Materials and Methods
Experimental samples Fresh deep-sea smelt (Glossanodon semifasciatus) caught by bottom trawl in 2019 was purchased from fish markets in Ishikawa, Toyama, and Niigata, Japan, and stored frozen at −30 °C for approximately 4 months until use. Salt (Namien, JT Co., Ltd., Tokyo, Japan) and soy sauce koji (Masuichi Brewery Co., Ltd., Toyama, Japan) were used as sub-materials. The koji was prepared by incubating a 1:1 mixture of steamed defatted soybean and roasted wheat with Aspergillus oryzae (Ichimurasaki: Bio'c; Co. Ltd., Toyohashi, Japan). Lactic acid bacterium T. halophilus for soy sauce (SC1) (Akita Konno Co., Ltd, Akita, Japan), T. halophilus strain 8-25 (SC2), and T. halophilus strain 14-1 (SC3) were used as SCs. SC2 and SC3 were isolated from marine fermented foods at The Fisheries Research Technology Institute, Japan Fisheries Research and Education Agency (FRA), and National Research and Development Agency and did not harbor and receive the hdc gene. The three types of SC were cultured at the Akita Konno Co. Ltd. and transported to the Umekama Co. Ltd. at 10 °C. The abundance of SC bacteria (most probable number, MPN/mL) in each culture before fish sauce production was 2.3 × 108, 4.3 × 108, 4.3 × 107, 2.3 × 107, 4.3 × 108, and 3.9 × 108 MPN/mL for SC1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2, respectively.
Production of deep-sea smelt fish sauces with SCs A 500 L polypropylene fermentation tank was cleaned prior to fish sauce production to remove any histamine-producing bacteria. The outside and inside of the tank drain were washed with a steel wool or brush. The insides of the cover and drain were disinfected using an ethanolic formulation for food additives (San Circle, Nippon Shinyaku Co., Ltd., Kyoto, Japan). Then, approximately 2 500 kg of frozen smelt was thawed overnight at 4 °C. The whole fish was minced in a meat grinder. The minced meat (120 kg) was mixed with 30.5 kg of salt, 18 kg of soy sauce koji, 6.5 kg of sugar, and 40 L of tap water. The prepared moromi was inoculated with SC1 (0.25 L), SC2 (1.5 L), and SC3 (1.5 L) and mixed well. One-half of each fermentation tank volume was mixed in two parts and transferred to the tank. The total amount of moromi per tank of SC1, SC2, and SC3 was approximately 430.5, 433.0, and 433.0 kg, respectively, with two replicates per SC. The total SC bacterial counts in the moromis were 2.7 × 105, 5.0 × 105, 3.0 × 105, 1.6 × 105, 3.0 × 106, and 2.7 × 106 MPN/g for SC-1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2, respectively. The tanks were covered, stacked, and left to ferment for approximately 6 months in a prefabricated house without air conditioning. The monthly average outer temperature from June to December in Toyama city in 2019 iv) ranged from approximately 12–32 °C. Tank positions were regularly rotated, and the tanks were stirred well using a paddle at 14-d intervals to achieve uniform fermentation. Each moromi was collected on days 0, 7, 14, 30, 61, 91, 122, and 188 of fermentation. After fermentation, the six moromis were placed in nylon bags and the sauce extracted in a compressor (KS-4, Shinkomagata Machinery, Tokyo, Japan). The extracted fish sauce was then heated to 85 °C for 30 min in a steam caldron (Hattori Industry, Okazaki, Japan), cooled to room temperature, and then filtered in a circulating filtration machine with 1.68 m2 membranes that had a 0.2 µm pore size (7M-1, Noritake, Nagoya, Japan). Part of the filtrate was bottled in plastic bottles as the final product.
Halophilic and histamine-producing bacterial counts Halophilic and histamine-producing bacterial counts in the moromi were measured by the MPN method (Satomi et al., 2008; Kimura et al., 2014). Starter cultures, halophilic, and histamine-producing bacterial counts in the moromi were measured using the MPN method to detect low levels of the latter two bacteria. Thirty-six milliliters of histidine broth (1 % glucose, 1 % peptone, 0.2 % yeast extract, 0.5 % L-histidine, and 10 % NaCl, pH 6.5) was added to 4 g of the moromi and mixed well using a stirrer. Serial dilution of this solution was conducted using histidine broth, and three test wells were used to culture each dilution stage at 30 °C for 14 d. The halophilic bacterial counts were calculated using the MPN table from the number of test wells that exhibited turbidity. The histamine-producing bacterial counts were calculated using the MPN table from the number of wells that showed a positive histamine reaction. The histamine-positive reaction was evaluated using a Check Color Histamine Kit (Kikkoman, Noda, Japan) (Sato et al., 2005).
Preparation of the assay sample Approximately 50 g of each moromi was collected over time during fermentation and centrifuged (10 000 ×g for 30 min at 4 °C). The supernatant was filtered using No. 5C filter paper (Toyo Roshi Co., Ltd., Tokyo, Japan), and the filtrate was used as the physicochemical analytical sample. The liquefaction ratio was calculated from the percentage of moromi weight after centrifugation and filtration relative to that before centrifugation.
Physicochemical analysis The color (L*, a*, and b* values) of the analytical sample was measured using an SA4000 spectrophotometer (Nippon Denshoku Industries Co., Ltd., Tokyo, Japan) and using the transmission method with a glass cell (2 mm × 40 mm × 50 mm). The pH of the sample was measured using a pH meter (HM-30R, DKK-TOA Cooperation, Tokyo, Japan) at room temperature. Titratable acidity and formol nitrogen level were determined according to the soy sauce test method (Japan Soy Sauce Institute, 1985). The content of soluble solids excluding salt in the samples was calculated after subtracting the salt content from the Brix value (Japan Soy Sauce Institute, 1985). The total nitrogen content was determined using the Kjeldahl method (Tsutsumi and Yasui, 1996). The protein degradation ratio (PDR) was calculated as the ratio (%) of formol nitrogen to the total nitrogen (Okazaki and Noguchi, 2008). Furthermore, the histamine level of the sample was determined using a Check Color Histamine Kit (Kikkoman).
Free amino acids The crude protein in the sample was removed by adding 2 % sulfosalicylic acid and filtering through a 0.22 µm nylon filter. The amino acid composition of each filtered sample was measured using an auto amino acid analyzer (L-8900, Hitachi High-Tech Corporation, Tokyo, Japan).
Organic acids Each analytical sample was diluted five times with ultrapure water. The diluted sample was mixed with a 5 % perchloric acid solution in a 1:1 (v/v) ratio. The sample was then filtered using a 0.45 µm cellulose acetate filter after centrifugation (9 167 × g for 15 min at 4 °C). Organic acids were determined by post-column labelling using high-performance liquid chromatography (HPLC). The HPLC analysis conditions were as follows: guard column, RSpak KC-G6B (Showa Denko K.K., Tokyo, Japan); separation columns, RSpak KC-811 (8.0 mm I.D. × 300 mm, Showa Denko K.K., Tokyo, Japan) ×2; eluent, 3 mM perchloric acid solution (pH 2.5); flow rate, 1.0 mL/min; reaction liquid, 0.2 mM BTB-15 mM sodium phosphate solution; column temperature, 60 °C; detection wavelength, 445 nm; and injection volume, 10 µL.
Taste sensor analysis The final fish sauce products were diluted 10 times with distilled water. Multiple taste tests of the diluted products were conducted using a taste sensor (TS) (TS5000Z, Intelligent Sensor Technology Inc., Atsugi, Japan). The taste of the sample was interpreted using five types of first taste (bitterness/food, bitterness/medicine, umami, saltiness, and sourness) and three kinds of aftertaste (bitterness/food, bitterness/medicine, and umami).
Statistical analysis Statistical analyses of the data obtained in the physicochemical, extracted component, and TS analyses were performed by one-way analysis of variance using JMP 14 (SAS Institute Japan Ltd.), and the statistical significance of the mean differences was determined using the Tukey-Kramer HSD test with a significance level of 95 %. The relationship among the physicochemical, extractive component, and TS analysis data was investigated using the same software. Principal component analysis (PCA) (covariance matrix) was performed using TS data and quality indicators to observe the above correlation, and the results were visualized in a PCA biplot using Ward's hierarchical cluster analysis.
Results and Discussion
Effect of SC type on physicochemical properties of moromis during fermentation The liquefaction ratio (Fig. 1A) of all samples increased sharply in the first 7 d and then increased to a lesser degree until the end of fermentation. The liquefaction ratio of all samples on day 188 was 52.0–54.0 %. Salt-soluble nitrogen levels (Fig. 1B) in all samples increased sharply up to day 7 and slowly increased over the next 14 d, fluctuating approximately 1.8 % until the end of fermentation. Formol nitrogen levels (Fig. 1C) in all samples increased sharply up to day 7, continued to increase over the next 54 d, decreased slightly until day 91, and then increased slowly toward the end of fermentation. The tendency of PDR (D) was almost the same as that of formol nitrogen except for a fluctuation on day 121. The PDR on day 188 was higher in SC2-1 and SC2-2 (72.0 %) than in the other samples (67.9–69.0 %). This difference in the PDR of the sample was likely caused by differences in the main components of the raw material, such as the intestine and meat. Consistent with our results, Furutani et al. (2012) reported that moromi protein degradation considerably progresses owing to protease originating from Aspergillus oryzae rather than from endogenous digestive enzymes in the viscera of fish during 7 d of fermentation, during which the liquefaction ratio of the sample increases dramatically.
The pH levels of the SC3-1 and SC3-2 samples decreased dramatically to below 5.0 in the first 14 days of fermentation (Fig. 2A). The pH levels of the SC1-1, SC1-2, SC2-1, and SC2-2 samples also rapidly decreased below 5.0 in the first 30 d, and all samples remained below pH 5.0 at the end of fermentation. Organic acid levels in the SC3-1 and SC3-2 samples dramatically increased from days 0 to 14 (Fig. 2B), and those of the SC1-1, SC1-2, SC2-1, and SC2-2 samples dramatically increased from days 14 to 30. The organic acid levels of all samples slowly increased from day 30 until the end of fermentation. The increase in organic acid levels from days 30 to 188 was greater in SC2-1 and SC2-2 than in the other samples, which caused a difference in the growth of halophilic bacteria in the moromi on the same days in SC2-1 and SC2-2 compared with that in other samples (Fig. 3A). Lactic acid levels in the samples for the first 30 d were almost the same as the total organic acid levels (Fig. 2C). However, the lactic acid levels of all samples slowly decreased from days 30 to 61 and then slowly increased toward the end of fermentation. Therefore, the rapid decrease in the pH from days 0 to 30 may have been associated with the rapid increase in lactic acid levels. The increase in organic acids (such as pyroglutamic and acetic acids) that occurred from days 30 to 188 may have influenced the fluctuating pH levels of all samples owing to the plateau of lactic acid production in all samples, as lactic acid accounted for approximately two-thirds of the total organic acids. A negative correlation was observed between pH and lactic acid levels in all samples (r = −0.9531, p < 0.0001). In this study, the lactic acid production of SC3-1 and SC3-2 was higher than that of the other SCs. This may be owing to the differences in the carbohydrate fermentation patterns of salt-tolerant strains of T. halophilus in various types of moromis (Higashi et al., 1998). The suppression of histamine accumulation in our study was almost the same as that reported in a previous study (Shozen et al., 2012), despite the different rates of pH decrease and increases in lactic acid.
Effect of SC on microbial properties and histamine levels of moromis during fermentation Halophilic bacterial counts in SC1-1 and SC1-2 decreased from days 30 to 122 and increased at the end of fermentation (Fig. 3A). Halophilic bacterial counts of SC2-1 and SC2-2 increased from days 0 to 30 and slowly decreased from days 30 to 188. Halophilic bacterial counts of SC3-1 and SC3-2 increased from days 0 to 14, slowly decreased from days 14 to 122, and then increased from days 122 to 188. Moreover, histamine-producing bacterial counts in SC1-1 and SC1-2 increased from day 14, peaked on days 61 and 30, respectively, and then slowly decreased until day 122 (Fig. 3B). Histamine-producing bacterial counts in SC2-1 and SC2-2 increased from day 14, reached a peak at day 61, and slowly decreased from 122 d. Histamine-producing bacteria were slightly detected in SC3-1 during fermentation (0.4–0.7 MPN/g) and were only slightly detected in SC3-2 on day 30 (2.3 MPN/g). The histamine-producing bacterial abundance in all samples was less than 9.3 × 10 MPN/g during the fermentation period.
The abundance ratios of the halophilic bacterial counts/histamine-producing bacterial counts of the SC1, SC2, and SC3 groups were 10 000:1–100 000:1, 100 000:1, and 10 000:1–1 000 000:1, respectively, during the fermentation period associated with the highest number of histamine-producing bacteria. The halophilic bacterial counts/histamine-producing bacterial counts of all samples were maintained above 10 000:1 throughout fermentation, although growth suppression of histamine-producing bacteria during fermentation differed from that of the SCs. In a previous study, the histamine level in fish sauce mash was approximately 200 mg/kg during fermentation when the SC bacterial count/histamine-producing bacterial count was over 1 000:1 (Kimura et al., 2015). In the present study, histamine levels were below 115 mg/L in moromis following inoculation with SCs; an increasing trend of histamine levels was observed in all samples until the end of fermentation, although histamine levels fluctuated from 2 to 115 mg/L throughout fermentation (Fig. 3C).
Effect of SC on the properties of the physicochemical and extractive components of the final fish sauce products The L* values (brightness) ranged from 84.73 to 88.20 for all samples. The a* (redness) and b* values (yellowness) of all the samples decreased between −0.38 to 2.63 and 42.18–61.63, respectively (Table 1). There was a significant difference (p < 0.05) in the titratable acidity and pH levels between the samples with SCs. The titratable acidity was the lowest in SC1-1 and SC1-2, followed by SC2-1 and SC2-2, and finally SC3-1 and SC3-2; the pH showed the opposite trend. The content of soluble solids excluding salt and total nitrogen levels in all samples was approximately 16 % and 1.82–1.93 g/100 mL, respectively, corresponding to the standard values of special grade soy sauce v) set by the Japanese Agricultural Standard. The PDR of all samples was 59.5–61.8 %, and the histamine levels of all samples were 43–113 mg/L, in compliance with the CODEX iii) limit for fish sauce. Histamine levels were lower than those (141–187 mg/L) in commercial fish sauces or fish sauce products obtained from factories in Thailand (Yongsawatdigul et al., 2004). High level of histamine accumulation (over 400 mg/L) have been found in the fish sauce without lactic acid bacteria (Taira et al., 2007). Therefore, no problematic histamine accumulation was detected during fish sauce production in this study. Furthermore, in this study, final products containing a low histamine level (below 113 mg/L) were obtained by a combination of sanitation and control of SC inoculation level because the suppression effect of histamine accumulation was higher when sanitation and fermentation control were combined (Satomi, 2016). The details are under consideration.
Table 1
Physicochemical properties of various types of fish sauce products.
|
|
SC1-1 |
SC1-2 |
SC2-1 |
SC2-2 |
SC3-1 |
SC3-2 |
|
L* |
86.45a |
85.63a |
87.51a |
84.73a |
88.20a |
86.57a |
| Color |
a* |
0.96bc |
2.07ab |
−0.38d |
2.63a |
−0.08cd |
0.82cd |
|
b* |
52.68c |
57.10a |
43.42c |
61.63a |
42.18c |
54.00b |
| pH |
|
4.93a |
4.91ab |
4.89ab |
4.83bc |
4.76c |
4.77c |
| Titrative acidity (m L) |
19.0c |
20.0c |
20.3bc |
21.5ab |
21.8a |
21.5ab |
| Salt content (g/100 mL) |
18.0a |
18.3a |
17.7a |
18.2a |
17.4a |
18.1a |
| Soluble solids excluding salt (%) |
16a |
16a |
16a |
16a |
16a |
16a |
| Total nitrogen (g/100 mL) |
1.86ab |
1.85ab |
1.87ab |
1.89ab |
1.82b |
1.93a |
| Formol nitrogen (g/100 mL) |
1.11a |
1.13a |
1.13a |
1.14a |
1.12a |
1.15a |
| Protein degradation ratio (%) |
59.7a |
60.7a |
60.4a |
60.2a |
61.8a |
59.5a |
| Histamine (mg/L) |
95a |
113a |
87a |
101a |
43a |
61a |
The data are expressed as the mean values (n = 3). Superscript letters indicate statistically significant differences (p < 0.05). See Fig.1 for SC1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2.
Five organic acids were detected in the samples (Table 2). The total organic acid levels were the lowest in SC1-1 and SC1-2, followed by SC3-1 and SC3-2, and finally SC2-1 and SC2-2. The primary organic acid was lactic acid. Malic acid was not detected in SC1-1 or SC1-2 but was identified in the other samples. Succinic, acetic, and pyroglutamic acids were detected in all samples. The lack of malic acid in SC-1 and SC-2 may be attributed to malic acid metabolism by lactic acid bacteria, which has been previously described to occur during soy sauce production (Kanbe et al., 1978).
Table 2
Organic acid compositions of various types of fish sauce products (mg/100 mL)
| Acids/sample |
SC1-1 |
SC1-2 |
SC2-1 |
SC2-2 |
SC3-1 |
SC3-2 |
| Succinic |
32bc |
38a |
23d |
19e |
30c |
34ab |
| Lactic |
902b |
917b |
1153a |
1126a |
1079a |
1082a |
| Acetic |
133c |
182ab |
197a |
190ab |
178b |
136c |
| Pyroglutamic |
206cd |
203cd |
244b |
280a |
194d |
222c |
| Malic |
ND |
ND |
76a |
72a |
43ab |
50b |
| Total |
1273c |
1340c |
1694a |
1687a |
1524b |
1523b |
The data are expressed as the mean values (n = 3). Superscript letters indicate statistically significant differences (p < 0.05). ND: not detected. See Fig.1 for SC1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2.
Twenty-one types of free amino acids were detected in the samples (Table 3). The total levels of free amino acids ranged from 78.64 to 83.51 mg/mL, and the main free amino acids were Asp, Glu, Ala, Val, Ile, Leu, Lys, and Arg. Similarly, taste-active components found in a Vietnamese fish sauce (Park et al., 2002) included Glu, Thr, Pro, Ala, Val, Cys, Tyr, Met, His, and pyroglutamic acid. The concentrations of these components were 36.52–37.96, 37.80–38.26, and 35.87–36.60 mg/mL in SC1-1 and SC1-2, SC2-1 and SC2-2, and SC3-1 and SC3-2, respectively. The taste strength of these products was almost identical to that of fish sauce products with soy sauce koji (33.13–37.34 mg/mL), nuocman (36.61 mg/mL), and soy sauce (38.11 mg/mL) (Taira et al., 2007). Thus, the strength of the products made with SCs was likely to be sufficient to exhibit deep and complex tastes (Park et al., 2002).
Table 3
Free amino acid compositions of various types of fish sauce products (mg/100 mL).
| Amino acid/sample |
SC1-1 |
SC1-2 |
SC2-1 |
SC2-2 |
SC3-1 |
SC3-2 |
| Tau |
99ab |
90c |
101ab |
103a |
94bc |
97b |
| Asp |
727a |
691ab |
509e |
550d |
651c |
681bc |
| Thr |
420a |
406ab |
405ab |
416a |
392b |
400ab |
| Ser |
432a |
415ab |
416ab |
424a |
386c |
396bc |
| Asn |
135a |
120bc |
104d |
104d |
126b |
114c |
| Glu |
1237a |
1181ab |
1122b |
1113b |
1133b |
1146b |
| Gly |
250a |
242a |
238a |
240a |
216b |
222b |
| Ala |
566b |
550b |
683a |
679a |
543b |
547b |
| Pro |
329a |
319a |
324a |
327a |
328a |
331a |
| Val |
543a |
520a |
524a |
537a |
522a |
532a |
| Cys |
6c |
4e |
12a |
4e |
11b |
5d |
| Met |
249a |
240a |
244a |
239a |
243a |
246a |
| Ile |
499a |
487a |
492a |
491a |
494a |
502a |
| Leu |
778a |
749ab |
758a |
708b |
760a |
769a |
| Tyr |
62a |
59ab |
55c |
56bc |
57bc |
60ab |
| Phe |
438a |
414a |
417a |
426a |
423a |
429a |
| Trp |
13a |
13a |
ND |
ND |
ND |
16a |
| Orn |
86b |
77c |
94a |
70d |
35e |
36e |
| Lys |
825a |
776b |
792ab |
813ab |
784ab |
800ab |
| His |
178a |
169ab |
168ab |
173ab |
164b |
171ab |
| Arg |
480b |
447c |
407d |
458bc |
515a |
521a |
| Total |
8351a |
7969a |
7864a |
7933a |
7876a |
8021a |
The data are expressed as the mean values (n = 3). Superscript letters indicate statistically significant differences (p < 0.05). ND: not detected (detection limit: < 0.3 mg/100 mL). See Fig.1 for SC1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2.
Gly, Ala, Thr, Pro, and Ser free amino acids (FAAs) are related to a sweet taste; Phe, Tyr, Arg, Ile, Leu, Val, Met, and Lys elicit a bitter taste; and Glu and Asp are related to umami and sourness (Ninomiya, 1999). Lys, Arg, and His are basic amino acids, whereas Glu and Asp are acidic amino acids (Kubota and Morimitsu, 2009). In this study, the sweet amino acid ratio (AAR) was determined by comparing the combined abundance of Gly, Ala, Thr, Pro, and Ser relative to the total FAAs. The bitter AAR was determined using the ratio of the sum of Phe, Tyr, Arg, Ile, Leu, Val, Met, and Lys to the total FAAs. The umami AAR:acidic AAR was determined by the percentage of the sum of Glu and Asp relative to the total FAAs. The basic AAR (BAAR) was determined using the percentage of the sum of Lys, Arg, and His to total FAAs. Umami AAR was the highest in SC1-1 and SC1-2 (23.5 %), followed by SC3-1 and SC3-2 (22.7–22.8 %), and finally SC2-1 and SC2-2 (20.7–21.0 %). Sweet AAR was the highest in SC2-1 and SC2-2 (26.3 %), followed by SC1-1 and SC1-2 (23.9–24.2 %), and SC3-1 and SC3-2 (23.6–23.7 %). Bitter AAR was the highest in SC3-1 and SC3-2 (45.0–45.1 %), followed by SC2-1 and SC2-2 (43.8–44.0 %) and finally SC1-1 and SC1-2 (43.3–43.4 %). The BAAR was the highest in SC3-1 and SC3-2 (18.6 %), followed by SC2-1 and SC2-2 (17.4–18.2 %), and finally SC1-1 and SC1-2 (17.5–17.8 %). Therefore, SC1, SC2, and SC3 were the strongest in umami AAR, sweet AAR, and bitter AAR and BAAR tastes, respectively. Considerable differences in umami AAR can occur depending on the fish species, according to a survey of umami AAR in Shottsuru (Tsukamoto et al., 2017). In one study, umami AAR and sweet AAR decreased, and bitter AAR slightly increased when the salt content decreased from 28 % to 20 % (Takahashi et al., 2023).
Relationship between taste, quality indicators, and SC in final products Food TS analysis is performed to measure taste in food manufacturing facilities (Wu et al., 2020; Toko, 2022). Some reports describe differences in the tastes of various fish or meat sauce products produced using different fermentation methods based on TS analysis (Funatsu et al., 2015; Funatsu et al., 2021; Harata et al., 2022; Takahashi et al., 2023). However, the relationship between taste data and indicators related to taste is unclear. In this study, the multivariate correlation between indicators (pH, titratable acidity, salt content, total nitrogen, formol nitrogen, PDR, histamine, lactic acid, umami AAR, sweet AAR, bitter AAR, and BAAR) related to taste (Tables 1–3) and TS data was investigated (Table 4). Positive (n = 8) and negative (n = 20) correlations were observed between 10 types of indicators except for salt content, formol nitrogen, and TS data (Table 5). The PCA of the 10 indicators and TS data showed that the fish sauce products were separated into three groups (Group I: SC1-1 and SC1-2, Group II: SC2-1 and SC2-2, Group III: SC3-1 and SC3-2) according to a hierarchical cluster analysis (Fig. 4). The distance was closer in Groups II and III than in Group I, with Groups II and III showing similar characteristics. The contribution ratios of PC1 and PC2 were 95.3 and 4.68 %, respectively, and the cumulative contribution was 99.98 % according to the PCA biplot. The taste characteristics of Group I included umami (first taste) saltiness (first taste) and bitterness/food (bitterness originated from raw materials). The indicators related to these tastes were mainly umami AAR and pH. The taste characteristics of Groups II and III were sourness (first taste), umami (after taste), and bitterness/medicine. These tastes were stronger in Group III than in Group II based on the direction and distance of the vectors that show the tastes. The indicators related to the tastes were mainly bitter AAR, BAAR, titratable acidity, and lactic acid content. Bitter AAR positively correlated (r = 0.8993, p = 0.0147) with bitterness/medicine (first taste). Bitterness/medicine (first taste) may originate from bitter AAR.
Table 4
Taste properties of various types of fish sauce products estimated using taste sensor analysis.
| Samples/taste |
First taste |
After taste |
| Bitterness/medicine |
Bitterness/food |
Umami |
Saltiness |
Sourness |
Bitterness/medicine |
Bitterness/food |
Umami |
| SC1-1 |
2.64cd |
0.10a |
2.81a |
−0.36a |
−5.87e |
1.26a |
−0.01ab |
0.18a |
| SC1-2 |
2.62c |
0.15a |
2.65b |
−0.40b |
−5.49d |
1.41a |
0.01a |
0.22a |
| SC2-1 |
2.76bc |
0.07a |
2.57b |
−0.66d |
−5.44d |
1.46a |
−0.08b |
0.21a |
| SC2-2 |
2.68cd |
0.15a |
2.13c |
−0.68d |
−4.56c |
1.33a |
−0.01ab |
0.21a |
| SC3-1 |
3.07a |
0.11a |
2.04cd |
−0.98e |
−4.40b |
1.47a |
−0.06ab |
0.25a |
| SC3-2 |
2.85b |
0.11a |
1.96d |
−0.60c |
−4.19a |
1.25a |
−0.07ab |
0.28a |
The data are expressed as the mean values (n = 3). The synthetic standard solution of soy sauce was set to zero.
Superscript letters indicate statistically significant differences (p < 0.05).
See Fig.1 for SC1-1, SC1-2, SC2-1, SC2-2, SC3-1, and SC3-2.
Table 5
Correlation between two pair variables.
| Variable |
vs. Variable |
Correlation coeffect |
p value* |
| Saltiness (first taste) |
Bitterness/medicine (first taste) |
−0.8806 |
0.0025 |
| Sourness (first taste) |
Umami (first taste) |
−0.9985 |
< 0.0001 |
| Umami (after taste) |
Umami (first taste) |
−0.8155 |
0.0479 |
| Umami (after taste) |
Sourness (first taste) |
−0.8326 |
0.0397 |
| pH |
Bitterness/medicine (first taste) |
−0.8323 |
0.0398 |
| pH |
Sourness (first taste) |
−0.9683 |
0.0015 |
| pH |
Umami (first taste) |
0.9725 |
0.0011 |
| pH |
Umami (after taste) |
−0.8502 |
0.032 |
| PDR |
Bitterness/medicine (after taste) |
−0.858 |
0.0288 |
| PDR |
TN |
−0.8246 |
0.0434 |
| Histamine |
Bitterness/medicine (first taste) |
−0.9613 |
0.0022 |
| Histamine |
pH |
0.8268 |
0.0424 |
| UmamiAAR |
Lactic acid |
−0.8674 |
0.0252 |
| SweetAAR |
UmamiAAR |
−0.9085 |
0.0122 |
| BitterAAR |
Bitterness/medicine (first taste) |
0.8993 |
0.0147 |
| BitterAAR |
Umami (first taste) |
−0.9041 |
0.0134 |
| BitterAAR |
Sourness (first taste) |
0.8952 |
0.0159 |
| BitterAAR |
Umami (after taste) |
0.8484 |
0.0327 |
| BitterAAR |
pH |
−0.9727 |
0.0011 |
| BitterAAR |
Histamine |
−0.9273 |
0.0077 |
| BAAR |
Umami (first taste) |
−0.8796 |
0.0209 |
| BAAR |
Sourness (first taste) |
0.8791 |
0.021 |
| BAAR |
pH |
−0.9105 |
0.0177 |
| Titratable acidity |
Umami (fitst taste) |
−0.9669 |
0.0016 |
| Titratable acidity |
Saltiness (first taste) |
−0.819 |
0.0462 |
| Titratable acidity |
Sourness (first taste) |
0.9619 |
0.0021 |
| Titratable acidity |
pH |
−0.9339 |
0.0462 |
| Titratable acidity |
BitterAAR |
0.8433 |
0.0349 |
UmamiAAR: umami amino acid ratio; SweetAAR: sweet amino acid ratio; BitterAAR: bitter amino acid ratio; BAAR: basic amino acid rario; PDR: protein degradation ratio; TN: total nitrogen; *: significant difference.
In recent years, there have been reports that the metabolism of bacteria contaminating fish sauce contributes to its flavor. Han et al. (2023) investigated the association between microbiota and the characteristic flavor of different fish sauces and reported that the top 10 genera related to flavor generation, such as Lactobacillus, Staphylococcus, and Enterobacter, appeared to play a prominent role in the flavor formation of fish sauce. In natural fermentation, halophilic lactic acid bacteria were identified as the most important bacteria that contribute to flavor and aroma in fish sauce (budu) (Sakpetch et al., 2022). Moreover, a previous study reported improvement in the taste of fish sauce when using halophilic lactic acid bacteria as the fermentation starter. Wakinaka et al. (2019) reported that the use of T. halophilus, possessing aspartate decarboxylase, as a fish sauce fermentation starter converts aspartate to Ala, thus causing taste alteration, and a milder tasting fish sauce. In this study, Group II could be milder than Groups I and III owing to the Ala level being higher in SC2-1 and SC2-2 than in the other SCs (Tables 3 and 4).
A Shottsuru product produced with T. halophilus strain 14-1 containing 20 % salt showed lower Arg levels and higher Orn levels than a product produced without T. halophilus strain 14-1 under the same salt conditions (Takahashi et al., 2023) due to amino acid decomposition (Tanaka, 2012). However, this phenomenon was not observed in the present study (Table 3). This discrepancy is attributed to the difference in fermentation conditions, such as the use of koji, the fermentation temperature, and salt concentration.
Fukami et al. (2004) reported that the fishy, sweaty, fecal, and rancid notes of the fish sauce treated with the bacterium, Staphylococcus xylosus, were all weaker than those of the non-treated fish sauce using quantitative descriptive analysis (QDA). In the present study, the differences in sensory attributes among the samples are not obvious, as QDA sensory evaluation was not conducted; therefore, further studies of the sensory attributes among the samples are needed. Moreover, a survey about consumer preference for fish sauce products produced with diverse types of SCs should be conducted to potentially spread interest in the production of fish sauce using SCs to meet consumer interest abroad vi). Furthermore, future studies should evaluate the relationship between consumer preference and quality indicators that were identified in the present study.
Collectively, our findings show that histamine accumulation in moromi during fermentation is suppressed below the CODEX limit by ensuring factory hygiene (tank cleaning) before fish sauce production and using appropriate amounts of SC inoculant for fish sauce production. Various fish sauce products with multiple tastes might be produced through the selective use of SCs.
), SC2-1 (▲), SC2-2 (
), SC3-1 (■), and SC3-2 (
) for 
