Investigating the differences of flavor profiles between two types of soy sauce by heat-treatment

ABSTRACT The difference of flavor profiles and physicochemical properties between the two types of raw soy sauces and heated soy sauces were investigated in the present research. The results indicated that no significant changes of physicochemical properties, organic acids (except succinic acid) or CIELAB color space among the samples after heat-treatment were observed. However, the content of volatiles was increased by about 30.89% in the fortified sample (PP1) by heat-treatment. The contents of 26 and 15 constituents in PP1 and PP2 (natural fermented sample) were enhanced compared with their respective control (P1 and P2), respectively. According to the changes of the constituents with odor activity value (OAV), the contribution of heat-treatment on soy sauce flavor profiles was related closely to the raw soy sauce type. The spicy, caramel-like and fruity note of heated soy sauces were higher than that of raw soy sauces, and the increasements of spicy, caramel-like and fruity note of PP1 were higher than that of PP2. These results may contribute to optimizing the heat-treatment parameters and improve the flavor of soy sauce.


Introduction
As an essential condiment in the oriental region, soy sauce not only has a long history but also contained various nutrient and bioactive substances that are responsible for its intense taste, distinct flavor and healthy function. The consumption of soy sauce has been increasing worldwide in the last three decades. [1] The soy sauce manufacture process involved in four main steps which were koji-making, mash fermentation, squeezing and blending as well as heat-treatment. [2] The quality and distinctive flavor of soy sauce are related to the raw materials and the regulations of these main four processes. [3,4] It is necessary to inoculate spore of Aspergillus oryzae, Aspergillus sojae or the mixture of Aspergillus. spp. (koji starter) during the koji-making process, which is benefit for the increase of utilization ratio of raw material and the formation of unique flavor of soy sauce. [5,6] The intensities of some dominant constituents, such as 1-octen-3-ol, benzeneacetaldehyde, 3-methylthio propanal, were closely related to the growth and sporulation of Asp. oryzae. [7][8][9] It was confirmed that the dominant microbes and their succession during the mash fermentation by culture-dependent and culture-independent methods. [9] These results depended on the foundation for fortified technology of soy sauce as well as the production of soy sauce in large-scale (180 m 3 fermentor). Therefore, bio-fortifying/bio-augmenting technology based on pure culture has been widely concerned in the last two decades. Cui et al. [10] had investigated the effect of Tetragenococcus halophilus, Zygosaccharomyces rouxii and Candida versatilis on soy sauce moromi flavor compounds in the single or combined way. And the results reported by Harada et al. [1] suggested that lactic acid bacterium affected the production of cyclotene, furfural, furfuryl alcohol, and methional. Wah et al. [11] and Udomsil et al. [12] also reported that the addition of halotolerant microorganisms could improve the flavor of Thailand soy sauce and fish sauce. Devanthi et al. [13] had demonstrated that the inoculation sequence of T. halophilus and Z. rouxii impacted on volatile profile during moromi fermentation.
Heat-treatment or sterilization was also an important procedure during soy sauce process, which is responsible for the safety, uniformity, stability, and the formation of flavor compounds of the final soy sauce. [5,14] The influences of heat-treatment on the major odorants in natural-fermented soy sauce were reported in the previous papers. [5,15,16] In general, it was inferred that heat-treatment is an important factor that influences the formation and increases or decreases of the aromatic components, colors and flavor components, such as the formation of Amadori compounds [17] as well as 4-vinylphenols. [16] Currently it is widely accepted that the Maillard reaction, one of the dominant reactions occurring during thermal processing of food, leads to the formation of aroma, color and taste compounds. [18] The Maillard reaction is the chemical reaction between reducing sugars and amino compounds. It was first described by Louis-Camille Maillard in 1912, [19] and Hodge proposed the first comprehensive scheme for the Maillard reaction in 1953. [20] In addition, recent studies have shown that certain categories of Maillard reaction products have a nonnegligible hazardous impact on human health, including acrylamide [21] produced in starch-rich processed foods, heterocyclic amines occurred in protein-rich cooked foods [22] and advanced glycation end products in baking and dairy products. [23,24] As we know, more and more large-scale reactors had been used in soy sauce manufacture, and fortification was proved to be a necessary measure to improve the quality of soy sauce. However, there is no report about the effect of heat-treatment on the flavor profiles of raw soy sauce produced by bio-fortifying so far. The objectives of the present investigation were to clarify the influence of heat-treatment on physicochemical properties, CIELAB color space and flavor profiles of fortified and conventional soy sauce.

Soy sauce samples
Soy sauce fermentation was performed as previously described. [10] Soy sauce mash, a mixture of koji and brine solution (25% w/v) at ratio of 1:1.8, was loaded into a 250 L of stainless-steel tank (200 kg/ tank) and fermented for 180 days at ambient temperature (Qianhe Condiment Co., Ltd, Meishan City, Sichuan province China). The matured moromi was pressed to retrieve fresh soy sauce. Samples labeled as PP2 and P2 were natural fermented soy sauce with/without heat-treatment. Samples labeled as PP1 and P1 were fortified soy sauce with/without heat-treatment, which was inoculated with C. versatilis CGMCC3790, Z. rouxii CGMCC3791 and T. halophilus CGMCC 3792, the microorganisms were inoculated according to the method described by Cui et al. [10] Raw soy sauces were heated at 125°C for 15 min to obtain heated soy sauces. And each sample was done in triplicates.

Physicochemical properties analysis
Titration acidity, amino nitrogen, reducing sugar and pH were determined according to the method described by Feng et al. [25] Ethanol was measured by gas chromatography with FID according to the methods described previously. [26] Organic acids of soy sauces were analyzed according to the method described by Marina et al. [27] with some modifications. Briefly, samples were diluted and adjusted pH to 7.0, then treated with C18 SPE column (Swell scientific instruments Co., Ltd. Chengdu, China) and subsequently filtered through 0.22 um filter (Micron Separation Inc., Westborough, MA). The filtered samples were injected into the Agilent 1260 HPLC (Agilent Technologies Inc., California, USA) system equipped with an Alltech OA-1000 organic acid column (300 × 7.8 mm) maintained at 75 o C. Degassed H 2 SO 4 (4.5 mmol/L) was used as mobile phase, and the organic acids, including citric, tartaric, malic, succinic, lactic, acetic and L-pyroglutamic acid, were detected using UV detector (215 nm). A 10 uL injection volume was used for both samples and standards. Organic acids were quantified using external standard.

Volatile compounds analysis
The volatile compounds of samples were detected according to the method described by Zheng et al. [28] 2 mL of the soy sauce sample was transferred into 20 mL headspace vials and saturated with NaCl. Before analysis 10 μl of internal standard (a mixture of 0.44 mg/mL 2-octanol and 0.91 mg/mL methyl octanoate solution) was added and mixed. A DVB/CAR/PDMS fiber (Supelco, Inc., Bellefonte, PA, USA) was used for the headspace solid-phase microextraction (SPME) of volatile compounds. Before extraction, vials containing samples were pre-equilibrated for 15 min at 60ºC. 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 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 absorbed on the fiber were transferred into the GC system with a splitless mode with a purge-off time of 90 s, and injector temperature was set at 250ºC. The initial temperature of the GC oven was kept at 40ºC for 5 min, then raised to 220ºC at a rate of 5ºC/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 230ºC and 250ºC, 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 index (RI) of each compound was calculated by using C 8 -C 20 n-alkanes mixture (Sigma-Aldrich) which was analyzed under the same chromatography conditions. [29] Relative amounts (μg/L) of certain volatiles were calculated by the peak area ratio to the internal standard on GC total ion chromatograms.

CIELAB color space analysis
Soy sauce samples were centrifuged (4°C, 10000 rpm) for 10 min prior to the spectrophotometric analysis. The difference in color space among sauce samples was determined by using spectrophotometer Color i5 (X-Rite, USA). The color parameters of soy sauce were 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 [30] :

Statistical analysis
All the analysis was conducted in triplicates. Data are reported as means ± 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) at the 95% confidence level by Duncan' s test.

Results and discussion
Effect of heat-treatment on the physicochemical properties As shown in Table 1, no significant difference of physicochemical properties among these samples was observed. The reducing sugars content in P2 was higher than that in P1 (p < .05), which caused by the inoculation of microorganisms. Ethanol in PP1 was significantly decreased compared with P1 (Table 1, p < .05). According to the previous researches, the thermal effect on reducing sugars and the titration acidity depended on the temperature and heat treatment times. [5] The higher the temperature and the longer the time, the more reducing sugars would be decreased. In the present research there were no significant differences in reducing sugars between raw soy sauces and heated soy sauces (Table 1, p < .05). The increase of titration acidity may be caused by the concentration effect during the heating process.
The detailed changes of organic acids in soy sauce after heat-treatment were reviewed by HPLC analysis. Compared to P2, the contents of respective organic acids in P1 were higher except that of lactic acid and malic acid. The content of succinic acid was significantly decreased by heat-treatment, while no significant differences of other organic acids between raw soy sauces and heated soy sauces (p < .05, Table 2). The results obtained above indicated that there was a slight influence of heattreatment on physicochemical properties and the composition of organic acids.

Effect of heat-treatment on CIELAB color space
The a* and b* values were calculated as the difference between color parameters of heated soy sauce and raw soy sauce. The C * value indicated the difference between the metric chroma of heated soy sauce and raw soy sauce, and L* indicated the difference between the brightness of heated soy sauce and raw soy sauce. As shown in Figure 1, no difference of color space indexes between P1 and P2 was observed. There were slight changes in color between heated soy sauces (PP1 and PP2), b* and C* values were slightly increased in the two types of raw soy sauces by heat-treatment, while the changes of L* and a* values were different, which were depended on the raw soy sauce type. Moreover, soy sauce in PP2 was redder, but the brightness was weaker in PP2 than that in PP1 ( Figure 1). The melanoidins formed during the heat-treatment process was the major contributor to the color of soy sauce. [31] Besides, the change of color was also related to phenolic composition. [32] The present results indicated that the changes of color depended on the metabolic composition of raw soy sauce so that it was closely related to the fermentation pattern.

Effect of heat-treatment on volatile compounds
Total of 66 volatile constituents was identified and quantitated in these samples and summarized in Table 3. These constituents were divided into 8 different groups according to the chemical structure, including 24 esters, 10 alcohols, 9 acids, 7 phenols, 5 pyrazines, 4 aldehydes, 4 ketones, and 3 others. The volatiles content of both P1 and P2 was almost the same, but the difference of respective group content was significant when heated. The phenols and esters content in P1 was higher than that of P2, while the other six groups of volatile compounds were lower in P1. The contents of volatiles in PP1 were increased by 30.89%, from 1461.62 ug/L to 1912.18 ug/L, while that of PP2 was almost unchanged (Figure 2a). The effect of heat-treatment on volatile profiles was shown in Figure 2. Volatile content of seven groups in PP1, except esters, which was decreased by 26.86%, were increased, in which the contents of acids, pyrazines, phenols, aldehydes, ketones, and others were significantly enhanced. While that of five groups volatiles in PP2 was enhanced, in which acids, pyrazines, and phenols were significantly increased. Further, the increment of acids, pyrazines, ketones, and others of the former was significantly higher than that of the latter. The contents of 3-methyl-1-butanol, phenyl ethyl alcohol and b-ethyl-benzene ethanol in P1 were higher than that of P2, which should be a contribution by the synergistic effect of inoculated yeast and lactic acid bacteria. [10] However, the contents of 2, 3-butanediol, 2-furanmethanol, and maltol were lower than that of P2 (Table 3). Among these constituents, maltol was a key aroma compound in Japanese soy sauce. [14] While 2-furanmethol was determined in Chinese soy sauce. [33] The content of 3-methylbutanol and 1-octen-3-ol in PP1 were decreased by 13.13% and 60.32% via heat-treatment, respectively. Besides, the contents of 2-furanmethanol and b-ethyl-benzene ethanol were increased by 112.44% and 17.39% in PP1, but that of PP2 was increased by 11.27% and 14.29% compared with respective control (Table 3). It is worth noting that the content of 3-methylthiopropanol was enhanced in PP1, but that of PP2 was decreased, which was one of the dominant flavor components in Japanese type soy sauce, which the threshold value was only 1.5 ug/L, endowed to mushroom-like note. [34]   Among nine acids identified, hexadecanoic acid in P1 was higher than that of P2, and other seven constituents, except nonanoic acid, were opposite. The contents of 2-methyl propanoic acid, 4-hydroxy-butanoic acid, and hexanoic acid were significantly enhanced in P1 and P2 when heated. The thresholds of these components were too high so that their contribution to the soy sauce flavor was weak.
The pyrazines content of P1 was significantly lower than that of P2, but the increment of PP1 was higher than PP2, which were enhanced by 166.04% and 94.85%, respectively ( Figure 2). Among these constituents, the contents of 2-ethenyl-6-methyl-pyrazine and 2-ethyl-5-methyl-pyrazine were increased by 376.82% and 102.43% in the former, while only 2-ethenyl-6-methyl-pyrazine was increased by 243.51% in the latter (Table 3). These components, formed via Maillard reaction during the heat-treatment process, endowed strong fragrant flavor to soy sauce. [35] A significant difference between benzaldehyde and benzene acetaldehyde content among P1 and P2 was observed. The contents of these aldehydes were significantly increased via heat-treatment, and their content of PP1 was higher than that of PP2. Higher content of benzaldehyde and benzeneacetaldehyde would enhance the intensity of unique flavor for soy sauce. [36] The content of Mesifurane in P1 was lower, but that of gamma-Nonanolactone and megastigmatrienone was higher than P2 ( Table 3). The contents of these four constituents in the former were increased, but that of gamma-Nonanolactone and megastigmatrienone was only increased in the latter. The  contents of gamma-nonanolactone and megastigmatrienone were increased by 67.57% and 133.52% in PP1, which were enhanced by 28.96% and 109.68% in PP2, respectively (Table 3). In P1 sample, the content of 4-ethylguaiacol and 2, 4-di-tert-butylphenol were higher than that of P2, while that of guaiacol and 4-ethylphenol were opposite (Table 3). The content of guaiacol, phenol, eugenol, and 4-vinlyguaiacol of P1 were significantly enhanced, while that of guaiacol, eugenol, and 4-vinlyguaiacol of P2 were enhanced by heat-treatment. The threshold values of these constituents, except 4-ethylphenol (1010.1 ug/L), were very low. Meanwhile, guaiacol, 4-ethylguaiacol, and 4-vinlyguaiacol were also major aromatic flavor constituents in soy sauce. [37] Therefore, it was beneficial for improving the quality of soy sauce that enhancing these constituents in soy sauce.
After heat-treatment, the content of 1, 2-benzisothiazole was decreased, while that of 2-acetylpyrrole and 2, 3-dihydrobenzofuran were decreased in P1 and P2 (Table 3). Among twenty-four esters identified, nine constituents in P1 were lower than that of P2. The contents of 8 constituents in PP1 were enhanced compared with P1, while that of 5 constituents in PP2 were also increased compared with P2, in which ethyl dodecanoate and ethyl pentadecanoate were significantly enhanced compared to P1 and P2. Besides, the contents of 4 constituents in PP1 were remarkably increased, which were Ethyl 15methylheptadecanoate, Methyl tetradecanoate, Ethyl 3-pyridinecarboxylate, and ethyl lactate, but were unchanged or decreased in PP2. However, the content of ethyl 9-Hexadecenoate, ethyl hexanoate, ethyl linoleate, ethyl hexadecanoate and ethyl heptadecanoate in PP2 were increased (Table 3).
Evaluating the difference of flavor profiles between PP1 and PP2 based on Odor Activity Value (OAV) analysis The contribution of each aroma constituents to soy sauce flavor depended on its odor activity value (OAV), which has been used to evaluate the characteristic flavor of various food in the last two decades. [10,28,38,39] The constituent with OVA>1 contributes to the overall aroma significantly. [40] In order to know the effect of heat-treatment on the flavor profiles of soy sauce, the OAVs of volatiles were calculated. Among 66 volatiles identified, the OAVs of 11 constituents were higher than one (Table 4 and Figure 3). OAVs of 1-octen-3-ol, 2, 3-butanediol, ethyl oleate, and ethyl linoleate in PP1 were decreased compared with P1. OAVs of 1-octen-3-ol and 2, 3-butanediol in PP2 were decreased compared with P2. OAV of gamma-nonalactone, megastigmatrienone, guaiacol, and 4-vinlyguaiacol were increased in both types of soy sauce via heat-treatment. Although limited research reported that megastigmatrienone was detected in soy sauce, it was widely focused on an aromatic flavor constituent of tobacco with its threshold and characteristic fragrant was related to its structure. [41] The content of 4-vinlyguaiacol was increased, but 4-ethylguaiacol was slightly decreased by heat treatment, which was consistent with the result reported previously. [14] As shown in Figure 3, the difference of characteristic flavor between PP1 and PP2 was observed. The spicy and burnt note endowed by 4-vinlyguaiacol for PP1 was stronger than that of PP2, but the fruity and caramel-like note impacted by mesifurane [42] for PP2 was stronger than that of PP1. It resulted in reducing the intensity of caramel-like as the result of 4-ethylguaiacol and phenols were decreased by heat-treatment.

Conclusion
The different effect of heat-treatment on the physicochemical properties, organic acids, and CIELAB color space of two types of soy sauce were investigated in the present study. Ethanol and succinic acid were significantly decreased by heat-treatment, while there were no significant changes of total acid, amino acid nitrogen, reducing sugars and pH in heated soy sauces compared with their respective control (p < .05). The color of soy sauces was slightly influenced by heat-treatment. However, significant divergences of 8 groups of volatile compounds were observed in PP1 and PP2. The total content of volatiles in PP1 was increased by 30.89%, compared with P1 except esters, and that of PP2 was almost unchanged, although that of acids, pyrazines, and phenols was increased compared with P2. The contents of 26 and 15 constituents were enhanced in PP1 and PP2 compared with their respective control. The OVA analysis indicated that the contribution of heat-treatment on the flavor profiles related closely to the raw soy sauce type. The spicy, caramel-like and fruity note of heated soy sauces were higher than that of raw soy sauces, moreover, the increasement of the above aroma were higher in fortified soy sauce.