Dried Kelp (Kombu), Scallop, and Dried Bonito (Katsuobushi)

Aqua-BioScience Monographs, Vol. 10, No. 1, pp. 1–22 (2017) www.terrapub.co.jp/onlinemonographs/absm/

Flavor Constituents in Savory Seafood: Dried Kelp (Kombu), Scallop, and Dried Bonito (Katsuobushi)

Yoichi Ueda1* and Kenji Fukami2

1Japan Food Additives Association Kodenma-cho Shin Nihonbashi Building, 4-9, Nihonbashi-Kodenma-cho, Chuo-ku, Tokyo 103-0001, Japan 2Ajinomoto Co., Inc., 1-15, Kyobashi 1-chome, Chuo-ku, Tokyo 104-8315, Japan *e-mail: [email protected]

Abstract Received on Omission tests of the synthetic water extract of dried kelp (Kombu-dashi) revealed that November 10, 2015 the water solution prepared by mannitol, monosodium glutamate (MSG), potassium chlo- Accepted on April 13, 2016 Online published on ride (KCl), and sodium chloride (NaCl) represented the whole taste characters of the March 10, 2017 synthetic extract. Succeeding study suggested that the characteristic Kombu-dashi like kokumi flavors (continuity, mouthfulness, and thickness) were generated by flavor inter- Keywords action among mannitol, MSG, and KCl. Some studies of taste interactions among NaCl, • dried kelp KCl, sweet amino acids, and 5¢-inosinate (IMP) showed that NaCl strengthened umami • garlic and sweetness of some amino acids and also suggested that IMP enhanced umami of Gly, • onion • scallop L-Ser, and L-Ala synergistically. The sulfur-containing compounds in garlic and onion, • dried bonito such as alliin, S-propenyl-L-cysteinesulfoxide and reduced glutathione (GSH), had no • mannitol tastes by themselves but they gave rise to kokumi flavors in umami solution or soups. • glutamate Scallop and beef contained a high amount of GSH and the peptide strengthened kokumi • sweet amino acid flavors of synthetic extracts of these foods. The examination of volatile compounds in • sulfur compound dried bonito by using GC-MS and GC-sniffing methods revealed that acetol and 2,3- • glutathione pentandione reacted with amino acids or protein in boiled bonito meat, generated pyrazines. • pyrazin Straight-chain aldehydes might be involved in the deterioration of crushed dried bonito. • hydrogen sulfide • umami Hydrogen sulfide, one of the key flavor components, decreased during preservation of • flavor interaction 3+ crushed dried bonito catalytically by the complex of Fe and histidine. This study, in- • kokumi cluding new findings, would lead to improve the quality of processed foods and contrib- • deterioration odor ute to future studies of taste phisiology and flavor chemistry.

1. Introduction modalities are called “kokumi flavors” in Japan. Aroma or odors of foods are generated by volatile compounds. Various kinds of food flavoring materials are used In the case of several foods, such as fruits, the whole for traditional dishes or cuisine in the world. Many aroma character can be reconstructed by a few kinds studies on flavors of foodstuffs have been done. Many of volatile compounds. In many cases a lot of volatile of their subjects were volatile odor components or non- components, however, comprise whole flavor charac- volatile taste active compounds which have basic tastes teristics of foods. For the components easily disappear (sweetness, sourness, saltiness, bitterness, or umami). or change to other compounds by the reactions among On the other hand many foods have flavor characters them, it has been a big issue how to control unstable expressed by terms such as complexity, body, continu- volatile compounds generated after enzymatic or ther- ity, mouthfulness, thickness, etc. It is difficult to de- mal reaction. fine these terms clearly, but these flavors sometimes Consumers request various kinds of high quality govern palatability of the foods. These flavor charac- processed foods which are convenient, have attractive teristics are presumed to be generated by the results of tastes or flavors, natural feeling, etc. To reply to those interactions among components in foods. These flavor demands from industry, basic studies are inevitable to

Table 1. Role of each component on the flavor of synthetic seafood extracts. Cited from Trends in Food Science & Technol- ogy, December 1996 (Vol. 7), Fuke S and Ueda Y, Interactions between umami and other flavor characteristics, 407–411, Table 1, „ 1996, with permission from Elsevier.

Component removed Effect on the synthetic extract Sea urchin Snow crab Scallop Short-necked clam Dried skipjack

Glutamate Umami Ø Umami Ø Umami Ø Umami Ø Umami Ø Sweetness ≠ Sweetness Ø Sweetness Ø Sweetness Ø Sweetness Ø Character Ø Palatability Ø Palatability Ø

AMP No effect Umami Ø Umami Ø Umami Ø No effect Sweetness Ø Sweetness Ø Palatability Ø Palatability Ø

IMP Umami Ø No effect No effect No effect Umami Ø Aftertaste Ø Sweetness Ø Palatability Ø

Glycine Sweetness Ø Umami Ø Sweetness Ø Sweetness Ø No effect Character Ø Sweetness Ø Palatability Ø Bitterness ≠

Alanine Sweetness Ø Sweetness Ø Sweetness Ø No effect No effect Bitterness ≠

Arginine Umami ≠ Character Ø Character Ø Character Ø No effect Sweetness ≠ Palatability Ø

+ Na NA Sweetness Ø Umami Ø Sweetness Ø Sourness Ø Umami Ø Character Ø Umami Ø Character Ø Character Ø Palatability Ø Sourness Ø Palatability Ø

Cl- NA Palatability Ø Sweetness Ø Sweetness Ø Sourness Ø Umami Ø Umami Ø Character Ø Palatability Ø Palatability Ø Palatability Ø

“Character” refers to the specific characteristic flavor of the seafood. The descriptions used were: umami, bitterness, specific flavor, aftertaste, sourness, saltiness, palatability, no effect. Effects on saltiness are not shown. ≠ indicates flavor characteristic increased when component removed. Ø indicates flavor characteristic decreased when components removed. NA, Not analyzed.

develop new ingredients and technology. We have en- tion will deal with the flavor characteristics of glutath- gaged in the research to get basic new knowledge on ione contained in scallop and other food materials, fo- kokumi flavors and to control unstable volatile com- cusing on the flavor interaction with umami substances pounds for quality improvement of processed foods, or other components extractive from some foods. The by the combination of instrumental analyses and sen- thermal degradation mechanism of glutathione will also sory evaluation. In this monograph, we will show three be mentioned. The third section will give an overview categories of our studies on the flavor constituents in of the mechanisms of increase or decrease in key vola- seafood and other foodstuffs. First of all, we will de- tile flavor components of dried bonito. scribe the studies on the key non-volatile components of “Kombu-dashi”, popular Japanese soup stock pre- 2. Flavor interaction among active components pared from dried kelp (Laminariaceae). The taste or in seafood flavor interaction between amino acids and other taste active compounds will also be shown. The next sec- L-Glutamic acid was found in 1908 as the key taste

Table 2. Synthetic extract of Kombu-dashi. Unpublished data.

Compound Conc. (mg/100 ml)

Glu◊Na (MSG) 37 Asp◊Na 16 Lactic acid◊Na 0.4 Formic acid 0.3 Malic acid 0.3 Citric acid 1.1 Succinic acid◊Na 0.2 Acetic acid 0.3 Pyroglutamic acid 0.9 Fig. 1. Concept of flavor. KCl 255 KI 1 2 NaCl 135

CaCl2 5

MgCl2 16

NaH2PO4 19 Mannitol 1200 Alginic acid◊Na 7

Prepared by mixing commercial chemical compounds to match the compositional profile as determined analytically. pH 6.3.

taste amino acid enhanced urchin-like taste and whole flavors. Taste active components in snow crab were investigated by using omission tests from synthetic extract made of commercial chemical compounds and Fig. 2. Dried kelp (kombu). Provided from UMAMI Infor- arginine was determined to enhance flavors under the mation Center, „ 2016. condition of coexisting with glutamic acid, inosinic acid, adenyric acid and glycine (Hayashi et al. 1981). Glycogen was found to constitute characteristic flavors constituent of “Kombu-dashi”, Japanese common soup of scallop (Watanabe et al. 1990). stock, prepared from dried kelp (Ikeda 1912). After industrialization of its salt, monosodium glutamate 2-1. Determination of non-volatile key flavor com- (MSG), this taste active crystal has been used for many ponents in Kombu-dashi cuisines as an inevitable seasoning worldwide. Many studies have been done on the flavor characteristics of As mentioned above, Kombu-dashi is used widely MSG. Especially the findings of synergistic effect be- for Japanese cuisines. This common soup stock itself tween glutamate and 5¢-ribonucleotides on umami taste has characteristic aroma and complex tastes. Although was a big event worthy of special mention (Kuninaka the key taste component was revealed to be monoso- 1964; Yamaguchi 1967). Glutamate does not only add dium glutamate, other constituents are necessary to umami to food but also enhances kokumi flavors such reconstruct whole Kombu-dashi characters. We inves- as continuity, mouthfulness, and thickness of many tigated the other flavor active non-volatile components foods (Yamaguchi and Kimizuka 1979; Yamaguchi in Kombu-dashi. There are some kinds of dried kelp in 1979, 1987). This phenomenon suggested that the Japan and it is believed that the quality of Kombu-dashi flavor interaction between glutamate and other com- depends on cultivation location or rating (size, shape, ponents would be developed in various foods. etc.) of kelp. Several studies about taste active components in At the beginning of our study we determined the savory seafood revealed that taste active compounds, authentic extraction condition for preparation of stand- such as amino acids, interact with other components ard Kombu-dashi. We investigated the preparation con- and strengthen the whole taste and character (Fuke and dition, focusing on temperature, resources of water, and Ueda 1996). Methionine was identified as a key flavor flavor characters, and we finally determined that the component of sea urchin (Komata 1964). This bitter best condition was the forty minutes extraction with

Table 3. Simplified synthetic extract of Kombu-dashi composed of taste-active components. Unpublished data.

Compounds Conc. (mg/100 ml) Mannitol 1200 MSG 3 7 KCl 255 NaCl 135

The solution was prepared by commercial chemical compounds. The key taste-active compounds were determined by a series of omission tests of synthetic extract of Kombu- dashi (Table 2) by four expert panelists. Each compound was omitted independently from the synthetic extract and the taste character of every omitted solution was compared to the synthetic solution. If the difference was recognized in the two solutions, we judged the eliminated component should Fig. 3. Generation of Kombu-dashi like kokumi flavors by be contribute for the whole tastes or flavors. mixing of mannitol, MSG, and KCl (molar ratio). Four expert panelists were employed for the judgements. Cited from Jpn. J. Taste Smell Res., 4(2), Ueda Y, The studies on the flavor characteristics “Koku” and “Atsumi” in some tasty distilled water at 50 C. Dried Kombu (30 g) was ex- ∞ foodstuffs, 197–200, Fig. 1, „ 1997, with permission from tracted with 1.0 L of distilled water under this condi- the Japanese Association for the Study of Taste and Smell. tion and filtered through a filter cloth. We employed omission tests to investigate taste-active components in accordance with the established methods (Konosu 1979; Hayashi et al. 1981; Watanabe et al. 1990). Synthetic reconstructed extract was prepared by mixing commercial chemical compounds through the analytical data of the standard extraction (Table 2). The content of Cl occupied 90% of total inorganic anion (Cl, PO4 and I). The synthetic extract had kokumi flavors equal to the original extract. After a series of omission tests, we finally determined the five compounds (mannitol, Glu, K+, Na+, and Cl–) as effective. Table 3 shows composition of simplified synthetic essence which exhibited Kombu-dashi like taste and kokumi flavors. In order to reconfirm the effec- tiveness of the synthetic essence, we further obtained the secondary extract from the Kombu residue after Fig. 4. A proposed flavor interaction among mannitol, MSG, preparation of the original Kombu-dashi and added the KCl, and NaCl. Changes of flavor impression were profiled four chemicals (mannitol, MSG, KCl, and NaCl) to the by four expert panelists. secondary extract according to the analytical data. The results of sensory examination showed that the standard and secondary extracts were not discriminated sta- model solution containing mannitol, MSG, and KCl tistically by the highly trained panelists (Triangle dis- by eliminating NaCl from the simplified synthetic ex- tinction test). We hence concluded that mannitol, Glu, + + – tract. Although whole flavors of the model solution K , Na , and Cl were key taste components of Kombu- were weaker than those of simplified synthetic extract, dashi. it still had Kombu-dashi like kokumi flavors. To clarify the flavor interactions among these three 2-2. Flavor interactions among mannitol, MSG, compounds by sensory evaluations, we prepared vari- and KCl ous kinds of the 40 solutions which had different com- positions of these three chemicals but the total weight It is well known that sodium and chloride ions are concentrations were adjusted to almost equal levels of taste active components in various kinds of seafood the simplified synthetic extract. The sensory tests (see Table 1). It can be said that three other compounds, + + showed that Kombu-dashi like kokumi flavors should mannitol, Glu , and K , should generate characteristic be active under the restricted conditions (Fig. 3). Pro- Kombu-dashi like kokumi flavors. We prepared the posed formula of the condition is shown as follows:

0.67

0.11

0.44

0.33

*1.22

*1.44 -

*1.33

*1.22

**1.67

**1.78

**1.44

**1.67 **1.78

Total taste

0.22

*1.33

**1.67 **1.89

**1.78 **1.89 **1.89

**1.78 *1.44

**1.56 *1.33

**1.33

**1.44 *1.44

**1.44

0.11

0.22

0.67 0.44

0.00 0.44 *1.00 1.00

0.56 0.44

0.33 0.33

0.44

0.44 0.56

0.89

0.67 0.89 0.56

0.44 0.44

Jpn.

2001,

„

Other taste

< 0.05.

0.22

0.11 0.33

0.44

0.22

0.44

0.11 0.44

able,

< 0.01; *

0.22

0.56

0.44 0.44 0.44 0.56

Umami

0.22 0.00 0.11

0.22 0.44 0.00 0.33 1.00 0.78

*1.22 *1.00

**1.33

0.11

0.22

0.56 0.33 0.11 0.00

0.67 0.33 0.11 0.00

0.56 0.44 0.22 0.11 0.33 0.56

0.67 0.00 0.00 0.44 0.44 1.11

0.22 0.44 0.00 0.78 0.33 0.22 0.00 0.67 1.11 0.00 0.33 0.11 0.44

0.00 0.56 0.00

0.78 0.56 0.22 0.00 0.33 0.67 0.56 0.33 0.00 0.00

= 9). –, decreased. **

*0.78 *0.89 0.22

*0.67 0.22 0.11 0.22

*0.78 0.22 0.11 0.22

*0.89 0.56 0.11 0.22 0.33 0.33 0.67

*0.78 0.56 0.11 0.00 0.44

*0.89 **0.78 -

*1.11 *0.78 0.11 0.00 0.44 1.33

**1.22

Bitterness

1.33

1.67

1.56

1.00 1.44 1.33

1.33

1.44

1.67

1.67 1.56

0.33

0.67

0.33

0.44

0.67

0.11

0.22

0.00

0.00 *0.67 0.56 0.33

0.11 0.11 0.22 0.00 0.11 0.33

0.00 0.44 0.11 0.22 0.00 0.00 0.33

0.00 0.56 0.33

- - -

- -

** ** **

aste and Smell.

1.00

1.33

1.56

1.22

0.11

0.22

0.11

0.33 *

0.22 **

0.67

0.11 **

0.67

0.11 **

0.00

0.00 0.00 0.00

0.00

Sourness

1.89 **

1.44 **

1.33 **

0.56 0.44

0.11

0.11 0.89 **

0.33

0.33 **

0.56 *

0.11

0.11 * 0.67 **

0.00

0.00 0.00

- -

., The interactions between various amino acids and Na/KCl, pp. 491–494, .,

et al

0.56 0.00 0.00

0.56

0.00

0.11 0.00 0.00 0.56 0.00 0.22

0.00

0.44 0.00 0.22

0.56

0.44 0.00 0.00 0.67 0.00

*0.89 0.11

Saltiness

ell-trained panel evaluated by labeled magnitude scale (

The Japan Association for Study of The Japan

(3), Kawai

*0.78

**1.67

**1.89 *1.00

**1.67 *0.78

**1.89 **1.56 **1.44 **1.11 **1.22 **1.22 *0.78 **1.33 *0.78

**1.22

**1.67 *0.67 **1.89 **1.89 **1.22 *0.78

**1.67

**1.22 **1.00

**1.67 **1.89

**1.11 **1.44

0.11

0.11 0.11

0.33

0.78

0.22

0.11

- -

. Summary of the changes in taste quality of amino acid solutions by the addition of NaCl or KCl. Modified from

able 4

Sweetness

J. Taste Smell Res J. Taste T

with permission from

0.11

0.22

0.67

0.22 0.11

0.33 0.33

0.44 0.11 0.00 0.11

0.33 0.22 0.11 0.22

0.22 0.11

0.00 0.33

0.11

0.67 0.22 0.11 0.44

NaCl KCl NaCl KCl NaCl KCl NaCl KCl NaCl KCl NaCl KCl NaCl KCl

*0.67 0.00 **1.44

**1.11

**0.78

**1.33 0.89

5 1

20 40 50

50 10

250 *1.11 0.78 **1.11 *0.67

200 0.22 100 *1.00 0.44 250 0.89 0.56

250 *1.22 0.22 *1.00 0.00 0.00

250 0.33 0.67 0.78 *0.89

250

250 *0.78 0.56 **1.22 *0.78

(mM)

HCl

◊

-Ala 250

Phe

Glu

Gly

His Glu Ilo

Amino acid Conc.

His Asp Hyp Ala Asp Lys D

Leu Met

Asn

Pro

pGlu 3 Gln

Arg 10 Arg Ser

Thr

Val Trp

Addition conc., NaCl (50 mM), KCl (50 mM).

[Condition under which three compounds generated acids. The mineral did not affect sweetness of all the “kokumi flavors” (by molar ratio)] amino acids tested, but it enhanced umami of MSG, 1) [KCl]/[MSG] = 2.5–20 suggesting that KCl would be a useful ingredient un- 2) [mannitol]/([mannitol]+[KCl]+[MSG]) = 0.25–0.75. der the combination with MSG.

In Japan dried Kombu products were generally 2-4. Taste interactions between amino acids and ranked to low or high grade classes according to their IMP flavors. We analyzed contents of mannitol, Glu, and K in several high grade kombu samples and demonstrated Synergistic effect between MSG and 5¢-inosine that the high grade Kombu-dashi contained three key monophosphate (IMP) on umami taste is familiar. Dried flavor components well in accordance with the results bonito (katsuobushi) contains high amount of IMP and in Fig. 3. Japanese people traditionally use the soup stock pre- Figure 4 shows a conceptual diagram of the flavor pared in combination of katsuobushi and kombu. interactions of four compounds. KCl itself had bitter, Nowadays MSG and IMP are industrially produced and astringent, and salty tastes. Bitter taste of KCl were used for various kinds of foods in the world. We trust decreased by MSG. Kombu-dashi like kokumi flavors in the fact that people acquire flavor rich and savory generated under the condition mannitol, KCl, and MSG processed foods manufactured by using the synergistic coexisted. Mannitol itself had a sweet taste, but this effect. On the other hand, many savory foods contain sugar alcohol generated characteristic kokumi flavors amino acids other than glutamic acid and then we in- in the coexistence of MSG and KCl. It was the first vestigated the flavor interaction of IMP with other description about flavor characteristics of mannitol, amino acids. suggesting that this sugar alcohol would be a useful Low concentration of IMP (0.5 mM, approximate its material for quality improvement of processed foods. threshold) affected the tastes of many amino acids ex- Subsequently we compared the flavor characteristics cept for some bitter taste amino acids (Table 5). Sev- of three kinds of sweet sugar alcohols (sorbitol, eral sweet amino acids (Gly, L-Ser, and L-Ala) and the maltitol, and xylitol) with that of mannitol. The three other amino acids showed significant umami enhance- sugar alcohols had the flavor character similar with ment. It is known that only acidic amino acids have mannitol, suggesting that they will be applicable as been reported to enhance umami by addition of purine food additives to enhance kokumi flavors of processed 5¢-monophosphates (Yamaguchi et al. 1971; Furukawa foods. 1991), but our study showed that the addition of IMP affected the tastes of various amino acids. 2-3. Taste interactions between amino acids and To examine the effects of IMP to three sweet amino NaCl/KCl acids (Gly, L-Ser, and L-Ala) in detail, another experi- ment was carried out, using sweetness inhibitor ±2-(p- Sodium and several amino acids had been known as methoxyphenoxy) propanoic acid (PMP) and 80 ses- key flavor components of savory seafood as shown in sions of triangle distinction tests by 4 expert panel Table 1. Although several studies on the taste interac- members (Kawai et al. 2002). The result suggested that tion of NaCl with other compounds were performed, the enhanced taste characters of the three amino acids the characteristics of KCl had not been studied much. were not sweetness but umami. We then studied the taste interactions between these In order to reveal whether these enhancement effects minerals and broad amino acids by sensory tests using were synergistic or not, addition effects of IMP to four a highly trained panel. amino acids (Gly, L-Ser, L-Ala, and D-Ala) were exam- NaCl strengthened sweetness of many amino acids ined by modulus-free magnitude estimation by 8 ex- (D-Ala, Gly, His, Hyp, Met, Pro, Ser, and Thr) and pert panel members. Potentiation ratio (PR) values were strengthened umami of some amino acids (Asp·Na, employed to judge on the synergistic effect between Glu, Glu·Na, Gln, and Thr) as shown in Table 4. These each amino acid and IMP (Fig. 5). The results showed results supported the previous studies by omission tests that PR values of Gly, L-Ser, and L-Ala were greater (in Table 1). NaCl also diminished sourness and/or bit- than 1, suggesting that the amino acids exhibited terness of many amino acids. In addition, this study umami taste enhancement synergistically with IMP, also confirmed the report by Ugawa et al. (1992) that though D-Ala did not exhibit the effect. NaCl added sweetness to the solution of Gly, Ala, and This was the first report that these three sweet amino Ser. The addition effect of NaCl to amino acids with acids had umami taste. We speculated that these amino regards to sweetness, umami, sourness, and bitterness acids would bind to both sweet and umami receptors, is supposed to be one of the reasons why NaCl is used and IMP would strengthen the affinities between these as an important seasoning ingredient. amino acids and the umami receptor. Just at that time, KCl, a salty and bitter mineral, had not exhibited the umami receptor (T1R1/T1R3) was found to respond remarkable addition effects to the tastes of broad amino to L-Glu synergistically with IMP, and respond to L-

0.386

0.084

0.715 0.708 0.053

0.021 0.603

0.149

0.603

0.855

0.67

0.56

0.22

0.22 0.78

1.00* 0.22

0.78

0.22

0.11

- - -

- -

0.317 0.89* 0.046

0.564 1.22* 0.029 0.655 0.44 0.420

0.655

0.334 1.000

< 0.01, statistically significant

Chem.

ther taste Total taste

0.22

0.22 0.11

0.11

0.22 0.11

0.11 0.317 0.67 0.196

- -

< 0.05; **

2002, with permis-

„

. Modified from

0.655 0.11 0.317

0.317 0.22 0.317 0.11 0.715

0.11 0.11

, 739–745, Table 1, Table , 739–745,

0.380 0.78* 0.038 0.89* 0.038 0.89 0.107

0.107 0.11 0.655 0.00 0.317 1.000 1.44** 0.007

0.317 1.89** 0.004 0.655 0.11 0.577 0.00 0.317 0.44 0.248 0.317 1.56* 0.020 0.257

0.729 1.00* 0.024 0.22 0.131 0.44 0.357 0.890 0.00 0.317 0.053 0.22 0.317 0.00 1.000

0.584 0.00 0.317 0.00 0.317 0.11 0.720

0.317 0.00 1.000 0.22 0.317

0.396 2.00* 0.011 0.00 1.000 1.22 0.080 0.414 0.00 1.000

1.000 1.56** 0.006 0.00 0.157 1.22* 0.036

0.465 0.00 1.000 0.22 0.317

0.655 0.17 0.317 0.00 1.000 0.11 0.564

ve, average. Wilc, Wilcoxon. * Wilcoxon. Wilc, ve, average.

Bitterness Umami O

0.33

0.67 0.22 0.11

0.11 0.22 0.78 0.33 0.22

0.22

0.11

0.22 0.44

0.11

0.44

0.06

- - - -

- -

0.317 0.00 0.317 1.33** 0.006 0.00 0.317 1.44** 0.006 0.206 0.408 0.11 0.408 0.33 0.102 0.00 0.317 1.00* 0.013

0.888 0.020 0.053 0.386 0.11 0.655 0.89* 0.038 0.22 0.157

0.059 1.000 0.086

0.603

0.050 0.11 1.000 2.00** 0.003 0.00 0.655 1.78** 0.006 0.025

0.009 0.11 0.317 1.78** 0.005 0.083 0.22 0.748 0.11 1.000 0.00 1.000 0.56 0.351

0.014

0.020

0.157

0.180 0.11 0.773 0.00 1.000 0.11 0.317

0.046

Sourness

1.26

0.33 0.56 0.33

0.56

1.00

0.11

1.56* 0.78

0.56

0.22 0.22

1.11 1.11*

1.44**

1.11*

1.22*

0.22

0.33

0.22*

- -

- - -

- - - -

0.317

were presented randomly to 10 subjects.

Saltiness

0.11

., Taste enhancement between various amino acids and IMP Taste .,

et al

0.317 0.00 1.000

0.414 0.00 1.000 0.00 1.000 0.22 0.566 1.000 0.33* 0.046 0.33 0.102

0.180 0.317 0.44 0.102

0.201 0.00 0.564

, Kawai

. Summary of the changes in taste quality of amino acid solutions by the addition of IMP

Sweetness

Ave. Wilc. Ave. Wilc. Ave Wilc. Ave. Wilc. Ave. Wilc. Ave. Wilc. Ave. Wilc.

0.11

0.22 0.11

0.11

0.44

0.44 0.157 0.11 0.655

0.00 1.000 0.78 0.059

0.11 0.317 0.11 0.317

0.22 0.317

0.28* 0.025 0.06 0.317

- -

able 5

Senses sion from Oxford University press.

ilcoxon’s signed rank test. w/o, without. ilcoxon’s

3.0

75.0 1.00* 0.014 0.00 1.000

25.0 0.00 1.000 0.00 1.000 0.00 1.000

60.6 0.22 0.317 0.00 1.000

50.0 1.00* 0.024 0.00 1.000 0.00 1.000 0.11 0.773 0.00 1.000

40.0 40.0 0.11 0.317 0.00 1.000 0.22 0.317 20.0

50.0 0.33 50.0

10.0 1.11* 0.013 0.00 0.317 0.00 0.317

50.0 0.44 0.157 0.00 1.000 10.0 0.22 0.157 0.00 1.000

(mM)

Conc.

250.0 1.33* 0.010 0.22 0.180

250.0 1.33** 0.009 0.00250.0 0.317 0.11 0.633 0.11 0.180

100.0 1.22* 0.018 0.00 0.317

250.0 0.44 0.417 0.00 0.317

250.0 250.0 0.67 0.272 0.44 0.276 250.0 0.56 0.270 0.00 0.317

-Pro

-Glu 3.6

HCL

-pyro Glu 3.0

Leu

-A

-Lys

-Ilo

-Cys

-A

-Phe

-H

-M -Asn -P

-G

-Ala

-Arg

-Thr -Val -Trp

-Ser

-Glu-

-Glu 1.0

Amino acid

L L L 4-OH- D L Gl L Gly-Gly

L L L L L L Succinic acid 2.0 DL L L L

L L L

w/o amino acid

differences based on differences

Pairs of stimulus solutions with and without 0.5 mM IMP

Fig. 5. The potentiation ratio (PR) of amino acid-IMP solutions derived from mean taste intensities. The PR was calculated by dividing the taste intensity of the mixture by the sum of taste intensities of the individual components in the mixture. The larger PRs at low stimulus concentrations might have been due to quasi-zero ratings for both amino acids and IMP unmixed solutions. Modified from Chem. Senses, 27, Kawai et al., Taste enhancement between various amino acids and IMP, 739–745, Fig. 2, „ 2002, with permission from Oxford University Press.

Ser and L-Ala (Li et al. 2002; Nelson et al. 2002). The studies supported our speculation. As generally accepted, glutamic acid and other amino acids, which enhance umami taste perception, take a role of indica- tor for humankind to recognize protein and contrib- utes to human nutritional intake.

3. Flavor characteristics of glutathione in scallop and other foodstuff

Glutathione (g-L-glutamyl-L-cysteinylglycine, GSH) is a well known tripeptide and exists in a wide range of organisms and foodstuffs. The representative physi- ological activity of the peptide is detoxification of Fig. 6. Scallop, garlic, and onion. methylglyoxal, nitrobenzene derivatives, etc., owing to its reductive action. Although many studies on the biological functions of the peptide have been done, little was known about its flavor characteristics. After a have several specific sulfur-containing compounds. S- series of studies on the flavor components in garlic and alkyl-L-cysteine sulfoxides are known as the popular onion, it was revealed that GSH has a so-called kokumi precursors of the specific odors through enzyme reac- flavor expressed by the terms of continuity, tions. Once we cut these plant tissues, pungent and hot mouthfulness and thickness. At the beginning of the taste other than the odors are generated by results of review about GSH, the studies on flavor components specific enzymes such as alliinase. of the Allium plants, which were the starting point of Cooked garlic is used for various kinds of meals, like the study of GSH taste characteristic, will be discussed. soup, fried meat, grilled meat etc., all over the world. The characteristic additional flavor effects on the food 3-1. Flavor characteristics of sulfur-containing can be expressed by terms such as continuity, components in garlic and onion mouthfulness, and thickness (kokumi flavors). Table 6 shows the addition effects of water extract of garlic Garlic (Allium sativum) and onion (Allium cepa) be- on the flavor profiles of soups. Key effective compo- long to the Allium genus. There are many studies about nents were found out through both traditional column the components in the plants from the perspective of chromatography and sensory tests. A cation exchange their bioactivities or characteristic odor. The plants resin absorbed fraction, contained sulfur-containing

Table 6. Addition effects of garlic extract on flavor profiles Table 7. Kokumi flavors of each sulfur-containing compo- of soups. Modified from Biosci. Biotech. Biochem., 54(1), nent. Cited from Biosci. Biotech. Biochem., 54(1), Ueda et Ueda et al., Characteristic flavor constituents in water ex- al., Characteristic flavor constituents in water extract of tract of garlic, 163–169, Table I, „ 1990, with permission garlic, 163–169, Table IV, „ 1990, with permission from from Japan Society for Bioscience, Biotechnology, and Japan Society for Bioscience, Biotechnology, and Agrochemistry. Agrochemistry.

Chinese soup Curry soup Component* Kokumi flavors** Other flavor AROMA Alliin +++ Garlic-like Whole Aroma ææ Cycloaliin + MeCSO ++ Leak-like BASIC TASTE GAC ++ Garlic-like GACSO + Garlic-like Saltiness ææ Sweetness Glutathione +++ ææ Cys + Sourness ææ Met + Bitterness ææ Umami ææUmami solution contained 0.05% MSG and 0.05% IMP. *0.2% (w/v) of each compound was added to the umami FLAVOR CHARACTER solution. **Continuity, mouthfulness, and thickness in the Continuity (0.84)* (0.84)** solution. +++, strongly recognized; ++, apparently recog- Mouthfulness (0.90)* (0.78)* nized; +, recognized. Thickness (0.84)* (0.74)*

The addition effects of garlic extract were evaluated by a et al et al five-point rating scale (–2 ~ +2). n = 19. *,**Stronger than ods we established (Ueda . 1990; Kuroda . control significantly. *p < 0.01 and **p < 0.001 versus con- 1997). The average content of alliin in 13 garlic sam- trol (Student’s t-test). —, not significant. ( ), Average score. ples was 2515 mg (/100 g dry weight) and the content was as much as that of arginine (2184 mg), the most principal component among amino acids, while average GSH content was small (19 mg/100 g dry weight) amino acids and peptides, exhibited characteristic (Ueda et al. 1991). In the case of onion, PeCSO (47.3– kokumi flavors in soups. This fraction also generated 82.3 mg/100 g dry weight) and g-Glu-PeCSO (50.3– kokumi flavors when it was added to the umami solu- 147.3 mg/dry weight) were major sulfur-containing tion prepared from MSG and IMP (Ueda et al. 1990). amino acids, while GSH contents were relatively small This fraction contained S-allyl-L-cysteine sulfoxide (al- (0.4–1.8 mg/100 g dry weight) (Ueda et al. 1994). liin), S-methyl-L-cysteine sulfoxide (MeCSO), g-L- These results suggested that these specific sulfur-con- glutamyl-S-allyl-L-cysteine (GAC), g-L-glutamyl-S-al- taining amino acids and peptides would be kokumi lyl-L-cysteine sulfoxide (GACSO), and 3-(S)-methyl- flavor constituents in garlic and onion. The contribu- 1,4-thiazane-5-(R)-carboxylic acid (cycloalliin), iden- tion of GSH, common sulfur-containing peptide, was tified by using 1H-NMR, 13C-NMR, and FD-MS (Fig. seemed not so large for the whole flavors of the foods. 7). As shown in Table 7, we determined some sulfur- Garlic and onion were often used for cuisines by heat containing amino acids and peptides as key kokumi cooking methods, such as fry, roast, and boil. We were constituents of garlic. GSH, one of sulfur-containing interested in thermal stability and degradation of al- components in garlic, also exhibited strong kokumi liin, PeCSO, and g-Glu-PeCSO in water. Alliin was flavors as alliin did. very stable in boiled water. PeCSO and g-Glu-PeCSO Onion is also used for cuisine all over the world. We were not so stable when they were heated for long time. successfully isolated S-propenyl-L-cysteine sulfoxide PeCSO gradually changed to cycloalliin during heat (PeCSO) and g-L-glutamyl-S-propenyl-L-cysteine treatment. g-Glu-PeCSO degraded to 5-oxo-2- sulfoxide (g-Glu-PeCSO), representative sulfur-con- pyrolidinecarboxylic acid (PCA) and PeCSO, and the taining compounds in onion, by the same purification resulting PeCSO was further converted to cycloalliin methods employed in the case of garlic. These isolated (Ueda et al. 1994). sulfur-containing compounds exhibited kokumi flavor in the umami solution (Ueda et al. 1994). 3-2. Flavor characteristics of glutathione in We determined contents of these sulfur-containing umami solution compounds in garlic and onion. Alliin, MeCSO, PeCSO, and g-Glu-PeCSO were analyzed by using an As stated above GSH, known as the most popular amino acid analyzer. GSH and cycloalliin were tripeptide in seafood and other foods, also exhibited analyzed by these moments, using specific HPLC meth- kokumi flavors in umami solution. GSH itself had sour-

Fig. 7. Major sulfur-containing components found in garlic. MeCSO, S-methyl-L-cysteine sulfoxide; GAC, g-L-glutamyl-S- allyl-L-cysteine; GACSO, g-L-glutamyl-S-allyl-L-cysteine sulfoxide.

ness in a water solution because of its acidic nature. Table 8. Threshold values of glutathione in water and umami But in neutral pH condition GSH did not have any taste solutions. Modified from Biosci. Biotech. Biochem., 61(12), in water. We then examined effects of GSH on four Ueda et al., Flavor characteristics of glutathione in raw and basic taste perceptions (sweetness, saltiness, sourness, cooked foodstuffs, 1977–1980, Table II, „ 1997, with per- and umami) by addition tests. The results of sensory mission from Japan Society for Bioscience, Biotechnology, and Agrochemistry. examination showed that GSH did not affect the intensities of each basic taste in the tested solutions Solution Threshold value (w/v) (Ueda et al. 1997). Although the threshold value of GSH in water was Water 0.04 40 mg/100 ml in the sensory evaluation, the value de- MSG 0.05% 0.04 creased to 10 mg/100 ml in the umami solution con- MSG 0.80% 0.02 MSG 3.1% 0.01 taining MSG and IMP (Table 8), suggesting that GSH MSG 0.05 + IMP 0.05% 0.01 had an interaction with umami substances on its threshold. The results showed that kokumi flavors of these The test solutions were adjusted to pH 7.0. Triangle distinc- dishes were strengthened by addition of 10 mg/100 g tion tests were used to measure the threshold values (n = 20, GSH in various kinds of cooked foods, such as Chi- p < 0.001). nese soup, hamburger steak, shaomai, etc., suggesting that the peptide would be a new promising food ingredient for processed foods. GSH at a high level concentration that exceeds its 3-3. Additional effects of glutathione in the syn- threshold value, proposing that GSH would be one of thetic beef and scallop extracts the flavor constituents in scallop. GSH was detected in beef meats at relatively high There was not much knowledge about contents of concentration. The synthetic beef extract was prepared GSH in foodstuffs, and then we determined GSH con- by 36 commercial chemical compounds (amino acids, tents in a broad range of foodstuffs. GSH was detected sugars, nucleotides, carbonic acids, salts, and others) in almost all the tested foods, 13–40 mg/100 g in beef according to analytical data of beef (Ueda et al. 1997). meats, 3–38 mg/100 g in wines and high levels of 9.6– This solution had rather beef like taste though it did 25 mg/100 g in scallop (Ueda et al. 1997). Table 9 not have any aroma. Effects of the addition of 20 mg/ showed the results of more specific analyses in fishes 100 ml GSH to the synthetic extract were investigated and other seafoods. GSH was detected at a lower level by sensory tests. As shown in Table 10, kokumi flavors in fishes, but adductor muscle of scallop contained were significantly enhanced in the presence of GSH,

Table 9. Reduced glutathione (GSH) and oxidized glutathione (GSSG) contents in seafood. Cited from Nippon Suisan Gakkaishi, 64(4), Ueda et al., Contents of glutathione in seafoods and its flavor characteristics, 710–714, Table 1, „ 1998, with permission from The Japanese Society of Fisheries Science.

Samples GSH (mg/100 g) GSSG* (mg/100 g) Fishes Bigeye tuna Thunnus obesus 0.81 3.75 Bluefin tuna Thunnus thynnus 5.46 nd Skipjack tuna Katsuwonus pelamis 2.58 4.13 Chub mackerel Scomber japonicus 0.29 nd Horse mackerel Trachurus japoncus 2.48 nd Sardine Sardinops melanosticus 6.51 1.06 Yellow tail Seriola quinqueradiata 0.17 nd Coho salmon Onchorhynchus kisutch 9.22 0.38 Flatfish Paralichthys olivaceus 2.57 nd

Other seafoods Northern green sea urchin Strongylocentrotus intermedius 7.36 3.24 Black tiger prawn Penaeus mondon 0.19 nd Northern shrimp Pandalus borealis 4.45 nd Neon flying squid Ommastrephes batrami nd nd Japanese flying squid Todarodes pacificus nd nd Short-neck clam Ruditapes philippinarum nd nd Scallop Pactinopecten yessoensis 29.05 nd

*Concentration of GSSG were determined by the differences between the total amount of GSH + GSSG and the amount of GSH. The total amount of GSH were determined after reduction with GSH reductase. nd, Not detected.

Fig. 8. Effects of GSH on the flavor of synthetic extract of scallop. Added GSH concentration was 29 mg/100 ml. The flavor of GSH-added synthetic scallop extract was evaluated by a five-point rating scale (–2 ~ +2, very weak~very strong), compared with that of GSH-free extract (n = 28). Each bar diagram includes a line indicating the 95% confidence level. Cited from Nippon Suisan Gakkaishi, 64(4), Ueda et al., Contents of glutathione in seafoods and its flavor characteristics, 710–714, Fig. 1, „ 1998, with permission from The Japanese Society of Fisheries Science. suggesting that GSH would contribute to generation fied synthetic extract containing (mg/100 ml) Gly 1925, of kokumi flavors of beef. It was noteworthy that the Ala 256, Arg 323, Glu·Na·H2O 179, AMP·Na2 195, meet-like flavor of the extract was also strengthened KOH 232, KCl 109, NaCl 71. The pH was adjusted to together with kokumi flavors. 6.1 by HCl. GSH was added to the solution at the con- Konosu et al. (1988) and Watanabe et al. (1990) al- centration of 29 mg/100 ml. GSH significantly en- ready revealed some key flavor components of scal- hanced kokumi flavors as well as sweetness and umami lop, Gly, Ala, Arg, Glu, AMP, K+, Na+, and Cl–, by a (Fig. 8), suggesting that GSH would be one of the flavor series of omission tests. We then employed the simpli- constituents in scallop.

Fig. 9. Degradation curve for GSH in aqueous solution (98∞C). ᮀ, GSH; ᭿, GSSG; ᭡, PCA. Modified from Biosci. Biotech. Biochem., 61(12), Ueda et al., Flavor characteristics of glutathione in raw and cooked foodstuffs, 1977–1980, Figure, „ 1997, with permission from Japan Society for Bioscience, Biotechnology, and Agrochemistry.

Table 10. Addition effects of glutathione on flavor profiles mined the degradation products of GSH. GSH was of umami solution and model beef extract. Modified from heated in hot water (98∞C, 5 hours) and three com- Biosci. Biotech. Biochem., 61(12), Ueda et al., Flavor char- pounds were isolated from the mixture of the degrada- acteristics of glutathione in raw and cooked foodstuffs, 1977– tion products. The main products were determined as 1980, Table IV, „ 1997, with permission from Japan Soci- oxidized glutathione (GSSG), PCA, and ety for Bioscience, Biotechnology, and Agrochemistry. cyclocysteinylglycine (cyclo Cys-Gly) disulfide. PCA and cyclo Cys-Gly were derived from glutamyl resi- Umami solution Model beef extract due and cysteinylglycine residue, respectively. Ap- Whole aroma ææproximately 90% of GSH remained under 30 min heat- a Basic tastes ææing (pH 5, 98∞C), though 80% of the peptide degraded Continuity 0.80b* 0.95** after heating for 5 hours (Fig. 9), suggesting that GSH Mouthfulness 0.65* 0.85* could be applied as useful materials for many proc- Thickness 0.65* æ essed food products, except for the retort sterilization Meat-like flavor æ 0.70** foods. In addition, it was the new finding that GSH changed to two kinds of cyclic compounds in water Umami solution contained 0.05% MSG and 0.05% IMP. The (Fig. 10). model beef extract was prepared by 36 commercial chemi- On the other hand GSH is known as a precursor of cal compounds (amino acids, sugars, nucleotides, carbonic reaction flavors when it was heated with reducing sugar acids, salts, and others). GSH concentration was 0.02% w/v. Well-trained 20 panel members evaluated on a five-point rat- (Zhang et al. 1988; Zhang and Ho 1989, 1991a, 1991b). ing scale (–2 ~ +2). *,**Significantly stronger than control. We assured the generation of MFT (2-methyl-3- *p < 0.01 and **p < 0.001 versus control (t-test). —, not furanthiol) and FFT (2-furanmethanethiol), known as significant. a saltiness, sweetness, bitterness, sourness, and the key flavor compounds of roasted meat or coffee, umami. b Average score. after the reaction of GSH and xylose. GSH could be the multiple functional food ingredient as a kokumi flavor and a base of reaction flavor.

3-4. Thermal stability of glutathione 4. Generation and decreasing mechanism of key volatile flavor compounds in dried bonito We investigated the heat stability of GSH in water to reveal the possibility of the future application of this Dried bonito (katsuobushi) has been used for Japa- peptide for food processing industry. At first we deter- nese cuisine for a long time as the savory food mate-

Fig. 10. A proposed degradation mechanism of glutathione in water.

rial of soup stock (dashi). It is well-known that dried bonito contains a large amount of IMP and enhances savory flavors of many kinds of dishes, and also its savory odor makes various foods preferable. Although some Japanese scientists investigated key volatile components (Sakakibara et al. 1990), the complete disso- lution of key compounds and their changes during cooking had remained big issues for the prevention of deterioration of savory odors. Ishiguro et al. (2001) studied the changes in volatile compounds during smoking process and evaluation of Fig. 11. Dried bonito (katsuobushi). Provided from UMAMI odor-effective constituents of dried bonito. They ex- Information Center, „ 2016. amined the changes of major volatile compounds in the production process (from boiled meats to katsuobushi) by using gas chromatography-mass spectrometry (GC-MS) and gas chromatography-olfac- 4-1. The generation of pyrazines and changes in tometry (GC-O). The study suggested that some the amounts of volatile compounds in smoke ethyldimethylpyrazines and phenols (guaiacol and 4- during preparation of dried bonito methylguaiacol) contributed to the developments of roasty and smokey aromas, respectively, and the sulfur- The generation mechanisms of pyrazines and containing compounds were also essential for whole phenols, which had been already confirmed to be the aroma characters of dried bonito. key flavor components of dried bonito, were examined Taking these results into account we sequentially by GC-MS analysis of various volatile compounds. At investigated generation and decrease mechanism of first the changes in the amounts of pyrazines was mea- some key volatile compounds, such as pyrazine, sured in the model dried bonito prepared by heating of phenols, and sulfur-compounds using GC-O methods boiled bonito with and without smoking. The amounts established by Ishiguro et al. (2001). Although Aroma of 2,5-dimethylpyrazine, 2,6-dimethylpyrazine, 2- Extract Dilution Analysis (AEDA) method had been ethyl-6-methyl pyrazine, 2-ethyl-5-methyl pyrazine, employed as a popular method for determination of and trimethylpyrazine in the dried bonito with smoke effective aromatic components in many foods (Mistry were larger than those in the dried bonito without et al. 1997), there still remained the possibility to over- smoke. Both 3-ethyl-2,5-dimethylpyrazine and 2-ethyl- look unstable or low boiling point components. The 3,5-dimethylpyrazine were not detected in the model method of Ishiguro et al. seemed effective to examine dried bonito (Kawaguchi et al. 2001). We prepared odor active components in savory foods because we crushed dried bonito (crushed by a rotary cooking cut- could analyzed the amounts of volatile components of ter) as model powder and reheated it with and without the comparable samples which had actual olfactory smoke. Figure 13 shows the changes in the amounts odors (Fig. 12). of pyrazines in the model powders. Almost all the

Fig. 12. Analytical method of volatile compounds in dried bonito.

Fig. 13. Changes in the amounts of pyrazines in dried bonito prepared with and without smoking by heating the powder of dried bonito at 70∞C for 4 hours. Modified from Nippon Shokuhin Kagaku Kogaku Kaishi, 48(12), Kawaguchi et al., Study on the generation of pyrazines and change in amounts of volatile compounds in smoke during preparation of dried bonito, 899– 905, Fig. 2, „ 2001, with permission from The Japanese Society for Food Science and Technology. pyrazines were detected in the model powders and their generated during heat treatment of the boiled bonito amounts increased by heating with smoke. 3-ethyl-2,5- meat without smoking, but their amounts and varieties dimethylpyrazine and 2-ethyl-3,5-dimethylpyrazine were less than those in the ordinary dried bonito. This were not detected in the model powder reheated with- suggested that some unknown substances in smoke out smoke. All the studies showed that pyrazines were could be related to the generation of pyrazines.

There had been many studies on the generation mechanism of pyrazines, focusing on Strecker degradation between a-dicarbonyl compounds and amino acids (Hwang and Hartman 1994; Cerny and Grosch 1994; Huang et al. 1996). In addition it was known that pyrazines were generated by the reaction between a-ketoalcohol and ammonia (Rizzi 1988). We presumed the generation mechanisms of pyrazines in the manufacture process of dried bonito to be as shown in Fig. 14. We added acetol and 2,3-pentanedione to the model powder and examined the addition effects on the changes in the amounts of pyrazines. As shown in Fig. 15, the amounts and varieties of pyrazines increased markedly in the heated model powder. From Fig. 14. A presumed mechanism of pyrazine generation in the result we concluded that acetol and 2,3-pentandione dried bonito. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 48(12), Kawaguchi et al., Study on the generation reacted with amino acids or protein in boiled meat, and of pyrazines and change in amounts of volatile compounds generated pyrazines. in smoke during preparation of dried bonito, 899–905, Fig. Phenols were not detected in the dried bonito pre- 3, „ 2001, with permission from The Japanese Society for pared by heating without smoking, proposing that Food Science and Technology. phenols in dried bonito would be mainly derived from

Fig. 15. Comparison of the amounts of pyrazines by heating the powder of dried bonito at 70∞C for 4 hours with and without acetol and 2,3-pentanedione. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 48(12), Kawaguchi et al., Study on the generation of pyrazines and change in amounts of volatile compounds in smoke during preparation of dried bonito, 899–905, Fig. 4, „ 2001, with permission from the Japanese Society for Food Science and Technology.

Fig. 16. GC chromatograms of volatile compounds in crushed dried bonito kept under different conditions. Volatile compounds in the stored samples were absorbed on Tenax-TA Polymer and applied for analysis. 1, propanal; 2, acetone; 3, butanal; 4, 2-butanone; 5, 2-methylbutanal; 6, 3-methylbutanal; 7, 2-ethylfuran; 8, pentanal; 9, 2,3-pentanedione; 10, hexanal; 11, 1- penten-3-ol. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 49(5), Kawaguchi et al., Changes in the volatile compounds related to deterioration of crushed dried bonito during storage, 312–319, Fig. 1, „ 2002, with permission from The Japanese Society for Food Science and Technology.

smoke. The incubation of the mixture prepared from of fresh crushed bonito odor had been a big issue for crushed dried bonito together with cellulose powder, processed food industry, the changes of the key vola- which contained substances in smoke, resulted in de- tile compounds during storage were not studied satis- creasing the amounts of furfural, 2-cyclopenten-1-one fyingly. and 5-methylfurfural. Possibly, these compounds were We preserved crushed dried bonito samples under converted to non-volatile compounds in the produc- three conditions (–20∞ for 24 days, 25∞C for 12 days, tion process of dried bonito (Kawaguchi et al. 2001). and 70∞C for 4 hrs) and examined their volatile compounds. The sample kept at –20∞C had strong acidic 4-2. Changes in the volatile compounds related odor and the one kept at 70∞C had weak roast odor to deterioration of crushed dried bonito dur- together with weak acidic odor. The acidic odor of the ing storage sample kept at 25∞C was less than that of the sample kept at –20∞C. Figure 16 showed GC chromatogram So far it had been known that savory odor of dried of volatile compounds. The amounts of carbonyl com- bonito was stable before crushed. The odors dramati- pounds increased during storage in all samples but the cally change after crushing and savory odor is lost af- variation of carbonyl compounds depended on the con- ter long preservation. Though repress of deterioration ditions (Table 11). The amounts of straight-chain al-

Table 11. Volatile compounds increased in crushed dried bonito during storage under different conditions. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 49(5), Kawaguchi et al., Changes in the volatile compounds related to deterioration of crushed dried bonito during storage, 312–319, Table 2, „ 2002, with permission from The Japanese Society for Food Science and Technology.

No. -20∞C, 24 d 25∞C, 12 d 70∞C, 4 h 1 propanal 2 acetone acetone 3 butanal 4 2-butanone 2-butanone 5 2-methylbutanal 2-methylbutanal 6 3-methylbutanal 3-methylbutanal 7 2-ethylfuran 2-ethylfuran 8 pentanal pentanal 9 2,3-pentanedione 2,3-pentanedione 10 hexanal 11 1-penten-3-ol 1-penten-3-ol 1-penten-3-ol

dehydes such as pentanal and hexanal increased at –20∞C, though these aldehydes were not detected in the samples kept at 70∞C. On the other hand, the amounts of branched chain aldehydes and ketones increased in the sample kept at 70∞C. These results suggested that degradation of fat (first reaction) and reaction among degradation products and other components (secondary reactions) would proceed simultaneously and their reaction velocities were different each other. When some of these compounds were added to crushed dried bonito and heated, hexanal and 2,3-pentadione decreased in the greatest degree (Fig. 17). This suggested that straight-chain aldehydes and a-diketones should have high reactivity. To investigate the generation mechanism of increased carbonyl volatile compounds, a series of experiments were carried out using crushed dried bonito and labeled oxygen compounds. The labeled model head space gas 18 was prepared by mixing of O2 (17.6%) and N2 (82.4%). Both of these crushed dried bonito samples (with ordinary head space gas and with the labeled Fig. 17. Recovery of the amount of volatile compounds af- model gas) were stored at –20 C and the volatile com- ∞ ter heating crushed dried bonito at 70∞C for 4 hours. Each pounds were analyzed by GC-MS (Fig. 18). The re- compound (2 ml) was added to crushed dried bonito (20 g) sults showed that 18O was incorporated to 1-penten-3- before heating. Cited from Nippon Shokuhin Kagaku Kogaku ol during storage. When water contained in crushed Kaishi, 49(5), Kawaguchi et al., Changes in the volatile com- 18 dried bonito was replaced to labeled water (H2 O) and pounds related to deterioration of crushed dried bonito dur- kept at –20∞, labeled oxygen was incorporated to ing Storage, 312–319, Fig. 2, „ 2002, with permission from propanal, butanal, pentanal, 2,3-pentadione, and The Japanese Society for Food Science and Technology. hexanal, which increased during storage (Kawaguchi et al. 2002b). It was clarified that these principal carbonyl compounds were derived from the head space All the results showed that acidic odor was gener- gas or water in dried bonito during storage. The addi- ated by formation of straight-chain aldehydes under tion tests of these carbonyl compounds to crushed dried cool condition (–20∞C), but acidic odor was diminished bonito samples showed that straight-chain aldehydes according to decreases in straight-chain aldehydes un- had strong sensual addition effects (adding rancid der higher temperatures. These findings suggested that odor), though branched chain aldehydes and ketones straight-chain aldehydes might be the cause of flavor did not have remarkable addition effects (Kawaguchi deterioration in crushed dried bonito. et al. 2002b).

Fig. 19. Changes in the amount of volatile sulfur compounds in crushed dried bonito during storage. COS, carbonyl sulfide; DMS, dimethyl sulfide. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 49(2), Kawaguchi et al., The mechanism of decrease of hydrogen sulfide in crushed dried bonito during storage, 99–105, Fig. 2, „ 2002, with permission from The Japanese Society for Food Science and Technology.

Fig. 18. Mass spectra of 1-penten-3-ol formed during storage of crushed dried bonito with labeled or unlabeled oxy- methanethiol, and dimethyl sulfide to the deteriorated gen. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, crushed dried bonito made the flavor as fresh as just 49(5), Kawaguchi et al., Changes in the volatile compounds after crushing. These results suggested that sulfur com- related to deterioration of crushed dried bonito during stor- pounds, such as hydrogen sulfide, should be the key age, 312–319, Fig. 3, „ 2002, with permission from The Japanese Society for Food Science and Technology. flavor components of crushed dried bonito.

4-4. The mechanism of decrease of hydrogen sulfide in crushed dried bonito during stor- 4-3. Changes in the volatile sulfur compounds in age crushed dried bonito during storage The volatile sulfur compounds were known to have Sakakibara et al. (1988) studied the changes of vola- low threshold and important key flavor components of tile compounds in dried bonito during storage under many foods, such as seafood. But little was known 30∞C. They speculated that decrease of sulfur-contain- about the generation or decrease mechanism of hydro- ing compounds, such as methanethiol and dimethyl gen sulfide (H2S) in dried bonito because its boiling sulfide, was the major deterioration factor of odors. point is very low and analysis method had not been We investigated the changes in major volatile sulfur- established. containing compounds during storage under various We investigated the decrease mechanism of hydro- conditions by GC analyses. The static head space gas gen sulfide in crushed dried bonito. We prepared ether- trapping method was employed in place of the purge extractive fraction and water-soluble fraction from and trap method in order to avoid volatilization of dried bonito, and examined reactivity of each fraction sulfur compounds. The GC-FPD analyses revealed the with H2S. After preservation of the aqueous solution amounts of volatile sulfur compounds (hydrogen composed of water-soluble fraction and H2S in the sulfide, carbonyl sulfide, methanethiol, carbon closed vessel for a few days at 25∞C, H2S disappeared disulfide, and dimethyl sulfide) during storage of and generated the pale precipitate of sulfur (Kawaguchi crushed dried bonito. The sulfur compounds decreased et al. 2002a). We investigated the key water soluble with time and hydrogen sulfide decreased most rap- components which caused the decrease in H2S, focus- idly (Fig. 19). Addition of hydrogen sulfide, ing on the coexist-effect of FeCl3 with amino acids or

Table 12. The reactivities of amino acids with H2S under various pH conditions. Cited from Nippon Shokuhin Kagaku Kogaku Kaishi, 49(2), Kawaguchi et al., The mechanism of decrease of hydrogen sulfide in crushed dried bonito during storage, 99– 105, Table 2, „ 2002, with permission from The Japanese Society for Food Science and Technology.

No. Amino acid pH Generation of Decrease of Browning of

precipitate H2S solution Before adjustment After adjustment

1 Gly 5.4 --- 2 L-Ala 5.4 --- 3 L-Asp 3.1 --- 4 L-Glu 3.7 --- 5 L-His 7.7 + + + 6 L-Arg◊HCl 5.4 --- 7 L-Lys◊HCl 5.3 --- 8 Tau 5 . 1 --- 9 Gly 5.4 5.8 --- 10 L-Ala 5.4 5.8 --- 11 L-Asp 3.1 5.8 --- 12 L-Glu 3.7 5.8 --- 13 L-His 7.7 5.8 + + + 14 L-Arg◊HCl 5.4 5.8 --- 15 L-Lys◊HCl 5.3 5.8 --- 16 Tau 5.1 5.8 ---

+, Recognized. –, not recognized. 480 ml of 1.0 mM FeCl3 and 80 mg of each amino acid were dissolved in 40 ml of distilled water. 5 ml of this solution and 2 ml of H2S were mixed in the closed vessel and stored at 25∞C. pH values of the solution 1~8 were not adjusted. pH values of the solution 9~16 were adjusted at 5.8.

organic acids. Referencing the analytical data by Fuke many seasoning products and processed foods in Ja- et al. (1989), we prepared model solutions contained pan. mixture of amino acids or mixture of organic acids. We also revealed that NaCl or IMP influenced on When FeCl3 and amino acids coexisted, H2S apparently taste characters of some sweet amino acids. The re- decreased and precipitate generated (Kawaguchi et al. sults would catch interest of scientists who are engaged 2002a). in studies on amino acids or taste physiology. Toda et Then we prepared several water solutions contain- al. (2013) proposed the molecular mechanism for ing several amino acids and FeCl3. The reactivities of amino acids and IMP recognition of T1R1/T1R3 (the amino acids with H2S under various pH conditions were umami receptor) by using chimeric human-mouse mu- shown in Table 12. Among tested amino acids, the ef- tants of the receptor. In near future the synergistic effect by histidine was the largest. The result suggested fects between these amino acids and IMP will be stud- that H2S in dried bonito would be oxidized to sulfur ied in more detail by using modern molecular biologi- powder catalytically by the complex of Fe3+ and histi- cal research method. dine (Fig. 20). Yamaguchi and Takahashi (1984) studied the interaction of MSG and NaCl on saltiness and palatability 5. Summary and future perspective of a clear soup. They revealed that a high palatability score was retained even though NaCl concentration was We successfully revealed the key flavor constituents reduced, as long as MSG was added at its optimal level. in Kombu-dashi by a series of omission tests. The mix- The study suggested that NaCl intake level could be ture of mannitol, monosodium glutamate, potassium reduced by using MSG. Our findings would also con- chloride and sodium chloride exhibited Kombu-dashi tribute to the development of low salt ingredients and like tastes and characteristic kokumi flavors after the reduction of sodium intake from processed foods by interactions among these compounds. We got the ba- the results of flavor interactions among MSG, amino sic knowledge leading to new applications of KCl and acids, and salts. sugar alcohols for quality improvements of food. These We found a new kind of non-volatile flavor compo- findings have already contributed for developments of nents in foods, focusing on the flavor interaction with

Fig. 20. A presumed mechanism of oxidation of H2S in crushed dried bonito. Modified from Nippon Shokuhin Kagaku Kogaku Kaishi, 49(2), Kawaguchi et al., The mechanism of decrease of hydrogen sulfide in crushed dried bonito during storage, 99– 105, Fig. 4, „ 2002, with permission from The Japanese Society for Food Science and Technology.

umami substances. The sulfur-containing compounds in garlic and onion, such as alliin, S-propenyl-cysteine sulfoxide and GSH, had no tastes themselves but they gave rise to kokumi flavors in the umami solution or soups. Scallop contained GSH in a high amount and the peptide suggested to contribute to characteristic Fig. 21. Structure of g-L-glutamyl-L-valylglycine. kokumi flavors of the food. The research on the chemical nature of the peptide revealed that it would be valu- able ingredient for food industry. On the bases of our dried bonito. We examined the changes in the amounts studies, a novel kokumi peptide, g-L-glutamyl-L- of some volatile sulfur compounds, key compounds for valylglicine (g-Glu-Val-Gly, Fig. 21), was found out fresh dried bonito odors, during storage of crushed by using modern biological technique (Ohsu et al. dried bonito. All these compounds decreased much 2010). They succeeded in the cloning of sensory during storage and hydrogen sulfide almost disap- receptor of GSH and determined several compounds peared. The results of our study suggested that hydro- which combined to the receptor. g-Glu-Val-Gly exhib- gen sulfide would be oxidized to sulfur by ferric ion its strong kokumi flavors in various foods and recently and histidine in dried bonito. There are various kinds it was approved by Japanese government as a speci- of processed foods which are made of dried bonito in fied food additive, expecting to be applied for proc- Japan. Effective generation of fresh dried bonito odors essed foods. Few studies have been done on kokumi and prevention of odor deterioration during total pro- flavors. The difference between GSH and g-Glu-Val- duction processes are essential issues for development Gly will be an interesting subject for scientists in fu- of high quality products. We believe all the findings in ture. this study should contribute to the resolutions of these The generation and decreasing mechanism of some issues and for future studies of volatile flavor compo- key volatile flavor components of dried bonito were nents in foods. determined by using GC-MS and GC-Sniffing methods. We found the mechanism of generating pyrazines, Acknowledgments the key components of savory odors of dried bonito. We are grateful to Dr. S. Watabe for helpful advices. This Pyrazines were presumed to be produced after the re- study was realized in cooperation with many persons of action between some smoke derived compounds, such Ajinomoto Co., Inc.: especially Mr. M. Sakaguchi, Mr. T. as acetol and 2,3-pentandione, and the other compo- Shimizu, Dr. A. Okiyama, Mr. N. Miyamura, Dr. H. nents in bonito meat. The studies on the changes of Wakabayashi, Mr. M. Yonemitsu, Dr. M. Kawai, Dr. M. aldehydes and ketones during storage under different Kuroda, and Dr. H. Kawaguchi. temperatures suggested that the compounds were produced by the complex reactions among low molecular References Cerny C, Grosch W. Precursor of ethyldimethylpyrazine iso- components other than degradation of fat. Straight mers and 2,3-diethyl-5-methylpyrazine formed in roasted chain aldehydes, such as hexanal and pentanal, might beef. Z. Lebensm. Unters, Forsch. 1994; 198: 210. be involved in the deterioration of flavor in crushed

Fuke S, Ueda Y. Interactions between umami and other flavor Li X, Staszewski L, Xu H, Durick K, Zoller M, Adler E. characteristics. Trends Food Sci. Technol., Elsevier Trends Human receptors for sweet and umami taste. Proc. Natl Journals 1996; 7: 407–411. Acad. Sci. USA 2002; 99: 4692–4696. Fuke S, Watanabe K, Sakai H, Konosu S. Extractive compo- Mistry BS, Reineccius T, Olson I. In: Marsili, R (ed). Tech- nents of dried skipjack (Katsuobushi). Nippon Shokuhin niques for Analyzing Food Aroma. Dekker, New York. Kagaku Kogaku Kaishi 1989; 36: 67–70. 1997; 265. Furukawa H. Synergistic effects in the taste of a-amino- Nelson G, Chandrashekar J, Hoon MA, Feng L, Zhao G, dicarboxylic acids and L-homocysteinate with 5¢-inosinate. Ryba NJP, Zuker CS. An amino-acid taste receptor. Na- Nippon Nogeikagaku Kaihi 1991; 65: 163–169. ture 2002; 416: 199–202. Hayashi T, Yamaguchi K, Konosu S. Sensory analysis of Ohsu T, Amino Y, Nagasaki H, Yamanaka T, Takeshita S, taste-active components in the extract of boiled snow crab Hatanaka T, Maruyama Y, Miyamura N, Eto Y. Involve- meat. J. Food Sci. 1981; 46: 479–483 and 493. ment of the calcium-sensing receptor in human taste per- Huang TC, Fu HY, Ho CT. Mechanistic studies of ception, J. Biol. Chem. 2010; 285: 1016–1022. tetramethylpyrazine formation under weak acidic condi- Rizzi GP. Formation of pyrazines from acyloin precursors tions and high hydrostatic pressure. J. Agric. Food under mild conditions. J. Agric. Food Chem. 1988; 36: Chem.1996; 44: 240. 349. Hwang H-I, Hartman TG. Formation of pyrazines from the Sakakibara H, Yanai T, Yajima I, Hayasihi K. Changes in Maillard reaction of glucose and lysine-a-amine-15N. J. volatile flavor compounds of powdered dried bonito Agric. Food Chem. 1994; 42: 1000. (Katsuobushi) during storage. Biosci. Biotech. Biochem. Ikeda K. On the taste of salt of glutamic acid. Int. Congr. 1988; 52: 2731–2739. Appl. Chem. 1912; 18: 147 (in Japanese). Sakakibara H, Ide J, Yanai T, Yajima I, Hayashi K. Volatile Ishiguro K, Wakabayashi H, Kawaguchi H. Changes in vola- flavor compounds of some kinds of dried and smoked fish. tile compounds during smoking process and evaluation of Biosci. Biotech. Biochem. 1990; 54: 9–16. major aroma constituents of dried bonito (Katuo-Bushi). Toda Y, Nakagita T, Hayakawa T, Okada S, Narukawa M, Nippon Shokuhin Kagaku Kogaku Kaishi 2001; 48: 570– Imai H, Ishimaru Y, Misaka T. Two distinct determinants 577. of ligand specificity in T1R1/T1R3 (the umami taste Kawaguchi H, Ishiguro K, Wakabayashi H, Ueda Y. Study receptor). J. Biol. Chem. 2013; 288: 36863–36877. on the generation of pyrazines and change in amounts of Ueda Y. Studies on the flavor characteristics “Koku” and volatile compounds in smoke during preparation of dried “Atsumi” in some tasty foodstuffs. Jpn. J. Taste Smell Res. bonito. Nippon Shokuhin Kagaku Kogaku Kaishi 2001; 48: 1997; 4: 197–200 (in Japanese). 899–905. Ueda Y, Sakaguchi M, Hirayama K, Miyajima R, Kimizuka Kawaguchi H, Ishiguro K, Wakabayashi H, Ueda Y. The A. Characteristic flavor constituents in water extract in mechanism of decrease of hydrogen sulfide in crushed garlic. Biosci. Biotech. Biochem. 1990; 54: 163–169. dried bonito during storage. Nippon Shokuhin Kagaku Ueda Y, Kawajiri H, Miyamura N, Miyajima R. Content of Kogaku Kaishi 2002a; 49: 99–105. some Sulfur-containing components and free amino acids Kawaguchi H, Ishiguro K, Wakabayashi H, Fukami K, Ueda in various strains of garlic. Nippon Shokuhin Kagaku Y. Changes in the volatile compounds related to deterio- Kgaku Kaishi 1991; 38: 429–434. ration of crushed dried bonito during storage. Nippon Ueda Y, Tsubuku T, Miyajima R. Composition of sulfur-con- Shokuhin Kagaku Kogaku Kaishi, 2002b; 49: 312–319. taining components in onion and their flavor characters. Kawai M, Okiyama A, Ueda Y. The interactions between Biosci. Biotech. Biochem. 1994; 58: 108–110. various amino acids and Na/KCl. Jpn. J. Taste Smell Res. Ueda Y, Yonemitsu M, Tsubuku T, Sakaguchi M, Miyajima 2001; 8: 491–494. R. Flavor characteristics of glutathione in raw and cooked Kawai M, Okiyama A, Ueda Y. Taste enhancements between foodstuffs. Biosci. Biotech. Biochem. 1997; 61: 1977– various amino acids and IMP. Chem. Senses 2002; 27: 739– 1980. 745. Ueda Y, Hibino G, Kohmura M, Kuroda M, Watanabe K, Komata Y. Studies on the extractives of “Uni”—IV. Taste of Sakaguchi M. Contents of glutathione in seafoods and its each component in the extractives. Nippon Suisan flavor characteristics. Nippon Suisan Gakkaishi 1998; 64: Gakkaishi 1964; 30: 749–756. 710–714. Konosu S. The tastes of fish and shellfish. In: Boudreau JC Ugawa T, Konosu S, Kurihara K. Enhancing effects of NaCl (ed). Food Taste Chemistry. ACS Symposium Series 115. and Na phosphate on human gustatory responses on amino ACS, Washington, D.C. 1979; 185–203. acids. Chem. Senses 1992; 17: 811–815. Konosu S, Watanabe K, Koriyama T, Shirai T, Yamguchi K. Watanabe K, Lan H-L, Yamaguchi K, Konosu S. Role of Extractive components of scallop and identification of its extractive components of scallop in its characteristic taste taste-active components by omission test. Nippon development. Nippon Shokuhin Kagaku Kogaku Kaishi Shokuhin Kagaku Kogaku Kaishi, 1988; 35: 252–258. 1990; 37: 439–445. Kuninaka A. History and development of flavor nucleotides. Yamaguchi S. Synergistic taste effect of monosodium gluta- Food Technol. 1964; 18: 29–35. mate and di-sodium 5¢-inosinate. J. Food Sci. 1967; 32: Kuroda M, Nagata H, Tsubuku T, Kawajiri H, Ueda Y. Si- 473–478. multaneous determination of glutathione, g- Yamaguchi S. The umami taste. In: Boudreau JC (ed). Food Glutamylcysteine and cysteine in commercial yeast ex- Taste Chemistry. American Chemical Society. 1979; 33– tract by HPLC with fluorometric detection. Food Sci. Tech. 35. Int. Tokyo 1997; 3: 239–240. Yamaguchi S. Fundamental properties of umami in human

taste sensation. In: Kawamura Y, Kare MR (eds). Umami. Zhang Y, Ho CT. Volatile compounds formed from thermal A Basic Taste. Marcel Dekker, Inc., New York. 1987; 41– interaction of 2,4-decadienal with cysteine and glutath- 73. ione. J. Agric. Food Chem. 1989; 37: 1016–1020. Yamaguchi S, Kimizuka A. Psychometric studies on the taste Zhang Y, Ho CT. Comparison of the volatile compounds of monosodium glutamate. In: Filer LJ, Garattini S, Kare formed from the thermal reaction of glucose with cysteine MR, Reynolds WA, Wurtman RJ (eds). Glutamic Acid: and glutathione. J. Agric. Food Chem. 1991a; 39: 760– Advances in Biochemistry and Physiology. Raven Press, 963. New York. 1979; 35–54. Zhang Y, Ho CT. Formation of meat-like aroma compounds Yamaguchi S, Takahashi C. Interactions of monosodium from thermal reaction of inosine 5¢-monophosphate with glutamate and sodium chloride on saltiness and palatabil- cysteine and glutathione. J. Agric. Food Chem. 1991b; 39: ity of a clear soup. J. Food Sci. 1984; 49: 82–85. 1145–1148. Yamaguchi S, Yoshikawa T, Ikeda S, Ninomiya T. Measure- Zhang Y, Chen M, Ho CT. Comparison of the volatile com- ment of the relative taste intensity of some L-a-amino ac- pounds obtained from thermal degradation of cysteine and ids and 5¢-nucleotides. J. Food Sci. 1971; 36: 846–849. glutathione. J. Agric. Food Chem. 1988; 36: 992–996.