-
As an important intelligent sensing technology, E-Nose can identify and judge different samples utilizing the rapid response characteristics of different types of sensors to specific volatile compounds. However, the sensor is prone to sample contamination, resulting in reduced sensor response. Compared with traditional E-Nose, the GC-E-Nose realizes the rapid separation of volatile compounds by combining MXT-1701 (medium polar) and MXT-5 (non-polar) columns, and qualitative characterization of volatile compounds by assembling the FID detectors.
In the study, the radar fingerprints of GE and GET were shown in Supplemental Fig. S1. Each peak represents a volatile compound, the peak area indicates the content of a volatile component, and the higher the peak area, the higher the content. It could be found that the number of peaks on two chromatographic columns in GE was more than that in GET. Moreover, there were more peaks on the MXT-5 column than on the MXT-1701 column. As an unsupervised statistical method, PCA is the most widely used data dimensionality reduction algorithm. In this study, PCA was adopted to analyze the volatile fingerprints of GE and GET. As shown in Supplemental Fig. S2, obvious spatial distribution characteristics was observed between GE and GET. The contribution rate of PC1 reached 77.31%, the contribution rate of PC2 reached 9.43%, and the cumulative contribution rate reached 86.74%. The above results showed that GC-E-Nose combined with PCA could realize the difference comparison and rapid differentiation between GE and GET.
Characterization of volatile compounds in GE and GET by GC × GC-TOFMS
Identification of volatile compounds in GE and GET
-
GC × GC-TOFMS, which combines the advantages of two-column amplification of peak capacity and high resolution of TOF-MS, is one of the most important analytical technologies in food flavor[19,20]. The total ion chromatograms of GE and GET were shown in Supplemental Fig. S3. Through the spectrum library search and retention index authentication, a total of 188 volatile compounds (including 25 alcohols, 21 aldehydes, 10 alkanes, 31 alkenes, 21 aromatic compounds, 57 esters, nine heterocyclic compounds, 11 ketones, and three acids) were identified in GE, while 202 volatile compounds (including 26 alcohols, 20 aldehydes, 10 alkanes, 32 alkenes, 21 aromatic compounds, 66 esters, 10 heterocyclic compounds, 15 ketones, and two acids) were identified in GET by GC × GC-TOFMS (Fig. 1a & Supplemental Table S1). Esters, alkenes and alcohols were the dominant categories in both samples, which accounted for 32.67%, 15.84% and 12.87% in GET, and 30.32%, 16.49% and 13.30% in GE, respectively (Fig. 1b). It is worth mentioning that 179 common volatile compounds were detected in both GE and GET (Fig. 1c).
Figure 1.
Information of volatile compounds in GE and GET identified by GC × GC-TOFMS. (a) Specific number of volatile categories in GE and GET. (b) Proportion of volatile categories in GE and GET. (c) Venn diagram of volatile compounds in GE and GET.
Esters were the main volatile compounds, showing the highest contents in GE (13,138.64 μg/L) and GET (19,956.62 μg/L). Among them, hexyl tiglate was the most abundant compound in GE (3,422.36 μg/L) and GET (3,178.78 μg/L). Benzoic acid, methyl ester also presented a higher content in GE (3,195.76 μg/L) and GET (3,549.00 μg/L). Moreover, a total of 37 important esters showed significant differences (p < 0.05) in GE and GET. For example, octanoic acid, methyl ester, 5-hexyldihydro-2(3H)-furanone and decanoic acid, methyl ester presented higher contents in GE than those in GET, while methyl tiglate, ethyl tiglate, (Z)-3-hexen-1-ol, propanoate and (E)-butanoic acid, 3-hexenyl ester showed the opposite trend.
Alkenes were the second largest category after esters. As shown in Fig. 2, the content of alkenes in GE (7,690.78 μg/L) was significantly higher than that in GET (3,143.27 μg/L) (p < 0.05). Among them, the content of theaspiran in GE (2,542.11 μg/L) and GET (1,684.29 μg/L) was the highest. A majority of alkenes such as (E,E)-1,3,5-heptatrien, α-phellandrene, α-terpinene, α-myrcene, α-farnesene and cis-calamenene were found higher in GE than those in GET. On the contrary, some alkenes such as styrene and (E,Z)-2,6-dimethyl-2,4,6-octatriene showed higher contents in GET than in GE.
Figure 2.
Content comparisons of volatile categories in GE and GET. Different lowercase letters indicate significant difference at p < 0.05
Alcohols play an important contribution to the sweet, floral, and grass odors of tea[21,22]. In the study, the content of alcohols in GE (6,685.48 μg/L) was significantly higher than that in GET (1,918.66 μg/L) (p < 0.05). Among them, 3-hexen-1-ol, benzyl alcohol, and linalool were the dominant alcohols in GE and GET. Benzyl alcohol was reported as an important volatile component in black tea and oolong tea, with rose-like and fruity aromas[23]. As an important monoterpene alcohol[24], linalool was associated with citrus-like and floral aromas[25]. In addition, phenylethyl alcohol presented a higher content in GE (1,906.91 μg/L) than that in GET (126.05 μg/L), which was mainly derived from the hydrolysis reaction of β-glucoside or β-primrose glycoside[26].
Ketones are known as important volatile components in tea flavor. In this study, the content of ketones in GE and GET was 1,405.45 and 378.38 μg/L, respectively. Geranylacetone, with magnolia and green odors, was reported to be a vital odorant in black, green and oolong tea[27]. In this study, it was found that geranylacetone was the most abundant ketone in GE (1,162.16 μg/L) while (E,E)-3,5-octadien-2-one was the highest ketone in GET (90.74 μg/L).
Aldehydes play an important role in the aroma quality of GE and GET. The content of aldehydes in GE (2,758.81 μg/L) was significantly higher than that in GET (756.61 μg/L) (p < 0.05). Aldehydes such as 2-methyl-butanal, pentanal, hexanal, benzaldehyde and benzeneacetaldehyde were detected in both GE and GET. Pentanal and hexanal were reported as importance contributors to tea aroma, presenting grassy and green odors. They were formed mainly through lipid degradation[27]. Benzeneacetaldehyde showed a higher content in GE (611.44 μg/L) than GET (95.47 μg/L) (p < 0.05), which was mainly generated by the Maillard reaction[26]. In addition, 2-methyl-pentanal and α-cyclocitral were detected only in GET, while (E)-2-hexenal, (Z)-2-heptenal and (E,Z)-2, 6-nonadienal were detected only in GE.
Pyrazine, pyrrole, furan and their derivatives are important heterocyclic compounds, presenting roasted and caramel-like aromas[28]. In this study, the furans and corresponding derivatives were the main heterocyclic compounds, and their contents in GE and GET were 631.55 and 559.65 μg/L, respectively.
In addition, the content of aromatic hydrocarbons in GE (4,333.28 μg/L) was significantly higher than that in GET (829.70 μg/L) (p < 0.05). Some aromatic hydrocarbons such as butylated hydroxytoluene, naphthalene, o-cymene, p-xylene, and toluene presented high amounts in both GE and GET.
Multivariate statistical analysis
-
To characterize the differences between GE and GET, PLS-DA model was constructed. A data matrix of 211 (volatile compounds) × 6 (samples) was obtained. As shown in Fig. 3a, a clear discrimination between GE and GET was obtained. The parameters of the model (R2Y = 0.991, Q2 = 0.977) demonstrated its good explanatory and predictive ability. To verify the robustness of the model, 200 permutation tests were carried out. The result showed that R2 and Q2 were (0, 0.838) and (0, −0.295) respectively, indicating that the model was robust and there was no overfitting (Fig. 3b). To further assess the specific volatile components that explain the difference, the corresponding loading plot was performed (Fig. 3c). The distance between the individual variable and the main cluster is positively correlated with its effect on classification. It could be found that dihydro-5-pentyl-2(3H)-furanone (164), heptanal (171), and nonanoic acid, methyl ester (59) were closer to GE than GET. On the contrary, acetic acid, pentyl ester (34), hexanoic acid, ethyl ester (40), butanoic acid, 2-methylbutyl ester (44), ethyl tiglate (38), 4-methyl-3-penten-2-one (140), 2,2,6-trimethyl-cyclohexanone (144) and isophorone (145) were closer to GET than GE.
Figure 3.
PLS-DA results of GE and GET using GC × GC-TOFMS. (a) Score plot of PLS-DA (R2Y = 0.991, Q2 = 0.977). (b) Plot of 200 permutation test (R2Y = 0.838, Q2 = −0.295). (c) Loading plot. (d) Fifty five key variables with VIP > 1.2. The numbers of volatile components correspond to Supplemental Table S1.
In order to further understand the crucial volatile compounds to distinguish GE and GET, the variable importance in the projection (VIP) was investigated. In general, when the VIP value is greater than 1, it can be judged that the variable plays an important role in group classification. In this study, a total of 55 volatile components with VIP > 1.2 were screened out (Fig. 3d). To visualize the differences, heat map analysis was carried out. Notably, the red color of the square indicates a higher concentration, while the yellow color indicates a lower concentration. As shown in Fig. 4, some volatile compounds such as (Z)-2-penten-1-ol, methyl tiglate, hexanoic acid, ethyl ester, (Z)-3-hexen-1-ol, acetate, humulene, α-cyclocitral, β-cyclocitral, 2-heptanone, and α-ionone showed higher concentrations in GET than those in GE. In contrast, some typical compounds such as 1-heptanol, terpinen-4-ol, decanoic acid, methyl ester, hexanal, heptanal, citral, nonanoic acid showed higher concentrations in GE than those in GET.
OAV analysis of volatile compounds in GE and GET
-
OAV was an important indicator to evaluate the contribution of volatile compounds to the overall aroma. In general, volatiles with OAV ≥ 1 are considered as important contributors responsible for the entire aroma profile. In this study, a total of 57 volatile components with OAV greater than 1 were screened in GE and GET, among which 37 common volatile compounds were found in both GE and GET (Supplemental Fig. S4 & Supplemental Table S2). Among them, propanoic acid, 2-methyl-, ethyl ester (No. = 28, OAV = 1102.36), methyl tiglate (No. = 32, OAV = 2.76), hexanoic acid, ethyl ester (No. = 40, OAV = 11.08), butanoic acid, 2-methylbutyl ester (No. = 44, OAV = 1.87), butanoic acid, hexyl ester (No. = 52, OAV = 2.34), 1-butanol, 3-methyl-, benzoate (No. = 84, OAV = 1.04,), styrene (No. = 97, OAV = 1.13), cis-jasmone (No. = 151, OAV = 2.89), 2-methyl-pentanal (No. = 167, OAV = 3.26), β-cyclocitral (No. = 184, OAV = 5.98) showed OAVs greater than 1 only in GET, while hexanoic acid, methyl ester (No. = 36, OAV = 1.12), α-terpinene (No. = 106, OAV = 4.46), terpinolene (No. = 108, OAV = 3.66), α-farnesene (No. = 122, OAV = 4.84), 6-methyl-5-hepten-2-one (No. = 142, OAV = 1.19), (E)-2-hexenal (No. = 170, OAV = 14.93), (Z)-2-heptenal (No. = 172, OAV = 2.63), (E, Z)-2,6-nonadienal (No. = 180, OAV = 13,325.75), p-xylene (No. = 190, OAV = 1.15), nonanoic acid (No. = 211, OAV = 3.86) showed OAVs greater than 1 only in GE. It is worth mentioning that linalool (No. = 15, OAVGE = 805.97, OAVGET = 712.35), phenylethyl alcohol (No. = 17, OAVGE = 5,448.33, OAVGET = 360.15), α-myrcene (No. = 103, OAVGE = 73.78, OAVGET = 50.80), (E,E)-3,5-octadien-2-one (No. = 146, OAVGE = 89.56, OAVGET = 181.49), benzeneacetaldehyde (No. = 175, OAVGE = 509.53, OAVGET = 79.56) showed higher OAV values in both samples, indicating that these substances may be the characteristic components of GE and GET.
Aroma wheel construction of GET
-
Through the evaluation by panelists, six odor attributes of GET were determined after discussion, which were 'fruity', 'green', 'floral', 'other', 'woody', and 'roasted', respectively. In order to explore which key volatile compounds contributed to the aroma profile of GET, GC-O-MS sniffing was conducted. A total of 41 volatile components, including 21 esters, nine alcohols, four aldehydes, four ketones, two alkenes and one heterocyclic compound, were selected as aroma-active components, and their aroma descriptions are described in Supplemental Table S3. Most odorants were recorded as fruity, floral, green, and sweet odors, with an aroma intensity (AI) between 1.0 and 4.0. Notably, linalool and cis-jasmonone have AI values as high as 4.0, indicating that they were the most important contributors to the aroma profile of GET.
To further identify the key odorants that contributed significantly to GET aroma characteristics, 24 key odorants were screened based on GC-O-MS sniffing and OAV ≥ 1 (as listed in Table 1). And aroma wheel was constructed to visualize their contributions to GET. As shown in Fig. 5, 24 volatile compounds were mainly divided into six odor attributes by combining with the results of sensory evaluation.
Table 1. A total of 24 key odorants in GET based on OAV ≥ 1 and GC-O-MS results.
Compounds RT (min)a Aroma descriptionsb Aroma intensitiesc Aroma attributesd Propanoic acid, 2-methyl-, ethyl ester 8.58 Fruity, banana-like 3.2 Fruity Butanoic acid, 2-methyl-, methyl ester 9.96 Fruity, apple, green pear 2.0 Fruity Hexanoic acid, ethyl ester 19.25 Fruity, sweet 2.2 Fruity Butanoic acid, hexyl ester 27.79 Green, fruity 3.0 Fruity (E,E)-3,5-Octadien-2-one 34.20 Fruity, green grassy 2.0 Fruity Benzoic acid, methyl ester 36.51 Fruity 3.0 Fruity Benzyl alcohol 45.31 Fruity, rose-like 3.0 Fruity Nerolidol 50.96 Fresh, floral, fruity 2.0 Fruity Pentanal 9.08 Fermented bready 1.7 Green Acetic acid, hexyl ester 21.11 Fruity, green 3.0 Green (Z)-3-Hexen-1-ol, acetate 23.22 Green, sweet, fruity 2.0 Green 3-Hexen-1-ol 26.13 Green, leafy, grassy 2.2 Green Geranylacetone 44.74 Magnolia, green 3.2 Green Linalool 33.08 Citrus, floral, sweet 4.0 Floral α-Ionone 44.70 Floral, violet 2.0 Floral Phenylethyl alcohol 47.05 Rose, honey 3.2 Floral cis-Jasmone 47.88 Jasmine-like, herbal, floral, woody 4.0 Floral Indole 59.77 Floral, animal-like 2.0 Floral Methyl tiglate 17.88 Ethereal 3.2 Other 1-Octanol 34.17 Matallic 2.3 Other Methyl salicylate 42.33 Mint-like 3.2 Other 2-Methyl-butanal 9.00 Musty, chocolate, nutty 2.5 Roasted Benzaldehyde 32.33 Bitter, almond-like 2.3 Roasted Safranal 37.36 Woody, spicy, phenolic 2.2 Woody a RT: The retention time of compounds identified by GC-O-MS. b Aroma descriptions obtained from GC-O-MS. c Aroma intensity perceived by panelists. d Aroma attributes referred to the results of sensory evaluation. The 'fruity' attribute of the aroma wheel was composed of eight volatile compounds, including two alcohols, five esters, and one ketone. Among them, esters such as benzoic acid, methyl ester, hexanoic acid, ethyl ester, propanoic acid, 2-methyl-, ethyl ester, butanoic acid, 2-methyl-, methyl ester and butanoic acid, hexyl ester were important contributors to the 'fruity' attribute. For alcohols, benzyl alcohol (fruity and rose-like aroma) and nerolidol (fruity, fresh and floral aroma) were reported to be important odorants for the fruity and floral aroma of black tea[29]. (E,E)-3,5-Octadien-2-one with a low threshold in water (0.5 μg/L), was also a significant contributor to the 'fruity' attribute.
The 'green' attribute of the aroma wheel mainly included one aldehyde (pentanal), one ketone (geranylacetone), one alcohol (3-hexen-1-ol) and two esters (acetic acid, hexyl ester and (Z)-3-hexen-1-ol, acetate) with carbon atomic number of 5 to 12, presenting green and grassy aromas. Among them, pentanal was the main contributor to the green and grassy flavor in tea[30]. As an important alcohol, 3-hexen-1-ol had an important contribution to the green odor of tea infusion. These compounds were mainly derived from the oxidative degradation of unsaturated fatty acids such as linoleic acid and linolenic acid[31].
Five volatile compounds including linalool, phenylethyl alcohol, α-ionone, indole and cis-jasmone were important volatile components responsible for the 'floral' attribute in aroma wheel. As a crucial odorant, linalool was widely distributed in a variety of plants, with citrus-like, floral and sweet aromas[32]. α-Ionone with violet-like aroma was reported to play an importance role in corn-like aroma of green tea[33]. Phenylethyl alcohol with floral and rose aromas, was mainly derived from phenylalanine degradation[31]. It is worth noting that linalool and cis-jasmone had strong aroma intensity and long duration during the sniffing process.
Benzaldehyde and 2-methyl-butanal were the main contributors to the 'roasted' attribute. Benzaldehyde was considered to present an almond-like aroma. 2-Methyl-butanal presenting a nutty aroma, was reported to play an important role in Darjeeling black tea[25].
As an important aroma attribute of GET, the 'woody' attribute was composed of safranal, which was mainly derived from the degradation of carotenoids. The 'other' attribute was mainly composed of methyl salicylate (mint-like), 1-octanol (metallic), and methyl tiglate (ethereal). Among them, methyl salicylate was reported to be an important volatile compound for the unique floral aroma in 'Oriental Beauty' oolong tea[34].
-
All data generated or analyzed during this study are included in this published article and its supplementary information files.
-
About this article
Cite this article
Xie J, Wang Q, Cui H, Wang L, Deng Y, et al. 2024. Characterization of Gardenia tea based on aroma profiles using GC-E-Nose, GC-O-MS and GC × GC-TOFMS combined with chemometrics. Beverage Plant Research 4: e001 doi: 10.48130/bpr-0023-0034
Characterization of Gardenia tea based on aroma profiles using GC-E-Nose, GC-O-MS and GC × GC-TOFMS combined with chemometrics
- Received: 23 August 2023
- Revised: 10 October 2023
- Accepted: 24 October 2023
- Published online: 02 January 2024
Abstract: Gardenia tea (GET) is one of the typical representatives of Chinese scented tea and is loved by consumers for its pleasant aroma. In the present study, the volatile profiles of GET were characterized by gas chromatography electronic nose (GC-E-Nose), comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOFMS) and gas chromatography-olfactory-mass spectrometry (GC-O-MS) combined with chemometrics. Satisfactory discrimination was obtained by GC-E-Nose combined with principal component analysis, with the cumulative contribution rate reaching 86.74%. A total of 202 volatile compounds were identified in GET by GC × GC-TOFMS, among which esters, alkenes and alcohols were the dominant volatile components. Moreover, 24 key odorants were screened out from GET based on odor activity value ≥ 1 and GC-O-MS results. The aroma wheel of GET with six attributes of 'fruity', 'green', 'floral', 'other', 'woody', and 'roasted' was constructed to visualize the contributions of those key volatile compounds. The results provide a new strategy to elucidate the volatile profiles and aroma wheel of scented tea.
-
Key words:
- GC-E-Nose /
- GC × GC-TOFMS /
- GC-O-MS /
- Gardenia tea /
- Aroma wheel