Fortification of Mucilage and GABA in Hovenia dulcis Extract by Co-fermentation with Bacillus subtilis HA and Lactobacillus plantarum EJ2014

Abstract

Hovenia dulcis Thunb. extract (HDE) was co-fermented by Bacillus subtilis HA and Lactobacillus plantarum EJ2014. The co-fermentation was optimized by adding glucose (3% for the first fermentation, 1.5% for the second fermentation), MSG (5%), and skim milk (5%). The HDE broth fermented by B. subtilis HA showed pH 7.3, 0.09% acidity, 0.52 mg/g tyrosine equivalents, and viable bacterial counts of 1.4×10⁹ CFU/mL. The viscous broth obtained indicated a consistency index of 3.90 Pa.sⁿ and 32.70 mg/mL mucilage after 2 days. Subsequently, after the second fermentation, the broth indicated pH 4.5, 1.08% acidity, 0.87 mg/g tyrosine equivalents and viable bacterial counts of 2.0×10⁶and 3.0×10⁹ CFU/mL for B. subtilis HA and L. plantarum EJ2014, respectively. The MSG substrate was effectively converted into GABA, which showed a final concentration of 16.08 mg/mL in the broth. Therefore, the co-fermented HDE was effectively fortified with mucilage, GABA, peptides, and probiotics, and could be used as a functional ingredient for nutraceuticals.

Introduction

Hovenia dulcis Thunb., also known as Oriental raisin, is a pseudofruit used as a food supplement and traditional medicine against liver diseases and alcoholic intoxication in its native East Asia (Hyun et al., 2010). H. dulcis extract has been shown to exert anti-inflammatory effects on liver cells (Yao et al., 2017) as well as blood sugar-lowering effects (Yang et al., 2002), suggesting that it could be used in the treatment of alcoholic liver disease and diabetes. Furthermore H. dulcis has demonstrated to be effective for relieving alcoholic intoxication as it lowers blood alcohol levels (Hyun et al., 2010) and could be useful to treat hangovers. Flavonoids including hovenitins as well as polysaccharides are present in the H. dulcis pseudofruit (Yoshikawa et al., 1997; Wang et al., 2012) and are likely to contribute to the observed bioactivities.

Fermentation of plant products can enrich them with bioactive compounds, enhancing their health benefits (Lim et al., 2016). Solid-state fermentation with Bacillus subtilis produces mucilage that is mainly composed of the edible biopolymer γ-polyglutamic acid (γ-PGA), which is responsible for the sticky consistency of natto. γ-PGA exhibits immune-boosting, anticancer, and antihypertensive activities (Kim et al., 2014; Lee et al., 2014), and is also biodegradable and non-toxic, having applications in the food, medicine, and wastewater industries (Ogunleye et al., 2015). γ-PGA has been efficiently produced in mucilage by solid-state fermentation using B. subtilis and L. plantarum (Park et al., 2012; Lim et al., 2016). Thus, alkaline fermentation could be applied to enhance the content of this metabolite in natural products.

γ - Aminobutyricacid (GABA) is one of the major neuroinhibitors and also exhibits hypotensive and anticancer effects; it has been used for treating sleeplessness, depression and for stimulating immune cells (Joye et al., 2011; Dhakal et al., 2012). GABA is synthesized by lactic acid bacteria (LAB) as an alternative source of energy and as defense against acid environments (Kook et al., 2010). In addition to their fermentative action, LAB are beneficial as probiotics, improving gut health, boosting immunity, and reducing the incidence of diarrhea among other effects (Parvez et al., 2006). Thus, various products such as tea, rice, and herbs have been fermented with LAB to enrich them with both GABA and probiotics (Watanabe et al., 2011; Kwon et al., 2016).

In this study, the co-fermentation of H. dulcis extract (HDE) was optimized to produce both mucilage and GABA from Monosodium glutamate (MSG) as a substrate. The aim was to produce an HDE with enhanced health benefits, which would not only contain the bioactive compounds from the plant extract, but would also be enriched with functional compounds from the fermentation. The fermented product was analyzed to determine the production of bioactive compounds and probiotics.

Materials and Methods

Materials    Hovenia dulcis extract (HDE) was purchased from Cheonginuseol Food Co. (Gapyeong, Korea). MSG and glucose were obtained from Yakuri Pure Chemicals (Kyoto, Japan). Skim milk was purchased from Seoul Milk ICA (Seoul, Korea). All chemicals used were of analytical grade. Lactobacillus plantarum EJ2014 was isolated from rice bran and deposited in the Korean Culture Collection Center (KCCM 11545P), and Bacillus subitilis HA was isolated from fermented soybean and deposited with code KCCM 10775P.

Analytical assays    The physicochemical properties of the HDE were determined using a standard methodology (Bradley, 2010). The mineral content was determined using inductively coupled plasma optical emission spectrometry (ICP-OES; Optima 700DV, Perkin Elmer, Waltham, MA, USA). As the pretreatment, the HDE sample was heated at 550–600°C for 12 h, solubilized with 10 mL of HCl, and filtrated to obtain a soluble fraction. The concentrations were calculated using a standard curve.

Co-fermentation process    The HDE was sterilized by heating at 121°C for 15 min, and was then mixed with sterilized 30% MSG and 30% glucose solutions, adjusting the final concentrations to 5% MSG and 3% glucose, respectively. The starter of B. subtilis HA was grown in nutrient broth (NB), cells were harvested and suspended in distilled water in a 1:10 ratio. The starter culture was inoculated at a 5% concentration and incubated by shaking (160 rpm) at 42°C for 3 days. Subsequently, for the second fermentation, glucose and skim milk were added at concentrations of 1.5% and 5%, respectively. The L. plantarum EJ2014 starter, grown in MRS broth, was inoculated at a 1% concentration, followed by incubation at 30°C for 7 days.

pH and acidity    The pH of the fermented broth was determined by pH-meter (Model 420+, Thermo Orion, Beverly, MA, USA). Titratable acidity (%, w/v) was determined as lactic acid by measuring the amount of 0.1 N NaOH required to reach a pH of 8.3.

Mucilage and consistency    The fermented broth (5 mL) was centrifuged at 15,000 rpm for 20 min. The supernatant was mixed with twice the volume of isopropanol, and the precipitate was washed with 95% ethanol and dried at 50°C for 24 h to yield crude mucilage.

The consistency of the fermented broth was determined using a rheometer (HAKKE RheoStress 1, Karlsruhe, Germany) fitted with a cone plate device (Plate PP35Ti, diameter 3.5 cm, 1.0 mm gap) at 25°C. The sample (1 mL) was loaded between the plate and the cone plate device.

Viable bacterial counts    To determine viable cells of B. subtilis, the HDE broth was serially diluted with sterilized water. Subsequently, 20 µL of diluted culture from the first fermentation was plated onto MRS agar plates and cultured at 42°C for 24 h to yield colony forming units (CFU). To selectively determine the viable bacterial counts of L. plantarum EJ2014 after the co-fermentation, the MRS agar plate with mixed strains was cultured at 30°C for 24 h to delay the growth of B. subtilis (Kook and Cho, 2013), and the CFU were determined.

Peptide content    A modified Anson's method was applied. Briefly, the fermented broth (4 mL) was diluted 10-fold using distilled water and centrifuged at 13,000 rpm for 15 min. The supernatant was mixed with 0.7 mL of 0.44 M trichloroacetic acid (TCA) to stop the reaction and incubated at 37°C for 30 min. The precipitate was removed by centrifugation at 13,000 rpm for 10 min and the supernatant was mixed with 2.5 mL of 0.55 M Na₂CO₃ and 0.5 mL of Folin-phenol reagent. After incubating at 37°C for 30 min, absorbance was measured using a spectrophotometer (Ultrospec^®2100 pro, Amersham Biosciences, Piscataway, NJ, USA) at 660 nm. The peptide content in the supernatant was determined using tyrosine as a standard.

GABA and amino acid analysis    Thin-layer chromatography was performed to monitor the production of GABA from the glutamate precursor. The viscous co-fermented broth was diluted 3-fold and spotted on a TLC plate (60 F₂₅₄, Merck KGaA, Darmstadt, Germany). The developing solvent was n-butanol: glacial acetic acid: distilled water (3: 1: 1, v/v/v). The TLC plate was dried in an oven at 50°C, sprayed with a 0.2% ninhydrin solution, and developed in an oven at 105°C for 3 min or until the spots appeared clearly. The GABA and glutamic acid in the co-fermented broth were determined by high-performance liquid chromatography (HPLC).

The γ-PGA content in crude mucilage was determined as the total glutamic acid. The crude mucilage was hydrolyzed in HCl. To determine the free amino acids, the culture broth (30 µL) was completely dried and then reacted with 20 µL of a phenylisothiocyanate (PITC) solution (MeOH:H₂O:TEA:PITC=7:1:1:1) at room temperature for 30 min. Free amino acids were analyzed using a Hewlett Packard 1100 Series HPLC (Palo Alto, CA, USA) fitted with a C18 column (Waters Nova-Pak 4 µm, Milford, MA, USA). The mobile phase was eluted at a flow rate of 1 mL/min in gradient mode, starting with 100% Solvent A (140 mM sodium acetate, 0.1% triethanolamine, 6% CH₃CN; pH 6.1) and ending with 100% Solvent B (60% CH₃CN), during 25 min. The elution was monitored with absorbance at 254 nm.

Statistical analysis    Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS Version 20.0, SPSS Inc. Chicago, IL, USA). The mean and standard error were calculated. Significant differences among means were determined by one way-ANOVA and Duncan's multiple range test (p < 0.05).

Results and Discussion

Physicochemical properties of the HDE    The HDE exhibited a dark color and slightly sour typical herbal taste. The physicochemical properties of the HDE are summarized in Table 1. The HDE indicated a pH of 5.85, 0.16% acidity, soluble solid content of 21.10 mg/g, and it also contained various minerals, among which K, P, Ca, and Mg predominated.

Table 1. Physicochemical properties of the Hovenia dulcis extract

Component		Value
pH		5.85±0.0
Acidity (%, w/v)		0.16±0.0
Soluble solid content (mg/g)		21.10±0.45
Reducing sugar (mg/g)		8.53±0.0
Polyphenols (mg/g)		0.86±0.11
Flavonoids (mg/g)		7.21±0.21
Minerals (µg/g)	Na	17.95
	Ca	48.52
	K	89.75
	Mg	37.37
	P	50.42
	Fe	1.42
	Mn	0.62

The growth of microorganisms is correlated with the availability of minerals, particularly Na, Mg, Fe, Ca, and Mn. A complex of mineral salts exerted a positive effect on the biomass accumulation of Bacillus thuringiensis (Drehval et al., 2003). Thus, the various minerals present in the HDE may provide the mineral nutritional requirement for microbial growth during the fermentation.

Viable bacterial counts after the first alkaline fermentation    The effect of glucose concentration on the viable bacterial counts of the HDE broth fermented by B. subtilis is shown in Table 2. The HDE proved to be a suitable medium for the fermentation as B. subtilis HA grew well even without glucose addition, although supplementation with this carbon source was necessary to produce γ-PGA and GABA. The fermented HDE is likely to retain its bioactivities as flavonoids and complex polysaccharides are not used as nutrients by the bacterial strains used, and the enhanced hepatoprotective effects of fermented HDE have been reported (Choi et al., 2014; Wang et al., 2016).

Table 2. Effect of glucose content on the physicochemical properties and viable bacterial counts of HDE broth fermented by B. subtilis HA. Values are mean ± SD (n > 3); different letters in the same column indicate significant differences by Duncan's multiple range test (p < 0.05)

Glucose concentration (%)	Fermentation time (days)	pH	Acidity (%, w/v)	Consistency index (Pa·sn)	n value of (Pa·sn)	Mucilage content (mg/mL, w/v)	Viable cell counts (CFU/mL, ×106)
0	0	5.74±0.01d	0.21±0.02a	0.00±0.00c	0.89	0.00±0.00e	35±7.6c
	1	7.74±0.01c	0.05±0.02b	0.01±0.00ab	0.84	0.00±0.00e	230±86b
	2	8.18±0.01b	0.04±0.01b	0.01±0.00bc	0.85	0.00±0.00e	400±8.7a
	3	8.51±0.02a	0.00±0.00c	0.01±0.01a	0.79	0.00±0.00e	230±1.4ab
3	0	5.72±0.02c	0.23±0.03a	0.00±0.00d	0.81	0.00±0.00d	35±7.6b
	1	6.41±0.01b	0.09±0.00b	0.52±0.01c	0.56	14.97±0.35c	1400±58a
	2	7.55±0.08a	0.10±0.02b	3.90±0.04a	0.27	32.70±0.48a	1700±144a
	3	7.30±0.11a	0.09±0.03b	3.34±0.01b	0.28	29.64±0.44b	1325±104a
5	0	5.72±0.01d	0.23±0.03a	0.00±0.00d	0.78	0.00±0.00c	35±7.6c
	1	6.51±0.01c	0.11±0.01b	0.44±0.04c	0.56	15.84±0.05b	2250±212b
	2	6.60±0.00a	0.09±0.03b	0.69±0.01b	0.57	25.49±0.08a	2825±530a
	3	6.57±0.01cb	0.08±0.01b	0.52±0.08a	0.56	23.07±0.21a	1750±304c

The alkaline fermentation was performed in the presence of 5% MSG as this concentration yielded the highest amount of mucilage. The viable bacterial counts tended to increase as the concentration of glucose increased, indicating that glucose was effectively used as a nutrient for growth. After 1 day of fermentation, the culture broth containing 3% and 5% glucose exhibited viable bacterial counts above 1×10⁹ CFU/mL, which increased to 2.82×10⁹ CFU/ mL after 2 days in the broth containing 5% glucose, although the number of bacteria decreased afterward. Bacterial numbers remained high and fairly stable in the broth containing 3% glucose. By contrast, the number of bacteria in the broth lacking glucose remained lower, reaching 3.95×10⁸ CFU/mL after 2 days.

pH and acidity of the HDE after the first fermentation    The pH and acidity of the HDE broth fermented by B. subtilis for 3 days is shown in Table 2. The pH of the HDE without glucose notably increased during the fermentation, showing the highest value (8.51) after 3 days. By contrast, the HDE with 3% and 5% glucose exhibited a gradual pH increase, indicating 7.30 and 6.57, respectively, after 3 days. Likewise, the acidity of the HDE containing glucose exhibited a gradual decrease during the fermentation, whereas that of the HDE without glucose showed an abrupt decrease from 0.21% at the start to 0% acidity at the end of the alkaline fermentation. The lowered pH in the broths containing glucose reflect the increase in the number of bacteria and suggest that glucose boosted microbial growth, thus increasing mucilage production (Ogawa et al., 1997). In addition, even though the fermentation with B. subtilis is an alkaline fermentation, this bacterium is also capable of producing lactate; thus, increased bacterial counts would result in more lactate production and a decreased pH (Cruz Ramos et al., 2000).

Mucilage and consistency of the HDE after the first fermentation    The effect of glucose on the mucilage production and consistency index of the fermented HDE is shown in Table 2. No mucilage production was observed in the HDE broth fermented without glucose, and thus its consistency index was very low. By contrast, the HDE with 3% glucose exhibited the highest amount of mucilage, indicating 32.7 mg/mL mucilage and a consistency index of 3.90 Pa·sⁿ after 2 days of fermentation, although further fermentation resulted in slightly lowered values. In addition, HPLC analysis revealed that the crude mucilage had a γ-PGA content of 595 mg/g. The amount of mucilage produced here was greater than that obtained by fermenting a defined medium (Kim et al., 2014). The content and composition of the mucilage produced in an alkaline fermentation depends on the type of strain and nutrient culture conditions (Kunioka and Goto, 1994). For instance, the mucilage of natto contains a mixture of γ-PGA and fructan produced by Bacilllus subtilis (Candela and Fouet, 2006). Nevertheless, mucilage with a high content of γ-PGA was produced in a defined medium containing limited nutrients and MSG (Seo et al., 2008). Here, the HDE containing 5% glucose exhibited an intermediate mucilage content of 25.5 mg/mL after 2 days of fermentation, although its consistency was substantially lower (0.69 Pa·sⁿ) than that obtained using 3% glucose. These results indicate that the mucilage production during the alkaline fermentation of the HDE was dependent upon the concentration of glucose. As described in a previous report (Chen et al., 2010), glucose is a necessary source of energy for the production of γ-PGA by B. subtilis; thus, the addition of glucose resulted in an increased mucilage content and consistency index, although excess glucose had a downregulating effect.

It is likely that the major component of the produced mucilage is γ-PGA as MSG acted as the substrate for production (Kim et al., 2014). The glucose and nitrogen (C:N) ratio in the culture broth is a crucial factor for regulating the production of metabolites during the fermentation (Touratier et al., 1999). The fermenting sugar in the HDE is a limiting factor, thus addition of glucose boosts microbial growth and metabolite production during the alkaline fermentation, but glucose excess would skew the C:N ratio, hindering metabolite production. Furthermore, the decrease in consistency observed with a greater glucose content is likely to be related to the decrease in pH as a lower pH induces changes in the tertiary structure of the γ-PGA molecule, decreasing its hydrodynamic molecular weight (Seo et al., 2008). Consequently, addition of 5% MSG and 3% glucose to the HDE followed by fermentation for 3 days yielded a high amount of viscous mucilage and a high and stable number of viable bacteria.

pH and acidity of the co-fermented HDE    The HDE broth fermented by B. subtilis HA was subjected to a second fermentation using L. plantarum EJ2014, which is able to produce GABA. After the first fermentation, 1.5% glucose and various concentrations of Skim milk were added to the viscous broth, followed by the co-fermentation for 7 days. The pH, acidity, viable bacterial counts, and peptide contents were then evaluated. As shown in Fig. 1a, the pH of the HDE without skim milk decreased gradually from 7.13 to 4.96 during the second fermentation. However, addition of skim milk resulted in a notable pH decrease after 1 day of fermentation, although the pH remained fairly constant during the rest of the second fermentation, except for the broth containing 1% skim milk, which exhibited a slight increase in pH by the end of the co-fermentation. These results suggest that addition of skim milk for the second fermentation facilitates microbial growth as it provides lactose as a fermenting sugar, and thus boosts the production of lactic acid. Afterward, the protons derived from the organic acid may be consumed by the intracellular enzymatic conversion of MSG to GABA (Feehily and Karatzas, 2013; Kook and Cho, 2013).

Fig. 1.

Changes in the (a) pH, (b) acidity, (c) viable bacterial counts, and (d) peptide content during the co-fermentation by B. subtilis HA and L. plantarum EJ2014 of the HDE broth containing various skim milk (SM concentrations). Error bars represent SD.

The changes in the acidity of the co-fermented broth are shown in Fig. 1b. The acidity of the HDE without skim milk increased gradually from 0.09% to 0.54% during the second fermentation. By contrast, the acidity of the HDE supplemented with skim milk notably increased after 1 day of fermentation from 0.16%–0.18% to 0.7%–0.9%. The HDE with 5% skim milk exhibited the greatest increase in acidity, indicating 1.08% after 7 days of co-fermentation. On the other hand, the HDE with 1% skim milk exhibited a decrease in acidity after 5 days of fermentation. Generally, a decrease in acidity is expected during the lactic acid fermentation if MSG is used as a substrate for GABA production (Lee and Lee, 2014).

However, the increase in acidity observed here suggests that the supplementation with lactose through skim milk could be an essential growth factor for acid production during the co-fermentation (Kim et al., 2014).

Viable bacterial counts of the co-fermented HDE    The effect of the skim milk content on the viable bacterial counts of the HDE co-fermented by L. plantarum EJ2014 is displayed in Fig. 1c. In the HDE co-fermented without skim milk, the viable cells of B. subtilis HA gradually decreased during 5 days; however, the viable cells of L. plantarum EJ2014 gradually increased, reaching 7.4×10⁸ CFU/mL after 7 days of fermentation. Nevertheless, in the presence of skim milk, the viable cells of both bacterial strains exhibited opposite patterns. After 1 day of fermentation, B. subtilis greatly decreased to 1×10⁷ CFU/mL, whereas L. plantarum increased above 1×10⁹ CFU/mL. Afterward, the viable cells of L. plantarum remained at a constant level, whereas the numbers of B. subtilis gradually decreased during 5 days, although slightly increased by the end of the co-fermentation. Final viable bacterial counts were 2.0×10⁶ and 3.0×10⁹ CFU/mL for B. subtilis HA and L. plantarum EJ2014, respectively. Thus, the growth of B. subtilis was inhibited during the co-fermentation, resulting in a lower population compared to that of L. plantarum. This inhibition may be due to the acidic environment generated by the action of the lactic acid-producing L. plantarum as well as to the more anaerobic conditions resulting from the production of mucilage (Kwon et al., 2016). Therefore, the co-fermentation of the HDE by L. plantarum could be modulated by adding skim milk, resulting in the inhibition of B. subtilis.

Peptide content of the co-fermented HDE    The effect of skim milk concentration on the peptide production in the HDE broth fermented by L. plantarum EJ2014 is shown in Fig. 1d. Co-fermentation of the HDE tended to increase the peptide content during 7 days, although this increase depended on the concentration of skim milk. The peptide content of the HDE co-fermented without skim milk slightly increased, whereas that of the HDE with 5% skim milk showed the highest value (0.79 mg/g) after 1 day and gradually increased reaching 0.87 mg/g by the end of the co-fermentation. Peptides derived from milk protein are known to be bioactive, decreasing the risk of chronic diseases (Clare and Swaisgood, 2000). It is thus expected that the peptides generated by the co-fermentation of the HDE could have valuable functional properties.

GABA and amino acid content in the co-fermented HDE    The TLC plate showing GABA production in the co-fermented HDE with various concentrations of skim milk is displayed in Fig. 2. MSG addition during the first fermentation is necessary for GABA production. Nevertheless, if only MSG is added, the conversion to GABA is incomplete as shown in the TLC plate, suggesting that enrichment with other components is required for an efficient GABA production. Here, fortification with skim milk boosted GABA production, which is likely to be due to modulation of the pH and microbial growth, thus facilitating the bioconversion of MSG to GABA (Lee and Lee, 2014). GABA is involved in the acid stress response of bacteria; thus, it is likely that the increased acid production attained by supplementation with skim milk stimulated L. plantarum to use the glutamate precursor to synthesize GABA.

Fig. 2.

TLC plate showing changes in the GABA production in the co-fermented HDE containing various skim milk (SM) concentrations.

Free amino acids in the co-fermented culture broth were determined using HPLC, revealing the complete conversion of glutamic acid to GABA after 5 days of fermentation. The initial glutamic acid content of 20.01 mg/mL in the broth containing 5% skim milk decreased to zero after 5 days, whereas the GABA content increased from none at the beginning to 16.08 mg/mL, which is higher than that reported for a co-fermentation of a soybean/turmeric mixture (Lim et al., 2016).

We also observed a decrease in the viscosity of the mixture as the fermentation progressed, which may suggest some degradation of the γ-PGA as B. subtilis also produces a poly-γ-glutamate depolymerase, resulting in fragmentation of the polymer. Nevertheless, the resulting fragments are unlikely to be used by L. plantarum as GABA precursors as they are γ-glutamyl peptides and not glutamate monomers (Ashiuchi et al., 2003).

In summary, the MSG substrate in the HDE was successfully converted to functional ingredients by the serial mixed fermentation using B. subtilis and L. plantarum.

Conclusion

In this study, Hovenia dulcis extract was co-fermented using B. subtilis HA and L. plantarum EJ2014 for fortifying it with mucilage and GABA. The HDE with 3% glucose and 5% MSG was efficiently converted to a viscous culture, having high mucilage content and viable bacterial counts. Supplementing the first culture broth with 5% skim milk resulted in increased GABA and peptide production. The bacterial strains acted in a synergistic fashion, with B. subtilis producing the viscous environment required by L. plantarum to produce GABA. Therefore, the co-fermented H. dulcis extract could be used as a functional ingredient in the production of nutraceuticals.

Acknowledgements    This work was supported by the Korean Institute of Planning and Evaluation for Technology in Food, Agriculture, Forestry and Fisheries (IPET) through the Technology Commercialization Support Program, funded by the Ministry of Agriculture, Food and Rural Affairs (No. 314082-3).

References

•  Ashiuchi,  M.,  Nakamura,  H.,  Yamamoto,  T.,  Kamei,  T.,  Soda,  K.,  Park,  C.,  Sung,  M.-H.,  Yagi,  T., and  Misono,  H. (2003). Poly-γ-glutamate depolymerase of Bacillus subtilis: production, simple purification and substrate selectivity. J. Mol. Catal. B-Enzym., 23, 249-255.
•  Bradley,  R.L. (2010). “Moisture and total solids analysis”. In: “Food Analysis”, ed. by  S.S.  Nielsen. Springer, USA, pp. 85-104.
•  Candela,  T. and  Fouet,  A. (2006). Poly-gamma-glutamate in bacteria. Mol. Microbiol., 60, 1091-1098.
•  Chen,  J.,  Shi,  F.,  Zhang,  B.,  Zhu,  F.,  Cao,  W.,  Xu,  Z.,  Xu,  G., and  Cen,  P. (2010). Effects of cultivation conditions on the production of γ-PGA with Bacillus subtilis ZJU-7. Appl. Biochem. Biotechnol., 160, 370-377.
•  Choi,  J.-Y.,  Kim,  J.-H.,  Kim,  G.,  Kim,  C.-K., and  Choi,  M.-S. (2014). Effect of fermented Hovenia dulcis Thunb fruit water extract on biomarker for liver injury and body weight changes in rats given oral administration of ethanol. Korean J. Food Preserv., 21, 412-420.
•  Clare,  D.A. and  Swaisgood,  H.E. (2000). Bioactive milk peptides: a prospectus1. J. Dairy Sci., 83, 1187-1195.
•  Cruz Ramos,  H.,  Hoffmann,  T.,  Marino,  M.,  Nedjari,  H.,  Presecan-Siedel,  E.,  Dreesen,  O.,  Glaser,  P., and  Jahn,  D. (2000). Fermentative metabolism of Bacillus subtilis: physiology and regulation of gene expression. J. Bacteriol., 182, 3072-3080.
•  Dhakal,  R.,  Bajpai,  V.K., and  Baek,  K.-H. (2012). Production of GABA (γ-aminobutyric acid) by microorganisms: a review. Braz. J. Microbiol., 43, 1230-1241.
•  Drehval,  O.A.,  Chervatiuk,  N.V.,  Cherevach,  N.V., and  Vinnikov,  A.I. (2003). Effect of mineral nutrition sources on the growth and toxin formation of the entomopathogenic bacteria Bacillus thuringiensis. Mikrobiolohichnyi zhurnal (Kiev, Ukraine : 1993), 65, 14-20.
•  Feehily,  C. and  Karatzas,  K.A.G. (2013). Role of glutamate metabolism in bacterial responses towards acid and other stresses. J. Appl. Microbiol., 114, 11-24.
•  Hyun,  T.K.,  Eom,  S.H.,  Yu,  C.Y., and  Roitsch,  T. (2010). Hovenia dulcis- an Asian traditional herb. Planta Med., 76, 943-949.
•  Joye,  I.J.,  Lamberts,  L.,  Brijs,  K., and  Delcour,  J.A. (2011). In situ production of γ-aminobutyric acid in breakfast cereals. Food Chem., 129, 395-401.
•  Kim,  J.E.,  Kim,  J.S.,  Song,  Y.C.,  Lee,  J., and  Lee,  S.P. (2014). Novel bioconversion of sodium glutamate to γ-poly-glutamic acid and γ-amino butyric acid in a mixed fermentation using Bacillus subtilis HA and Lactobacillus plantarum K154. Food Sci. Biotechnol., 23, 1551-1559.
•  Kook,  M.-C. and  Cho,  S.-C. (2013). Production of GABA (gamma amino butyric acid) by lactic acid bacteria. Korean J. Food Sci. An., 33, 377-389.
•  Kook,  M.-C.,  Seo,  M.-J.,  Cheigh,  C.-I.,  Lee,  S.-J.,  Pyun,  Y.-R., and  Park,  H. (2010). Enhancement of γ-aminobutyric acid production by Lactobacillus sakei B2-16 expressing glutamate decarboxylase from Lactobacillus plantarum ATCC 14917. J. Korean Soc. Appl. Bi., 56, 816-820.
•  Kunioka,  M. and  Goto,  A. (1994). Biosynthesis of poly(γ-glutamic acid) from l-glutamic acid, citric acid, and ammonium sulfate in Bacillus subtilis IFO3335. Applied Microbiol. Biotechnol., 40, 867-872.
•  Kwon,  S.-Y.,  Garcia,  C.V.,  Song,  Y.-C., and  Lee,  S.-P. (2016). GABA-enriched water dropwort produced by co-fermentation with Leuconostoc mesenteroides SM and Lactobacillus plantarum K154. LWT - Food Sci. Technol., 73, 233-238.
•  Lee,  E.-J. and  Lee,  S.-P. (2014). Novel bioconversion of sodium glutamate to γ-amino butyric acid by co-culture of Lactobacillus plantarum K154 in Ceriporia lacerata culture broth. Food Sci. Biotechnol., 23, 1997-2005.
•  Lee,  N.-R.,  Go,  T.-H.,  Lee,  S.-M.,  Jeong,  S.-Y.,  Park,  G.-T.,  Hong,  C.-O., and  Son,  H.-J. (2014). In vitro evaluation of new functional properties of poly-γ-glutamic acid produced by Bacillus subtilis D7. Saudi J. Biol. Sci., 21, 153-158.
•  Lim,  J.-S.,  Garcia,  C.V., and  Lee,  S.-P. (2016). Optimized production of GABA and γ-PGA in a turmeric and roasted soybean mixture co-fermented by Bacillus subtilis and Lactobacillus plantarum. Food Sci. Technol. Res., 22, 209-217.
•  Ogawa,  Y.,  Yamaguchi,  F.,  Yuasa,  K., and  Tahara,  Y. (1997). Efficient production of γ-polyglutamic acid by Bacillus subtilis (natto) in jar fermenters. Biosci. Biotechnol. Biochem., 61, 1684-1687.
•  Ogunleye,  A.,  Bhat,  A.,  Irorere,  V.U.,  Hill,  D.,  Williams,  C., and  Radecka,  I. (2015). Poly-γ-glutamic acid: production, properties and applications. Microbiology, 161, 1-17.
•  Park,  M.-J.,  General,  T., and  Lee,  S.-P. (2012). Physicochemical properties of roasted soybean flour bioconverted by solid-state fermentation using Bacillus subtilis and Lactobacillus plantarum. Prev. Nutr. Food Sci., 17, 36-45.
•  Parvez,  S.,  Malik,  K.A.,  Ah Kang,  S., and  Kim,  H.Y. (2006). Probiotics and their fermented food products are beneficial for health. J. Appl. Microbiol., 100, 1171-1185.
•  Seo,  J.-H.,  Kim,  C.-S., and  Lee,  S.-P. (2008). Physicochemical properties of poly-γ-glutamic acid produced by a novel Bacillus subtilis HA isolated from cheonggukjang. Prev. Nutr. Food Sci., 13, 354-361.
•  Touratier,  F.,  Legendre,  L., and  Vézina,  A. (1999). Model of bacterial growth influenced by substrate C:N ratio and concentration. Aquat. Microb. Ecol., 19, 105-118.
•  Wang,  M.,  Zhu,  P.,  Jiang,  C.,  Ma,  L.,  Zhang,  Z., and  Zeng,  X. (2012). Preliminary characterization, antioxidant activity in vitro and hepatoprotective effect on acute alcohol-induced liver injury in mice of polysaccharides from the peduncles of Hovenia dulcis. Food Chem. Toxicol., 50, 2964-2970.
•  Wang,  N.,  Du,  S.-k.,  Feng,  X.-c.,  Yang,  Z.,  Xue,  K.-l., and  Li,  Z.-x. (2016). Protective effect of lactic acid-fermented beverage prepared from the peduncles of Hovenia dulcis on alcohol-induced liver damage in mice. Mod. Food Sci. Technol., 8, 005- (in Chinese).
•  Watanabe,  Y.,  Hayakawa,  K., and  Ueno,  H. (2011). Effects of co-culturing LAB on GABA production. J. Biol. Macromol., 11, 3-13.
•  Yang,  J.,  Shan,  C.,  Kuirong,  Z., and  Wenjun,  W. (2002). Effects of Hovenia dulcis Thunb on blood sugar and hepatic glycogen in diabetic mice. J. Chin. Med. Mat., 3, 190-191 (in Chinese).
•  Yao,  K.,  Feng,  C., and  Zhu,  X. (2017). Regulatory roles and mechanisms of Hovenia dulcis Thunb toward polarization of macrophage RAW264. 7. Int. J. Clin. Exp. Med., 10, 2137-2146.
•  Yoshikawa,  M.,  Murakami,  T.,  Ueda,  T.,  Yoshizumi,  S.,  Ninomiya,  K.,  Murakami,  N.,  Matsuda,  H.,  Saito,  M.,  Fujii,  W., and  Tanaka,  T. (1997). Bioactive constituents of Chinese natural medicines. III. Absolute stereostructures of new dihydroflavonols, hovenitins I, II, and III, isolated from hoveniae semen seu fructus, the seed and fruit of Hovenia dulcis THUNB.(Rhamnaceae): inhibitory effect on alcohol-induced muscular relaxation and hepatoprotective activity. Yakugaku zasshi: Journal of the Pharmaceutical Society of Japan, 117, 108-118 (in Japanese).