Analysis of aroma compounds derived from heated tsukudani seasoning liquid models containing rare sugars

Abstract

Three tsukudani seasoning liquid models of soy sauce and amino acid seasoning liquid mixed with rare sugar syrup (RSS, main rare sugars: d-allulose, d-sorbose, d-tagatose, d-allose), glucose-fructose liquid sugar (GF), and white sugar (WS) were heated in a headspace sampler to create a delicious aroma. The aroma components generated were then analyzed using HS-GC/MS. As a result of the Maillard reaction, brown color intensity measured at absorbance 420 nm in the RSS liquid was significantly higher than in the GF and WS liquids. Principal component analysis of the quantitative data of the aroma compounds revealed 1.1–4.5 times more furans such as furfural, furfuryl alcohol, 2,5-diethyl tetrahydrofuran, and 3-phenylfuran were released from RSS than from GF and WS, which contributed to the characteristic aroma components of the tsukudani seasoning liquid. These results suggest that the four main rare sugars in the RSS are involved in the aroma formation.

Introduction

To make tsukudani, large amounts of sweeteners with amino acid seasoning liquids have been used not only for adding seasoning flavors but also for reducing water activity and imparting roasting aromasⁱ⁾. In particular, the aroma generated by the Maillard reaction caused by heating the seasoning liquid is important for enhancing palatability. Meng et al. (2015) reported that the flavor of food is qualitatively changed by cooking and heating with a seasoning liquid containing soy sauce, mirin, and sugars. Because tsukudani is cooked using a lot of sugar, developing new tsukudani with reduced sugar and some health benefits is required.

Rare sugar syrup (RSS) with functional claims of reducing blood sugar level (Yamada et al., 2017) contains 44.3 % glucose, 31.9 % fructose, and about 15 % rare sugars (D-allulose, D-sorbose, D-tagatose, D-allose). Moreover, using RSS in foods has additional merits such as improving richness in taste and adding flavor (Uchiyama, 2020), which have come to be expected in various cooked and processed foods. In particular, when RSS is used in cooked products that are simmered for a long time, such as tsukudani, it tastes more flavorful than when other sweeteners are used.

Aroma formation during the Maillard reaction of D-allulose (Tamura et al., 2008) or D-tagatose (Cho et al., 2010) has been reported. However, RSS containing D-allulose, D-sorbose, D-tagatose, and D-allose has not been investigated for the formation of aroma components during cooking. Tsukudani, which has a high sugar content and is cooked for a long time, produces many characteristic aroma components through the Maillard reaction (Okumura, 1993). Therefore, the generation and control of aroma is an important issue.

In this study, tsukudani seasoning liquid models using white sugar (WS), glucose-fructose liquid sugar (GF), or RSS were investigated. The aroma compounds generated during the cooking process were quantified using a headspace-gas chromatograph/mass spectrometer to clarify the aroma generation by adding RSS as an ingredient as well as the characteristics of the cooking aroma.

Materials and Methods

Chemicals　Rare sugar syrup (Concentration of each sugar: 44.3 % glucose, 31.9 % fructose, 6.0 % allulose, and 17.8 % other saccharides) (Hayashi et al., 2014) was purchased from Matsutani Chemical Industry Co., Ltd. (Hyogo), and rare sugars (D-allulose, D-allose, D-tagatose, and D-sorbose) were supplied by the International Institute of Rare Sugar Research and Education, Kagawa University (Kagawa). White sugar (Molasses has been completely removed in the refining process; Mitsui Sugar Co., Ltd., Tokyo) and koikuchi soy sauce (Morita Co., Ltd., Aichi) were purchased from a supermarket. Amino acid seasoning liquid (diamond amino acid liquid, reduced salt NK-12) obtained by hydrolyzing plant proteins was supplied by Bansyu Chomiryo Co., Ltd. (Hyogo). Glucose-fructose liquid sugar (i.e. corn-syrup, glucose:fructose = 6:4) was manufactured by Nihon Shokuhin Kako Co., Ltd. (Tokyo). Unless otherwise specified, all other sugars and chemicals were obtained from Fujifilm Wako Pure Chemical Corporation (Osaka).

Preparation and heat treatment of the tsukudani seasoning liquid　The tsukudani seasoning liquid was prepared according to the method of Aoyama et al. (2000). Soy sauce (30 g) and amino acid seasoning liquid (8.0 g) were mixed with RSS, GF, or WS until 18.0 % Brix concentration was reached, following by an adjustment of the final weight to 100.0 g with distilled water. The prepared tsukudani seasoning liquid (2 mL) was dispensed into a 20 mL vial, 500 µL of 2 mg/mL 1-propanol was added as an internal standard solution, and the vial was sealed. Model cooking was performed by shaking and heating the vial containing the seasoning liquid at 85 °C for 60 min using the oven built into the HS-20 Trap headspace sampler (Shimadzu Corporation, Kyoto).

Determination of color intensity and pH　The color intensity was determined by diluting the tsukudani seasoning liquid heated at 85 °C for 60 min with distilled water 10 times and measuring the absorbance at 420 nm using a V-730 spectrophotometer (Jasco Corporation, Tokyo) with a 10 mm cell. The pH of each sample was determined using a B-712 compact pH meter (Horiba, Ltd., Kyoto).

Sugar composition　Sugars analysis was performed on a Thermo Scientific Dionex UltiMate 3000 System with a Corona Veo RS Charged Aerosol Detector (CAD; Thermo Fisher Scientific K.K., Tokyo). The sugar composition of the tsukudani seasoning liquid before and after heating was determined based on the method reported by Miyoshi et al. (2019) by a direct coupling of two CARBOSep CHO 882 lead form columns (7.8 mm i.d. × 300 mm; Tokyo Chemical Industry Co., Ltd., Tokyo). For the analysis, the sample was desalted using a G1 Micro Acilyzer (Asahi Kasei Co., Ltd., Tokyo) equipped with an AC-110-10 electrodialysis membrane (Astom Corporation, Tokyo).

Amino acid composition　The tsukudani seasoning liquid before and after heating was diluted 1000-fold with distilled water and filtered through a centrifugal ultrafiltration filter (Amicon Ultra NMWL 3K; Merck KGaA, Germany). The filtrate obtained was analyzed using a high-performance liquid chromatograph (HPLC) by a ortho-phthalaldehyde (OPA)-post column derivatization system (Jasco) equipped with a Hitachi #2619PH column (4.0 × 150 mm; Hitachi High-Tech Fielding Corporation, Tokyo).

Headspace-gas chromatography/mass spectroscopy analysis　The aroma components were analyzed using a headspace-gas chromatograph/mass spectrometer (HS-GC/MS, GCMS-QP2010 Ultra; Shimadzu) equipped with an HS-20 Trap headspace sampler and an SLB-5 ms fused silica capillary column (30 m, 0.25 mm i.d., film thickness 0.25&x#00B5;m; Merck). The operating conditions of the headspace sampler were as follows: After heating the vial, headspace gas flowed at 100.0 kPa for 6 s, and then the aroma compounds were trapped on a hydrophobic Tenax adsorbent at 50 °C. Finally, the aroma compounds were released at 200 °C. The transfer line was heated at 150 °C. The operating conditions of the GC were as follows: Using ultra-pure helium as a carrier gas, analysis was performed in a splitless manner at 37.1 kPa. The oven was maintained at 50 °C for 10 min, programed at 3 °C/min until 300 °C, and then held for 5 min. Mass spectrometry was performed in scan mode in an ion source of 200 °C, an interface of 250 °C, and a mass range of 40–400 m/z using the electron ionization method (EI method).

Data processing and statistical analysis　All experiments were performed in triplicate, and various aroma components detected by GC/MS analysis were expressed as mean ± standard deviation. A relative quantitative value was calculated using the ratio of the peak area of the detected compound to the peak area of the internal standard 1-propanol (RT: 7.2 min). Using the FFNSC2 mass spectral database (Shimadzu) and NIST Mass Spectral Library 11 (NIST, Gaithersburg, MD, USA), the retention indices (RI) were calculated from the retention times of eluents and an alkane mixture (C₇-C₂₂) solution (Hayashi Pure Chemical Ind., Ltd., Osaka) as a standard substance. The HS-GC/MS analytical data were subjected to principal component analysis (PCA) using Pirouette ver. 4.0 (GL Sciences Inc., Tokyo) to evaluate the aroma characteristics of the tsukudani seasoning liquids.

Results and Discussion

Heating-induced changes in brown color intensity and amino acid and sugar composition of the tsukudani seasoning liquids　Figure 1 shows the heating-induced changes in brown color intensity and pH values of the tsukudani seasoning liquids containing different sugars. The brown color intensity of the tsukudani seasoning liquid containing RSS was 2 % (p < 0.05) and 4 % (p < 0.05) higher than that of the seasoning liquids containing GF and WS, respectively, and the pH of the seasonings were 5.03, 5.10, and 5.13 for RSS, GF, and WS, respectively. The pH value (5.03) of RSS was the lowest but there was no statistically significant difference among those pH values. Because H⁺ ions are formed during the Maillard reaction, the pH of the reaction mixture may decrease (Davies and Labuza, 1997), suggesting that the tsukudani seasoning liquid with RSS was more susceptible to the Maillard reaction than those with GF and WS. Table 1 shows the heating-induced changes in the sugar and amino acid compositions of the various tsukudani seasoning liquids. The total sugar decrease in the RSS liquid after heating was 56.12 mg/mL (individual reductions in glucose, fructose, sorbose, allose, tagatose, and allulose were 33.05, 18.88, 1.73, 0.52, 0.43, and 1.51 mg/mL, respectively). The retaining rate of sugars during the heating may not be related to the kind of sugar such as rare sugars, glucose, and fructose in RSS but rather related to their concentrations as reported by Oshima et al. (2014). In the GF liquid, the total amount of sugar decreased by 60.52 mg/mL (glucose and fructose: 34.81 and 25.71 mg/mL, respectively), and in the WS liquid, glucose and fructose were found before heat treatment, which would be derived from the soy sauce used to mix the seasoning liquids. The total amount of sugar in the WS liquid decreased by 29.14 mg/mL (sucrose, glucose, and fructose: 27.16, 1.27, and 0.71 mg/mL, respectively). Because the tsukudani seasoning liquids were at approximately pH 5, the sucrose in the WS liquid may be partially hydrolyzed to glucose and fructose during heat treatment and then changed into Maillard reaction products with various amino acids. Reyes et al. (1982) reported that sucrose was readily hydrolyzed under the reaction conditions (60 °C, pH 3.5) and underwent Maillard browning reactions.

Fig. 1

Heating-induced changes in brown color intensity and pH of the tsukudani seasoning liquids containing different sugars.

RSS, rare sugar syrup; GF, glucose-fructose liquid sugar; WS, white sugar.

RSS, GF, and WS were sweeteners in the tsukudani seasoning liquids, and the browning reaction is described in the text in detail. Values are mean ± SD (standard deviation) of three repeats. Asterisks indicate significant difference obtained by one-way analysis of variance and multiple comparison tests (Bonferroni–Dunn) at p < 0.05.

Table 1. Heating-induced changes in the sugar and amino acid compositions of the tsukudani seasoning liquids containing different sugars.

Compound (mg/mL)	RSS					GF					WS
	Before heating		After heating		DAb	Before heating		After heating		DA	Before heating		After heating		DA
	Average	SDa	Average	SD		Average	SD	Average	SD		Average	SD	Average	SD
< Sugar >
Glucose	107.93	3.80	74.88	8.91	33.05	129.40	0.24	94.59	8.75	34.81	6.79	0.28	5.52	0.69	1.27
Fructose	63.31	1.87	44.43	5.35	18.88	98.29	2.20	72.58	7.70	25.71	4.90	0.37	4.19	0.36	0.71
Sorbose	9.85	0.07	8.12	0.12	1.73	-	-	-	-	-	-	-	-	-	-
Aliose	2.46	0.09	1.94	0.02	0.52	-	-	-	-	-	-	-	-	-	-
Tagatose	2.39	0.10	1.96	0.09	0.43	-	-	-	-	-	-	-	-	-	-
Allulose	8.78	0.16	7.27	0.10	1.51	-	-	-	-	-	-	-	-	-	-
Sucrose	-	-	-	-	-	-	-	-	-	-	186.25	13.04	159.09	2.82	27.16
					f
< Amino acid >
Aspartic acid	3.93	0.06	3.29	0.21	0.64	4.04	0.23	3.59	0.09	0.45	4.21	0.11	3.50	0.07	0.71
Threonine	1.50	0.02	1.22	0.07	0.28	1.47	0.08	1.30	0.02	0.17	1.53	0.01	1.27	0.02	0.26
Serine	2.05	0.06	1.68	0.11	0.37	2.06	0.17	1.82	0.07	0.24	2.11	0.03	1.80	0.07	0.31
Glutamic acid	6.02	0.09	4.86 c, d	0.09	1.16	6.14	0.30	5.37c, e	0.12	0.77	6.38	0.16	5.20d, e	0.08	1.18
Proline	2.14	0.10	1.62	0.08	0.52	2.13	0.03	1.70	0.06	0.43	2.11	0.03	1.64	0.09	0.47
Glycine	1.26	0.03	1.05	0.08	0.21	1.27	0.08	1.11	0.04	0.16	1.30	0.01	1.10	0.03	0.20
Alanine	1.90	0.03	1.69	0.22	0.21	1.94	0.09	1.73	0.03	0.21	1.98	0.01	1.74	0.10	0.24
Valine	1.70	0.02	1.47	0.18	0.23	1.71	0.08	1.52	0.03	0.19	1.77	0.02	1.50	0.06	0.27
Methionine	0.23	0.01	0.19	0.03	0.04	0.25	0.01	0.19	0.00	0.06	0.26	0.00	0.18	0.01	0.08
Isoleucine	1.46	0.03	1.28	0.16	0.18	1.49	0.06	1.31	0.03	0.18	1.53	0.02	1.30	0.04	0.23
Leucine	2.30	0.04	2.03	0.25	0.27	2.35	0.11	2.09	0.04	0.26	2.43	0.03	2.03	0.05	0.40
Tyrosine	0.31	0.01	0.32	0.09	−0.01	0.25	0.03	0.22	0.02	0.03	0.22	0.01	0.24	0.05	−0.02
Phenylalanine	2.94	0.07	2.65	0.14	0.29	2.91	0.12	2.63	0.11	0.28	2.78	0.05	2.51	0.05	0.27
Lysine	1.57	0.05	1.29	0.11	0.28	1.63	0.27	1.65	0.15	−0.02	1.75	0.05	1.37	0.04	0.38
Histidine	0.51	0.01	0.39	0.05	0.12	0.52	0.03	0.44	0.04	0.08	0.57	0.07	0.45	0.08	0.12
Arginine	1.27	0.05	1.10	0.16	0.17 <4.96>g	1.28	0.08	1.13	0.14	0.15 <3.66>	1.38	0.12	1.19	0.18	0.19 <5.29>

^a  SD: Standard deviation

^b  DA: Difference in the amount of individual compounds between “Before heating” and “After heating”

^c–e  Values in the same row with the same letter are one-way analysis of variance and multiple comparison tests (Bonferroni–Dunn) at p < 0.05

^f  ⃞ : Total amount of sugar reduction

^g  <> : Total amount of amino acid reduction

The total amounts of amino acid in the RSS, GF, and WS liquids changed to −4.96 mg/mL (−16 %), −3.66 mg/mL (−12 %), and −5.29 mg/mL (−16 %), respectively. Comparing the RSS and GF liquids, the amount of individual amino acids except methionine and tyrosine in the RSS liquid was the same or lower than that in the GF liquid. Because rare sugars in the RSS liquid were more reactive, these rare sugars may be closely involved in the Maillard reaction rather than glucose and fructose.

There were changes in the amino acid composition of the seasoning liquids after heating, with significant differences observed in the residual amounts of glutamic acid (p < 0.05) in the RSS, GF, and WS liquids. In tsukudani seasonings, glutamic acid is a major amino acid component, suggesting that it is closely involved in the browning of seasonings after heating. In addition, its amino acid importance in terms of taste suggested that it should be considered in the use of RSS. Motai (1973) prepared melanoidins by heating various amino acids and peptides with sugars, and then reported that amino acids such as lysine, histidine, glutamic acid, and aspartic acid showed the highest browning color changes among the amino acids tested.

Various aroma compounds associated with heating the tsukudani seasoning liquids　The Maillard reaction of the tsukudani seasoning liquids occurred in an oven built into the HS-20 Trap headspace sampler (85 °C, 60 min) coupled with GC/MS. Aroma components generated from the headspace sampler were immediately injected into the GC/MS, and 43 aroma compounds were identified using the RI database and GC/MS libraries (Table 2).

Table 2. Aroma compounds produced by heating the tsukudani seasoning liquids containing different sugars.

No.	Aroma compound	RI aa	RSS		GF		WS		Odor quality	Reference
			RI bb	Peak area (×103)	RI bb	Peak area (×103)	RI bb	Peak area (×103)
×	Alcohols
1	isopentyl alcohol	729	732	28301.3	734	27513.8	739	28779.5
2	vinyl amyl carbinol	978	979	41.4	980	37.4	981	29.6
3	2-ethylhexanol	1030	1028	214.3a*	1028	140.2	1029	93.2a*	green	Hayase and Watanabe, 2014
4	phenethyl alcohol	1113	1109	968.6	1109	892.8	1109	866.9
5	4-ethylguaiacol	1275	1271	223.7	1271	238.9	1271	238.4
6	4-vinylguaiacol	1309	1307	34.0	1307	44.9	1307	58.8	roasted peanut	Starowicz and Zieliñski, 2019
Δ	Aldehydes
7	3-methylcrotonaldehyde	780	783	1758.2	786	1953.2	790	2914.9
8	benzaldehyde	960	958	4776.5 b**, c**	959	5283.8 b**, d**	960	6704.0 c**, d**	sharp, sweet	Starowicz and Zieliñski, 2019
9	phenylacetaldehyde	1045	1040	9522.4	1040	10046.3	1041	11800.3	green	Starowicz and Zieliñski, 2019
10	nonanal	1104	1101	127.6	1102	117.9	1102	138.8
11	2-phenylpropenal	1148	1152	353.2 e*, f*	1152	437.3 e*	1152	1368.9 f*	sharp aromatic odor	Huang et al., 2019
12	decanal	1208	1202	29.4	1203	43.3	1203	43.7
13	alplza-ethylidenebenzeneacetaldehyde	1265	1265	21.1 g*	1266	139.1	1266	159.6 g*	caramel-like, smoky, nutty	Jati and Ariza, 2011
14	dodecanal	1410	1406	21.1	1406	35.4	1406	45.5
15	5-methyl-2-phenyl-2-hexenal	1485	1481	13.9	1482	14.7	1482	33.0
▲	Esters
16	2-methylpropyl chloroacetate	946	948	89.3	948	115.2	948	116.9
17	2-ethylallyl butyrate	1013	1012	235.3	1013	212.1	1013	220.8
18	ethyl benzoate	1170	1167	11.4	1167	17.0	1167	28.4	flower-like	Horiguchi, 1955
19	diethyl succinate	1183	1176	130.2	1176	119.5	1177	136.4	odorless	Horiguchi, 1955
20	ethyl phenylacetate	1246	1239	87.8	1239	74.0	1251	38.8	sweet nectar	Horiguchi, 1955
21	2-phenethyl acetate	1257	1251	23.1	1251	29.6	1251	38.8
22	methyl para-tert-butylphenylacetate	1502	1503	121.6 h**	1455	1200.6 h**, i**	1454	162.1 i**	green	(ii) URL cited
23	ethyl palmitate	1993	1993	36.0 j*	1993	35.6 k**	1993	95.5 j*, k**	waxy	(iii) URL cited
♦	Acid
24	isovaleric acid	842	836	392.3	837	457.2	838	597.7	valeriana-like	Horiguchi, 1955
+	Ketones
25	2-heptanone	853	849	593.6 l*	850	405.9l*	852	466.8	strong fruit-like	Horiguchi, 1955
26	ethyl butyl ketone	885	882	149.7 m**	883	229.4 m**, n*	884	177.3 n*	fruity	(iii) URL cited
27	5-methyl-3-heptanone	938	940	240.1	940	218.5	941	294.0
28	3-octanone	986	983	21.5 o**, p**	984	62.1 o**, q*	984	112.7 p**, q*	fruit-like	Horiguchi, 1955
29	hexyl methyl ketone	989	988	9.1	990	6.7	990	16.0	carnation, reseda-like	Horiguchi, 1955
30	6-methylhept-5-en-2-one	986	991	230.7r*	991	295.1 r*	991	249.9	citrus, green	(iii) URL cited
31	acetophenone	1068	1062	124.1	1062	144.8	1062	167.9	bitter almond oil-like	Horiguchi, 1955
●	Furans
32	furfural	831	825	1930.5	826	1031.8	828	877.9	sweet, woody	Starowicz and Zieliñski, 2019
33	furfuryl alcohol	849	846	584.5 s*, t**	848	130.8 s*	850	198.2 t**	sulfurous, estery	Starowicz and Zieliñski, 2019
34	2,5-diethyltetrahydrofuran	896	897	328.7	897	302.3	899	307.4
35	2-pentylfuran	991	987	10.6 u**, v**	987	116.5 u**, w**	988	45.8 v**, w**	fruity, green, earthy	Starowicz and Zieliñski, 2019
36	tetrahydro-2,2-dimethyl-5-(1-methylpropyl)furan	1018	1024	100.0	1024	88.2	1025	81.4
37	3-phenylfuran	1216	1219	445.2 x**, y**	1219	328.4 x**	1219	352.9 y**	sweet	Ono et al., 2015
○	N-compounds
38	bipropionyl	801	801	557.4	803	590.8 z**	804	993.6 z**	buttery	(iii) URL cited
39	2-ethyl-6-methylpyrazine	1000	1000	330.9	1000	317.6	1001	287.5	green nutty, caramel, potato, cocoa	Starowicz and Zieliñski, 2019
40	2(or 3)-ethyl-3,5(or 2,5)-dimethylpyrazine	1081	1078	47.3	1078	59.6	1078	67.2
41	2-butyl-3,5-dimethylpyrazine	1306	1310	18.9	1310	20.8	1310	24.5
◊	S-compounds
42	3-methylthiopropionaldehyde	909	902	2890.1	905	3077.8	906	3054.2	cooked potatoes and soy sauce	John, 2014
43	dimethyl trisulfide	969	964	620.5	964	798.1	965	606.7	fresh onion	Starowicz and Zieliñski, 2019
Total amount of peak area				56867.2		57425.0		63092.5

Retention indices (RI) were determined using n-alkans, C₇-C₂₂ as external references. 1-Propanol (1000 µg) peak area was 4493770 as an internal standard. Each value is presented as mean ± SD (n = 3). The underline is maximum peak areas with significant difference.

RI ^aa, Library value by FFNSC2, NIST11; RI ^bb, Measured value; ^a–z, Values in the same row with the same letter are one-way analysis of variance and multiple comparison tests (Bonferroni–Dunn) (** p < 0.01; * p < 0.05).

The aroma components in the RSS liquid were quantified and then compared with the peak areas of those in the GF and WS liquids. 2-Ethylhexanol (green; Hayase and Watanabe, 2014), 2-heptanone (strong, fruit-like; Horiguchi, 1955), furfuryl alcohol (sulfurous, estery; Starowicz and Zieliñski, 2019), and 3-phenylfuran (sweet; Ono et al., 2015) in the RSS liquid were significantly higher than those in the GF (p < 0.01) and WS (p < 0.05) liquids as shown in Table 2 (n = 3). The sulfur-containing amino acid 3-(methylthio)propionaldehyde (smell of cooked potatoes and soy sauce; John, 2014), which has a high possibility of being derived from methionine, was detected and had a larger peak area than the other peak components in the chromatogram. However, the decreasing rate of methionine in the RSS liquid was the lowest (Table 1, RSS: 17.4 %; GF: 24.0 %; WS: 30.8 %), and also the peak area of 3-(methylthio)propionaldehyde from the RSS liquid was the smallest among the RSS, GF, and WS liquids (Table 2, RSS: 2890070; GF: 3077816; WS: 3054166). Therefore, we concluded that the generation of 3-(methylthio)propionaldehyde was suppressed in the RSS liquid. Because the odor threshold is 0.2 ppb and the amount in the three model solutions was not so different, the smell of cooked potatoes and soy sauce may contribute to all the samples.

In the GF liquid, methyl p-tert-butylphenylacetate (green)ⁱⁱ⁾, ethyl butyl ketone (fruity)ⁱⁱⁱ⁾, 6-methylhept-5-en-2-one (citrus, green)ⁱⁱⁱ⁾, and 2-pentylfuran (fruity, green, earthy; Starowicz and Zieliñski, 2019) were significantly higher than in the RSS and WS liquids at levels of p < 0.01 and p < 0.05 as shown in Table 2. Benzaldehyde (sharp, sweet; Starowicz and Zieliñski, 2019), 2-phenylpropenal (sharp aromatic odor; Huang et al., 2019), benzeneacet aldehyde alpha-ethylidene-(caramel-like, smoky, nutty; Jati and Ariza, 2011), ethyl palmitate (waxy)ⁱⁱⁱ⁾, 3-octanone (fruit-like; Horiguchi, 1955), and bipropionyl (buttery)ⁱⁱⁱ⁾ in the WS liquid were significantly higher than in the RSS and GF liquids at levels of p < 0.01 and p < 0.05 as shown in Table 2.

PCA of the aroma compounds from each seasoning liquid with a different sugar　PCA was performed on the 43 components identified in Table 2. Individual concentrations of the identified compounds calculated from the peak area of 1-propanol as the internal standard were set as the parameters for the explanatory variables. The score plot of the PCA is shown in Fig. 2. The first principal component (PC 1) contributes to 39.9 % and the second principal component (PC 2) was 18.1 %. PC 1 and PC 2 could explain 58.0 % of the total variance. The WS liquid is located to the right of PC 1, and the most browning RSS is plotted to the left of PC 1. Therefore, PC 1 may correlate to lightness of the browning intensity of the three sugar liquids, being associated with the activity of the Maillard reaction (RSS > GF > WS, as shown in Fig. 1). Because the loss of total amino acid content with heating in the RSS, GF, and WS liquids was 4.96, 3.66, and 5.29 mg/mL, respectively (Table 1), PC 2 was considered to be related to the loss of total amino acid content in the three sugar liquids.

Fig. 2

Principal component analysis score plot derived from the aroma compounds of three tsukudani seasoning liquids after heating.

○, experimental results with RSS; ☓, experimental results with GF; Δ, experimental results with WS; The numbers next to the legend indicate the number of independent experiments. In the figure, the horizontal axis is PC1, and the vertical axis is PC2.

Next, Fig. 3 shows the loading plot in the PCA to classify the compounds that determine the characteristics of the aroma profiles of each sugar liquid. The WS liquid was characterized by the score plots of many aldehydes, positioned in the positive area of PC 1 (Fig. 3). The GF liquid was characterized by the score plots of the S-containing compounds, positioned in the positive area of PC 2. On the other hand, the furans were located on the negative side of the loading plot of PC 1, and the position of the loading plots of the furans could be characterized for the RSS liquid. The furans in the RSS liquid are furfural, furfuryl alcohol, 3-phenylfuran, and 2,5-diethyl tetrahydrofuran. In particular, significantly more furfuryl alcohol and 3-phenylfuran were produced in the RSS liquid than in the GF and WS liquids (Table 2). Furthermore, 2-pentylfuran was positioned in the positive area of PC 2, corresponding to the GF liquid in the score plot (Fig. 2), and had more significant peak areas than the other compounds (Table 2).

Fig. 3

Principal component analysis loading scatter plot derived from the aroma compounds of the tsukudani seasoning liquids containing different sugars after heating.

×, alcohols; A, aldehydes; ▲, esters; ♦, acid; +, ketones; ●, furans; ○, N-compounds; ◊, S-compounds. The numbers next to the symbols correspond to the number of aroma compounds in Table 2. In the figure, the horizontal axis is PC1, and the vertical axis is PC2.

Fructose is known as a precursor of various aroma compounds in the Maillard reaction (Kraehenbuehl et al., 2010). Shibamoto (1989) examined the distribution of heterocyclic compounds identified in the Maillard reaction and reported that the highest number of furans was formed in the dehydration reaction of the basic sugar backbone. Limacher et al. (2008) reported that under the cooking conditions of roasting and pressure cooking at pH 4 and 7, fructose promoted the production of 2-methylfuran more than glucose. Shen et al. (2021) also reported that using fructose in strawberry jam produced about twice as much furan compared with glucose or sucrose. Previously, fructose was reported to promote furan production, but in this study, we confirmed that rare sugars contained in the RSS promoted more furan production than did fructose. Additionally, Cho et al. (2010) reported that glucose, galactose, fructose, and tagatose undergo Maillard reactions with several amino acids, that tagatose produces more furans than other monosaccharides, and that a sweet, fruity, caramellike aroma was formed in the reaction. In addition, Sawettanun and Ogawa (2022) reported that using allulose in bread produced more furans than did sucrose, resulting in characteristic roasted and caramelized odors.

Moreover, serine, cysteine, alanine, threonine, and aspartic acid have been reported as substrate amino acids for furan formation (Nie et al., 2013). However, no significant difference was observed for these amino acids in the seasonings with the respective sugars added, and only glutamic acid, a major amino acid, showed a significant difference in RSS against GF and WS, suggesting that the amino acid concerned had an effect on the formation of aroma. Fujimaki and Kurata (1971) reported that glutamic acid reacts with glucose, and furfural and 5-hydroxy methyl furfural are formed. Thus, the rare sugars in RSS and glutamic acid may also be closely involved in the formation of furan. These results suggest that using the rare sugars contained in the RSS in the tsukudani seasoning liquid results in the formation of a characteristic furan-like aroma during the cooking process.

Reactive amino acid in the GF liquid may be coupled and combined with GF sugars (total amino acid loss, 3.66 %, Table 1), but its loss is too small when compared with the large loss of sugars (60.52 %, Table 1). Because there was little color formation, melanoidin and flavor formations were weak. Therefore, aroma formation from the three seasoning liquids (RSS, GF, and WS) can be mainly regulated by two parameters, browning color and amino acid loss, during the cooking procedure, but those parameters are not directly related to aroma formation. Aroma formation may occur after multiple reactions of sugars and amino acids, and then the aroma is a kind of final product of the Maillard reaction.

In future research, we plan to conduct a more detailed study on the relationship between sensory evaluation and aroma components using odorant GC/MS, and to clarify the mechanism of formation for aroma components in foods using RSS containing rare sugars.

Conclusion

HS-GC/MS was used to analyze the aroma components of tsukudani seasoning liquids made with WS, GF, and RSS, and the data were subjected to PCA. The results showed that furans were formed as characteristic aroma components in the tsukudani seasoning liquid using RSS, suggesting that RSS has an aroma-enhancing effect.

Acknowledgements　We thank Shigehiko Onishi from the Kagawa Prefectural Industrial Technology Center for his advice on HS-GC/MS analysis and PCA.

Conflict of interest　There are no conflicts of interest to declare.

Abbreviations

GC/MS

gas chromatography/mass spectrometry

GF

glucose-fructose liquid sugar

HS-GC/MS

headspace-gas chromatography/mass spectrometry

HPLC

high-performance liquid chromatography

PCA

principal component analysis

RSS

rare sugar syrup

WS

white sugar

References

•  Aoyama,  O.,  Okazaki,  T., and  Yoneda,  T. (2000). Influence of water activity, yeast concentration and storage temperature on spoilage of konbu-tsukudani. Bulletin of Hiroshima Pref. Food Tech. Res. Ctr., 22, 7–12. (in Japanese)
•  Cho,  I.H.,  Lee,  S.,  Jun,  H.-R.,  Roh,  H.-J., and  Kim,  Y.-S. (2010). Comparison of volatile Maillard reaction products from tagatose and other reducing sugars with amino acids. Food Sci. Biotech., 19, 431–438.
•  Davies,  C.G.A., and  Labuza,  T.P. (1997). The Maillard reaction: application to confectionery products. Confect. Science, 33, 35–66.
•  Fujimaki,  M., and  Kurata,  T. (1971). Aroma of heated foods. Kagaku to Seibutsu, 9, 85–96. (in Japanese)
•  Hayase,  F., and  Watanabe,  H. (2014). Volatile components contributed to koku taste. J. Brew. Soc. Japan, 109, 161–167. (in Japanese)
•  Hayashi,  N.,  Yamada,  T.,  Takamine,  S.,  Iida,  T.,  Okuma,  K., and  Tokuda,  M. (2014). Weight reducing effect and safety evaluation of rare sugar syrup by a randomized double-blind, parallel-group study in human. J. Funct. Foods, 11, 152–159.
•  Horiguchi,  H. (1955). Fragrances dictionary (Translated from Japanese). Kyoritsu Shuppan Co., Ltd., Tokyo. (in Japanese)
•  Huang,  X.H.,  Zheng,  X.,  Chen,  Z.H.,  Zhang,  Y.Y.,  Du,  M.,  Dong,  X.P.,  Qin,  L., and  Zhu,  B.W. (2019). Fresh and grilled eel volatile fingerprinting by e-Nose, GC-O, GC-MS and GCxGC-QTOF combined with purge and trap and solvent-assisted flavor evaporation. Food Res., Int., 115, 32–43.
•  Jati,  M., and  Ariza,  B.T.S. (2011). Use of gas chromatography-olfactometry in combination with solid phase micro extraction for cocoa liquor aroma analysis. Int. Food Res. J., 18, 829–835.
•  John,  W. (2014). 3-(Methylthio)propionaldehyde. Perfumer & Flavorist, 39, 18–20.
•  Kraehenbuehl,  K.,  Davidek,  T.,  Devaud,  S., and  Mauroux,  O. (2010). Basic and acidic sugars as flavor precursors in the maillard reaction. Exp. of Multidisciplinary Flavour Science, 267–270 (https://www.researchgate.net/publication/236846371).
•  Limacher,  A.,  Kerler,  J.,  Davidek,  T.,  Schmalzried,  F., and  Blank,  I. (2008). Formation of furan and methylfuran by Maillard-type reactions in model systems and food. J. Agric. Food Chem., 56, 3639–3647.
•  Meng,  Q.,  Imamura,  M.,  Oubata,  A., and  Sugawara,  E. (2015). Changes in aromatic compounds in the seasoning prepared from soy sauce and mirin by cooking. J. Cookery Sci. Japan, 48, 249–254. (in Japanese)
•  Miyoshi,  M.,  Kimura,  I.,  Inazu,  T., and  Izumori,  K. (2019). High sensitivity analysis and food processing stability of rare sugars. Food Sci. Technol. Res., 25, 891–901.
•  Motai,  H. (1973). Color tone of various melanoidins produced from model systems. Agr. Biol. Chem., 37, 1679–1685.
•  Nie,  S.,  Huang,  J.,  Hu,  J.,  Zhang,  Y.,  Wang,  S.,  Li,  C.,  Marcone  M., and  Xie,  M. (2013). Effect of pH, temperature and heating time on the formation of furan in sugar-glycine model systems. Food Sci. Hum. Wellness, 2, 87–92.
•  Okumura,  J. (1993). The Maillard reaction and flavor formation. J. Brew. Soc. Japan, 88, 178–187. (in Japanese)
•  Ono,  T.,  Usami,  A.,  Nakaya,  S.,  Shinpuku,  H.,  Yonejima,  Y.,  Ikeda,  A., and  Miyazawa,  M. (2015). Agroecosystem development of industrial fermentation waste—characterization of aroma-active compounds from the cultivation medium of Lactobacillus brevis—. J. Oleo Sci., 64, 585–594.
•  Oshima,  H.,  Kimura,  I.,  Kitakubo,  Y.,  Hayakawa,  S., and  Izumori,  K. (2014). Factors affecting psicose formation in food products during cooking. Food Sci. Technol. Res., 20, 423–430.
•  Reyes,  F.G.R.,  Pooncharoen,  B., and  Wrolstad,  R.E. (1982). Maillard browning reaction of sugar-glycine model systems: changes in sugar concentration, color and appearance. J. Food Sci., 47, 1376–1377.
•  Sawettanun,  S., and  Ogawa,  M. (2022). Physicochemical parameters, volatile compounds and organoleptic properties of bread prepared with substituted sucrose with rare sugar D-allulose. Food Sci., Technol., 57, 5931–5942.
•  Shen,  M.,  Liu,  L.,  Zhang,  F.,  Jiang,  Y.,  Huang,  Z.,  Xie,  J.,  Zhong,  Y.,  Nie,  S., and  Xie,  M. (2021). Effects of processing parameters on furan formation in canned strawberry jam. Food Chem., 358.
•  Shibamoto,  T. (1989). Volatile flavor chemicals formed by the Maillard reaction. In “Thermal Generation of Aromas,” ACS Symp. Ser. 409 edited by  T.H.  Pariment,  R.J.  Mcgorrin, and  C.-T.  Ho. ACS, Washington, D.C. pp. 134–142.
•  Starowicz,  M., and  Zieliñski,  H. (2019). How Maillard reaction influences sensorial properties (color, flavor and texture) of food products? Food Rev. Int., 35, 707–725.
•  Tamura,  H.,  Izumori,  K., and  Tokuda,  M. (2008). Method for improving aromatic odor of food or drug medicine. Japan Kokai Tokkyo Koho, P2008–48685A, Mar. 6. (in Japanese)
•  Uchiyama,  A. (2020). Taste quality improvement effect of rare sugar syrup (Translated from Japanese). A Tech. J. Food Chem. & Chem., 7, 61–66. (in Japanese)
•  Yamada,  T.,  Shintani,  T.,  Iida,  T.,  Kishimoto,  Y., and  Okuma,  K. (2017). Effect of ingestion of rare sugar syrup on the blood glucose response in humans. J. Jpn. Soc. Nutr. Food. Sci., 70, 271–278. (in Japanese)
• i)  https://sugar.alic.go.jp/japan/user/user0810a.htm (Mar. 14, 2022)
• ii)  https://www.sigmaaldrich.com/JP/ja/product/aldrich/w269018 (May 23, 2023)
• iii)  http://www.thegoodscentscompany.com/index.html (May 23, 2023)