Characterization of Key Aroma Compounds in Fermented Bamboo Shoots Using Gas Chromatography-Olfactometry-Mass Spectrometry, Odor Activity Values, and Aroma Recombination Experiments

Guangxi fermented bamboo shoots (GFBS) are widely appreciated by consumers in China because of their unique aroma. In this study, the dominant aroma compounds of GFBS were investigated using gas chromatography-olfactometry-mass spectrometry, odor-activity values, and aroma recombination. The results show that 70 aroma compounds, including alcohols, esters, aldehydes, acids, phenols, ethers, ketones, alkenes, benzene derivatives, and furans, were identified in GFBS. Among them, 15 aroma compounds with odor-activity values (OAVs) > 1 were identified. Aroma-recombination-omission experiments and sensory evaluation demonstrated that octanal, (E)-2-octenal, acetic acid, guaiacol, phenylethyl alcohol, creosol, 4-ethylguaiacol, and p-cresol significantly contributed to the characteristic aroma of GFBS. Most importantly, p-cresol (34,997.95 ≤ OAV ≤ 71,409.51) and acetic acid (2155.79 ≤ OAV ≤ 3872.09) significantly contributed to its aroma (p < 0.001). The major aroma profile of GFBS included a strong fermented odor, which was pungent and sour. This study provides a theoretical basis for improving the flavor of GFBS.


Introduction
Fermented bamboo shoots (FBS) have tremendous health benefits, including such properties as anti-cancer, anti-oxidant, anti-aging, cardioprotective, weight loss, and probiotic, to name a few. In addition, FBS are an important functional food with both industrial and economic value [1][2][3][4]. Nowadays, their value is appreciated worldwide. In India, fermented bamboo shoots are referred to as "green gold" for their high nutritional value [3], and they have become part of an everyday diet in Taiwan, China [5]. GFBS is a unique FBS produced in the Guangxi Zhuang Autonomous Region, China. GFBS is also the main ingredient and flavor of many characteristically Chinese foods in Guangxi (in a similar manner as Luosifen, which is a kind of famous noodles). Notably, aroma is regarded as one of the most important factors for evaluating the quality and consumer acceptance of GFBS.
Despite this, only a few studies have been conducted on the structure of the characteristic flavor components of GFBS. For example, Guo et al. [6] compared the extraction effects of liquid-liquid extraction, simultaneous distillation extraction, and headspace solid-phase microextraction (HS-SPME) using gas chromatography-mass spectrometry (GC-MS). The results showed that HS-SPME was more suitable for extracting odorants from fermented bamboo shoots resulting in the largest peak area and the greatest number of substances

Quantitative Assessment of Aroma Compounds
The aroma compounds in GFBS were semi-quantified using 2-methyl-3-heptanone as an internal standard. Briefly, 1.5 µL of 2-methyl-3-heptanone was added to the sample and analyzed using GC-MS. The concentration of aroma compounds was calculated according to the ratio between the peak area and the concentration of 2-methyl-3-heptanone.
The aroma compounds identified through GC-MS-O were quantified, according to the literature, with several modifications [13,15,16]. Prior to quantification, an odorless artificial matrix was prepared by further treatment with GFBS. The sample was eluted with an organic solvent using HS-SPME-GC-O-MS until nothing was detected. The specific method was as follows: first, 1-butanol and di-isopropyl ether (B-DIPE) were added to the sample (1-butanol, B-DIPE, and sampled at a ratio of 2:3:2.5, w/w/w). Then, the mixture was shaken for 8 h, and the organic solvent was removed. This was repeated three times. Furthermore, the organic solvent in the mixture was removed by rotary evaporation at least three times. Samples were subsequently treated with liquid nitrogen (99.99%) and frozen in an FD-1D-5D vacuum freeze dryer (Beijing Biocool Lab Apparatus Co., Ltd., Beijing, China) at −60 • C for 48 h.

OAVs
The OAVs of the quantified compounds on the calibration curves were calculated using the following formula: OAV = C i /OT i , where C i is the concentration of compound I, and OT i is the odor threshold value in the water of compound i.

Quantitative Descriptive Sensory Analysis
Quantitative descriptive sensory analysis was carried out in a sensory room (the room temperature was 25 ± 1 • C), which was followed by ISO 8589 standard [17]. Nine sensory evaluators (five female and four male) who were 20-30 years old were recruited to perform GFBS sensory evaluation [18]. The sensory evaluators were selected from 53 candidates based on their performance in several olfactory tests and on the absence of olfactory allergies [19]. The sensory evaluators were trained with standard solutions and Guangxi fermented bamboo shoot products twice a week for six weeks to familiarize them with the tasks and underlying mechanisms of the testing and evaluation procedure. Each participant provided full informed consent. Five relevant flavor attributes-fermented, pungent, sour, alcoholic, and moldy-were selected after panel discussion. The intensity of each attribute was rated by the assessors on a 10-point scale (0: none to 10: very strong). The average results obtained from the sensory panel were ultimately mapped to a sensory profile.

Recombination and Omission Experiments
Aroma recombination was conducted to verify that the volatile compounds with high OAVs were the key odorants of GFBS. LZ-FBS have the richest flavor and were selected as the recombinant control. Recombination and omission experiments were performed as previously reported [20,21]. Recombination model 1 was formulated with an odorless matrix, authentic flavor standards with OAVs > 1, and ultrapure water. Recombination model 2 consisted of recombination model 1 with one flavor standard omitted. The aroma components of recombination model 2 were different from those of GFBS and recombination model 1, indicating that the missing aroma components of recombination model 2 are the key aroma components. Recombination model 3 consisted of key aroma compounds, an odorless matrix, and ultrapure water. Recombination model 3 was evaluated on the basis of whether it was consistent with GFBS according to panelists in a triangle test, according to the previously described method [21].

Statistical Analysis
All experimental results were based on three replicates. Statistical analysis was conducted using SPSS software (version 19.0; IBM Corporation, Armonk, NY, USA). The cluster heatmap of aroma compounds in GFBS was created using R software (version 4.0.2; R Foundation for statistical computing, Vienna, Austria) with the heatmap package. Results are expressed as mean ± standard deviation from three measurements.

Sensory Analysis
Quantitative description analysis was performed to determine the aroma characteristics of GFBS from the four regions of Guangxi. As shown in Figure 1, the aroma profiles of GFBS were described using five attributes, namely "fermented", "pungent", "sour", "alcoholic", and "moldy". Analysis of variance was employed to distinguish differences in sensory evaluation scores, and the results showed that the intensity of two (fermented and sour) attributes significantly differed among the different GFBS samples (p < 0.05). The BS-FBS samples only scored highly in the "moldy" category and had very low scores for the other four sensory attributes. This might be due to the relatively high phenol content and low levels of other aroma-active compounds. NN-FBS and GL-FBS had significantly different "sour" scores but similar scores for the other attributes. In addition, LZ-FBS was characterized most strongly by a sour note with a score of 7.48, followed by "fermented" (7.17), "pungent" (6.00), and "alcoholic" (3.44). for the other four sensory attributes. This might be due to the relatively high phenol content and low levels of other aroma-active compounds. NN-FBS and GL-FBS had significantly different "sour" scores but similar scores for the other attributes. In addition, LZ-FBS was characterized most strongly by a sour note with a score of 7.48, followed by "fermented" (7.17), "pungent" (6.00), and "alcoholic" (3.44).

Identification and Quantification of GFBS with GC-MS
As shown in Table 1, 70 aroma compounds were identified from GFBS from four regions of the Guangxi Zhuang Autonomous Region. These compounds were divided into ten groups: alcohols, esters, aldehydes, acids, phenols, ethers, ketones, alkenes, benzene derivatives, and furans. Among these, the highest number of chemical components detected in one group was alcohols; 13 to 16 alcohols were detected in LZ-FBS, GL-FBS, NN-FBS, and BS-GFBS. More importantly, the chemical class with the highest content was phenols, accounting for 59. 8

Identification and Quantification of GFBS with GC-MS
As shown in Table 1, 70 aroma compounds were identified from GFBS from four regions of the Guangxi Zhuang Autonomous Region. These compounds were divided into ten groups: alcohols, esters, aldehydes, acids, phenols, ethers, ketones, alkenes, benzene derivatives, and furans. Among these, the highest number of chemical components detected in one group was alcohols; 13 to 16 alcohols were detected in LZ-FBS, GL-FBS, NN-FBS, and BS-GFBS. More importantly, the chemical class with the highest content was phenols, accounting for 59. 8

Identification of Aroma-Active Compounds by GC-MS-OSME and Their OAV Analysis
OSME is widely used to determine the compounds responsible for aroma perception. The results of the sensory evaluation demonstrated that LZ-FBS had the richest aroma and was most representative of GFBS in general. The LZ-FBS was therefore used for GC-O analysis, which revealed 18 aroma compounds. As shown in Table 1, the perceptible aroma compounds were propanal, ethyl acetate, octanal, nonanal, 1-octen-3-ol, (E)-2-octenal, acetic acid, benzaldehyde, 1-nonanol, methyl salicylate, guaiacol, geranylacetone, phenylethyl alcohol, creosol, 4-ethylguaiacol, anisaldehyde, p-cresol, and 4-ethyl-phenol. These were considered the primary contributors to the aroma of GFBS.

Identification of Aroma-Active Compounds by GC-MS-OSME and Their OAV Analysis
OSME is widely used to determine the compounds responsible for aroma perception. The results of the sensory evaluation demonstrated that LZ-FBS had the richest aroma and was most representative of GFBS in general. The LZ-FBS was therefore used for GC-O analysis, which revealed 18 aroma compounds. As shown in Table 1, the perceptible aroma compounds were propanal, ethyl acetate, octanal, nonanal, 1-octen-3-ol, (E)-2-octenal, acetic acid, benzaldehyde, 1-nonanol, methyl salicylate, guaiacol, geranylacetone, phenylethyl alcohol, creosol, 4-ethylguaiacol, anisaldehyde, p-cresol, and 4-ethyl-phenol. These were considered the primary contributors to the aroma of GFBS.
The contribution of the individual aroma compounds to the overall aroma of GFBS depends not only on the content of the compounds but also on their OAVs. To accurately calculate the OAV values of each aroma compound, their concentrations were calculated according to the calibration curves. As shown in Table 2, the calibration curves and characteristic ion fragments (m/z) were determined. The correlation coefficients (R 2 > 0.99) revealed high linearity for aroma compounds in GFBS.
As shown in Table 3  The contribution of the individual aroma compounds to the overall aroma of GFBS depends not only on the content of the compounds but also on their OAVs. To accurately calculate the OAV values of each aroma compound, their concentrations were calculated according to the calibration curves. As shown in Table 2, the calibration curves and characteristic ion fragments (m/z) were determined. The correlation coefficients (R 2 > 0.99) revealed high linearity for aroma compounds in GFBS.
As shown in Table 3 (1.64-5.33) also had high OAVs in most GFBS samples. It is important to note that the OAVs of acetic acid and p-cresol were much higher than those of the other compounds, and they were the most decisive compounds. In summary, 15 aroma compounds with OAVs > 1 were detected in GFBS.  x is the peak area relative to that of the internal standard, 2-methyl-3-heptanone, and y is the concentration (ng/g) in the sample relative to that of the internal standard, 2-methyl-3-heptanone. c Sniffed odors determined by GC-O analysis.  [22]. c Odor threshold from Lu et al. [23]. d Odor threshold from Liu et al. [13]. e Odor threshold from Zhao et al. [9]. f Odor threshold from Ren et al. [16]. g Odor threshold from Rothe et al. [24]. h Odor threshold from Stevens [25]. i Odor threshold from Abe et al. [26]. j Odor threshold from Wang et al. [27].

Determination of Key Aroma Compounds Using Recombination Experiments
To evaluate the overall importance of each aroma compound, aroma omission was performed with 15 aroma compounds with OAVs > 1. A triangle test was used for sensory evaluation. After comparing the odor properties of recombinant samples and GFBS samples, the main characteristics were described using five sensory attributes: fermented, pungent, sour, alcoholic, and moldy.
By comparing the aroma characteristics of recombinant model 1 and LZ-FBS, it was found that the sensory properties of the samples obtained by recombination of the 15 aroma compounds (recombinant model 1) were in accordance with the original sample. Fifteen omission models were prepared, and the contribution of each aroma compound was evaluated using a triangle test under the guidance of recombinant model 1. As shown in Table 4, eight aroma compounds, including aldehydes, acids, phenols, and alcohols, significantly influenced the aroma profile of the recombination model. Notably, acetic acid, guaiacol, and p-cresol exhibited highly significant differences (p ≤ 0.001) when they were omitted from the mixture. The omission of phenylethyl alcohol and creosol caused a highly significant difference (p ≤ 0.01). Moreover, when octanal, (E)-2-octenal, and 4ethylguaiacol were omitted, the difference was also significant (p ≤ 0.05). We then carried out a recombination experiment with eight compounds that had a significant effect on the aroma of the profile of recombination model 1. The results are shown in Figure 3, and the aroma characteristics of recombinant model 3 were similar to those of GFBS. In terms of overall similarity, compared with the original GFBS, the recombined aroma model was rated 4.3 out of 5 points.

Discussion
In this study, 70 aroma compounds were identified from GFBS using GC-MS, and these were divided into 10 categories. Guo et al. identified 43 volatiles in GFBS using the HS-SPME method with a coating fiber (65 µm DVB/PDMS). In our study, we extracted volatiles from GFBS using a coating fiber (50/30 µm CAR/PDMS/DVB) and identified more volatile compounds.

Dominant Aroma Compounds in GFBS
As shown in Tables 3 and 4, p-cresol had the highest OAVs and had a highly significant effect on the aroma of GFBS; this was therefore designated the most important aroma compound in GFBS. This is consistent with Guo et al. [6], in which p-cresol made the greatest contribution to odor composition in fermented bamboo shoots. In addition, pcresol has been identified as an important aroma compound in fermented bacterial food [28]. Other phenolic compounds, including guaiacol, 4-ethylguaiacol, and creosol, were

Discussion
In this study, 70 aroma compounds were identified from GFBS using GC-MS, and these were divided into 10 categories. Guo et al. identified 43 volatiles in GFBS using the HS-SPME method with a coating fiber (65 µm DVB/PDMS). In our study, we extracted volatiles from GFBS using a coating fiber (50/30 µm CAR/PDMS/DVB) and identified more volatile compounds. Tables 3 and 4, p-cresol had the highest OAVs and had a highly significant effect on the aroma of GFBS; this was therefore designated the most important aroma compound in GFBS. This is consistent with Guo et al. [6], in which p-cresol made the greatest contribution to odor composition in fermented bamboo shoots. In addition, p-cresol has been identified as an important aroma compound in fermented bacterial food [28]. Other phenolic compounds, including guaiacol, 4-ethylguaiacol, and creosol, were identified as the key aroma compounds in recombination tests. These phenolic compounds have also been reported to contribute to the characteristic aromas of some fermented foods. For example, guaiacol was reported to be the key aroma compound in Japanese miso [29], and 4-ethylguaiacol was identified as the key aroma compound of pixian broad bean paste (PBBP) and traditional Chinese soy sauce [9,27]. It has been reported that 4-ethylguaiacol and creosol contribute to spicy and smoky odors [30], which may play an important role in the production of pungent odors by GFBS. In summary, phenolic compounds were the dominant aromatic compounds in GFBS.

As shown in
Moreover, acetic acid also had a highly significant effect on the aroma of GFBS. It was previously reported that the strongest odors of bamboo shoots from Taiwan were caused by p-cresol, acetic acid, 2-heptanol, and 1-octen-3-ol, and it is speculated that the main odor may be formed by the synergistic action of p-cresol and acetic acid [5]. Based on the recombination tests and OAV analysis, we speculate that the synergistic effect of p-cresol and acetic acid may be the reason for the main aroma of GFBS. Acetic acid is the most important fatty acid in GFBS and contributes to a strong sour odor. Therefore, acids were also the dominant aromatic compounds in GFBS. Additionally, acetic acid has been reported as the principal aroma-active compound in Japanese miso [29]. The synergistic effect of acetic acid with 3-methylbutanoic acid and 4-methylpentanoic may be the reason for the characteristic sour aroma of PBBP [9].

Subdominant Aroma Compounds in GFBS
As shown in Table 3, five aldehydes are important aroma components of GFBS, specifically, octanal and (E)-2-octenal. (E)-2-octenal exists as a key aroma component, which has also been reported in other foods, such as Millet Huangjiu [31] and Beijing roasted duck [13]. Octanal has been reported to have a significant influence on roasted duck aroma [13]. Most aldehydes are formed via yeast metabolism or carbon chain oxidation of unsaturated fatty acids [32]. Aldehydes in GFBS may be related to yeast fermentation. In terms of alcohols, 13-16 alcohols were detected per GFBS sample, comprising the largest number of compounds in one group. Moreover, in this study, phenylethyl alcohol was identified as the most important alcohol and it exhibited a floral odor, followed by 1-octen-3-ol and 1-nonanol. The important aroma compounds 1-octen-3-ol and phenylethyl alcohol were identified in GFBS, which is consistent with Guo et al. [8]. Alcohols in GFBS are mainly formed through bacterial and fungal fermentation [33].

Other Substances with Important Contributions to GFBS Odor
In this study, ethers, ketones, alkenes, benzene derivatives, and furans were also detected in GFBS. However, they accounted for a small proportion of the total number of compounds, but their contribution is indispensable. Among these, esters are important. A total of 12 esters were detected; ethyl acetate is an important aroma compound, which reportedly plays an important role in the aroma of other fermented foods, such as wine [30]. Esters are mainly formed by the esterase-catalyzed reaction of alcohols and acids produced from glucose and amino acids during microbial metabolism [31].

Conclusions
This study provided a comprehensive investigation into the identification of potent compounds contributing to the characteristic aroma of GFBS through recombination and omission experiments. A total of 70 aroma compounds were identified by GC-MS, and 18 aroma-active compounds were identified from LZ-GFBS by GC-MS-O. Fifteen aroma-active compounds with OAVs > 1 were further recognized as important aroma compounds. Finally, omission tests further confirmed octanal, (E)-2-octenal, acetic acid, guaiacol, phenylethyl alcohol, creosol, 4-ethylguaiacol, and p-cresol as the key odorants. Informed Consent Statement: Informed consent was obtained from all participants involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with this manuscript. We have no financial or personal relationships with other people or organizations that may inappropriately influence our work.