JOURNAL OF BIOSCIENCE AND BIOENGWEERING Vol. 88, No. 6, 676-678. 1999

Production of D-Lyxose from D-Glucose by Microbial and Enzymatic Reactions

ZAKARIA AHMED,’ HIROYUKI SASAHARA,2 SHAKHAWAT HOSSAIN BHUIYAN,’ TETSUYA SAIKI,’ TSUYOSHI SHIMONISHI,’ GORO TAKADA,’ AND KEN IZUMORI’*

Department of Biochemistry and Food Sciences, Faculty of Agriculture, Kagawa University, Miki-cho, Kagawa 761-0795’ and Food Research Institute, Kagawa Prefectural Government, Goto-cho, Takamatsu, Kagawa 761-8031,2 Japan

Received 5 July 1999/Accepted 8 September 1999

o-Arabitol was first prepared from o-glucose using Candidafamata R28. The reaction gave 5.0% o-arabitol from 10.0% o-glucose. o-Arabitol was then almost completely converted to o-xylulose using Acetobacter aceti IF0 3281. Finally, o-lyxose was prepared from o-xylulose enzymatically using L-ribose isomerase from tol- uene-treated cells of Acinetobacter sp. strain DL-28. The isomerization reaction progressed steadily and the concentration of o-xylulose increased from 1.0 to 10.0%. About 70% of o-xylulose was converted to u-lyxose in all cases. Separation of residual o-xylulose from the reaction mixture is very difficult to achieve by column chromatography, but u-xylulose could be selectively degraded easily using Saccharomyces cerevisike IF0 0841. The product was crystallized and was confirmed to be o-lyxose by HPLC, 13C-NMR spectra, IR spectra analysis, and optical rotation measurement.

[Key words: D-glucose, D-arabitol, D-xylulose, D-lyxose]

The configurations of various diastereoisomeric sugars are related through a common aldopentose (such as D-

arabinose, D-arabinitol, and D-lyxose) and this relation- ship offers a simple route for the synthesis of uncommon aldoses from common ones (1). The aldopentose D-

lyxose has promise for use as a starting material for the production of antitumoral and immunostimulatory a- galactosylcer-amide agents that have been found to be active against several murine tumors (2, 3): Agelaspin- 9b (AGL-9b), for example, can be useful in cancer treat- ment. To allow this sugar to be utilized for various other important purposes, a convenient production method needs to be established. D-Lyxose is comparatively rare in nature and has only been obtainable by methods in- volving chemical synthesis (1, 4, 5). However, chemical methods suffer from several drawbacks: many steps and long reaction times are necessary for product formation, a number of expensive chemicals are required, the proce- dures are laborious, and unnecessary by-products are formed. As a result, these methods are not suitable for mass production of this rare sugar. To obtain D-lyxose in amounts large enough for it to be used as a starting material to formulate useful products for biological ac- tivities, a simple and less expensive mass-production method with high yield is essential. Earlier wide-ranging studies on the production of polyols such as ribitol, erythritol, xylitol, D-arabitol, and D-mannitol during the fermentation of soy sauce by halotolerant yeasts (6, 7) revealed that Candida sp. is one of the most potent microorganisms for D-arabitol production (8, 9). We previously isolated a halotolerant yeast, Candida famata R28, from soy sauce mash which can produce D-talitol from D-psicose (10). Previous reports have shown that D-xylulose can be prepared by microbial and enzymatic methods (ll-14), and in the course of studies on the production of rare ketoses, Acetobacter sp. was found to be a potent oxidizer for polyols (15, 16). We have also reported the constitutive L-ribose isomerase (L-RI)

* Corresponding author.

activity of Acinetobacter sp. strain DL-28 and shown that the enzyme can catalyze the isomerization of D-man- nose, D-lyxose and L-ribose (17). Here, we describe a sim- ple method for the production of D-lyxose from D-glu- cose by a three-step reaction sequence involving two microbial conversions followed by enzymatic isomeriza- tion. As shown in Fig. 1, the first reaction is the micro- bial conversion of D-glucose to D-arabitol by the halotol- erant yeast, C. famata, the second is the microbial oxida- tion of D-arabitol to D-xylulose by Acetobacter aceti strain IF0 3281, and finally, D-xylulose is enzymatically isomerized to produce D-lyxose by L-RI. As far as we know, this is the first report on the production of D- lyxose using a microbial and enzymatic method.

C. famata was cultivated aerobically in the medium described previously (10) at 30°C with shaking for 24 h. Cells were then harvested by centrifugation, washed twice with 0.05 M glycine-NaOH buffer @H9.0), and used for the production of D-arabitol from D-glucose. The C. famata cell reaction was carried out at 30°C with shaking in the following reaction mixture: lO.Og D-

glucose, 50.0ml of 0.1 M glycine-NaOH buffer @H 9.0), and 50.0ml of a washed cell suspension prepared in the same buffer which had an absorbance of 80 at 600nm (total volume, 100.0 ml). The rate of transformation and accumulation of D-arabitol in the reaction mixture were detected by the glucose oxidase method and HPLC analy- sis (Nihonbunko HPLC 880 PU liquid chromatograph; Shimadzu RID-6A refractive index detector; and Shimadzu C-R6A chromatopac; Hitachi GL-611 HPLC column). Separation was achieved at 60°C using lob4 M NaOH at a flow rate of 1.0 ml/min. The conversion of 10.0% D-glucose to 5.0% D-arabitol was achieved after 75 h (data not shown). The D-arabitol was then used as a substrate for the production of D-xylulose by the washed cell reaction of the potent microorganism Acetobacter sp. IF0 3281. After being cultivated in the medium described previously (15), cells were harvested and washed twice with 0.05 M glycine-NaOH buffer (pH 9.0). The reaction for D-xylulose production was carried out

676

VOL. 88, 1999 NOTES 677

YHO

yH,OH HVOH

HOCH HOTH HCOH Candida fimata R-28

H&OH HFoH

&H,OH HFOH

CH,OH D-Arabitol DGlucose

3 CH,OH ?HO &O

HObH - L-Ribose isomerase

HOFH HOFH

H&OH 30:70 -

&H,OH

H<OH CH,OH

bxylulose DLyxose

FIG. 1. Schematic diagram for the production of D-lyxose from D-glucose.

at 30°C with shaking in the following reaction mixture: 5.0% D-arabitol, 0.05 M glycine-NaOH buffer (PH 9.0), and a washed cell suspension prepared in the same buffer which had an absorbance of 10 at 600nm (total volume, lOO.Oml). The formation of D-xylulose was de- tected by the cysteine-carbazole reaction (18) and HPLC analysis. Almost complete conversion from D-arabitol to D-xylulose was achieved after 20 h without any by-prod- uct formation (data not shown). Finally, D-lyxose was synthesized from D-xylulose by the enzymatic reaction of L-RI from Acinetobacter sp. DL-28. To prepare the L-RI, cells were grown in the medium described previ- ously (17), harvested, washed twice with 0.05 M glycine- NaOH buffer (pH 9.0), and treated with toluene in the following manner: in a reaction mixture containing 10 ml toluene and 90 ml of 0.05 M glycine-NaOH buffer (pH 9.0), 9.Og of washed cells were incubated at 30°C with shaking for 15 min (final volume, lOO.Oml). The reaction for the production of D-lyxose was carried out in a reaction mixture composed of 9.Og (18 U/ml) tol- uene-treated cells of Acinetobacter sp. DL-28 and a 5.0% solution of D-xylulose with shaking at 30°C (total volume, lOO.Oml). The transformation of D-xylulose to D-lyxose was determined by the method of Dische and Borenfreund (18) and HPLC analysis. At substrate con- centrations from 1.0 to IO%, the reaction reached equilibrium after 12 h with an overall yield of D-lyxose from D-xylulose of about 70% (data not shown). After the equilibrium state was reached, the reaction mixture was collected by centrifugation and was then treated with washed cells of Saccharomyces cerevisiae IF0 0841 for the selective degradation of ketose (D-xylulose) from the resultant reaction mixture and isolation of the product (D-lyxose). The general conditions for degrada- tion of ketose were as follows. A reaction mixture was prepared containing 2.Og of washed S. cerevisiae cells in

0 10 15 20 25 30

Time (h)

FIG. 2. Degradation of D-xylulose from the reaction mixture during the reaction by Saccharomyces cerevisiae IF0 0841.

100.0 ml of a mixed solution of D-xylulose and D-lyxose. The reaction was carried out at 30°C with shaking for 27 h. The rate of degradation of D-xylulose was deter- mined by the cysteine-carbazole reaction (18) by measur- ing the decrease in absorbance at 540 nm and the reduc- ing sugar was assayed by the method of Somogyi and Nelson (19) as well as by HPLC analysis. The time course of ketose degradation (D-xylulose) from the reac- tion mixture using S. cerevisiue is shown in Fig. 2. The Somogyi and Nelson analysis revealed that the mixture of D-lyxose and D-xylulose decreased while the HPLC analysis indicated that the amount of D-lyxose remained the same during the ketose degradation (data not shown). After the ketose was completely degraded the reaction mixture was centrifuged and the supernatant was collected as the isolated product. The product was deionized by ion-exchange resins (Diaion SKIB, H+ form and Amberlite IRA-411, COj2- form) and concen- trated by evaporation under vacuum at 35°C until a vis- cous mass was obtained. Then, a very small amount of authentic D-lyxose crystal was added to the viscous mass and it was kept in a desiccator. After a few days, the crystals that formed were washed and dried in a desicca- tor. All the physical data (HPLC, IR spectra, W-NMR spectra, optical rotation) for this product were identical to those of authentic D-lyxose.

D-Glucose is the most abundant aldose occurring in nature. Although it has been reported previously that D-arabitol can be produced from D-glucose by various yeasts, by-products were concomitantly produced (6, 20). In contrast, our strain of C. famata produced only D-arabitol from D-glucose without any by-product forma- tion. It should be noted that although all the D-glucose (10.0%) was fermented, our analysis revealed that about 5.0% D-arabitol was produced, the rest being consumed by the cells (data not shown). Microbiological and en- zymatic production of D-xylulose from D-arabitol have been reported previously (10, 16). However, Acetobacter sp. IF0 3281 used in this study demonstrated an ex- tramely high level of activity in the transformation of D-arabitol to D-xylulose at a relatively high substrate concentration, and showed no product or substrate consumption or by-product formation tendencies. The enzyme L-RI exhibited excellent D-xylulose to D-lyXOSe isomerization ability even at a high substrate concentra- tion, and in every case the reaction reached an equilibri- um of 70 : 30 (D-lyxose : D-xylulose) after 12 h (data not

678 AHMED ET AL. J. BIOSCI. BIOENG.,

shown), illustrating the advantage of using L-RI for the production of D-lyxose. Because the separation of D-xylu- lose from a reaction mixture containing D-xylulose and D-lyxose is very difficult using column chromatography, we resorted to microbial degradation of D-xylulose in the reaction mixture. Previous studies have reported that pentose can be fermented by yeasts (21, 22), and it has also been shown that D-xylulose can easily be degraded by S. cerevisiae (23, 24). In our simple method, D-lyxose was isolated from the reaction mixture by a S. cerevisiae washed cell reaction (Fig. 2). Other yeast and bacterial strains that we have so far tried have been unable to degrade D-xylulose from the reaction mixture like 5’. cerevisiae can. The success of our method of D-lyxose synthesis is due to three factors: (a) all the reactions are selective only for product formation; (b) the production of D-lyxose from D-glucose is a continuous process, which means that product separation or purification at each step is not needed; and (c) most importantly the final product (D-lyxose) is very easily separated from the reaction mixture with no effect on the desired product and without producing by-products. This new technique is an example of interconversion between a hexose and a pentose by biochemical means. From our results, we esti- mate that about 35% D-lyxose can be produced from D-

glucose. The method described here, which involves only three reaction steps and is considerably simpler than reported chemical methods (1, 4, 5), should prove feasi- ble for the mass production of D-1yXOSe.

REFERENCES

1. Giudici, T. A. and Griffin, J. J.: The interconversion of mono- saccharide configurations: arabinose to lyxose. Carbohy. Res., 33, 287-295 (1974).

2. Morita, M., Sawa, E., Yamaji, K., Sakai, T., Natori, T., Koezuka, Y., Fukushima, H., and Akimoto, K.: Practical total synthesis of (2&3S,4R)-l-O-(alpha-D-galactopyranosyl)-IV-hex- acosanoyl-2-amino-1,3,4-octadecanetriol, the antitumoral and immunostimulatory a-galactosylcer-amide, KRN7000. Biosci. Biotech. Biochem., 60, 288-292 (1996).

3. Takagi, Y., Nakai, K., Tsuchiya, T., and Takeuchi, T.: A 5’- (trifluoromethyl) anthracycline glycoside: synthesis of an- titumor-active 7-0-(2,6-dideoxy-6,6,6-trifluoro-alpha-l-lyxo-hex- opyranosyl) adriamycinone. J. Med. Chem., 39, 1582-1588 (1996).

4. Fletcher, H. G., Jr., Diehl, H. W., and Hudson, C. S.: lmprove- ments in the preparation of D-arabinose from calcium D-

gluconate and of D-lyxose from calcium D-galactonate. J. Ameri. Chem. Sot., 72, 4546 (1950).

5. Fletcher, H. G., Jr.: a-o-Lyxose. Ruff degradation of D-galac- tonate. Methods Carbohyd. Chem., 1, 77-79 (1962).

6. Onishi, H. and Suzuki, T.: The production of xylitol, L-

arabinitol and ribitol by yeasts. Agric. Biol. Chem., 30, 113% 1144 (1966).

7. Onishi, H. and Suzuki, T.: Production of erythritol, D- arabitol, D-mannitol and a heptitol-like compound from glycerol by yeasts. J. Ferment. Technol., 48, 563-566 (1970).

8. Jennings, D.H.: Polyol metabolism in fungi. Adv. Microb. Physiol., 25, 149-193 (1984).

9. Bernard, E. M., Christiansen, K. J., Tsang, S., Kiehn, T. E., and Armstrong, D.: Rate of arabinitol production by pathogen- ic yeast species. J. Clin. Microbial., 146, 353-359 (1982).

10. Sasahara, H., Mine, M., and Izumori, K.: Production of u- talitol from D-psicose by Can&da famata R28. J. Ferment. Bioeng., 85, 84-88 (1998).

11. Schwartz, D., Stein, M., Schneider, K. H., and GifThorn, F.: Synthesis of D-xylulose from D-arabitol by enzymatic conver- sion with immobilized mannitol dehydrogenase from Rhodobac- ter sphaeroides. J. Biotechnology, 33, 95-101 (1994).

12. Mitsuhashi, S. and Lampem, J. 0.: Conversion of D-xylose to D-xylulose in extracts of Lactobncillus pentosus. J. Biol. Chem., 204, 1011-1018 (1953).

13. Pronk, J. T., Bakker, A. W., Van Dam, H. E., Straathuf, A. J. J., Scheffers, W. A., and Van Dijken, J. P.: Preparation of D-xylulose from D-xylose. Enzyme Microb. Technol., 10. 537-542 (1988).

14. Doten, R. C. and Mortlock, R. P.: Production of D- and L-xyh- lose by mutants of Klebsiella pneumoniae and Erwina uredo- voru. Appl. Environ. Microbial., 49, 158-162 (1983).

15. Bhuiyan, S. H., Ahmed, Z., Utamura, M., and Izumori, K.: A new method for the production of L-lyxose from ribitol using microbial and enzymatic reactions. J. Ferment. Bioeng., 86, 513-516 (1998).

16. Moses, V. and Ferrier, R. J.: The biochemical preparation of o-xylulose and L-ribulose. Biochem. J., 83, 8-14 (1962).

17. Shimonishi, T. and Izumori, K.: A new enzyme, L-ribose isomerase from Acinetobacter sp. strain DL-28. J. Ferment. Bioeng., 81, 493-497 (1996).

18. Dische, Z. and Borenfreund, E.: A new spectrophotometric method for the detection of keto sugars and trioses. J. Biol. Chem., 192, 583-587 (1951).

19. Somogyi, M.: Notes on sugar determination. J. Biol. Chem., 195, 19-23 (1952).

20. De Wulf, P., Soetaert, W., Schwengers, D., and Vandamme, E. J.: Screening and mutational improvement of a D-ribose secreting Candidu pelliculosa strain. J. Ferment. Bioeng., 82, l-7 (1996).

21. Jeppsson, H., Yu, S., and Hahn-H, B.: Xylulose and glucose fermentation by Saccharomyces cerevisiae in chemostat culture. Appl. Environ. Microbial., 62, 1705-1709 (1996).

22. Yu, S., Jeppsson, H., and Hahn-H, B.: Xylulose fermentation by Saccharomyces cerevisiae and xylose-fermenting yeast strains. Appl. Microbial. Biotechnol., 44, 314-320 (1995).

23. Maleszka, R. and Schneider, H.: Fermentation of D-xylose, xylitol and D-xylulose by yeasts. Can. J. Microbial., 28, 360- 363 (1982).

24. Suihko, M. L. and Drazik, M.: Pentose fermentation by yeasts. Biotechnol. Lett., 5, 107-I 12 (1983).