Development of β-conglycinin quantitative method in soybean foods using capillary electrophoresis-based immunoassay

Kazuhiro Fujita; Megumi Suzuki; Kazushi Mizukoshi; Yushi Takahashi; Toshiaki Yokozeki; Izumi Yoshida; Mari Maeda-Yamamoto

doi:10.3136/fstr.FSTR-D-23-00135

Abstract

β-Conglycinin (β-CG), a major soy protein, has been associated with the reduction of body fat and triglycerides in the blood. Quantitation of the α, α′, and β subunits in β-CG is desirable to establish the functional properties of β-CG in soybean. Therefore, in this study, we used the Jess System, a capillary electrophoresis-based immunoassay system, for quantifying β-CG in soybean foods. The validity of this method was confirmed using soybeans and processed soy milk. The within-laboratory reproducibility of the total β-CG amount was < 15 %, and its trueness was > 80 %. The β-CG content in the commercially available soybeans, Kori-tofu, soy milk, soy yogurt, Okara powder, and soy meat was determined to be 40.3–148.5, 65.2–97.2, 6.1–7.8, 3.5, 18.0, and 57.5 mg/g, respectively. For the first time in the literature, this method enables the quantitation of individual subunits of β-CG, despite the large variation in the amount of each subunit.

Introduction

Soybean (Glycine max L.) protein is used in a variety of foods, including traditional Asian and Western cuisine. In recent years, soy meat has become a common alternative protein source for livestock and marine products (Nakano, 2021; Zhang et al., 2021; Messina et al., 2022; Ou et al., 2023). Soy proteins have unique physiological functions, including serum lipid-lowering effects (Sirtori et al., 1977; Nagata et al., 1982; Sugano et al., 1993; Anderson et al., 1995; Kito et al., 1993; Fukui et al., 2002). β-Conglycinin (β-CG), a major soy protein, has been reported to reduce body weight in mice. Moreover, in a randomized, double-blind, placebo-controlled human study, candy containing 88 % pure β-CG significantly reduced body weight and visceral fat in subjects compared to the placebo (Cope et al., 2008; Udenigwe et al., 2015). In addition, based on this evidence, foods containing β-CG as a functional ingredient have been submitted in Japan as “Foods with Function Claims” (Sato et al., 2008). On the other hand, β-CG has been reported as a putative soybean allergen in some patients with soybean allergy; therefore, a quantitation method for β-CG in soybean foods must be developed (Li et al., 1991; Ogawa et al., 1995; Tryphonas et al., 2003).

β-CG is a heterotrimer composed of α (72 kDa), α′ (76 kDa), and β (53 kDa) subunits with molecular heterogeneity. Six different isomeric species have been reported: α′β2, αβ2, αα′β, α2β, α2α′, and α3 (Thanh et al., 1976). However, the functional properties of α, α′, and β subunits are different. The α′ subunit has been associated with a reduction in cholesterol in plasma and with triglyceride levels in hypercholesterolemic rats (Scarafoni et al., 2007). In addition, residues 173–185 of the α′ subunit, known as soymatide, exhibited immunostimulatory properties (Guijarro-Díez et al., 2013). Therefore, the selective quantitation of the α, α′, and β subunits of β-CG is crucial for evaluation of the functional properties of β-CG in soybeans.

Enzyme-linked immunosorbent assay (ELISA) and liquid chromatography with tandem mass spectrometry (LC-MS/MS) have previously been used to analyze β-CG in soybean foods. In direct ELISA, β-CG is immobilized on a microtiter plate and bound to a polyclonal antibody (Moriyama et al., 2005). In sandwich ELISA, a polyclonal antibody is immobilized on a microtiter plate, and β-CG is sandwiched between enzyme-labeled monoclonal antibodies for measurement (Hei et al., 2012). A commercially available ELISA kit for allergen analysis has also been used to quantitate β-CG (Moriyama et al, 2020). The reported LC-MS/MS method (Ippoushi et al, 2019) is a useful method for selecting and quantifying specific peptide fragments from the tryptic digest of β-CG; however, peptide fragments specific for the α′ subunit cannot be obtained. Therefore, the quantitative value of the α′ subunit was evaluated as the total amount with the α subunit. Recently, a device for automatically performing Western blotting (WB) was developed for biochemical applications (Lee et al., 2021; Li et al., 2021; Kylies et al., 2022; Crawford et al., 2022; Mir et al., 2022; Mimatsu et al., 2023). This device separates proteins by capillary electrophoresis, immobilizes them on the capillary via irradiation with ultraviolet light, and binds to a primary antibody specific to the target protein. Next, an enzyme-labeled secondary antibody was bound to the primary antibody and reacted with a luminescent reagent for detection. In addition to conventional qualitative analysis, this device is applicable to quantitative analysis but is less accurate than ELISA.

To overcome all the above-mentioned issues, in this study, we developed a quantitation method for α, α′, and β subunits of β-CG using a capillary electrophoresis-based immunoassay. Primary antibodies specific to each subunit were prepared using peptide antigens selected by in silico analysis. Commercially available high-purity β-CG, typically used for allergen analysis, was used as the analytical standard.

Materials and Methods

Samples Four varieties of soybeans (Fukuyutaka, Tsurunoko, Nanahomare, and Toyoshirome), two types of Kori-tofu (brands 1 and 2), two types of unprocessed soy milk (brands 3 and 4), two types of processed soy milk (brand 3), soy meat, soy milk yogurt, and a soy drink were used. Nanahomare was provided by Minami-Sangyo Co., Ltd. The other soy products were purchased from retail outlets in Japan.

Solid foods such as soybeans were ground in a mill and passed through a sieve with an opening of 250 μm, and liquid foods such as soy milk were mixed well.

Apparatus The Jess System (ProteinSimple, Bio-Techne, San Jose, CA, USA) was used as the capillary electrophoresis-based immunoassay system. Capillary cartridges (ProteinSimple, Bio-Techne; 2-40 kDa, 8 × 25 and 12-230 kDa, 8×25) were used for the luminescence separation module.

Chemicals and reagents Natural Gly m 5 (purity: > 95 % by SDS-PAGE, 1.3 mg/mL by BCA; 0.25 mg as β-CG containing 28 % α,18 % α′, and 52 % β subunits) was purchased from Indoor Biotechnologies Inc. (Charlottesville, VA, USA) and used as the reference standard. Note that Gly m 5 is the allergen name for β-CG. SDS solution (10 %) and 1 mol/L Tris-HCl (pH 8.0) were purchased from NIPPON GENE Co. Ltd. (Tokyo, Japan). Can Get Signal Immunoreaction Enhancer Solution 1 (CIES1, for primary antibody) was purchased from TOYOBO Co., Ltd. (Osaka, Japan). Anti-Rabbit Detection Module was purchased from ProteinSimple, Bio-Techne, and Normal Goat Serum (10 %) was purchased from Kirkegaard & Perry Laboratories, Inc. (Gaithersburg, MD, USA). Analytical-grade sodium sulfite, acetone, and glycerin were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan).

The water used was double-distilled and purified using a Milli-Q system (Merck KGaA, Darmstadt, Germany).

The extract buffer solution was prepared using 10 mL of 1 mol/L Tris-HCl buffer (pH 8.0), 10 mL of 1 mol/L sodium sulfite, 8 mL of 10 % SDS, and water to make 100 mL.

The dilution buffer solution was prepared by diluting the extract buffer solution 8-fold with 0.1 mol/L Tris-HCl buffer (pH 8.0).

Stock solutions (500 μg/mL) of the reference standard were prepared by diluting Natural Gly m 5 to 0.5 mL by adding 0.1 mol/L Tris-HCl buffer (pH 8.0). This solution was divided into 12.5 μL aliquots and stored frozen. Working standard solutions were prepared daily by diluting the aliquots with the dilution buffer solution.

Preparation of anti-β-CG antibodies Three polyclonal rabbit antibodies were prepared against β-CG (α, α′, and β subunits) and used as primary antibodies in the Jess system.

These polyclonal antibodies were produced by designing antigens whose epitopes were peptide sequences specific to each subunit by in silico analysis and commissioned to COSMO BIO Co., Ltd. (Tokyo, Japan). The same volume of glycerol was added to each of the obtained antibodies and mixed. The resultant solution was divided into 0.5 mL aliquots and stored at −20 °C. This stock solution was diluted 40-fold with CIES1 and used for the measurements.

An anti-rabbit horseradish peroxidase (HRP) secondary antibody (ProteinSimple, Bio-Techne) was used as the secondary antibody in the Jess System.

Sample preparation For solid samples, 0.2 g of sample was weighed and mixed with 0.5 mL of water using a vortex mixer. For liquid samples, 2 g of sample was weighed. Then, 30 mL of acetone was added to the weighed sample. The resultant dispersion was shaken (20 min) to remove fat and cooled to 5 °C for 30 min. The dispersion was centrifuged (10 000 × g, 10 min, 5 °C), the supernatant was removed by decantation, and the residue was left to dry in a fume hood for 4 h. The extract buffer solution (20 mL) was added to the residue, mixed with a vortex mixer, and shaken (ambient, 16 h) on a horizontal shaker. The resultant solution was subjected to ultrasonic treatment (30 min) and centrifuged (10 000 × g, 5 min, 20 °C). The supernatant was filtered through a syringe filter (0.45 μm, PP). The filtrate was diluted 8-fold with 0.1 mol/L Tris-HCl buffer (pH 8.0) and further diluted appropriately with the dilution buffer solution to prepare the test solution.

Measurements using the Jess System The measurement was performed according to the “Jess Protocol Flowchart” attached to the apparatus. The reference standard and test solutions were mixed in a 4:1 ratio with a fluorescence master mix (ProteinSimple, Bio-Techne) and heated to 95°C for 5 min. The samples (reference standard and test solutions), blocking reagent, primary antibodies, HRP-conjugated secondary antibodies, and chemiluminescent substrate were dispensed into the designated wells of a 25-well plate. The plate was subjected to separation electrophoresis and immunodetection in a fully automated system. Analysis in the Jess System was conducted under ambient conditions, and the default instrument settings were used with the exception of the blocking time. Capillaries were filled with a separation matrix, followed by a stacking matrix and ~40 nL of sample loading. During electrophoresis, proteins were separated based on molecular weight through stacking and separation matrices at 375 volts for 25 min and immobilized on the capillary wall using proprietary photo-activated capture chemistry. The matrices were washed again. Then, the capillaries were incubated with a blocking reagent for 30 min. The target proteins were probed with primary antibodies and incubated with HRP-conjugated secondary antibodies. Luminol and peroxide (ProteinSimple, Bio-Techne) were added to generate chemiluminescence, which was captured using a CCD camera. Digital images were analyzed using Compass software (ProteinSimple, Bio-Techne), and the quantified data of the detected proteins were reported as molecular weights and signal/peak intensities. One lane was used for α and β subunit measurements, and one lane was used for α′ subunit measurement. In addition, normal goat serum (10 %) was used as the blocking reagent, and the dispensed volume for line A was 4 μL.

Quantitation The calibration curve for each β-CG subunit was drawn from the concentration and peak area of the electropherogram using a quadratic regression curve. The obtained calibration curve was used to calculate the concentration in the test solution. The concentration of each subunit in the sample was calculated by considering the sample amount and dilution ratio. The results were used to obtain the β-CG content.

Method validation Precision and accuracy were evaluated to confirm the validity of the developed method using soybeans (Fukuyutaka) and processed soymilk (brand 3). The precision was determined by evaluating two samples per day on five different days. Trueness was determined via fortified recovery tests, where two samples of ~40 mg/g of soybeans and ~4 mg/g of processed soymilk were evaluated per day on five different days. For trueness, as the analytical standard Natural Gly m 5 is expensive, β-CG was purified from a mixture of Nanahomare and Tsurunoko soybeans (purified β-CG) using a previously reported method (Samoto et al., 2007), and the concentration was confirmed under the measurement conditions of this method before use. Recovery tests were performed by aliquoting 1.67 mL of purified β-CG solution (5.0 mg/mL) into 50 mL centrifuge tubes, concentrating to dryness using a centrifugal evaporator, and then weighing the sample. The subsequent steps were performed according to the sample preparation procedure described above.

Precision results were calculated using one-way analysis of variance (ANOVA). Recovery rates were calculated by subtracting the average precision values.

Results and Discussion

Preparation of anti-β-CG antibodies Five peptide antigen sequences for each subunit were obtained using in silico analysis (Table 1). Two peptides with a highly hydrophilic region (to avoid hydrophobic bonding in solution) and a disordered region (whose structure is less affected by environmental factors) were selected. The selected peptides, α-Rank4 (α-4), α-Rank5 (α-5), α′-Rank2 (α′-2), α′-Rank5 (α′-5), β-Rank1 (β-1), and β-Rank5 (β-5), were combined with the carrier protein keyhole limpet hemocyanin (KLH) and rabbits were immunized with these to obtain antisera. The applicability of the obtained serum (polyclonal antibodies) to the Jess System was confirmed using purified β-CG.

Table 1. Peptide antigen sequences for each subunit by in silico analysis

Subunit	Rank	Position	Length	Sequence (N→C)	M.W.	a/h ^a	c/d ^b
α	1	584–598	15	KKKEEGNKGRKGPLS	1655.88	−1.987	0.625
	2	395–411	17	SRKTISSEDKPFNLRSR	2022.26	−1.317	0.500
	3	70–86	17	EGEIPRPRPRPQHPERE	2080.31	−2.228	0.611
	4	127–141	15	RKEEKRGEKGSEEED	1805.87	−2.769	0.875
	5	160–176	17	HQKEERKQEEDEDEEQQ	2214.20	−3.250	0.833
α′	1	599–611	13	PQQKEEGNKGRKG	1455.56	−2.429	0.571
	2	135–151	17	KGSEEEQDEREHPRPHQ	2088.14	−2.733	0.722
	3	199–213	15	SQREPRRHKNKNPFH	1932.13	−2.525	0.625
	4	175–187	13	EEEEEDQDEDEEQ	1652.48	−3.071	0.929
	5	123–139	17	HEWHRKEEKHGGKGSEE	2060.15	−2.428	0.778
β	1	420–434	15	KEEGSKGRKGPFPSI	1617.82	−1.213	0.500
	2	222–234	13	SRRAKSSSRKTIS	1463.65	−1.229	0.500
	3	196–208	13	LFGEEEEQRQQEG	1579.61	−1.729	0.571
	4	312–326	15	KEQQQKQKQEEEPLE	1899.03	−2.625	0.625
	5	24–36	13	LKVREDENNPFYL	1637.81	−0.836	0.500

^a Average hydrophobicity

^b Charge density

α-4, α′-2, and β-5 antibodies were selected owing to their good sensitivity and selectivity. However, the β-5 antibody showed some cross-reactivity with the α subunit (Fig. 1). Furthermore, as the α and α′ subunits have similar molecular weights, separating them was difficult in the Jess System. Hence, the cross-reactivity of α-4 and α′-2 antibodies was evaluated using conventional Western blotting. They did not interact with each other (data not shown). Based on these results, we decided to measure the α and β subunits in the same lane and the α′ subunit in another lane on the Jess System.

Fig. 1

Comparison of the sensitivity and concentration of the prepared anti-β-conglycinin (CG) antibodies determined using the Jess System, a capillary electrophoresis-based immunoassay. Lane 1–3: antibody, α′-2; dilution, 1:1 000, 1:200, 1:40; Lane 4–6: antibody, α′-5, dilution, 1:1 000, 1:200, 1:40; Lane 7–9: antibody, α-4; dilution, 1:1 000, 1:200, 1:40; Lane 10–12: antibody, α-5; dilution, 1:1 000, 1:200, 1:40; Lane 13–15: antibody, β-1; dilution, 1:1 000, 1:200, 1:40; Lane 16–18: antibody, β-5; dilution, 1:1 000, 1:200, 1:40.

Variation between the luminescence separation module lanes Initially, a 12-230 kDa capillary cartridge was used for the luminescence separation module, considering the molecular weight of β-CG. Even though the manufacturer’s specifications stated that the RSD % was within 15 %, we observed that the variation between lanes became problematic. Therefore, the variation was confirmed by measuring the α and β subunits in all lanes for two lots of each capillary cartridge, including the 2-40 kDa capillary cartridge for small molecules. The marker molecular weights of the 2-40 kDa capillary cartridge are 2, 5, 12, 26, and 40 kDa, which do not cover the molecular weight of β-CG. However, even if the upper limit of the marker is exceeded, migration data can be obtained, and the setting for the molecular weight range to be analyzed by software can be changed. Therefore, we decided to determine the β-CG fraction from the migration pattern of a 12-230 kDa capillary cartridge, and change the molecular weight range for analysis of a 2-40 kDa capillary cartridge.

As a result, we chose to use the 2-40 kDa capillary cartridge because it showed less variation than the 12-230 kDa capillary cartridge, as shown in Table 2.

Table 2. Confirmation of lane-to-lane variation in the luminescence separation module.

Applicable molecular weight	Lot. No.	Subunit	RSD ^a (%)
12-230 kDa	93615	α	13.6
		β	14.1
12-230 kDa	73640	α	15.9
		β	19.8
2-40 kDa	26801	α	3.8
		β	3.4
2-40 kDa	48950	α	5.2
		β	3.2

^a Relative standard deviation, n=24

Optimization of the Jess System measurement conditions The optimum primary antibody concentration was estimated for the Jess System using 5, 10, 20, and 80 μg/mL of β-CG standard solutions (total β-CG). Good results were obtained for all antibodies when a solution of the primary antibody diluted 80 times with CIES1 was used. Typical electropherograms are shown in Fig. 2. Although a linear regression was not obtained for the calibration curve, a good correlation was obtained using quadratic curve regression (Fig. 3). Additionally, since the α′ subunit has low quantification at high concentrations, we decided to use quantitative values at concentrations up to 10 μg/mL.

Fig. 2

Typical electropherograms of β-CG standard solutions. (a) Lane 1, 14.4 μg/mL of α′ subunit; Lane 2, 7.2 μg/mL of α′ subunit; Lane 3, 3.6 μg/mL of α′ subunit; Lane 4, 1.8 μg/mL of α′ subunit; Lane 5, 22.4 μg/mL of α subunit and 41.6 μg/mL of β subunit; Lane 6, 11.2 μg/mL of α subunit and 20.8 μg/mL of β subunit; Lane 7, 5.6 μg/mL of α subunit and 10.4 μg/mL of β subunit; Lane 8, 2.8 μg/mL of α subunit and 5.2 μg/mL of β subunit. (b) 3.6 μg/mL of α′ subunit corresponding to Lane 3. (c) 5.6 μg/mL of α subunit and 10.4 μg/mL of β subunit corresponding to Lane 7.

Fig. 3

Typical calibration curve for β-CG.

Sample preparation Hexane, petroleum ether, and diethyl ether were used for degreasing samples in previous studies (Samoto et al., 2007; Hei et al., 2012; Moriyama et al., 2020). However, water-containing samples, such as soy milk, were separated into two layers and were difficult to handle. Therefore, we chose acetone, which is miscible with water. Unprocessed soy milk (2 g) was weighed and defatted by adding 20, 30, or 40 mL of acetone, and β-CG was quantified and compared. As a result, 30 mL of acetone was considered sufficient, as the highest value was obtained with 30 mL of acetone. After defatting, β-CG was extracted with 0.1 mol/L Tris-HCl buffer (pH 8.0) containing SDS and sodium sulfite by shaking at room temperature for 16 h, according to the allergen inspection method of the Japanese food labeling standardsⁱ⁾.

Method validation The precision results are presented in Table 3. The within-laboratory reproducibility (RSD_wr) of individual subunits (particularly the β subunit) in soybean (Fukuyutaka) was high. However, the total value, calculated from the sum of each subunit, was satisfactory (10.1 %). Notably, the RSD_wr of individual subunits in processed soy milk (brand 3) was high, but the total value was low (13.5 %), showing the same trend as for soybean. The reason for the small RSD_wr in the total value was thought to be that the obtained standard deviation was not very large compared to the standard deviation of each individual subunit, and the average of the total value, which served as the denominator, became large. Although the variation in values was large compared to that produced by chemical analytical methods such as chromatographic analysis (AOAC, 2012), these results were reasonable considering the use of Western blotting and the variation among the lanes of the capillary cartridge.

Table 3. Results of precision.

Sample	Subunit	Mean^a (mg/g)	RSD_r^b (%)	RSD_wr^c (%)
Soybean (Fukuyutaka)	α	13.3	11.3	11.3
	β	15.2	21.5	22.1
	α′	7.3	6.0	12.7
	Total	35.8	8.6	10.1
Processed soy milk-1 (Brand 3)	α	1.8	17.0	24.8
	β	1.4	26.8	26.8
	α′	0.9	10.7	15.5
	Total	4.1	11.5	13.5

^a Mean of 10 replicates

^b Relative standard deviation (Repeatability)

^c Relative standard deviation (Within-laboratory reproducibility)

The trueness results for soybeans (Fukuyutaka) and processed soy milk (brand 3) were 81.5 % and 88.4 %, respectively (Table 4). RSDwr, which was calculated in the same manner as precision, showed the same tendency as precision.

Table 4. Results of trueness (recovery tests).

Sample	Subunit	Mean of fortified^a (mg/g)	RSD_r^b (%)	RSD_wr^c (%)	Mean of not fortified^a (mg/g)	Fortified (mg/g)	Recovery (%)
Soybean (Fukuyutaka)	α	27.3	8.5	10.8	13.3	16.55	84.6
	β	26.6	20.3	20.3	15.2	14.45	78.9
	α′	16.0	9.0	14.5	7.3	10.85	80.2
	Total	69.9	8.5	8.8	35.8	41.85	81.5
Processed soy milk-1 (Brand 3)	α	3.3	9.9	15.5	1.8	1.655	90.6
	β	2.6	18.6	22.9	1.4	1.445	83.0
	α′	1.9	16.3	18.7	0.9	1.085	92.2
	Total	7.8	8.5	12.4	4.1	4.185	88.4

^a Mean of 10 replicates

^b Relative standard deviation (Repeatability)

^c Relative standard deviation (Within-laboratory reproducibility)

Application to commercial samples The results for the commercially available samples are shown in Table 5, and typical electropherograms are shown in Fig. 4. The β-CG content determined in soybeans, Kori-tofu, soy milk, soy yogurt, soy milk, soy yogurt, Okara powder, and soy meat was 40.3–148.5, 65.2–97.2, 6.1–7.8, 3.5, 18.0, and 57.5 mg/g, respectively. Moreover, individual subunits (especially the β subunit) exhibited large variability, similar to the validation results. However, the total RSD_wr value in all cases was within 15 %, except for that for soybean (Toyoshirome).

Table 5. Determination of β-conglycinin contents in the various soybean foods.

Sample	Subunit	Mean^a (mg/g)	RSD_r^b (%)
Soybean (Tsurunoko)	α	27.3	4.1
	β	26.7	26.9
	α′	19.6	14.3
	Total	73.6	7.8
Soybean (Nanahomare)	α	58.6	7.7
	β	47.2	26.6
	α′	42.7	12.0
	Total	148.5	6.4
Soybean (Toyoshirome)	α	12.7	10.4
	β	18.4	19.0
	α′	9.2	25.1
	Total	40.3	17.1
Kori-tofu (Brand 1)	α	31.6	21.9
	β	16.2	15.7
	α′	17.4	5.4
	Total	65.2	9.7
Kori-tofu (Brand 2)	α	41.8	3.9
	β	32.7	25.8
	α′	22.7	6.0
	Total	97.2	7.9
Soy meat	α	21.7	15.3
	β	26.4	16.4
	α′	9.5	14.1
	Total	57.5	11.6
Okara powder	α	7.4	4.0
	β	4.9	22.6
	α′	5.7	4.1
	Total	18.0	6.5
Unprocessed soy milk (Brand 3)	α	2.4	3.2
	β	2.2	19.0
	α′	1.6	3.5
	Total	6.3	5.8
Unprocessed soy milk (Brand 4)	α	3.4	6.7
	β	2.9	17.6
	α′	1.5	4.5
	Total	7.8	4.7
Processed soy milk-2 (Brand 3)	α	2.4	6.9
	β	2.2	16.4
	α′	1.5	5.6
	Total	6.1	6.7
Soy yogurt	α	1.6	23.8
	β	1.3	12.8
	α′	0.6	20.6
	Total	3.5	10.1
Soy drink	α	3.5	15.1
	β	2.5	18.6
	α′	1.9	9.1
	Total	7.9	5.1

^a Mean of 4 replicates

^b Relative standard deviation (Repeatability)

Fig. 4

Typical electropherograms of commercial samples. (a) α′ subunit of soybean (Tsurunoko), (b) α and β subunits of soybean (Tsurunoko), (c) α′ subunit of Kori-tofu (Brand 2), (d) α and β subunits of Kori-tofu (Brand 2), (e) α′ subunit of unprocessed soy milk (Brand 4), and (f) α and β subunits of unprocessed soy milk (Brand 4).

These quantitation results were slightly lower than those reported in a previous ELISA study (Moriyama et al., 2020), which could be attributed to two factors. High-purity β-CG was used as the analytical standard in this study. In contrast, in the previous study, β-CG was purified from soybeans, and the sample with a purity of ≥ 85 % as evaluated by SDS-PAGE was used as the analytical standard. This difference in purity explains the different quantitation results. Furthermore, individual subunits were used for quantitation in our method. Such quantitation was not performed for each subunit in the previous study, suggesting that the subunit composition may have differed between the analytical standard and the sample.

Conclusions

A quantitation method for β-CG in soybean foods was developed using the Jess System, a capillary electrophoresis-based immunoassay. Peptide antigens considered specific to each β-CG subunit were selected for in silico analysis. Polyclonal antibodies against these proteins were prepared and used as primary antibodies in the measurement system. Commercially available, high-purity β-CG was used as the analytical standard, and each subunit was quantified. The precision and trueness of the estimated total β-CG amount via our method were evaluated using soybean (Fukuyutaka) and processed soy milk, within a laboratory reproducibility of 15 % and trueness of > 80 %. To demonstrate the applicability of our quantitation method, the β-CG content in commercial samples was measured. The β-CG content in soybeans, Kori-tofu, soy milk, soy yogurt, Okara powder, and soy meat was determined to be 40.3–148.5, 65.2–97.2, 6.1–7.8, 3.5, 18.0, and 57.5 mg/g, respectively. These quantitative results were slightly lower than those of a previous ELISA report. This may be attributed to the high-purity β-CG analytical standard used in our study and the quantitation of each subunit. Even though our method exhibited a large variation in the amount of each subunit, the variation in the total β-CG amount is almost within the manufacturer’s specifications. Notably, our method enables unprecedented individual quantitation of each subunit.

Acknowledgements This study was supported by the Strategic Innovation Promotion Program (SIP) and “Technologies for Smart Bio-industry and Agriculture” (funding agency: Bio-oriented Technology Research Advancement Institution).

Conflict of interest There are no conflicts of interest to declare.

References

Corresponding author

Register with J-STAGE for free!