Food Science and Technology Research
Online ISSN : 1881-3984
Print ISSN : 1344-6606
ISSN-L : 1344-6606
Original papers
Effects of Alkaline Deamidation on the Chemical Properties of Rice Bran Protein
Jingtao GuanRyunosuke TakaiKohei TorayaTomohisa OgawaKoji Muramoto Satoshi MohriDaitaro IshikawaTomoyuki FujiiHyun ChiSeong-Jun Cho
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2017 Volume 23 Issue 5 Pages 697-704

Details
Abstract

The water solubility of rice bran protein (RBP) was improved by deamidation under alkaline conditions. The degree of deamidation was found to be a major factor in improving the solubility of RBP. The decrease in molecular mass or the degradation of peptide bonds was not detected in deamidated RBP under the conditions used. The thermal property and secondary structure of deamidated RBP measured by differential scanning calorimetry and Fourier transform infrared spectroscopy indicated that the secondary structure of RBP was well preserved during alkaline deamidation. By raising pH and temperature for deamidation, the generation rates of constituent amino acid racemization and lysinoalanine increased. The highest solubility (∼90%) of RBP was achieved by treatment at pH 12 and 120°C for 15 – 30 min by enduring side reactions. Moderate solubility (∼40%) could be achieved by deamidation at pH 8 and 100°C for 30 min to minimize side reactions.

Introduction

Rice (Oryza sativa) is one of the major staple foods, with an annual production of about 700 million metric tons world-wide. Rice bran, which is a by-product of the rice milling industry and constitutes around 10% of the total weight of rough rice, usually contains about 11 – 15% protein, 34 – 62% carbohydrates, and 15 – 20% oil. It also contains minerals, vitamins, and a variety of bioactive phytochemicals, indicating it is a possible candidate raw material for the preparation of nutritional foods and nutraceuticals (Gul et al., 2015). Rice bran protein (RBP) has long been recognized as a good source of well-balanced amino acids with low hypoallergenic activity (Han et al., 2015) as well as a resource for non-food applications such as surfactants, adhesives, coatings and plastics (De Graaf, 2000). Its favorable functional properties such as foaming and emulsifying properties have also been recognized (Wang et al., 1999; Chandi and Sogi, 2007; Khan et al., 2011). Moreover, in silico analysis has shown that RBP harbors several bioactive peptides within its primary structures (Cavazos and Mejia, 2013; Udenigwe, 2016). In fact, various functional peptides were purified and characterized from RBP hydrolysates (Adebiyi et al., 2008a, 2009b; Foong et al., 2015).

However, the utilization of RBP is still limited due to the difficulty in protein extraction and its low solubility. A number of methods have been developed to extract RBP, including alkaline (Gnanasambandan and Hettiarachchy, 1995), enzymatic (Hamada, 2000, Tang et al., 2003; Bandyopadhyay et al., 2008; Zhang et al., 2012), and physical treatments (Anderson and Guraya, 2001; Tang et al., 2002; Chittapalo and Noomhorm, 2009). Subcritical water hydrolysis has also been used in extraction (Sereewatthanawut et al., 2008). The most common method is alkaline extraction followed by isoelectric precipitation. The extracted RBP consists of several major proteins having different solubility in water. In the case of a Japanese rice variety (Hitomebore), RBP is composed of albumin, globulin, glutelin, and prolamin, at the ratios of 24 – 39%, 27 – 30%, 33 – 42%, and 1 – 3%, respectively (Adebiyi et al., 2009a).

To enhance the solubility of food proteins such as soybean proteins and wheat gluten, various modification approaches, such as physical, chemical, enzymatic and genetic modifications, have been applied (Li et al., 2005; Mirmoghtadaie et al., 2016). Most of these approaches are related to methods that introduce hydrophilic groups to the side chains of amino acid residues. Chemical modifications such as succinylation and acetylation improved the solubility and functional properties of protein (Kinsella et al., 1976). Deamidation is one of the most commonly used methods to improve the functional properties of food proteins. The conversion of Gln and Asn residues to their deamidated forms increases the solubility of proteins (Mirmoghtadaie et al., 2009). In addition, other physicochemical properties, such as foaming and emulsifying activity of proteins, can be improved by deamidation (Liao et al., 2010).

There are several methods for deamidation of food proteins, e.g., acidic, alkaline, and enzymatic deamidation (Wagner and Gueguen, 1995; Hamada, 1992). During the course of RBP fractionation (Adebiyi et al., 2009a), we found that alkaline deamidation, but not acidic deamidation, could effectively increase the solubility of RBP. However, it is known that severe alkaline treatments of proteins cause racemization of amino acids (Schwass and Finley, 1984; Friedman and Liardon, 1985) and yield Nε-(D,L-2-amino-2-carboxyethy)-l-lysine (lysinoalanine) (Chu et al., 1976). Lysinoalanine (LAL) has a nephrotoxic effect and is formed by the coupling reaction of the ε-amino group of Lys with dehydroalanine resulting from the β-elimination of half-cystine, serine or its derivatives. However, there have been no reports describing such side reactions caused by alkaline deamidation of RBP.

In this study, RBP was deamidated under various alkaline conditions of temperature, pH, and reaction time. The change in water-solubility was examined as a function of the degree of deamidation. The thermal properties and conformational changes of deamidated RBP were investigated by using differential scanning calorimetry (DSC) and Fourier transform infrared (FT-IR) spectroscopy. In addition, the effects of alkaline deamidation on the constituted amino acids of RBP were evaluated by analyzing the racemized amino acids and LAL.

Materials and Methods

Isolated RBP (commercial name: RiPro80), which had been extracted from Japonica rice bran with a weak alkaline solution (pH 9), isoelectric precipitated, and dried by spray drying, was obtained from CJ Cheijedang Foods R&D (Seoul, South Korea). The protein content was estimated to be ∼80% on the basis of total nitrogen content. All processes were carried out at < 60°C except for pasteurization at 135°C for 3 sec before spray drying. 7-Fluoro-4-nitrobenzo-2-oxa-1,3-diazole (NBD-F) and amino acid standard solution (H type) were purchased from Wako Chemicals (Osaka, Japan). All other reagents were of analytical grade from Wako Chemicals or Nacalai Tesque (Kyoto, Japan).

Deamidation of proteins    RBP (0.25 g) was suspended in 25 mL of 0.1 M NaHCO3 and pH was adjusted to 8.0, 10.0, and 12.0 with 1 M NaOH or 1 M HCl. The solutions were then heated at 80°C or 100°C for 30 or 60 min, or at 120°C for 15 or 30 min. After heat treatment, the RBP solutions were neutralized, dialyzed against water at 4°C for 48 h, and then lyophilized.

Solubility of RBP    Fifty milligrams of the protein samples were suspended in 1.0 mL of deionized water, mixed for 30 min at room temperature, and centrifuged at 15000 x g for 20 min. The supernatants were transferred to other tubes. After drying at 40°C, the weight of residual protein was measured.

Measurement of deamidation degree    The degree of deamidation was calculated as the ratio of ammonia generated in the deamidated sample to that of completely deamidated RBP. RBP samples (20.0 mg) and 1.0 mL of 1 M H2SO4 were placed in glass tubes. The glass tubes were sealed by a gas flame and heated at 110°C for 4 h for complete deamidation. The samples were neutralized with 2 M NaOH, transferred to 20-mL volumetric flasks, and filled up to 20.0 mL with water. Ammonia was quantitated using an ammonia assay kit (AA0100-1KT) (Sigma-Aldrich, St. Louis, MO, USA). The amount of NH3 recovered from 1 M H2SO4 hydrolysis was considered to be from complete deamidation.

Fourier transform infrared (FT-IR) spectroscopy    Each dried RBP sample was loaded on an attenuated total reflection (ATR) cell in the FT-IR spectrometer (FT-IR670/IRT-30, Jasco, Tokyo, Japan). The spectra were measured over the range of 400 – 7000 cm−1 with a resolution of 1 cm−1 for 254 scans against a background spectrum. The program IR-SSE (Jasco) was used to estimate the secondary structure of each derivative by principal component regression.

Differential scanning calorimetry (DSC)    The thermal characteristics of RBP samples were examined using a differential scanning calorimeter (SII DSC6200, Seiko Instruments, Chiba, Japan). Samples (10.0 – 12.0 mg) were sealed in aluminum pans and an empty pan was used as a reference. The temperature scanning range was 30 – 140°C at a heating rate of 5°C/min.

Size-exclusion HPLC    Deamidated RBP samples were dissolved in 0.1 M acetic acid/ 3 M urea/ 0.01 M cetyltrimethyl ammonium bromide (AUC solvent) at 10.0 mg/mL and centrifuged to remove insoluble materials at 10,000 x g for 10 min (Hamada, 1996). Twenty microliters of the supernatant was separated by size-exclusion HPLC on a Superdex 200 column (1 × 30 cm) (Amersham Pharmacia Biotech. Uppsala, Sweden) using AUC solvent as a mobile phase at a flow-rate of 0.7 mL/min. The HPLC column was calibrated using standard proteins of 12 – 660 kDa.

Amino acid analysis    RBP samples were hydrolyzed using 6 M HCl/3% (w/v) phenol at 110°C for 22 h in evacuated glass tubes and subjected to amino acid analysis (Muramoto and Kamiya, 1990). Amino acid compositions of RBP samples were determined by an amino acid analyzer (Shimadzu, Kyoto, Japan) using an ion exchange column (Shim-pack Amino-Na, 6 × 100 mm). Separated amino acids were post-column labeled with o-phthalaldehyde and detected by a fluorescence detector (Ex: 350 nm; Em: 450 nm), and the amino acids were quantified by comparing with standard amino acid profiles. For cystine/cysteine determination, samples were first oxidized with performic acid in an ice-water bath for 2 h, and dried under vacuum before hydrolysis.

D/L-Amino acid analysis    RBP samples were hydrolyzed as described above, and the dried hydrolysates were dissolved in 0.1 M sodium borate buffer (pH 8.0) containing 0.01% EDTA. An aliquot of the sample solution (20 µL) was placed into a microtube and mixed with the same volume of NBD-F (50 mM in ethanol, freshly prepared), and the tube was heated at 60°C for 1 min (Miyano et al., 1985). After cooling in ice-water, the reaction mixture was diluted with 460 µL of 5 mM HCl and subjected to reversed-phase (RP)-HPLC on an ODS Capcell pack AG 120A (Shiseido, Tokyo, Japan) (5 µm particles, 4.6 × 250 mm) column maintained at 40°C. The elution solvent systems were (A) acetonitrile (MeCN)-75 mM H3PO4 (13:87, v/v), and (B) MeCN-methanol (MeOH)-50 mM KH2PO4 (21:39:40, v/v/v). The flow rate was 1.0 mL/min. NBD-amino acids were detected by a fluorescence detector (Ex: 470 nm; Em: 530 nm). Since NBD-Trp had no fluorescence, it was detected by visible absorption at 470 nm.

d- and l-isomers of authentic amino acids (Ser, Asp, Glu, Ala, Pro, and Trp) (47 – 61 µM) were derivatized with NBD-F as described and separated by RP-HPLC. The amino acid derivatives separated by RP-HPLC were collected in microtubes and concentrated by a centrifugal evaporator. The dried samples were dissolved in ethanol and injected in chiral-HPLC to separate each d- and l-amino acid. Chiral-HPLC was carried out by using a Sumichiral OA-2500-S (6 µm particles, 4.6 × 250 mm) column (Sumika Chemical Analysis Service, Tokyo, Japan) eluted with 5 mM citric acid-methanol solution at 25°C. The flow rate was 1.0 mL/min.

Statistical analysis    Experiments were carried out in duplicate or triplicate and results are presented as the mean ± standard deviation of the mean (SD). The differences between the means were compared by Duncan's multiple-range test. Differences with p < 0.05 were considered to be significant.

Results and Discussion

In this study, a commercial RBP, prepared by alkaline extraction and isoelectric precipitation, was used as a starting material. The RBP sample was a mixture of albumin, globulin, glutelin and prolamin, and contained greater than 20% non-protein components. RBP was deamidated by changing the heating temperature, pH, and period. The degree of deamidation was increased by raising these parameters (Fig. 1). The highest deamidation degree, ∼80%, was obtained with the sample deamidated at pH 12 and 120°C for 30 min. As the deamidation degree increased, the solubility increased as expected (Table 1). That is, Asn and Gln residues in RBP converted to Asp and Glu residues upon deamidation, so that the surface polarity of the protein increased to improve its solubility. The solubility of non-deamidated RBP was only 18% and increased by alkaline deamidation to 30% even at an acidic pH. The minimum solubility of RBP was observed at pH 4–5, which might be due to the isoelectric range. The solubility increased on either side of this pH region. The highest solubility (93%) was achieved by deamidation at pH 12 and 120°C for 15 min. The deamidated RBP maintained high solubility even at pH 6 and showed slightly lower solubility at pH 3-5 (Fig. 2).

Fig. 1.

Deamidation degree of rice bran proteins. All measurements are expressed as means ± SD of three independent experiments. The different letters above the bars indicate significant differences at p < 0.05 level between treatments.

Table 1. The solubility and denaturation temperature of rice bran proteins deamidated under various conditions.
pH Temperature (°C) Time (min) Solubility (%) DT (°C)
No treatment 18 105
8 80 30 36 100
60 36 98
100 30 44 94
60 41 95
120 15 54 96
30 52 96
10 80 30 34 101
60 30 84
100 30 39 96
60 49 93
120 15 46 72
30 71 95
12 80 30 51 98
60 39 101
100 30 36 90
60 72 109
120 15 93 97
30 92 105

Values represent the average of duplicate measurements. DT: Denaturating temperature measured by DSC.

Fig. 2.

Solubility of deamidated rice bran proteins treated at pH 10 and 120°C for 30 min. Each point represents the average of duplicate measurements.

ART FT-IR spectroscopy is a powerful tool to determine the secondary structure of insoluble proteins by analyzing the amide I band: α-helices (1653 cm−1), β-sheets (1620, 1635, and 1683 cm−1), β-turn (1669 and 1675 cm−1), and random coils (1645 cm−1) (Glassford et al., 2013). The bands at 1683 and 1920 cm−1 are associated with the aggregation process. The program IR-SSE was produced according to the method of Sarver et al. (1991). Although this method was based on the IR data and X-ray data from globular proteins, it was still useful to compare the secondary structures of aggregated forms (Murakami et al., 2003). Therefore, it was applied to assess the changes in the secondary structures of RBP following deamidation (Table 2). RBP showed 15% α-helix, 34% β-sheet, 25% β-turn, and 26% others before deamidation. The aggregated proteins are generally considered to contain a significant amount of β-sheet structures (Hu et al., 2006). Alkaline deamidation did not change these values, suggesting that the secondary structure of RBP was very stable against the deamidation conditions.

Table 2. Secondary structures of rice bran proteins deamidated under various conditions for 30 min.
pH Temperature (°C) α-helix (%) α-sheet (%) α-turn (%) Other (%)
No treatment 15 34 25 26
8 80 17 32 26 25
100 17 37 24 22
120 15 32 28 25
10 80 17 39 22 22
100 21 32 23 24
120 19 34 23 24
12 80 16 38 23 23
100 16 39 22 23
120 14 36 26 24

Values represent the average of duplicate measurements.

In contrast to FT-IR spectroscopy, DSC thermograms of RBP showed different patterns upon deamidation. Although the DSC thermogram did not change following deamidation at pH 8 and 100°C for 30 min, the thermograms changed considerably by raising pH and temperature (Fig. 3). The denaturation temperature of non-deamidated RBP was about 105°C, indicating that RBP had high thermostability (Table 1). The denaturation temperatures of isolated albumin, globulin, glutelin, and prolamin fractionated from RBP were measured to be 46, 79, 74, and 79°C, respectively (Adebiyi et al., 2009a). The high thermostability of RBP might be due to its compact structure, which also causes poor solubility. The distinct structure of RBP may be attributed to the industrial processing conditions, particularly, spray drying (Jerez et al., 2007). The denaturation temperature of RBP decreased upon deamidation, which is known to partially unfold the proteins. However, deamidation conditions such as higher temperature and pH seem to affect the denaturation more strongly. Under moderate conditions, e.g. pH 8–10 and 80 – 100°C, the denaturation temperature decreased with increasing heating period. In contrast, the denaturation temperature was raised again under harsh conditions, e.g., pH 12 and 120°C, which might bring about an apparently stable structure, such as an aggregate form. This is suggested to be the reason why the DSC thermograms of RBP deamidated at 100°C or at pH 10 do not necessarily show changes related to pH and temperature (Fig. 3). The deamidation at pH 10 was effective in decreasing the denaturation temperature, and the lowest denaturation temperature was observed with the deamidation at pH 10 and 120°C for 15 min.

Fig. 3.

DSC thermograms of rice bran proteins deamidated under various conditions. (a) Treated at pH 8, 10, or 12 at 100°C for 30 min; (b) treated at 80, 100, or 120°C at pH 10 for 30 min.

Size-exclusion HPLC was utilized to examine if RBP was destroyed during deamidation (Fig. 4). AUC solvent was used as the elution solvent to dissociate and solubilize deamidated RBP. RBP gave two major peaks corresponding to 33 kDa and 13.7 kDa. The molecular mass distribution of RBP used in this study was shifted to lower regions compared with the samples prepared on a laboratory scale (Adebiyi et al., 2009). We do not know the exact reason for this shift at present. The molecular mass did not change following deamidation under the conditions employed, indicating that the peptide bonds of RBP were not broken during the deamidation process. Furthermore, the RBP samples of higher degree of deamidation gave larger peaks in the chromatograms, due to their higher solubilities in AUC. It is concluded that the improvement in solubility was mainly caused by deamidation and not by the hydrolysis of peptide bonds.

Fig. 4.

Size-exclusion HPLC of deamidated rice bran proteins. RBP was deamidated at pH 8, 10, or 12 at120°C for 30 min. Deamidation pH was 8 (a), 10 (b), or 12 (c), respectively. (d) nontreated RBP.

Column: Superdex 200 (1 x 30 cm); mobile phase: 0.1 M acetic acid/3 M urea/0.01 M cetyltrimethyl ammonium bromide; flow rate: 0.7 mL/min; sample volume: 20 µL.

As shown in Table 3, RBP was rich in Glu/Gln and Asp/Asn, which accounted for 16% and 9%, respectively. High contents of hydrophobic amino acids, such as Val, Leu, Ile, Pro, Phe, and Met may cause low protein solubility. The total percentage of hydrophobic amino acids in RBP was 30%; therefore, the low solubility of RBP samples cannot be attributed to its amino acid composition. The amino acid compositions of the deamidated RBP samples did not show significant decreases of Cys and Lys contents due to the formation of LAL in this study.

Table 3. Amino acid compositions of rice bran proteins deamidated at 120°C for 30 min at pH8, 10, or 12.
No treatment pH8 pH10 pH12
Cys 1.1±0.1 1.1±0.1 1.1±0.1 1.1±0.1
Asx 8.9±0.2 8.8±0.4 9.0±0.2 9.1±0.2
Thr 3.9±0.2 3.9±0.5 4.1±0.1 3.6±0.4
Ser 3.6±0.3 3.9±1.2 4.2±0.1 3.3±0.6
Glx 16.5±0.4   16.6±0.6   17.0±0.2   17.1±0.8  
Pro 4.4±2.5 4.8±0.6 5.9±2.1 4.2±2.5
Gly 9.7±0.4 9.7±1.0 9.5±0.1 9.4±0.6
Ala 8.8±0.6 8.8±1.1 8.3±1.2 9.1±0.3
Val 7.6±0.1 7.5±0.3 7.7±0.1 8.0±0.1
Met 3.8±0.1 3.9±0.3 4.0±0.1 3.2±0.2
Ile 4.8±0.1 4.4±0.4 4.9±0.2 5.0±0.2
Leu 9.2±0.1 9.1±0.1 9.3±0.1 9.9±0.1
Tyr 1.2±0.3 1.2±0.2 1.2±0.2 0.9±0.3
Phe 4.3±0.4 4.6±0.3 4.5±0.4 4.8±0.4
His 2.6±0.2 2.0±0.1 2.0±0.1 2.1±0.1
Trp 0.2±0.1 0.2±0.1 0.2±0.1 0.2±0.1
Lys 5.2±0.1 4.8±0.1 4.7±0.1 4.7±0.3
Arg 4.3±0.7 4.8±1.4 4.5±0.3 4.3±1.0

Deamidated RBPs were hydrolyzed at 110°C for 22 h using 6 M HCl containing 3% phenol. Values are the mean ± SD of three replicates.

It is known that severe alkaline treatments of proteins result in the racemization of amino acids (Schwass and Finley, 1984; Friedman and Liardon, 1985). For the analysis of d-amino acids generated during alkaline deamidation of RBP samples, the HCl hydrolysates of the samples were first labeled with NBD-F to form fluorescent derivatives and separated by RP-HPLC. Each separated NBD-amino acid was collected and subjected to chiral-HPLC to separate d/l isomers (Hamase et al. 2010). In this study, Ser, Asp, Glu, Ala, Pro, and Trp derivatives were collected for chiral analysis. Racemization of amino acids increased by raising the pH and temperature for deamidation (Fig. 5). Especially, d-Ser increased from 7% to 41% when deamidated at pH 8 to 12 at 120°C. d-Asp increased from 16% to 33%, d-Glu from 4% to 25%, and d-Ala from 3% to 12% by increasing the pH for deamidation. Racemization of amino acids is postulated to proceed by an abstraction of the protein from the asymmetric carbon atom to give a negatively charged, optically inactive planar carbanion (Friedman and Liardon, 1985). This carbanion can then be reprotonated on either side, regenerating an equal mixture of both D and L isomers. Relative susceptibilities of racemization seems to be correlated with electron-donating inductive effects of amino acid side chains, as observed by the susceptibility of Ser, Asp and Glu to racemization in this study. Pro, which is the only cyclic amino acid containing a secondary amino group, and Trp resisted racemization. Little racemization of hydrophobic amino acids and Pro has been shown with alkali-treated wheat gluten (Schwass et al., 1984). Since racemization may impair the nutritional value of food proteins due to the non-digestable and non-utilizable forms of d-amino acids, it is preferable to deamidate at pH 8 and 100°C to improve the solubility of RBP samples used as a food ingredient.

Fig. 5.

Some D-amino acid contents of deamidated rice bran proteins.

RBP (0.25 g/15 mL) was treated at various pH at 100 or 120°C for 30 min. D-amino acids were determined by chiral-HPLC method. All measurements are expressed as means ± SD of three independent experiments.

The production of LAL is another disadvantage of heating proteins under alkaline conditions (Schwass et al., 1984; Hou et al., 2017). LAL could be detected by the amino acid analysis system described above. The LAL peak came after the Phe peak and just before the His peak. As for the other amino acids, the amount of LAL was determined according to the peak area. LAL content of non-deamidated RBP was 0.4 µmol/g and increased with the deamidation process (Fig. 6). LAL was determined to be 0.8 – 5 µmol/g following heating at 100°C or 120°C at pH 8-10. When heating at pH 12, the LAL content increased to 11 µmol/g at 100°C and 19 µmol/g at 120°C. In this case, it was calculated that approximately 5% of Lys was consumed in the generation of LAL. The decrease of Lys or other amino acids did not affect the amino acid composition of RBP samples, as shown in Table 3. On the basis of these results, it is concluded that the solubility of RBP can be increased by deamidation at pH 8 and 100°C for 30 min with minimal amino acid racemization and LAL production. Moreover, high contents of d-amino acids and LAL of highly deamidated RBL should be advantageous for non-food applications such as surfactants, adhesives, coating and plastics because of its resistance to enzymatic digestion (De Graaf, 2000). Deamidated RBL of various degrees of water solubility may broaden both its food and non-food applicability as an under-utilized protein resource. These applications include matrices for bioplastics, controlled-release devices, or enzyme immobilization, as well as for films and coatings (Jerez et al., 2007).

Fig. 6.

LAL contents of deamidated rice bran proteins. RBP (0.25 g/15 mL) was treated at pH 8, 10, or 12 at 100 or 120°C for 30 min. All measurements are expressed as means ± SD of three independent experiments.

In this study, the solubility of RBP was improved by deamidation under alkaline conditions. Although solubility was improved by increasing the pH for deamidation, the rates of constituent amino acid racemization and LAL generation also significantly increased. The degree of deamidation was found to be a major factor in improving the solubility of RBP. The decrease of molecular mass or the degradation of peptide bonds could not be detected in alkaline deamidated RBP. The thermal property and secondary structure of deamidated RBP measured by DSC and FTIR suggest that the compact structure of RBP was well preserved during the deamidation process. Taken together, the optimal condition for deamidation was heating at pH 8 and 100°C for 30 min to give deamidated RBP of moderate solubility with minimal side reactions.

Acknowledgement    This work was supported by JSPS KAKENHI grant no. 26292111.

References
 
© 2017 by Japanese Society for Food Science and Technology
feedback
Top