studies have shown an unexpectedly high nutrient requirement for efficient ethanol production by ethanologenic recombinants of Escherichia coli B such as LY01 which contain chromosomally integrated Zymomonas mobilis genes (pdc,adhB) encoding the ethanol pathway. The basis for this requirement has been identified as a media-dependent effect on the expression of the Z. mobilis genes rather than a nutritional limitation. Ethanol production was substantially increased without additional nutrients simply by increasing the level of pyruvate decarboxylase activity. This was accomplished by adding a multicopy plasmid containing pdc alone (but not adhB alone) to strain LY01, and by adding multicopy plasmids which express pdc and adhB from strong promoters. New strong promoters were isolated from random fragments of Z. mobilis DNA and characterized but were not used to construct integrated biocatalysts. These promoters contained regions resembling recognition sites for 3 different E. coli sigma factors: sigma(70), sigma(38), and sigma(28). The most effective plasmid-based promoters for fermentation were recognized by multiple sigma factors, expressed both pdc and adhB at high levels, and produced ethanol efficiently while allowing up to 80% reduction in complex nutrients as compared to LY01. The ability to utilize multiple sigma factors may be advantageous to maintain the high levels of PDC and ADH needed for efficient ethanol production throughout batch fermentation. From this work, we propose that the activation of biosynthetic genes in nutrient-poor media creates a biosynthetic burden that reduces the expression of chromosomal pdc and adhB by competing for transcriptional and translational machinery. This reduced expression can be viewed as analogous to the effect of plasmids (plasmid burden) on the expression of native chromosomal genes.

Isolation and characterization of ethanol-tolerant mutants of Escherichia coli KO11 for fuel ethanol production.

Genetically engineered Escherichia coli KO11 is capable of efficiently producing ethanol from all sugar constituents of lignocellulose but lacks the high ethanol tolerance of yeasts currently used for commercial starch-based ethanol processes. Using an enrichment method which selects alternatively for ethanol tolerance during growth in broth and for ethanol production on solid medium, mutants of KO11 with increased ethanol tolerance were isolated which can produce more than 60 g ethanol L-1 from xylose in 72 h. Ethanol concentrations and yields achieved by the LY01 mutant with xylose exceed those reported for recombinant strains of Saccharomyces and Zymomonas mobilis, both of which have a high native ethanol tolerance.

Isolation and molecular characterization of high-performance cellobiose-fermenting spontaneous mutants of ethanologenic Escherichia coli KO11 containing the Klebsiella oxytoca casAB operon.

Escherichia coli KO11 was previously constructed to produce ethanol from acid hydrolysates of hemicellulose (pentoses and hexoses) by the chromosomal integration of Zymomonas mobilis genes encoding pyruvate decarboxylase (pdc) and alcohol dehydrogenase (adhB). Klebsiella oxytoca P2 was constructed in an analogous fashion for the simultaneous saccharification and fermentation of cellulose and contains PTS enzymes for cellobiose. In this study, KO11 was further engineered for the fermentation of cellulose by adding the K. oxytoca casAB genes encoding Enzyme IIcellobiose and phospho-beta-glucosidase. Although the two K. oxytoca genes were well expressed in cloning hosts such as DH5 alpha, both were expressed poorly in E. coli KO11, a derivative of E. coli B. Spontaneous mutants which exhibited more than 15-fold-higher specific activities for cellobiose metabolism were isolated. The mutations of these mutants resided in the plasmid rather than the host. Three mutants were characterized by sequence analysis. All contained similar internal deletions which eliminated the casAB promoter and operator regions and placed the lacZ Shine-Dalgarno region immediately upstream from the casA Shine-Dalgarno region. KO11 harboring mutant plasmids (pLOI1908, pLOI1909, or pLOI1910) rapidly fermented cellobiose to ethanol, and the yield was more than 90% of the theoretical yield. Two of these strains were used with commercial cellulase to ferment mixed-waste office paper to ethanol.

Extracellular melibiose and fructose are intermediates in raffinose catabolism during fermentation to ethanol by engineered enteric bacteria.

Contrary to general concepts of bacterial saccharide metabolism, melibiose (25 to 32 g/liter) and fructose (5 to 14 g/liter) accumulated as extracellular intermediates during the catabolism of raffinose (O-alpha-D-galactopyranosyl-1, 6-alpha-D-glucopyranosyl-beta-D-fructofuranoside) (90 g/liter) by ethanologenic recombinants of Escherichia coli B, Klebsiella oxytoca M5A1, and Erwinia chrysanthemi EC16. Both hydrolysis products (melibiose and fructose) were subsequently transported and further metabolized by all three organisms. Raffinose catabolism was initiated by beta-fructosidase; melibiose was subsequently hydrolyzed to galactose and glucose by alpha-galactosidase. Glucose and fructose were completely metabolized by all three organisms, but galactose accumulated in the fermentation broth with EC16(pLOI555) and P2. MM2 (a raffinose-positive E. coli mutant) was the most effective biocatalyst for ethanol production (38 g/liter) from raffinose. All organisms rapidly fermented sucrose (90 g/liter) to ethanol (48 g/liter) at more than 90% of the theoretical yield. During sucrose catabolism, both hydrolysis products (glucose and fructose) were metabolized concurrently by EC16(pLOI555) and P2 without sugar leakage. However, fructose accumulated extracellularly (27 to 28 g/liter) at early stages of fermentation with KO11 and MM2. Sequential utilization of glucose and fructose correlated with a diauxie in base utilization (pH maintenance). The mechanism of sugar escape remains unknown but may involve downhill leakage via permease which transports precursor saccharides or novel sugar export proteins. If sugar escape occurs in nature with wild organisms, it could facilitate the development of complex bacterial communities which are based on the sequence of saccharide catabolism and the hierarchy of sugar utilization.

Soy-based medium for ethanol production by Escherichia coli KO11.

An optimized soy-based medium was developed for ethanol production by Escherichia coli KO11. The medium consists of mineral salts, vitamins, crude enzymatic hydrolysate of soy and fermentable sugar. Ethanol produced after 24 h was used as an endpoint in bioassays to optimize hydrolysate preparation. Although longer fermentation times were required with soy medium than with LB medium, similar final ethanol concentrations were achieved (44-45 g ethanol L-1 from 100 g glucose L-1). The cost of materials for soy medium (excluding sugar) was estimated to be $0.003 L-1 broth, $0.06 L-1 ethanol.