Direct production of allitol from d -fructose by a coupling reaction using d -tagatose 3-epimerase, ribitol dehydrogenase and formate dehydrogenase

JOURNALOF BIOSCIENCE AND BIOENGINEERING Vol. 90, No. 5, 545-548. 2000

Direct Production of Allitol from D-Fructose by a Coupling Reaction Using D-Tagatose 3-Epimerase, Ribitol Dehydrogenase and Formate Dehydrogenase KEI TAKESHITA,

YUTAKA

ISHIDA,

GORO TAKADA,

AND KEN IZUMORI*

Department of Biochemistry and Food Science,Faculty of Agriculture, Kagawa University, Miki-cho, Kagawa 761-0795, Japan Received 8 May 2OOOIAccepted 21 August 2000

Agitol was produced from o-fructose via a new NADH-regenerating enzymatic reaction system using Dtagatose 3-epimerase (D-TE), ribitol dehydrogenase (RDH), and formate dehydrogenase (FDH). u-Fructose was epimerized to o-psicose by the D-TE of Pseudomonas cichorii ST-24 and the o-psicosewas subsequently reduced to alfitol by the RDH of an RDH-constitutive mutant, X-22, derived from ZCZebsieZZupneumoniae IF0 3321. NADH regeneration for the reduction of o-psicose by the RDH was achieved by the irreversible formate dehydrogenase reaction, which allowed the o-psicose produced from o-fructose to be successivelytransformed to allitol with a production yield from D-fructose of almost 100%. The reactions progressed without any byproduct formation. After separation of the product from the reaction mixture by a simple procedure, a crystal of allitol was obtained in a yield exceeding 90%. This crystal was characterized and determined to be allitol by HPLC analysis, its IR and NMR spectra, its melting point, and optical rotation measurement. [Key words: allitol, n-fructose,

ribitol

dehydrogenase,

Rare polyols, such as L-iditol, allitol, and n-talitol, are not abundant in nature and are therefore very expensive. They are useful not only as sweeteners (e.g. xylitol) or as raw materials of chemical compounds, but also as research reagents. For this reason, we have been studying the microbial production of rare polyols, and have already reported the production of several types from various ketoses by microbial methods; r.-iditol from L-sorbose using Candida famata 234B (l), n-talitol from o-tagatose using Aureobasidium pullulans strain 113b (2), allitol from D-psicose using Enterobacter agglomerans strain 221e (3), D-talitol from D-psicose using Candida famata R28 (4), and D-iditol from D-sorbose using Rhodotolura rubra RYlO (5). Instead of using microbial methods, polyols can also be produced from ketoses enzymatically by coupling a polyol dehydrogenase with a coenzyme regeneration system. For example, Kulbe and Chmiel (6) reported the production of n-mannitol from D-fructose by coupling the enzymes o-mannitol dehydrogenase and formate dehydrogenase (FDH). In this reaction, shown in Fig. 1, the NADH regeneration reaction of FDH is irreversible, so theoretically, all of the n-fructose can be converted to D-mannitol. As noted above, allitol can be produced from o-psicase (3), which in turn can be prepared from D-fructose using D-tagatose 3-epimerase (D-TE) (7). However, there are two problems with this method: (i) the equilibrium ratio of the epimerization reaction in preparing the starting material-o-psicose-from D-fructose, and (ii) the relatively low substrate concentration in the bacterial reduction reaction of D-psicose. To overcome these drawbacks, we have devised a new enzymatic reaction system that uses D-TE, ribitol dehydrogenase (RDH), and FDH. In this system, shown in Fig. 2, NADH is regenerated using the FDH reaction, facilitating the conversion of Dpsicose to allitol. As the FDH reaction is irreversible, all the D-psicose is transformed to allitol by RDH, thereby * Corresponding

n-tagatose 3-epimerase, formate dehydrogenase]

removing the equilibrium between D-fructose and D-psicase. Using this new enzyme reaction system, all the Dfructose could be transformed to allitol directly, eliminating the need for processes to separate D-psicose from the reaction mixture. As far we know, this is the first report of allitol production using an enzymatic method. MATERIALS

AND METHODS

Chemicals All carbohydrates were purchased from Sigma Chemical Company (St. Louis, MO, USA). All other chemicals used were of reagent grade and were obtained from Wako Pure Chemical Industries (Osaka). Formate dehydrogenase (FDH) was purchased from Boehringer Mannheim (Mannheim, Germany). D-Tagatose 3-epimerase (D-TE) was prepared from recombinant Escherichia coli strain pIK-01 by the method of Ishida et al. (8). Culture conditions for the microbe and preparation of crude RDH Cells of Klebsiella pneumoniae strain X-

22, ribitol dehydrogenase a (RDH)-constitutive mutant strain derived in this study from K. pneumoniae IF0 3321, were grown aerobically in a yeast extract/PolyC&OH A=0

CHzOH HO J H

HO c! H 1

HO l! H H A OH

.

HgjH D-Fructose CO:! =

f NADH h

A-

H&OH NAD+ I

D-Manuitol

HCOOH FDH FIG. 1. Scheme for n-mannitol synthesis from o-fructose using n-mannitol dehydrogenase (MDH) in combination with FDH.

author. 545

546 TAKESHITA ET AL. CHzOH

t

D-TE

6 &OH D-Fructose

FIG. 2.

C&OH

CHzOH A=0

L=o HO c!H

A HOH HA OH

J.

H A OH

H A OH

&

:iiioHp;

iii,,

D-P&c;

~

\

Scheme for allitol production

TE, RDH, and FDH.

Tm

HIr;i

FDH from n-fructose using

D-

pepton medium containing 1.0% yeast extract, 0.2% Polypepton, 0.1% MgS04. 7Hz0, and 1.O% xylitol with shaking at 260 strokes per min on a reciprocal shaker at 28°C for 24 h. After cultivation, the cells were collected by centrifugation at 8200 x g for 10 min and then washed with 50mM Tris-HCl buffer (PH 8.0). The washed cells were used to prepare the cell-free extract. Washed X-22 cells were suspended in 50mM Tris-HCl buffer (pH 8.0), disrupted by ultrasonication at 4”C, and the supernatant obtained after centrifugation (12,000 x g for 30 min) was used as the crude enzyme (RDH). Assay of RDH and FDH Both RDH and FDH were assayed by measuring the absorbance of the NADH produced at 340nm. The reaction mixture for each dehydrogenase assay consisted of 0.7 ml 0.05 M Tris-HCl buffer (pH 8.0), 0.1 ml 0.025 M NAD+, 0.1 ml enzyme solution, and 0.1 ml 0.1 M substrate (ribitol or sodium formate). The reaction was initiated by addition of the substrate and the rate of NADH production was measured as the absorbance at 340nm. One unit of enzyme is defined as the amount required to convert 1.0 pmol of NAD+ to NADH in one min at 30°C. Analytical methods for sugars The reduction of ketose to polyol in the reaction mixture was detected by measuring the decrease in ketose using the method of Dische and Borenfreund (9). The polyol produced was assayed by high-performance liquid chromatography (HPLC). The HPLC system consisted of Nihonbunko 880 PU liquid chromatograph (Tokyo), a Shimadzu (Kyoto) RIDdA refractive index detector and Shimadzu CR-6A Chromatopac) using a Hitachi GL-611 column. Separation was achieved at 60°C using 10p4M NaOH at a flow rate of 1.0 ml/min. The 13C NMR spectra of the product were measured using a Nihonbunko NMR spectrophotometer according to the method reported previously (3). The infrared spectrum was measured using a Nihonbunko model A-302 infrared spectrophotometer. The melting point was measured with a Yanagimoto Micro melting point apparatus (Kyoto). RESULTS Selection of suitable polyol dehydrogenase for reduction of n-psicose to allitol As shown in Fig. 2, D-psicase is reduced to allitol by a polyol dehydrogenase whose cofactor is NADH. If this polyol dehydrogenase was active on D-fructose leading to the production of Dmannitol or D-sorbitol, it will be unsuitable for use in our system, even if the enzyme had high activity for the production of allitol. This is because, n-mannitol or o-sorbitol might be produced from D-fructose by the dehydro-

BIOSCI.

BIOENG.,

genase and remain in the reaction mixture as impurity. It was therefore essential to select a suitable polyol dehydrogenase that is active on D-psicose producing allitol but not active on D-fructose. On the basis of these criteria, we selected the RDH from K. pneumoniae IF0 3321 as a suitable polyol dehydrogenase for the system. Among polyol dehydrogenases examined thus far, only the RDH from K. pneumoniae IF0 3321 has a substrate specificity suitable for this allitol-producing system. Table 1 shows the substrate specificities of the RDHs from K. pneumoniae IF0 3321 and Enterobacter agglomerans strain 221e. Both are active on allitol, but the RDH of strain 221e is also active on D-sorbitol. If the strain 221e RDH were used in this system, D-sorbitol would be produced from D-fructose as a by-product. Isolation of RDH-constitutive mutant strain X-22 from K. pneumoniae IF0 3321 The RDH of K. pneumoniae IF0 3321 is inducible with ribitol as the inducer. However, if this RDH were to be used for mass production of allitol, the cost of a culture medium containing sufficient ribitol to obtain the necessary amount of RDH would be high. We therefore attempted to isolate an RDH-constitutive mutant. Strain IF0 3321 was unable to utilize xylitol as a growth substrate. The strain was cultivated aerobically at 28°C in an 18 x 180mm test tube with 5 ml of a xylitol/mineral salt medium containing 0.26% (NH4)$04, 0.24% KH2P04, 0.56% K2HP04, 0.01% MgS04.7H20, 0.05% yeast extract, and 1.0% xylitol, pH 7.0. After about 72 h incubation, sudden cell growth occurred and the cells reached the stationary growth phase after 120 h. The cells were subcultured three times-once every 48 h. Cells in the second and third subcultures, grew well after 24 h incubation (Fig. 3). Cells of the third cultivation were streaked on to the plates of a mineral salts medium containing 1% xylitol and 36 colonies were isolated. Distinct colonies isolated on the agar plates were cultivated on a yeast extract/Polypepton liquid medium containing no ribitol. All the colonies produced RDH constitutively. One colony, which showed about twice the activity of the parent strain, was picked up, and named strain X-22. This strain produced RDH constitutively at about 180U per 1OOml medium (the parent strain produced about 90 U per lOOm1 of a medium containing 1% ribitol). The

Time (h) FIG. 3. Isolation of mutant through sequential subcultures of the parent strain, IF0 3321, in a mineral salts medium containing 1% xylitol. Symbols: 0, 1st subculture; A, 2nd subculture; 0, 3rd subculture. Cell growth was measured as the absorbance at 660 nm using Taitec (Saitama) Miniphoto 518 photometer.

VOL.

DIRECT PRODUCTION

90, 2000

OF ALLITOL

FROM D-FRUCTOSE

547

TABLE 1. Substrate specificity of ribitol dehydrogenases from K. pneumoniae IF03321 and E. aggIomerans strain 221e Relative activities (%)” K. pneumoniae E. agglomerans IF03321 strain 221e (13) 100 100

Substrate (polyol) Ribitol L-Talitol Allitol r-Mannitol Xylitol r..-Arabitol n-Arabitol D-Talitol D-Mannitol D-Sorbitol Galactitol

52

27 66 45 43 57 0

11 2

1 0 0 0


0 0 0

0

0 5 3

a Activity against ribitol: 100%. The reaction mixture for the dehydrogenase assays consisted of 0.7ml 0.05 M Tris-HCL buffer (pH 8.0), 0.1 ml 0.025 M NAD+, 0.1 ml enzyme solution, and 0.2 ml 0.1 M polyol.

crude enzyme from strain X-22 showed high activity on ribitol, L-talitol, allitol and was not active on D-mannitol, D-sorbitol, D-talitol and galactitol. The relative activities of crude enzyme from strain X-22 against various polyols were quite same as that of RDH from strain IF0 3321 shown in Table 1. This result indicated that we could use this enzyme to produce allitol from D-fructose without any by-product formation, because this RDH crude preparation did not have any activity against Dfructose. Production of allitol from D-psicose We first investigated the production of allitol from D-psicose by a coupling reaction using RDH and FDH. The reaction conditions (concentrations of NAD+, n-psicose, and HCOONa; units of RDH and FDH; buffer pH; reaction temperature) were optimized according to the process used to produce D-mannitol from n-fructose (Fig. 1) (6). The reaction mixture optimized for allitol production from n-psicose contained 20 mg D-psicose, 20mg HCOONa, 2.5 mM NAD+, 1.0 U RDH, and 1.0 U FDH in 2.0 ml of 50 mM Tris-HCl buffer (pH 8.0). The mixture was incubated at 30°C. As shown in Fig. 4, almost all D-psicose was transformed to allitol in 48 h, and no by-product formation was detected by HPLC analysis after the reaction.

0

10

20

30

40

50

60

10

20

30

40

50

60

70

(h) FIG. 5. Allitol production from n-fructose using D-TE, RDH, and FDH. Symbols: 0, n-fructose; 0, allitol. The production of allitol and decrease in n-fructose are shown by the transformation yield (%). The initial concentration of n-fructose was lOmg/ml, as described in the text. Time

Production of allitol from D-fructose Finally, we conducted an experiment to produce allitol directly from D-fructose by the addition of D-TE to the coupling reaction system. The reaction was carried out under the same conditions as those used for allitol production from n-psicose, except that 20mg n-fructose was used as the raw material instead of n-psicose and 2.OU D-TE was added to the reaction mixture. The reaction was again completed in 48 h, and the transformation rate was almost 100% with 1% substrate concentration (Fig. 5). No by-product formation was observed during the reaction. Crystal allitol was prepared from D-fructose using 1OOml of the same reaction mixture. After the reaction had finished, the mixture was treated with activated charcoal for 1 d at room temperature, centrifuged to remove the charcoal and filtrated. The sample was then deionized by passing it through Diaion SKIB (H+ form) and Amberlite IRA-41 1 (C032-) ion-exchange resins. After evaporation and concentration under a vacuum at 35”C, a small allitol crystal was added to the syrup, which solidified on being kept in a desiccator at @C: About 0.9 g of crystal was-obtained. Authentic

allitol

70

Time (h) FIG. 4. Allitol production from n-psicose using RDH and FDH. Symbols: 0, n-psicose; 0, allitol. The production of allitol and decrease in o-psicose are shown by transformation yield (%). The initial concentration of n-psicose was 10 mg/ml, as described in the text.

4000

2000

1500

Wave number

FIG. 6.

1000

500

(td)

Infrared spectra of authentic allitol and the product.

548

TAKESHITA TABLE

2.

Peak no. 1 2 3

J. BIOSCI. BIOENG.,

ET AL.

Chemical shifts of 13C-NMR of authentic allitol and the product Chemical shifts (ppm) Authentic allitol

Product

63.2 73.0 73.2

63.2 73.0 73.2

a Internal reference: 1,4-dioxane (67.4 ppm).

Identification of the crystal product Authentic allitol, the crystal produced, a mixture of the two had the same melting point of 150°C. The retention time as determined by HPLC analysis, the IR spectrum (Fig. 6), and the i3C-NMR spectrum (Table 2) of the isolated crystal were identical to those of authentic allitol. The specific optical rotation was determined to be 0. On the basis of these results, the crystal produced from D-fructose using the described enzyme coupling system with RDH, D-TE, and FDH was judged to be as allitol. DISCUSSION To establish an enzymatic system for the production of allitol from D-fructose without any by-product formation using an enzyme coupling system, it was first essential to select an NAD+-dependent polyol dehydrogenase that is active on allitol but not on D-sorbitol or D-mannitol; otherwise n-fructose could be reduced to D-sorbitol or D-mannitol, which would remain as an impurity in the product. The RDH of K. pneumoniae IF0 3321 was found to have suitable properties for the system. The mutant X-22, which produces RDH constitutively at a high level (180 units per 100 ml medium), was subsequently isolated by the method previously used to obtain an L-ribose isomerase-constitutive mutant of Acinetobacfer sp. (10) and an L-rhamnose isomerase-constitutive mutant of Pseudomonas sp. (11). In these cases, Dlyxose and L-lyxose were respectively used as carbon sources, for growth. The epimerization reaction between o-fructose and D-psicose mediated by D-TE is reversible, and the ratio of D-fructose to D-psicose at equilibrium is 75 : 25 (12)-which means that when D-psicose is prepared from D-fructose using D-TE, the maximum of production yield of D-psicose is only 25%. In the new system, the irreversible reaction of FDH resulted in a very high yield of allitol from n-fructose because all of the D-

psicose epimerized from D-fructose was successively reduced to allitol. The system developed in this study should be applicable to the production of various poly01s from a variety of aldoses coupled with aldose isomerases instead of D-TE. REFERENCES 1. Sasahara, H. and Izumori, K.: Reduction of L-sorbose by halotolerant yeast, Candida famata 234B. Seibutsu-kougaku, 72, 299-304 (1994). 2. Muniruzzaman, S., Kobayashi, H., and Izumori, K.: Production of D-talitol from D-tagatose by Aureobasidium pullulans strain 113B. J. Ferment. Bioeng., 78, 346-350 (1994). 3. Maniruzzaman, S., Tokunaga, H., and Izumori, K.: Conversion of D-psicose to allitol by Enterobacter aggtomerans strain 221e. J. Ferment. Bioeng., 79, 323-327 (1995). 4. Sasahara, H., Mine, M., and Izumori, K.: Production of Dtahtol from n-psicose by Candida famata R28. J. Ferment. Bioeng., 85, 84-88 (1998). 5. Sasahara, H. and Izumori, K.: Production of D-iditol from Dsorbose by Rhodotolura rubra RYlO isolated from miso paste. J. Biosci. Bioeng., 87, 548-550 (1999). 6. Kulbe, K. D. and Chmlel, H.: Coenzyme-dependent carbohydrate conversions with industrial potential. Ann. New York Acad. Sci., 542, 444-464 (1988). 7. Itoh, H., Sato, T., and Izumorl, K.: Preparation of p-psicose from D-fructose by immobilized D-tagatose 3-epimerase. J. Ferment. Bioeng., 80, 101-103 (1995). 8. Ishida, Y., Kamiya, K., and Izumori, K.: Production of Dtagatose 3-epimerase of Pseudomonas cichorii ST-24 using recombinant Escherichia coli. J. Ferment. Bioeng., 84, 348-350 (1997). 9. Dische, Z. and Borenfreund, E.: A new spectrophotometric method for the detection of keto sugars and trioses. J. Biol. Chem., 192, 583-587 (1951). 10. Shimonishi, T. and Izumori, K.: A new enzyme, L-ribose isomerase from Acinetobacter sp. strain ~~-28. J. Ferment. Bioeng., 81, 493-497 (1996). 11. Bhoiyan, S. H., Itami, Y., and Izamori, K.: Isolation of an Lrhamnose isomerase-constitutive mutant of Pseudomonas sp. strain LL172: purification and characterization of the enzyme. J. Ferment. Bioeng., 84, 319-323 (1997). 12. Itoh, II., Okaya, H., Khan, A. R., Tajima, S., Hayakawa, S., and characterization of Dand Izumori, K.: Purification tagatose 3-epimerase from Pseudomonas sp. ST-24. Biosci. Biotech. Biochem., 58,2168-2171 (1994). 13. Mmdruzzaman, S., Kuuihiia, Y., Ichiraku, K., and Izumori, K.: Purification and characterization of a ribitol dehydrogenase from Enterobacter agglomerans 221e. J. Ferment. Bioeng., 79, 496-498 (1995).