Comparison of Aroma-Active Compounds and ... - ACS Publications

Jan 25, 2015 - The Twister bar was constantly stirred for 50 min at a speed of 600 rpm ...... of volatile compounds and polyphenols in cherry wines Fo...
0 downloads 0 Views 736KB Size
Article pubs.acs.org/JAFC

Comparison of Aroma-Active Compounds and Sensory Characteristics of Durian (Durio zibethinus L.) Wines Using Strains of Saccharomyces cerevisiae with Odor Activity Values and Partial LeastSquares Regression JianCai Zhu,† Feng Chen,†,§ LingYing Wang,# YunWei Niu,† Chang Shu,† HeXing Chen,† and ZuoBing Xiao*,† Downloaded via UNIV OF KANSAS on January 22, 2019 at 08:17:04 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



Department of Perfume and Aroma Technology, Shanghai Institute of Technology, Shanghai, China Department of Food, Nutrition, and Packaging Sciences, Clemson University, Clemson, South Carolina 29634, United States # Shanghai Cosmax (China) Cosmetics Company, Ltd., Shanghai, China §

ABSTRACT: The study evaluated the effects of five different strains (GRE, RC212, Lalvin D254, CGMCC2.4, and CGMCC2.23) of the yeast Saccharomyces cerevisiae on the aromatic characteristics of fermented durian musts. In this work, 38 and 43 compounds in durian juices and wines were analyzed by gas chromatography−mass spectrometry (GC-MS) and GC− pulsed flame photometric detection (GC-PFPD) with the aid of stir bar sorptive extraction (SBSE), respectively. According to the measured odor activity values (OAV), only 11 and 15 aroma compounds had OAVs >1 in durian juices or wines, among which 2,3-butanedione, 3-methylbutanol, dimethyl sulfide, dimethyl disulfide, methyl ethyl disulfide, ethyl 2-methylbutanoate, ethyl butanoate, and ethyl octanoate were major contributors to the aroma of juices and wines. Partial least-squares regression (PLSR) was used to detect positive correlations between sensory analysis and aroma compounds. The results showed that the attributes were closely related to aroma compounds. KEYWORDS: Saccharomyces cerevisiae, durian wine, aroma, OAV, PLSR



INTRODUCTION The aroma profile is one of the major characteristics that define quality differences among various fruit wines. The profile is a complex mixture of compounds that are the result of the microbiological conversion of sugars, amino acids, and other chemical components to ethanol, carbon dioxide, and secondary metabolites.1 These metabolites, along with the intrinsic compounds in the fruit, are responsible for characterization and differentiation of aroma in fruit wines. The fruit wine aroma compounds can be influenced by many factors such as fruit variety, geography, and growing circumstances but also depends on yeast strain and the pH of the medium.2 Although each of these factors exerts an important influence on the quality of the fruit wine, the yeast strain plays a key role in the development of aroma in fruit wines during alcoholic fermentation. Different strains of Saccharomyces cerevisiae can produce significantly different aroma profiles when fermenting the same musts. This is a consequence of the differential ability of wine yeast stains in synthesizing yeast-derived volatile compounds.3 Therefore, the selection of the proper yeast strain can be critical for the development of the desired fruit wine style. Durian (Durio zibethinus L.) is a popular and expensive tropical fruit widely grown in Southeast Asia. Durian is called the “king of tropical fruits” due to its superlative flesh, which is rich in nutritional components.4 However, some problems persist in the processing of durian: on the one hand, durian fruit has an excellent, unique flavor, but has a strong, distinctive aroma, which makes it difficult to transport and store. On the © 2015 American Chemical Society

other hand, the price of durian has declined sharply due to an oversupply of the fruit during the durian season. Thus, attempts have been made to add value to the durian fruit crop. One method for solving these problems is turning it into fruit juices or wines. Tropical fruit juices or wines have become popular because an increasing number of people are aware of the health benefits of natural fruit juice.5 Therefore, fruit wines have become alternatives to traditional caffeine-containing beverages such as coffee, tea, or carbonated soft drinks. Several studies on the aroma fractions of durian showed great variability in the concentration of aroma compounds.4,6,7 Durian fruits possess two distinct odor notes, that is, a strong sulfur/onion-like odor and a slight fruit-like odor.6,7 Moreover, a previous study8 demonstrated the impact of nitrogen supplementation on durian wine fermentation by S. cerevisiae. However, to the best of our knowledge, the aroma of durian wine fermented with strains of Saccharomyces and the correlation between samples, sensory attributes, and aroma compounds had not yet been characterized. Therefore, the aims of the present study were (1) to characterize the aroma compounds in durian wine using SBSE followed by capillary GC-MS and GC-PFPD analysis for the first time; (2) to evaluate the influence of different yeast strains of Saccharomyces on the analytical and sensory properties of Received: Revised: Accepted: Published: 1939

November 26, 2014 January 25, 2015 January 25, 2015 January 25, 2015 DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

Article

Journal of Agricultural and Food Chemistry

then filtered with cellulose filters and stored at 5 °C in sealed glass bottles to avoid oxygen entrance. Standard Chemical Analysis. The official methods of the Office Internationale de Vigne et Vin (OIV)10 were employed for the conventional determinations, such as pH, ethanol content, total acidity, free SO2, and total reducing sugar. To analyze glycerol, the Boehringer-Mannheim (Germany) test kit was used. Yeast growth was followed spectrophotometrically (Shimadzu, Kyoto, Japan) by absorbance at 600 nm. Viable cells were determined by plating and counting of colonies on YPG agar. SBSE Adsorption of Aroma Compounds. Each wine sample (5 mL) was placed in a 20 mL vial, in which 20 μL of internal standard solution (2-octanol, 50 mg/L in the ethanolic solution) was added. A stir bar (Twister) coated with polydimethylsiloxane (PDMS) phase (1 cm length, 0.5 mm thickness, Gerstel Inc., Baltimore, MD, USA) was used to extract the aroma compounds from the sample. The Twister bar was constantly stirred for 50 min at a speed of 600 rpm and 40 °C for extraction temperature. After sampling, the Twister bar was rinsed with distilled water, dried with a Kimwipe tissue paper, and placed into the sample holder of the thermal desorption unit (TDU) (Gerstel, Inc.). The analyses were performed using a TDU sampler (Gerstel, Inc.) mounted on an Agilent GC-MS system (7890-5975, Agilent Technologies, Santa Clara, CA, USA). The analytes were thermally desorbed at the TDU in splitless mode, ramping from 45 to 240 °C at a rate of 5 °C/min, and held at the final temperature for 5 min. The desorbed analytes were cryofocused (−80 °C) in a programmed temperature vaporizing (PTV) injector (CIS 4, Gerstel, Inc.) with liquid nitrogen. After desorption, the PTV was heated from −80 to 240 °C at a rate of 10 °C/s and held at 240 °C for 5 min. The solvent vent injection mode was employed. GC-MS Identification of Aroma Compounds. The analyses were performed using a Hewlett-Packard 7890 GC with a 5975 mass selective detector (MSD) (Agilent Technologies) instrument operating under electron ionization (EI) mode (70 eV, ion source temperature 230 °C) with the quadrupole in a scanning mode (scan range was m/z 30−450 at a scan rate of 1 scan/s). Separation of compounds was achieved on Innowax-Wax (60 m × 0.25 mm i.d. × 0.25 μm film thickness, Agilent Technologies) and DB-5 (60 m × 0.25 mm i.d. × 0.25 μm film thickness, Agilent Technologies). Helium (purity = 99.999%) was used as a carrier with a constant flow velocity of 1 mL/min. The quadrupole mass filter was at 150 °C. The transfer line temperature was operated at 250 °C. The oven temperature was held at 35 °C for 3 min, then ramped to 60 °C at the rate of 2 °C/min, and ramped at the rate of 6 °C/min to 240 °C for the last 5 min. The volatile compounds were identified by comparing retention indices and retention times with those obtained for authentic standards, or those of literature data, or with mass spectra in the Wiley and NIST11 libraries. The RIs were determined via sample injection with a homologous series of alkanes (C5−C30) (Sigma-Aldrich, St. Louis, MO, USA). GC-MS Quantitation of Aroma Compounds. To obtain a matrix similar to that of durian wine, model wine was prepared containing 15.0 g/L malic acid, 0.5 g/L lactic acid, 2.0 g/L citric acid, 10.0 g/L fructose, 8.0 g/L glucose, and 12% of ethanol in Milli-Q deionized water. The pH was adjusted to 4.0. Quantitation of the major aroma compounds was carried out by standard curves obtained by each compound from six different concentrations in ethanol. The levels of the aroma compounds were normalized by 2-octanol equivalents. The ethanolic solution of internal standard 2-octanol (20 μL of 50 mg/L) was introduced to the 5 mL of model wine in a 20 mL vial and then extracted by SBSE, as was performed for the durian wines. The standard curves were shown in the research, where y represented the peak area ratio (peak area of volatile standard/peak area of internal standard) and x represented the concentration ratio (concentration of volatile standard/concentration of internal standard). The calibration curves were obtained from Chemstation software (Agilent Technologies Inc.) and used for calculation of volatiles in juices and durian wines.

durian wines; and (3) to establish the relationship between samples, sensory attributes, and aroma compounds using multivariate analysis of PLSR. A better understanding of these points will be helpful for improving the characteristic aroma of durian wine by adjusting fermentation parameters or compensating for typical aroma compounds after alcoholic fermentation.



MATERIALS AND METHODS

Yeast Strains. Five different commercial yeast (Saccharomyces cerevisiae) strains were used in this research. Strains GRE (Y1), Lalvin RC212 (Y3), and Lalvin D254 (Y4) were supplied by Lallemand (France). The result showed that strain GRE was capable of producing high amounts of volatile esters, whereas strains RC212 and D254 showed excellent capability in the production of alcohol compounds.9 Two strains were isolated from Sichuan province, China, named CGMCC2.23 (Y2) and CGMCC2.4 (Y5), and preserved in the Northeast Institute of Science. Chemicals. Acetaldehyde, methanethiol, ethanethiol, dimethyl sulfide, 1-propanol, 2,3-butanedione, propane-1-thiol, ethyl acetate, 2-butenal, acetic acid, 1-butanol, 2-pentanone, dimethyl disulfide, methyl propanoate, (Z)-2-but-2-ene-1-thiol, methyl butanoate, 3methylbutanol, 2-methylbutanol, 2-octanol (internal standard), 2methyl-2-butenal, ethyl 2-methylpropanoate, butanoic acid, 2,3butanediol, ethyl butanoate, propyl propanoate, butyl acetate, 3methylbut-2-ene-1-thiol, 2-methyl-2-pentenal, methyl ethyl disulfide, ethyl 2-butenoate, ethyl thiolacetate, ethyl 2-methylbutanoate, ethyl 3methylbutanoate, diethyl disulfide, propyl 2-methylbutanoate, 3methylbutyl propanoate, 3-methylthio-1-propanol, ethyl hexanoate, propyl 2-methyl-2-butenoic acid, ethyl 2-hexenoate, propyl hexanoate, ethyl heptanoate, dipropyl disulfide, ethyl 2-methylhexanoate, methyl octanoate, 3,5-dimethyl-1,2,4-trithiolane, diethyl trisulfide, and ethyl octanoate were purchased from Sigma-Aldrich (St. Louis, MO, USA). All of them were analytical reagents. Durian Musts. Fresh durian fruits (Durio zibethinus L.) cultivar ‘Monthong’ originating in Malaysia were purchased from Wal-Mart Stores in Shanghai. Only fully ripened fruits without any split on the husk were selected for the study. Five kilograms of durian fruits was washed in clean water to remove plant residue. Next, the pulp was extracted manually by mechanical pressure. Seeds and pulp residue were separated from the juice by centrifugation (RCF = 11000, 12 min, 18 °C). Then, the musts were pasteurized for 30 min at 65 °C, cooled, and poured into 10 L bottles. The initial values of musts were as follows: total sugars, 163 g/L; pH, 6.3. The fruit musts were mixed with a sucrose solution to adjust the total sugars to 210 g/L. The pH was adjusted to 3.5 with 1 mol/L DL-malic acid. Sulfur dioxide, in the form of potassium pyrosulfite, was added into the musts to maintain the concentration of 50 mg/L free SO2 to inhibit bacterial growth. All experiments were carried out in triplicate. Durian Wine Production. The fermentation temperature for durian wine production was approximately 20 °C, and no stirring was performed during any stage of the fermentation. Yeast strains were grown in YPG medium (1% yeast extracts, 2% peptone, and 2% glucose). With a platinum loop, yeasts were inoculated into tubes containing 200 mL of YPG broth at 20 °C until the cell density reached approximately 107 cells/mL. The cells were counted, and an equal amount of cells per strain was resuspended in the same medium for the fermentation. Each vat was then inoculated with 10 mL of prepared suspension to obtain the final cell density of 106 cells/mL. Four liters of durian musts was utilized for durian wine production. All vinifications were carried out in a 10 L bioreactor at 20 °C. Fermentation was monitored through measuring viable cells and total reducing sugars. The fermentation was considered to be ended when the sugar content was below 1 g/L. After fermentation, fermented durian musts were transferred to the 10 L bottles and stored at 5 °C for biomass sedimentation. After 24 h, the durian wines were transferred to new bottles without aeration. After 10 days, wines were 1940

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

Article

Journal of Agricultural and Food Chemistry

Figure 1. Evolution of yeasts (A) (as viable cell counts) and total reducing sugar (B) in durian wine during fermentation.

Table 1. Means (n = 3) of General Parameters of Wines Fermented by Five Different Yeast Strainsa pH ethanol content (% vol) total acidity (g/L) free SO2 (mg/L) glycerol (g/L) reducing sugar (g/L) a

Y1

Y2

Y3

Y4

Y5

3.95 ± 0.01ab 11.2 ± 0.1a 6.2 ± 0.1a 14.2 ± 0.3b 1.4 ± 0.1b 0.64 ± 0.04c

4.08 ± 0.01a 11.3 ± 0.2a 5.9 ± 0.2ab 17.5 ± 0.2b 1.5 ± 0.2a 0.76 ± 0.08b

3.87 ± 0.02b 11.4 ± 0.1a 5.4 ± 0.2b 13.8 ± 0.4b 1.6 ± 0.2a 0.81 ± 0.11ab

4.01 ± 0.01ab 11.5 ± 0.2a 5.7 ± 0.1ab 17.3 ± 0.3a 1.3 ± 0.1c 0.91 ± 0.12a

3.89 ± 0.01b 11.6 ± 0.1a 5.8 ± 0.1ab 17.8 ± 0.2a 1.1 ± 0.3d 0.82 ± 0.09ab

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

Gas Chromatography−PFPD. A HP-5890 series II GC from Agilent equipped with a 5380 PFPD detector from OI Analytical (College Station, TX, USA) was used in the sulfur mode. Separation of compounds was achieved on Innowax-Wax (60 m × 0.25 mm i.d. × 0.25 μm film thickness, Agilent Technologies) and DB-5 (60 m × 0.25 mm i.d. × 0.25 μm film thickness, Agilent Technologies) columns. The oven temperature was held at 35 °C for 3 min, then ramped to 60 °C at the rate of 2 °C/min, and ramped at the rate of 6 °C/min to 240 °C for the last 5 min. GC was operated in a constant flow mode (1 mL/min) with helium as the carrier gas. The PFPD detector was set at 250 °C, and the PMT voltage was set at 500 V. The sulfur-containing compounds were confirmed by comparison with authentic standards on both columns and RI value matching. The quantitation method was identical to the GC-MS analysis as described in the above section. The samples were run in triplicate. Sensory Analysis. A quantitative descriptive sensory analysis was applied for evaluating five fruit wine samples using a 18 cm line scale by a well-trained panel consisting of 20 members (10 females and 10 males, ages 20−30 years). Before the quantitative descriptive analysis, 30 mL of fruit wine was put in a 100 mL volume white china cup covered with a plastic Petri dish and was served to a panelist in a laboratory at a room temperature (25 °C). Twenty judges had discussed aroma compositions of samples through three preliminary sessions (each needs 3 h), until everyone agreed to use them as the attributes. Then, the quantitative descriptive analysis was executed using five sensory attributes (“fruity”, “sulfur”, “sweety”, “off-flavor”, and “harmony”) for all five samples that were randomly divided into two sessions. In every session, samples were randomly presented for every member to avert causing a so-called order effect. All of the samples were evaluated in triplicate. Odor Activity Values (OAV). The contribution of each odor to the overall fruit wine aroma was evaluated by the OAV, which was measured as the ratio of the concentration of each compound to its detection threshold. The threshold values were taken from information available in the references (shown in Table 3). Statistical Analysis. The chemical data and quantitative descriptive sensory analysis were submitted to variance analysis (ANOVA). Duncan’s multiple-comparison tests were applied to

determine significant differences between the samples and sensory attributes. All of the analyses was carried out employing XLSTAT ver. 7.5 (Addinsoft, New York, NY, USA). Partial least-squares regression (PLSR) was employed to explore the correlations between samples, aroma compounds, and sensory attributes using Unscrambler version 9.7 (CAMO ASA, Oslo, Norway). All variables were centered and standardized (1/SD) so as to make each variable have a unit variance and zero mean before applying PLS analyses to finally obtain the unbiased contribution of each variable to the criterion. All PLSR models were validated using full cross-validation.



RESULTS AND DISCUSSION Microvinifications. The five S. cerevisiae (Y1, Y2, Y3, Y4, and Y5) commercial fruit wine yeast strains showed growth indication when incubated at 20 °C (Figure 1). Fermentation parameters were calculated for these strains. Among the five strains, the growth rate of strain Y1 was higher than that of the other strains, and the cell density reached 6.4 × 107 cells/mL on day 6 of growth (Figure 1A). Then the viability decreased, and viability loss was measured at 25.8% after day 8 in culture Y1. Strains Y3 and Y5 presented a cell density increase in the first 4 days of incubation and reached their maximum cell densities of 5.1 × 107 and 4.9 × 107 cells/mL,respectively;thereafter, the cell densities remained almost constant at 1 × 107 cells/mL. Strain Y4 had a 2 day adaptation period (lag phase) and then started to grow at a slower rate than the other strains. Although strain Y4 showed delayed growth, it remained a strain with a rapid growth rate overall. The sugar values in five S. cerevisiae strains displayed similar patterns of rapid initial reduction (Figure 1B). Strains Y1, Y2, and Y5 showed a gradual reduction in sugar value over the 10 day fermentation period and reached a stable sugar level around 1.1−1.2 g/L, indicating that almost all of the sugars were consumed and that the fermentation process was considered to 1941

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

code

b

compound

acetaldehyde methanethiol ethanethiol dimethyl sulfide 1-propanol 2,3-butanedione propane-1-thiol ethyl acetate 2-butenal acetic acid 1-butanol 2-pentanone dimethyl disulfide methyl propanoate (Z)-2-but-2-ene-1-thiol methyl butanoate 3-methylbutanol 2-methylbutanol 2-methyl-2-butenal ethyl 2-methylpropanoate butanoic acid 2,3-butanediol ethyl butanoate propyl propanoate butyl acetate 3-methylbut-2-ene-1-thiol 2-methyl-2-pentenal methyl ethyl disulfide ethyl 2-butenoate ethyl thiolacetate ethyl 2-methylbutanoate ethyl 3-methylbutanoate diethyl disulfide propyl 2-methylbutanoate 3-methylbutyl propanoate 3-methylthio-1-propanol ethyl hexanoate propyl 2-methyl-2-butenoic acid ethyl 2-hexenoate propyl hexanoate ethyl heptanoate

1) for the Main Volatile Compounds in Durian Juices and Wines Fermented by Five Different Yeast Strains code

Oth (mg/L)

juices

Y1

Y2

Y3

Y4

Y5

62.520 15.018

9.680 11.488

30.440 9.140

26.120 8.665

15.240 4.734

22.200 6.767

26 26 27, 28 27, 28 27, 28 26

2.842 51.705 1.636 3.715 1.024 4.367

7.925 11.220 12.993 17.466 0.054 12.549

2.653 44.086 9.362 17.679 0.520 5.792

4.721 36.759 6.250 8.790 0.162 18.486

2.528 12.895 9.222 6.371 0.353 8.484

2.145 21.478 1.485 1.542 0.082 10.578

1

26

7.452

30.764

12.188

17.060

9.834

4.888

propyl propanoate

0.23

26

43.490

32.965

35.611

10.038

9.940

methyl ethyl disulfide ethyl 2methylbutanoate diethyl disulfide 3-methylbutyl propanoate ethyl hexanoate ethyl octanoate

0.062

7

46.068

32.198

197.434

143.174

67.029

109.679

26

14.603

171.807

92.744

109.837

72.865

46.171

sweet caramel, grape

0.0043 0.23

18 27, 28

106.192

14.011 11.598

69.811 6.995

55.062 9.145

18.682 6.946

41.543 5.892

sulfury, roasty, cabbage-like odor pineapple-apricot-like odor

2.3 2

26 27, 28

1.637 1.500

21.653 25.749

13.201 18.673

19.783 20.311

6.620 17.526

4.426 7.567

dimethyl sulfide 2,3-butanedione

C3 C4 C5 C6 C7 C8

5 0.03 8.5 0.1 30 0.2

C9

ethyl acetate dimethyl disulfide methyl propanoate methyl butanoate 3-methylbutanol ethyl 2methylpropanoate ethyl butanoate

C10 C11

C13 C14 C15 C16

IDa 18 26

C1 C2

C12

a

compound

0.027 0.1

0.2

descripors cooked cabbage, onion, corn creamy, sweety, caramel, butter scotch fruity, buttery, orange cooked cabbage, onion fruity odor reminiscent of rum apple-like odor herbaceous and cacao strawberry, fruity, sweet fruity, papaya, butter, sweetish, apple, perfumed complex fruity odor reminiscent of apple banana cooked cabbage, sulfur,onion

fruity, green apple sulfury, heavy, cocoa odor

Code representing the reference number.

methylbutanol, ethyl 2-methylpropanoate, ethyl butanoate, propyl propanoate, methyl ethyl disulfide, ethyl 2-methylbutanoate, diethyl disulfide, 3-methylbutyl propanoate, ethyl hexanoate, 3,5-dimethyl-1,2,4-trithiolane, and ethyl octanoate. Among these compounds, 2,3-butanedione, 3-methylbutanol, dimethyl sulfide, dimethyl disulfide, methyl ethyl disulfide, ethyl 2-methylbutanoate, ethyl butanoate, and ethyl octanoate were major contributors to aroma of juices and wines. 3-Methylbutanol, which was responsible for vinous, herbaceous, and cacao aromas, is formed during fermentation by deamination and decarboxylation reactions from isoleucine.17 Compared to five durian wines, the concentration (30.734 mg/ L) of this compound only slightly surpassed the odor threshold (30 mg/L) in durian juice. In other words, this compound decreased with the progress of fermentation. Obviously, the conclusion was in contradiction with other literature.17 The reason might be that 3-methylbutanol was consumed by the esterification reaction. Correspondingly, the concentration of ethyl 3-methylbutanoate increased during the fermentation process. Ester compounds were responsible for the fruity aroma in fruit juice and wines; in particular, those compounds were correlated with the freshness and fruitiness of new wines.13 These compounds in fruit wines were mainly synthesized from their corresponding precursors by enzymatic ethanolysis with the aid of acyl-CoA during yeast fermentation. Their concentrations were influenced by many parameters, such as yeast strain, fermentation temperature, degree of aeration, and sugar content.18 On the one hand, propyl propanoate, ethyl 2-butenoate, ethyl 2-methylhexanoate, 3-methylbutyl propanoate, propyl 2methyl-2-butenoic acid, ethyl 2-methylhexanoate, and propyl hexanote were detected only in the fermented wines. The phenomenon demonstrated that the function of yeasts in the production of aroma was to synthesize yeast-derived aroma compounds during fermentation. On the other hand, as shown

Figure 2. Quantitative descriptive analysis of durian wines produced by five yeast strains of Saccharomyces cerevisiae. In sensorial parameters indicated by (∗∗∗) a difference among some trials is verified for p < 0.001.

analysis to represent the attributes in samples might be useful. As reported by previous literature,14 the low concentrations of several fermentative aroma compounds could not actually reflect the influence on the perceived aroma intensity in samples due to their low detection thresholds. This phenomenon was particularly evident in tropical fruits. Therefore, the OAV was usually used to provide a rough evaluation of the real contribution of each aroma compound to the global aroma.15 Compounds with OAVs > 1 were commonly at the perception level and considered important aroma compounds contributing to fruit wine aroma.16 According to OAVs (Table 3), only 11 and 15 aroma compounds were analyzed with the OAVs > 1 in juice or durian wines: dimethyl sulfide, 2,3-butanedione, ethyl acetate, dimethyl disulfide, ethyl propanoate, methyl butanoate, 31944

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

Article

Journal of Agricultural and Food Chemistry

Figure 3. Overview of the variation found in the mean data from partial least-squares regression (PLSR) correlation loading plot for five samples. The model was derived from aroma compounds (OAV > 1) as the X-matrix and samples and sensory variables as the Y-matrix. The concentric circles represent R2 = 0.5 and 1.0, respectively.

fermentation, which had also variously been linked to cysteine, cystine, or glutathione metabolism in yeast.18 What is worth mentioning was that 3,5-dimethyl-1,2,4trithiolane seemed to make a significant contribution to the durian wine, but it was likely an artificial compound formed during the desorption process with the method of SBSE.21−23 Thus, the compound 3,5-dimethyl-1,2,4-trithiolane had been excluded as an aroma-active compound in the juices and wines. From Table 2, methanethiol, ethanethiol, propanethiol, and (2Z)-but-2-ene-1-thiol were present only in juices. Although the concentrations of these compounds were relatively low, the contribution to the aroma of the juice might be great due to the low threshold. The result was consistent with the findings of the literature,19 which showed that these sulfur-containing compounds presented a significant contribution to durian juice aroma. The reason that these compounds were absent from fermented durian wines might be the result of strong volatilization or utilization by the yeast strains. Although sulfur compounds were character-impact compounds for durian fruit, these compounds were usually perceived as offensive due to the unpleasant aroma.7,18 On the one hand, it is well-known that the quality and value of fruit wines was strongly related to the characteristic aroma profile developed during fermentation process, which contributed to the strengthening of its characteristic organoleptic aroma. On the other hand, there is no doubt that the off-flavor compounds should be controlled in the fermentation process. Therefore, the balance between characteristic aroma and off-flavor attribute was the key to durian fermentation. Sensory Analysis. Sensory analysis was performed by the evaluation of the organoleptic qualities of durian wines fermented by the different S. cerevisiae strains, using five descriptors including “fruity”, “sulfur”, “sweety”, “off-flavor”, and “harmony” for their aromas. ANOVA was used to distinguish different durian wines by their sensory evaluation scores. The statistical analysis demonstrated that samples fermented by yeast strains showed significant differences in

in Table 3, ethyl 2-methylbutanoate presented the highest value of OAVs, which ranged from 14.603 to 171.807 in juice and five durian wines. As a result, this compound contributed to the juices and wines with banana and pear aroma.17 Another important ester compound was ethyl octanoate, which was far above its threshold in durian wines, indicating that this compound was a significant contributor to fruity aroma of durian wines. Ethyl octanoate showed its highest OAV (25.749) in Y1-derived durian wine. From the above-mentioned, major aroma compounds such as ethyl 2-methylbutanoate and ethyl octanoate were synthesized during the fermentation process. Moreover, strain Y1 led to durian wines with the highest OAVs of ester compounds, suggesting that Y1 strain possessed a relatively higher ability for ester synthesis. Five sulfur-containing aroma compounds were identified across the range of durian wines analyzed, including dimethyl sulfide, dimethyl disulfide, methyl ethyl disulfide, diethyl disulfide, and 3,5-dimethyl-1,2,4-trithiolane. Most of those compounds (dimethyl sulfide, dimethyl disulfide, and diethyl disulfide) decreased rapidly with the progress of fermentation. The reduction of sulfur-containing compounds varied with fermentations, where strain Y1 had the slowest reduction followed by strains Y4 and Y5. The result was according with the findings of literature, which showed that most sulfurcontaining compounds might be consumed or degraded to other compounds during the fermentation.19 Despite the fact that some sulfur aroma compounds were present in only low relative concentrations, as in the case of dimethyl sulfide and diethyl disulfide in this study, these compounds contributed a strong onion-like odor to the fruit wines due to their extremely low thresholds, 0.027 and 0.0043 μg/L, respectively. The sulfur aroma compounds found in this study might serve as character-impact compounds in durian that contributed to its sulfur note.7,20 Dimethyl sulfide was considered as a beneficial compound in low concentrations. The formation of dimethyl sulfide in juice and wine existed not only from its fruit and fruit wine maturation but also during 1945

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

Article

Journal of Agricultural and Food Chemistry

validated variance (Figure 3). The result demonstrated that the optimal number of components in the PLSR model presented was determined as two principal components (PC2): the issue of PC2 versus PC3 results was not presented here, as no additional information was gained by its examination. The estimated regression coefficients from the jack-knife uncertainty test showed that all of the aroma compounds, except 3-methylbutanol (C7) and ethyl 3-methylpropanoate (C8), were significant for one or more of the five samples and five significant sensory descriptors. Five of the sensory attributes were placed between the inner and outer ellipses, R2 = 0.5 and 1.0, respectively. The result indicated that they were well explained by the PLSR model. From Figure 3, samples appeared to be divided into three groups according to strain. Among these samples, the fruit wine fermented by strain Y1 was located on the positive region of PC1 and the negative region of PC2. The fruit wines produced by strains Y4 and Y5, which were located on the negative region of PC1 and PC2, were clearly differentiated from samples fermented by other strains. The wines fermented by strains Y2 and Y3 lay in upper part of PC2. The first PC was mainly defined by the aroma descriptors showing a contrast between “fruity”, “harmony”, and “sweety” attributes on the positive dimension and “sulfur” and “offflavor” attributes on the negative dimension. The fruit wines produced by strains Y4 and Y5 were negatively correlated to all of the sensory variables and all aroma compounds even though they showed a certain aroma profile in the judges’ evaluations thereof. This phenomenon might have been caused by sensory evaluation error, or it might be that those fruit wines could not possess characteristic, or distinguishing, aromas marking them apart from other fruit wines. As compared with other wines, fruit wine fermented by strain Y1 was highly correlated with the sensory attributes of “fruity”, “harmony”, and “sweety”. This was related to aroma compounds such as 2,3-butanedione (C2), ethyl acetate (C3), ethyl propanoate (C5), methyl butanoate (C6), ethyl butanoate (C9), ethyl 2-methylbutanoate (C12), and ethyl octanoate (C16). This phenomenon indicated that strain Y1 should have a higher ethyl acetate yield ability. Some aroma compounds are directly derived from chemical components of the musts, whereas many fruit-derived compounds were released or modified by the action of aroma-active yeast, and a further substantial portion of fruit wine aroma substances resulted from the metabolic activities of these fruit wine microbes.2,25 Therefore, the selection of yeasts was central to the development of fruit wine aroma. In contrast, “sulfur” and “off-flavor” attributes were strongly correlated with each other and showed high loadings in the opposite direction from other attributes. These attributes, which were pronounced in fruit wines fermented by strains Y2 and Y3, were strongly connected with the following compounds: dimethyl sulfide (C1), dimthyl disulfide (C4), methyl ethyl disulfide (C11), and diethyl disulfide (C13).

all attributes, which indicated that these fruit wines had different aroma intensities (p < 0.001). Although judges exhibited significant subjectivity in their use of “fruity”, “sulfur”, “sweety”, “off-flavor”, and “harmony” as attribute descriptors, this result was inevitable in such a quantitative descriptive analysis. The reason might be because the judges expressed their perceptions against different scoring criteria, due to their differences in age, background, and olfactory sensitivity. From the research, no significant interaction between judge and replication was found, indicating that all of the judges were reproducible with regard to the scoring of all attributes in triplicate. Similarly, there was no significant interaction between sample and replication (in any attribute). However, according to the statistical analysis, the interaction between sample and judge showed significant differences for the “fruity” (p < 0.01), “sulfur” (p < 0.001), “off-flavor” (p < 0.01), “sweety” (p < 0.01), and “harmony” (p < 0.001) attributes. This result suggested that the judges were scoring samples not consistent with each attribute. Sensory analysis highlighted that some descriptors were statistically influenced by yeast starters (Figure 2). The durian wines fermented by strains Y1 and Y3 were accompanied by fruity notes more than in other strains. This phenomenon indicated that strains Y1 and Y3 yielded the highest content of compounds able to influence the fruity aroma of the corresponding fruit wine. It is common knowledge that the “fruity” attribute was the fundamental part of the overall perception of the aroma of durian wine. Therefore, the “fruity” attribute was an important symbol of fruit wine quality. The durian wines fermented by strains Y1 and Y2 were mostly associated with a greater “sweety” attribute than wines produced by other strains. The fruit wines fermented by strains Y2 and Y3 were rated as having the highest values of sulfur aroma, whereas durian wine produced by strain Y1 presented low sensorial scores for sulfur aroma. Traditionally, this sulfur aroma was typically associated with a negative sensory contribution to fruit wine (5).7 However, the “sulfur” attribute remained the characteristic aroma of tropical fruit, which could obviously be distinguished from other fruits by this specific property. In addition, a similar behavior was observed for the “off-flavor” attribute in five fruit wines. The reason might be that sulfur-containing aroma compounds were typically perceived as offensive, they had very low detection thresholds, and they generally conferred a negative sensory contribution to fruit wine.5,7,18 The result was consistent with findings of the literature,24 which showed that the sulfurcontaining aroma compounds were the main source of “offflavor” attributes in fruit wines. The highest score under the “harmony” attribute was found in fruit wine produced by strain Y1, whereas the lowest score was found in that produced by strain Y5. Interestingly, by comparing the sensory analysis of “harmony” and “off-flavor” attributes, we found that these two attributes presented completely opposite scores when evaluated by our judges. Relationship between Samples, Aroma Compounds, and Sensory Attributes. PLSR was used to process the mean data accumulated from sensory evaluation by the judges, aroma compounds (OAVs > 1), and samples. The X-matrix was designated as representing the aroma compounds of fruit wines; the Y-matrix was designated as representing the sensory variables and fruit wine samples. The derived PLSR model included two significant PCs explaining 83% of the cross-



AUTHOR INFORMATION

Corresponding Author

*(Z.X.) Mail: School of Perfume and Aroma Technology, Shanghai Institute of Technology, 100 Haiquan Road, Shanghai 200233, People’s Republic of China. Phone: 0086-02160873424. Fax: 0086-021-54487207. E-mail: flavorsit@163. com. 1946

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947

Article

Journal of Agricultural and Food Chemistry Funding

compounds of two related commercial wine yeast strains. Food Chem. 2009, 117, 189−195. (17) Sun, S. Y.; Che, C. Y.; Sun, T. F.; Lv, Z. Z.; He, S. X.; Gu, H. N.; Shen, W. J.; Chi, d. C.; Gao, Y. Evaluation of sequential inoculation of Saccharomyces cerevisiae and Oenococcus oeni strains on the chemical and aromatic profiles of cherry wines. Food Chem. 2013, 138, 2233− 2241. (18) Swiegers, J. H.; Pretorius, I. S. Yeast modulation of wine flavor. Adv. Appl. Microbiol. 2005, 57, 131−175. (19) Lee, P.-R.; Saputra, A.; Yu, B.; Curran, P.; Liu, S.-Q. Biotransformation of durian pulp by mono- and mixed-cultures of Saccharomyces cerevisiae and Williopsis saturnus. LWT−Food Sci. Technol. 2012, 46, 84−90. (20) Dantes, P. T. G.; Maninang, J. S.; Elepaño, A. R.; Gemma, H. In Analysis of Aroma Volatile Profile of Philippine Durian Pulp (Durio zibethinus Rumph. ex Murray) using Headspace Solid Phase Microextraction (HS-SPME) Coupled with GC-MS, 13th ASEAN Food Conference, Singapore; September 9−11, 2013. (21) Granvogl, M.; Christlbauer, M.; Schieberle, P. Quantitation of the intense aroma compound 3-mercapto-2-methylpentan-1-ol in raw and processed onions (Allium cepa) of different origins and in other Allium varieties using a stable isotope dilution assay. J. Agric. Food Chem. 2004, 52, 2797−2802. (22) Wu, C. M.; Wang, Z. Volatile compounds in fresh and processed shiitake mushrooms (Lentinus edodes Sing.). Food Sci. Technol. Res. 2000, 6, 166−170. (23) Granvogl, M.; Schieberle, P. Facts and ‘artefacts’ in the flavour chemistry of onions. Dev. Food Sci. 2006, 43, 359−362. (24) Mestres, M.; Busto, O.; Guasch, J. Analysis of organic sulfur compounds in wine aroma. J. Chromatogr. A 2000, 881, 569−581. (25) Ferreira, V.; López, R.; Cacho, J. F. Quantitative determination of the odorants of young red wines from different grape varieties. J. Sci. Food Agric. 2000, 80, 1659−1667. (26) Giri, A.; Osako, K.; Okamoto, A.; Ohshima, T. Olfactometric characterization of aroma active compounds in fermented fish paste in comparison with fish sauce, fermented soy paste and sauce products. Food Res. Int. 2010, 43, 1027−1040. (27) Näf, R.; Velluz, A. Sulfur compounds and some uncommon esters in durian (Durio zibethinus Murr.). Flavour Fragrance J. 1996, 11, 295−303. (28) van Gemert, L. J. Complilations of Odour Threshold Values in Air, Water and Other Media; Van Setten Kwadraat: Houten, The Netherlands, 2003.

This research was funded by the Natural Science Foundation of Shanghai Institute of Technology (Grant YYY-11607), the National Natural Science Foundation of China (Grants 21476140 and 21306114), and the “Twelfth Five Year” National Science and Technology Support Program Topic (Grant 2011BAD23B01). Notes

The authors declare no competing financial interest.



ABBREVIATIONS USED GC-MS, gas chromatography−mass spectrometry; PFPD, pulsed flame photometric detection; SBSE, stir bar sorptive extraction; OAV, odor activity value; PLSR, partial least-squares regression; PDMS, polydimethylsiloxane; TDU, thermal desorption unit; PTV, programmed temperature vaporizing



REFERENCES

(1) Saberi, S.; Cliff, M. A.; Van, H. J. J. Impact of mixed S. cerevisiae strains on the production of volatiles and estimated sensory profiles of Chardonnay wines. Food Res. Int. 2012, 48, 725−735. (2) Câmara, J. S.; Alves, M. A.; Marques, J. C. Multivariate analysis for the classification and differentiation of Madeira wines according to main grape varieties. Talanta 2006, 68, 1512−1521. (3) Swiegers, J.; Francis, I.; Herderich, M.; Pretorius, I. Meeting consumer expectations through management in vineyard and winery. Wine Ind. J. (Aust. N. Z.) 2006, 21, 34−42. (4) Weenen, H.; Koolhaas, W. E.; Apriyantono, A. Sulfur-containing volatiles of durian fruits (Durio zibethinus Murr.). J. Agric. Food Chem. 1996, 44, 3291−3293. (5) Norjana, I.; Noor Aziah, A. Quality attributes of durian (Durio zibethinus Murr) juice after pectinase enzyme treatment. Food Res. Int. J. 2011, 18, 1117−1122. (6) Li, J.-X.; Schieberle, P.; Steinhaus, M. Characterization of the major odor-active compounds in Thai durian (Durio zibethinus L. ‘Monthong’) by aroma extract dilution analysis and headspace gas chromatography−olfactometry. J. Agric. Food Chem. 2012, 60, 11253− 11262. (7) Voon, Y. Y.; Abdul Hamid, N. S.; Rusul, G.; Osman, A.; Quek, S. Characterisation of Malaysian durian (Durio zibethinus Murr.) cultivars: relationship of physicochemical and flavor properties with sensory properties. Food Chem. 2007, 103, 1217−1227. (8) Lee, P.-R.; Toh, M.; Yu, B.; Curran, P.; Liu, S.-Q. Manipulation of volatile compound transformation in durian wine by nitrogen supplementation. Int. J. Food Sci. Technol. 2013, 48, 650−662. (9) Sun, S. Y.; Jiang, W. G.; Zhao, Y. P. Evaluation of different Saccharomyces cerevisiae strains on the profile of volatile compounds and polyphenols in cherry wines. Food Chem. 2011, 127, 547−555. (10) Compendium of International Methods of Wine and Must Analysis; OIV: Paris, France, 2009. (11) Birch, A. N.; Petersen, M. A.; Arneborg, N.; Hansen, Å. S. Influence of commercial baker’s yeasts on bread aroma profiles. Food Res. Int. 2013, 52, 160−166. (12) Pretorius, I. S. Tailoring wine yeast for the new millennium: novel approaches to the ancient art of winemaking. Yeast 2000, 16, 675−729. (13) Torrens, J.; Urpí, P.; Riu-Aumatell, M.; Vichi, S.; LópezTamames, E.; Buxaderas, S. Different commercial yeast strains affecting the volatile and sensory profile of cava base wine. Int. J. Food Microbiol. 2008, 124, 48−57. (14) Lambrechts, M.; Pretorius, I. Yeast and its importance to wine aroma − a review. S. Afr. J. Enol. Vitic. 2000, 21, 97−129. (15) Guth, H. Quantitation and sensory studies of character impact odorants of different white wine varieties. J. Agric. Food Chem. 1997, 45, 3027−3032. (16) Molina, A. M.; Guadalupe, V.; Varela, C.; Swiegers, J. H.; Pretorius, I. S.; Agosin, E. Differential synthesis of fermentative aroma 1947

DOI: 10.1021/jf505666y J. Agric. Food Chem. 2015, 63, 1939−1947