Typical Aroma of Merlot Dry Red Wine from Eastern Foothill of Helan Mountain in Ningxia, China

Aroma is an important aspect of wine quality and consumer appreciation. The volatile organic compounds (VOCs) and olfactory profiles of Merlot dry red wines from the Eastern Foothill of Helan Mountain (EFHM) were analyzed using gas chromatography-mass spectrometry and quantitative descriptive analysis. The results showed that Merlot wines from EFHM were characterized by intense flavors of drupe and tropical fruits compared with the Gansu region. Nineteen VOCs were defined as essential compounds contributing to the aroma characteristics of the Merlot wines through gas chromatography–olfactometry/mass spectrometry and odor activity value analysis. Predominantly, geranyl isovalerate, which contributed to the herbal odors of the Merlot wines, was detected in the grape wine of EFHM for the first time. The addition experiment revealed that geranyl isovalerate influenced the aroma quality of wine by increasing herbal odors and enhancing the olfactory intensities of tropical fruits. These results are helpful for further understanding the aroma of Merlot wines from EFHM and improving the quality of wine aromas.


Introduction
Aroma is an essential sensory aspect of wine and is highly valued by connoisseurs and consumers. Volatile organic compounds (VOCs) are the chemical basis of the aroma profiles of wine, which exhibit a nonrandom distribution pattern across different regions and contribute to the distinctiveness of wine [1][2][3]. The VOCs deriving from grape berries, alcoholic fermentation, and the aging process can be defined as primary, secondary, and tertiary aromas of wine, respectively [4]. Although hundreds of VOCs have been identified in grapes and wines, only a few help mold the typicality of wine flavor [5][6][7]. Some VOCs are found in low concentrations in wine, but they may have a considerable effect on the aroma quality of wine because of their low sensory threshold at the level of ng·L −1 [4,8]. Nevertheless, it is reported that only a subset of volatiles interacts with the olfactory receptors in the human nose to cause an aroma perception in the brain. A sensory-guided analysis, such as gas chromatography olfactometry (GC-O), revealed that some aroma compounds can be separated from the majority of odorless volatiles [9]. VOCs with odor activity values (OAV) > 1 or odor intensity values greater than 3 in GC-O are generally considered to contribute to the overall aroma of a wine and are considered the key aroma substances in wine [10,11]. Mayr et al. [12] analyzed Syrah wines from Australia using GC-O, quantitation, and aroma reconstitution techniques. According to OAV, ethyl octanoate, ethyl hexanoate, ethyl-3-methylbutanoate, ethyl-2-methylbutanoate, ethyl

VOCs in Merlot Wines
The variation in aroma characteristics arises from different concentrations of VOCs [8]. The headspace solid-phase micro-extraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS) technique detected 51 VOCs in the 72 Merlot wine sam-

VOCs in Merlot Wines
The variation in aroma characteristics arises from different concentrations of VOCs [8]. The headspace solid-phase micro-extraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS) technique detected 51 VOCs in the 72 Merlot wine samples from different regions (Table 1). Each region was considered a distinct entity, and the content of VOCs was aggregated accordingly, as depicted in Figure 2. The total VOC amount was significantly lower in EFHM and GS wines compared with XJ. Nonetheless, the sensory characteristics of the wine are not solely reliant on the cumulative presence of VOCs, but are also influenced by the specific arrangement of distinct categories [21]. Mountain; XJ, Xinjiang; GS, Gansu. Different letters for each aroma attribute indicate significant differences (Duncan's test, p < 0.05), with the color and order of the letters corresponding to the legend.

VOCs in Merlot Wines
The variation in aroma characteristics arises from different concentrations of VOCs [8]. The headspace solid-phase micro-extraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS) technique detected 51 VOCs in the 72 Merlot wine samples from different regions (Table 1). Each region was considered a distinct entity, and the content of VOCs was aggregated accordingly, as depicted in Figure 2. The total VOC amount was significantly lower in EFHM and GS wines compared with XJ. Nonetheless, the sensory characteristics of the wine are not solely reliant on the cumulative presence of VOCs, but are also influenced by the specific arrangement of distinct categories [21]. The 51 VOCs were grouped into six types: esters (29), higher alcohols (8), fatty acids (5), carbonyl compounds (7), terpenes (1), and volatile phenols (1), according to the similarity of chemical structures ( Figure 3). Although most esters were below the olfactory thresholds, they may contribute to the fruity characteristics of wine due to the synergistic The 51 VOCs were grouped into six types: esters (29), higher alcohols (8), fatty acids (5), carbonyl compounds (7), terpenes (1), and volatile phenols (1), according to the similarity of chemical structures ( Figure 3). Although most esters were below the olfactory thresholds, they may contribute to the fruity characteristics of wine due to the synergistic effects [22,23]. Unfortunately, the total content of esters in EFHM Merlot dry red wine was significantly lower than in the other two regions.
Higher alcohols in wine are produced through two pathways during alcoholic fermentation, the Harris pathway of glycolysis and the Ehrlich pathway in amino acid degradation [24]. Studies have shown that higher alcohols can increase the aroma complexity of wine when it is below 300 mg·L −1 [24]. From Figure 3, the alcohol contents of all the Merlot wine samples were less than this level, and the dissimilarity among regions was not significant.
Fatty acids in wine are mainly produced during the alcoholic fermentation stage, which is a by-product of fatty acid metabolism [25]. Fatty acids consist of straight-chain and branched-chain structures depending on whether branches exist in the carbon chain. Wine has cheese-like fragrances from straight-chain fatty acids, while rotten and oily odors come from branched-chain fatty acids [26]. The fatty acid content of Merlot dry red wine in EFHM was significantly lower than in the XJ and GS wines (Figure 3).
Aldehydes and ketones may contribute distinctive flavors to grape wine because their olfactory thresholds are typically 100-10,000 times lower than the corresponding alcohols [26]. Examples include methional with boiled potato smells, nonanal with citrus aromas, and diacetyl with butter odors [26,27]. From Figure 3, the dissimilarity of aldehydes and ketones among regions was not significant.
Terpenes are aroma compounds deriving from grape berries, which are synthesized through the methylerythritol 4-phosphate pathway and the mevalonate pathway [28]. Terpenes contribute floral and fruity characteristics to wine, whose concentrations in EFHM wines were compatible with other regions. Excessive volatile phenols impair the fruity intensity of wine by bringing about smells described as "animal", "stable flavor", "leather", "spicy" [29], and "medicinal" [30]. However, volatile phenols are believed to enhance the aroma complexity at a concentration below 420 µg·L −1 [30,31]. The total content of volatile phenols in Merlot dry red wine ( Figure 3) in EFHM was significantly higher than in XJ wines.  The data are presented in the form of "mean ± standard deviation". In the rows, different letters represent significant differences between samples (Duncan's test, p < 0.05). LRI: linear retention indices on the DB-Wax column obtained from the NIST Chemistry WebBook (https://webbook.nist.gov/ (accessed on 23 July 2023)). LRI *: linear retention indices calculated according to the retention times of C8-C20 n-alkanes and the retention time of each compound on the DB-Wax column [32]. The compound marked "Q" indicates that it was quantified using a standard curve, while the compound labeled "SQ" indicates it was semi-quantified.

of 14
wines were compatible with other regions. Excessive volatile phenols impair the fruity intensity of wine by bringing about smells described as "animal", "stable flavor", "leather", "spicy" [29], and "medicinal" [30]. However, volatile phenols are believed to enhance the aroma complexity at a concentration below 420 µg·L −1 [30,31]. The total content of volatile phenols in Merlot dry red wine ( Figure 3) in EFHM was significantly higher than in XJ wines.

VOCs above Olfactory Threshold
A substance with OAV higher than 1 has an essential effect on the aroma quality of wine, and a higher OAV indicates more essential contributions to wine aroma. Nine VOCs  (Table 3).

Reconstruction of Aroma of Merlot Wine
Considering the sample HL2 having aroma characteristics close to the average of the EFHM wine samples, it was chosen as a control in the reconstruction experiment. The aroma of Merlot dry red wine was imitated using 19 standards of VOCs (Table 4) with OAVs larger than 1, and with olfactory intensities greater than 3. The reconstructed wine was assessed using the QDA method. The aroma profiles on fruity, flower, cream, herbs, and herbaceous plants can be reproduced by the 19 key VOCs, but the smells of nut, marmalade, preserved fruit, spices, and baked flavor of the reconstructed sample deviated from the original wine (Figure 4).

Omission Tests for Geraniol Isovalerate
To investigate the significance of the geraniol isovalerate contribution to Merlot wine, an omission model was prepared to compare with the reconstitution model by a triangle

Omission Tests for Geraniol Isovalerate
To investigate the significance of the geraniol isovalerate contribution to Merlot wine, an omission model was prepared to compare with the reconstitution model by a triangle test. Table 5 showed that the omission of geraniol isovalerate, responsible for the herbal odors of wine, was successfully perceived by all panelists with high significance (p ≤ 0.05).

Addition Experiments for Geraniol Isovalerate
The sensory threshold for geraniol isovalerate was 60 µg·L −1 . To evaluate the perceptual interactions in Merlot wine, geraniol isovalerate at various concentrations was added according to the threshold value and the contents in real wines. Aroma addition experiments highlighted that geraniol isovalerate had significant effects on the herbs and tropical fruit aroma intensity of Merlot wine at the concentrations measured in a Merlot wine sample ( Figure 5). The aromatic intensity of herbs and tropical fruits exhibited consistent changing patterns, initially increasing and reaching its peak around TJ7, followed by a subsequent decrease (Figure 6).

Wine Sample Collection
Seventy-two bottles of Merlot dry red wine (2020 vintage) were collected from 12 wineries located at the EFHM, XJ, and GS wine-producing regions ( Table 6). None of these wines underwent oak barrel aging. After completing alcoholic fermentation and malolactic fermentation, the

Wine Sample Collection
Seventy-two bottles of Merlot dry red wine (2020 vintage) were collected from 12 wineries located at the EFHM, XJ, and GS wine-producing regions ( Table 6). None of these wines underwent oak barrel aging. After completing alcoholic fermentation and malolactic fermentation, the wines were stabilized for a period of 3 months in stainless steel storage tanks. Subsequently, they were filtered, bottled, and promptly sent to the laboratory for analysis.

Conventional Analysis
Alcoholicity (%, v/v), dry matter (g/L), titratable acidity (expressed as g/L of tartaric acid), residual sugar (g/L), and volatile acidity (expressed as g/L of acetic acid) were measured according to the OIV Compendium of International Methods of Wine and Must Analysis (2008). The pH was measured with a PHS-2F pH meter (INESA, Shanghai, China). From Table 7, all the wine samples were dry wines (residual sugar less than 4 g·L −1 ). The EFHM wines exhibited higher alcohol content and lower pH. The volatile acidity content of all samples was below 0.5 g/L, thus avoiding any potential interference in subsequent olfactory sensory analysis. The data are presented in the form of "mean ± standard deviation". In the columns, different letters represent significant differences between samples (Duncan's test, p < 0.05).

Sensory Analysis
The sensory panel consisted of six wine professionals (three males and three females, 20-30 years old). All the panelists were trained for four weeks before the formal sniffing. At the end of the training, a set of wine samples was provided to the panel for description and discussion. A total of 14 aroma descriptors (red fruits, black fruits, citrus fruits, drupe, tropical fruits, herbaceous plant, fresh floral scent, nut, marmalade, preserved fruit, spices, cream, herbs, and baked flavor) were identified for QDA analysis [5,8,11,15].
The formal experiment was conducted in a blind sniffing in a standard wine tasting room (ISO 8589-1998) at 16 • C, using standard wine glasses (ISO 3591-1997) covered with aluminum foil, thereby preventing the panelists from being influenced by the appearance of the wine and affecting the evaluation of its aroma. Each panelist was required to rate the intensity of each descriptor on a scale of 0-10 twice (i.e., two repetitions: once in the morning and once in the afternoon). Each repetition consisted of six sessions, during which the panelist compared the aromas of 12 wine samples from different regions. Although all 12 samples were presented simultaneously within each session, the panelists were asked to take a break of 30-60 s between smelling every two glasses to prevent olfactory fatigue.

GC-MS Analysis
The headspace solid-phase microextraction combined with a 7890B gas chromatography-7000D mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) was used to extract and analyze the aromatic substances [37]. The divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (50/30 µm, 1 cm) was previously conditioned in a baking out unit according to the manufacturer's recommendations (250 • C × 10 min). Then, 1.5 g of NaCl, 5 mL of the wine sample and 10 µL of 4-methyl-2-pentanol (1.008, 3 g/L) were added to the headspace vial. The mixture was preheated at 40 • C for 5 min and extracted using the fiber at 40 • C for 30 min on a CTC PAL autosampler (CTC Analytics, Zwingen, Switzerland). The fiber was desorbed at 240 • C for 10 min in the splitless mode. The carrier gas was high-purity helium (purity ≥ 99.999%), with a flow rate of 1 mL·min −1 . The initial oven temperature for the DB-WAX column (film thickness of 30 m × 0.25 mm id × 0.25 µm; J&W Scientific, Folsom, CA, USA) was 40 • C, held for 5 min, then raised to 97 • C for 7 min at 3 • C·min −1 , then raised to 120 • C at 2 • C·min −1 , raised to 150 • C at 3 • C·min −1 , and finally raised to 220 • C at 8 • C·min −1 for 10 min. The temperature of the transferred line was set at 230 • C. In the MS detector, the full scan mode (35-300 m/z) and electron ionization source were used, with a source temperature of 230 • C and electron energy of 70 eV.

GC-O/MS Analysis
GC-O/MS analysis was performed using the 7890B GC-7000D MS equipped with an olfactory detection port (ODP-4, Gerstel, Mülheim, Germany). EFHM wine samples were extracted by the liquid-liquid extraction method. Twenty milliliters of wine were added to a 50-mL centrifuge tube, 15 mL of dichloromethane and 2 g of NaCl were added, and 80 µL of 4-methyl-2-pentanol (18 µg·mL −1 ) was added as a standard internal reference. The sample was vortexed for 10 min, sonicated for 20 min, and then left at 4 • C for stratifying. The organic phase was collected with a disposable needle tube, and 4 g of sodium sulfate was added for dehydration overnight. Finally, the extract was concentrated with pure nitrogen and stored at −20 • C for analysis.
In a splitless mode, one microliter of the enriched extract was delivered into the front inlet of the 7890 B gas chromatography. The oven temperature for the DB-WAX column was originally held at 40 • C for 5 min, then escalated to 230 • C at a rate of 5 • C/min, and kept at that temperature for 5 min. A flow of 1 mL·min −1 of helium was used as the carrier gas. The split ratio between ODP and MS was 1:1. The temperature of the sniffing transmission line and the sniffing port were 260 • C and 220 • C, respectively, and the humidifier flow rate was 12 mL·min −1 . The six panelists were asked to record the intensity and smell of each VOC on a 1-4 scale (1, weak; 2, medium; 3, strong; 4, extremely strong).

Qualitative and Quantitative Analysis
VOCs were identified through the NIST 17 standard spectral library and further verified with linear retention indices (LRIs) of Alkanes C8 to C20 (Sigma-Aldrich, Shanghai, China) on the DB-Wax column. Stock solutions of standards were prepared volumetrically in absolute ethanol and stored at −20 • C until dissolved in synthetic wines (15% (v/v) of alcohol, 5 g·L −1 of tartaric acid, and pH 3.6) to prepare the calibration data. For quantification, 10-point calibration curves were prepared for each compound by employing a progressive dilution method with a twofold decrease at each step (Table S1). The compounds without standards were semi-quantified, as described by Xia et al. [38].

Aroma Reconstruction Experiments
The VOCs used for aroma reconstruction were only those that simultaneously satisfy the following two conditions: (1) OAV greater than 1.0 and (2) intensity determined by GC-O no less than 3.0 [8][9][10][11]16]. They were added to a synthetic wine (prepared as describe in Section 3.4) according to their concentrations in real wines. The panelists were asked to evaluate the intensities of 14 aroma attributes of the reconstructed wine on a scale from 0 (not perceivable) to 10 (strongly perceivable).

Aroma Omission Experiments
An omission trial was performed to determine the role of geranyl isovalerate in the herbal odors of Merlot wines. The geranyl isovalerate omission model was compared with two completely reconstructed samples, as described in Section 3.5, using a triangle test, and the samples were assigned random numbers [16]. A group of expanded panelists consisting of ten males and ten females, aged between 20 and 30 years, were asked to sniff these samples and indicate one that exhibited differences.

Aroma Addition Experiments
In the addition experiment, we opted for HL2 as the base wine, primarily due to its VOC content closely resembling the average of all collected Merlot wines from the EFHM. The range of added geranyl isovalerate varied between 10 and 90 µg·L −1 . The 4 mL mixture of geranyl isovalerate and HL2 or HL2 alone were kept in an odor-proof container (5 mL) until evaluation [39]. The expanded panelists assessed the 14 aroma attributes described in Section 3.3 on a scale of 0 (absent) to 10 (highly intense). The assessments were conducted in triplicate.

Measurement of Odor Threshold
A synthetic wine (prepared as described in Section 3.4) served as the base wine for the odor threshold tests. Ten concentrations of geranyl isovalerate ranging from 0 to 90 µg·L −1 were evaluated. Each concentration was presented as a set of three glasses and arranged in ascending order of potency. The accuracy of the expanded 20 panelists was evaluated through a triangle test. Correct responses were recorded and plotted against the concentration, and the concentration at which 50% of the judges answered correctly was designated the odor threshold [40].

Statistical Analysis
Data calculations were performed using Microsoft Excel 2016 software (Microsoft Office, Redmond, WA, USA). One-way ANOVA and the Duncan test were applied to determine the variances of basic oenological parameters, volatile aroma components, and sensory scores. The bar chart and radar map were drawn with Origin 2017 software (OriginLab Corporation, Northampton, MA, USA).

Conclusions
In this study, the sensory description analysis and GC-MS analysis were used to investigate the aroma of Merlot wines in three regions of China under different terroirs. The Chinese Merlot wines were described as having intense flavors of black and red fruits. The aroma components of Merlot wine showed differences among different regions. The contents of esters and fatty acids in wine samples from EFHM were the lowest. Due to wine typicality related to terroir, this result could be an indication of different specific soil, topography, climate, landscape characteristics, and biodiversity features of each winemaking region [41,42].
Based on GC-O/MS analysis and OAV analysis, nineteen odor-active compounds were selected to reconstruct the aroma profile of EFHM Merlot wines. The results of GC-O/MS analysis, omission tests, and reconstruction experiments further indicated that geranyl isovalerate contributed to the herb aroma. Geranyl isovalerate was detected for the first time in Merlot wine from EFHM. The initial documentation of geranyl isovalerate occurred in a medical journal publication in September of 2021, where its biological properties were elucidated. Upon ingestion by the human body, this compound can directly interact with the epithelial cells of the gastrointestinal tract and exhibit anti-cancer properties [43].
The additional experiments showed that perceptual interactions among the key aroma compounds in Merlot wine vary with different geranyl isovalerate concentrations. However, the factors behind the shifting patterns in the intensity of aromas found in herbs and tropical fruits remain uncertain and warrant additional investigation. Additionally, these factors may interact synergistically with other compounds [39], highlighting the need for further research in this area.