Introduction

Soy sauce, as an indispensable condiment in the oriental region (China, Japan, Korea and Southeast Asia), is also becoming popular all over the world (Harada et al. 2016). The history of soy sauce manufacture could be traced back to more than 3000 years ago (Zhou Dynasty) in China (Cui et al. 2014). The production quality and distinctive flavor of soy sauce were related not only to four main processes, including koji-making (Gao et al. 2009; Ito et al. 2013), moromi fermentation (Harada et al. 2016, 2018; Qi et al. 2014; Song et al. 2015; Tanaka et al. 2012), heat-treatment/sterilization (Gao et al. 2009; Shu et al. 2013) and storage/packaging, but also to raw material (Kinoshita et al. 1998; Liang et al. 2019). The volatiles formed in the koji-making stage was responsible for the unique flavor of soy sauce and the role of microbes is crucial in fermentation stages (Gao et al. 2009; Ito et al. 2013). A lot of research on the relationship between major microbes and metabolic components have been carried out in the last several decades (Harada et al. 2016, 2018; Qi et al. 2014; Song et al. 2015; Tanaka et al. 2012). From the above literature, it was confirmed that Tetragenococcus halophilus, Zygosaccharomyces rouxii, Candida etchellsii and Candida versatilis were the dominant microbial strains for various soy sauce fermentation technology, although the microbial community and dynamics presented some differences. In general, Z. rouxii is not only responsible for the production of ethanol but also for the formation of higher alcohols (such as 3-methyl-1-butanol and 2-methyl-1-propanol), which were important aroma active constituents of soy sauce (Sluis et al. 2002). Moreover, it was reported that 4-hydroxy-2(or 5)-ethyl -5(or 2)-methyl-3(2H)- furanone (HEMF) and 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) could also be synthesized by Z. rouxii (Sluis et al. 2001). While 4-ethyl guaiacol and 4-ethylphenol, endowed unique flavor to soy sauce, were the metabolites of C. etchellsii and C. versatilis (Yasuhiko et al. 2006). pH was reduced by the metabolites of T. halophilus, such as lactic acid and acetic acid at initial stage of fermentation. The declined pH was also profitable for the propagation of Z. rouxii in the moromi (Devanthi et al. 2017). The production of cyclotene, furfural, furfuryl alcohol and methional was also influenced by T. halophilus (Harada et al. 2016). It has been indicated that the contents of volatiles could be significantly enhanced by sequentially inoculating the pure culture of T. halophilus, Z. rouxii and C. versatilis during moromi fermentation (Devanthi et al. 2017; Natteewan et al. 2011; Wah et al. 2013).

Heat-treatment/sterilzation was also an important process for the unique taste and quality control of soy sauce. Maillard reaction products were remarkably increased during heat-treatment, which had an impact on the over-all taste of soy sauce (Gao et al. 2009; Shu et al. 2013). Packaging is the last process of soy sauce production. Plastic materials have been widely used for soy sauce packaging since 1976 (Shirakura et al. 2006). In comparison to other liquid foods, such as carbonated soft drinks, tea, mineral water and edible oil, soy sauce possess high permeability as well as corrosiveness of salt and acid. However, there is limited information about the changes of the constituents of soy sauce bottled in the plastic bottle. Besides, fresh soy sauce, as a new type of finished product, also have attracted attention from consumers in recent years (Shu et al. 2013). In the same manner, the effect of storage conditions (including packaging material and storage temperature) on the physicochemical properties and volatiles of soy sauce has not been reported so far.

In the present article, the influence of different storage condition, such as storage temperature and packaging materials on the major physicochemical properties and volatiles profile of the two types of soy sauces was investigated and the differences between the two type of soy sauces were revealed. And we also try to clarify the interrelation between major flavor profiles and storage factors through hierarchical cluster analysis. To the best of our knowledge, it is the first time that the effect of various storage factors on the unique flavor profiles of finished products was estimated. The study was aimed at understanding the main factors affecting the characteristic flavors of soy sauce during storage process. The purpose of this study is to optimize the storage conditions of soy sauces.

Materials and method

Samples collection

In our present research, two different types of raw soy sauce (P1 and P2) were used. Raw soy sauce P1 was sampled from the moromi inoculated with T. halophilus CGMCC3792, Z. rouxii CGMCC3791 and C. versatilis CGMCC3790 according to the process described by Cui et al. (2014). Raw soy sauce P2 was sampled from the spontaneously fermented moromi. Row soy sauce Heat-treatment was conducted at 121 °C for 15 min.

The unheated and heated soy sauce samples were bottled in glass bottles (B) and polyethylene terephthalate (PET) bottles (S) and then stored at ambient temperature (AT, 20–25 °C) and low temperature (LT, 4–10 °C, which was similar to the temperature ranges of cold chain) for 90 days, respectively. The detailed information of samples is listed in Table 1.

Table 1 Detailed information of samples used in the present research
Full size table

Physicochemical property analysis

Titration acidity (TA), amino nitrogen (AN) and reducing sugars (RS) were measured according to the method described by Feng et al. (2012). pH was measured directly using a Leici model PHS-25 pH meter (Leici Instrument, Shanghai, China).

Ethanol was detected and quantified by gas chromatography with FID according to the methods described by Miyagi et al. (2013).

Volatile compounds analysis

The volatile compounds of soy sauces were detected according to the method described by Zheng et al. (2013). Soy sauce was transferred into 20-ml headspace vials and saturated with NaCl (2.00 ml for each parallel test). Prior to analysis 10 μl of internal standard solution (a mixture of 0.91 mg/ml 2-octanol and 0.4 mg/ml methyl octanoate solution) were added to the tested samples. A DVB/CAR/PDMS fiber (Supelco, Inc., Bellefonte, PA, USA) was used for the extraction of volatile compounds. Before extraction, vials containing samples were pre-equilibrated for 15 min at 60 °C. Subsequently, the SPME fiber was inserted and maintained for another 45 min to adsorption. Then the loaded fiber was inserted into the injector of GC for 3 min to desorption.

The volatile compounds were analyzed by using GC–MS (Thermo Electron Corporation, Waltham, USA) equipped with a DB-INNOWAX capillary column (30.0 m × 0.25 mm, 0.25 μm, Agilent, Santa Clara, USA). Volatile compounds adsorbed on the fiber were transferred into the GC system with a splitless model with a purge-off time of 90 s and injector temperature was set at 250 °C. The initial temperature of GC oven was kept at 40 °C for 5 min and then raised to 220 °C at a rate of 5 °C/min (held for 10 min). Helium was used as carrier gas at a constant flow of 1.0 ml/min. Mass spectrometer conditions were as follows: Electron impact was 70 eV. Ion source and transfer line temperatures were 230 °C and 250 °C, respectively. Mass range was 40 to 400 amu. Each volatile compound was identified by comparing their mass spectrum with those in the NIST05 library database (Finnigan Co. USA), Compounds were reported on the basis of their similarity (> 800). At the same time, Kovát retention indices (RI) of each compound were calculated by using C8–C20n-alkanes mixture (Sigma-Aldrich) which was analyzed under the same chromatography condition (Dool and Kratz 1963). Relative contents (ug/l) of certain volatiles were calculated by the peak area ratio to the internal standard on GC total ion chromatograms.

CIELAB color space determination

The difference of color space among sauce samples was determined by using spectrophotometer Color i5 (X-Rite, USA). The color parameter of soy sauce was expressed as L* (lightness from 0 = black, to 100 white), a* and b* indicate chromaticity corresponding to green (negative value) to red (positive value) for a* and blue (negative value) to yellow (positive value) for b*. The metric chroma (C*) is defined as follows CIE (2004).

$${\text{C}}^{*} = \sqrt {a*^{2} + b*^{2} }$$

Statistics analysis

All the analysis was conducted in triplicate. Data are reported as mean ± standard deviation (SD). Analysis of variance (ANOVA) and significant differences among means were tested by one-way ANOVA using SPSS Software (SPSS 19.0,SPSS Inc., Chicago, IL, USA). P < 0.05 (Duncan’s test) was considered as statistically significant in all test. Odor activity value (OAV) was calculated by dividing the concentration of each compound by its odor threshold concentration and was often used to evaluate the flavor profiles characteristics of food (Cui et al. 2014; Petra and Peter 2007; Reboredo-Rodríguez et al. 2013).

Results and discussion

Effect of packaging materials and storage temperatures on physicochemical properties of soy sauce

The discrepancy of physicochemical properties among 20 different samples was investigated as shown in Table 2. The content of ethanol was decreased significantly after storage for 90 days in all tested samples (P < 0.05). while the changes of other physicochemical properties varied with types of soy sauce and storage conditions. Compared with P1, TA contents were decreased in CSP1, CBP1, LSP1 and LBP1 after storage for 90 days, while there were no effect of storage on pH, AN and RS (P < 0.05). An increment of the contents of pH and AN and a decrease of RS in CSP2, CBP2, LSP2 and LBP2 were observed with respect to P2 (Table 2, P < 0.05).

Table 2 Effect of packing material types and store temperature on physicochemical properties of soy sauce
Full size table

For the heated samples, ambient temperature (AT) storage had slightly influence on TA, pH, AN and RS compared with their respective control (PP1 and PP2, Table 2, P < 0.05). while the changes of TA, pH, AN and RS varied with types of soy sauce at LT. for example, pH was increased in LBPP1 and LSPP1 compared with PP1, while the pH values of LBPP2 and LSPP2 were decreased compared with PP2 (Table 2, P < 0.05). The results indicated that the changes of physicochemical properties depended on the types of raw soy sauce, heat-treatment and storage conditions.

Effect of packaging materials and storage temperatures on CIELAB color space of soy sauce

The a* and b* values were calculated to differentiate the color parameters among soy sauce samples. The C* value indicated the difference of metric chroma, while the L* value indicated the difference of the brightness. As shown in Fig. 1, the nuance of color parameters, metric chroma and brightness varied with samples. The intensities of reddish yellow were enhanced for LBPP1 and LSPP1 at LT, but that of lightness was slightly increased in CBPP1 and CSPP1 under ambient temperature (AT) compared with control (PP1). The color of CSP1 and LSP1 became to yellowish green after storage at AT with respect to P1.

Fig. 1
figure 1

Difference of CIELAB color space among various samples

Full size image

Furthermore, the b* value and L* value of CBPP2 sample were increased compared with PP2 after stored at AT. The color of soy sauce is regarded as one of the the important sensory characteristics, which is believed to be mainly attributable to melanoidins formed during the heat-treatment process (Miki et al. 2011). Besides, the change of color was affected by the changes of phenolic composition (Recamales et al. 2006). The present result indicated that the changes of soy sauce color depended on the metabolized composition, packaging material and storage temperature.

Effect of packaging materials and storage temperatures on volatiles profile of soy sauce

Among the 66 volatile compounds identified and quantified, as listed in Table S1. The contents of total volatiles in CBP1 and CBPP1 were increased after stored at AT, while, that a decrease was observed in LBP1 and LBPP1 after stored at LT compared with their respective control (P1 and PP1). The volatiles in the rest samples were decreased after storage, ranged from 11.57 to 19.89% with respect to control, except that of LSP1 and CSP1. The results indicated that storage temperature was the key factor affecting the volatiles of two types of soy sauce. Compared with the glass bottle, it was observe that PET bottle showed a weak sorption of flavor compounds in the present research, although it was reported the sorption of flavor compounds was one of the most important limitations of plastic bottles (Sajilata et al. 2010). As shown in Fig. 2, the changes of the volatile profiles among samples depended on the raw soy sauce types, packaging materials and storage temperatures. The abundances of alcohols were increased after storage with respect to their respective control (P1 and PP1, Fig. 2). Besides, the abundances of acids in LSP1, LBP1, LSPP1 and LBPP1 were also increased (Fig. 2). It may be related with hydrolysis of esters since five esters were degraded after storage. However, the abundance of phenols was significantly decreased, which could be caused by oxidation reaction. In general, oxygen may be permeated into soy sauce easily through PET material, which could result in a stronger oxidation reaction of phenols in PET bottle than that of glass bottle (Ghidossi et al. 2012). However, no significant difference of induced proportion of phenols was observed among samples bottled in different packaging materials (Fig. 2). The result suggested that the effect of oxygen carried into soy sauce during bottling on phenols was stronger than that of permeated oxygen during storage. As shown in Fig. 2, the contents of pyrazines were higher in sterilized sample PP1 and PP2. Although, the decrease of pyrazines was observed in all stored samples, the reduction among samples were different. As we know, pyrazines could be synthesized chemically, especially via thermal reaction or biosynthesis. While there was little information about the abiotic degradation of pyrazines (Müller and Rappert 2010). Therefore, it may be speculated that the microorganisms in fresh soy sauce were responsible for the decrease of pyrazines, but the detailed reason need to be explored further.

Fig. 2
figure 2

Effect of packaging materials and storage temperatures on the volatile profiles

Full size image

Four components appeared after storage in all tested samples. These components included 1-hydroxy-2-propanone, butyl hexanoate, ethyl heptanoate and heptanoic acid. Twelve components disappeared after storage for 90 days. These components included acids (2-methyl-propanic acid), esters (methyl nonanoate, ethyl nonanoate, ethyl dodecanoate, 1-methylethyl tetradecanoate, ethyl tetradecanoate, ethyl heptadecanoate and ethyl-15-methyl heptadecanoate), pyrazines (2-methylpyrazine and 2,6-dimethylpyrazine), others (1,2-benzisothiazole) and aldehydes (2,3-dihydro-1H-indene-4-carboxaldehyde). Besides, ethyl-9-hexadecenoate vanished in the CBP2, LBP2, LBPP1 and LBPP2 samples. 2-Ethyl-5-methyl-pyrazine also disappeared in the unsterilized samples stored at AT. Benzyl alcohol was detected in stored PP2 samples (CBPP2, CSPP2, LBPP2 and LSPP2). The findings above may be involved in the flavor scalping of soy sauce during the storage process, but it was difficult to understand the mechanism because it was so complex that we could not reveal the interaction relationship among them.

Evaluation for the differences of flavor profiles among samples

In order to evaluate the storage conditions on the flavor profiles of soy sauces, 9 constituents were selected, which made the prominent contribution to the overall aroma of soy sauce since their odor activity values (OAVs) were beyond 1 in the present experiments.

The difference of the OAVs of constituents among samples was observed (Table S2). OAV of 4-vinlyguaiacol and ethyl linoleate decreased significantly, but their decrement was different among the inoculated and sterilized samples. However, OAV of 4-ethylguaiacol, eugenol and 4-vinlyguaiacol reduced distinctly, in comparison to ethyl octanoate which increased in the heated samples without inoculation. Both 4-ethylguaiacol and 4-vinlyguaiacol were the characteristic flavor constituents of soy sauce, which endowed spicy and burnt as well as smoky note (Feng et al. 2015), eugenol, as volatile phenols which gave clove note (Imamura 2016). OAV of ethyl hexanoate in the samples based on P2 also increased after storage. Ethyl hexanoate was one of the key aroma constituent in Baijiu, endowed the ferment products with fruity-note (Fan et al. 2011). As shown in Fig. 3, a significant difference of mesifurane, with the threshold of 0.3 ug/l, among samples was observed, which was one of the key aroma constituents (Petra and Peter 2007) endowed the fruity and caramel aroma (Prat et al. 2014). Besides, 1-octen-3-ol was also an aroma-impact constituent with mushroom-like note (Petra and Peter 2007). Hence, the effect of the packaging material, the types of raw soy sauce and storage temperature resulted in changing the intensities of fruity, caramel aroma, mushroom-like note as well as a smoky note of soy sauce (Fig. 3).

Fig. 3
figure 3

Comparison of the difference of aroma features of soy sauce samples by OAVs. a Represents the differences among samples including P1, CSP1, LSP1 and LBP1; b Represents the differences among samples including P2, CSP2, LSP2 and LBP2; c Represents the differences among samples including PP1, CSPP1, LSPP1 and LBPP1; d Represents the differences among samples including PP2, CSPP2, LSPP2 and LBPP2

Full size image

The differences of flavor feature among samples based on the results revealed by principle component analysis (PCA)

In order to elucidate the differences among soy sauce samples after storage, PCA was conducted. As shown in Fig. 4, a total of two significant principal components accounting for 70.23% (Fig. 4a) and 64.61% (Fig. 4b) of the total variance of unsterilized or sterilized soy sauces were extracted, respectively. In Fig. 4a, PC1 explained 41.80% of the total variance and a general separation between P2 and P1 group was separated. In Fig. 4b, PP1 and PP2 group was also separated successfully.

Fig. 4
figure 4

Principal components analysis of characteristic flavor compounds of soy sauce from different fermentations patterns. a P1 and P2 groups; b PP1 and PP2 groups

Full size image

In contrast to P1 and PP1, 1-octen-3-ol, ethyl hexanoate and ethyl octanoate were dominant volatile compounds in the inoculated samples (CSP1, CBP1, CSPP1 and CBPP1) after stored at AT, while the samples stored at LT were characterized by 3-methyl-1-butanol, phenylethyl alcohol, γ-nonalactone, megastigmatrienone, 4-ethylguaiacol, eugenol, 4-vinylguaiacol, ethyl oleate and ethyl linoleate (Fig. 4a, b). Compared with P2, these samples stored at LT were characterized by 2, 3-butanediol, mesifurane and guaiacol. But CBPP2, CSPP2, LBPP2 and LSPP2 were only characterized by 2, 3-butanediol with respect to PP2 (Fig. 4b). The results of PCA suggested that the 9 volatile compounds selected can be used as markers to distinguish the soy sauce types and storage conditions.

Conclusion

In the present research, the effect of storage conditions (packaging material and storage temperature) on two types of soy sauce was investigated. The results shown that ethanol content decreased significantly in all tested samples after storage (P < 0.05). while the changes of physicochemical properties and CIELAB color space varied with soy sauce types, packaging material and storage temperature. The changes of volatiles and volatile profiles after storage indicated that storage temperature was a key factor resulting in flavor scalping. It also suggested that there was no significant difference of flavor compounds sorption between glass and PET bottle. The abundances of acids and alcohols increased after stored at AT and LT, but phenols decreased. The contents of pyrazines were higher in sterilized sample PP1 and PP2. Although, the decrease of pyrazines was observed in all stored samples, the reduction among samples were different, which suggested that the decrease of pyrazines might be related to the effect of microbes. According to odor activity value analysis, the effect of the packaging material, raw soy sauce types and storage temperature resulted in the changes of intensities of fruity, caramel aroma, mushroom-like note as well as a smoky note. It is the first time that a decrease of 4-vinylguaiacol during the storage was reported. For the inoculated soy sauces, 1-octen-3-ol, ethyl hexanoate and ethyl octanoate in the samples when stored at AT were dominant, while the samples stored at LT were characterized by multiple components according to the results of PCA. These results indicated the benefits for understanding the main factors affecting the flavor profiles and quality of soy sauce during storage, as well as optimizing the storage condition.