A comparison of the influence of eight commercial yeast strains on the chemical and sensory profiles of freshly distilled Chinese brandy

Transcription

1 Research article Received: 26 July 2012 Revised: 23 September 2012 Accepted: 4 October 2012 Published online in Wiley Online Library: 7 November 2012 (wileyonlinelibrary.com) DOI /jib.44 A comparison of the influence of eight commercial yeast strains on the chemical and sensory profiles of freshly distilled Chinese brandy Yu Ping Zhao, 1 *JiMingLi, 2 Bao Chun Zhang, 2 Ying Yu, 2 Chun Hua Shen 2 and Pu Song 1 The current study examined eight commercial Saccharomyces cerevisiae strains [Lalvin ICV-D80, Lalvin FC9 EDV, Lalvin QA23, Lalvin RHST, Uvaferm 43, Enoferm Burgundy (BGY), Lalvin EC1118 and Lalvin M69] for their influence on young brandies, with a special emphasis on chemical, volatile and sensory characteristics. Results of the chemical analysis of the fermented wines showed that all of the strains exhibited a similar performance for ethanol production, but titratable acidity was more variable, with the highest being found in the yeast BGY-derived wine and the lowest in the yeast QA23-derived wine. Spirits produced using yeast FC9 EDV showed a significantly higher content of volatile alcohols, esters and acids, and conveyed to the brandy typical fruity and rosy notes. Brandies derived using yeast Uvaferm 43 presented the highest content of total benzene compounds and this brandy was characterized by rosy and onion attributes. Higher levels of varietal compounds and a medium rosy and slightly acidic and rancid nuances were produced when yeast QA23 was used. When using principal component analysis to classify the samples, there were four groups: group 1 ( ICV-D80, Uvaferm 43 and BGY), group 2 (FC9 EDV), group 3 (QA23, RHST and EC1118) and group 4 (M69). This work sheds some light on the flavour complexity owing to the use of different commercial yeasts and provides useful information for the brandy-maker regarding the choice of yeast for the fermentation based on the volatile profile. Copyright 2012 The Keywords: young brandy; yeasts; volatile compounds; sensory analysis; PCA Introduction Brandy is one of the oldest distilled spirits in the world. It originated from the French area of Charentes, and spread to China as early as the 1890 s. The first Chinese brandy was produced in Yantai (Shandong Province), which is located at the same latitude as Bordeaux (France) and is now one of the largest winemaking grape regions in Asia (1). Today brandy is an important segment of the Chinese wine industry, with an annual production of 50,000 tons and sales revenue of over 5 billion Yuan. As brandy has become more popular over the past decades in China, its characteristic and distinct flavour has begun to receive more scrutiny from consumers. Currently, the brandies made by Chinese wineries are consumed mainly by domestic customers (1). Aroma is one of the main characteristics that determine a brandy s organoleptic quality and style. This is the result of the contribution of hundreds of volatile compounds, including higher alcohols, esters, acids, aldehydes, ketones, terpenes, norisoprenoids and volatile phenols that are derived from the volatile chemical compounds arising from the grapes, and the wine-making and distillation process; some are oak-derived (2,3). Undoubtedly the alcoholic fermentation significantly influences the flavour and final quality of the brandy. A great number of volatile components are formed and modulated by yeast during this process, through the release of varietal volatile compounds from the grape precursors, and the synthesis of de novo yeast-derived volatile compounds (4). However, the profiles and production levels of these by-products vary depending on the yeast species and the particular yeast strain (5). Both indigenous and commercial yeasts can be used as starters for the fermentation process. The use of indigenous yeasts has the advantage of conferring on the wine a significant and favourable sensory effect, but spontaneous fermentations sometimes cease prematurely or proceed too slowly. In addition, there is a lack of reproducibility/predictability, which is an important problem when conducting spontaneous wine fermentations at an industrial level. Starters of selected Saccharomyces cerevisiae strains possess the advantage of assuring a rapid and reliable grape juice fermentation and give wines a consistent and predictable quality. Nowadays, there are a number of wine yeast strains that are commercially available. These are useful strains for the brandy-makers and researchers to study in order to understand the potential differences in the production of brandies by using the particular strains. * Correspondence to: Yu Ping Zhao, College of Life Science, Yantai University, Yantai, Shandong , People s Republic of China. water15689@ 163.com 1 College of Life Science, Yantai University, Yantai, Shandong, , People s Republic of China 2 Technology Centre of Changyu Pioneer Wine Company Limited, Yantai, Shandong, , People s Republic of China 315 J. Inst. Brew. 2012; 118: Copyright 2012 The

2 316 However, there have been limited studies on the evaluation of the effect of S. cerevisiae strains on the volatile profiles of brandy. Furthermore, the volatile composition of Chinese brandies is still an area that is under-researched. Thus, in the present study, eight commercial Saccharomyces yeasts were assessed for their influence on the chemical, volatile and sensory characteristics of brandies produced from the same pool of grape must. The aim of this study was to identify the relationship between the chemical composition, volatile profile and sensory characteristics of freshly distilled brandy, and the employment of commercial yeast strains, with a view to comparing the influence of these strains, which not only meet the technological specifications placed upon Chinese brandy, but also possess the most optimal sensory profile for the production of a high-quality brandy. Based on the present investigations, the goal is to offer prospects for the development of wine yeasts with aroma-producing capability, and assist brandy-makers in selecting the proper yeast starters to produce brandy with definable flavour specifications and styles. Materials and methods Yeast strains and grapes A total of eight commercial wine yeast strains Lalvin ICV-D80, Lalvin FC9 EDV, Lalvin QA23, Lalvin RHST, Uvaferm 43, Enoferm Burgundy (BGY), Lalvin EC1118 and Lalvin M69 were included in the present study. The grapes used were a white variety of Ugni Blanc, which is one of the main white grape varieties grown in China. The fruits were picked at the stage of full ripeness during the month of September 2010 in plantations of the Yantai Region (China). Fruits were manually selected and transported to the laboratory on the day of collection. The fruits were immediately frozen and stored at 20 C until use. Alcoholic fermentation Initial grape must was composed of the following: total reducing sugars g/l, total soluble solids 16 Brix, ph 3.13, titratable acidity g/l. The must was inoculated with S. cerevisiae as recommended by the producer (Lallemand, France). Briefly, the dried commercial yeast strains were rehydrated in water (5 volume) with sucrose (10%, w/v) at 40 C for at least 20 min. The yeasts were stirred gently to break up any clumps and added to the must. Fermentations were conducted in triplicate, in 20 L stainless steel vessels at 15 C for 10 days (until the sugar concentration was reduced to <4 g/l). Fermentation progress was monitored by measuring the total reducing sugars (TRS). At the end of the fermentation, natural sedimentation of yeast cells was allowed to take place for 14 days at 15 C. Distillation When alcoholic fermentations ceased, the fermented liquids were immediately distilled in a pot still by double distillation. Distillation took place in two steps. First, the fresh wines were distilled in a pre-concentration step to obtain an intermediate distillate product at an alcohol concentration of ~28% (v/v). During the second distillation, four fractions were collected: heads (1 2% of distillate), heart 1 (47%), heart 2 (41%) and tails (10%), so that in the final brandies (combined heart 1 and 2) an alcohol concentration of ~70% (v/v) was reached (6). In order Y. P. Zhao et al. to avoid the loss of aroma, all the fractions were kept at 4 C until analysed. The young brandies obtained were designated as brandy samples A H, which corresponded to the wines before distillation inoculated with the yeasts of Lalvin ICV-D80, Lalvin FC9 EDV, Lalvin QA23, Lalvin RHST, Uvaferm 43, Enoferm Burgundy (BGY), Lalvin EC1118 and Lalvin M69. Analytical methods After fermentation, the total sugars, reducing sugars, ethanol content, titratable acidity (expressed as g/l of malic acid) and ph were determined according to the methods described in the International Organization of Vine and Wine guidelines (7). All of the analyses were performed in triplicate and the results are expressed as means with their respective standard deviations. Solid-phase microextraction Before solid-phase microextraction (SPME) analysis, brandy samples were diluted to ~14% (v/v) ethanol with deionized water, saturated with sodium chloride. Then, a portion of the 5 ml sample and 5 ml of 3-octanol (internal standard, 100 mg/l in ethanol) were added to a 15 ml headspace vial. A 50/30 mm Divinylbenzene/Carboxen/ Polydimethylsiloxane (DVB/CAR/PDMS) fibre (Supelco, Bellefonte, PA, USA) was applied to extract the volatile compounds from the headspace of the prepared glass vial, which was placed in a thermostatted bath adjusted to 50 C with constant agitation for 30 min. After extraction, the fibre was inserted into the injection port of a GC at 250 C for thermal desorption for 5 min. GC-MS analysis Volatile compounds in the samples were analyzed on an Agilent 6890 N gas chromatograph equipped with an Agilent 5975 mass selective detector on a DB-Wax column (60 m 0.25 mm i.d., 0.25 mm film thickness, J&W Scientific) with splitless injection. The column carrier gas was helium at a constant flow rate of 2 ml/min. The injector and detector temperature were both 250 C. The oven temperature was held at 40 C for 2 min, and then raised to 230 C at a rate of 4 C/min, and held for 5 min. The ion source was set as 230 C, and the electron impact mass spectra were recorded at 70 ev ionization energy. The GC-MS analysis was carried out in the scanning mode in the am mass range. All analyses were conducted in triplicate. Compound identification Compound identification was performed by comparing the mass spectral data with the Wiley 275.L Database (Agilent Technologies, Inc., Santa Clara, CA, USA). Retention indices (RI) of unknown compounds were calculated in accordance with a modified Kovats method (8). Identification of unknown compounds was achieved by comparing mass spectra and RI of the standards or from the literature (9,10). Quantitative analysis For volatile quantification, semi-quantitative analysis by the internal standard method was used. The internal standard (IS) solution of 3-octanol was individually prepared in ethanol and stored at 4 C. Dilutions were made with ethanol at the same wileyonlinelibrary.com/journal/jib Copyright 2012 The J. Inst. Brew. 2012; 118:

3 Effect of yeast on the production of brandy temperature. The integral for all chromatogram peaks used the selective ion method (SIM). Semi-quantitative concentrations = (peak area/is peak area) IS concentration (11). Quantitative descriptive analysis An expert panel composed of nine tasters (five females and four males), ranging in age from 22 to 43 years, carried out the quantitative descriptive analysis. Panellists were recruited according to their motivation and availability. The standards for the initiation and training of assessors in the detection and recognition of odors were prepared in the way described in UNE (1996). These experts took part regularly in sensory analysis of brandy held on a weekly basis. During the test, a brandy sample of 20 ml was presented to judges in a dark glass covered with a plastic dish. Nine sensory terms were selected by the panel to describe the samples: rosy, fruity, acid, almond, grass, hay, rancid, butter and onion. Standards for the selected descriptors were prepared, and a scale from 0 to 5 was used to score each attribute s intensity, where 0 indicated that the descriptor was not perceived, and the values 1 5 indicating that the intensity was very low, low, medium, high and very high, respectively. Statistical analysis Statistical procedures were carried out using the SPSS version 16.0 statistical package (SPSS Inc., Chicago, Illinois). Principal component analysis (PCA) was applied to investigate the possible differences amongst the young brandies. Results and discussion Fermentations of the musts by eight strains The eight S. cerevisiae commercial wine yeast strains were grown under wine fermentation conditions at 15 C, and the results obtained from the controlled fermentations are shown in Table 1. The wines were tested for parameters of technological interest, based on their compliance with the following criteria: residual sugar at a concentration of less than 4 g/l within 14 days of inoculation; and ethanol at a content of 7 8% (v/v). The eight commercial S. cerevisiae strains all finished the alcoholic fermentation within 10 days. The most rapid fermentation occurred in musts inoculated with the yeast Uvaferm 43 and M69. These needed less than 210 h to complete the fermentation. Yeasts ICV-D80, FC9 EDV and EC1118 were characterized by a slower fermentation rate. They needed over 220 h to reach a stable level of residual sugar of below 4 g/l. Similar behaviour amongst all the strains was recorded for ethanol production, with strain QA23-derived wine being the highest (7.49%, v/v) and BGY fermented wines being the lowest (7.01%, v/v). In addition, the ph value of the fresh musts was 3.13, and after fermentation the value changed slightly, varying from 3.10 to The titratable acidity, reported as grams of malic acid per litre of wine, was more variable, with the highest being found in the yeast BGY-derived wine (10.26 g/l) and the lowest in the yeast QA23-derived wine (9.46 g/l). In all the samples, the titratable acidity decreased during fermentation, with an average content of 0.09 g/l, indicating that the majority of the strains of S. cerevisiae used in the current study possessed the capability to metabolise malic acid, in the presence of glucose or other assimilable carbon sources (12). Table 1. The chemical composition of fresh and fermented musts (n = 3) Item Must ICV-D80 a FC9 EDV QA23 RHST Uvaferm 43 BGY EC1118 M69 Total sugars de ab cd a bc f f ef Reducing sugars (g/l) b ab ab b a c d ab ph a a c ab a a ab a Titratable acidity b (g/l) a a a a a a a a Ethanol (% v/v) a a a a a a a a Time needed to complete fermentation (h) a Values with different letters (a f) in the same row are significantly different according to the Duncan test (p < 0.05). b Expressed as g/l of malic acid. 317 J. Inst. Brew. 2012; 118: Copyright 2012 The wileyonlinelibrary.com/journal/jib

4 318 Volatile compounds in young Chinese brandy SPME offers a rapid, solventless and relatively inexpensive method for extracting volatile compounds from different substrates (13 15). It has been applied to many fields of food science, including flavour chemistry and food analysis. In the current study, headspace solid-phase microextraction (HS-SPME) with a 50/30 mm DVB/CAR/ PDMS fibre was applied to extract the volatiles from Chinese brandies. The optimization of the HS-SPME conditions was reported in our preliminary research paper (1), where important parameters including adsorption time and temperature were evaluated according to the total peak areas and peak numbers of volatiles in the brandies, and the best condition involved maintaining the headspace vial at 50 C with constant agitation for 30 min. To obtain brandy samples, fermented musts were distilled, the raw spirits (28% vol.) obtained were fractionated, and three fractions, that is, the heads, the heart fraction and the tails, were collected. The chemical characteristics of the heart fraction (brandy) are presented in Table 2. As seen in this table, a total of 98 volatiles were detected by qualitative analysis. Of these, 86 major volatile compounds were listed, including 30 esters, 18 alcohols, 14 acids, 3 aldehydes and ketones, 6 terpenes and C 13 -norisoprenoids, 12 benzene compounds, and 3 sulphur compounds. Esters, alcohols and acids were the most abundant groups in the young brandies, followed by benzene compounds, terpenes and C 13 -norisoprenoids, and aldehydes and ketones. The B brandies (yeast FC9 EDV) showed a significantly higher total content of volatile alcohols, esters and acids compared with the other samples, while the E brandies (yeast Uvaferm 43) presented the highest content in total volatile benzene compounds among the tested samples. In regard to quantitative and qualitative composition, the brandies obtained were characterized by statistically significant differences. Alcohols Alcohols dominate the group of volatile compounds in brandies and have a significant effect on their sensory characteristics and quality (16). Most of these are formed from amino acids by yeast via the Ehrlich metabolic pathway (17,18), and small amounts are made by the yeast through the reduction of the corresponding aldehydes. The content of volatile alcohols in the tested samples varied significantly, with the highest amount being found in the B brandy and the lowest in A brandy. The total amount of volatile alcohols in the samples was strongly associated with the level of 3-methyl-1-butanol, which constituted from 33.5% (G brandy) to 95.2% (F brandy) of the total alcohols. The compound 2-methyl- 1-propanol also covered a high concentration across the samples, with yeast EC1118 being the highest producer ( mg/l). However, A, E and F spirits had an absence of this compound. The C6-alcohols (1-hexanol, trans-3-hexen-1-ol and cis-3-hexen-1-ol) belong to the group of C6-compounds, which are formed during pre-fermentative steps, including harvesting, transport, crushing and pressing, as well as during must heating and grape maceration (19). Among the C6-alcohols, 1-hexanol is a typical heart product (20), the concentration of which depends on the raw material employed in the distillation (21). This compound usually has a positive influence on the aroma of a distillate when its concentration reaches 20 mg/l (22), but it also will negatively affect the flavour of brandies by imparting a grass-like note, when the concentration exceeds 100 mg/l (2). Taking into account the highest amount of 1-hexanol found in H brandy ( mg/l), Y. P. Zhao et al. which surpassed by almost 6-fold the amount of this compound in the samples fermented by yeast A (the lowest producer), it could be considered that the presence of this compound in the current studies did not damage the brandy flavour based on the concentrations present. Additionally, the spirits obtained with yeast FC9 EDV and QA23 were characterized by a higher amount of heptanol ( mg/l) compared with the samples obtained after the fermentation of musts by yeast BGY (39.01 mg/l), as well as yeast EC1118 (38.05 mg/l). Simultaneously, yeast ICV-D80 produced significantly lower amounts of dodecanol (89.6 mg/l) compared with the other cultures. Esters Esters are mainly produced during ethanol fermentation by yeast, in the reaction between alcohols, and are acetyl-coa catalyzed (12). Most esters have a typical fruity and floral descriptor, and contribute to fruity, sweet, apple, pineapple and floral odours in the brandy (23). In this study, the total concentration of the volatile esters in the B brandy ( mg/l) was clearly greater than the others, almost 4-fold higher than that in F brandy ( mg/l), and ethyl acetate, ethyl hexanoate, 3-methylbutyl acetate, ethyl octanoate, ethyl decanoate, ethyl die-9-enoate and ethyl dodecanoate were identified as the important esters in the samples owing to their higher concentrations. The presence of ethyl acetate makes a significant contribution to the volatile profile of distilled alcoholic beverages. The amount of ethyl acetate quantified in the B brandy ( mg/l) exceeded by over 7 times that detected in the brandies produced using yeast BGY ( mg/l). Ethyl octanoate, responsible for a cooked fruit-like aroma (24), wasdetectedin higher amounts in the B and F brandies. Samples inoculated with M69 contained a higher concentration of ethyl hexanoate ( mg/l), which is responsible for fruit-like aromas in brandies. Furthermore, ethyl decanoate, which gives a grape odour, has been found to be an important aroma compound in these young spirits. There were significant differences with the highest levels in samples of B brandy (2082 mg/l) and the lowest in F brandy ( mg/l). In addition, yeast FC9 EDV-derived brandy was also characterized by a higher amount of ethyl 2-hydroxypentanoate, ethyl tetradecanoate and ethyl hexadec-9-enoate. The H-spirits also produced higher amounts of diethyl propanedioate, ethyl tridicanoate and ethyl hexadecanoate. The C brandy presented the highest levels of hexyl acetate. Furthermore, ethyl propanoate had a notable content in the B, C and D spirits, and can be considered as specific to these brandies. Likewise, methyl hexadicanoate ( mg/l) was observed to be specific to B, C, D and G brandies. Acids The main origin of volatile acids in wines is from the alcoholic fermentation and the content depends on the fermentation conditions, the must nutrient levels and the yeast used (25). Among the acids identified, hexanoic acid, octanoic acid and decanoic acid presented relatively higher amounts (>500 mg/l in most samples). Acetic acid is produced by oxidation of acetaldehyde and its content in an alcoholic beverage depends mainly on the yeast strain applied and, to a lesser extent, on the raw material used (26). The highest content was found in samples using yeast M69 (743.4 mg/l). A smaller amount was detected in the spirits obtained using yeast ICV-D80 (48.65 mg/l). Octanoic wileyonlinelibrary.com/journal/jib Copyright 2012 The J. Inst. Brew. 2012; 118:

5 Effect of yeast on the production of brandy Table 2. Concentrations of volatile compounds in eight spirit samples (in micrograms per litre, n = 3) A a B C D E F G H RI Compounds Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Esters 884 Ethyl acetate b d c b b a a cd Ethyl propanoate a a a n-propyl acetate 15.25a b c b Methylpropyl acetate 77.96c b b a b Ethyl butanoate Methylbutyl acetate cd e d d a b c c Ethyl but-2-enoate 7.03b a Methylbutyl propanoate 10.47c ab b a Ethyl hexanoate b d cd cd a a bc e Hexyl acetate 7.08a a b a a a Hex-3-enyl acetate 3.81a b Ethyl 2-hydroxypropanoate 26.76b a a a ab Ethyl octanoate cd d bc b a a b d Ethyl oct-4-enoate 23.11b a Ethyl 2-hydroxypentanoate 7.04a d b c a a c c Diethyl propanedioate 62.97a b ab b c Ethyl decanoate d e bc cd a a b bcd Methylbutyl octanoate 39.82b a a a Diethyl butanedioate 33.73c ab b b a Ethyl dec-9-enoate d e c c a ab b ab Methylpropyl decanoate b b b a b ab a c bis(1-methylpropyl) methyl-butanedioicate 19.03ab d c b ab Ethyl dodecanoate b d c c a a c d Ethyl tridecanoate 26.49ab cd bc d ab a e f Ethyl tetradecanoate 10.42a d bc b b c Methyl hexadecanoate 41.77d c b a Methyl hexadec-9-enoate 55.89b a Ethyl hexadecanoate 11.54a c b b a a b d Ethyl hexadec-9-enoate c ab b a ab Ethyl linoleate 16.28a c b b d 9.19 Totals cd g ef de b a c f Alcohols Propanol 8.30a cd e d b c Methyl-1-propanol a a a b b Butanol 6.93a b a a c Methy-1-lbutanol a e de b cd c a a Heptan-2-ol 22.13d ab a b c 0.33 (Continues) 319 J. Inst. Brew. 2012; 118: Copyright 2012 The wileyonlinelibrary.com/journal/jib

6 Y. P. Zhao et al. Table 2. (Continued) A a B C D E F G H RI Compounds Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Methyl-1-pentanol, 9.40a b ab ab c Hexanol 82.13a c b bc a a bc d (E)-Hex-3-en-1-ol 3.74a b a Ethoxy-1-propanol 16.63b c b a b (Z)-Hex-3-en-1-ol 18.61b a a a c Heptanol 48.91a cd d c a a a b Nonan-2-ol 17.76c b b a d ,3-Butanediol 11.55a b bc bc a a c d Octanol 10.16a b c cd a a d a Nonanol 46.09a a b a c Dodecanol 89.60a b b c a a d e Decanol 20.26a c b c b a e d Tetradecanol 55.26bc bc ab c a 3.98 Totals a e e cd d cd b bc Acids 1453 Acetic acid 48.65a b b b a a b c Propanoic acid 13.74a Methylpropanoic acid 65.70c a b b c Butanoic acid 5.42a d bc c b bc Methylbutanoic acid 72.60b b b a b Hexanoic acid ab e bc d a a cd d Heptanoic acid 53.40c b b a Octanoic acid c d c b b a Nonanoic acid 25.73a b b b a a b b Decanoic acid c d b b a a a b Decenoic acid c d c c b ab Dodecanoic acid a b c b a a b b Tridecanoic acid 63.83a b Tetradecanoic acid 69.24b a c Totals d e c ab b ab a c Aldehydes and ketones 703 Acetaldehyde 79.29a b c b Methylhept-5-en-2-one 14.80b b b a c Hydroxybutan-2-one 7.63a ab ab b c 1.44 (acetoin) Totals b c c c a 6.30 Terpenes and C13- norisoprenoids 1698 a-terpineol 44.55c a a a b b-citronellol 95.66e b bc c a d e wileyonlinelibrary.com/journal/jib Copyright 2012 The J. Inst. Brew. 2012; 118:

7 Effect of yeast on the production of brandy Table 2. (Continued) A a B C D E F G H RI Compounds Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD Concentration SD 1790 Nerol 10.82c b a trans-b-damascenone 26.49ab bc e cd bc d d Nerolidol 47.23b b a a b Farnesol 46.89b c a a a 3.21 Totals 26.49ab cd e c b a cd d Benzene compounds 1519 Benzaldehyde 4.23a c b c a a d cd Ethyl benzeneacetate 64.91b a a Phenylethyl acetate bc e d de a ab d c Ethyl 3-phenylpropanoate 39.33c d a a b Benzyl alcohol 70.28b b a a a Phenylethanol bc c c ab e d a c Methyl-5-(1-methylethyl)-phenol 51.73c b b a b ,4-bis(1,1-Dimethylethyl)- phenol 81.64b a a ,5-Dimethylanisole 44.31a b a Phenylethyl octanoate 30.32b a c d Benzoic acid 34.25a b b ,5-di-tert-Butyl-4- hydroxybenzaldehyde 37.77b a a 3.51 Totals a b c a d b a a Sulphur compounds 1514 Dihydro-2-methyl-3(2 H) a b a a c 4.31 thiophenone (Methylthio)propanol 55.22e d bc e a b cd Tetrahydro-2- methylthiophene 18.93a 2.99 Totals 98.50d c b c a b c 5.88 a Values with different letters (a f) in the same row are significantly different according to the Duncan test (p < 0.05). 321 J. Inst. Brew. 2012; 118: Copyright 2012 The wileyonlinelibrary.com/journal/jib

8 acid accounted for 59% (E sample) to 28.2% (H sample) of the total acid concentration in these samples and it contributed candy and fruity notes to the brandies (24). However, the D brandy and the G brandy were found to lack this compound. Decanoic acid, which imparts a fatty odour, was also present in higher amounts in the study, with yeast FC9 EDV being the highest producer ( mg/l). Although fatty acids usually confer undesirable odours, they only do this at concentrations above 20 mg/l (2). Thus, in the current investigation, these fatty acids did not appear to damage the flavour profile of young brandies. Y. P. Zhao et al. from yeast Uvaferm 43 were distinguished by a higher amount of 2-phenylethanol ( mg/l), while the others were characterized by a fairly uniform level of this compound ( mg/l). This component is an aroma carrier, which may contribute to the floral nuance of alcoholic beverages, and is produced from L-phenylalanine via metabolic transformations with the participation of yeast under anaerobic conditions (26). Results obtained confirmed that the majority of the S. cerevisiae yeast used in this study were relatively good producers of this aromatic compound. Aldehydes and ketones With regard to aldehydes and ketones, the total content of this volatile group determined in the C, D and G brandies showed higher levels were produced using yeasts QA23, RHST and EC1118 ( mg/l). This was related to the higher concentration of acetaldehyde present. This compound is a potent flavour compound and is commonly present in many alcoholic beverages (27). It is produced by the decarboxylation of pyruvate during the alcoholic fermentation by yeast (28). The ability to produce this compound is a property of different yeast strains, and data available indicate that the majority of the S. cerevisiae used in this study produced relatively high levels of acetaldehyde (29). Terpenes and C 13 -norisoprenoids Terpenes and C 13 -norisoprenoids play a significant role in the sensorial differentiation of wines, because they can influence the wine style by reflecting the flavour characteristics of the grape variety (30). These compounds have a very pleasant aroma and a very low olfactory threshold, so they are readily perceived, even at low concentrations (1). Total values of this group showed higher levels in the C brandy ( mg/l) and the lowest level in the F brandy (10.81 mg/l), while the others had fairly moderate levels. Among the six terpenes and C 13 -norisoprenoids detected, b-damascenone, identified in various types of wines and brandy (31), was detected in all of the tested samples except for the F brandy. This compound has a very low sensorial threshold (0.05 mg/l), and the highest content was achieved in the C brandy ( mg/l). a-terpineol and b-citronellol impart mint and rose notes, respectively, and are also important aromatic compounds in wines and brandies. In the current study, they never exceeded their thresholds. In general, terpenes and C 13 -norisoprenoids are released from their non-odorous precursors (in the form of glycosides) in wine making from grapes. Their contents in the brandies are reported to be connected to the activity of b-glycosidase in the yeast strains. The current investigation revealed a large variation among the tested samples, probably indicating that the b-glycosidase activity varied significantly across the yeast strains tested. PCA of young distilled Chinese brandies When PCA was applied to normalize the relative amounts of the 86 volatile compounds and 24 brandy samples, the first two principal compounds (PC) were extracted, explaining cumulative percent (cum%) of the total variance of the initial data set. The observation of the loading scores suggested that the 64 redundant variables (having coefficients magnitude <0.9) were insufficient to adequately describe the samples according to the yeast strains, so they were removed from the matrix. A new set with 22 volatiles (data matrix 24 22) accounting for 93.7 cum% of the total variance was used. As seen in Fig.1, the first principal component (PC1) was plotted against the second (PC2), and the separation among different samples from this PC1 PC2 scatter point plot was clear. The young brandies produced by yeast QA23 and RHST were located in the first quadrant of the scatter point plot and were more influenced by the variables related to PC1. The compounds 3-hydroxybutan-2-one (0.992), diethyl propanedioate (0.99), acetic acid (0.979), ethyl linoleate (0.965) and hexanol (0.964) were highly oriented towards the positive PC1. The spirits produced by yeast M69 were also strongly associated with the positive PC1; therefore, they were classified into the same group with C and D. The B brandy was located in the second quadrant, which was characterized by a significantly higher amount of the major volatile compounds, especially volatile esters, alcohols and acids in comparison with other samples, and the yeast FC9 EDV can be considered as the highest volatile producer. Samples 322 Benzene compounds For benzene compounds, as seen in Table 2, the total amount of this group presented a significantly higher value in the young brandies produced using the yeast Uvaferm 43 ( mg/l) than samples fermented by other strains. Among this group, 2-phenylethyl acetate and 2-phenylethanol were the major representatives, constituting from 99.81% (E brandy) to 61.66% (C brandy) of the total levels. In particular, the spirits obtained Figure 1. PC1 vs. PC2 scatter point plot of young brandies resulting from eight commercial yeasts and 22 volatile compounds by PCA. V1, 3-hydroxybutan-2-one (acetoin); V2, diethyl propanedioate; V3, 2,3-butanediol; V4, acetic acid; V5, 3-methylpentanol; V6, ethyl linoleate; V7, hexanol; V8, (Z)-hex-3-en-1-ol; V9, 6-methylhept-5-en-2-one; V10, ethyl hexanoate; V11, octanol; V12, ethyl tridecanoate; V13, nonanol; V14, butanol; V15, ethyl hexadecanoate; V16, methyl hexadecanoate; V17, 3,5-di-tert-butyl-4-hydroxybenzaldehyde; V18, 2,4-bis(1, 1-dimethylethyl)- phenol; V19, 9-decenoic acid; V20, heptanoic acid; V21, ethyl oct-4-enoate; V22, methyl hexadec-9-enoate. wileyonlinelibrary.com/journal/jib Copyright 2012 The J. Inst. Brew. 2012; 118:

9 Effect of yeast on the production of brandy butter grass onion Figure 2. almond hay rancid acid fruity rosy Descriptive sensory analysis of eight brandies. A, E and F were characterized by a significantly lower amount of the various volatile components (28 35 in total), including volatile esters, alcohols, acids, aldehydes, terpenes and C 13 -norisoprenoids and benzene compounds. They were located in the third quadrant and were negatively influenced by the variables generated from the PCA analysis. Finally, H brandy produced by yeast M69 was situated in the fourth quadrant, which was strongly associated with higher amounts of acetic acid, ethyl hexanoate and 2,3-butanediol. Quantitative descriptive analysis of young distilled Chinese brandies In order to evaluate the sensory differences of the brandies conveyed by diverse yeast strains, the aroma of the samples was orthonasally assessed by a trained panel composed of nine tasters. The attributes used by the panellists to define the brandies included rosy, fruity, acid, almond, grass, hay, rancid, butter and onion, which were previously agreed upon as the most suitable attributes for the description of sensory characteristics of the tested brandies. Figure 2 shows the average brandy aroma-intensity attributes results obtained using quantitative descriptive analysis. As can be seen, the samples presented differences amongst most sensory attributes, especially in the case of rose, acid, almond, rancid and grass. The H brandies produced from yeast M69 were characterized mostly by fruity, almond and grass attributes, but also presented some aromas of butter and hay, two negative characters. The G spirits revealed similar sensory characteristics as the H brandy in the nuances including almond, hay and butter, but obtained lower scores for the fruity and grass descriptors. The rose nuance stood out in the E brandy obtained from yeast Uvaferm 43; however, this sample was also notable for its onion attribute. All of the judges described the B brandy made with yeast FC9 EDV as the most intense in the fruity and rose characters, but the high acid, rancid and onion nuances could to some extent mask the strong fruit and floral nuances. The C sample was characterized by a middle sense of rosy, but simultaneous detection of acid and rancid nuances masked the rosy character. Moreover, the sensory panel found that the D sample, obtained from yeast RHST, and the F brandy resulting from yeast BGY, showed moderate scores for most of the sensory attributes. The D sample exhibited a stronger fruity sense A B C D E F G H and a moderate sense of almond and grass, while the F brandy gave a greater rancid sensation. Finally, the A brandy produced using yeast ICV-D80 received the lowest scores for almost all of the attributes, with the exception of the fruity nuance, but this observation also revealed positive findings in that the negative nuances (hay and butter) were not perceived by the panellists. Conclusions In conclusion, the volatile profiles of young brandies fermented with different yeast strains varied significantly depending on the particular yeast used for the fermentation. Each strain produced a unique profile of volatile compounds and thus produced wines with diverse sensory characteristics. Yeast FC9 EDV was considered to be a generally high volatile producer, and its corresponding brandy was rated as the highest in rosy and fruity nuances. The young brandy made with yeast M69 contained the highest concentration of acetic acid, ethyl hexanoate and 2,3-butanediol, and its sensory profile was defined by notable fruity, almond and grass characters. Yeast QA23, RHST and EC1118 produced similar amounts of total volatile components, but were characterized by a diverse sensory profile. Brandies A, E and F produced from yeast ICV- D80, Uvaferm 43 and BGY, respectively, contained relatively lower contents of volatile components, especially the A sample, which obtained the lowest scores for almost all of the attributes. Generally, this work has shed more light on the flavour complexity of the different commercial wine yeasts. It helps wine-makers to select a yeast strain for fermentation that offers the potential to modulate the aroma profile to match predetermined specifications or to allow them to more reliably produce wines in the desired style. Acknowledgements This work was financially supported by The Nature Science Foundation of Shandong Province (ZR2011CM026) and The Key Technologies R & D Program of Shandong Province (2009GG ). References 1. Zhao, Y., Xu, Y., Li, J., Fan, W., and Jiang, W. (2009) Profile of volatile compounds in 11 brandies by headspace solid-phase microextraction followed by gas chromatography mass spectrometry. J. Food Sci., 74, C90 C Satora, P., and Tuszynski, T. (2010) Influence of indigenous yeasts on the fermentation and volatile profile of plum brandies. Food Microbiol., 27, Fan, W., Xu, Y., and Yu, A. (2006) Influence of oak chips geographical origin, toast level, dosage and aging time on volatile compounds of apple cider. J. Inst. Brew., 112, Molina, A. M., Guadalupe, V., Varela, C., Swiegers, J. H., Pretorius, I. S., and Agosin, E. (2009) Differential synthesis of fermentative aroma compounds of two related commercial wine yeast strains. Food Chem., 117, Chen, S., and Xu, Y. (2010) The influence of yeast strains on the volatile flavour compounds of Chinese rice wine. J. Inst. Brew., 116, Leaute, R. (1990) Distillation in alambic. Am. J. Enol. Viticulture, 41, Organization of Vine and Wine (2005) Compendium of International Methods of Wine and Must Analysis, Organisation Internationale de la Vigne et du Vin, Paris. 8. Cates, V. E., and Meloan, C. E. (1963) Separation of sulfones by gas chromatography. J. Chromatogr. A, 11, Ledauphin, J., Saint-Clair, J. F., Lablanquie, O., Guichard, H., Founier, N., Guichard, E., and Barillier, D. (2004) Identification of trace volatile 323 J. Inst. Brew. 2012; 118: Copyright 2012 The wileyonlinelibrary.com/journal/jib

10 compounds in freshly distilled calvados and cognac using preparative separations coupled with gas chromatography mass spectrometry. J. Agric. Food Chem., 52, Fan, W., and Qian, M. C. (2006) Identification of aroma compounds in Chinese Yanghe Daqu liquor by normal phase chromatography fractionation followed by gas chromatography/olfactometry. Flav. Frag. J., 21, Caven-Quantrill, D. J., and Buglass, A. J. (2006) Comparison of microscale simultaneous distillation extraction and stir bar sorptive extraction for the determination of volatile organic constituents of grape juice. J. Chromatogr. A, 1117, Sun, S. Y., Jiang, W. G., and Zhao, Y. P. (2011) Evaluation of different Saccharomyces cerevisiae strains on the profile of volatile compounds and polyphenols in cherry wines. Food Chem., 127, Ozhan, D., Anli, R. E., Vural, N., and Bayram, M. (2009) Determination of chloroanisoles and chlorophenols in cork and wine by using HS-SPME and GC-ECD detection. J. Inst. Brew., 115, Wang, L., Xu, Y., Zhao, G., and Li, J. (2004) Rapid analysis of flavour volatiles in apple wine using headspace solid-phase microextraction. J. Inst. Brew., 110, Horak, T., Culik, J., Kellner, V., Jurkova, M., Cejka, P., Haskova, D., and Dvorak, J. (2010) Analysis of selected esters in beer: comparison of solid-phase microextraction and stir bar sorptive extraction. J. Inst. Brew., 116, Plutowska, B., Biernacka, P., and Wardencki, W. (2010) Identification of volatile compounds in raw spirits of different organoleptic quality. J. Inst. Brew., 116, Wondra, M., and Berovic, M. (2001) Analyses of aroma components of chardonnay wine fermented by different yeast strains. Food Technol. Biotechnol., 39, Li, X., Yu, B., Curran, P., and Liu, S.-Q. (2012) Impact of two Williopsis yeast strains on the volatile composition of mango wine. Int. J. Food Sci. Technol., 47, Oliveira, J. M., Faria, M., Sá, F., Barros, F., and Aráujo, I. M. (2006) C6-alcohols as varietal markers for assessment of wine origin. Anal. Chim. Acta, 563, Flouros, A. I., Apostolopoulou, A. A., Demertzis, P. G., and Akrida- Demertzi, K. (2003) Influence of the packaging material on the major volatile compounds of Tsipouro, a traditional Greek distillate. Food Sci. Technol. Int., 9, Y. P. Zhao et al. 21. Dieguez, S. C., De la Pena, M. L. G., and Gomez, E. F. (2005) Volatile composition and sensory characters of commercial Galician orujo spirits. J. Agric. Food Chem., 53, Apostolopoulou, A. A., Flouros, A. I., Demertzis, P. G., and Akrida- Demertzi, K. (2005) Differences in concentration of principal volatile constituents in traditional Greek distillates. Food Control, 16, Ferrari, G., Lablanquie, O., Cantagrel, R., Ledauphin, H., Payot, T., Fournier, N., and Guichard, E. (2004) Determination of key odorant compounds in freshly distilled Cognac using GC-O, GC-MS, and sensory evaluation. J. Agric. Food Chem., 52, González Álvarez, M., González-Barreiro, C., Cancho-Grande, B., and Simal-Gándara, J. (2011) Relationships between Godello white wine sensory properties and its aromatic fingerprinting obtained by GC-MS. Food Chem., 129, Rodríguez-Bencomo, J. J., Cabrera-Valido, H. M., Pérez-Trujillo, J. P., and Cacho, J. (2011) Bound aroma compounds of Gual and Listán blanco grape varieties and their influence in the elaborated wines. Food Chem., 127, Satora, P., Sroka, P., Duda-Chodak, A., Tarko, T., and Tuszyński, T. (2008) The profile of volatile compounds and polyphenols in wines produced from dessert varieties of apples. Food Chem., 111, Geroyiannaki, M., Komaitis, M. E., Stavrakas, D. E., Polysiou, M., Athanasopoulos, P. E., and Spanos, M. (2007) Evaluation of acetaldehyde and methanol in Greek traditional alcoholic beverages from varietal fermented grape pomaces (Vitis vinifera L.). Food Control, 18, Cabaroglu, T., and Yilmaztekin, M. (2011) Methanol and major volatile compounds of Turkish Raki and effect of distillate source. J. Inst. Brew., 117, Erten, H. (2002) Relations between elevated temperatures and fermentation behaviour of Kloeckera apiculata and Saccharomyces cerevisiae associated with winemaking in mixed cultures. World J. Microbiol. Biotechnol., 18, Fan, W. L., and Qian, M. C. (2005) Headspace solid phase microextraction and gas chromatography olfactometry dilution analysis of young and aged Chinese Yanghe Daqu liquors. J. Agric. Food Chem., 53, Xu, Y., Fan, W. L., and Qian, M. C. (2007) Characterization of aroma compounds in apple cider using solvent-assisted flavour evaporation and headspace solid-phase microextraction. J. Agric. Food Chem., 55, wileyonlinelibrary.com/journal/jib Copyright 2012 The J. Inst. Brew. 2012; 118: