Efficient Production of 2-Deoxyribose 5-Phosphate From

Biosci. Biotechnol. Biochem., 70 (6), 1371–1378, 2006

Efficient Production of 2-Deoxyribose 5-Phosphate from Glucose and Acetaldehyde by Coupling of the Alcoholic Fermentation System of Baker’s Yeast and Deoxyriboaldolase-Expressing Escherichia coli

y Nobuyuki HORINOUCHI,1 Jun OGAWA,1; Takako KAWANO,1 Takafumi SAKAI,1 Kyota SAITO,1 Seiichiro MATSUMOTO,2 Mie SASAKI,2 Yoichi MIKAMI,2 and Sakayu SHIMIZU1

1Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kitashirakawa-oiwakecho, Sakyo-ku, Kyoto 606-8502, Japan 2Tokyo Laboratory, Yuki Gosei Kogyo Co., Ltd., 3-37-1 Sakashita, Itabashi-ku, Tokyo 174-0043, Japan

Received December 5, 2005; Accepted February 19, 2006; Online Publication, June 23, 2006 [doi:10.1271/bbb.50648]

2-Deoxyribose 5-phosphate production through cou- of Klebsiella pneumoniae B-4-4 with triosephosphates pling of the alcoholic fermentation system of baker’s such as D-glyceraldehyde 3-phosphate (G3P) and dihy- yeast and deoxyriboaldolase-expressing Escherichia coli droxyacetone phosphate (DHAP) and acetaldehyde as was investigated. In this process, baker’s yeast generates starting materials.8) Triosephosphates are the intermedi- fructose 1,6-diphosphate from glucose and inorganic ates of glycolysis, so if triosephosphates are supplied phosphate, and then the E. coli convert the fructose 1,6- through glycolysis from cheap sugars, the process diphosphate into 2-deoxyribose 5-phosphate via D-glyc- should become more practical. In a previous study,9) eraldehyde 3-phosphate. Under the optimized condi- we investigated the glycolytic function of DERA- tions with toluene-treated yeast cells, 356 mM (121 g/l) expressing E. coli with several intermediates of glycol- fructose 1,6-diphosphate was produced from 1,111 mM ysis, such as glucose, fructose, fructose 6-phosphate, and glucose and 750 mM potassium phosphate buffer (pH fructose 1,6-diphosphate (FDP), and found that the 6.4) with a catalytic amount of AMP, and the reaction phosphorylated intermediates, especially FDP, served as supernatant containing the fructose 1,6-diphosphate was suitable precursors for G3P. used directly as substrate for 2-deoxyribose 5-phosphate Processes for the production of various substances production with the E. coli cells. With 178 mM enzy- through coupling with the powerful fermentative ability matically prepared fructose 1,6-diphosphate and 400 of baker’s yeast as the energy source were established by mM acetaldehyde as substrates, 246 mM (52.6 g/l) 2- Tochikura and colleagues.10–12) The sugar-fermentative deoxyribose 5-phosphate was produced. The molar yield system of baker’s yeast was used as the ATP donor in of 2-deoxyribose 5-phosphate as to glucose through the these processes. Temporary accumulation of FDP in the total two step reaction was 22.1%. The 2-deoxyribose course of ATP regeneration has been reported.13–15) 5-phosphate produced was converted to 2-deoxyribose Based on these reports, we designed a novel meta- with a molar yield of 85% through endogenous or bolic and enzymatic DR5P production system consisting exogenous phosphatase activity. of glucose fermentation and DERA-catalyzed aldol condensation with glucose and acetaldehyde as starting Key words: 2-deoxyribose 5-phosphate; 2-deoxyribose; materials. In this study, we tried to optimize FDP deoxyriboaldolase; alcoholic fermentation; production from glucose and inorganic phosphate with 20-deoxyribonucleoside toluene-treated baker’s yeast (toluene-treated yeast), and DR5P production from enzymatically prepared FDP and The synthesis of antiviral 20-deoxyribonucleosides acetaldehyde by DERA-expressing E. coli (Fig. 1). such as AZT requires a deoxyribose component as a Further convenient conversion of DR5P to 2-deoxyri- starting material, but chemical synthesis of the deoxy- bose (DR) was also examined. ribose frame is a tedious process with many protection and deprotection steps.1–4) Materials and Methods We performed 2-deoxyribose 5-phosphate (DR5P) production by means of deoxyriboaldolase 5–7) (DERA) Preparation of a toluene-treated yeast. Pressed

y To whom correspondence should be addressed. Tel: +81-75-753-6122; Fax: +81-75-753-6128; E-mail: [email protected] Abbreviations: DR5P, 2-deoxyribose 5-phosphate; DERA, deoxyriboaldolase; FDP, fructose 1,6-diphosphate; G3P, D-glyceraldehyde 3-phosphate; DHAP, dihydroxyacetone phosphate; LB, Luria-Bertani; DR, 2-deoxyribose 1372 N. HORINOUCHI et al.

ATP ATP ADP ADP AMP AMP Glucose FDP ADP ATP CH OH 2 O O P O HO O P OH OH Ethanol, CO2 HO OH OH HO Pi

Pi Baker's yeast

FDP

O OH DERA CHO CH OH P O 2 CHOH C=O

HO CH2O P CH2O P DR5P G3P DHAP

DERA-expressing CH3CHO E. coli

Fig. 1. DR5P Synthesis from Glucose and Acetaldehyde through the Fermentation Energy of Baker’s Yeast and DERA-Expressing E. coli. baker’s yeast (Oriental Yeast, Tokyo) was incubated resulting supernatants were measured as described with 16.6% (v/v) toluene in 176 mM potassium phos- below after checking the accumulation of DR by TLC. phate buffer (pH 7.0) at 37 C for 1 h with standing. The The averages of three separate experiments, which were suspension was centrifuged (3;500 g, 20 min), and the reproducible within 10%, are presented in the text resulting pellet was used as the source of glycolytic and figures. enzymes (toluene-treated yeast). DR production: DR production was carried out by means of temperature shift or phosphatase addition. As Culture conditions for DERA-expressing E. coli. for temperature shift, the complete reaction mixture for DERA-expressing E. coli 10B5/pTS8 was cultivated DR5P production was further incubated at 47 C for at 37 C in Luria-Bertani (LB) medium (1% peptone, 22 h and centrifuged (15;000 g, 15 min), and then the 0.5% yeast extract, 1% NaCl, pH 7.2) for 12 h. Cells amount of DR in the resulting supernatant was assayed. were harvested by centrifugation (8;000 g, 10 min). As for phosphatase addition, the reaction mixture for After washing with a 0.85% NaCl solution, the cells DR5P production was centrifuged (15;000 g, 15 min), were used for DR5P synthesis. and the resulting supernatant was incubated at 80 C for 1 h, and then 2% (w/v) of Sumizyme PM (Shin Nihon Reaction conditions. FDP production: The standard Chemical, Anjo, Japan) was added. After 4 h incubation reaction mixture comprised, in 10 ml, 1,111 mM glucose, at 37 C, the reaction mixture was centrifuged (15;000 500 mM K2HPO4,30mM MgSO4.7H2O, 15 mM AMP. g, 15 min) and the amount of DR in the supernatant was 2Na and 60% (w/v) toluene-treated yeast. The reactions determined. The averages of three separate experiments, were carried out at 37 C for 3–7 h with standing. The which were reproducible within 10%, are presented supernatants obtained on centrifugation (15;000 g, in the text. 15 min) were subjected to FDP analysis and then used for DR5P production as an enzymatically prepared FDP Analytical methods. FDP analysis: FDP was measured solution. The averages of three separate experiments, enzymatically by monitoring a decrease in absorbance at which were reproducible within 10%, are presented 340 nm of NADH through coupled reactions catalyzed in the text and figures. by FDP aldolase, triosephosphate isomerase (TPI), and DR5P production: The standard reaction mixture a-glycerophosphate dehydrogenase (GDH). Twenty ml comprised, in 0.5–60 ml, 50% (v/v) enzymatically of FDP-containing sample solution was added to 140 ml prepared FDP solution, 200 mM acetaldehyde, 200 mM of a reaction mixture comprising 150 mM Tris/HCl potassium phosphate buffer (pH 7.0), 15 mM MgSO4. (pH 7.4) and 0.5 mM NADH. Then 20 ml of a TPI/GDH 7H2O, 0.4% (v/v) polyoxyethylenelaurylamine, 1.0% solution (Sigma, St. Louis, MO) containing 50 U of TPI (v/v) xylene, and 12.5% (w/v) wet cells of E. coli and 5 U of GDH was added to decompose contaminating 10B5/pTS8. The reactions were carried out at 28 C for triosephosphates. After 10 min incubation at 30 C, 20 ml 2–4 h with shaking (120 rpm), followed by centrifuga- of FDP aldolase solution (Sigma) containing 10 U of tion (15;000 g, 15 min). The amounts of DR5P in the FDP aldolase from rabbit muscle was added, and then Enzymatic 2-Deoxyribose 5-Phosphate Production 1373

) A M 250 mM Pi 500 mM Pi 750 mM Pi1000 mM Pi 1250 mM Pi 450

300

150

FDP production (m 0 357 7 357 357 35 357 Reaction time (h) )

M B 250 mM Pi 500 mM Pi 750 mM Pi1000 mM Pi 1250 mM Pi 160 120 80 40 0 DR5P production (m 3 h 5 h 7 h 3 h 5 h 7 h 3 h 5 h 7 h 3 h 5 h 7 h 3 h 5 h 7 h

Reaction time for FDP production

Fig. 2. Effects of the Inorganic Phosphate Concentration on FDP (A) and DR5P (B) Production. A, The reactions were carried out in 10 ml under the standard conditions, except for the inorganic phosphate concentration, for 7 h. B, The reaction mixtures, comprising in 300 ml 25% (v/v) enzymatically prepared FDP solution (with various initial inorganic phosphate concentrations and reaction times), 200 mM KH2PO4,15mM MgSO4.7H2O, 0.4% (v/v) polyoxyethylenelaurylamine, and 1% (v/v) xylene, were incubated at 28 C for 2 h with shaking (120 rpm). Pi, inorganic phosphate. the decrease in NADH was monitored at a wavelength of Result 340 nm with a Spectra Max 190 (Molecular Devices, Sunnyvale, CA) after 20 min incubation at 30 C. The Optimization of FDP production with baker’s yeast FDP concentration was calculated based on the decrease Reaction pH and temperature: The effect of the in NADH using a calibration curve obtained with reaction pH (in a range of 6.0–9.0) and temperature (in a authentic FDP solutions of known concentrations. range of 18–60 C) were examined. The FDP production DR5P analysis: Qualitative analysis of DR5P and DR proceeded well at pH 6.4 (500 mM potassium phosphate was performed by TLC with Kieselgel 60 F254 (Merck, buffer) and 37 C (data not shown). Darmstadt, Germany). The developing system consisted Inorganic phosphate concentration: The highest FDP of n-butanol, 2-propanol, and H2O in a ratio of 3:12:4 production was attained with 1,000 mM inorganic phos- (v/v/v). DR5P and DR were detected with 1% (v/v) phate in 5 h (Fig. 2A). The enzymatically prepared FDP anisaldehyde and 2% (v/v) H2SO4 in acetic acid as solutions obtained with various phosphate concentra- purple spots. Quantitative analysis of DR5P was tions and different reaction times were centrifuged, and performed with cysteine-sulfate as described previous- the resulting supernatants were used directly for DR5P ly,16) after checking for the absence of DR by TLC. production with E. coli 10B5/pTS8. When the enzy- DR analysis: DR was measured by HPLC with a matically prepared FDP solution obtained with 750 mM refractive index detector (Shimadzu RID-6A, Kyoto, phosphate (initial concentration) was used as the Japan) (column, Shodex Sugar KS-801 [8:0 300 mm], substrate, the highest production of DR5P was achieved Showa Denko, Tokyo; eluent, H2O; flow rate, 0.5 ml/ (Fig. 2B). As a result, the optimum initial phosphate min; temperature, 55 C). concentration for FDP production was determined to be Glucose analysis: The amounts of glucose in the 750 mM in relation to DR5P production. reaction mixtures were determined by glucose oxidase Energy carrier: The addition of an energy carrier was (Glucose C2 Test WAKO; Wako Pure Chemical effective, although approximately 130 mM of FDP was Industries, Osaka, Japan). produced even without the addition of an endogenous Inorganic phosphate analysis: The amounts of inor- energy carrier. Among the energy carriers tested (ATP, ganic phosphate in the reaction mixtures were deter- ADP, AMP, adenosine, and adenine, each 15 mM), ATP, mined with a Phospha-C Test WAKO (Wako Pure ADP, and AMP enhanced FDP production (Fig. 3). Chemical Industries). AMP was selected for further optimization. Sugar type and concentration: Monosaccharides (glu- 1374 N. HORINOUCHI et al.

Adenine ) M Adenosine

AMP

ADP

ATP Energy carrier (each 15 m None

0 100 200 300 400

FDP production (mM)

Fig. 3. Effect of Energy Carrier on FDP Production. Reactions were carried out in 10 ml under the standard conditions for 3 h, except that 15 mM of energy carrier (ATP, ADP, AMP, adenosine, or adenine) was used instead of AMP.

Lactose N.D.

Maltose ) M Sucrose

Sorbitol N.D.

Mannose

Rhamnose

Galactose N.D. Saccharides ( each 700 m Fructose

Glucose

0 100 200 300 400

FDP production (mM)

Fig. 4. Production of FDP from Various Saccharides. Reactions were carried out in 10 ml under the standard conditions for 3 h, except that 700 mM of the indidated saccharides were used. N.D., not detected. cose, fructose, galactose, rhamnose, mannose, and Mg2þ concentration: The effect of the Mg2þ concen- sorbitol) and disaccharides (sucrose, maltose, and tration on ATP generation was examined. FDP produc- lactose) were examined. Glucose, fructose, mannose, tion increased with increasing Mg2þ concentrations up sucrose, and maltose served as substrates and energy to 30 mM, but decreased with higher concentrations of sources for FDP production, while galactose, sorbitol, Mg2þ (data not shown). and lactose did not (Fig. 4). Glucose was selected for Acetaldehyde: The addition of acetaldehyde to the further investigation, and the effect of its concentrations reaction mixture enhanced FDP production. FDP pro- was examined. FDP production increased with increas- duction increased with increasing concentrations of ing initial concentration of glucose up to 1,111 mM (data acetaldehyde up to 250 mM, but decreased with higher not shown). The initial glucose concentrations did not concentrations of acetaldehyde (data not shown). have any effect on DR5P productivity. Based on these results, the optimum reaction con- Enzymatic 2-Deoxyribose 5-Phosphate Production 1375

1200

1000

Glc

) 800 M 356 mM (121 g/l) 600 FDP

400 Pi 200 Glucose, Pi, FDP (m

0 1 3 5 7 Reaction time (h)

Fig. 5. Time Courses of FDP Production. The reaction mixture, comprising, in 200 ml, 1,111 mM glucose, 750 mM potassium phosphate buffer (pH 6.4), 15 mM AMP.2Na, 30 mM MgSO4.7H2O, 250 mM acetaldehyde, and 60% (w/v) toluene-treated yeast, was incubated in a 200-ml Erlenmeyer flask at 37 C with standing. Symbols; , FDP; , glucose (Glc); , inorganic phosphate (Pi). ditions for FDP production with toluene-treated yeast were determined to be as follows: The reaction mixture comprised 1,111 mM glucose, 750 mM potassium phos- DR phate buffer (pH 6.4), 15 mM AMP.2Na, 250 mM acet- aldehyde, 30 mM MgSO4.7H2O, and 60% (w/v) tol- uene-treated yeast. Two hundred ml of the reaction mixture was poured into a 200-ml Erlenmeyer flask and then incubated at 37 C with standing. Shaking was less DR5P effective, because soluble oxygen inhibited anaerobic alcoholic fermentation with toluene-treated yeast. As shown in Fig. 5, 356 mM FDP (121 g/l) accumulated in 18 28 37 47 57 5 h, and the inorganic phosphate added was almost completely consumed (the yield of inorganic phosphate Temperature (°C) was 95.9%). Fig. 6. Effect of Reaction Temperature on DR5P Production. These reaction mixtures contained, in 500 ml, 25% (v/v) enzy- Optimization of DR5P production from enzymatically matically prepared FDP solution, 200 mM acetaldehyde, 200 mM prepared FDP and acetaldehyde with E. coli 10B5/ potassium phosphate buffer (pH 6.0), 15 mM MgSO4.7H2O, 0.4% pTS8 (v/v) polyoxyethylenelaurylamine, 1.0% (v/v) xylene and 12.5% The FDP-containing reaction mixture prepared above (w/v) wet cells of E. coli 10B5/pTS8. They were incubated for 2 h was centrifuged, and the resulting supernatant was used with shaking (120 rpm) at the indicated temperatures. directly as the FDP source for DR5P production with DERA-expressing E. coli 10B5/pTS8. The reaction proceeded well under both aerobic (shaking) and about 6.0 without the addition of any buffer. Hence the anaerobic (standing) conditions. To mix the reaction reactions were carried out without any buffer. The effect mixtures well, the reactions were carried out with of reaction temperature (in a range of 18–57 C) on shaking (120 rpm). DR5P was examined. At higher than 28 C, dephos- Reaction pH and temperature: The effect of the phorylation of DR5P to DR was observed, which reaction pH (in a range of 5.0–6.3 with 200 mM resulted in low DR5P accumulation (Fig. 6). Hence potassium phosphate buffer) on DR5P production was the reaction temperature was kept at 28 C. examined. The reaction proceeded well at pH 6.0 (data Substrate concentration: DR5P production increased not shown). The effect of potassium phosphate buffer with increasing concentrations of acetaldehyde up to (pH 6.0) concentration was examined. The reaction was 400 mM (data not shown). With 400 mM acetaldehyde, inhibited with increasing concentrations of potassium the best amount of the enzymatically prepared FDP phosphate buffer and proceeded well without the buffer. solution was equivalent to half the volume (50%, v/v) of Furthermore, the pH of the reaction mixture remained at the reaction mixture (FDP concentration, 178 mM). A 1376 N. HORINOUCHI et al.

250

200 ) M

150

100 FDP production (m 50

0 0 1 2 3 4

Reaction time (h)

Fig. 7. Effect the Enzymatically Prepared FDP Solution on DR5P Production. The reaction mixtures comprised, in 500 ml, 400 mM acetaldehyde, 15 mM MgSO4.7H2O, 0.4% (v/v) polyoxyethylenelaurylamine, 1.0% (v/v) xylene, 12.5% (w/v) wet cells of E. coli 10B5/pTS8, and 25–75% (v/v) enzymatically prepared FDP solution (FDP concentration, 89–267 mM). Reactions were carried out at 28 C for 3 h with shaking (120 rpm). Symbols: , 75% (267 mM FDP); , 50% (178 mM FDP); , 25% (89 mM FDP).

higher amount of FDP solution was less effective for 300 DR5P production (Fig. 7). ) 250 Based on these results, the optimum conditions for M DR5P production from enzymatically prepared FDP and 200 DR5P acetaldehyde with E. coli 10B5/pTS8 were determined 246 mM (52.6 g/l) to be as follows: The reaction mixture comprised 50% 150 (v/v) enzymatically prepared FDP solution (178 mM 100 FDP FDP), 400 mM acetaldehyde, 15 mM MgSO4.7H2O,

0.4% (v/v) polyoxyethylenelaurylamine, 1.0% (v/v) DR5P, FDP, Pi (m 50 xylene, and 12.5% (w/v) wet cells of E. coli 10B5/ Pi 0 pTS8. Improvement of the FDP-permeability of the 1342 catalyst E. coli cells by detergent treatment was effec- Reaction time (h) tive. Xylene and polyoxyethylenelaurylamine were used for this purpose.17) Sixty ml of the reaction mixture was Fig. 8. Time Courses of DR5P Production. poured into a 200-ml Erlenmeyer flask, followed by The reaction mixture, comprising, in 60 ml, 50% (v/v, 178 mM) incubation at 28 C for 2 h with agitation by magnetic enzymatically prepared FDP solution, 400 mM acetaldehyde, 15 mM MgSO4.7H2O, 0.4% (v/v) polyoxyethylenelaurylamine, 1.0% (v/v) stirring. Under these conditions, 246 mM DR5P (52.6 xylene, and 12.5% (w/v) wet cells of 10B5/pTS8, was incubated in g/l) accumulated in 2 h (Fig. 8). Theoretically, 2 mol of a 200-ml Erlenmeyer flask at 28 C for 4 h with magnetic stirring. DR5P is produced from 1 mol of FDP, so the yield of Symbols: , FDP; , DR5P; , inorganic phosphate (Pi). DR5P to FDP in this reaction was calculated to be 69.1%. natant obtained on centrifugation with the addition of Transformation of DR5P into DR exogenous phosphatase (Sumizyme PM-L, 2% w/v) at Dephosphorylation of DR5P into DR was carried out 37 C. by incubating the intact DR5P-containing reaction mixture prepared as above at 47 C to trigger endoge- Discussion nous phosphatase activity (Fig. 6). After 10 h incuba- tion, about 50% of the DR5P was dephosphorylated to In this study, we optimized FDP production using DR, and after 22 h, 210 mM (28.1 g/l) DR was obtained, toluene-treated yeast, and also further transformation to with a molar yield of DR5P of 85%. Almost the same DR5P and DR with DERA-expressing E. coli. amount of DR (approximately 210 mM) was obtained on In this two-step reaction system, it is important that 4 h incubation of the DR5P-containing reaction super- almost all the inorganic phosphate is consumed in the Enzymatic 2-Deoxyribose 5-Phosphate Production 1377 first step (FDP production) because the second step The DR5P produced was easily converted to DR (DERA reaction) is inhibited by inorganic phosphate. through phosphatase-catalyzing reactions. Furthermore, 0 When FDP production was carried out with 750 mM the DR5P was enzymatically converted into 2 -deoxy- inorganic phosphate, almost all the inorganic phosphate ribonucleoside derivatives. Investigation of the latter was converted into FDP, so the subsequent DR5P enzymatic process will be reported elsewhere. production proceeded well. The addition of acetalde- hyde was effective for FDP production. In the presence Acknowledgments of a downstream intermediate of alcoholic fermentation, acetaldehyde, FDP degradation might be depressed, This work was partially supported by the Industrial resulting in an increase in FDP accumulation. A stand- Technology Research Grant Program (no. 02A07001a ing reaction with an almost fully filled Erlenmeyer flask, to J.O.), and the Project for the Development of a to maintain a low soluble oxygen concentration, was Technological Infrastructure for Industrial Bioprocesses effective. Under the preparative reaction conditions, on R&D of New Industrial Science and Technology 356 mM FDP (121 g/l) accumulated in 5 h. The molar Frontiers (to S.S.) of the New Energy and Industrial yields of FDP to glucose and inorganic phosphate were Technology Development Organization of Japan, and by 32% and 95.9% respectively, and the apparent AMP Grants-in-Aid for Scientific Research (no. 16688004 to turnover was 47.5. J.O.) and COE for Microbial-Process Development For optimization of DR5P production from the Pioneering Future Production Systems (to S.S.) from enzymatically prepared FDP solution and acetaldehyde the Ministry of Education, Culture, Sports, Science and with DERA-expressing E. coli 10B5/pTS8, surfactants, Technology of Japan. polyoxyethylenelaurylamine and xylene, were added for improvement of the permeation of phosphorylated References compounds.17) For efficient accumulation of DR5P, incubation of the reaction mixture at 28 C was better 1) Aoyama, H., Stereoselective synthesis of anomers of than at higher temperatures. If the reaction temperature 5-substituted 20deoxyuridines. Bull. Chem. Soc. Jpn., 60, is higher, phosphatases of the E. coli host might become 2073–2077 (1987). active, and in this case DR5P will be dephosphorylated 2) Kawakami, H., Matsushita, M., Naoi, Y., Itoh, K., and 0 into DR. This is a disadvantage for DR5P production, Yoshikoshi, H., The synthesis of 2 -deoxyadenosine via but is favorable for DR production. Under the prepara- a stereospecific coupling reaction. Chem. Lett., 235–238 (1989). tive reaction conditions, 246 mM DR5P (52.6 g/l) was 3) Park, M., and Rizzo, C. J., Stereocontrolled de novo produced from the enzymatically prepared FDP solution synthesis of 20-deoxynucleosides. J. Org. Chem., 61, and acetaldehyde in 2 h. The molar yields of DR5P to 6092–6093 (1996). FDP and acetaldehyde were 69.1% and 61.5% respec- 4) Komatsu, H., Awano, H., Ishibashi, H., Oikawa, T., tively. The molar yield to glucose through the total two- Ikeda, I., and Araki, T., Chemo-enzymatic synthesis of step reaction was 22.1%. If the further transformation of natural and unnatural 20-deoxynucleosides. Nucl. Acids DR5P to 20-deoxyribonucleosides proceeded with al- Res. Suppl., 101–102 (2003). most 100% yield, the yield of DR5P to glucose obtained 5) Barbas, F. C., Wang, Y. F., and Wong, C. H., here (22.1%) might be comparable the yield of ribonu- Deoxyribose 5-phosphate aldolase as a synthetic cata- cleoside fermentation from glucose (7–18%).18–20) lyst. J. Am. Chem. Soc., 112, 2013–2014 (1990). In a previous paper,9) we reported DR5P production 6) Gijsen, J. H., and Wong, C. H., Sequential one-pot aldol from glucose and acetaldehyde with a great deal of ATP reactions catalyzed by 2-deoxyribose-5-phosphate aldo- lase and fructose-1,6-diphosphate aldolase. J. Am. Chem. as the energy carrier for acceleration of glycolysis by Soc., 117, 2947–2948 (1995). E. coli itself. One hundred mM ATP was required for the 7) Greenberg, A. W., Varvak, A., Hanson, R. S., Wong, K., production of 100 mM DR5P. But from the viewpoint of Huang, H., Chen, P., and Burk, J. M., Development of industrial production, ATP is an expensive compound. It an efficient, scalable, aldolase-catalyzed process for is noteworthy that coupling of the glycolytic pathway of enantioselective synthesis of statin intermediates. Proc. baker’s yeast to DR5P synthesis greatly reduced the Natl. Acad. Sci. USA, 101, 5788–5793 (2004). amount of endogenous energy required. The same 8) Ogawa, J., Saito, K., Sakai, T., Horinouchi, N., Kawano, coupling of the alcoholic fermentation system of baker’s T., Matsumoto, S., Sasaki, M., Mikami, Y., and Shimizu, yeast and aldolase-expressing E. coli might be promis- S., Microbial production of 2-deoxyribose 5-phosphate from acetaldehyde and triose phosphate for the synthesis ing for other aldolases and transketolases that use G3P 0 or DHAP as a substrate, such as FDP aldolase, tagatose of 2 -deoxyribonucleosides. Biosci. Biotechnol. Bio- chem., 67, 933–936 (2003). 1,6-diphosphate aldolase, L-fuculose 1-phosphate aldo- 9) Horinouchi, N., Ogawa, J., Sakai, T., Kawano, T., lase, rhamnulose 1-phosphate aldolase, and 1-deoxy-D- Matsumoto, S., Sasaki, M., Mikami, Y., and Shimizu, S., 21,22) xylulose 5-phosphate synthase etc. Using the meth- Construction of deoxyriboaldolase-expressing Esche- od presented here, glucose can be used as a starting richia coli and its application to 2-deoxyribose 5- material for these aldolase-catalyzing reactions instead phosphate synthesis from glucose and acetaldehyde for of expensive phosphorylated compounds as substrates. 20-deoxyribonucleoside production. Appl. Environ. Mi- 1378 N. HORINOUCHI et al. crobiol., 69, 3791–3797 (2003). technol. Biochem., 69, 784–789 (2005). 10) Tochikura, T., Production of a nucleotide-related sub- 16) Stumpf, P. K., A colorimetric method for the determi- stance by yeast and utilization of fermentation energy nation of deoxyribonucleic acid. J. Biol. Chem., 169, in the biosynthetic process. Hakko Kogaku Kaishi (in 367–371 (1974). Japanese), 56, 508–526 (1978). 17) Fujio, T., and Maruyama, A., Enzymatic production of 11) Tochikura, T., Kuwahara, M., Yagi, S., Okamoto, H., pyrimidine nucleotides using Corynebacterium ammo- Tominaga, Y., Kano, T., and Ogata, K., Fermentation niagenes cells and recombinant Escherichia coli cells: and metabolism of nucleic acid-related compounds in enzymatic production of CDP-choline from orotic acid yeast. J. Ferment. Technol., 45, 511–529 (1967). and choline chloride (part I). Biosci. Biotechnol. Bio- 12) Tochikura, T., Kariya, Y., Yano, T., Tachiki, T., and chem., 61, 956–959 (1997). Kimura, A., Study of the utilization of fermentation and 18) Ishii, K., and Shiio, I., Improved inosine production and respiratory energy in the biosynthetic process: I. Fer- derepression of purine nucleotide biosynthetic enzymes mentative production of phosphoric acid anhydrides and in 8-azaguanine resistant mutants of Bacillus subtilis. phosphate esters. Amino Acids Nucleic Acids (in Japa- Agric. Biol. Chem., 36, 1511–1522 (1972). nese), 29, 59–74 (1974). 19) Fruya, A., Kato, F., and Nakayama, K., Accumulation of 13) Harden, A., and Young, W., The alcoholic fermentation 6-azauridine by mutants of Brevibacterium ammonia- of yeast-juice. Proc. R. Soc. Lond. Ser. B, 77, 405–420 genes. Agric. Biol. Chem., 39, 767–771 (1975). (1905). 20) Ogata, K., The microbial production of nucleic acid- 14) Wakisaka, S., Ohshima, Y., Ogawa, M., Tochikura, T., related compounds. Adv. Appl. Microbiol., 19, 209–247 and Tachiki, T., Characteristics and efficiency of (1975). glutamine production by coupling of a bacterial gluta- 21) Takayama, S., McGarvey, G. J., and Wong, C. H., mine synthetase reaction with the alcoholic fermentation Microbial aldolases and transketolases: new biocatalytic system of baker’s yeast. Appl. Environ. Microbiol., 64, approaches to simple and complex sugars. Annu. Rev. 2953–2957 (1998). Microbiol., 51, 285–310 (1997). 15) Yamamoto, S., Wakayama, M., and Tachiki, T., The- 22) Fessner, W. D., and Helaine, V., Biocatalytic synthesis anine production by coupled fermentation with energy of hydroxylated natural products using aldolases and transfer employing Pseudomonas taetrolens Y-30 gluta- related enzymes. Curr. Opin. Biotechnol., 12, 574–586 mine synthetase and baker’s yeast cells. Biosci. Bio- (2001).