Abstract
The present study is to investigate the non-volatile and volatile profiles in radish according to the different oven roasting processing. In non-volatile compound profiles, different roasting temperatures (140–200 °C) and times (5, 10, 15, and 20 min) influenced non-volatile compounds in radishes, and high temperature roasted radish represented obvious changes than low temperature roasted radish. In volatile profiles, high temperature roasted radish were generated a higher number of Maillard reactions-related volatiles, including furfurals and 2-ethyl-5-methylpyrazine, than low temperature roasted radish. In chemometrics results, a radish roasted at 200 °C for 20 min was the highest dissimilarity compared with the other roasted radishes. This study is believed to be the first research demonstrating comprehensive identification of changes in non-volatile/volatiles profiles in radish by various processes (different times and temperatures) of oven roasting for food applications.
Keywords: Wintering radish, Chemosensory characteristics, E-tongue, E-nose, Multivariate analysis
Introduction
Radish has been cultivated in Korea, Japan, China, the United States, and Europe, and is classified as a Cruciferae plant. Radish, produced in Jeju island, is responsible for the dominant amount of cultivation in Korea and is used for Kimchi's main ingredients. Jeju radish is divided into four types, such as spring, summer, autumn, and wintering radishes, following the sowing and harvesting season (Hong et al., 2021a). Jeju wintering radish contains higher glucosinolates (GLS) contents compared with Europe and US radishes, and GLS is associated with nutritional attributes as well as flavor attributes (Kim et al., 2018, 2019a).
The roasting may change in volatile and non-volatile compound profiles. In general, thermal processing generates various volatile compounds and aroma perception in the heated samples, which are not detected in the raw radish samples. Although a number of volatiles are generated via Maillard reactions, only a few volatiles attributes aroma perception. Maillard reaction is mainly caused by the decomposition of fatty acids, amino acids, sugars, and lipids. Furthermore, Maillard reactions generally occur ranged from 140 to 230 °C (Kim et al., 2000; Yin et al., 2021a; Zhang et al., 2021). Previous research showed that volatile compounds, which are not detected in the raw radish sample, were detected in the radish through the oven roasting, and Maillard-related volatiles were generated. In addition, non-volatile compound profiles were also changed via oven roasting and other thermal processing (Hong et al., 2021a). Moreover, roasting temperature (180 °C) occurs in Maillard-related volatile compounds (Kim et al., 2019b).
To date, human sensory evaluation has been conducted for flavor characteristics in the food matrix. Flavor attribute is one of the important factors in the assessment of food quality (Yin et al., 2021a). Additionally, flavor attributes affect the overall acceptability and sensory cognition to choose the food matrix and are influenced by non-volatile compounds and volatile profiles (Hong et al., 2021a; Yin et al., 2021a). Even though it can provide the information about non-volatile compounds and aroma attributes, human sensory evaluation is necessary for expensive and time-consuming, and influenced by individual conditions of sensory panels. Accordingly, electronic sensors, which include electronic tongue (E-tongue) and electronic nose (E-nose), have been applied to analyze non-volatile and volatile profiles. Electronic sensors have advantages, such as rapid response, high sensitivity, low-cost, high productivity, and non-destructive method. Therefore, electronic sensors enable the evaluation of non-volatile compounds and volatile properties in various food samples. E-tongue can mimic the human perception of non-volatile compounds processing, and E-nose can mimic the human perception of the aroma precessing. Thus, these instruments could respond consistently to non-volatile compounds and volatile compounds similar to the results of human sensors. Furthermore, electronic sensors can distinguish the food samples which was difficult to be distinguished from human sensory evaluation. However, the limitation of electronic sensors is the relatively lower identification capacity of volatile compounds compared to GC–MS, and it has a lower perception capacity of complex taste and aroma compounds compared to human sensory analysis (Hong et al., 2022; Liu et al., 2017; Yin et al., 2021b; Zhang et al., 2021).
The aim of this work was to investigate the changes in non-volatile and volatile profiles in Jeju-wintering radishes using the chemosensory instruments (E-tongue and E-nose) via Maillard reaction through the low (140 °C) and high (180 °C) temperatures oven roasting.
Materials and methods
Materials
Jeju wintering radishes were purchased from a local grocery store (Jinju, Republic of Korea). The sample for E-tongue and E-nose analyses was used as slurry status using a plastic plate to prevent oxidation of radish by steel. During the oven roasting process, the slurry (30 g) was roasted at 140 °C for 5, 10, 15, and 20 min, respectively. Additionally, the slurry (30 g) was also roasted at 200 °C for 5, 10, 15, and 20 min, respectively, and these roasting processes were conducted using a complexed roasting device (EONC200F, SK Magic, Seoul, Korea). After each roasting process, non-volatile compounds and volatile profiles were analyzed (Hong et al., 2021a).
E-tongue analysis
To measure the non-volatile compound profiles, raw and oven-roasted radishes were analyzed using E-tongue system (ASTREE II, Alpha MOS, Toulouse, France). The radish slurry for E-tongue analysis was made by mixing between purified water (100 mL) and radish (30 g) for 30 min at 400 rpm in order to extract non-volatile compounds properties in radish. After the mixing process, radish extracts were used for E-tongue analysis. In the E-tongue system, E-tongue was composed of seven electronic sensors, including five sensors related to basic non-volatile compounds (SWS: saltiness, STS: sweetness, SRS: sourness, BRS: bitterness, UMS: umami) and two reference sensors (GPS: metallic, SPS: spiciness). Extracted radish sample was subjected to an E-tongue module and was repeated six times for each sample. The five basic non-volatile compound sensors were represented as sensor values of intensity from 1 to 12 and the general non-volatile compound distribution was given in a descriptive manner (Hong et al., 2021a).
E-nose analysis
To measure volatile profiles, raw and oven-roasted radishes were analyzed using E-nose system (HERACLES Neo, Alpha MOS, Toulouse, France). The radish slurry (1.3 g) for E-nose analysis was weighed in the headspace vial (22.5 × 75 mm, PTFE/silicone septum, aluminum cap) and volatiles in the headspace of vial was collected during stirring (500 rpm) at 50 °C for 20 min by the incubator in E-nose. After collecting the volatile in the headspace, volatile (1000 μL) were collected by an automatic sampler and into E-nose. In E-nose system, the flow rate of hydrogen gas was set to 1 mL/min, and flame ionization detectors (FID) and MTX-5 column (2 m × 0.18 mm) were used. Volatile compounds were analyzed under the condition of trap absorption temperature 40 °C, trap desorption temperature 250 °C, and acquisition time 227 s.
Identification of each compound was conducted using Kovat’s index (the retention index in terms of carbon number) library-based AroChembase (Alpha MOS), and volatile compounds corresponding to each peak were analyzed (Hong et al., 2021a).
Statistical analysis
In this study, all experiments were conducted in triplicate and the results were shown as mean ± standard deviation. Tukey’s multiple range test was examined to compare the means utilizing SAS 9.2 (Statistical Analysis System, Version 9.0, SAS Institute Inc., Cary, NC, USA). p < 0.05 was confirmed to be statistically significant differences between samples (p < 0.05). Furthermore, principal component analysis (PCA) was examined for non-volatile compounds and volatile profile data set using XLSTAT software ver 9.2 (Addinsoft, New York, NY, USA) to distinguish non-volatile compounds and volatile profiles, respectively.
Results and discussion
E-tongue analysis
Non-volatile compound profiles (SRS, UMS, STS, SWS, BRS) in raw and oven-roasted radishes were measured using E-tongue, and the results are shown in Fig. 1A, B. In the case of non-volatile compound profiles in raw and oven-roasted radishes heated at 140 °C (lower temperature), SRS was the highest value in raw radish among all samples, and SRS decreased under oven-roasting processes compared to raw radish. Among oven-roasted radishes, roasted for 20 min is higher SRS compared with other roasting times (5, 10, 15 min). Similar to SRS results, UMS was the highest value in raw radish among all samples, and oven-roasting induced decreasing UMS compared to raw radish. In addition, roasting for 20 min is higher UMS compared with other roasting times (5, 10, 15 min), these results were in accordance with SRS results in oven-roasted radishes. Unlike the results of SRS and UMS, STS was the lowest value among all samples, and STS increased after the oven-roasting process. Under oven-roasting conditions, oven-roasted for 5 and 15 min were lower STS compared with oven-roasted for 10 and 20 min. Similar to the results of STS, SWS and BRS were the lowest values in raw radishes. All oven-roasted radishes were higher SWS and BRS compared with raw radishes, regardless of roasting times. Under oven-roasting conditions, SWS and BRS in radishes decreased SWS and BRS in a time-dependent manner.
Fig. 1.
Radar plot for non-volatile profiles in radish at 140 °C (A) and at 200 °C (B) using electronic tongue technique
In the case of non-volatile compound profiles in raw and oven-roasted radishes heated at 200 °C (higher temperature), SRS was the highest value in raw radish among all samples. SRS decreased after the oven-roasting process and roasting for 15 min was higher compared with other roasting times. UMS also decreased after oven-roasting, and raw radish showed the highest value among all radishes. Under roasting conditions, UMS in oven-heated radishes increased from 5 to 15 min, however, UMS in radish roasted for 20 min was decreased. STS was the lowest in raw radish among all radishes, and STS increased upon oven-roasting. Among STS in oven-roasted radishes, roasted for 5 and 10 min radishes were higher values compared with roasted 15 and 20 min radishes. Similar to the results of STS, SWS in radishes increased after oven-roasting processing. On the other hand, there were no obvious variations of SWS during oven-roasting. In BRS, there were also no obvious variations during oven-roasting processing. Furthermore, there were obvious variations among raw and oven-roasted radishes.
E-tongue responds to electrical signals into non-volatile compounds signals to classify each non-volatile compound of foodstuffs and this instrument contains the sensor array utilized to measure non-volatile compounds profiles (Hong et al., 2022). Furthermore, E-tongue mimics the human tongue system and processing of non-volatile compounds perception. Therefore, the results of E-tongue analysis showed a similar tendency compared to human sensory analysis (Liu et al., 2017; Phat et al., 2016). Based on the non-volatile compound profiles, oven roasting processing showed a similarity between roasting at 140 °C and 200 °C. In the case of SRS, STS, and UMS, slight variations were identified during oven roasting at 140 °C, however, obvious variations were identified during oven roasting at 200 °C compared with oven roasting processing at 140 °C. A previous study reported that SRS and UMS sensory profiles in roasted samples decreased via roasting processing, and our results are in accordance with previous research data (Lee et al., 2019). Another previous study reported that STS sensory profiles in thermally processed samples decreased during thermal processing, and our results are in accordance with previous research data (Zhang et al., 2021). Furthermore, roasting processes with different temperatures can induce variations of non-volatile compound profiles based on E-tongue data, and roasting at a higher temperature may contribute to more variations compared to roasting at a lower temperature (Kim et al., 2019b).
PCA-biplot and cluster analysis of non-volatile compound profiles using E-tongue in raw and oven-roasted Jeju wintering radishes
PCA-biplot and cluster analysis (CA) were conducted based on non-volatile compound profiles using E-tongue and the results are represented in Fig. 2A, B. In PCA-biplot (Fig. 2A), PC1 has shown 75.66% variance and PC2 has shown 17.66% variance. Thus, a total of PCs have shown 93.32% variances. Among raw, lower temperature (140 °C) oven-roasted radishes, and higher temperature (200 °C) oven-roasted radishes, most samples were mainly isolated by the axis of PC1. In addition, most of the oven-roasted radishes were located on the negative axis of PC1 by SWS and STS, except for radishes (roasted at 140 °C for 5 and 10 min), however, raw radish and radishes roasted at 200 °C for 15 and 20 min were located on the positive axis of PC1 by SRS and UMS. Unlike PC1 results, raw radish and radishes roasted at 140 °C for 5 and 10 min were located on the positive axis of PC2, however, most oven-roasted radishes were located on the negative axis of PC2. In radishes roasted at 140 °C, radishes have shifted to the negative axis of PC2 and were higher influenced by SWS and STS according to the oven roasting times. In radishes roasted at 200 °C, radishes roasted for 5 and 10 min were mainly influenced by SWS and STS, however, radishes roasted for 15 and 20 min were mainly influenced by SRS and UMS. In CA results (Fig. 2B), a total of four clusters were represented. Raw radish was classified as cluster I, and radishes roasted at 140 °C for 5 and 10 min were classified as cluster II. Radishes roasted at 200 °C for 15 and 20 min were classified as cluster III, and the rest radishes were classified as cluster IV.
Fig. 2.
PCA plot (A) and dendrogram (B) for non-volatile profiles in radish using electronic tongue technique
PCA-biplot and CA are widely used to distinguish and classify foodstuffs. PCA-biplot is commonly used in chemometric research and is considered as the multivariate modeling and analysis method (Shin et al., 2010). A previous study reported that radish was measured using E-tongue and PCA-biplot and CA were used to classify between raw and thermally processed radishes (Hong et al., 2021a). Another previous study reported that cooked food samples were measured using E-tongue and the samples were classified using PCA (Zhang et al., 2021). Furthermore, the variations of non-volatile compound profiles in roasted coffee beans were identified using E-tongue, and these samples were classified using PCA and CA (Dong et al., 2019).
E-nose analysis
Volatile profiles in raw and oven-roasted (140 °C) radishes were measured using E-nose and these results are shown in Table 1. A total of 13 volatile compounds were detected, including 3 sulfur-containing compounds, 1 heterocyclic compound, 3 acids and esters, 2 aldehydes, and 4 ketones. In the sulfur-containing compounds among raw and oven-roasted (140 °C) radishes, raw radish only contained methanethiol and this compound was the dominant amount in the raw one compared with oven-roasted (140 °C) radishes. In particular, oven-roasting (140 °C) induced significantly decreased contents (p < 0.05), and methanethiol was not found in the oven-roasted radish (20 min). Otherwise, ethanethiol and carbon disulfide were found in the oven-roasted radish (15 min), however, the oven-roasted radish (20 min) only contained carbon disulfide. Carbon disulfide was generated from 15 min, and oven roasting (20 min) significantly increased the amount of this volatile compound (p < 0.05).
Table 1.
Volatile profiles in radish according to the oven roasting time at 140 °C using E-nose (Peak area × 103)
| Compounds | RT(s) (RI) | Sensory description | Unheated | 140 °C | |||
|---|---|---|---|---|---|---|---|
| MXT-5 | 5 min | 10 min | 15 min | 20 min | |||
| Sulfur-containing compounds(3) | |||||||
| Methanethiol | 15.73 (451) | Cabbage, Garlic, Sulfurous | 13.72 ± 0.80a | 0.21 ± 0.01b | 0.16 ± 0.01b | 0.36 ± 0.04b | ND |
| Ethanethiol | 18.61 (516) | Garlic, Onion, Sulfurous | ND | ND | ND | 0.33 ± 0.02 | ND |
| Carbon disulfide | 20.23 (552) | Burnt, Sweet, Sulfurous | ND | ND | ND | 0.22 ± 0.01b | 1.38 ± 0.08a |
| Heterocyclic compound(1) | |||||||
| 2-Methylfuran | 22.42 (600) | Burnt, Sweet | ND | 0.09 ± 0.02a | 0.10 ± 0.02a | 0.13 ± 0.01a | ND |
| Acids and esters(3) | |||||||
| Ethyl isobutyrate | 37.77 (754) | Fruity, Sweet | 13.02 ± 0.72a | 0.87 ± 0.14b | 0.83 ± 0.01b | 1.53 ± 0.04b | 0.72 ± 0.04b |
| Pentanoic acid | 53.83 (901) | Cheese, Putrid | 0.08 ± 0.01 | ND | ND | ND | ND |
| Decyl acetate | 85.77 (1407) | Fresh, Fruity, Oily | 0.06 ± 0.01 | ND | ND | ND | ND |
| Aldehydes(2) | |||||||
| 2-Methylbutanal | 28.21 (665) | Burnt, Almond, Green | ND | ND | ND | 0.30 ± 0.01b | 1.35 ± 0.10a |
| Nonanal | 69.74 (1110) | Fruity, Sweet, Citrus | ND | 0.06 ± 0.01b | 0.07 ± 0.01ab | 0.08 ± 0.01a | 0.08 ± 0.01a |
| Ketones(4) | |||||||
| Propan-2-one | 17.40 (489) | Fruity, Sweet | ND | 0.12 ± 0.01ab | 0.13 ± 0.01ab | 0.11 ± 0.02b | 0.15 ± 0.01a |
| 1-Hydroxy-2-propanone | 27.13 (653) | Pungent, Sweet, Caramelized | ND | ND | ND | ND | 0.44 ± 0.03 |
| 2-Octanone | 61.41 (991) | Floral, Fruity, Green | 0.09 ± 0.01 | ND | ND | ND | ND |
| δ-Octalactone | 79.82 (1287) | Fatty, Sweet | ND | 0.17 ± 0.06a | 0.16 ± 0.07a | 0.16 ± 0.04a | 0.17 ± 0.07a |
RT retention time, RI retention indices, ND not detected
Means with different letters(a–d) are significantly different by Tukey’s multiple range test (p < .05)
Data represent the mean ± SD in triplicate
In the heterocyclic compound (1) among raw and oven-roasted (140 °C) radishes, 2-methylfuran was only found in radishes roasted for 5, 10, and 15 min. This volatile compound was not found in the raw radish, however, 2-methylfuran may be generated via oven roasting processes. On the contrary, radish roasted for 20 min has not contained 2-methylfuran.
Except for sulfur-containing compounds and heterocyclic compound, oven-roasting (140 °C) induced decomposition of volatile compounds and/or generation of volatile compounds. Especially, raw radish had dominant content of ethyl isobutyrate, while ethyl isobutyrate significantly decreased through the oven roasting process (p < 0.05). Furthermore, some volatile compounds were generated via oven roasting processing. For instance, 2-methylbutanal, propane-2-one, and δ-octalactone were generated after oven roasting processing.
Volatile profiles in raw and oven-roasted (200 °C) radishes were measured using E-nose and these results are shown in Table 2. A total of 22 volatile compounds were detected, including 3 sulfur-containing compounds, 4 heterocyclic compounds, 5 acids and esters, 1 alcohol, 4 aldehydes, and 5 ketones. In the sulfur-containing compounds among raw and oven-roasted (200 °C) radishes, methanethiol in raw radish was the dominant content compared with the others (p < 0.05). This volatile compound significantly decreased during the oven roasting processes and these results are in accordance with the variations of sulfur-containing compounds during the oven roasting (140 °C). Similar to variations of sulfur-containing compounds during the oven roasting (140 °C), ethanethiol and carbon disulfide were also found after oven roasting (200 °C). These volatiles were found at 10 min and increased at 15 min (p < 0.05). In the case of ethanethiol, this volatile compound was not found at 20 min, however, carbon disulfide was found at 20 min. In general, sulfur-containing compounds were mainly generated from GLS decomposition and/or degradation. GLS decomposition and degradation were generally caused by the activation of an enzyme (myrosinase), external energy, and thermal energy (Hong et al., 2022; Oerlemans et al., 2006). In particular, a previous study reported that thermal processing results in GLS decomposition, and thus sulfur-containing compounds were generated and reduced (Blažević and Mastelić, 2019). Sulfur-containing compounds are mainly attributed to the aroma activation of Brassicaceae vegetables, such as radish and cabbage (Hong et al., 2021b; Oerlemans et al., 2006). Previous studies reported that methanethiol has a low threshold (Jia et al., 2019) and the amount of methanethiol in the Brassicaceae vegetable decreased after oven roasting and other roasting processes. Besides, this volatile compound could be converted to other sulfur-containing compounds (Hong et al., 2022). In this study, methanethiol also decreased via roasting processing, and sulfur-containing compounds, which are ethanethiol and carbon disulfide, were generated via oven-roasting. Ethanethiol has a high odor threshold and was found in the Jeju-wintering radish using E-nose (Cannon and Ho, 2018; Hong et al., 2021b) Carbon disulfide is known as a radish-like aroma and was found in the radish sample using E-nose (Sung et al., 2016). In addition, carbon disulfide has a low odor threshold (Cheng et al., 2020).
Table 2.
Volatile profiles in radish according to the oven roasting time at 200 °C using E-nose (Peak area × 103)
| Compounds | RT(s) (RI) | Sensory description | Unheated | 200 °C | |||
|---|---|---|---|---|---|---|---|
| MXT-5 | 5 min | 10 min | 15 min | 20 min | |||
| Sulfur-containing compounds(3) | |||||||
| Methanethiol | 15.73 (451) | Cabbage, Garlic, Sulfurous | 13.72 ± 0.80a | 1.22 ± 0.04b | 0.06 ± 0.01c | 0.10 ± 0.01c | 0.23 ± 0.01c |
| Ethanethiol | 18.61 (516) | Garlic, Onion, Sulfurous | ND | ND | 0.02 ± 0.04b | 0.31 ± 0.03a | ND |
| Carbon disulfide | 20.23 (552) | Burnt, Sweet, Sulfurous | ND | ND | 0.11 ± 0.02c | 2.71 ± 0.05a | 0.61 ± 0.01b |
| Heterocyclic compounds(4) | |||||||
| 2-Methylfuran | 22.42 (600) | Burnt, Sweet | ND | 0.10 ± 0.01b | 0.10 ± 0.01b | ND | 1.85 ± 0.07a |
| Furfural | 47.55 (841) | Baked, Almond, Sweet | ND | ND | ND | 0.22 ± 0.01b | 0.30 ± 0.01a |
| 5-Methylfurfural | 60.02 (974) | Almond, Burnt, Caramelized | ND | ND | ND | ND | 0.17 ± 0.01 |
| 2-Ethyl-5-methylpyrazine | 63.18 (1014) | Fruity, Nutty, Sweet | ND | ND | ND | ND | 0.04 ± 0.03 |
| Acids and esters(5) | |||||||
| Methyl propanoate | 24.99 (629) | Apple, Fresh, Sweet | ND | ND | ND | 0.07 ± 0.01b | 0.11 ± 0.01a |
| Ethyl isobutyrate | 37.77 (754) | Fruity, Sweet | 13.02 ± 0.72a | 1.84 ± 0.08b | 0.18 ± 0.01c | 1.57 ± 0.07b | 0.46 ± 0.04c |
| Pentanoic acid | 53.83 (901) | Cheese, Putrid | 0.08 ± 0.01 | ND | ND | ND | ND |
| Hexanoic acid | 61.35 (990) | Cheese, Sour, Sweaty | ND | ND | ND | 0.05 ± 0.05a | 0.04 ± 0.03a |
| Decyl acetate | 85.77 (1407) | Fresh, Fruity, Oily | 0.06 ± 0.01 | ND | ND | ND | ND |
| Alcohol(1) | |||||||
| Pentan-2-ol | 30.79 (694) | Nutty, Sweet | ND | ND | ND | ND | 0.11 ± 0.01 |
| Aldehydes(4) | |||||||
| 2-Methylpropanal | 18.67 (517) | Burnt, Malty, Bread | ND | ND | ND | ND | 0.75 ± 0.01 |
| Butanal | 21.73 (585) | Green, Malty, Pungent | ND | ND | ND | ND | 0.07 ± 0.01 |
| 2-Methylbutanal | 28.21 (665) | Burnt, Almond, Green | ND | ND | 0.21 ± 0.06c | 4.27 ± 0.06a | 1.03 ± 0.01b |
| Nonanal | 69.74 (1110) | Fruity, Sweet, Citrus | ND | 0.07 ± 0.01a | 0.08 ± 0.01a | 0.08 ± 0.01a | 0.08 ± 0.01a |
| Ketones(5) | |||||||
| Propan-2-one | 17.40 (489) | Fruity, Sweet | ND | 0.08 ± 0.01c | 0.08 ± 0.01c | 0.26 ± 0.02b | 1.06 ± 0.04a |
| 1-Hydroxy-2-propanone | 27.13 (653) | Pungent, Sweet, Caramelized | ND | ND | ND | 0.73 ± 0.02a | 0.15 ± 0.01b |
| 2,3-Pentanedione | 32.53 (710) | Almond, Burnt, Sweet | ND | ND | ND | ND | 0.36 ± 0.02 |
| 2-Octanone | 61.41 (991) | Floral, Fruity, Green | 0.09 ± 0.01 | ND | ND | ND | ND |
| δ-Octalactone | 79.82 (1287) | Fatty, Sweet | ND | 0.15 ± 0.06a | 0.16 ± 0.08a | 0.16 ± 0.05a | 0.16 ± 0.06a |
RT retention time, RI retention indices, ND not detected
Data represent the mean ± SD in triplicate
Means with different letters(a–d) are significantly different by Tukey’s multiple range test (p < .05)
In the heterocyclic compounds (4), 2-methylfuran was found after oven roasting processes, and this result is in accordance with oven roasting at 140 °C. Unlike the oven roasting at 140 °C, 2-methylfuran was found at 20 min, and a significantly increased amount. Except for 2-methylfuran, rest volatiles were only found during oven roasting at 200 °C. Furfural was found after oven roasting at 15 min, this volatile compound increased in a time-dependent manner (p < 0.05). 5-Methylfurfural and 2-ethyl-5-methylpyrazine were only found at 20 min. Heterocyclic compounds are commonly produced via Maillard reactions, which induce reducing sugars, lipids, amino acids, and fatty acids (Farmer et al., 1989; Lee et al., 2019). In addition, heterocyclic compounds generally contribute to distinct odor activation and description (Hong et al., 2022; Yin et al., 2021a). The results of volatile profiles in this study indicated that 2-methylfuran was generated via two kinds of oven roasting steps (140 and 200 °C). On the contrary, furfural, 5-methylfurfural, and 2-ethyl-5-methylpyrazine were only generated via higher temperature oven roasting (200 °C). Furans are commonly formed by the thermal decomposition of sugar polymers and D-glucose (Yang et al., 2016). Furthermore, furans also can be generated from the oxidation of lipids, which are from the decomposition of thiamine and nucleotides breakdown, via thermal treatment (Lee et al., 2019; Mottram, 2007; Yang et al., 2016;). Furfural is commonly considered as one of the Amadori rearrangement products, and this volatile compound is mainly formed from deoxyosones. Additionally, furfurals can be generated from the oxidation of furfuryl alcohol, and 5-methylfurfural is also generated and increased from these reactions (Caporaso et al., 2018). Total furan concentrations, which include furans and furan derivatives, may be associated with roasting degree (Caporaso et al., 2018; Yang et al., 2016). Pyrazines are commonly associated with Strecker reactions, which are reacted with aminoketones and aldehydes, and these reactions-related products commonly contribute to various aroma activation (Caporaso et al., 2018). A previous study identified that higher temperature generated more heterocyclic compounds and odor activation (Kim et al., 2019b), and another study identified that most of the furans and furfurals increased in a cooking time-dependent manner (Zhang et al., 2021). Our results also indicated that roasting at 200 °C for 20 min had the highest amount of furan, furfurals, and pyrazine. Thus, Maillard reactions in radishes are easily likely to occur at 200 °C compared to 140 °C.
Except for sulfur-containing compounds and heterocyclic compounds, oven-roasting (200 °C) generated volatile compounds, and some volatile were decomposed. Volatiles were generated via oven roasting, and these compounds were found in alcohol, aldehydes, ketones, and acids and esters. Among generated volatiles, 2-methylbutanal at 15 min was the dominant amount compared with other generated volatile compounds via oven roasting. The decomposition of volatile compounds was only identified in acids and esters, except for alcohol, aldehydes, and ketones. Ethyl isobutyrate in raw radish was the dominant amount among all volatiles, however, the amount of this compound decreased via oven roasting processes. These results are similar to the variations during oven roasting processes (140 °C).
Multivariate analysis of volatile profiles using E-nose in raw and oven-roasted Jeju wintering radishes
PCA-biplot and CA were conducted based on volatile profiles using E-tongue and the results are represented in Fig. 3A, B. In PCA-biplot (Fig. 3A), PC1 has shown 48.84% variance, and PC2 has shown 29.83% variance. Thus, a total of PCs have shown 78.67% variances. PC1 (48.84%) has a higher influence on the separation of radishes than PC2 (29.83%). Most radishes, which include raw radish, were located on the negative axis of PC1, however, radishes roasted at 200 °C for 15 and 20 min were located on the positive axis of PC1. Unlike the results of PC1, most of the oven-roasted radishes were located on the positive axis of PC2, however, raw and oven-roasted radishes (200 °C, 20 min) were located on the negative axis of PC2. These results indicated that oven roasting induced a shift to the upper location (positive axis of PC2) while oven roasting (200 °C) over 15 min induced a shift to the bottom location (negative axis of PC2). Moreover, oven roasting (200 °C) also induced a shift to the right location according to the roasting times. In detail, raw radish was mainly influenced by ethyl isobutyrate, 2-octane, methanethiol, decyl acetate, and pentanoic acid, however, most of the oven-roasted radishes, except for radish roasted at 200 °C for 20 min, were mainly and/or slightly influenced by δ-octalactone and nonanal. Meanwhile, radish roasted at 200 °C for 20 min was mainly influenced by furfural, methyl propanoate, propan-2-one, 2-methyl furan, pentan-2-ol, 2-methylpropanal, butanal, 5-methylfurfural, 2,3-pentanedione, and 2-ethyl-5-methylpyrazine, and slightly influenced by hexanoic acid. In CA results (Fig. 3B), a total of four clusters were represented. The radish roasted at 200 °C for 20 min was classified as cluster I, and the raw radish was classified as cluster II. The radish roasted at 200 °C for 15 min was classified as cluster III, and the rest radishes were classified as cluster IV. Similar to the results of PCA-biplot and CA based on non-volatile compounds profiles, radishes roasted at 200 °C for 15 and 20 min were separated by the other oven-roasted radishes.
Fig. 3.
PCA plot (A) and dendrogram (B) for non-volatile profiles in radish using electronic nose technique
In conclusion, non-volatile and volatile compound profiles in Jeju wintering radishes, which were oven-roasted under different temperatures and times, were identified using E-tongue and E-nose. In addition, the results of non-volatile compounds and volatile compounds in raw and oven-roasted radishes were distinguished by chemometrics (PCA and CA). For the E-tongue analysis, non-volatile compound profiles were changed during oven roasting processes, and thus five non-volatile compounds (SWS, STS, SRS, UMS, BRS) were influenced and changed. Between two kinds of roasting temperatures (140 °C and 200 °C), all temperatures of oven roasting increased STS and SWS compared to raw one and decreased SRS and UMS. In particular, BRS decreased during oven roasting (140 °C), however, oven roasting (200 °C) showed no obvious variations of BRS compared to lower temperature of oven roasting. On the other hand, four non-volatile compounds (SWS, STS, SRS, UMS) showed obvious variations during the higher temperature of oven roasting compared with a lower temperature of oven roasting. For the E-nose analysis, all roasting processes showed that methanethiol was decreased, and ethanethiol and carbon disulfide were generated. Thus, sulfur-containing compounds showed similar tendencies between lower and higher temperatures of oven roasting. In contrast, Maillard reactions-related volatiles were mainly generated via the higher temperature of oven roasting. For the PCA and CA, four non-volatile compounds (SWS, STS, SRS, UMS) mainly influenced all radishes, and thus four clusters were represented and identified. In the case of PCA and CA based on the results of E-nose, raw and oven-roasted radishes were obviously separated and oven-roasted radish at 200 °C for 20 min showed the obvious variation among all oven-roasted radishes, regardless of roasting temperatures. Therefore, radish roasted at 200 °C for 20 min was classified as cluster I, and the rest radishes were classified as other clusters. Accordingly, this study identified the variations of the non-volatile compound and volatile compound profiles according to the different roasting temperatures and times, and further investigation seems to be necessary for optimal oven roasting conditions and physiological effects in non-volatile compounds and volatile profiles in roasted radishes by response surface methodology and clinical trials, respectively. Additionally, the results of this work on chemosensory characteristics (non-volatile and volatile compounds) can provide the basic research data (roasting conditions and thermal effects on the formation of volatile and non-volatile profiles in Jeju wintering radish) to academic circles and/or industry,
Acknowledgements
This research was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education (NRF-2022R1I1A3066192).
Declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Seong Jun Hong and Seong Min Jo have contributed equally to this work.
Contributor Information
Seong Jun Hong, Email: 01028287383a@gmail.com.
Seong Min Jo, Email: jojo9875@naver.com.
Sojeong Yoon, Email: dbsthwjd0126@naver.com.
Hyangyeon Jeong, Email: giddus9967@naver.com.
Youngseung Lee, Email: youngslee@dankook.ac.kr.
Sung-Soo Park, Email: foodpark@jejunu.ac.kr.
Eui-Cheol Shin, Email: eshin@gnu.ac.kr.
References
- Blažević I, Mastelić J. Glucosinolate degradation products and other bound and free volatiles in the leaves and roots of radish (Raphanus sativus L.) Food Chemistry. 2019;113(1):96–102. doi: 10.1016/j.foodchem.2008.07.029. [DOI] [Google Scholar]
- Cannon RJ, Ho CT. Volatile sulfur compounds in tropical fruits. Journal of Food and Drug Analysis. 2018;26:445–468. doi: 10.1016/j.jfda.2018.01.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Caporaso N, Whitworth MB, Cui C, Fisk ID. Variability of single bean coffee volatile compounds of Arabica and robusta roasted coffees analyzed by SPME-GC-MS. Food Research International. 2018;108:628–640. doi: 10.1016/j.foodres.2018.03.077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cheng Y, Rouseff R, Li G, Wu H. Methanethiol, an off-flavor produced from the thermal treatment of mandarin juices: a study of citrus sulfur volatiles. Journal of Agricultural and Food Chemistry. 2020;68:1030–1037. doi: 10.1021/acs.jafc.9b06647. [DOI] [PubMed] [Google Scholar]
- Dong W, Hu R, Long Y, Li H, Zhang Y, Zhu K, Chu Z. Comparative evaluation of the volatile profiles and non-volatile compounds properties of roasted coffee beans as affected by drying method and detected by electronic nose, electronic tongue, and HS-SPME-GC-MS. Food Chemistry. 2019;272:723–731. doi: 10.1016/j.foodchem.2018.08.068. [DOI] [PubMed] [Google Scholar]
- Farmer LJ, Mottram DS, Whitfield FB. Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid. Journal of the Science of Food and Agriculture. 1989;49:347–368. doi: 10.1002/jsfa.2740490311. [DOI] [Google Scholar]
- Hong SJ, Boo CG, Lee J, Hur SW, Jo SM, Jeong H, Yoon SJ, Lee Y, Park SS, Shin EC. Chemosensory approach supported-analysis of wintering radishes produced in Jeju island by different processing methods. Food Science and Biotechnology. 2021;30:1033–1049. doi: 10.1007/s10068-021-00948-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hong SJ, Boo CG, Heo SU, Jo SM, Jeong H, Yoon S, Shin EC. Investigation of non-volatile compounds and flavor properties of radish varieties harvested in Korea using electronic tongue and electronic nose. Korean Journal of Food Science and Technology. 2021;53:375–381. [Google Scholar]
- Hong SJ, Yoon S, Lee J, Jo SM, Jeong H, Lee Y, Park SS, Shin EC. A comprehensive study for non-volatile compounds and odor characteristics using electronic sensors in broccoli floret with different methods of thermal processing. Journal of Food Processing and Preservation. 2022;2022:e16435. [Google Scholar]
- Jia X, Zhou Q, Wang J, Liu C, Huang F, Huang Y. Identification of key aroma-active compounds in sesame oil from microwaved seeds using E-nose and HS-SPME-GC×GC-TOF/MS. Journal of Food Biochemistry. 2019;43(10):e12786. doi: 10.1111/jfbc.12786. [DOI] [PubMed] [Google Scholar]
- Kim HW, Park KM, Choi CU. Studies on the volatile flavor compounds of sesame oils with roasting temperature. Korean Journal of Food Science and Technology. 2000;32:238–245. [Google Scholar]
- Kim MK, Lee MA, Lee KG. Determination of compositional quality and volatile flavor characteristics of radish-based Kimchi suitable for Chinese consumers and its correlation to consumer acceptability. Food Science and Biotechnology. 2018;27:1265–1273. doi: 10.1007/s10068-018-0387-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim B, Cho YJ, Kim M, Hurh B, Baek HH. Changes in volatile flavor compounds of radish fermented by lactic acid bacteria. Korean Journal of Food Science and Technology. 2019;51:24–329. [Google Scholar]
- Kim DS, Lee JT, Hong SJ, Cho JJ, Shin EC. Thermal coursed effect of comprehensive changes in the flavor/non-volatile compounds of Cynanchi wilfordii. Journal of Food Science. 2019;84(10):2831–2839. doi: 10.1111/1750-3841.14797. [DOI] [PubMed] [Google Scholar]
- Lee J, Kim DS, Cho J, Hong SJ, Pan JH, Kim JK, Shin EC. Perilla frutescens Britton: a comprehensive study on flavor/non-volatile compounds and chemical properties during the roasting process. Molecules. 2019;24:1374. doi: 10.3390/molecules24071374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu D, Li S, Wang N, Deng Y, Sha L, Gai S, Liu H. Xu XEvolution of non-volatile compounds compounds of Dezhou-braised chicken during cooking evaluated by chemical analysis and an electronic tongue system. Journal of Food Science. 2017;82:1076–1082. doi: 10.1111/1750-3841.13693. [DOI] [PubMed] [Google Scholar]
- Mottram DS. Flavours and fragrances. Berlin: Springer; 2007. The Maillard reaction: source of flavour in thermally processed foods; pp. 269–283. [Google Scholar]
- Oerlemans K, Barrett DM, Suades CB, Verkerk R, Dekker M. Thermal degradation of glucosinolates in red cabbage. Food Chemistry. 2006;95(1):19–29. doi: 10.1016/j.foodchem.2004.12.013. [DOI] [Google Scholar]
- Phat C, Moon B, Lee C. Evaluation of umami non-volatile compounds in mushroom extracts by chemical analysis, sensory evaluation, and an electronic tongue system. Food Chemistry. 2016;192:1068–1077. doi: 10.1016/j.foodchem.2015.07.113. [DOI] [PubMed] [Google Scholar]
- Shin EC, Craft BD, Pegg RB, Phillips RD, Eitenmiller RR. Chemometric approach to fatty acid profiles in Runner-type peanut cultivars by principal component analysis (PCA) Food Chemistry. 2010;119:1262–1270. doi: 10.1016/j.foodchem.2009.07.058. [DOI] [Google Scholar]
- Sung NY, Park WY, Kim YE, Cho EJ, Song HY, Jun HK, Park JN, Kim MH, Rhy GH, Byun EH. Increase in anti-oxidant components and reduction of off-flavors on radish leaf extracts by extrusion process. Journal of the Korean Society of Food Science and Nutrition. 2016;45(12):1769–1775. doi: 10.3746/jkfn.2016.45.12.1769. [DOI] [Google Scholar]
- Yang N, Liu C, Liu X, Degn TK, Munchow M, Fisk I. Determination of volatile marker compounds of common coffee roast defects. Food Chemistry. 2016;211:206–214. doi: 10.1016/j.foodchem.2016.04.124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yin X, Wen R, Sun F, Wang Y, Kong B, Chen Q. Collaborative analysis on differences in volatile compounds of Harbin red sausages smoked with different types of woodchips based on gas chromatography–mass spectrometry combined with electronic nose. LWT. 2021;143:111144. doi: 10.1016/j.lwt.2021.111144. [DOI] [Google Scholar]
- Yin X, Lv Y, Wen R, Wang Y, Chen Q, Kong B. Characterization of selected Harbin red sausages on the basis of their flavour profiles using HS-SPME-GC/MS combined with electronic nose and electronic tongue. Meat Science. 2021;172:108345. doi: 10.1016/j.meatsci.2020.108345. [DOI] [PubMed] [Google Scholar]
- Zhang L, Hu Y, Wang Y, Kong B, Chen Q. Evaluation of the flavour properties of cooked chicken drumsticks as affected by sugar smoking times using an electronic nose, electronic tongue, and HS-SPME/GC-MS. LWT. 2021;140:110764. doi: 10.1016/j.lwt.2020.110764. [DOI] [Google Scholar]