Characterization of Aroma Volatiles in Xilin Fire Ginger Oils by HS-SPME-GC-MS

ABSTRACT Ginger is widely used as either a food product or a herbal medicine around the globe. In the current study, we used headspace solid-phase microextraction and gas chromatography-mass spectrometry technology (HS-SPME-GC-MS) with n-butyl acetate as an internal standard to characterize three kinds of ginger oil in Xilin fire ginger, which included ginger essential oil (GEO), ginger oleoresin extracted with petroleum ether (PEGO) and ginger oleoresin extracted with absolute ethanol (AEGO). Cluster analysis of heatmap was used to reveal the differences in concentration in these oils. Odor activity value, in combination with principal component analysis, was further used to analyze the contribution to the aroma. This study demonstrated that despite the similarities in the aroma compounds and content of three kinds of ginger oil, GEO exhibited a better aroma quality, followed by AEGO and PEGO.


INTRODUCTION
Ginger oils are an essential kind of extract of ginger, with vital phytochemical characteristics that reflect the sensory characteristics and biological activity of ginger. Based on the extraction method, it can be classified into two categories, ginger essential oil (GEO) and ginger oleoresins (GOs). [1,2] Ginger essential oil is mainly prepared by hydrodistillation, while ginger oleoresins are prepared by various methods, such as pressing method, liquid CO 2 extraction method, supercritical CO 2 extraction method and solvent extraction method that can also be combined with ultrasonic method and microwave. [3,4] Due to different extraction methods, the compounds and aroma of ginger oils are quite different. Volatiles, including sesquiterpene and monoterpenoid hydrocarbons in GEO, are the main constituents to provide distinct aromas and tastes. GOs, which also contain nonvolatile phenolics known as gingerols, embody pungent characteristics. [5][6][7][8] The aroma is one of the vital categories for the sensory evaluation of ginger oils, and plays a substantial role in defining the product's quality, and greatly influences consumers' attraction and purchasing. [9] The headspace compounds are more similar to that of inhaled or sniffed aroma. [10] HS-SPME is also one of the main measurement methods of aroma compounds, with a simple device, easy operation, fast sampling, good reproducibility and avoiding chromatographic column contamination, etc. [11] It in combination with GC-MS has been widely used in many oils for quantifications, such as grape seed oil, [12] camellia seed oil, [13] garlic oil, [14] tuna oil [15] and ginger oleoresins. [16] Previously, The dried ginger slices were pulverized into a 40 mesh molecular sieve and were used for extraction. The essential oil was extracted by hydrodistillation process using a Clevenger's type apparatus with material ratio of 1:16 (g: mL), and the volume of oil kept constant indicating extraction end. The clear and transparent yellow oil was collected from the upper layer and dried with a minimum amount of anhydrous sodium sulfate to remove any remaining moisture traces.
Two oleoresins were extracted from the Soxhlet apparatus using petroleum ether and absolute ethanol at a temperature of 85 °C and 95 °C, respectively, and the material ratio of 1:8 (g: mL). The extraction process was continued until the reflux liquid became colorless and transparent and concentrated at 50 °C. The essential oil and oleoresins were stored at low temperature (4 ± 2 °C) in the dark for further use.
HS-SPME The best HS-SPME conditions as follows were optimized in our previous studies. 0.1500 g ginger essential oil (GEO), 0.4000 g ginger oleoresin extracted with petroleum ether (PEGO) and 0.4000 g ginger oleoresin extracted with absolute ethanol (AEGO) were, respectively, added to 0.5500 g, 2.0000 g, and 1.000 g of n-butyl acetate methanol internal standard solution (0.0010 g/g), respectively. Then, each solution was placed into a 15 mL headspace vials and thoroughly mixed. After adsorption, the SPME fiber was inserted through the septum and exposed to the headspace for desorption and further analyses. GEO was adsorbed for 3.5 min and desorpted for 3 min, each oleoresin was adsorbed for 7.5 min and desorpted for 2 min.

GC-MS analysis
A gas chromatograph (Hewlett Packard 7890) coupled with HP 5973 C MS detector (Agilent Technologies, USA) with a BR-5 ms capillary column (selectivity similar to 5% diphenyl/95% dimethyl polysiloxane, 60 m × 0.25 mm, 0.25 μm film thickness) was used to analyze volatile aroma compounds. Helium (99.999%) was used as a carrier gas at a constant flow of 1.20 mL/min. The split injection was applied, and the split ratio was set at 10:1. The oven temperature parameter was set as: 50 °C for 1 min; 3 °C/min ramp to 150 °C and holding for 2 min; 20 °C/min ramp to 250 °C and holding for 2 min. The mass detector was conducted with electronic impact (EI) mode at 70 eV, and the ion-trap manifold temperature of 230 °C and source temperature of 250 °C, the scanning rate of 1 scan s −1 and mass acquisition range was set between 35 and 550 amu.
The volatile compounds were identified by NIST 11 Database, which required positive and negative matching degrees greater than 800 and comparing their retention indices (RI) with those previously reported in the literatures. The RI of volatile compounds was calculated by sample injection with a homologous series of straight-chain n-alkanes (C8-C40) under the same conditions. Each volatile concentration in ginger oils was calculated by comparing it with the concentration of the internal standard.

Statistical analysis
To quantify the volatiles, each sample was analyzed in triplicate, and the results are expressed as the mean ± standard deviation of three measurements. Heatmap analysis was performed by an online platform using Metabo Analyst 3.5 at https://www.metaboanalyst.ca. Data were compared using a one-way analysis of variance followed by Duncan's post-hoc tests (among groups) using SPSS 25.0. Principal component analysis (PCA) was conducted using SIMCA14.1, and the data were preprocessed by log2-transformed and UV scaling prior to analysis.

Aroma compounds analysis of ginger oils
Lowercase letters a, b, and c in the same row represented significant differences (P < .05).

Heatmap clustering analysis
Heatmap can simply and intuitively display the similarity and difference of multiple samples at various classification levels through the color gradient and degree of similarity, it is often used to study the similarity in data between samples. [25] The heatmap of 55 aroma compounds' concentrations in Table 1 expecting δ -Eiemene, α -selinene and (Z,E)-α -farnesene who were automatically filtered out because only little was detected in AEGO of the three kinds of ginger oil are shown in Figure 3.
(GEO (A1, A2, A3), PEGO (B1, B2, B3), AEGO (C1, C2, C3), and each number corresponding to substance is consistent with that in Table 1) As shown in Figure 3 the color orange (blue) represents a larger (smaller) value. The differences in the concentration of 52 aroma compounds were clearly shown. And the closer the color is, the closer the clustering relationship is. As a result, the three kinds of ginger oil also successfully carried out hierarchical clustering through these compounds. First, the heatmap accurately clustered three kinds of ginger oil represented by each parallel sample (A, B, and C) and then clustered the ginger oleoresins (B and C) together, and finally clustered with ginger essential oil.
The content of all monoterpenes and most sesquiterpenes aroma compounds in ginger essential oil were much higher compared with the ginger oleoresins. This may be due to the hydrodistillation method, which is the main extraction method to extract volatile compounds, and ginger essential oil has good fluidity that is more conducive to aroma compounds volatilization. In addition, the presence of citronellol acetate, α-bergamotene, γ-elemene, (E)-β-famesene, aromandendrene, (-)-α-muurolene and (+)-δ-cadinene in ginger oleoresins were not detected in ginger essential oil. Therefore, the clustering relationship was far.
In ginger oleoresins, the species of monoterpenes were rare and their content was lower, while the sesquiterpenes were abundant and the content was higher in AEGO, such as the content of (+)cyclosativene, copaene, trans-α-bergamotene, (-)-zingiberene, (E,E)-α-farnesene, β-bisabolene and βsesquiphellandrene in AEGO were significantly higher than PEGO. And citronellol acetate, trans-αbergamotene and (-)-α-muurolene were not detected in PEGO. However, the content of camphene and eucalyptol in PEGO was significantly higher than AEGO, and tricyclene, α-pinene, (-)-verbenone  were not detected in AEGO. This may be due to sesquiterpenes and total terpenoids, which increase with the increasing polarity of the extraction solvent. While monoterpenes decreased with the increasing polarity of the extraction solvent. However, the total terpenoid content of AEGO was lower than that of PEGO, which may be due to high viscosity of AEGO than PEGO, which affects the volatilization of aroma compounds. The contents of other aroma compounds were slightly different, so they showed an excellent clustering relationship.
Our result showed that β-myrcene, β-phellandrene, eucalyptol, β-linalool, (-)-β-elemene and geraniol had significant aroma contributions in the three kinds of ginger oil, suggesting that these were important characteristic aroma compounds of fire ginger oils. Based on the odor descriptions of  aroma compounds in Table 1, these compounds contributed a similar aroma profile to that of fresh, fruity, spicy, woody and sweet balsamic smells. These findings were consistent with the conclusions of Yingngam et al. [27] and Manuhara et al.. [28] Principal component analysis (GEO (A1, A2, A3), PEGO (B1, B2, B3), AEGO (C1, C2, C3) and each number corresponding to substance is consistent with that in Table 1) Although the overall aroma contour of the three kinds of ginger oil was very similar, but the aroma of GEO was significantly more refreshing, fruity, and woody. PEGO was bland and mild, while AEGO was stimulating, spicy, and herbal. PCA can reduce complex data to low dimensions on the principle of ensuring full utilization of data, and reveals hidden simple and useful important information, which is objective, fair, and scientific. [29] Therefore, PCA of OAVs was performed using the SIMCA14.1 software to explore the fragrance differences of various ginger oils. Our results show that the initial 39 variables were reduced to 2 principal components, variance with 82.6% and 15% were explained by the first principal component (PC1) and second principal component (PC2), respectively (Figure 4). The contribution rate of the principal component's cumulative variance was 97.6%, replicating most of the information and successfully achieving dimension reduction. In Figure 4, PEGO and AEGO were closely clustered together, GEO and GOs (PEGO and AEGO) were located in obviously different regions, which suggests that the extraction method caused a big difference in aroma, while the extraction solvent caused a small difference. The main loading summary of ginger oils are shown in Table 2. And the main aroma compounds of different ginger oils were analyzed by combining Table 2 and Figure 4.
α-Pinene, camphene, β-pinene, 6-methyl-5-hepten-2-one, 2-undecanone, geranyl acetate, and (-)-β-elemene had significant absolute load values in both PC1 and PC2, therefore they were the main compounds responsible for the difference in aroma among three kinds of ginger oil. There are 16 aroma compounds, which could not use OAV for aroma evaluation due to the lack of odor threshold. Previous studies have found that (-)-zingiberene, β-bisabolene, β-sesquiphellandrene, and αcurcumene are highly significant for aroma characteristics of ginger. [30,31] In addition, they also had high content in this study, so the effect of aroma was analyzed based on the results of one-way analysis of variance of their concentrations. The concentration of α-curcumene shows a significant difference among GEO, PEGO and AEGO, while (-)-zingiberene, β-bisabolene and β-sesquiphellandrene did not show significant differences between GEO and AEGO, which were significantly different when comparing GEO with PEGO. These aroma compounds are mainly responsible for the expression of stimulating, spicy, and herbal smells in all compounds (Table 1). So, they were the main reasons why the aroma of GEO and AEGO were more stimulating, spicy, and herbal compared with the PEGO.

CONCLUSION
In this study, HS-SPME-GC-MS combined with multivariate statistical techniques made aroma compounds of Xilin fire ginger oils qualitative and quantitative, and made a detailed material analysis on aroma presentation and difference. GEO, PEGO, and AEGO, respectively, were determined 44, 41, and 44 aroma compounds and were clearly reflected in their differences and divided into ginger essential oil and ginger oleoresins by heatmap based on the concentration. The OAVs of 39 aroma compounds were used to analyze the aroma contribution. β-Myrcene, β-phellandrene, eucalyptol, βlinalool, β-elemene and geraniol were important characteristic aroma compounds of ginger oils. PCA was used to further explore the aroma differences of ginger oils, 32 aroma compounds, such as terpinolene, citronellal, 2-heptanol, β-citral, α-citral, (-)-camphor, geraniol, 2-heptanol, acetate, terpinen-4-ol, and α-terpineol sufficiently explained the large difference in aroma of ginger essential oil and ginger oleoresins due to different extraction methods. And, 14 aroma compounds, consisting of (E,E)α-farnesene, copaene, geraniol, (Z,E)-α-farnesene, citronellol acetate, α-selinene, (-)-β-elemene, βpinene, geranyl acetate, 2-undecanone, α-pinene, camphene, 6-methyl-5-hepten-2-one, and eucalyptol showed that different extraction solvent of two kinds of ginger oleoresin had small difference in aroma. The results of single-factor analysis of variance of (-)-zingiberene, β-bisabolene, βsesquiphellandrene, and α-curcumene whose OAVs cannot be calculated, showed that they had an important effect on the aroma of GEO and AEGO. The findings of the current study can provide significant applications in food, spices, perfumes, and other products. Nevertheless, using OAV to analyze aroma contribution is only limited to known aroma threshold volatiles, so further studies will be required to measure and supplement the threshold value of these substances or combined with other methods.