Introduction

Rice (Oryza Sativa L.) is one of the world’s major cereal crops, and staple food for nearly half of the world’s population. Commonly, brown rice is consumed after polishing. However, a large quantity of nutrition is lost when it gets milled, which causes nutritional deficiency diseases in the population feeding on white rice. In comparison with white rice, brown rice is not only rich in the basic nutrition components such as starch, protein and fat, but also contains more bioactive components, such as dietary fibers, phytic acids, E and B vitamins, γ-aminobutyric acid, ferulic acid, calcium, iron, magnesium, phosphorus and zinc [1]. Recently, people have paid special attention to brown rice [2]. However, because of its poor texture, smelly odor of bran, low digestibility and the not-easy-to-cook characteristic, brown rice is not considered as a suitable table rice [3]. As stated above, it is of importance to search for a process that can improve the texture, digestibility and the health-care function of brown rice.

Germination is a widely used processing method for cereals, especially in the Orient and far east Asia [4]. During germination, hydrolytic enzymes are activated and major reserve substances (starch and protein) are mobilized and converted into micro-molecules for root and shoot growth. High molecular weight reserve proteins would be degraded into smaller peptides, which are subsequently hydrolyzed to free amino acids by peptidases [5–7]. In this way, the nutritional quality of proteins is enhanced. Germination has positive effects on the amino acid composition, protein availability and sensory properties, and causes improvement in the contents of total sugars and certain B-group vitamins and a decrease in anti-nutrients [8]. After germination, brown rice may be cooked easily, while its nutrition is easy to uptake and the texture is softer than that of ungerminated brown rice [3, 9]. That is why a limited germinating process has been suggested as a technological procedure for improving the nutritional quality of the cereal [8]. Therefore, germinated brown rice (GBR) could become a popular health food.

Peptides are better absorbed than single amino acids during gastrointestinal digestion and may play a dominant role in the absorption of amino acids [10]. Moreover, many food-derived peptides have been found to have some physiological activities, including antimicrobial properties, antihypertensive effect, cholesterol-lowering ability, antithrombotic and antioxidant activities, enhancement of mineral absorption/bioavailability and immunomodulatory effect [11]. Rice protein is as valuable as fish, shrimp and beef. Moreover, it has been reported to have unique hypoallergenic properties [12]. Many rice-derived bioactive peptides have been identified as angiotensin-I-converting enzyme inhibitory peptides in sake and sake lees, opioid-antigonist peptides and immunomodulating peptides [13–16]. As stated above, rice-derived peptides are potential substances for functional food material.

Protease plays a critical role in the metabolism of protein during brown rice germination. Both plant hormone and metal ions can effectively activate or inhibit the protease, and in this way protein metabolism will be induced [5, 17]. On one hand, rice bran has homologous protein with brown rice; on the other hand, rice bran has more soluble protein than brown rice that makes it easier to be cleaved. In this study, we added plant hormone, metal ions and rice bran extract to the culture solution to improve the degradation efficiency of storage protein and peptides accumulation.

Statistical methodologies, such as Plackett–Burman design (PBD), have been shown to be efficient and effective approaches to systematically investigate the target factors. Uniform design (UD) has the attractive advantage of needing much fewer trials compared to other statistical designs of experiments such as response surface design [18]. There are many reports on the influence of germination on γ-amino butyric acid, phytic acid, starch, α-amylase and β-amylase in brown rice [2, 3, 8]; however, there are few reports on the change of peptides content. The objective of this study was to optimize the culture solution to enhance the peptides accumulation in GBR. To this aim, PBD and UD were combined to optimize the compositions of culture solution for brown rice germination. We hope that this research will throw light on the possibility of GBR serving as a staple food, combined with the advantages of taste and nutrition.

Materials and methods

Materials

Paddy rice (Oryza stativa L.) was purchased from Jintan Rice Co. Ltd. (Jiangsu, China). These grains were harvested in 2006, stored at −20 °C and kept at 4 °C for thaw before use, and then dehusked by the grain sheller (JGMJ8098,Shanghai Jiading Oils and Grains Apparatus Co. Ltd., China). The defatted rice bran was pulverized (80 mesh) and extracted with deionized water (10 g/50 mL) at 25 °C for 2 h. The homogenate was centrifuged at 4,000 rpm for 15 min and the supernatant was used for further study (rice bran extract). All reagents were of analytical grade.

Routine assays

The moisture, crude fat, ash, crude fiber and starch contents were estimated by standard AOAC methods [19]. Release of α-amino nitrogen was measured by the ninhydrin method and total nitrogen by micro-Kjeldahl method. Brown rice protein was calculated from Kjeldahl N by multiplying with 5.95. Soluble protein was determined by Coomassie blue G-250 dye binding using bovine serum albumin as standard [20]. The proximate composition of brown rice was determined and the results are presented in Table 1.

Table 1 Proximate composition of brown rice
Full size table

Seed germination

Brown rice was cleaned manually to remove the disfigured and other extraneous materials. About 20 g of seeds were surface-sterilized by immersion in sodium hypochlorite solution having 1 mL/100 mL available chlorine for 30 min, followed by a thorough washing in deionized water. A steeping schedule (6 h wet steep in deionized water and 15 min air rest at 32 °C) was adopted. After that, the seeds were transferred to a cultivated pot with lid (φ 7.5 cm × 8 cm) for a period of 2 days of germination at 32 °C, where they were soaked in 100 mL of culture solution in an incubator. The culture solution was aerated by a pump with an air flow rate of 1.2 L/min and replaced with fresh solution once at 24 h intervals. The rice germinated in deionized water was taken as the control. Then, the germinated seeds were carried out from the incubator and washed with deionized water, frozen in liquid nitrogen and stored in a refrigerator at −20 °C until analysis [21].

Selection of significant variables by Plackett–Burman design

For the selection of significant variables affecting the accumulation of rice peptides, 11 variables including gibberellic acid (GA3), inorganic salts and rice bran extract (RBE) in the culture solution were evaluated. Based on the PBD, each factor was examined at two levels: −1 for a low level and +1 for a high level. Table 2 illustrates the factors under investigation as well as the levels of each factor used in the experimental design, whereas Table 3 represents the design matrix.

Table 2 Experimental variables at different levels used for the peptides content using Plackett–Burman design
Full size table
Table 3 Twelve-trial Plackett–Burman design matrix for 11 variables with symbol code along with the observed peptides content
Full size table

Optimization by uniform design

On the basis of the results of the screening with PBD, the experiment for optimization of components in the culture solution was arranged according to the UD table U13*(134) [18]. The four investigated factors were GA3, NaCl, CuSO4 and MnSO4, respectively. Their concentrations were varied at 13 levels (Table 5), the peptides content was taken as response and protease activity was also observed.

Determination of peptides content in germinated brown rice

About 2 g of frozen germinated seeds were homogenized in 8.0 mL of phosphate buffer solution (pH 7.0, 0.2 mol/L) by grinding. The homogenate was centrifuged at 5,000 rpm for 15 min. The supernatant was used for peptides assay according to the biuret method [22]. Macromolecule protein in the supernatant was deposited by an equal volume of 10 g/100 mL trichloroacetic acid (TCA), and then centrifuged at 10,000 rpm for 15 min. Into 3 mL of the supernatant, 2 mL aliquots of biuret reagent were added. After an incubation period of 10 min at 25 °C and a centrifugation at 10,000 rpm for 15 min, the peptides were determined by the absorbance at 540 nm using l-glutathione (reduced) as standard.

Protease activity assay

About 2 g of frozen germinated seeds were homogenized in 8.0 mL of McIlvaine buffer (pH 6.0, 0.02 mol/L and ice bath), containing 5 mmol/L 2-mercaptoethanol and 2.5 mmol/L EDTA. The homogenate was centrifuged at 10,000 rpm for 30 min at 4 °C. The supernatant was used for protease activity assay.

The determination of protease activity is based on the method described by Harvey and Oaks [23] with some modifications. Casein at a concentration of 20 g/L was prepared as substrate. Casein solution (1 mL) was incubated with 1 mL of enzyme solution at 40 °C for 10 min. The reaction was stopped by incubating in 90 °C water bath for 5 min. After adding 2 mL of TCA (0.4 mol/L), the suspension was allowed to stand for 15 min at room temperature and then centrifuged at 8,000 rpm for 10 min. Activity was routinely measured as the increase of absorbance at 275 nm of the trichloroacetic acid-soluble fraction in a UV/vis spectrophotometer (UNICO UV-2802, USA) and arbitrarily calibrated against the absorbance of tyrosine at 275 nm. One unit of protease activity was defined as an increase of one absorbance unit (1 cm light path) per minute in the assay conditions. As a control analysis, the enzyme was de-activated by incubating in 90 °C water bath for 5 min before being determined.

Statistical analysis and modeling

The data from PBD trials were subjected to manual regression analysis using the statistical software package “Design-Expert” (Version 7.0 Stat-Ease Inc., Minneapolis, MN, USA) to select the variables with statistically significant and positive effect for response value. The statistical significance of the model equation and the model terms were evaluated via computing the F value at a probability (P) of 0.05.

The UD experimental data were fitted to a Taylor second-order approximating function:

$$ Y = {{\upbeta}}_{0} + \sum\limits_{i = 1}^{4} {{{\upbeta}}_{i} X_{i} } + \sum\limits_{i = 1}^{4} {{{\upbeta}}_{ii} X_{i}^{2} } + \sum\limits_{i = 1}^{3} {\sum\limits_{i < j}^{{}} {{{\upbeta}}_{ij} X_{i} X_{j} } } $$
(1)

where Y stands for the response function. β0 denotes the center point of the system, β i , β ii and β ij represent the coefficients of the linear, quadratic and interactive effect, respectively, and X i , X ii and X i X j represent the linear, quadratic and interactive effect of the independent variables, respectively. A backward regression analysis of UD experimental data was carried out using the statistical analysis system (version 9.0, SAS Institute Inc., Cary, NC, USA) and the second-order polynomial regressed equation was established. The optimal combination of variables was found using LINGO Optimization Toolbox (version 10.0, LINDO System Inc., Chicago, USA). All trials were carried out in triplicate and all the data were reported as mean ± SD (standard deviation). All chemical and physical analyses were expressed on a dry weight basis.

Results

Screening of significant variables using Plackett–Burman design

A total of 11 variables were analyzed with regard to their effects on peptides content using a PBD. The design matrix selected for the screening of significant variables and the corresponding responses are shown in Table 3. Of these, only the most effective factors with positive significance would be selected for further optimization, while those showing high negative effects on the bioprocess may be dropped in all further experiments [24].

The data in Table 3 indicated that there was a wide variation from 5.02 to 9.10 mg/g of peptides content in 12 trials. This variation reflected the importance of culture solution optimization in attaining higher peptides yield. The adequacy of the model was calculated and the variables evidencing statistically significant effects were screened via the F values for ANOVA (Table 4). The lower P values indicate the more significant factors on peptides content; “effect” implies the positive or negative effect. The P values of variables were obtained. GA3, NaCl, CuSO4 and MnSO4 exerted significant positive effects on peptides content, whereas KCl, ZnSO4 and RBE contributed negatively; MgSO4, CaCl2, FeSO4 and l-cysteine had insignificant effects.

Table 4 Estimated effect, regression coefficient, sum of squares, contribution and corresponding P values for peptides content in Plackett–Burman design experiment
Full size table

Consequently, based on the results from PBD experiment, the optimum levels of the four variables, statistically significant variables (GA3, NaCl, CuSO4, and MnSO4) with positive effect were further determined by a UD experiment.

Optimization of significant variables by uniform design

The design matrix and the corresponding result of UD experiment to determine the effects of four independent variables (GA3, NaCl, CuSO4 and MnSO4) are shown in Table 5. The independent and dependent variables were fitted to the second-order model equation and examined for the goodness of fit. The ANOVA for regression models on the responses is shown in Table 6. The statistical analysis indicated that the proposed model was adequate with very satisfactory values of the R 2, probability, model significance (F value) and C p for the responses. The R 2 values for peptides content and protease activity were 0.8759 and 0.9740, and the P values were 0.0468 and 0.0011, respectively. The C p values shown in Table 6 are the finals through the backward regression analysis. These results indicated that the models were statistically acceptable at the 5% level. Hence, these models could be used to predict the response variables at the experimental region covered in the study.

Table 5 Uniform design with the observed response
Full size table
Table 6 ANOVA for the regression models
Full size table

Regression coefficients and significance for dependent variables of GBR are summarized in Table 7. The significance of each coefficient was determined via computing the F value at P values of 0.01, 0.05 and 0.1. The corresponding variables will be more significant if the absolute F value becomes larger and the P value becomes smaller [25]. Peptides content is the response value that we care about most in this research. It was affected by the concentration of components in the culture solution. It can be seen that the variable with the most significant effect was GA3 (P < 0.01), followed by the interaction effect of NaCl and MnSO4, the linear terms of MnSO4 and NaCl (P < 0.05), and then the interaction effect of GA3 and MnSO4, significant at P < 0.1. CuSO4 has insignificant effect (P > 0.1) on peptides content. The model was then submitted to statistical analysis to neglect all terms that were statistically insignificant (P > 0.1; Table 7). Consequently, the polynomial model describing the correlation between peptides content and the four variables is presented as follows:

Table 7 Regression coefficients and statistical significance for the models
Full size table
$$ Y = 3.93 + 0.15X_{1} - 0.76X_{3} + 4.14X_{8} - 0.001X_{1}^{2} - 1.85X_{8}^{2} - 0.02X_{1} X_{8} + 0.63X_{3} X_{8} $$
(2)

where Y stands for the predicted value of peptides content, mg/g; X 1, X 3 and X 8 are the concentration of GA3 (μmol/L), NaCl (mmol/L) and MnSO4 (mmol/L), respectively.

The regression model Eq. 2 allowed the prediction of the effect of four parameters on peptides content. To aid visualization, the response surface for the effect of the independent variables on peptides content are illustrated in Figs. 1 and 2.

Fig. 1
figure 1

Response surface plot showing the interactive effect of NaCl (X 3) and MnSO4 (X 8) on peptides accumulation during brown rice germination. The germinating condition, an air flow rate of 1.2 L/min at 32 °C for a period of 2-day germination in the dark, GA3 = 55 μmol/L

Full size image
Fig. 2
figure 2

Response surface plot showing the interactive effect of GA3 (X 1) and MnSO4 (X 8) on peptides accumulation during brown rice germination. The germinating condition: an air flow rate of 1.2 L/min at 32 °C for a period of 2-day germination in the dark, NaCl = 5.0 mmol/L

Full size image

Figure 1 shows the response surface plot for the effect of NaCl (X 3) and MnSO4 (X 8) concentration in culture solution on peptides content. There was a significant (P < 0.05) cooperation between the two factors according to the results of the study (Table 7). NaCl concentration displayed a linear effect on the response. When the concentration of MnSO4 was lower than 1.2 mmol/L, it can be seen that the peptides content decreased while the NaCl concentration increased, and increased when higher than 1.2 mmol/L. The concentration of MnSO4 demonstrated a quadratic effect on the response.

Figure 2 shows the response surface plot for the effect of GA3 (X 1) and MnSO4 (X 8) concentration in culture solution on peptides content. It can be seen that peptides content showed a wide variation because of changeable concentration for GA3 and MnSO4. The overall effect was curvilinear in nature. Peptides content in GBR was found to increase rapidly with the increase of the GA3 and MnSO4 concentration until 60 μmol/L and 1.6 mmol/L, respectively, followed by a decline. It depends on GA3 and MnSO4 concentrations where the linear (P < 0.05) and quadratic (P < 0.05) effects are all significant, and the positive linear and negative quadratic effects of GA3 and MnSO4 concentration in the culture solution explain the observed nature of the curve shown in Fig. 2.

From the results of the study (Table 7), GA3 and Na+ had significant positive effect (P < 0.1) on protease activity, and Mn2+ exerted insignificant effect (P > 0.1) on protease activity of GBR. It was indicated that Cu2+ had insignificant effect (P > 0.1) on these two responses, as also the interaction effect of Cu2+ and other variables at the experimental region. Therefore, it was excluded from the culture solution.

The correlation between peptides content and protease activity was also investigated and we can conclude that a low correlation was found (0.244). This observations could be explained, but the degradation of protein at the early germination stage could not be explained simply by the amount of protease present (measured using casein as the substrate) [26].

Optimum conditions and model verification

From the model Eq. 2, optimum levels of components in the culture solution for enrichment of peptides in GBR were obtained as presented: GA3 55 μmol/L, NaCl 6.0 mmol/L and MnSO4 1.8 mmol/L. A maximum response of 10.66 mg/g peptides in GBR was predicted. An additional independent trial was carried out to assess the validity of the model. The observed experimental value in the optimum culture solution (10.53 mg/g) was only 1.23% lower than the predicted value. It suggested that there would be a high degree of fit between the values observed and the values predicted by Eq. 2. This revealed that the model was capable of describing the relation between mixture factors and response. An overall 1.90-fold increase in peptides accumulation was achieved in the optimized culture solution compared with the control (5.53 mg/g, germinated in deionized water), and 2.52-fold increase compared with ungerminated brown rice (4.18 mg/g). This also reflected the necessity of the optimization process.

Additionally, the optimal concentrations of GA3 and MnSO4 were within their range, arranged in UD, except for NaCl on the border (6.0 mmol/L). Hence, further studies need to be performed on the concentration optimization of NaCl, which could perfect the composition of the culture solution.

Discussion

As far as both brown rice and peptides are concerned, we put forward a hypothesis that peptides will be accumulated through brown rice germination and the accumulation will depend on the extent of germination period, which is based on the theory of seed physiology. Gibberellins are a large family of tetracyclic diterpenoid natural growth regulators in plants, which play essential roles in the control of various physiological processes in plant growth and development including seed germination, shoot growth, cell division and internode elongation [27]. Gibberellins produced in the embryos of germinating seeds of many cereals promote the formation of amylases, proteases and other hydrolases, which degrade the reserve materials in the seeds, making them available for the growing embryo [28]. In this study, we conclude that exogenous gibberellins can also accelerate germination and induce rice protein proteolysis by activating the protease. During this course, peptides as a kind of protein hydrolysate are enriched. This result also confirms the feasibility of our tentative idea.

Metal ions have been proved to be effective substances for improving the protease activity, the activities of EDTA-inhibited enzymes were restored by the addition of low concentrations of either Mn2+ or Zn2+ ions, but they were inhibited by higher concentrations of these ions [5, 17, 19]. In this research, it was showed that the effects of metal ions on peptides content didn’t coincide with that on protease activity. This result indicated that the protease activity we measured using casein as a substrate just didn’t account for the degradation of storage protein during germination [26]. It was suggested that “actual protease activity” may be affected significantly by Mn2+. At the same time, Na+ may be one of the activating agents of protease from brown rice. Similar observation was reported in barley malt [29]. However, Orlowski [30] reported that monovalent cations like Na+ and K+ inhibit protease activity, while divalent ions like Mg2+ and Ca2+ activate the protease activity. The similar effects about Mg2+ and Ca2+ were found in malted sorghum and barley respectively, while in sorghum and barley malts, metal ions, such as zinc, and ferrous ions inhibited the proteinase activity [17]. Stoknes and Rustad [31] investigated the effect of NaCl on the protease activity. The result showed that protease activity (hemoglobin-hydrolyzing, chymotrypsin-like and trypsin-like activities) with different substrate showed different effect (positive or negative). So, the observed protease activity in brown rice germination may not reflect the actual reaction when storage protein was hydrolyzed by protease [5, 17]. That can also explain the observed low correlation of peptides content and protease activity. The results should be treated with caution because the protease activity was assayed using specialized substrates (casein) [17].

Cu2+ is an essential micronutrient for plants, but it can be toxic to plant at higher concentration. Seed germination is more sensitive to metal pollution because of a lack of some defense mechanisms. In PBD experiment, Cu2+ showed significant and positive effect on peptides content, but insignificant in UD experiment. The reasonable explanation is that a significant interaction effect between Cu2+ and other materials existed in PBD experiment which does not describe interaction among factors, so Cu2+ showed significant and positive effect in statistical analysis arbitrarily. As the components in UD experiment were decreased Cu2+ showed insignificant effect for peptides accumulation with the disappearing of interaction effect among metal ions. Ahsan et al. [32] reported that a total of 25 proteins were differentially expressed in germinating rice seeds when they exposed to copper stress. Although Cu2+ showed insignificant effects on peptides content and protease activity in this study, it is considered that the metabolites of protein may have been induced by copper stress.

The proteases are responsible for the breakdown of large, generally insoluble, storage proteins into soluble proteins, peptides and amino acids [7, 33]. Typically, storage proteins are first cleaved by specific endoproteinases; the resulting peptides are then hydrolyzed to free amino acids by the action of multiple, less specific exopeptidases and/or endopeptidases [7]. Hence, if the endoproteinase, exopeptidases, endopeptidases, aminopeptidase and carboxypeptidase are all investigated respectively, it seems to be more convictive to understand this course. We don’t know the reason for this rise and fall in protease activity. A possible explanation is that the rise was caused by accumulation of hydrolyzed storage proteins, which are released as free amino acids and peptide nitrogen. The fall was likely caused by these hydrolyzed proteins (free amino acids and peptides) migrating from the storage endosperm to the living embryo, with some of these molecules partly lost through the roots and shoots [17]. Hough et al. [34] has shown that roots and shoots of germinating grains are rich in peptides. To understand the regulation of storage protein proteolysis, there is still a long way to go. Then, we may describe the theory of storage protein mobilization and peptides accumulation clearly.

Additionally, we focused on peptides enrichment of GBR in this experiment; however the bioactive functions of them were not investigated. In the future work, more research will be required to identify the bioactivity.

Nowadays, people are paying more and more attentions to the security, nutrition and function of food. In the near future, GBR would be in high demand for accumulating peptides as a result of proteolytic effect by endogenous protease during germination, not synthetic nutriment in food industry. The conclusion of the paper could promote functional food production by brown rice germinated in a particular culture solution.

Conclusions

In this study, eleven variables were tested using Plackett–Burman design and four variables (i.e. GA3, NaCl, MnSO4 and CuSO4) exerted significant (P < 0.05) and positive effect on peptides accumulation. Subsequently, uniform design was used for further optimization of these selected variables to improve peptides content. A second-order polynomial model was established to identify the relationship between the variables and peptides content. The most three effective components are GA3, NaCl and MnSO4. The optimum levels of the variables were obtained. It can be verified that peptides content of GBR in the optimized culture solution was improved significantly.