2017 年 23 巻 2 号 p. 349-353
1-Deoxynojirimycin (DNJ) is an effective sugar analogue that suppresses the elevation of postprandial blood glucose. Therefore, food products containing DNJ are expected to be promising functional foods in preventing the onset of diabetes. To examine the absorption kinetics of DNJ, we prepared 15N-labeled DNJ from the culture medium of a Bacillus amyloliquefaciens DSM7 strain. In the process, we found that lactose markedly increased the production of DNJ at a carbon/nitrogen ratio of 6.25:1, reaching 1140 mg/L with 5 days of cultivation. Comparisons between lactose and glucose on cell growth and DNJ production suggested that lactose made the DSM7 strain grow more slowly and maintained DNJ production in the long term. These findings may be useful for efficient microbial DNJ production or in the development of DNJ-containing fermented food products.
1-Deoxynojirimycin (DNJ), a potent alpha-glycosidase inhibitor, has the potential to prevent the onset of diabetes (Asano et al., 1994, 2000). It is also expected to be a clinically applicable lipid metabolism-improving agent (Nakagawa, 2013). DNJ was initially identified in mulberry leaves, and several animal and human research trials revealed that the administration of mulberry leaf extract suppressed the elevation of post-prandial blood glucose (Asai et al., 2011; Kimura et al., 2007; Vichasilp et al., 2012). At present, various food products containing DNJ are commercially available as functional foods in Japan and other countries, with mulberry leaves as the raw material. However, the low DNJ content in mulberry leaves (approximately 0.1% w/w) increases production costs. In addition, mulberry leaves are no longer available in sufficient quantities in Japan because of the considerable decline in Japanese silk production. Therefore, the application of DNJ-producing microorganisms is expected to decrease production costs and increase supplies.
The production of DNJ by Bacillus sp. and Streptomyces sp. has previously been reported (Ezure et al., 1985; Hardick et al., 1991, 1992; Hardick and Hutchinson, 1993; Kang et al., 2011; Onose et al., 2013; Paek et al., 1997; Schmidt et al., 1979; Stein et al., 1984; Yamaki et al., 2006). A gene cluster that initiates azasugar biosynthesis in Bacillus amyloliquefaciens was also identified (Clark et al., 2011; Kang et al., 2011). However, it is difficult to use microbially derived DNJ as food or food additives in Japan because of stringent food safety regulations. Mulberry leaves are recognized as a food product in light of their long consumption, unlike microbially derived DNJ. Therefore, extensive safety checks of DNJ derived from microorganisms are needed prior to its approved use as a food additive. In particular, whether microbially derived DNJ or its intermediates accumulate in specific organs (e.g., kidney or liver) should be examined in high-dose or chronic administration tests. To initiate these studies, we prepared large amounts of 15N-labeled DNJ by cultivating a B. amyloliquefaciens DSM7 strain in chemically defined media including 15N-ammonium sulfate as the sole nitrogen source. We found that lactose markedly increased DNJ production during the course of preparing the 15N-labeled DNJ. The contents of DNJ derivatives, deoxymannojirimycin (DMJ), nojirimycin (NJ), and 2-aminomannitol (ADM), were also examined in this study.
Chemicals Standard DNJ, DMJ, and NJ were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). The 15N-labeled ammonium sulfate (99.9% atom%) was purchased from SI Science Co., Ltd. (Saitama, Japan). We synthesized ADM (unpublished), and all other reagents were of the highest grade available.
Strain and preculture conditions B. amyloliquefaciens DSM7 (NBRC 15535) was obtained from the Biological Resource Center, National Institute of Technology and Evaluation (Chiba, Japan). Luria broth was used to maintain the culture stock and prepare the inoculum. DSM7 was precultured at 37°C in a test tube containing 3 mL of Luria broth with rotary shaking at 100 rpm for 8 h as the inoculum.
Production of DNJ by cultivation The composition of chemically defined media other than carbon sources and ammonium sulfate was previously described as minimal media (Onose et al., 2013). Carbon sources and ammonium sulfate were separately autoclaved to prevent a browning reaction. DSM7 was cultured at 37°C in 500 mL flasks with baffles containing 100 mL of chemically defined media with rotary shaking at 100 rpm. A 100 µL aliquot of the preculture medium per flask was added as inoculum. After 5 days of growth, the culture was centrifuged at 6,000 × g for 10 min to separate the cells from the culture supernatant. An equal volume of ethanol was added to the culture supernatant (to precipitate protein and polysaccharide) and centrifuged at 6,000 × g for 10 min. The supernatants were used for 15N-labeled DNJ preparation or estimation of DNJ contents.
Comparison of lactose and glucose as carbon sources for DNJ production Concentrations of glucose and lactose in the culture medium were measured using a glucose test kit II (Wako, Tokyo, Japan) or F-kit lactose/D-galactose (Roche Diagnostics GmbH, Mannheim, Germany). Samples for measuring ATP contents were prepared by disrupting centrifuged cell pellets with 2.5% trichloroacetic acid, in accordance with the manufacturer's instructions (ATP chemiluminescence kit “Kinshirow,” Toyo B-Net Co., Ltd., Tokyo, Japan).
Purification of 15N-labeled DNJ The 15N-labeled DNJ was purified following the method of Asano et al. (1994) with some modifications (Asano et al., 1994). Specifically, 4 L of ethanol was added to an equal volume of the culture supernatant and mixed well. After centrifugation at 10,000 × g for 10 min, the supernatant was collected and filtered through a GA-200 glass fiber filter (Toyo Roshi Kaisha, Ltd., Tokyo, Japan). Next, 8 L of the filtrate was applied to an Amberlite IR-120B column (Dow Chemical Co., Midland, MI; 800 mL resin, H+ form), and 0.5 N NH4OH elution was collected and applied to a Dowex 1X2 column (Dow Chemical Co.; 180 mL resin, OH− form). The pass-through fraction and water elution were collected and concentrated with a rotary evaporator. Methanol was added to the concentrate and recrystallized. The obtained crystal was used to confirm the contents of 15N-labeled DNJ, DMJ, NJ, and ADM by a hydrophilic interaction chromatography-tandem mass spectrometry (HILIC-MS/MS) analysis.
HILIC-MS/MS analysis of the culture medium or purified 15N-labeled DNJ A 5 µL aliquot of the culture medium or 15N-labeled DNJ was subjected to a HILIC-MS/MS analysis. The HILIC-MS/MS apparatus consisted of liquid chromatograph (Shimadzu, Kyoto, Japan) and 4000QTRAP MS/MS instruments (AB SCIEX, Tokyo, Japan). For the quantification of DNJ, NJ, and ADM, the MS/MS parameters were optimized with each standard preparation under positive ion electrospray ionization. DNJ (15N-labeled DNJ), ADM (15N-labeled ADM) and NJ (15N-labeled NJ) were determined using the selective reaction monitoring (SRM) mode as follows: DNJ (15N-labeled DNJ), m/z 164 (165) / 69 (69); ADM (15N-labeled ADM), m/z 182 (183) / 69 (69); NJ (15N-labeled NJ), m/z 180 (181) / 162 (163). The detection of DMJ followed the 9-fluorenylmethyl chloroformate (FMOC)-HPLC procedure as reported by Kim et al. (Kim et al., 2003). Briefly, the samples were derivatized with FMOC to differentiate the mobility between DNJ and DMJ followed by reversed-phase HPLC-fluorescence detection. The 1H-NMR spectra of 15N-DNJ were acquired on a Varian VNMRS-400 MHz spectrometer in deuterium oxide containing 1% 4,4-dimethyl-4-silapentane-1-sulfonic acid-d6 (DSS) (Sigma-Aldrich Japan, Tokyo, Japan).
Optimization of culture conditions for DNJ production To maximize the yield of 15N-labeled DNJ using chemically defined media, we examined the culture conditions. The B. amyloliquefaciens DSM7 strain was chosen because its complete genome sequence is available (Rückert et al., 2011), making it suitable for gene expression analysis in the future. Preliminary combined tests were performed to optimize three factors: the types of carbon sources and the concentrations of carbon and nitrogen sources in cultivation (Fig. 1). We have already reported that sorbitol and galactose markedly increased DNJ production when the B. amyloliquefaciens AS385 strain was cultured using 4% soybean peptone and 5% carbon sources as nutrients (Onose et al., 2013). On the other hand, lactose was the most effective in increasing DNJ production in this experiment. The effectiveness of lactose in DNJ production has already been reported for Streptomyces sp. SID9135 (Paek et al., 1997). The influence of the lactose and nitrogen source (ammonium sulfate) concentrations was examined in more detail (Fig. 2). When the lactose concentration was fixed at 2.5%, the optimum nitrogen concentration was approximately 0.3% – 0.4% (Fig. 2a). However, the influence of the nitrogen concentration was not so marked. In contrast, the quantities of lactose and ammonium sulfate at the same carbon/nitrogen (C/N) ratio (6.25:1) had a marked effect on DNJ production (Fig. 2b). Finally, the ideal concentrations of lactose and ammonium sulfate were determined to be 2.5% and 0.4%, respectively, which resulted in an optimal DNJ production of 1140 mg/L.
Combination of three factors (types of carbon sources and carbon and nitrogen concentrations) for DNJ production. Panels (a)–(e) show the concentrations of DNJ accumulated in the culture medium over 5 days. One flask containing 100 mL of the culture medium was used for each combination. White (1%), light gray (2.5%), dark gray (5%), and black (7.5%) bars represent the concentrations of each carbon source in chemically defined media.
Effects of carbon and nitrogen sources for DNJ production. Panel (a) shows the influence of ammonium sulfate concentration with a carbon source (lactose) concentration fixed at 2.5%. Panel (b) shows the influence of the quantities of nutrients at the same C/N ratio (6.25:1). Three flasks containing 100 mL of the culture medium were used for each condition. The bars indicate the averages ± standard deviation of DNJ concentration in culture medium after 5 d of cultivation.
Comparison between lactose and glucose on cell growth and DNJ production To obtain clues as to the cause of lactose-enhanced DNJ production, the growth and DNJ production of the DSM7 strain using 2.5% lactose or glucose as the carbon source was examined (Fig. 3). The ATP content in cells was measured as an indicator of the amount of living cells. Glucose was rapidly consumed in 24 h, and the amount of living cells (estimated from the ATP content) peaked at the same time. The culture medium became highly viscous at that time (data not shown), indicating that rapid bacteriolysis had occurred. DNJ concentration in the culture medium peaked from 24 h to 48 h, after which further DNJ production was aborted. The consumption of lactose was more moderate than that of glucose, and the living cells reached maximum at 72 h and the production of DNJ was maintained up to 120 h. These results suggested that lactose made the DSM7 strain grow more slowly and maintained DNJ production in the long term. When lower (1%) or higher (6.25%) concentrations of lactose were fed at the same C/N ratio (6.25:1), the DNJ accumulation was less than that of 2.5% lactose. These results suggested that appropriate nutrient quantities were essential to make lactose accelerate DNJ production by the DSM7 strain.
Comparison of lactose and glucose as carbon sources for DNJ production. Panel (a) shows the consumption of carbon sources, (b) ATP content in cells per 1 mL of culture medium, and (c) DNJ concentration in the culture medium. Three flasks containing 100 mL of the culture medium were used for each condition. White (glucose 2.5%), light gray (lactose 1%), dark gray (lactose 2.5%), and black (lactose 6.25%) bars indicate the averages ± standard deviation.
Preparation of 15N-labeled DNJ and purity evaluation The 15N-labeled DNJ was prepared using 15N-labeled ammonium sulfate as the sole nitrogen source. In a previous study, Streptomyces subrutilus was reported to produce DMJ (Ezure et al., 1989), whereas Bacillus did not appear to produce DMJ (Seo et al., 2013). Since DNJ and DMJ were not separated by HILIC-MS/MS, we derivatized the imino sugars with FMOC. The FMOC-derivatized DNJ (retention time: 4.85 min) and DMJ (5.14 min) were completely separated in the HPLC chart (data not shown). The HPLC analysis of FMOC-derivatized imino sugars showed that the B. amyloliquefaciens DSM7 strain did not produce DMJ (Fig. 4). From a food safety viewpoint, the existence of DMJ and other DNJ analogues should be minimal because those types of accessory products might cause unanticipated side effects. A small amount of ADM (retention time 4.52 min) was contained in the culture medium (Fig. 4). The HILIC-MS/MS analysis showed that the culture medium contained trace amounts of NJ, although its peak (retention time 5.47 min) was not observed. Sequential cation/anion exchange chromatography effectively purified 3.8 g of DNJ prepared from 4 L of the culture medium. An unknown peak (retention time 7.10 min) was also retained (data not shown). The 1H-NMR chart of purified 15N-labeled DNJ also indicated the presence of some impurities (Fig. 5). The purity of 15N-labeled DNJ (approximately 82%) was estimated using an internal standard. Further improvements are needed to minimize the concentrations of ADM and other impurities in the final products, by modifying culture conditions (e.g., the strain, nutrient composition, or culture temperature) or the purification process. However, the physiological function or health effects of ADM are unknown at present. Extensive safety checks of microbially derived DNJ are imperative. We are currently performing a kinetic analysis of 15N-labeled DNJ following oral administration in animal experiments.
The HPLC analysis of FMOC - derivatized imino sugars produced in the culture medium.
1H-NMR chart of purified 15N-DNJ. The noise peaks are indicated by arrows. Some signals derived from the internal standard (DSS) are marked with asterisks.
Acknowledgements We would like to thank Professor Shigefumi Kuwahara and M.S. Takafumi Hirokawa (Laboratory of Applied Bioorganic Chemistry, Graduate School of Agricultural Science, Tohoku University) for 1H-NMR analysis of 15N-labeled DNJ. Part of this study was supported by KAKENHI (C) (23580190, to T.K.) of the Japan Society for the Promotion of Science (JSPS), Japan.
