Contribution of 2-methyl-3-furanthiol to the cooked meat-like aroma of fermented soy sauce

Abstract

The cooked meat-like aroma compound, 2-methyl-3-furanthiol (2M3F), was detected in fermented soy sauce (FSS) by GC-olfactometry and GC-MS. 2M3F was present in FSS at a concentration considerably greater than the perception threshold, and the 2M3F concentration increased with heating temperature. Sensory analysis indicated that with the addition of only 0.2 μg/L of 2M3F to the soy sauce sample, the cooked meat-like aroma is significantly stronger than that of sample without the addition of 2M3F. Hence, 2M3F contributes to the cooked meat-like aroma of FSS, which constitutes the key aroma component of FSS. In addition, 2M3F was generated from the addition of ribose and cysteine in FSS by heating at 120 °C, but it was not detected in a phosphate buffer under the same condition. Furthermore, 2M3F was not detected in acid-hydrolyzed vegetable-protein-mixed soy sauce (ASS) and heated ASS. These results indicated that fermentation by micro-organisms facilitates the generation of 2M3F in FSS.

Graphical abstract

2-methyl-3-furanthiol (2M3F)contribute to the cooked meat-like aroma of fermented soy sauce, and this compound is a very important aroma component of fermented soy sauce.

Key words:

• fermented soy sauce
• 2-methyl-3-furanthiol
• aroma compound
• heating

Soy sauce, a traditional fermented seasoning used throughout East Asia, is widely used on a daily basis in Japan. The aroma of soy sauce is one of the most important factors that determine its quality as well as marketability to consumers. More than 300 flavor components have been identified in Japanese soy sauce.^1,2) Previous studies have demonstrated that the caramel-like aroma compound 4-hydroxy-2 (or 5)-ethyl-5 (or 2)-methyl-3(2H)-furanone (HEMF) is the key aroma component of FSS.^3,4) Previously, we have reported the mechanism for formation of HEMF by yeast during the fermentation of moromi;^5–7) we believe that another important compound of soy sauce in the unidentified odorant is formed by the Maillard reaction and/or yeast action.

Soy sauce is used as a condiment as well as in cooking. Particularly, soy sauce is typically used as seasoning for meat dishes, attributed to the aroma characteristics of cooked soy sauce. Heating soy sauce leads to a clear change in the aroma: heat-treated soy sauce exhibits a caramel-like, cooked meat-like, roasted and burnt aroma. The caramel-like compound, HEMF,^3,4) and the roasted compound, 2-furanmethanethiol (2FM), have already been suggested as key aroma constituents in heated soy sauce.⁸⁾ Kaneko et al. have reported that phenolic compounds exhibiting burnt aroma increase during the heating of soy sauce,⁹⁾ while the cooked meat-like aroma compound has been analyzed or reported to a less extent.

2-Methyl-3-furanthiol (2M3F) is an isomer of 2FM, which exhibits a cooked meat-like aroma at an extremely low odor threshold level of 4 ng/L in a model dilute alcohol solution.^10,11) Typically, the formation of 2M3F and 2FM has been simultaneously investigated, because these aroma compounds are formed from the same precursor compounds (hexoses or pentoses and sulfur amino acid) by heating.¹²⁾ 2FM has been reported to be a key aroma component of heat-treated soy sauce.⁸⁾ High concentrations of sulfur amino acids and carbohydrates are formed during the fermentation of soy sauce. Hence, 2M3F is anticipated to be the key aroma component of soy sauce, which contributes to the cooked-meat like aroma. As 2FM is generated by yeast during the fermentation of soy sauce,¹³⁾ 2M3F is possibly generated during the same production process. For proving our hypothesis, FSS samples were compared with acid-hydrolyzed vegetable-protein-mixed soy sauce (ASS) samples in this study.

This study describes (1) the identification of 2M3F in soy sauce and its contribution to the aroma of soy sauce, (2) the change of the 2M3F concentration in soy sauce by heating and (3) the possible pathways leading to the formation of 2M3F in soy sauce.

Materials and methods

Soy sauce samples

Five heat-treated and five raw soy sauce samples were purchased from a local market in Japan. For heating experiments, a raw soy sauce sample (A), raw sample (B) and heat-treated soy sauce sample (BH) were prepared by Kikkoman Corporation for this study. Prepared A and B had different lot numbers. BH was prepared by heating B for pasteurization by Kikkoman Corporation. Three ASS samples were purchased from local markets in the USA and Brazil.

Chemicals

Cysteine, 2M3F, and dichloromethane were purchased from Sigma-Aldrich (Tokyo, Japan). Glucose, ethyl acetate, and potassium dihydrogen phosphate were purchased from Kanto Chemicals (Tokyo, Japan). 4-Methoxy-2-methyl-2-mercaptobutane (an internal standard) was purchased from Oxford Chemicals (Bound Brook, NJ, USA). Ribose and sodium chloride were purchased from Nacalai Tesque (Kyoto, Japan).

Heat-treatment of the raw soy sauce samples and a phosphate buffer model experiment

In the first series of experiments, 200 mL of A was loaded into a glass cylinder and heated for 5 or 20 min at 80 °C or 95 °C in a dry stirring bath and at 120 °C in a laboratory autoclave. Next, the samples were cooled in water to 25 °C after attaining the final temperature. In the second series of experiments, 7.7 mM of ribose or 47 mM of glucose and 1.8 mM of cysteine were added to B and BH, respectively, followed by immediately heating at 120 °C for 5 min. In addition, 7.7 mM of ribose or 47 mM of glucose and 1.8 m M of cysteine were dissolved in a phosphate buffer (0.5 mol/L, 180 mL) at pH 5.0 and thermally treated for 5 min at 120 °C in a laboratory autoclave. The concentrations of ribose, glucose, and cysteine added in the second series of experiments represented the final concentrations in the sample solutions.

Analysis of volatile thiols in the soy sauce samples

Volatile thiols in each sample were extracted and assayed by the method described in our previous study.⁸⁾ First, sodium sulfite (1 g) and 0.5 mL of ethyl acetate were added to 200 mL of a 40% soy sauce solution, and 10.72 μg/L of 4-methoxy-2-methyl-2-mercaptobutane was added as the internal standard. Second, the sample solution was stirred two times with 5 mL of dichloromethane at 1000 rpm for 5 min. The resulting organic phase was dried with anhydrous sodium sulfate and then concentrated to 25 μL under a nitrogen flow in a glass tube. Each of the obtained samples was then subjected to GC-O and GC-MS analyses. GC-O analysis was conducted on a Shimadzu GC-17A instrument. For the DB-XLB capillary column (J&W; 60 m × 0.25 mm (i.d.), 0.25 μm film thickness), the oven temperature was maintained constant at 40 °C for the first 10 min, increased to 220 °C at a heating rate of 3 °C/min, and then maintained constant at 220 °C for 1 min. The flow rate of the He carrier gas was 1.5 mL/min, linear speed was 26.5 cm/s, and the injection and flame ionization detection (FID) temperatures were 210 and 230 °C, respectively. The effluent from the GC column was split 1:1 between the FID and sniffing port. 2M3F was detected on a Shimadzu QP-2010 GC-MS instrument, equipped with a DB-XLB capillary column (J&W; 60 m × 0.25 mm (i.d.), 0.25 μm film thickness). The oven temperature was programmed in a manner similar to that conducted for GC-O. The mass spectrometer was operated in the electron impact (EI) mode with an ionization voltage of 70 eV, and the ion source temperature was set to 200 °C. Injection for GC-MS analysis was conducted in the splitless mode (splitless time of 1 min). 2M3F was detected by GC-O and GC-MS analyses according to the spectrum obtained in the SIM mode, and the retention time determined relative to the aroma concentrate was compared with the reference compound. 2M3F was detected in the SIM mode, with selection for ions of m/z = 45, 85, 86, 113, and 114. The ions of m/z = 114 for 2M3F and 2FM, m/z = 134 for ethyl 2-mercaptopropionate (ET2MP), m/z = 124 for benzenemethanethiol (BM), and m/z = 134 for the internal standard were utilized for quantification. Four analyses were conducted using the heated raw soy sauce samples and phosphate buffer model samples, and the significance of the 2M3F concentration in these samples was assessed by the Student’s t test.

Sensory analysis

Olfactory discrimination tests on two soy sauce samples were conducted by using the paired comparison test, one with heat-treated soy sauce containing 0.5 μg/L of 2M3F, and the other with the same soy sauce sample added with different concentrations of 2M3F (0.2, 0.4, or 0.8 μg/L). In this study, 34 female panelists (27–55 years) participated in the paired comparison test; these panelists were selected as per the guidelines for the selection, training, and monitoring of selected panelist stipulated by ISO 8586 and their sensitivity for soy sauce.¹⁴⁾ The panelists were asked to smell both soy sauce samples and to choose the one that exhibited stronger cooked meat-like aroma. Fifteen milliliters of a sample was poured into a black plastic cup (60 mL) encoded with a 3-digit number. The strength of cooked meat-like aroma in samples was assessed by the χ² test.

Results and discussion

Identification and quantification of 2M3F in the fermented soy sauce (FSS) samples

Volatile thiols in the heat-treated and raw soy sauce samples were extracted and analyzed by GC-O and GC-MS. GC-O analysis with a DB-XLB column indicated cooked meat-like aroma at 20.6 min. According to the mass spectrum obtained from the same sample, the odor quality and retention time of the reference compound for 2M3F, which represents the 2M3F volatile thiol, was identified in the FSS sample. Kaneko et al. have reported the presence of 2M3F in raw soy sauce by GC-O, but 2M3F is not detected in the mass spectrum by GC-MS. ⁹⁾ In addition, we have previously reported volatile thiols as the key aroma component of FSS at levels of micrograms per liters,⁸⁾ and 2M3F is present in the FSS samples at a concentration as low as micrograms per liter. Hence, a standard curve for a 2M3F assay is prepared by GC-MS in the SIM mode. At concentrations ranging from 0 to 20 μg/L, the regression equation was linear in the form y = 0.5002x, r = 0.999 (y = 2M3F, μg/L; x represents the height of the thiol peak/the height of the internal standard peak). 2M3F was quantified in various FSS samples, to the best of our knowledge, this is the first report of the quantification of 2M3F in FSS by GC-MS analysis.

Table 1 summarizes the concentrations of volatile thiols in all FSS samples. Irrespective of the source, 2M3F was detected in the fermented form of five raw soy sauce samples at concentrations of 0.4–1.0 μg/L and in five heat-treated soy sauce samples at concentrations of 0.5–1.9 μg/L. As shown in Table 1, the 2FM, ET2MP, and BM concentrations in this study were in agreement with those previously reported.⁸⁾ The production process for soy sauce includes heating raw soy sauce at 70–90 °C for pasteurization, as a result, 2M3F and other volatile thiols are possibly formed by heating. Nevertheless, the perception threshold of 2M3F in soy sauce could not be directly determined. Alcoholic fermentation generates 2–3% ethanol during the aging of soy sauce, and Tominaga et al.¹¹⁾ have reported that the threshold for 2M3F in a model dilute alcohol solution is 4 ng/L. The concentration of 2M3F in all of the analyzed FSS samples was greater than this perception threshold. Thus, 2M3F contributes to the cooked meat-like aroma of fermented soy sauce (FSS). As such, sensory evaluation was imperative for investigating the contribution of 2M3F to the aroma of FSS.

Table 1. Concentrations of volatile thiols in FSS samples.

Compound	Threshold^11,17^) (ng/L)	Concentration (μg/L)
		Raw soy sauce		Heat-treated soy sauce
		Range^a	Average	Range	Average
2M3F	4	0.4–1.0	0.7	0.5–1.9	1.3
2FM	0.4	0.8–1.9	1.3	1.2–2.8	2.0
ET2MP	500	5.4–27.4	13.6	13.7–40.9	22.8
BM	0.3	0.1–0.8	0.3	0.3–0.7	0.6

^a Range of volatile thiol concentrations in five soy sauce samples.

Sensory analysis

For clarifying the contribution of 2M3F to the aroma of FSS, sensory analysis was performed using heat-treated soy sauce containing 0.5 μg/L of 2M3F; the concentration of 2M3F in this sample was maintained to be similar to the lowest value in the five heat-treated soy sauce samples. Paired comparison tests on soy sauce samples with and without the addition of 2M3F were conducted for examining whether the variation of 2M3F concentration would result in differences in the perception of cooked meat-like aroma. The results indicated that soy sauce samples with the addition of 0.2, 0.4, or 0.8 μg/L of 2M3F exhibited cooked meat-like aroma significantly stronger than those without the addition of 2M3F (Table 2). The majority of the FSS samples contained 2M3F in amounts greater than 0.5 μg/L, hence, 2M3F contributes to the cooked meat-like aroma of FSS. In addition, it is reasonable that, in view of its extremely low perception threshold, 2M3F contributed to the cooked meat-like character of raw soy sauce to a similar extent.

Table 2. Number of panelists who recorded stronger cooked meat-like aroma soy sauce (N = 34).

Concentration of 2M3F for addition to soy sauce (μg/L)		0.2	0.4	0.8
Number of panelists	Soy sauce	9	2	4
	Soy sauce with the addition of 2M3F	25	32	30
p-value（χ^2-test）		<0.01	<0.001	<0.001

Formation of 2M3F in the FSS samples

Soy sauce is used as seasoning during cooking, as a result, the aroma compounds are either newly formed or degraded. During the heat treatment of soy sauce, the 2FM concentration has been reported to be increased by greater than two times by the Maillard reaction of cysteine and furfural.^8,13) 2M3F and 2FM are formed by different pathways from the same precursor compounds: carbohydrate and sulfur amino acids.¹²⁾ Hence, 2M3F would be formed in FSS by heating.

Table 3 shows the changes in the concentrations of 2M3F and 2FM in A by heating. The 2M3F and 2FM concentrations significantly increased with temperature. Although their concentrations significantly increased with heating time at 120 °C, the concentrations of these thiols did not increase with heating time at 80–95 °C. Moreover, the concentration of 2FM in raw soy sauce was approximately two times greater than that of 2M3F, but as compared to 2FM, 2M3F was formed in higher amounts by heating. Notably, at 95 °C, almost 7 μg/L of 2M3F and 5 μg/L of 2FM were detected in the sample. These results imply that as compared to 2FM, 2M3F is formed more easily in FSS under the same heating condition.

Table 3. Concentrations of 2M3F and 2FM in raw soy sauce A after heating.

Compound	2M3F (μg/L)		2FM (μg/L)
Temperature	5 min^a^, ^d	20 min	5 min^e	20 min
–^b	0.3 ± 0.1^c		0.6 ± 0.1
80 °C	1.2 ± 0.1	2.4 ± 0.3	1.2 ± 0.1	1.4 ± 0.1
95 °C	6.6 ± 0.4	6.8 ± 0.3	5.0 ± 0.3	5.2 ± 0.1
120 °C	11.1^f ± 0.4	18.3^g ± 0.8	9.7^h± 0.7	18.2^i ± 0.6

^a Heating time.

^b Before heat treatment.

^c Values correspond to the mean ± SD (standard deviation) of four analyses.

^d,e Values in the same column are significantly different from each other (p < 0.05).

^f,g,h,i Values with different superscripts indicate significance (p < 0.01).

As shown in Table 4, the 2M3F concentration significantly increased by approximately four times in BH and by 19 times in B at 120 °C for 5 min. Furthermore, the 2M3F concentration tended to increase in B and BH with the addition of ribose and cysteine (0.05 < p < 0.1), but the 2M3F concentration did not increase by the addition of glucose and cysteine. In contrast, 2M3F was not detected in a pH 5 phosphate buffer by the addition of ribose or glucose and cysteine at 120 °C for 5 min (data are not shown).

Table 4. Changes of 2M3F and 2FM in FSS samples at 120 °C for 5 min^a.

Sample		Raw soy sauce (B, μg/L)		Heat-treated soy sauce (BH, μg/L)
Before heating		2M3F	2FM	2M3F	2FM
		0.7^b ± 0.1^d	1.2 ± 0.1	1.3^e ± 0.2	2.4 ± 0.2
After heating	Not added	13.5^c ± 0.3	8.9 ± 0.3	5.6^f ± 0.4	4.4 ± 0.1
	Added ribose + Cys	15.7 ± 1.1	12.1 ± 0.1	6.6 ± 0.2	6.8 ± 0.3
	Added glucose + Cys	12.4 ± 1.4	11.4 ± 0.1	4.3 ± 0.4	6.1 ± 0.5

^a Soy sauce samples (180 mL) were heated in a laboratory autoclave. For these heating experiments, cysteine (1.8 mM) and the corresponding carbohydrate (47 mM of glucose or 7.7 mM of ribose) were added.

^b,c,e,f Values with different superscripts indicate significance in raw or heat-treated soy sauce (p < 0.01).

^d Values correspond to the mean ± SD (standard deviation) of four analyses.

In the past 50 years, a large number of studies have been performed for clarifying the precursor compounds of 2M3F in food. Sugar and cysteine have been reported as effective precursors of 2M3F.¹⁵⁾ High concentrations of free amino acids and sugars (pentoses and hexoses) are formed during the fermentation of soy sauce. Therefore, when B or BH soy sauce was heated at 120 °C for 5 min, 2M3F is possibly formed by the Maillard reaction from these precursor compounds. In addition, it is possible that, when raw soy sauce was heated for pasteurization, the Maillard reaction caused the reduction in the amino acids and sugar concentrations. Thus, a higher amount of 2M3F is formed in B as compared with that in BH by heating at 120 °C for 5 min.

Typically, soy sauce contains more hexoses than pentoses; hence, it is possible that hexoses may be more important than pentoses for the formation of 2M3F in soy sauce. According to a study by Hofmann and Schieberle,¹²⁾ heating a phosphate buffer system by the addition of ribose and cysteine in the pH range of 3–7 at 145 °C for 20 min results in the decrease of 2M3F with increasing pH. The greatest yields for glucose and cysteine were obtained at pH 5. The pH of the FSS sample was approximately 4.7; hence, the results obtained herein for the increased 2M3F in FSS by heating with the addition of ribose and cysteine are in agreement with their study, although the reason for the changed of 2M3F concentration in FSS by heating glucose and cysteine is still not clear.

Jennifer et al.¹⁶⁾ have reported that 2M3F is detected from glucose and cysteine at pH 5.5 and 150 °C in a phosphate buffer, but 2M3F is not detected at 120 °C. The result obtained herein for 2M3F not being detected in the phosphate buffer at 120 °C is in agreement with that reported by Jennifer et al. To the best of our knowledge, there is no study on the generation of 2M3F by the heating of carbohydrate and cysteine at a temperature less than 145 °C for 5 min; hence, we believe that 2M3F is not formed in the phosphate buffer because of the low temperature and a short heating time. These results indicated that it is easier to generate 2M3F in FSS as compared to the phosphate buffer.

On the other hand, 2M3F was not detected in the phosphate buffer model by heating, but greater than 10 μg/L of 2M3F was formed in FSS under the same heating condition. Hence, it is interesting to check whether 2M3F can be detected in ASS. 2M3F was not detected in the aroma concentrates of three ASS samples analyzed by GC-O and GC-MS. Furthermore, 2M3F was not detected in any of the three ASS samples heated at 120 °C for 5 min.

2M3F was not detected in both the ASS samples and phosphate buffer model by heating, but 2M3F was detected in FSS and its concentration increased under the same heating condition. The largest difference between FSS and ASS or the phosphate buffer model is fermentation by micro-organisms. Hence, fermentation by micro-organisms in FSS possibly plays a key role in the formation of 2M3F.

Previously, we have reported that yeast plays an important role in the production of 2FM in FSS.¹³⁾ Although no studies have been reported on the microbial or enzymatic aspects associated with the production of 2M3F, to the best of our knowledge, 2M3F is a product typically obtained from the Maillard reaction by heating. Thus, further experiments should be conducted for clarifying the mechanism for the formation of 2M3F during the fermentation of soy sauce.

Author contributions

Qi Meng (QM) and Etsuko Sugawara designed the experiment. QM and Riho Kitagawa performed the experiments. Miho Imamura designed the sensory analysis and performed it by herself. QM wrote the manuscript. All authors contributed to the development of the manuscript. All authors have read and approved the final manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.