Introduction

Foxtail millet (Setaria italica L.) is a staple crop in Eurasia as food and forage [1]. Because of its strong vitality in drought prone areas and very poor soils where other crops fail to grow, millet has been extensively cultivated in semi-arid regions [2]. The main components of foxtail millet include starch, protein, lipid, vitamins and minerals [3–5]. Millet porridge is often used as weaning food in developing countries due to its flavor and high nutritional value. In north China, this porridge has been widely used as a nourishing gruel or soup for pregnant and nursing women, and has been applied to food therapy [6]. However, some antinutrient agents, such as phytic acid, tannins and polyphenols in millet affect its nutritional value [2]. Would there be any way to decrease the antinutritional agents and increase the nutritional value of millet?

Germination is a method that can modify the presence of nutrients and antinutrients of cereals [7, 8]. During germination, a large number of significant changes were observed in the biochemical and physical aspects [9], which were accompanied by interconversion and production of new compounds [10]. Especially, germination can increase the content of functional components of cereals and enhance their healthful functions. The capability of hydroxyl radical inhibition was enhanced in germinated millet [11]. It was reported that the content of γ-aminobutyric acid (GABA) in germinated brown rice is three times that of brown rice [12].

GABA, a non-protein amino acid with four-carbon components, is ubiquitously distributed in eukaryotes and prokaryotes. It is produced in plant tissues by the irreversible decarboxylation of l-glutamic acid which is catalyzed by glutamate decarboxylase (GAD, EC4.1.1.15). GABA has a series of healthful functions: as a major inhibitory neurotransmitter in animals [13, 14], regulation of cardiovascular functions, such as blood pressure and heart rate, and alleviation of pain and anxiety [14]. However, GABA content in plant tissues is rather low [ranging from 0.31 to 20.62 mg 100 g−1 fresh weight (FW)] [15, 16]. Therefore, GABA-enriched food is currently one of the hot topics in functional food research.

Researchers have indicated that stressful environmental conditions, such as heat or cold shock, mechanical stimulation and hypoxia [17–19], could strongly promote GABA accumulation among which anaerobiosis is the condition that induced the largest accumulation [20, 21]. Aurisano et al. [22] reported that accumulation of GABA was induced by anoxia in rice seedling through 24 h of anoxia and the yield was 6–8 times that of air treatments. GAD is a key enzyme for stimulating GABA accumulation. GAD activity is enhanced by increasing the concentration of cytosolic H+ [23]. GAD extracted from plant tissue has an optimal pH between 5.5 and 6.0 [21, 24, 25]. In vitro and in vivo data indicated that a reduced cytosolic pH stimulated GAD activity and promoted GABA accumulation [23].

Previous studies were mostly focused on phytic acid and mineral changes during germination of foxtail millet [2, 5, 9, 26]. However, to the best of our knowledge, there has been little literature on GABA accumulation in foxtail millet. Foxtail millet is rich in glutamic acid [3, 27], which makes this millet a potentially promoting source for GABA-enriched foods.

In the present study, the accumulation of GABA during millet germination under hypoxia was investigated. First, a buffer solution was selected by comparing the effects of citrate and acetate buffer solutions on GABA yield. Then, the effects of culture temperature, air flow rate and pH of culture solution on GABA accumulation were investigated using Box–Behnken experimental design of response surface methodology (RSM). The main objective is to find a way to improve GABA content in foxtail millet.

Materials and methods

Materials

Foxtail millet seeds (Jingu-34) were obtained from Shanxi Academy of Agricultural Science (Taiyuan, China). The materials were stored at 4 °C until used.

GABA standard sample was purchased from Sigma Chemical Co. (St Louis, USA). Acetonitrile was high performance liquid chromatography (HPLC)-grade, and other chemicals and reagents used were of analytical grade.

Seeds germination by soaking and gaseous treatment

Dried seeds were cleaned manually in distilled water to remove supernatant seeds, dust and other extraneous materials. About 20 g of seeds cleaned were sterilized by soaking in solution of 1% sodium hypochlorite for 30 min, followed by washing in sufficient distilled water, and then steeped in distilled water at 25 ± 1 °C for 8 h. After that, the steeped grains were drained and placed in cultivated pots with lids (φ 5.5 cm × 6 cm) where they were germinated with 100 ml of culture medium at a constant temperature in a dark incubator. The culture medium was aerated by a pump. The air flow rate was controlled by flowmeter (Yuyao jintai Meter Co. Ltd, Zhejiang, China). The culture solution was replaced with fresh solution at 24 h intervals until germination was complete. The germinated millet seeds were rinsed with distilled water, dried with a filter paper, frozen in liquid nitrogen, and stored in a refrigerator at −20 °C until analysis.

Selection of culture medium

According to the above mentioned method, millet seeds were germinated in a buffer solution of 10 mmol/L of Citrate and that of acetate (pH 5.8) respectively at the same time at a temperature of 32 °C and an air flow rate of 1.5 L/min for 60 h. Germinated seeds were collected from the two solutions to measure and compare their contents of GABA at 12 h intervals during germination. The solution that contributed to higher accumulation of GABA in germinated millet seeds was selected. Then, different concentrations of the selected buffer solution (pH 5.8) were prepared ranging from 5 to 200 mmol/L. Millet seeds were germinated in these solutions for 48 h. The buffer solution of a certain concentration that produced the highest GABA accumulation was determined as the culture medium for this study.

Box–Behnken design

The experimental design was performed using Stat-Ease software (Design-Expert version 6.0.10 Trial, Delaware, USA Echip, 1993). A three-factor three-level Box–Behnken design was chosen to evaluate the combined effects of three independent variables. On the basis of single factor experiments, the levels of culture temperature (X 1), air flow rate (X 2) and pH of culture solution (X 3) were determined (Table 1). The above selected buffer solution at different pH values were prepared for this design. GABA content (Y) was taken as response value. The completed design consisted of 17 combinations including five replicates of the center points (Table 2). The experimental results should agree with a second-order polynomial equation Eq. (1) by a multiple regression technique:

Table 1 Independent variables and their coded and actual values using a Box–Behnken design for optimization of culture conditions
Full size table
Table 2 Box–Behnken design and experiment date for GABA content from germinated foxtail millet
Full size table
$$ Y = \beta _{0} + \sum\limits_{{i = 1}}^{{\text{k}}} {B_{i} X_{i} } + \sum\limits_{{i = 1}}^{{\text{k}}} {B_{{ii}} X_{i} ^{2} } + \sum\limits_{{i > j}}^{{\text{k}}} {B_{{ij}} X_{i} } X_{j} $$
(1)

where, Y stands for GABA content, β 0 denotes the model intercept, X i and X j are the coded independent variables, B i , B ii and B ij represent the regression coefficients of variables for linear, quadratic and interaction regression terms, respectively, k equals to the number of the tested factors (k = 3). An analysis of variance (ANOVA) table is generated to determine individual linear, quadratic and interaction regression coefficients. The significances of polynomial relations are examined statistically by computing the F value at a probability (P) of 0.001, 0.01 or 0.05, respectively. The regression coefficients are then used to make statistical analyses and to generate contour maps of the regression models.

Determination of GABA

Germinated foxtail millet (1.00 g) was mulled with 6 ml of 4% acetic acid. The serosity was deposited for 1 h for extracting GABA sufficiently, and then centrifuged at 2,415 g for 15 min. The supernatant was collected and added with 4 ml of ethanol to remove macro-molecular polymers. The purified supernatant was evaporated (0.1 MPa, 45 °C) to volatilize the acetum acid and ethanol. The residues were dissolved with 0.5 ml of distilled water and centrifuged at 1,766 g for 10 min.

One hundred microlitres of the filtered supernatant was analyzed by HPLC (Agilent 1200, USA) with a Prodigy C18 reversed-phase column (5 µm), 4.6 × 250 mm i.d. as described by Rossetti and Lombard [28]. Mobile phase A consisted of buffer solution of 70 mmol/L sodium acetate (pH 5.8) added 0.5 ml triethylamine in 1 L, whereas mobile phase B was acetonitrile. The elution system involved 92% mobile phase A and 8% mobile phase B at the flow rate of 0.5 ml/min during the entire run. Twenty microlitres of each sample was injected, detected at 254 nm, column temperature at 27 °C.

Statistical analysis

All the trials were carried out in triplicate. The experimental results obtained were expressed as means ± SD. Statistical analysis was performed using the software STATISTICA 6.0. Data were analyzed by ANOVA (P < 0.05) and the means separated by Duncan’s multiple range test.

Results and discussion

Effects of buffer species on GABA yield during germination of foxtail millet

In this study, citrate and acetate buffer solutions were chosen as culture media according to the purpose of germinated foxtail millet as food. The effects of the two buffer solutions on GABA content of germinated foxtail millet were shown in Fig. 1. GABA yield in germinated foxtail millet increased with longer cultivated time in both buffer solutions, but there was no significant increase (P < 0.05) of GABA in both the solutions when the cultivated time was extended from 48 to 60 h, implying that GABA stopped accumulating at 48 h. Between 0 and 24 h of germination, there was no significant difference (P < 0.05) of GABA accumulation between the two solutions, but there was a significant difference (P < 0.05) between 36 and 60 h of germination. During the entire 60 h germinating period, GABA yields in germinated foxtail millet increased by 2.21 folds in citrate solution and 1.81 folds in acetate buffer solution. Citrate solution produced higher GABA accumulation.

Fig. 1
figure 1

Effects of different buffer solutions on GABA accumulation during germination of foxtail millet. Germination 0 h means 30 min sterilized in sodium hypochlorite added 8 h soaked in distilled water. The reaction was performed in 10 mmol/L buffer solution of citrate (filled diamond) and acetate (filled square)

Full size image

The pH value of culture solutions is enhanced during decarboxylation of glutamate, exceeding the optimum pH of GAD in plants. However, acidic environment is more favorable for GABA accumulation [29]. Therefore, buffer solutions were used to stabilize pH of reaction in this study. The results indicated that citrate buffer was more suitable for GABA accumulation than acetate buffer in this study. At the same time, sprout length of at least 20 millet seeds were measured with a centimeter ruler. The sprout lengths in citrate buffer were higher than that in acetate (the data were not shown), it indicated that citrate buffer promoted the growth of foxtail millet while acetate solution inhibited millet growth. Sadami et al. [30] also found that GABA accumulation in rice germ was influenced by buffer species, 2-morpholinoethane-sulfonic (MES) was most suitable for GABA production among the buffers, however, MES is inappropriate for food processing, citrate and malate was the second place, succinate was inappropriate for controlling pH during GABA production. The mechanism of the effect of different buffer solutions on GABA yield was yet to know.

Effects of citrate buffer at different concentrations on GABA yield

The results showed in Fig. 2 indicated that GABA accumulation increased significantly (P < 0.05) when the concentrations of the buffer solution were between 5 and 10 mmol/L. The accumulation continued increasing but not significantly in statistical analysis (P < 0.05) when the solution concentrations were between 10 and 50 mmol/L. It decreased significantly (P < 0.05) when the buffer concentrations were between 50 and 200 mmol/L. Sprout length of millet seeds was measured under different concentrations of citrate buffer, the sprout lengths of foxtail millet in 100 and 200 mmol/L buffer were zero, whereas, the sprout length of millet in other concentrations exceed 2 mm. These phenomena suggested that millet growth was inhibited when it was maintaining in higher concentrations of the buffer. The results indicated that the citrate buffer at a concentration of 10 mmol/L was a desirable solution as culture medium for ideal accumulation of GABA in foxtail millet germination.

Fig. 2
figure 2

Effects of citrate buffer solution (pH 5.8) at different concentration on GABA yield of germinated foxtail millet. Values marked by the same letter are insignificantly different (P < 0.05)

Full size image

Analysis of Box–Behnken experiment

The RSM design and the corresponding experimental data were shown in Table 2. By applying multiple regression analysis of the data, we found that the model of RSM design were consistent with the second-order polynomial equation referred to in Eq. (1). The second-order polynomial model describing the correlation between GABA content and the three variables in this study was obtained in Eq. (2) below:

$$ Y = - 760.68 + 21.44X_{1} + 16.58X_{2} + 144.51X_{3} - 0.27X_{1} ^{2} - 4.47X_{2} ^{2} - 10.65X_{3} ^{2} - 0.66X_{1} X_{3} . $$
(2)

The comparison of the observed values of GABA content with the predicted values (Fig. 3) showed that the two sets of values were very close, indicating the experimental model was valid. By ANOVA for Eq. (2), we obtained an F value of 30.42 which implied that the model was significant (P < 0.0001), a coefficient of determination (R 2) of 0.9594 which indicated a close agreement between experimental and predicted values of GABA content, and an adequate precision of 15.264 which showed an adequate signal. These results also proved the validity of the experimental model.

Fig. 3
figure 3

Correlation between predicted and observed values of GABA content

Full size image

Influence of culture temperature, air flow rate and pH of buffer on GABA accumulation

To determine the optimal levels of the variables for GABA accumulation in foxtail millet germination, three-dimensional surface plots were constructed, according to Eq. (2). By using the plots, interaction between two variables could be easily understood.

Figure 4 showed the effects of temperature and pH on GABA accumulation at an air flow rate of 1.5 L/min during millet germination. Temperature had most significant linear and quadratic effects (P < 0.0001) on GABA accumulation; pH value had significant quadratic effect (P = 0.0035) on the accumulation. Besides, there was a significant interaction (P = 0.0434) between temperature and pH (Table 3), which also influenced GABA accumulation. At a fixed pH, temperature increase led to a sharp increase in GABA accumulation. When the temperature was 33 °C, the accumulation reached its peak, indicating 33 °C was the optimal temperature for GABA accumulation. When the temperature was set, pH increase resulted in slow increase of GABA accumulation; and the optimal pH was observed at 5.8.

Fig. 4
figure 4

Effects of temperature and pH on the content of GABA in germinated foxtail millet

Full size image
Table 3 Analysis of variance (ANOVA) for the regression equation
Full size table

The effects of temperature and air flow rate at pH 5.8 on GABA content were illustrated in Fig. 5. The air flow rate had a significant linear effect (P < 0.0001) and quadratic effect (P = 0.0302) on GABA yield. However, temperature and air flow rate did not interact significantly (P > 0.05) (Table 3). When air flow rate increased from 1.0 to 1.9 L/min at fixed temperature, GABA content slightly increased. When air flow rate continued to increase, GABA content decreased. This suggested that 1.9 L/min was the optimal air flow rate. Our results indicated that GABA content in germinated foxtail millet was accumulated under hypoxia, which agreed with previous reports [20–22].

Fig. 5
figure 5

Effects of temperature and air flow rate on the content of GABA in germinated foxtail millet

Full size image

Figure 6 indicated the effects of pH and air flow rate on GABA content of germinated foxtail millet at 32 °C. The interaction between the two variables in this model was insignificant (P > 0.05) in GABA accumulation (Table 3). The GABA yield increased when pH increased from 5.4 to 5.8 but it did not continue to increased significantly when pH is higher than 5.8. Reducing the pH value of culture solution might lead to a decline of cytosolic pH, which stimulated GAD activity and promoted GABA accumulation of germinated foxtail millet. Our results also confirmed that GABA synthesis was in related to pH-regulating, which was also reported in previous papers [23, 29, 31].

Fig. 6
figure 6

Effects of pH and air flow rate on the content of GABA in germinated foxtail millet

Full size image

Verification of the model

According to the RSM test results, the optimal conditions obtained from the model was as follows: germinating temperature, 33 °C; air flow rate, 1.9 L/min; and the pH value of cultivating solution, 5.8. Under these conditions, the model predicted that a maximal content of GABA was 26.35 mg 100 g−1 FW.

Experiments with different conditions for GABA accumulation within the scopes of the variables investigated in Box–Behnken design in order to evaluate the validity of the above model. The arrangement and results of the confirmatory trials were shown in Table 4. The observed content of GABA in millet germination under the optimal conditions was 26.96 mg/100 g FW, which was quite close to the predicted value of 26.35 mg 100 g−1 FW in the model. The experimental results proved that the model was valid. In addition, the GABA yield (26.96 mg/100 g FW) in germinated foxtail millet in the present study was relatively satisfactory compared with the GABA accumulation (24.9 mg/100 g DW) in germinated brown rice under soaking and gaseous treatment by Komatsuzaki et al. [12].

Table 4 Arrangement and result of confirmatory trials
Full size table

Conclusions

Accumulation of GABA in germinated foxtail millet under hypoxia was investigated in this study. The effects of two buffer solutions—citrate solution and acetate solution—on GABA accumulation in germinated foxtail millet were compared. Citrate buffer solution at 10 mmol/L was found to be the most effective medium. By RSM, the optimal conditions for GABA accumulation in germinated foxtail millet were obtained: temperature, 33 °C; air flow rate, 1.9 L/min; and pH, 5.8. And the predicted highest GABA yield was 26.35 mg 100 g−1FW. Experiments under the designed conditions were carried out, and a highest GABA yield of 26.96 mg 100 g−1FW was obtained and was close to the predicted yield. The results prove the designed model to be valid. This present study indicates that foxtail millet can be used as a good source of GABA-enriched foods.