APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1982, p.

1064-1069

Vol. 43, No. 5

0099-2240/82/051064-06$02.00/0

Production of 2-Keto-L-Gulonic Acid from D-Glucose by TwoStage Fermentation

TAKAYASU SONOYAMA,* HIROYOSHI TANI, KATSUHIRO MATSUDA, BUNJI KAGEYAMA, MASAHIRO TANIMOTO, KOBEI KOBAYASHI, SHIGEO YAGI, HISASHI KYOTANI, AND KENJI MITSUSHIMA Department of Production, Shionogi and Company, Ltd., 192, Imafuku, Amagasaki, Hyogo 660, Japan

Received 16 September 1981/Accepted 30 December 1981

A practical method for the production of calcium 2-keto-L-gulonate (an intermediate in the Reichstein synthesis of L-ascorbic acid) from D-glucose has been established by using a two-stage fermentation system. D-Glucose was first converted to calcium 2,5-diketo-D-gluconate by a mutant strain of Erwinia sp. in a medium containing D-glucose, corn steep liquor, (NH4)2HP04, and CaCO3. After a 26-h cultivation, 328.6 mg of calcium 2,5-diketo-D-gluconate per ml was obtained, with a 94.5% yield from D-glucose. This broth was used directly for the next conversion without removal of cells by treatment with sodium dodecyl sulfate. The stereospecific reduction of calcium 2,5-diketo-D-gluconate to calcium 2-keto-L-gulonate was performed with a mutant strain of Corynebacterium sp. When the cell growth reached a maximum (about 16 h) in a medium containing Dglucose, corn steep liquor, NaNO3, KH2PO4, and trace elements, NaNO3 was added to the culture, and then the calcium 2,5-diketo-D-gluconate broth was fed over a period of about 50 h. Since the mutant strain requires a hydrogen donor for reduction, the calcium 2,5-diketo-D-gluconate broth was mixed with D-glucose before being fed. The results offour two-stage fermentations in 10-m3 conventional fermentors showed that an average of 106.3 mg of calcium 2-keto-L-gulonate per ml was obtained, with a 84.6% yield from D-glucose, the starting material of calcium 2,5-diketo-D-gluconate production. Calcium 2-keto-L-gulonate was stable in the broth. Neither 2-keto-D-gluconic acid nor 5-keto-D-gluconic acid was detected in the final broth.

It is well known that 2-keto-L-gulonic acid (2KLG) is readily converted to L-ascorbic acid (vitamin C) by a one-step chemical procedure in the Reichstein method (20). During the past 2 decades, considerable efforts have been made to find microbial methods for preparing 2KLG from various starting materials, including Dsorbitol, L-sorbose, L-sorbosone, and L-idonic acid (2, 3, 6, 7, 9, 15, 16, 19). Available information, however, indicates that neither of these methods has yet been used for the commercial production of 2KLG. Low productivity of 2KLG or difficulty in obtaining starting materials economically seems to be the reason for their impractical use (5). 2,5-Diketo-D-gluconic acid (25DKG) is produced effectively from D-glucose by Pseudomonas albosesamae, which was isolated by Wakisaka (17). In the course of studies on the use of 25DKG, our attention was directed toward finding microorganisms capable of converting 25DKG to 2KLG by stereospecific reduction at the C-5 position. Consequently, a number of microorganisms belonging to the coryneform bacteria were isolated from soil samples as the

most effective strains for 2KLG production. This discovery encouraged us to attempt the improvement of these strains and the development of a process based on the reduction of 25DKG because 25DKG is inexpensively produced from D-glucose. Detailed information concerning these results will be subjects for future reports. In this paper, we report a practical fermentation method, established by using 10-mi3 conventional stainless-steel fermentors, for the production of calcium 2KLG (Ca-2KLG) from Dglucose via calcium 25DKG (Ca-25DKG). The process consists of a two-stage fermentation: oxidative fermentation of D-glucose to Ca25DKG by a mutant strain of Erwinia sp., which was newly isolated from fruit, and then its reductive fermentation to Ca-2KLG by a mutant of Corynebacterium sp.
MATERIALS AND METHODS Chemicals. Ca-25DKG and calcium 5-keto-D-gluconate were obtained from D-glucose as the oxidation products of P. albosesamae and Acetobacter suboxydans (Shionogi), respectively, according to the meth-

1064

VOL. 43, 1982

ods described by Wakisaka (18) and Stubbs et al. (12). Calcium L-idonate was obtained from calcium 5-ketoD-gluconate as the reduction product of Brevibacterium ketosoreductum ATCC 21914 by the method of Sonoyama et al. (unpublished data). The 2KLG was prepared from the monohydrate of diacetone gulonic acid by acid hydrolysis. Calcium 2-keto-D-gluconate (Ca-2KDG; its free acid is referred to as 2KDG) was purchased from Nakarai Chemicals, Ltd. (Kyoto, Japan). For the assay of D-gluconic acid, gluconate kinase, P-NADP and 6-phosphogluconate dehydrogenase, and ATP were obtained from Boehringer Mannheim-Yamanouchi Co., Ltd. (Tokyo, Japan), Oriental Yeast Co., Ltd. (Tokyo, Japan), and Kyowa Hakko Kogyo Co., Ltd. (Tokyo, Japan), respectively. Other chemicals used for analyses were reagent grade. All components of the media used were commercial preparations, unless otherwise noted. Fermentation method for Ca-25DKG production. Strain SUS 2629001, a blocked mutant derived from Erwinia sp. SHS 2006 (ATCC 31626), was grown at 28C for 24 h on agar slants containing 0.5% glycerol, 0.5% yeast extract (Difco Laboratories, Detroit, Mich.), 0.3% peptone (Difco), 0.1% KH2PO4, 0.02% MgSO4-7H2O, and 2% agar. The preseed medium (600 ml in a 2-liter flask) and the seed medium (420 liters in a 1-m3 stainless-steel fermentor) consisted of 0.5% glycerol, 3% corn steep liquor, 0.1% KH2PO4, 0.01% p-2000 (Voranol-2120; The Dow Chemical Co., Freeport, Tex.) antifoam, and tap water. These media were adjusted to pH 6.8 before sterilization at 121C for 20 min. The preseed inoculum was prepared by inoculating a fresh agar slant of strain SHS 2629001 in the sterile medium and incubating at 28C for 20 h on a rotary shaker at 275 rpm. The entire contents of the five preseed flasks were added to the sterile seed medium, which was incubated at 28C for 10 h with an air flow rate of 0.42 m3/min (normal temperature and pressure) and agitation at 244 rpm. Ca-25DKG was prepared in a 10-m3 conventional stainless-steel fermentor. The fermentor medium was composed of (grams per liter): D-glucose, 58; corn steep liquor, 11.3; (NH4)2HP04, 5.6; CaCO3, 184; p2000 antifoam, 0.2; and tap water. D-Glucose was sterilized separately at 115C for 15 min. Corn steep liquor and CaCO3 were also sterilized separately at 121C for 20 min, whereas (NH4)2HP04 was sterilized by filtration. The pH of sterilized total medium was maintained at 6.8 by adjusting those of the medium components before sterilization. The fermentor (1.86 m3 of the medium) was inoculated with the entire contents of the 1-m3 seed fermentor and incubated at 28C. Agitation was maintained at 160 rpm, and air was sparged in at a rate of 3.6 m3/min (N.T.P.) at a tank pressure of 1.5 kg/cm2 (gauge). After 3 h of cultivation, a sterile 1-glucose solution was fed intermittently to the culture as described in Results. Bacterial growth was monitored by measuring optical density changes at 660 nm with a Hitachi spectrophotometer, model 139. After completion of Ca-25DKG fermentation, the broth was treated with sodium dodecyl sulfate as described in Results. Fermentation method for Ca-2KLG production. Ca2KLG was also produced in a 10-m3 conventional stainless-steel fermentor. Agar slants of strain SHS 752001, a blocked mutant derived from Corynebacterium sp. SHS 0007 (ATCC 31090), were prepared with

2KLG PRODUCTION

1065

medium containing 0.5% D-glucose, 0.5% yeast extract (Difco), 0.5% peptone (Difco), 0.1% KH2PO4, 0.02% MgSO4-7H2O, and 2.0% agar. They were grown at 28C for 40 h. The preseed medium (600 ml in a 2liter flask) consisted of 1% D-glucOse, 0.5% yeast extract (Difco), 0.5% peptone (Difco), 0.1% NaNO3, 0.1% KH2PO4, and 0.02% MgSO4-7H2O. The pH of the medium was adjusted to 7.0 before sterilization at 121C for 15 min. A shake flask was inoculated with a fresh slant of strain SHS 752001 and incubated at 28C for 24 h on a rotary shaker at 275 rpm. The seed medium (470 liters in a 1-m3 conventional stainlesssteel fermentor) contained 1.0% D-glucose, 2.0% corn steep liquor, 0.2% NaNO3, 0.1% KH2PO4, 0.02% MgSO4-7H2O, 0.005% p-2000 antifoam, and tap water. The pH of the medium was adjusted to 7.0 before sterilization at 121C for 20 min. The entire contents of 10 preseed flasks were transferred into the seed medium, which was incubated at 28C for 24 h with an air flow rate of 0.47 m3/min (N.T.P.) and agitation at 244 rpm. The fermentor medium consisted of (grams per liter): D-glucose, 20; corn steep liquor, 30; NaNO3, 3.45; KH2PO4, 0.67; p-2000 antifoam, 0.0167; trace elements, including ZnSO4 7H2O (4.9 mg), MnCl2-4H2O (0.8 mg), thiamine hydrochloride (0.22 mg), and calcium D-pantothenate (0.17 mg); and tap water. Trace elements were filter sterilized. D-Glucose was steam sterilized at 115C for 15 min. The rest of the medium was adjusted to a pH of 7.2 and steam sterilized at 121C for 20 min. The pH of sterilized total medium was approximately 6.9. The fermentor (4.23 m3 of the medium) was inoculated with the entire contents of the 1-m3 seed fermentor and incubated at 28C with agitation at 160 rpm, whereas air was admitted through a pipe sparger at a rate of 1.18 m3/ min (N.T.P.) and a tank pressure slightly over atmosphere. When the cell growth reached a maximum (16 h in Fig. 2), 9.7 kg of a sterile solution of 48.5% (wt/wt) NaNO3 was added to the culture, and then feeding of the Ca-25DKG broth was started as described in Results. Bacterial growth was measured by optical density in a manner similar to that of Ca-25DKG fermentation. Analytical methods. The Ca-25DKG content in both the Ca-25DKG and Ca-2KLG fermentation broths was determined with a Technicon Autoanalyzer by the following NH40H-HCl method devised by us. To 0.6 ml of suitably diluted sample (Ca-25DKG, 80 to 400 ,uM) of the fermentation broth, 0.6 ml of air, 0.6 ml of calcium chloride (12.0 mM), 0.32 ml of 1.5 N NH40H, and 0.34 ml of 4 N HCl were pumped per min and mixed in coils successively. The mixture (2.46 ml/min) was passed through a 28-ml-capacity coil (1.6-mm inside diameter) placed in a heating bath at 50C, and then the absorbance at 460 nm was measured in the flow cell. For the assay of Ca-2KLG, direct conversion of 2KLG to L-ascorbic acid was carried out in a reaction mixture containing 1 ml of suitably diluted sample (Ca-2KLG, 5 to 28 ,umol) of the fermentation broth with 3 ml of 8 N HCI. After incubation for 30 min at 90C, the reaction was stopped by adding 4 ml of 5 N NaOH, and the resulting solution was extracted with 3 ml of isoamyl alcohol. The amount of L-ascorbic acid remaining in the lower phase was then determined spectrophotometrically by a modified phospho-18tungstic acid method (1). The absorbance at 750 nm

1066

SONOYAMA ET AL.

APPL. ENVIRON. MICROBIOL.

feeding period of D-glucose

I 6
0.5
4-0

400

E
c-

'a)
0

300
0 0

conic acid, L-idonic acid, and other metabolic products. The trimethylsilylester derivatives of these compounds were prepared by the method of Sweely et al. (13). A 10-,u1 sample was injected into a coiled glass column (2,000 by 3-mm inside diameter), packed with 5% SE-52 on 80/100 Chromosorb W AW DMCS (General Electric Co., Waterford, N.Y.). Helium served as the carrier gas. The injection port and detector temperatures were 230 and 235C, respectively, and the initial column oven was set at 160C, followed by a 4C/min temperature increase to 210C. n-Heneicosane was used as an internal standard. Peaks were identified by comparison with authentic standards.

a
X

20
(0 CD

200 C)
0

'IO 10
a)

C\j
0

1000
0

cu
0.

3
LO

RESULTS Ca-25DKG production in a 10-m3 fermentor. The formation of D-gluconic acid, 2KDG, and 25DKG occurred simultaneously with D-glucose utilization during the cultivation. Figure 1 shows the results obtained when 14.4 kg of a sterile solution of 50% (wt/wt) D-glucose was fed at 4.5min intervals over a period of 12 h. The amounts of calcium D-gluconate and Ca-2KDG decreased

0,

1
0 10

30 20 Cultivation Time (h) FIG. 1. Time course of typical Ca-25DKG production by a mutant strain of Erwinia sp. in a 10-m3 conventional stainless-steel fermentor. Symbols: 0, optical density; 0, Ca-25DKG; A, Ca-2KDG; *, calcium D-gluconate (Ca-DG); A, pH.

measured. Ca-2KDG content in the Ca-25DKG fermentation broth was estimated as follows. The fermentation broth sample was divided into two portions. One portion of the sample was used to measure the Ca-25DKG concentration with the NH40H-HCl method. Another portion of the sample was subjected to a treatment similar to that described above for the Ca-2KLG assay, and the absorbance at 750 nm (Ca25DKG plus Ca-2KDG) was read. Ca-2KDG concentration was then determined by subtracting the Ca25DKG value (750 nm), which was obtained in comparison with a standard curve of Ca-25DKG concentrations, from the sample value (750 nm) and comparing the resulting value (750 nm) with a standard curve of Ca-2KDG concentrations. To ensure the accuracies of all assay systems used here, new standard curves were made up periodically. D-Gluconic acid was determined enzymatically by the method described by Mollering and Bergmeyer (8). D-Glucose was also measured enzymatically by using the Glucose-B-Test (Wako Pure Chemical Industries, Ltd., Osaka, Japan), based on a modified method of Sharp (10). A paper chromatographic system, according to the method described by Wakisaka (18), was used for the identification of 2KLG, 2KDG, 25DKG, 5-keto-Dgluconic acid, D-glucose, and other metabolic products. The cultures were also analyzed with a Shimazu gas chromatograph (model GC-5A) and a thermal conductivity detector for the identification of D-gluwas

rapidly after completion of D-glucose feeding, whereas the accumulation of Ca-25DKG continued and reached a peak, 328.6 mg/ml at 26 h of cultivation. The yield of 94.5 mol% Ca-25DKG was obtained from the total amounts of Dglucose added. In the final broth, 2.5 mg of calcium D-gluconate per ml remained unconverted, whereas D-glucose and Ca-2KDG were not detected. The pH, which was not regulated, began at 7.1, dropped to 4.7 at 3 h of cultivation, and finally reached 3.9. The volume of the broth was increased from 2.28 to 4.2 m3 by feeding of D-glucose solution. Precipitation of calcium Dgluconate, Ca-2KDG, or Ca-25DKG was not observed during the fermentation. No 5-keto-Dgluconic acid was detected by means of paper chromatography throughout the fermentation. Treatment of Ca-Z5DKG broth with sodium dodecyl sulfate. The Ca-25DKG broth (4.2 m3) was mixed with 7.27 kg of a sterile solution of 14.4% (wt/wt) sodium dodecyl sulfate and maintained at 28C for 6 h with mild agitation at 80 rpm. By this treatment, the total viable cell was decreased from 5 x 109 to 9 x population 10 cells per ml. Finally, the broth was cooled and maintained at approximately 10C to minimize the decomposition of Ca-25DKG. The Ca25DKG broth thus treated was used directly as the source of Ca-25DKG for the next conversion, without removal of cells. Ca-2KLG production in a 10-m3 fermentor. The mutant strain SHS 752001 requires a hydrogen donor for the reduction of 25DKG to 2KLG. Figure 2 presents the data obtained when Dglucose was used as the hydrogen donor. The Ca-25DKG broth prepared in the preceding stage was mixed with a sterile solution of 50% (wt/wt) D-glucose in the proportion of 0.2 kg of

VOL. 43, 1982

2KLG PRODUCTION

1067

feeding period of Ca-25DKG broth

9_
I

8
6

Z/ a&

120 '
E

100

'a
a)
0

80
cO

-0
L-

a J

60
40
5

E co
Lo)
C\J

cm

-10
co
c

C6 o

C6

0)

cs

5-keto-D-gluconate, and Ca-2KDG were not detected by means of paper and gas chromatography in the fermentation broth over the period of Ca-25DKG feeding. Total accumulated Ca-2KLG in the broth was isolated by K. Tokuyama, M. Kiyokawa, and T. Sonoyama as crystals in the form of the monohydrate of its sodium salt with purities of >99% (unpublished data). A portion of these crystals was recrystallized from distilled water and submitted to physicochemical analyses by M. Sawada (unpublished data). The melting point was 145.5C (decomposition), and the optical rotation was [oa]D22 = -25.60 (c = 2.0, water). Elemental analysis corresponds to C6H907Na H2O; calculated: C, 30.78; H, 4.77; Na, 9.82. Found: C, 30.80; H, 4.77; Na, 9.88. The infrared spectrum of the crystals in a KBr disk was identical to that of an authentic sample.
DISCUSSION Conversion of D-glucose to 25DKG via Dgluconic acid and 2KDG was first proved by Katznelson et al. (4) with Acetobacter melanogenus MA 6.2 and later by Wakisaka (18) with P. albosesamae. Recently, Stroshane and Perlman (11) reported that A. melanogenus ATCC 9937 converts D-glucose to 25DKG via D-gluconic acid and 5-keto-D-gluconic acid. Our strain SHS 2629001, a mutant strain derived from Erwinia sp., accumulated D-gluconic acid and 2KDG as intermediate products (Fig. 1). No 5-keto-Dgluconic acid could be detected in the broth throughout the cultivation. This observation suggests that strain SHS 2629001 converts Dglucose to 25DKG in a pathway similar to those of A. melanogenus MA 6.2 and P. albosesamae. The high-yield production of 25DKG from Dglucose (>90%) has already been reported with relatively low accumulation of 25DKG (about 90 mg/ml), although at shake flask scale (11, 18). In this work the production of large amounts of 25DKG was demonstrated with 10-mi3 fermenTABLE 1. Ca-2KLG production in fermentorsa
Parameter

3
C._

20- 4 2E
2
E

0,~ 1

0: 0.

O60 20 40 Cultivation Time (h)

Oa:

FIG. 2. Time course of typical Ca-2KLG production by a mutant strain of Corynebacterium sp. in a 10m3 conventional stainless-steel fermentor. Symbols: 0, optical density; 0, Ca-2KLG; 0, Ca-25DKG; A, pH.

D-glucose to 1 kg of 25DKG before being fed. The feeding of the Ca-25DKG broth was started at 18 h of cultivation and continued over a period of 48 h. The feed rate of the Ca-25DKG broth was adjusted to maintain the concentration of *Ca-25DKG in the Ca-2KLG fermentation broth below 5 mg/ml. The accumulation of Ca-2KLG began immediately after the Ca-25DKG broth was fed and finally reached 106.5 mg/ml (97.0 mg/ml as free acid). In the final broth, however, 1.6 mg of Ca-25DKG per ml remained unconverted. After completion of Ca-25DKG broth feeding, neither an increase nor a decrease of Ca-2KLG content in the broth was observed. The yield of Ca-2KLG from Ca-25DKG utilized was 92.5 mol%, whereas the overall yield of Ca2KLG from D-glucose, the starting material of Ca-25DKG production, was 86.0 mol%. Total volume of the broth increased gradually throughout the period of feeding from 4.7 to 7.43 mi3. The pH, which was not regulated, dropped from 8.3 to 6.1, whereas the cell population was held constant at around 22 (optical density at 660 nm). Results of four two-stage fermentations are shown in Table 1. In all cases, D-glucose, calcium L-idonate, calcium D-gluconate, calcium

10-m3

Accumulated
(mg/ml)

Ca-2KLG yield from

D-glucose (mol%) 86.0 84.6 82.9

106.9 Maximum 106.3 Mean 105.1 Minimum a Values were obtained in four two-stage fermentations. Fermentations were carried out in the cultural conditions described in the text. The Ca-25DKG broths were fed over a period of 50 h. b D-Glucose was the starting material of Ca-25DKG

production.

1068

SONOYAMA ET AL.

APPL. ENVIRON. MICROBIOL.

tors. Strain SHS 2629001 converted the added Dglucose to 25DKG with a yield of 94.5 mol% and

accumulated 328.6 mg of Ca-25DKG (297.6 mg/ ml as free acid) per ml in low-cost corn steep liquor medium (Fig. 1); the time required for this fermentation was 26 h. This high productivity was obtained by maintaining the dissolved-oxygen concentrations in the culture broth above zero throughout the fermentation and implementing timely feeding of D-glucose. The 25DKG is unstable, especially to heat and in alkaline solution (4). Our early attempts to isolate Ca-25DKG from the broth resulted in significant losses (more than 30%) of Ca25DKG. This phenomenon of the instability of 25DKG was also observed by Stroshane and Perlman (11) in their studies on the isolation of 25DKG from broth. Thus, the elimination of the isolation process of Ca-25DKG, which is an additional advantage in industrial applications, was important to the success of this work. Of about 20 chemical agents tested, sodium dodecyl sulfate was selected as the most effective one having bactericidal activity against SHS 2629001 (a gram-negative rod). The use of the 25DKG broth treated with sodium dodecyl sulfate (approximately 250 ,ug/ml) did not inhibit the conversion of Ca-25DKG to Ca-2KLG and rather enhanced Ca-2KLG production in comparison with the use of the Ca-25DKG powder (data not shown). The microbial reduction of 5-keto-D-gluconic acid to L-idonic acid was found first by Takagi (14) with Fusarium species and later by us (unpublished data) with B. ketosoreductum ATCC 21914. However, the selective and stereospecific reduction of 25DKG to 2KLG by microorganisms, described here, is a new finding to our knowledge. The mutant strain SHS 752001 could utilize various carbohydrates and low-cost molasses as hydrogen donors (data not shown). In this report we present the results obtained when D-glucose was used as the donor (Fig. 2; Table 1). All of four two-stage fermentations in 10-m3 fermentors resulted in excellent levels of Ca-2KLG production (Table 1): an average of 106.3 mg of Ca-2KLG (96.8 mg/ml as free acid) per ml was accumulated, with an 84.6 mol% yield from Dglucose, the starting material of Ca-25DKG production. The standard deviations from the mean were 0.8% for the accumulation and 1.5% for the yield. Thus, the reliability and reproducibility of the above fermentation method for Ca-2KLG production, both essential for the commercial production, were also demonstrated. Since the Ca-25DKG yield from D-glucose, the Ca-2KLG yield from Ca-25DKG utilized, and the overall yield of Ca-2KLG from D-glucose were 94.5, 92.5, and 86.0% (Fig. 1 and 2), respectively, the

TABLE 2. Ca-2KLG production (means) in 30-liter, 1-rm3, and 10-m3 fermentorsa Ca-2KLG yield from Accumulated Fermentor Ca-2KLG

(mg/ml)

D-glucose (mol%)b

83.6 108.3 30 liters 84.3 105.4 1 m3 84.6 106.3 10 m3 a Fermentations were carried out in the cultural conditions described in the text. The Ca-25DKG broths were fed over a period of 50 h. b D-Glucose was the starting material of Ca-25DKG

production.

decomposition of Ca-25DKG during the periods of the treatment with sodium dodecyl sulfate and the storage for feeding was estimated to be about 1.5%. Although the mutant strain SHS 752001 is strictly aerobic, low aeration rates below 0.25 vol/vol per min (N.T.P.) were suitable for Ca2KLG production. The onset of feeding of the Ca-25DKG broth was most favorable at the early stationary phase of cell growth, indicating the optimal phase of the overflow of reducing powers such as NAD(P)H to reduction of 25DKG in the cell's metabolism. For economic considerations, it may be noted here that these results of Ca-2KLG production were obtained with commercially feasible media for both Ca-25DKG and Ca-2KLG fermentation. In addition, the stability of Ca-2KLG allowed more flexibility in the isolation of the product. Furthermore, the absence of D-glucose, calcium D-gluconate, calcium L-idonate (the isomer of Dgluconate), calcium 5-keto-D-gluconate, and especially Ca-2KDG (the isomer of 2KLG and readily converted to isoascorbic acid in the same chemical procedure as that of L-ascorbic acid synthesis from 2KLG) in the final broth also allowed us to simplify the isolation process of 2KLG from the broth. Similar results on the kinetics of cell growth and Ca-2KLG production were obtained with 30-liter, 1-m3, and 10-m3 fermentors (Table 2), possibly indicating easy scale-up to industrialscale production. In conclusion, the method described here will provide a new way of commercially producing 2KLG, an intermediate in the synthesis of Lascorbic acid.
ACKNOWLEDGMENTS We thank Tahiko Honjo, former director of our department, for his continued interest and encouragement during the course of this work. We are also grateful to Shuichi Aiba, Osaka University, for valuable discussions. LITERATURE CITED 1. Fugita, A., and T. Ebihara. 1937. Kolorimetrische Bestimmung von Vitamin C mittels Phospho-18-wolframsaure. Biochem. Z. 290:182-191.

VOL. 43, 1982

2. Isono, M., I. Nakanishi, K. Sasajima, K. Motizuki, T. Kanzaki, H. Okazaki, and H. Yoshino. 1968. 2-Keto-Lgulonic acid fermentation. I. Paper chromatographic characterization of metabolic products from sorbitol and Lsorbose by various bacteria. Agric. Biol. Chem. 32:424431. 3. Kanzaki, T., and H. Okazaki. 1970. 2-Keto-L-gulonic acid fermentation. IV. L-Sorbose metabolism in Pseudomonas aeruginosa. Agric. Biol. Chem. 34:432-436. 4. Katznelson, H., S. W. Tanenbaum, and E. L. Tatum. 1953. Glucose, gluconate, and 2-keto-gluconate oxidation by Acetobacter melanogenum. J. Biol. Chem. 204:43-59. 5. Kulhanex, M. 1970. Fermentation processes employed in Vitamin C synthesis. Adv. Appl. Microbiol. 12:11-33. 6. Makover, S., G. B. Ramsey, F. M. Vane, C. G. Witt, and R. B. Wright. 1975. New mechanisms for the biosynthesis and metabolism of 2-keto-L-gulonic acid in bacteria. Biotechnol. Bioeng. 17:1485-1514. 7. Martin, C. K. A., and D. Perlman. 1976. Conversion of Lsorbose to 2-keto-L-gulonic acid by mixtures of immobilized cells of Gluconobacter melanogenus IFO 3293 and Pseudomonas species. Eur. J. Appl. Microbiol. 3:91-95. 8. Mollering, H., and H. U. Bergmeyer. 1974. D-Gluconate, p. 1243-1247. In H. U. Bergmeyer (ed.), Methods of enzymatic analysis, vol. 3, 2nd English ed. Academic Press, Inc., New York. 9. Okazaki, H., T. Kanzaki, K. Sasajima, and Y. Terada. 1969. 2-Keto-L-gulonic acid fermentation. III. Evaluation of the pathway of sorbitol metabolism in Gluconobacter melanogenus. Agric. Biol. Chem. 33:207-211. 10. Sharp, P. 1972. Interference in glucose oxidase-peroxidase blood glucose methods. Clin. Chim. Acta 40:115120. 11. Stroshane, R. M., and D. Perlman. 1977. Fermentation of glucose by Acetobacter melanogenus. Biotechnol. Bioeng. 19:459-465.

2KLG PRODUCTION

1069

12. Stubbs, J. J., L. B. Lockwood, E. T. Roe, B. Tabenkin, and G. E. Ward. 1940. Production of ketogluconic acids from glucose. Ind. Eng. Chem. 32:1626-1631. 13. Sweely, C. C., R. Bentley, M. Makita, and W. W. Wells. 1963. Gas-liquid chromatography of trimethylsilyle derivatives of sugars and related substances. J. Am. Chem. Soc. 85:2497-2507. 14. Takagi, Y. 1962. Reduction of 5-keto-D-gluconic acid to Lidonic acid by Fusarium species. Agric. Biol. Chem. 26:717-718. 15. Tengerdy, R. P. 1961. Redox potential changes in the 2keto-L-gulonic acid fermentation. II. Relationship between redox potential and product formation. J. Biochem. Microbiol. Technol. Eng. 3:255-260. 16. Tsukada, Y., and D. Perlman. 1972. The fermentation of L-sorbose by Gluconobacter melanogenus. I. General characteristic of the fermentation. Biotechnol. Bioeng. 14:799-810. 17. WakLsaka, Y. 1964. Carbohydrates oxidation by Pseudomonas nov. sp. I. Taxonomic studies on a newly isolated bacterium, Ps. albosesamae nov. sp. Agric. Biol. Chem. 28:369-374. 18. Wakisaka, Y. 1964. Carbohydrates oxidation by Pseudomonas albosesamae nov. sp. II. Chemical structure of the product on glucose and gluconate oxidation. Agric. Biol. Chem. 28:819-827. 19. Yamazaki, M. 1953. Production of Vitamin C by fermentation. II. Oxidation of Ca-L-idonate and Ca-D-gluconate by Pseudomonas sp. and isolation of the products. J. Ferment. Technol. 31:86-90. 20. Yamazaki, M. 1954. Production of Vitamin C by fermentation. III. The formation and isolation of 2-keto-L-gulonic acid from Ca-L-idonate and Ca-D-gluconate mixture by the submerged fermentation, and synthesis of Vitamin C from 2-keto-L-gulonic acid. J. Agric. Chem. Soc. Jpn. 28:890-894.