The quality changes and microflora analysis of commercial instant soya milk

Abstract Soya milk can be easily contaminated because of its rich nutritional profile and simple package form, which thus may lead to short shelf life and has been considered as a public health concern. The objective of this study is to investigate the changes of sensory quality, viscosity, pH values, bacteria, and protein denaturation in soya milk samples which were stored for 0, 4, 8, 12, and 24 h at 25 ± 2°C. The results showed that the sensory properties were on the decline along with the storage time. The viscosity value increased from 1.61 Pa.s to 2.50 Pa.s, while the pH value decreased from 6.87 to 6.61. In addition, the number of total bacteria and Lactobacillus increased and the protein in soya milk denatured continually. The 16S rDNA sequence analysis indicated that the main microbial strains were Rummeliibacillus, Acinetobacter, Enterobacter, Phyllanthus, Bergia, Zhihengliuella, and Nesterenkonia. In this study, statistics in producing, packaging, and stocking process of soya milk were given, which provided experience to controlling soya milk.


Introduction
Soya milk has been a traditional drink for breakfast in Southeast Asia. It is made from soaked soybeans by grinding, heating, and filtering. Soybean-based products are very popular and have attracted worldwide interest. China, the largest importing country to domesticate soybeans and a major global soybean grower and consumer, has extensive distributions of soybean accessions (Li et al. 2014). Recently, researchers are interested in the physiologically beneficial phytochemicals from soybeans and soya products (Jooyandeh 2011). Isoflavones, an important component of soybeans and soya products, has been reported to protect against cardiovascular diseases, breast cancer, prostate cancer, testicular cancer, uterine cancer and other hormonedependent cancers, and osteoporosis (Patisaul and Jefferson 2010). In addition, soya bean and soya products are good sources of phytosterols, polyamines, and tocopherols, which has been reported to help control cholesterol, protect against various types of injuries, and has showed strong antioxidant activity (Marangoni and Poli 2010;Bae et al. 2008;Niki and Noguchi 2004). Soya milk is the biggest soya-based product consumed in the world, not only because of its potential health benefits but also as an alternative to cow milk targeting for lactose-intolerant individuals, those allergic to milk proteins, or those avoiding consuming milk for other reasons (Reilly et al. 2006). Soya milk occupies an important position in China health program for breakfast. Commercial soy beverages in glass bottles are sterilized batch wise in retorts or continuously in a hydrostatic sterilizer to achieve commercial sterility (Kwok and Niranjan 1995). At present, soya milk in the market basically has instant soy milk powder and instant soya milk. When soy milk powder is dissolved in hot or boiled water, the protein contained was denatured and the reconstituted soya milk has beany flavor.
In China's health project for breakfast, plastic packaging of soya milk has been widely used. Soya milk, which is rich in nutrition, may promote bacteria growth when contamination happens. Lactobacillus was reported to be one of the most common spoilage and pathogenic microorganisms presented in soya products (Odu and Egbo 2012). The contamination of these microorganisms in soya milk during processing or stocking will result in a potential risk for public health and a reduced shelf life. So, the contamination of soya milk should be tightly controlled and monitored.
In this study, we aimed to identify the changes of sensory perception, pH, viscosity from soya milk samples kept at 25 ± 2°C for 0, 4, 8, 12, and 24 h. Moreover, bacteriological parameters were also monitored during the experimental period. The results may help both the producers and consumers to better understand the quality control of soya milk.

Soya milk samples treatments
Soya milk samples were purchased from Xi'an Commercial soya milk. The soya milk samples without opening the package were stored for 0, 4, 8, 12, and 24 h at 25 ± 2°C, respectively.

Sensory analysis of soya milk samples
To present the samples to the panelists, soy milk was placed in white cups, with blind codes with two digits. Each panelist in his/her white-lighted booths received his/her samples in a random presentation order and answered the questions at their computers. Between the samples given, the mouth was cleaned with water. To analyze flavor and texture, the panelists tasted an adequate amount of food samples (Ghosh and Chattopadhyay 2012). All the food samples were presented by pair evaluated in color, bean flavor, beany, quality, flavor, and overall acceptability. The principle of Quantitative Descriptive Analysis (QDA) is based on the ability to train panelists to measure specific attributes of a product in a reproducible manner to yield a comprehensive quantitative product description amenable to statistical analysis. Panelists were recruited from the staff members of College of Food Engineering and Nutritional Science, Shaanxi Normal University with a focus to identify key product attributes. This group of panelists is then trained to reliably identify and score product attributes. As panelists generate the attribute terms, the resulting descriptions are meaningful to consumers, and thus, analysis provide information amenable to modeling predictions of consumer acceptability. QDA results are analyzed statistically and then represented graphically.

Measurement of pH
The pH value was determined in triplicates for each sample by a digital pH meter (PHSJ-3F, ex Electric Chemical; Shanghai, China).

Viscosity analysis
The viscosity of soya milk sample was measured by viscometer (RVDV-II+Pro; Brookfield, USA). All tests were performed at 5.0 ± 0.02°C. The samples were presheared for 5 sec at a controlled shear rate of 2001/sec to create a dispersion of uniform properties allowed to equilibrate for 30 sec before measuring intrinsic viscosity (Purwandari et al. 2007). The measurements were then performed at a constant shear rate of 25.81/sec for 60 sec according to previous report (Burkus 2003). The viscosity values were monitored.

Bacteria and Lactobacillus enumeration
Colony-forming units (CFUs) of bacteria and molds were counted using a plate-counting method. Soya milk sample (1 mL) was dispersed in 9 mL of sterile saline solution and oscillated for 1 min to achieve a 10-fold dilution. A sterile cotton wool was wetted by sterile saline, the inside wall of a soya milk bottle cap was wiped, and then 10 mL sterile saline was added to the test samples. The dilution with sterile saline solution was repeated twice to achieve three continuous dilutions. A 0.1 mL aliquot from each serial dilution was spread over an agar plate with plate count agar (PCA) medium for molds, and over a plate of de Man, Rogosa, and Sharpe (MRS) for Lactobacillus. The plates were incubated at 37°C for 48 h. CFUs, which varied from 10 to 100 for molds and from 30 to 300 for bacteria, were counted.

Isolation of total microbial DNA
For microbial genomic DNA extraction, 10 mL of each sample were seeded in 100 mL of basic culture medium and incubated overnight in a shaker. Total DNA was isolated by using Bacteria Gen DNA Kit (Cwbio, Beijing, China) according to the manufacturer's instructions.

Electrophoretic and identification of bands
Denaturing gradient gel electrophoresis (DGGE) was performed using a DCode apparatus (Bio-Rad, Richmond, CA) at 60°C and employing 8% polyacrylamide gels with a denaturing range of 40-60% for total bacteria. Electrophoresis was performed at 75 V for 16 h and 130 V for 4.5 h for bacteria. Bands were visualized under UV light after staining with ethidium bromide (0.5 mg mL −1 ) and photographed.
Bands in the gels were identified by sequencing. Bands were excised from the gels and set to sequence (Sangon Biotech, Shanghai, China). The identity of the sequences was determined by the BLASTN algorithm in the GenBank database (http://www.ncbi.nlm.nih.gov/BLAST/).

16S rDNA gene amplification conditions
Amplifications were performed using the following PCR conditions: V3 region primers conditions: 94°C for 2 min, 30 cycles of 94°C for 30 sec, 60°C for 30 sec and 72°C for 30 sec, and a final extension step at 72°C or 2 min. V3-V5 region primers conditions: 94°C for 2 min, 30 cycles of 94°C for 30 sec, 50°C for 30 sec and 72°C for 30 sec, and a final extension step at 72°C for 2 min. The sequences obtained were uploaded at the NCBI Sequence Read Archive.

Sequence treatment and bioinformatics analysis
The MOTHUR program was also used to perform the Fast UniFrac test, which was employed to compare the phylogenetic structure of the libraries and to generate the Venn diagrams. A neighbor-joining tree was constructed with representative sequences of each Optical Transform Unit (OTU) selected by MOTHUR. These sequences were compared against the RDP database using the Seqmatch option to select for the nearest neighbors. All sequences were then aligned using MEGA 5.0 software (Tokyo, Japan) (Tamura et al. 2011) and the Jukes-Cantor model.

Protein analysis with SDS-PAGE
Buffer A: 0.2 mol/L Na 2 HPO 4 (31.2 g Na 2 HPO 4 . 12 H 2 O add distilled water to 1000 mL and mixed well); buffer B: 0.2 mol/L NaH 2 PO 4 (35.6 g NaH 2 PO 4 .2 H 2 O add distilled water to 1000 mL and mixed well); Phosphate buffer (PB): 0.2 mol/L PB (pH 7.4, 19 mL buffer A added to 81 mL buffer B and mixed well).
For total protein extraction, each soya milk sample (10 g) was transferred into a hydrolysis tube with 100 mL of PB, and stir-extracted for 60 min on ice, and then centrifuged (15,000 × g/min, 45 min) at 4°C. The supernatant was centrifuged again, and the supernatant was saved as soya milk protein samples.

SDS-PAGE analysis
Protein samples were added with 2% of 2-mercaptoethanol and heated in boiling water bath (100°C) for 3 min. The samples (8 or 16 μL) were loaded to SDS-PAGE with 16% separating gel according to the method by Schägger (2006). The samples were electrophoresed at constant voltage of 30 mV until all samples entered into the stacking gel, and then at constant voltage of 100 mV until end. After electrophoresis, the gel was fixed with a solution of 100 mmol/L ammonium acetate dissolved in methyl alcohol/acetic acid (5/1, v/v) for 2 h. After fixing, the gel was stained with 0.025% (w/v) Coomassie Blue G-250 in 10% (v/v) acetic acid for 2 h, and destained by 10% (v/v) acetic acid. The band intensities on gel were analyzed by Bio-Rad Image Lab Software.

Statistical analysis
Non-parametrical Kruskal-Wallis analysis was performed. The least significant difference test was employed to determine differences between means at a 5% significance level. Results were analyzed using the Statistical Software Package for Windows PASW Statistic 20.0 (SPSS, Chicago, IL).

Sensory analysis
Standards used in the sensory evaluation are listed in Table 1. Results listed in Table 2 showed no significant difference in scores of samples stored for less than 8 h, while significant differences (P < 0.05) in scores of samples stored for more than 8 h. There were significant differences between scores of samples stored for 12 h and 24 h, which meant sensory properties changed a lot from 12 h to 24 h. Scores of samples stored for 8 hours were about 10 points than that of samples stored for 12 h when taking each one of the six sensory properties into consideration. This showed sensory quality of soya milk drop a lot when its storage time was more than 8 h under room temperature. Thus, it is suggested that soya milk stored under room temperature should be sold out within 8 h.

Effect of storage time on viscosity, pH, and bacteria of soya milk samples
The viscosity of a food system is dependent on the volume fraction occupied by the contributing particles in combination with the inherent viscosity of the continuous phase (Anema et al. 2014). The viscosity of soya milk samples increased continually during the storage time (Fig. 1). It was in a stable state within the first 8 h, and after 8 h, it increased quickly. As pH value is an important factor that may cause the volume fraction of casein micelles (Anema and Creamer 1993), we investigated the pH values during the experimental time. The pH value decreased from 6.87 to 6.61 during the storage time, and there was an obvious drop after 8 h (Fig. 2). Thus, the soya quality of milk was basically stable in 8 h. For the concentrate from the treated soya milk, the pH decreased the consistency coefficients of the soya milk concentrate during the storage time. This is also consistent with previous studies on the apparent viscosity of soya milk concentrates (Snoeren et al. 1984;van Hooydonk et al. 1986). Thus, changes of bacteria in soya milk and bottleneck were investigated. Figure 3 shows obviously that the total bacterial count both in soya milk and bottleneck, and Lactobacillus count in soya milk increased with storage time. The increasing rates were fast within 8 h and became moderate after 12 h. Total bacterial count in bottleneck reached 4.7 log CFU/mL, which is the limit detection for milk according to Guangdong enter-exit inspection and quarantine compiled (2002) (Fig. 3). Lactobacillus was dominant bacteria in soya milk, which were more than 50% of the total bacteria count (Fig. 3). That means, the storage time should be less than 8 hours when soya milk is stored under 25 ± 2°C.

SDS-PAGE analysis of protein samples
The SDS-PAGE profiles of protein samples stored for 0 h, 4 h, 8 h, 12 h, and 24 h are shown in Figure 6. BIO-BEST 200 E gel imaging system was used to analyze the protein in the treated soya milk samples (Fig. 6). The corresponding subunit molecular weight was shown in details in Table 4. Migration rate and the molecular weight of the protein showed a negative correlation between logarithmic in SDS-PAGE gel electrophoresis system. Proteins of greater molecular weight showed smaller mobility. Comparison of these five sample bands, there was a subunit composition gap between samples. Soya milk protein contained 7-9 subunits and subunit molecular weight in the range of 13.08--106.8 kDa, while <3% in total protein content distributed among 90-107 kDa. Most of subunit molecular weight distributed from 20 to 90 kDa (Table 4). This shows that the treated soya milk protein samples mainly contain medium and small molecular weight subunit. It also shows the subunit in soya milk protein changed constantly during storage. And along with the extension of storage time, the amount of Lactobacillus in soya milk increased, the pH values decreased, leading to protein denaturation, and produce high-molecular weight subunit (Jarpa-Parra et al. 2014).

Conclusions
Collectively, we used quantitative descriptive analysis to describe the change of sensory quality, viscosity, pH, protein composition, and bacteria in both soya milk and bottleneck, in commercial soya milk stored for 0, 4, 8, 12, and 24 h at 25 ± 2°C. The results show that with the extension of storage time, the above indicators are changed in the direction of the soya milk deterioration. The contamination may be the main reason for soya milk's deterioration. The recommendation is that the best shelf life of commercial soya milk is less than 8 h at room temperature. These results may help both factories and consumers to better understand the quality control of soya milk products.