Improved xylanase production by<Emphasis Type="Italic"> Trichoderma reesei</Emphasis> grown on <Emphasis Type="SmallCaps">l</Emphasis>-arabinose and lactose or <Emphasis Type="SmallCaps">d</Emphasis>-glucose mixtures

Trichoderma reesei Rut C-30 was grown on eight different natural or rare aldopentoses as the main carbon source and on mixtures of an aldopentose with d-glucose or lactose. The fungal cells consumed all aldopentoses tested, except l-xylose and l-ribose. The highest total xylanase and cellulase activities were achieved when cells were grown on l-arabinose as the main carbon source. The total xylanase activity produced by cells grown on l-arabinose was even higher than that produced by cells grown on an equal amount of lactose. In co-metabolism of d-glucose (15 g l^–1) and l-arabinose (5 g l^–1), the total volumetric and specific xylanase productivities were improved (derepressed) approximately 23- and 18-fold, respectively, compared to a cultivation on only d-glucose (20 g l^–1). In a similar experiment, in which cells were grown on a mixture of lactose and l-arabinose, the xylanase productivity was approximately doubled, compared to a cultivation on only lactose. The cellulase productivities, however, were not improved by the addition of l-arabinose. Compared with a typical industrial fungal enzyme production process with lactose as the main carbon source, better volumetric and specific xylanase productivities were achieved both on a lactose/arabinose mixture and on a glucose/arabinose mixture.     

Purification and characterization of <Emphasis Type="SmallCaps">l</Emphasis>-arabinose isomerase from <Emphasis Type="Italic">Lactobacillus plantarum</Emphasis> producing <Emphasis Type="SmallCaps">d</Emphasis>-tagatose

l-arabinose isomerase (EC5.3.1.4. AI) mediates the isomerization of d-galactose into d-tagatose as well as the conversion of l-arabinose into l-ribulose. The AI from Lactobacillus plantarum SK-2 was purified to an apparent homogeneity giving a single band on SDS–PAGE with a molecular mass of 59.6 kDa. Optimum activity was observed at 50°C and pH 7.0. The enzyme was stable at 50°C for 2 h and held between pH 4.5 and 8.5 for 1 h. AI activity was stimulated by Mn²⁺, Fe³⁺, Fe²⁺, Ca²⁺ and inhibited by Cu²⁺, Ag⁺, Hg²⁺, Pb²⁺. d-galactose and l-arabinose as substrates were isomerized with high activity. l-arabitol was the strongest competitive inhibitor of AI. The apparent Michaelis–Menten constant (K _m), for galactose, was 119 mM. The first ten N-terminal amino acids of the enzyme were determined as MLSVPDYEFW, which is identical to L. plantarum (Q88S84). Using the purified AI, 390 mg tagatose could be converted from 1,000 mg galactose in 96 h, and this production corresponds to a 39% equilibrium.     

Biotechnological production of <Emphasis Type="SmallCaps">l</Emphasis>-ribose from <Emphasis Type="SmallCaps">l</Emphasis>-arabinose

l-Ribose is a rare and expensive sugar that can be used as a precursor for the production of l-nucleoside analogues, which are used as antiviral drugs. In this work, we describe a novel way of producing l-ribose from the readily available raw material l-arabinose. This was achieved by introducing l-ribose isomerase activity into l-ribulokinase-deficient Escherichia coli UP1110 and Lactobacillus plantarum BPT197 strains. The process for l-ribose production by resting cells was investigated. The initial l-ribose production rates at 39°C and pH 8 were 0.46 ± 0.01 g g⁻¹ h⁻¹ (1.84 ± 0.03 g l⁻¹ h⁻¹) and 0.27 ± 0.01 g g⁻¹ h⁻¹ (1.91 ± 0.1 g l⁻¹ h⁻¹) for E. coli and for L. plantarum, respectively. Conversions were around 20% at their highest in the experiments. Also partially purified protein precipitates having both l-arabinose isomerase and l-ribose isomerase activity were successfully used for converting l-arabinose to l-ribose.     

Fermentation of <Emphasis Type="SmallCaps">d</Emphasis>-glucose and <Emphasis Type="SmallCaps">d</Emphasis>-xylose mixtures by <Emphasis Type="Italic">Candida tropicalis</Emphasis> NBRC 0618 for xylitol production

The fermentation of d-glucose and d-xylose mixtures by the yeast Candida tropicalis NBRC 0618 has been studied under the most favourable operation conditions for the culture, determining the most adequate initial proportion in these sugars for xylitol production. In all the experiments a synthetic culture medium was used, with an initial total substrate concentration of 25 g L⁻¹, a constant pH of 5.0 and a temperature of 30 °C. From the experimental results, it was deduced that the highest values of specific rates of production and of overall yield in xylitol were achieved for the mixtures with the highest percentage of d-xylose, specifically in the culture with the initial d-glucose and d-xylose concentrations of 1 and 24 g L⁻¹, respectively, with an overall xylitol yield of 0.28 g g⁻¹. In addition, the specific rates of xylitol production declined over the time course of the culture and the formation of this bioproduct was favoured by the presence of small quantities of d-glucose. The sum of the overall yield values in xylitol and ethanol for all the experiments ranged from 0.26 to 0.56 g bioproduct/g total substrate.     

<Emphasis Type="SmallCaps">l</Emphasis>-Arabinose pathway engineering for arabitol-free xylitol production in <Emphasis Type="Italic">Candida tropicalis</Emphasis>

Xylose reductase (XR) is a key enzyme in biological xylitol production, and most XRs have broad substrate specificities. During xylitol production from biomass hydrolysate, non-specific XRs can reduce l-arabinose, which is the second-most abundant hemicellulosic sugar, to the undesirable byproduct arabitol, which interferes with xylitol crystallization in downstream processing. To minimize the flux from l-arabinose to arabitol, the l-arabinose-preferring, endogenous XR was replaced by a d-xylose-preferring heterologous XR in Candida tropicalis. Then, Bacillus licheniformis araA and Escherichia coli araB and araD were codon-optimized and expressed functionally in C. tropicalis for the efficient assimilation of l-arabinose. During xylitol fermentation, the control strains BSXDH-3 and KNV converted 9.9 g l-arabinose l⁻¹ into 9.5 and 8.3 g arabitol l⁻¹, respectively, whereas the recombinant strain JY consumed 10.5 g l-arabinose l⁻¹ for cell growth without forming arabitol. Moreover, JY produced xylitol with 42 and 16% higher productivity than BSXDH-3 and KNV, respectively.     

Xylose reductase activity in <Emphasis Type="Italic">Debaryomyces hansenii</Emphasis> UFV-170 cultivated in semi-synthetic medium and cotton husk hemicellulose hydrolyzate

To develop a new enzymatic xylose-to-xylitol conversion, deeper knowledge on the regulation of xylose reductase (XR) is needed. To this purpose, a new strain of Debaryomyces hansenii (UFV-170), which proved a promising xylitol producer, was cultivated in semi-synthetic media containing different carbon sources, specifically three aldo-hexoses (d-glucose, d-galactose and d-mannose), a keto-hexose (d-fructose), a keto-pentose (d-xylose), three aldo-pentoses (d-arabinose, l-arabinose and d-ribose), three disaccharides (maltose, lactose and sucrose) and a pentitol (xylitol). The best substrate was lactose on which cell concentration reached about 20 g l⁻¹ dry weight (DW), while the highest specific growth rates (0.58–0.61 h⁻¹) were detected on lactose, d-mannose, d-glucose and d-galactose. The highest specific activity of XR (0.24 U mg⁻¹) was obtained in raw extracts of cells grown on d-xylose and harvested in the stationary growth phase. When grown on cotton husk hemicellulose hydrolyzates, cells exhibited XR activities five to seven times higher than on semi-synthetic media.     

Production of <Emphasis Type="SmallCaps">d</Emphasis>-lactic acid by <Emphasis Type="Italic">Corynebacterium glutamicum</Emphasis> under oxygen deprivation

In mineral salts medium under oxygen deprivation, Corynebacterium glutamicum exhibits high productivity of l-lactic acid accompanied with succinic and acetic acids. In taking advantage of this elevated productivity, C. glutamicum was genetically modified to produce d-lactic acid. The modification involved expression of fermentative d-lactate dehydrogenase (d-LDH)-encoding genes from Escherichia coli and Lactobacillus delbrueckii in l-lactate dehydrogenase (l-LDH)-encoding ldhA-null C. glutamicum mutants to yield strains C. glutamicum ΔldhA/pCRB201 and C. glutamicum ΔldhA/pCRB204, respectively. The productivity of C. glutamicum ΔldhA/pCRB204 was fivefold higher than that of C. glutamicum ΔldhA/pCRB201. By using C. glutamicum ΔldhA/pCRB204 cells packed to a high density in mineral salts medium, up to 1,336 mM (120 g l⁻¹) of d-lactic acid of greater than 99.9% optical purity was produced within 30 h.     

Characterization of ribose-5-phosphate isomerase of <Emphasis Type="Italic">Clostridium thermocellum</Emphasis> producing <Emphasis Type="SmallCaps">d</Emphasis>-allose from <Emphasis Type="SmallCaps">d</Emphasis>-psicose

The rpiB gene, encoding ribose-5-phosphate isomerase (RpiB) from Clostridium thermocellum, was cloned and expressed in Escherichia coli. RpiB converted d-psicose into d-allose but it did not convert d-xylose, l-rhamnose, d-altrose or d-galactose. The production of d-allose by RpiB was maximal at pH 7.5 and 65°C for 30 min. The half-lives of the enzyme at 50°C and 65°C were 96 h and 4.7 h, respectively. Under stable conditions of pH 7.5 and 50°C, 165 g d-allose l⁻¹ was produced without by-products from 500 g d-psicose l⁻¹ after 6 h.     

The alternative<Emphasis Type="SmallCaps"> d</Emphasis>-galactose degrading pathway of<Emphasis Type="Italic"> Aspergillus nidulans</Emphasis> proceeds via<Emphasis Type="SmallCaps"> l</Emphasis>-sorbose

The catabolism of d-galactose in yeast depends on the enzymes of the Leloir pathway. In contrast, Aspergillus nidulans mutants in galactokinase (galE) can still grow on d-galactose in the presence of ammonium—but not nitrate—ions as nitrogen source. A. nidulans galE mutants transiently accumulate high (400 mM) intracellular concentrations of galactitol, indicating that the alternative d-galactose degrading pathway may proceed via this intermediate. The enzyme degrading galactitol was identified as l-arabitol dehydrogenase, because an A. nidulans loss-of-function mutant in this enzyme (araA1) did not show NAD⁺-dependent galactitol dehydrogenase activity, still accumulated galactitol but was unable to catabolize it thereafter, and a double galE/araA1 mutant was unable to grow on d-galactose or galactitol. The product of galactitol oxidation was identified as l-sorbose, which is a substrate for hexokinase, as evidenced by a loss of l-sorbose phosphorylating activity in an A. nidulans hexokinase (frA1) mutant. l-Sorbose catabolism involves a hexokinase step, indicated by the inability of the frA1 mutant to grow on galactitol or l-sorbose, and by the fact that a galE/frA1 double mutant of A. nidulans was unable to grow on d-galactose. The results therefore provide evidence for an alternative pathway of d-galactose catabolism in A. nidulans that involves reduction of the d-galactose to galactitol and NAD⁺-dependent oxidation of galactitol by l-arabitol dehydrogenase to l-sorbose.     

Characterization of <Emphasis Type="SmallCaps">d</Emphasis>-tagatose-3-epimerase from <Emphasis Type="Italic">Rhodobacter sphaeroides</Emphasis> that converts <Emphasis Type="SmallCaps">d</Emphasis>-fructose into <Emphasis Type="SmallCaps">d-</Emphasis>psicose

A non-characterized gene, previously proposed as the d-tagatose-3-epimerase gene from Rhodobacter sphaeroides, was cloned and expressed in Escherichia coli. Its molecular mass was estimated to be 64 kDa with two identical subunits. The enzyme specificity was highest with d-fructose and decreased for other substrates in the order: d-tagatose, d-psicose, d-ribulose, d-xylulose and d-sorbose. Its activity was maximal at pH 9 and 40°C while being enhanced by Mn²⁺. At pH 9 and 40°C, 118 g d-psicose l⁻¹ was produced from 700 g d-fructose l⁻¹ after 3 h. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Effect of pyruvate dehydrogenase complex deficiency on <Emphasis Type="SmallCaps">l</Emphasis>-lysine production with <Emphasis Type="Italic">Corynebacterium glutamicum</Emphasis>

Intracellular precursor supply is a critical factor for amino acid productivity of Corynebacterium glutamicum. To test for the effect of improved pyruvate availability on l-lysine production, we deleted the aceE gene encoding the E1p enzyme of the pyruvate dehydrogenase complex (PDHC) in the l-lysine-producer C. glutamicum DM1729 and characterised the resulting strain DM1729-BB1 for growth and l-lysine production. Compared to the host strain, C. glutamicum DM1729-BB1 showed no PDHC activity, was acetate auxotrophic and, after complete consumption of the available carbon sources glucose and acetate, showed a more than 50% lower substrate-specific biomass yield (0.14 vs 0.33 mol C/mol C), an about fourfold higher biomass-specific l-lysine yield (5.27 vs 1.23 mmol/g cell dry weight) and a more than 40% higher substrate-specific l-lysine yield (0.13 vs 0.09 mol C/mol C). Overexpression of the pyruvate carboxylase or diaminopimelate dehydrogenase genes in C. glutamicum DM1729-BB1 resulted in a further increase in the biomass-specific l-lysine yield by 6 and 56%, respectively. In addition to l-lysine, significant amounts of pyruvate, l-alanine and l-valine were produced by C. glutamicum DM1729-BB1 and its derivatives, suggesting a surplus of precursor availability and a further potential to improve l-lysine production by engineering the l-lysine biosynthetic pathway. This study is dedicated to Prof. Dr. Hermann Sahm on the occasion of his 65th birthday.     

Cloning and characterization of a novel <Emphasis Type="SmallCaps">l</Emphasis>-arabinose isomerase from <Emphasis Type="Italic">Bacillus licheniformis</Emphasis>

Based on analysis of the genome sequence of Bacillus licheniformis ATCC 14580, an isomerase-encoding gene (araA) was proposed as an l-arabinose isomerase (L-AI). The identified araA gene was cloned from B. licheniformis and overexpressed in Escherichia coli. DNA sequence analysis revealed an open reading frame of 1,422 bp, capable of encoding a polypeptide of 474 amino acid residues with a calculated isoelectric point of pH 4.8 and a molecular mass of 53,500 Da. The gene was overexpressed in E. coli, and the protein was purified as an active soluble form using Ni–NTA chromatography. The molecular mass of the purified enzyme was estimated to be ~53 kDa by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and 113 kDa by gel filtration chromatography, suggesting that the enzyme is a homodimer. The enzyme required a divalent metal ion, either Mn²⁺or Co²⁺, for enzymatic activity. The enzyme had an optimal pH and temperature of 7.5 and 50°C, respectively, with a k _cat of 12,455 min⁻¹ and a k _cat/K _m of 34 min⁻¹ mM⁻¹ for l-arabinose, respectively. Although L-AIs have been characterized from several other sources, B. licheniformis L-AI is distinguished from other L-AIs by its wide pH range, high substrate specificity, and catalytic efficiency for l-arabinose, making B. licheniformis L-AI the ideal choice for industrial applications, including enzymatic synthesis of l-ribulose. This work describes one of the most catalytically efficient L-AIs characterized thus far.     

<Emphasis Type="SmallCaps">L</Emphasis>(+)-Lactic Acid Production Using <Emphasis Type="Italic">Lactobacillus Casei</Emphasis> in Solid-State Fermentation

Lactobacillus casei was grown at 37 °C on sugarcane bagasse (5 g) soaked with cassava starch hydrolysate (final moistening volume 34 ml) containing 3 g reducing sugar in a solid-state condition. The maximum yield of l-lactic acid after various process optimisations was 2.9 g/5 g initial substrate corresponding to 97% conversion of sugar to lactic acid with initial substrate moisture of 72%.     

Cloning,sequencing, overexpression and characterization of <Emphasis Type="SmallCaps">l</Emphasis>-rhamnose isomerase from <Emphasis Type="Italic">Bacillus pallidus</Emphasis> Y25 for rare sugar production

The l-rhamnose isomerase gene (L -rhi) encoding for l-rhamnose isomerase (l-RhI) from Bacillus pallidus Y25, a facultative thermophilic bacterium, was cloned and overexpressed in Escherichia coli with a cooperation of the 6×His sequence at a C-terminal of the protein. The open reading frame of L -rhi consisted of 1,236 nucleotides encoding 412 amino acid residues with a calculated molecular mass of 47,636 Da, showing a good agreement with the native enzyme. Mass-produced l-RhI was achieved in a large quantity (470 mg/l broth) as a soluble protein. The recombinant enzyme was purified to homogeneity by a single step purification using a Ni-NTA affinity column chromatography. The purified recombinant l-RhI exhibited maximum activity at 65°C (pH 7.0) under assay conditions, while 90% of the initial enzyme activity could be retained after incubation at 60°C for 60 min. The apparent affinity (K _m) and catalytic efficiency (k _cat/K _m) for l-rhamnose (at 65°C) were 4.89 mM and 8.36 × 10⁵ M⁻¹ min⁻¹, respectively. The enzyme demonstrated relatively low levels of amino acid sequence similarity (42 and 12%), higher thermostability, and different substrate specificity to those of E. coli and Pseudomonas stutzeri, respectively. The enzyme has a good catalyzing activity at 50°C, for d-allose, l-mannose, d-ribulose, and l-talose from d-psicose, l-fructose, d-ribose and l-tagatose with a conversion yield of 35, 25, 16 and 10%, respectively, without a contamination of by-products. These findings indicated that the recombinant l-RhI from B. pallidus is appropriate for use as a new source of rare sugar producing enzyme on a mass scale production.     

Production of <Emphasis Type="SmallCaps">d</Emphasis>-lactic acid from sugarcane molasses,sugarcane juice and sugar beet juice by <Emphasis Type="Italic">Lactobacillus delbrueckii</Emphasis>

Lactobacillus delbrueckii was grown on sugarcane molasses, sugarcane juice and sugar beet juice in batch fermentation at pH 6 and at 40°C. After 72 h, the lactic acid from 13% (w/v) sugarcane molasses (119 g total sugar l⁻¹) and sugarcane juice (133 g total sugar l⁻¹) was 107 g l⁻¹ and 120 g l⁻¹, respectively. With 10% (w/v) sugar beet juice (105 g total sugar l⁻¹), 84 g lactic acid l⁻¹ was produced. The optical purities of d-lactic acid from the feedstocks ranged from 97.2 to 98.3%.     

<Emphasis Type="SmallCaps">L</Emphasis>-Tyrosine production by deregulated strains of <Emphasis Type="Italic">Escherichia coli</Emphasis>

The excretion of the aromatic amino acid l-tyrosine was achieved by manipulating three gene targets in the wild-type Escherichia coli K12: The feedback-inhibition-resistant (fbr) derivatives of aroG and tyrA were expressed on a low-copy-number vector, and the TyrR-mediated regulation of the aromatic amino acid biosynthesis was eliminated by deleting the tyrR gene. The generation of this l-tyrosine producer, strain T1, was based only on the deregulation of the aromatic amino acid biosynthesis pathway, but no structural genes in the genome were affected. A second tyrosine over-producing strain, E. coli T2, was generated considering the possible limitation of precursor substrates. To enhance the availability of the two precursor substrates phosphoenolpyruvate and erythrose-4-phosphate, the ppsA and the tktA genes were over-expressed in the strain T1 background, increasing l-tyrosine production by 80% in 50-ml batch cultures. Fed-batch fermentations revealed that l-tyrosine production was tightly correlated with cell growth, exhibiting the maximum productivity at the end of the exponential growth phase. The final l-tyrosine concentrations were 3.8 g/l for E. coli T1 and 9.7 g/l for E. coli T2 with a yield of l-tyrosine per glucose of 0.037 g/g (T1) and 0.102 g/g (T2), respectively.     

Enzymatic production of <Emphasis Type="SmallCaps">l</Emphasis>-amino acids from the corresponding <Emphasis Type="SmallCaps">dl</Emphasis>-5-substituted hydantoins by <Emphasis Type="Italic">Bacillus fordii</Emphasis> MH602

Bacillus fordii MH602 was newly screened from soil at 45 °C and exhibited high activities of hydantoinase and carbamoylase, efficiently yielding l-amino acids including phenylalanine, phenylglycine and tryptophan with the bioconversion yield of 60–100% from the corresponding dl-5-substituted hydantoins. Hydantoinase activity was found to be cell-associated and inducible. The optimal inducer was dl-5-methylhydantoin with concentration of 0.014 mol L⁻¹ and added to the fermentation medium in the exponential phase of growth. In the production of optically pure amino acids from dl-5-benylhydantoin, the optimal temperature and pH of this reaction were 45–50 °C and 7.5 respectively. The hydantoinase was non-stereoselective, while carmbamoylase was l-selective. The hydantoinase activity was not subject to substrate inhibition, or product inhibition by ammonia. In addition, The activities of both enzymes from crude extract of the strain were thermostable; the hydantoinase and carbamoylase retained about 90% and 60% activity after 6 h at 50 °C, respectively. Since reaction at higher temperature is advantageous for enhancement of solubility and for racemization of dl-5-substituted hydantoins, the relative paucity of l-selective hydantoinase systems, together with the high level of hydantoinase and carbamoylase activity and unusual substrate selectivity of the strain MH602, suggest that it has significant potential applications.     

Fungal arabinan and <Emphasis Type="SmallCaps">l</Emphasis>-arabinose metabolism

l-Arabinose is the second most abundant pentose beside d-xylose and is found in the plant polysaccharides, hemicellulose and pectin. The need to find renewable carbon and energy sources has accelerated research to investigate the potential of l-arabinose for the development and production of biofuels and other bioproducts. Fungi produce a number of extracellular arabinanases, including α-l-arabinofuranosidases and endo-arabinanases, to specifically release l-arabinose from the plant polymers. Following uptake of l-arabinose, its intracellular catabolism follows a four-step alternating reduction and oxidation path, which is concluded by a phosphorylation, resulting in d-xylulose 5-phosphate, an intermediate of the pentose phosphate pathway. The genes and encoding enzymes l-arabinose reductase, l-arabinitol dehydrogenase, l-xylulose reductase, xylitol dehydrogenase, and xylulokinase of this pathway were mainly characterized in the two biotechnological important fungi Aspergillus niger and Trichoderma reesei. Analysis of the components of the l-arabinose pathway revealed a number of specific adaptations in the enzymatic and regulatory machinery towards the utilization of l-arabinose. Further genetic and biochemical analysis provided evidence that l-arabinose and the interconnected d-xylose pathway are also involved in the oxidoreductive degradation of the hexose d-galactose.     

Engineering of pentose transport in Corynebacterium glutamicum to improve simultaneous utilization of mixed sugars

Corynebacterium glutamicum strains CRA1 and CRX2 are able to grow on l-arabinose and d-xylose, respectively, as sole carbon sources. Nevertheless, they exhibit the major shortcoming that their sugar consumption appreciably declines at lower concentrations of these substrates. To address this, the C. glutamicum ATCC31831 l-arabinose transporter gene, araE, was independently integrated into both strains. Unlike its parental strain, resultant CRA1-araE was able to aerobically grow at low (3.6 g·l⁻¹) l-arabinose concentrations. Interestingly, strain CRX2-araE grew 2.9-fold faster than parental CRX2 at low (3.6 g·l⁻¹) d-xylose concentrations. The corresponding substrate consumption rates of CRA1-araE and CRX2-araE under oxygen-deprived conditions were 2.8- and 2.7-fold, respectively, higher than those of their respective parental strains. Moreover, CRA1-araE and CRX2-araE utilized their respective substrates simultaneously with d-glucose under both aerobic and oxygen-deprived conditions. Based on these observations, a platform strain, ACX-araE, for C. glutamicum-based mixed sugar utilization was designed. It harbored araBAD for l-arabinose metabolism, xylAB for d-xylose metabolism, d-cellobiose permease-encoding bglF ^317A , β-glucosidase-encoding bglA and araE in its chromosomal DNA. In mineral medium containing a sugar mixture of d-glucose, d-xylose, l-arabinose, and d-cellobiose under oxygen-deprived conditions, strain ACX-araE simultaneously and completely consumed all sugars.     

<Emphasis Type="SmallCaps">d</Emphasis>-Psicose production from <Emphasis Type="SmallCaps">d</Emphasis>-fructose using an isolated strain, <Emphasis Type="Italic">Sinorhizobium</Emphasis> sp.

Sinorhizobium sp., which can convert d-fructose into d-psicose, was isolated from soil. The optimal pH, temperature, and cell concentration for d-psicose production with the isolated strain were 8.5, 40°C, and 60 mg/ml, respectively. The toluene-treated cells showed 2.5- and 4.8-fold increases in the d-psicose concentration and productivity compared with untreated washed cells. Under the optimal conditions, the toluene-treated cells produced 37 g d-psicose/l from 70% (w/v) (3.9 M) d-fructose after 15 h.