Aroma Composition and Sensory Quality of Cabernet Sauvignon Wines Fermented by Indigenous Saccharomyces cerevisiae Strains in the Eastern Base of the Helan Mountain, China

Abstract

In order to improve the regional flavor peculiarity of Cabernet Sauvignon wine in the eastern base of the Helan Mountain, an Appellation of Origin in China, eight indigenous Saccharomyces cerevisiae strains were studied and evaluated. Compared to commercial control F15, indigenous N11424 present significant difference in the concentration of hexyl acetate, isoamyl acetate, ethyl hexanoate, diethyl succinate, and methyl benzoate, and the highest content of esters and least of alcohols were found in the wine made by N11424. By sensory analysis, N11424 wine was characterized with honey, pepper, as well as more fruit flavor, such as peach, cherry, clove, redcurrant, and blackcurrant. By odor activity value analysis, most of the high impact odorants were esters and the odor activity values of esters in most of indigenous strains were higher than in control F15.

Keywords:

• Wine yeast
• Indigenous
• Volatiles
• Sensory quality
• Odor activity value

Introduction

As one of the important attributes in the determination of wine quality, aroma is the combination of up to 600–800 volatile compounds, including varietal compounds, fermentative compounds, and post-fermentative compounds. These aroma volatiles are originated from grape, produced by yeast and bacteria during fermentation and resulted from chemical reactions that occur during wine storage and ageing, respectively. Fermentative volatiles derived from the biosynthesis of yeast make up the largest percentage of total aroma composition of wine.^[1]

At present, commercial S. cerevisiae strains are popularly used in winemaking and most of them are imported. However, in some cases, the commercial S. cerevisiae strains inoculated could not compete successfully with indigenous strains and, therefore, do not dominate the fermentation as expected.^[2] While local selected strains of S. cerevisiae which are better acclimated to micro-area conditions of the wine production region, and easily dominated on the natural biota, is rather advisable as starters,^[3] and finally contribute to the regional characteristics of wine.^[4] The influence of autochthonous S. cerevisiae strains on the aroma profiles has been widely reported.^[5–8] They found some potential native strain with more ester production, and variable production of fermentative volatiles, which may be species- and strain-dependent.^[9,10]

Finally, the difference of aroma composition in terms of numbers and contents, leads to the different sensory quality of wine. It is well known that although plenty of volatile compounds are present in wine, only a few of compounds are actually important to wine aroma.^[11] General approaches to identifying “important” or high impact odorants are based on odor activity values (OAVs) and odorants with low OAVs are considered to be unimportant to the overall sensory perception.^[12] Nevertheless, the compounds with OAVs lower than 1, may also contribute to the wine aroma for the additive effect of similar compounds and synergy with other compounds with similar OAVs.^[13,14] Considering the separation, inhibiting, and masking effect of other aroma compounds, relating volatile compounds concentration and whole sensory data of wine is a complex task. Some correlations have been reported in the literatures, such as the positive correlation of ethyl esters and acetates with fruity aromas, monoterpenes with floral aromas;^[15] the wine organoleptic quality with ethyl acetate.^[16]

China as a wine developing country has 10 major viticultural regions and four of them were recognized as geographical indications of wine, one of which is the eastern base of the Helan Mountain, in Ningxia Autonomous Region located at the northwest of China. In this geographical area, the wine grape occupied 22,000 ha in the eastern base of the Helan Mountain, with an estimated annual wine production of 16,527 tons in 2012. Given the contribution of native yeasts to the geographic characteristics of wines,^[17] this work was based on the previous study of Li et al.^[18] to investigate the preselected Saccharomyces isolates and to select some local strains for improving wine quality which may be helpful for wine geographic characteristics in Ningxia region. In this work, fermentation aroma profiles and sensory characteristics of Cabernet Sauvignon (Vitis vinifera) wines fermented by these strains were investigated and compared.

Material and methods

Yeast Strains and Culture Media

Eight S. cerevisiae strains N12413, N3315, N8316, N349, N1134, N11424, N8422, and N9412, selected for this study were previously isolated from spontaneous fermentations of Cabernet Sauvignon in Ningxia Autonomous Region in 2009 according to Li and others.^[18] These strains were chosen for their good oenological traits in the previous small simulation fermentation (see Supplementary Table S1). The commercial yeast F15 (Lafford, France), popularly applied in the Ningxia wineries, was used as a control. Yeast strains were grown and stored in yeast peptone dextrose medium containing the following: 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% glucose (w/v).

Micro-Vinification

Micro-fermentations were carried out in duplicate at 25–28°C in 1 L Erlenmeyer flasks containing 700 mL sterile filtered (0.45 μm, nylon membrane, Whatman, NJ, USA) must of Cabernet Sauvignon grapes sourced from vineyards in the eastern base of the Helan Mountain by hand. In order to keep aroma precursors homogeneous, Cabernet Sauvignon grapes were selected one by one grain for similar size. Before the inoculation, Cabernet Sauvignon mashes were kept in cold storage (about 4°C) overnight for more aroma precursors. After the separation of grape skins, Cabernet Sauvignon musts were distributed into equivalent amounts. The must contained 229.52 ± 0.57 g/L of initial sugar, 6.43 ± 0.04 g/L of total acidity as tartaric acid, 3.58 ± 0.01 of pH, and SO₂ was added to reach a final concentration of 60 mg/L. Yeasts were cultured aerobically in yeast peptone dextrose (YPD) broth in 100 mL shake cultures at 28°C overnight. After centrifugation at 5000 × g for 5 min at 15°C, the cell sediment was harvested, washed twice with sterile water, and then inoculated to each flask at a rate of 10⁶ CFU/mL. The static batch fermentations were performed at 26°C in duplicate. During fermentation, temperature and reduced sugar were monitored every half day. At the end of the alcoholic fermentation (residual sugar below 4 g/L), all wines were separated from lees and clarified by freezing for 2 days at 2–4°C. After this, the wines were bottled and stored at −20°C until required for analysis.

Chemical Analysis

In addition to sugars and total acidity in musts, general wine parameters (alcohol content, residual sugars, pH, dry extract, total and volatile acidity, free and total SO₂) were also determined in triplicate according to International Organizatin of Vine and Wine (OIV).^[19]

Aroma Volatiles Analysis

Headspace solid-phase microextraction (HS-SPME) was used to extract the wine volatiles following the method proposed and validated by Yue et al.^[20] The Agilent 6890 gas chromatography (GC) was equipped with an Agilent 5975 MS and fitted with a 60 m × 0.25 mm id HP-INNOWAX capillary column with 0.25 μm film thickness (J&W Scientific, Folsom, CA) was employed to separate and identify the aromatic volatiles. The carrier gas was helium at a flow rate of 1 mL/min. Samples were injected by placing the solid-phase microextraction (SPME) fiber at the GC inlet for 8 min with the split-less mode. The oven’s starting temperature was 50°C, which was held for 1 min, then raised to 220°C at a rate of 3°C/min and held at 220°C for 5 min. The mass spectrometry in the electron impact mode (MS/EI) at 70 eV was recorded in the range m/z 20 to 450 U. The mass spectrophotometer was operated in the selective ion mode under auto tune conditions. Volatiles were identified by the mass spectra comparison and the retention time to the pure standards, and the mass spectra in the standard NIST05 library; the quantification of volatile compounds detected was obtained on the calibration curves of respective standards. All pure standards were purchased from Aldrich (Milwaukee, WI, USA) and Fluka (Buchs, Switzerland), with purity greater than 99%.

Wine Sensory Evaluation

The aroma analysis was performed as described by Tao and Zhang.^[21] A panel of tasters, consisted of 18 students had been trained for 4 h per week over 2 months, with “Le Nez du Vin” aroma kit (54 aromas, Yixiangle, Hong Kong) to conduct the wine sensory analysis. In the analysis of a balanced and completed block design, the wine aroma profiles were described by each panelist using five or six terms of “Le Nez du Vin.” And they also were asked to score the intensity of each term using 5-point scale (0: none detected, 1: weak, hardly recognizable note, 2: clear, but inadequate note, 3: clear and regular note, 4: clear and strong note, 5: intense note). The data processed was a mixture of intensity and frequency of detection, which was calculated with the formula proposed:

where F (%) is the detection frequency of an aromatic attribute expressed as percentage of 18 (the number of judges) and I (%) is the average intensity expressed as percentage of the maximum intensity.

Data Analysis

The chemical, sensory, and instrumental data were performed using the Statistical Product and Service Solutions (SPSS) statistical package version 17.0 for Windows (SPSS Inc., Chicago, IL, USA). In order to test significant differences among wine composition (enological parameters and volatile compounds), analysis of variance (one-way ANOVA) Duncan’s multiple range tests were performed. Principal component analysis (PCA) was used to study the typical sensory characteristics of these wines. The correlation analysis was applied to verify the relationship between the sensory descriptors and OAVs of aroma compounds.

Result and discussion

Fermentation Ability

As Fig. 1 showed, all musts inoculated with different strains showed similar fermentation curves. The sugar consumption rate increased in the initial stage of fermentation because of the greatly grown of the inoculated strains; the sugar consumption rate reached the peak value at the third day of wine fermentation; then the rate gradually decreased. The commercial strain F15 fermented fast and completed the wine fermentation for 7 days, with average sugar consumption rate 32.8 g/L·d which was higher than the native yeast strains except of N11424. Concerned to N11424, it also completed the fermentation for 7 days with the highest sugar consumption rate 34.9 g/L·d, slightly faster than other native strains. The sugar consumption was decreasing as the fermentation proceeded; however, the strains N12413, N3315, and N1134 had increasing sugar consumption at the third and fourth day and most of them at the sixth day. Considering the curve of sugar consumption and the time of wine fermentation completed, N11424 was similar to F15 in ferment ability with higher average rate of sugar consumption.

FIGURE 1 (a) The reduced sugar concentration and (b) consumption rate curves of S. cerevisiae strains.

Physicochemical Analysis

The chemical parameters were summarized in Table 1. The result showed that eight native strains could complete fermentation with residual sugar lower than 4 g/L with the exception of N12413 and N349. Among wines fermented by native strains, the difference of total acid, volatile acid, total and free SO₂ among these strains was significant. N11424 produced the significant lower concentration of total acid than others, while N8316 led to the lowest concentration of volatile acid but highest pH. The variation of SO₂ may be resulted by the dissolved oxygen in wine, and chemical intense oxidations of their phenolic and certain aroma compounds.^[22]

TABLE 1 Physicochemical parameters of wines obtained from different S. cerevisiae strains (mean ± SD)

Strains	Ethanol (vol%)	Residual sugars (g/l)	Average sugar consumption rate (g/d)	Total acid (g/l)	Volatile acid (g/l)	Dry extract (g/l)	pH	FSO2 (mg/l)	TSO2 (mg/l)
F15	13.5 ± 0.14a	3.05 ± 0.07b	32.8	8.35 ± 0.07b	0.24 ± 0.01c	33.45 ± 0.07a	3.55 ± 0.01a	13.5 ± 2.1f	42.0 ± 0.0b
N12413	13.9 ± 0.42a	4.60 ± 1.13a	27.4	8.10 ± 0.28b	0.40 ± 0.04a	31.35 ± 1.06a	3.42 ± 0.01a	19.0 ± 2.8a	56.0 ± 4.2c
N3315	13.2 ± 0.00a	3.70 ± 0.14ab	27.1	8.70 ± 0.14b	0.33 ± 0.04ab	31.50 ± 0.14a	3.50 ± 0.01a	27.0 ± 0.0b	65.0 ± 2.8d
N8316	13.95 ± 0.07a	3.70 ± 0.00ab	27.2	8.60 ± 0.14b	0.26 ± 0.00b	32.25 ± 0.35a	3.57 ± 0.02a	14.0 ± 1.4f	46.5 ± 0.7b
N349	13.4 ± 0.28a	4.05 ± 0.78a	27.2	8.70 ± 0.14b	0.32 ± 0.02ab	32.15 ± 0.07a	3.44 ± 0.01a	30. 5 ± 14.9c	66.5 ± 16.3e
N1134	13.9 ± 0.28a	3.80 ± 0.00ab	26.6	8.55 ± 0.07b	0.35 ± 0.06ab	32.80 ± 0.14a	3.52 ± 0.01a	20.0 ± 18.4a	65.0 ± 38.2e
N11424	13.9 ± 0.28a	3.65 ± 0.21ab	34.9	7.60 ± 0.00a	0.40 ± 0.04a	30.50 ± 0.14a	3.52 ± 0.01a	13.5 ± 0.7f	42.0 ± 0.0b
N8422	13.3 ± 0.00a	3.65 ± 0.07ab	30.6	8.60 ± 0.14b	0.37 ± 0.01a	31.40 ± 0.28a	3.44 ± 0.01a	24.0 ± 1.4ab	62.5 ± 0.7d
N9412	14.05 ± 0.21a	3.70 ± 0.14ab	32.9	8.55 ± 0.21b	0.40 ± 0.01a	31.30 ± 0.71a	3.42 ± 0.01a	21.0 ± 4.2a	62.0 ± 5.7e

Values with different roman letters (a–d) in the same column are significantly different according to the Duncan test (p < 0.05).

Volatile Composition of Wines

Yeast metabolism during the alcohol fermentation is the main origin of aroma volatiles and over 60 compounds are known to have an impact on wine flavor (e.g., higher alcohols, esters, acids, carbonyls, thiols).^[23–25] Quantitative data of the volatile compounds analyzed in the experimental wines was shown in Table 2, including esters, alcohols, acids, and carbonyls. As it has been demonstrated in previous researches,^[26,27] the differences in the volatile composition seemed to be quantitative rather than qualitative. Higher alcohols are quantitatively the largest group of volatile compounds in these wines, and they are strictly related with yeast metabolism during fermentation. According to the previous studies, the wine strains affect the concentrations of higher-alcohol in wine, so the difference of the higher-alcohol content in wine could be used to differentiate wine strains.^[28] Among the eight native strains, N349 made the wine with highest level of alcohols, followed by N8316. N9412 and N11424 produced lower alcohol levels than other strains, which would be beneficial to highlight the grape variety aroma with weak intensity of the fusel aroma.^[29,30] Isoamyl alcohol was most abundant consisting almost half content of alcohols in all testing wines; however, no significant difference was observed among these strains. Regarding thiols, methionol is derived from the sulfur amino acid methionine and impart a powerful odor reminiscent of soup, meat, and onions.^[31] Methionol in the treatments with native strains were above the perception threshold 1 mg/L and the wine fermented by N9412 contained the lowest content 1235.0 μg/L, followed by the wine fermented by N11424.

TABLE 2 Concentration of fermentation volatile compounds identified in wines (mean ± SD)

Compounds(μg/l)	F15	N12413	N3315	N8316	N349	N1134	N11424	N8422	N9412	p value
Esters (15)
ethyl acetate	30673.2 ± 595.1ab	30705.9 ± 1658.1ab	31731.4 ± 2334.2ab	31168.9 ± 2326.7ab	34389.2 ± 2229.6ab	32152.3 ± 310.9ab	34712.6 ± 322.5ab	31963.2 ± 4349.6ab	29142.0 ± 149.9ab	0.01
isobutyl acetate	89.7 ± 13.1ab	114.9 ± 18.3bc	126.7 ± 4.1c	103.0 ± 5.8bc	108.1 ± 22.1bc	88.2 ± 6.8ab	93.5 ± 18.4abc	94.9 ± 20.8bc	59.3 ± 0a	0.039
isoamyl acetate	663.1 ± 61.3ab	588.9 ± 61.2ab	651.7 ± 15.1ab	837.6 ± 43.6ab	934.3 ± 61.5ab	271.6 ± 40.5a	1765.3 ± 118c	756.5 ± 49.7ab	398.8 ± 17.4a	0.292
hexyl acetate	7.6 ± 1.3a	13.9 ± 3.2ab	11.3 ± 0.5a	12.3 ± 0.1a	10.2 ± 1.5a	11.3 ± 0.7a	22.7 ± 8.7b	13.9 ± 6.3ab	8.8 ± 0.1a	0.084
phenethyl acetate	14.4 ± 5.7ab	12.7 ± 1.3a	21.1 ± 0.9b	17.1 ± 2.9ab	17.4 ± 2.7ab	11.4 ± 1.3a	13.3 ± 3.1a	16.6 ± 3.7ab	11.6 ± 2.0a	0.116
ethyl butanoate	128.9 ± 15.4ab	176.1 ± 26.9bc	194.0 ± 6.3c	155.3 ± 10.9bc	166.0 ± 35bc	134.8 ± 10.7ab	143.1 ± 35abc	145.4 ± 31.5abc	91.2 ± 1.3a	0.044
ethyl hexanoate	123.7 ± 14.5a	337.5 ± 36.2ab	968.2 ± 41.4c	415.2 ± 48.3ab	714.6 ± 13.3bc	88.8 ± 19.3a	718.6 ± 35.3bc	683.0 ± 17.9bc	425.5 ± 7.6ab	0.03
ethyl lactate	37501.5 ± 1806.5ab	41805.5 ± 766.6a	29981.8 ± 1283.4c	37000.9 ± 1173.8ab	33779.9 ± 6885.8bc	35782.4 ± 1733.5abc	34531.6 ± 1881.3bc	35043.1 ± 539.1bc	30091.3 ± 467.8c	0.001
ethyl octanoate	1080.3 ± 267.8ab	1540.1 ± 168.1ac	1505.0 ± 64.4ac	1547.5 ± 11.6ac	1680.6 ± 269.6c	1234.6 ± 170.7abc	1300.2 ± 333.4abc	1176.2 ± 244.2abc	922.6 ± 101.3b	0.044
ethyl decanoate	80.2 ± 5.1abc	89.3 ± 12.5ab	104.0 ± 14.5bd	129.2 ± 23.2d	124.1 ± 2.1d	74.0 ± 6.5ac	73.2 ± 4.3ac	52.5 ± 6.8c	61.3 ± 13.3ac	0.001
diethyl succinate	2913.7 ± 777.4a	908.6 ± 260.6b	1335.9 ± 57.2b	1116.5 ± 189.5b	1239.9 ± 159.8b	884.8 ± 1.5b	2777.7 ± 98.7b	1096.0 ± 137.2b	1218.9 ± 59.4b	0.001
ethyl dodecanoate	316.2 ± 0.7a	316.2 ± 0.4a	317.3 ± 0.1b	317.4 ± 0.5b	317.1 ± 0.1b	315.9 ± 0.1ac	315.6 ± 0.3ac	315.9 ± 0.4ac	315.3 ± 0.0c	0.134
methyl octanoate	1.2 ± 0.3a	1.7 ± 0.1b	1.8 ± 0.1b	1.8 ± 0.1b	1.7 ± 0.3b	1.4 ± 0.2ab	1.4 ± 0.3ab	1.4 ± 0.2ab	1.0 ± 0.1a	0.023
methyl benzoate	38.5 ± 13.2a	30.5 ± 3.4ab	34.9 ± 1.5ab	30.2 ± 3.8ab	20 ± 15.7ab	29.6 ± 5ab	36.8 ± 6b	21.9 ± 5ab	33.8 ± 0.7ab	0.196
methyl salicylate	0.7 ± 0.3a	0.5 ± 0.0a	0.8 ± 0.0a	0.6 ± 0.1a	0.8 ± 0.0a	0.5 ± 0.0a	0.5 ± 0.2a	0.6 ± 0.2a	0.7 ± 0.1a	0.435
Alcohols (21)
1-propanol	22089.9 ± 278.8a	20832.3 ± 4975.1a	13972.7 ± 626.6bc	22323.5 ± 1857.8a	18271.7 ± 229.6ab	20452.9 ± 1101.0a	18966.4 ± 916.2a	19406.2 ± 1416.4a	12890.1 ± 900.8c	0.007
isobutanol	57765.3 ± 273.2a	56046.5 ± 1635.8a	53393.1 ± 4146.9a	54802.5 ± 1278.3a	71627.6 ± 1938.2a	61454.7 ± 2636.5a	56725.9 ± 31139.0a	55332.6 ± 3735.4a	63924.1 ± 3467.9a	0.059
1-butanol	2187.2 ± 50.8a	8693.5 ± 2514.6e	4450.1 ± 190.5bc	7312.4 ± 402.1de	4471.1 ± 588.0bc	6122.6 ± 853.1cd	2279.0 ± 188.6ab	6716.3 ± 2.4de	2359.0 ± 92.8ab	0
isoamyl alcohol	188023.5 ± 14355.1a	272159.2 ± 98088.2b	252728.5 ± 20680.3ab	239797.6 ± 39043.4ab	243229.0 ± 23293.2ab	210910.0 ± 20130.9ab	205113.9 ± 11049.6ab	242270.1 ± 38737.1ab	169211.3 ± 16535.6a	0.345
1-pentanol	97.8 ± 0.2ab	140.4 ± 20.8c	133.1 ± 5.7c	123.2 ± 4.8ac	126.5 ± 15.4ac	81.9 ± 5.8a	101.9 ± 0.8ab	123.6 ± 13.5ac	89.9 ± 18.6a	0.008
4-methyl-1-pentanol	287.7 ± 54.8a	249.9 ± 39.3a	552.5 ± 62.5a	218.6 ± 6.3a	182.0 ± 16.9a	182.7 ± 21.4a	89.9 ± 6.8a	232.6 ± 34.9a	94.3 ± 11.5a	0.585
2-heptanol	2.4 ± 0.4ab	3.0 ± 0.5ab	11.2 ± 0.5a	6.4 ± 6.7ab	7.5 ± 2.8ab	1.4 ± 0.2b	2.4 ± 0.2ab	8.4 ± 0.1ab	2.4 ± 0.3ab	0.254
(E)-3-hexen-1-ol	75.0 ± 20.1ab	64.7 ± 7.9abc	64.0 ± 5.5abc	57.5 ± 1.9abc	56.2 ± 5.1bc	50.0 ± 9.4c	80.7 ± 9.6a	51.9 ± 8.2bc	46.5 ± 6.1c	0.065
1-hexanol	3164.1 ± 187.5ab	4135.7 ± 474.9cd	4362.1 ± 186.7d	3380.4 ± 65.4ab	3452.4 ± 237.6ac	2696.1 ± 422.8be	2222.9 ± 296.5e	3559. 9 ± 200.4ac	2362.5 ± 330.7e	0
(Z)-3-hexen-1-ol	100.2 ± 25ab	131.2 ± 16.3a	111.4 ± 4.8ab	111.1 ± 3.6ab	101.4 ± 7.3ab	82 ± 13.1bc	119.7 ± 14.4a	112.1 ± 18.8ab	60.8 ± 9.5c	0.02
(E)-2-hexen-1-ol	19.2 ± 4.7a	5.0 ± 1.6b	6.1 ± 0.3bc	5.0 ± 0.4b	3.3 ± 4.7b	1.9 ± 2.7b	11.8 ± 0.9c	4.7 ± 0.9b	5.0 ± 0.9b	0.002
(Z)-2-hexen-1-ol	11.9 ± 11.9a	4.7 ± 0.3a	5.4 ± 0.2a	4.9 ± 0.9a	5.3 ± 1.9a	4.6 ± 1.3a	8.4 ± 0.3a	6.7 ± 0.5a	6.8 ± 1.1a	0.487
1-octen-3-ol	3.4 ± 1ab	3.8 ± 0.3ab	4.2 ± 0.2a	3.1 ± 0.3ab	3.8 ± 0.6ab	2.7 ± 0.5b	2.4 ± 0.4b	3.3 ± 0.9ab	2.6 ± 0.3b	0.132
1-heptanol	33.4 ± 14.4ab	23.7 ± 5.8abc	39.3 ± 1.7a	10.3 ± 0.6c	20.8 ± 13.8abc	27.9 ± 5.7abc	21.9 ± 6abc	26.2 ± 10abc	17.6 ± 4.9bc	0.145
2-ethyl-1-hexanol	0.9 ± 0.2a	1.2 ± 0.4a	1.6 ± 0.1ab	1.9 ± 0ab	2.9 ± 0.6cde	3.2 ± 0.7de	3.7 ± 0.7e	2.5 ± 0.6bcd	2.4 ± 0.2bcd	0.003
2-nonanol	119.8 ± 18.5a	99.7 ± 46.1a	138 ± 0.7a	135.1 ± 34.7a	130.1 ± 28.9a	109.0 ± 16.3a	89.9 ± 1.8a	137.6 ± 8.5a	94.6 ± 3.9a	0.322
1-octanol	11.0 ± 4.8a	9.8 ± 0.1a	11.4 ± 0.5a	8.7 ± 0.5a	9.3 ± 0.3a	7.6 ± 2.0a	9.0 ± 2.3a	9.7 ± 3.2a	7.0 ± 1.0a	0.616
methionol	2287.6 ± 44.8ab	2375.3 ± 155.9a	3050.2 ± 130.6cd	2889.5 ± 198.6c	3242.4 ± 42.5de	2019.3 ± 272.6b	1235.0 ± 140.4f	3406. 1 ± 105.2e	1041.1 ± 4.0f	0
benzyl alcohol	233.7 ± 78.4a	133.3 ± 48.7b	196.7 ± 8.4ab	175.3 ± 34.3ab	201.1 ± 46.4ab	142.1 ± 5.9ab	134.6 ± 18.7b	178.5 ± 7.1ab	123.1 ± 1.9b	0.149
2-phenylethanol	51802.2 ± 3487.1a	46423.3 ± 2048.4a	65020.0 ± 2783.2b	67317.0 ± 2861.5bc	65972.2 ± 2534.0bc	49258.4 ± 1791.7a	37131.0 ± 1234.4d	71765.4 ± 4359.8c	40235.9 ± 841.0d	0
Acids (4)
isobutyric acid	1214.2 ± 245.1a	1036.2 ± 273.4a	1196.9 ± 51.2a	1422.7 ± 213.3a	1402.3 ± 425.4a	1093.0 ± 109.0a	1242.5 ± 678.1a	1329.4 ± 97.3a	762.8 ± 39.5a	0.555
isovaleric acid	3109.7 ± 702.6a	2266.9 ± 716.9ab	2695.3 ± 103.7ab	3165.3 ± 416.7a	2863.9 ± 804.6a	2715.6 ± 42.6ab	2537.0 ± 2.6ab	3328.1 ± 176.5a	1702.6 ± 40.7b	0.105
hexanoic acid	1131.3 ± 48.2a	524.1 ± 67.0ed	1088.6 ± 46.6ab	691.0 ± 136.2cde	948.8 ± 444.4abc	699.2 ± 32.9bcde	485.2 ± 61.3e	892.8 ± 56.6abcd	332.8 ± 33.6e	0.007
octanoic acid	285.6 ± 6.2a	110.4 ± 5.7b	313.7 ± 13.4c	162.0 ± 3.5d	299.3 ± 13.4ac	212.8 ± 16.6e	122.4 ± 5.2b	260.0 ± 14.2f	82.0 ± 5.6g	0
Carbonyl compounds (1)
acetoin	8627.0 ± 553.3a	6131.2 ± 89.2b	3730.9 ± 74.3cd	4283.8 ± 256.5de	5818.7 ± 360.2bg	4673.8 ± 130.2e	3306.5 ± 51.4f	5528.9 ± 180.6g	6828.6 ± 68.2h	0.007

Values with different superscript roman letters (a–h) in the same row are significantly different according to the Duncan test (p < 0.05); the value in the parenthesis is the number of compounds in the related group.

Esters are also one of important volatile compounds related to yeast metabolism. The 15 esters were divided into three groups, including 5 acetate esters, 7 ethyl esters, and 3 methyl esters. N12413 produced the highest quantities of esters (76.64 mg/L), while the maximum of ethyl esters mainly responsible for the fruity flavor was also found in this wine. Nine esters were significantly different among these strains, including two acetate esters (ethyl acetate, isobutyl acetate), six ethyl esters (ethyl butanoate, ethyl hexanoate, ethyl lactate, ethyl octanoate, ethyl decanoate, diethyl succinate), and methyl benzoate. It was demonstrated that the wine yeast exerted a critical impact on the contents of ethyl esters, agreed with Álvarez-Pérez et al.^[32] The dominant esters were ethyl acetate and ethyl lactate in these wines. Ethyl acetate at concentrations lower than 150 mg/L was considered the favorable compound and added complexity to the aroma of wines.^[6] Hexyl acetate related to fruity aroma of green apple and strawberry,^[33] was more abundant in indigenous strains treatment than in commercial strain, and the highest value was found in the wine fermented by N11424. Besides hexyl acetate, there were also four esters (isoamyl acetate, ethyl hexanoate, diethyl succinate, and methyl benzoate) in the wine fermented by N11424 significantly different from the wine by F15.

There were four organic acids quantified in wines obtained with different S. cerevisiae. Among these acids, isobutyric acid, and isovaleric acid are produced as a result of amino acids metabolism of yeast strains, while hexanoic acid and octanoic acid are mainly formed from fatty acids by yeast and bacteria.^[9] Although the production of these organic acids varies with yeast species and strains,^[32] in this case only hexanoic acid and octanoic acid had significant difference among these yeast strains. Acetoin metabolized from diacetyl,^[23] the only ketone detected in the wines, had significant difference among these strains and was highest in the wine fermented by F15.

In recent years, new selection criteria for wine yeast have emerged, and one of these selection criteria is to appropriately enhance wine aroma via the production of volatile compounds such as esters and higher alcohols.^[29] S. cerevisiae strains that produce only small amounts of higher alcohols reduce the intensity of the wine aroma, helping to highlight the varietal aromas. Figure 2 revealed the total concentrations of esters and alcohols among these yeast strains. Compared to the control F15, N12413, and N11424 produced a higher total concentration of esters, 76.64 and 76.48 mg/L, respectively. Only N11424 had a lower total concentration of alcohols (370.64 mg/L) than F15 (397.14 mg/L).

FIGURE 2 The content of total esters and alcohols in the wines fermented by different strains.

Sensory Analysis

By modified frequency (MF) values calculation, there were 28 descriptors with the MF > 20, at least in one wine sample (see Supplementary Table S2). Furthermore, PCA was used to integrate the sensory properties of wines issued with different selected local yeast strains. After deleting some characters with little importance in loading, 13 characters (pear, plum, cherry, strawberry, redcurrant, blackcurrant, clove, honey, green, pepper, filbert, smokey, and caramel) were selected to build the principle components. Figure 3 showed the first two components accounted 34.5 and 30.5% of total variance, respectively. Some floral and fruity terms (pear, cherry, strawberry, redcurrant, clove, and honey) located in the positive part of PC1. Pepper, filbert, and blackcurrant laid in the positive part of PC2. Smokey and green were located in the left quadrant of PC1; while plum, caramel, and strawberry were located in the lower quadrant of PC2. Among these characters, “pepper” is always regarded as varietal aromatic characteristics of Cabernet Sauvignon and related to methoxypyrazines.^[34] However, aroma compounds in wine are very complex. Research on the aroma compounds’ contributions to wine should take into account not only the number of odorants and their relative importance but also the possible existence of synergetic interactions between those odors and with the matrix constituents. So it is possible to hypothesize that the descriptor “pepper” in these wines was influenced by some substances with a depressing or masking effect.^[35]

FIGURE 3 PCA analysis: Global analysis of sensory descriptors of wines fermented with selected strains.

Yeast starters play an unnegligible role on wine sensory properties. Blanco et al.^[36] assessed the influence of two autochthonous strains of S. cerevisiae (XG1 and XG3), a commercial yeast (QA23) and spontaneous fermentation on the sensory properties of wines from Godello and Albariño. They found that aroma descriptors allowed the distinction between wines made with different yeast strains. Albariño made with XG3 and Godello from spontaneous fermentation were the most appreciated wines. Cortés and Blanco^[5] evaluated the influence of five S. cerevisiae strains on the aroma and sensory properties of wine from Treixadura. The results of the sensory analysis confirmed that the aroma of Treixadura wines depended on the inoculated yeast strain. In our study, some sensorial descriptors were also statistically influenced by yeast starters, such as peach, plum, cherry, strawberry, redcurrant, blackcurrant, pepper, and honey (Fig. 3). Indigenous S. cerevisiae N11424 was in right upper quadrant of PC1, which revealed indigenous N11424 fermented the wine was characterized with honey, pepper, as well as more fruit flavor, such as peach, cherry, clove, redcurrant, and blackcurrant. The control F15 and N9412 were both on the positive quadrant of PC1 and negative quadrant of PC2. They were rated high in the attributes of caramel and plum. But they present low flavor of the pepper and blackcurrant, which were typical aromatic characteristics of Cabernet Sauvignon.^[21] Other indigenous strains N349, N1134, N12413, N8316, and N3315 were all on the left quadrant of PC1. Compared to N11424, these strains were rated high in the attributes of typical aromatic characteristics of Cabernet Sauvignon and were scored as less intense in fruit attributes (strawberry, redcurrant, and peach), in concordance with their low concentrations of esters.

OAVs Analysis of Volatile Compounds

Twenty-one volatile compounds, odor threshold (OTH)^[37] for every compound and their OAVs were displayed in Table 3, including 10 esters, eight alcohols, and three acids. Among these odorants, esters are the principle group of wine flavor, which provides positive odor to wine aroma, such as floral flavor, apple, banana, kinds of fruity flavor.^[38] The volatile with highest OAV was ethyl octanoate (pineapple aroma) followed by isoamyl acetate and ethyl hexanoate, in agreement with Losada et al.^[39] in white Verdejo wines. Table 3 showed that some indigenous strains played better roles in some esters production than F15, for example the OAV of ethyl hexanoate produced by native strains was much higher than F15, moreover, the highest OAV of isoamyl acetate was produced by N11424 and the highest OAV of ethyl octanoate was produced by N349. With respect to alcohols, only five compounds (isobutanol, isoamyl alcohol, 2-nonanol, methionol, and 2-phenylethanol) were predicated as positive odorants with OAVs above 1, most of them contributing unpleasant flavor to wine, except of 2-phenylethanol with roses fragrance. The OAVs of 2-phenylethanol were more than one in all wines and the highest one was found in the wine fermented by N8422, least in the wine by N11424. It was noted that some native strains, such as N12413, N8316, N349, N1134, N8422, and N9412, led to the lower content of diethyl succinate, methyl benzoate, isovaleric acid, and/or hexanoic acid than their OTHs. The quantity of esters with OAVs ≥ 1 was more in F15, N11424, and N9412 wines than others, in accordance with the result of their more fruity and floral characters. However, wine aroma was not simple added by aroma of these active odorants. Some volatile compounds with an OAV low than 1, could also contribute to wine flavor by additive effect and synergy with other compounds. So the “green” and “pepper” characteristic in these wines could be explained by additive effect of six carbon atoms compounds with low OAVs (1-hexanol and (Z)-3-hexen-1-ol).^[32]

TABLE 3 OAVs of compounds which concentrations above their corresponding thresholds (OTH) in, at least, one wine

Compounds	Odor descriptor	OTH^a (mg/L)	F15	N12413	N3315	N8316	N349	N1134	N11424	N8422	N9412
Esters (10)
ethyl acetate	pineapple, fruity balsamic	7.5^1	4.1	4.1	4.2	4.2	4.6	4.3	4.6	4.3	3.9
ethyl butanoate	strawberry, apple banana	0.02^2	6.5c	8.8bc	9.7b	7.8c	8.3bc	6.7c	7.2c	7.3c	4.6d
isoamyl acetate	banana,fruity sweet	0.03^2	22.1a	19.6a	21.7a	27.9a	31.1b	9.1c	58.8d	25.2a	13.3ac
ethyl hexanoate	fruity, green apple; floral, violet	0.005^1	24.7d	67.5a	193.6c	83.0b	142.9c	17.8d	143.7c	136.6c	85.1b
ethyl lactate	lactic, raspberry	14^3	2.7	3.0	2.1	2.6	2.4	2.6	2.5	2.5	2.2
ethyl octanoate	fruity, pineapple pear, floral	0.002^1	540.2b	770.0e	752.5e	773.8e	840.3d	617.3c	650.1c	588.1b	461.3a
diethyl succinate	light fruity	1.2^2	2.4b	0.8a	1.1a	0.9a	1.0a	0.7a	2.2b	0.9a	1.0a
methyl benzoate	floral, honey	0.03^3	1.3a	1.0a	1.2a	1.0a	0.7b	1.0a	1.2a	0.7b	1.1a
ethyl decanoate	fruity, pleasant	0.2^3	0.4c	0.5a	0.5a	0.7a	0.6a	0.4c	0.4c	0.3c	0.3c
ethyl dodecanoate	flowery, fruity	1.5^3	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2
Alcohols (8)
isobutanol	fusel, alcohol	40^2	1.4	1.4	1.3	1.4	1.8	1.5	1.4	1.4	1.6
1-propanol	fresh, alcohol	50^3	0.4	0.4	0.3	0.4	0.4	0.4	0.4	0.4	0.3
isoamyl alcohol	whiskey, nail polish	30^2	9.6b	12.4a	11.8a	11.3a	11.4a	10.4a	10.4a	11.4a	9.0b
(Z)-3-hexen-1-ol	green grass, herb	0.4^4	0.3	0.3	0.3	0.3	0.3	0.2	0.3	0.3	0.1
1-pentanol	alcohol, balsamic	0.064^3	1.5b	2.2c	2.1c	1.9bc	2.0c	1.3a	1.6b	1.9bc	1.4b
1-hexanol	green, grass	8^2	0.4	0.5	0.5	0.4	0.4	0.3	0.3	0.4	0.3
2-nonanol	–	0.058^3	2.1ab	1.7a	2.4b	2.3b	2.2b	1.9ab	1.6a	2.4b	1.6a
methionol	raw potato, garlic	1^5	2.3a	2.4a	3.1b	2.9b	3.2b	2.0a	1.2c	3.4b	1.0c
2-phenylethanol	roses	14^2	3.7a	3.3a	4.6b	4.8b	4.7b	3.5a	2.7c	5.1b	2.9c
Acids (3)
isovaleric acid	blue cheese	2^5	1.6a	1.1b	1.3a	1.6a	1.4a	1.4a	1.3a	1.7a	0.9b
hexanoic acid	cheese, rancid	0.42^5	2.7c	1.3a	2.6c	1.7a	2.3c	1.7a	1.2a	2.1ac	0.8a
octanoic acid	rancid, harsh, cheese, fatty acid	0.5^2	0.6b	0.2a	0.6b	0.3a	0.6b	0.4ab	0.2a	0.5b	0.2a

^aOTHs were not determined in this work; the appropriate reference is given as a superscript number. 1: Guth (1997). The matrix was a 10% water/ethanol solution. 2, 4, 5: Cullere et al. (2004) Goméz et al. (2007) and Ferreira et al. (2000). The matrix was an 11% water/ethanol solution containing 7 g/l glycerol and 5 g/l tartaric acid, with the pH adjusted to 3.4 with 1 M NaOH. 3: Li (2006). The matrix was a 12% ethanol/water mixture containing 5 g/l tartaric acid at pH 3.2.

The Relationship of Descriptive Sensory Analysis and OAVs of Aroma Compounds

To verify the relationship between the descriptive sensory descriptors and OAVs of aroma compounds of the wines, Pearson’s correlation coefficient (two-tailed) was used. The OAV of ethyl acetate showed a negative correlation with the sensory character “plum” (r =–0.816; p < 0.05), and positive correlation with blackcurrant (r = 0.712; p < 0.05), honey (r = 0.722; p < 0.05). A strong correlation was detected between the OAV of ethyl octanoate and most of sensory descriptors. This correlation was significant for strawberry (r = –0.88; p < 0.01), redcurrant (r = –0.854; p < 0.01), and blackcurrant (r = 0.877; p < 0.01).

The study about the relationship between aroma compounds in wines and sensory evaluation is a challenge because of the complex interaction among volatiles in wines, including additive, synergy, separation, and masking effects.^[1] In our work, the aroma character with positive correlation to some volatile was somewhat different from the flavor of the individual compound, such as the positive relationship of (Z)-3-hexen-1-ol to cherry term. The similar unexpected result was also found by San-Juan and others,^[40] who reported a negative relationship between (β-damascenone and β-ionone) and fruit note, as well as the positive contribution of branched acids to fruity note. In order to investigate this surprising finding, they made further experiment with aroma reconstitution and omission. They found the unexpected result could be interpreted in terms of aroma profiles, rather than of aroma intensities: the odor concepts may be linked to the existence of well-defined ratios of odorants; compounds of non-fruity or unpleasant aroma, could be effective and positive contributors to the perception of wine fruitiness. From the practical point-of-view, correlation analysis was just a first approximation, the exact contribution of volatiles to some aroma character or wine flavor should be evaluated by experimentation with aroma reconstitution and omission.

Conclusion

In consideration of the chemical and sensorial data, it can be concluded that indigenous S. cerevisiae strains N11424, isolated from vineyards in the eastern base of the Helan Mountain, is able to produce Cabernet Sauvignon wines with high quality in a suitably short time, in no way inferior to the control commercial strain F15. Indigenous S. cerevisiae strains N11424 fermented the wine with higher concentration of esters and lower concentration of alcohols. Moreover, N11424 demonstrated more floral and fruity aromatic characteristics, such as clove, honey, peach, cherry, plum, and strawberry, as well as the typical aromatic characteristics (redcurrant, blackcurrant, pepper) of Cabernet Sauvignon wines in Ningxia Region. Therefore, the autochthonous N11424 has potential to be used in industrial wine production.

Acknowledgments

The authors would like to thank the Center for Viticulture and Enology, China Agricultural University for technical assistance in the completion of the gas chromatography–mass spectrometry (GC–MS) experiment.

FUNDING

The authors are grateful to China Agriculture Research System [CARS-30-jg-3], the National Natural Science Foundation of China (31301541), and the Fundamental Research Funds of the Central Universities (QN2013021, Z109021201) for providing generous financial support for this work.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

The authors are grateful to China Agriculture Research System [CARS-30-jg-3], the National Natural Science Foundation of China (31301541), and the Fundamental Research Funds of the Central Universities (QN2013021, Z109021201) for providing generous financial support for this work.