Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates

Research article Received: 16 February 2016 Revised: 26 July 2016 Accepted: 2 August 2016 Published online in Wiley Online Library: 28 October 2016 (wileyonlinelibrary.com) DOI 10.1002/jib.374 Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates Katarzyna Pielech-Przybylska, 1 Maria Balcerek, 1 * Agnieszka Nowak, 2 Piotr Patelski 1 and Urszula Dziekońska-Kubczak 1 The character of plum brandies depends on a unique aroma profile of the plum and the microbiota present on the surface of the fruits, as well as yeast used for fermentation. In this study, an evaluation of the effect of microorganisms applied for the fermentation of Węgierka Zwykła var. plum mashes and processing temperature (18 C, 30 C) on its efficiency and volatile profile, as well as taste and flavour of distillates obtained was performed. An estimation of the odour activity values (OAVs) of the volatile compounds was also conducted. Regardless of whether the fermentation was carried out using Saccharomyces bayanus wine yeast or by native microflora present on plums as well as raisins, the efficiency of this process was high and ranged between 91.7 and 96.7% of the theoretical efficiency. Especially rich in esters (among others ethyl acetate and isoamyl acetate) was the distillate derived after fermentation with the microflora of plums and raisins, at 18 C. An evaluation of the individual aromatic effect of chemical compounds present in tested distillates, in terms of their OAVs, revealed that the highest OAVs were reached with isovaleraldehyde. Other compounds that showed aroma values >1 and possibly had an effect on the overall aroma of tested plum distillates were the following: hexanal, benzaldehyde, ethyl acetate, isoamyl acetate, ethyl benzoate, ethyl hexanoate, 1-propanol, 2-methyl-1-butanol, 3-methyl-1-butanol and 1-hexanol. The performed sensory ranking showed that the best rated distillate was the one obtained after fermentation with the indigenous microflora of plums and raisins, at 18 C. Copyright 2016 The Keywords: plum; fermentation; raisins; yeast; odour activity value 612 Introduction Like many stone fruits, plums (Prunus domestica L., Rosaceae) are appreciated by consumers all over the world and, consequently, have gained great economic importance. The fruits are cultivated in all temperate regions of the world. The crop is marketed fresh, used as a dried fruit or used to produce juice. It is also commonly used to make jams and other recipes or is fermented to produce wine and brandy (1). Plum brandy (slivovitz) is the popular spirit prepared from fresh Węgierka plums (P. domestica), manufactured in eastern and central Europe, both commercially as well as domestically. Poland has a long tradition of making slivovitz. One of the more recognized of such products is Śliwowica Łącka, which is produced in a submontane region of Poland with specific climatic and soil conditions by means of the spontaneous fermentation of Węgierka Zwykła plums(2,3). Slivovitz can be made using indigenous yeasts found on plums, or with selected pure cultures of yeast. Its character depends on a unique aroma profile of the plum fruits, as well as the yeast used for fermentation or the diverse microbiota present during spontaneous fermentation. Microorganisms isolated from the surface of plums influence the chemical composition of the manufactured spirits. Satora and Tuszyński (4) found that blue plum fruits are colonized mainly by the yeast-like fungi of genus Aureobasidium sp. and Kloeckera apiculata yeasts, which constitute >80% of the fungal microbiota. These microorganisms enter through the must during fruit processing and start the fermentation process. The first phase of fermentation is dominated by representatives of the species K. apiculata and Candida pulcherrima. Also, the above-mentioned authors (5) observed that, as the fermentation progresses, the non-saccharomyces species successively die off, leaving Saccharomyces cerevisiae to dominate and complete the fermentation. Plum distillates, apart from ethanol and water, contain numerous compounds, the concentrations of which vary over an average of 0.5 1.0% (6). During the manufacturing process, the quality of the plum brandy is influenced by many factors such as the characteristics of the fruit varieties, the soil and the climate characteristics, as well as technological procedures, among others (4,7 9). Among numerous volatile compounds, there are components that are important and desirable for the quality of plum brandy (6) as well as undesirable components (10). Chemical compounds that give a beverage its characteristic flavour and aroma can be determined * Correspondence to: Maria Balcerek, Department of Spirit and Yeast Technology, Institute of Fermentation Technology and Microbiology, Faculty of Biotechnology and Food Sciences, Lodz University of Technology, 90-924 Lodz, Wolczanska 171/173, Poland. E-mail: maria.balcerek@p.lodz.pl 1 Department of Spirit and Yeast Technology, Institute of Fermentation Technology and Microbiology, Faculty of Biotechnology and Food Sciences, Lodz University of Technology, 90-924 Lodz, Wolczanska 171/173, Poland 2 Department of Technical Microbiology, Institute of Fermentation Technology and Microbiology, Faculty of Biotechnology and Food Sciences, Lodz University of Technology, 90-924 Lodz, Wolczanska 171/173, Poland J. Inst. Brew. 2016; 122:612 623 Copyright 2016 The

Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates and used to classify the beverage as to the type and country of origin (6). The compounds that must be limited are hydrocyanic acid, methanol (11), as well as ethyl carbamate (12). However, certain amounts of methanol are present in fermented plum distillates and its presence is considered to be proof ( i.e. an indicator) of authentic natural fruit origins (13). The aim of this study was an evaluation of the effect of the microorganisms used for the fermentation of plum mashes and the temperature on process efficiency, the chemical composition and the taste and flavour of the distillates obtained. An estimation of the odour activity values (OAVs) of the volatile compounds present in the distillates was also performed. Experimental Raw materials, microorganisms, and supplements The raw materials used for the production of distillates were mashes prepared from plums var. Węgierka Zwykła that were purchased from a Polish fruit processing factory. Their chemical composition is shown in Table 1. To prepare mashes for fermentation, the stoned plums were homogenized into a pulp. Comminuted stones in the amount of 10% by weight were added to the plum pulp. All mashes were supplemented with (NH 4 ) 2 HPO 4 (0.2 g/kg fruit pulp) as a nitrogen source for the yeast. Fermentations were initiated using: (a) indigenous microflora present on the plums (so called spontaneous fermentation); (b) indigenous microflora present on plums and raisins (1.5 g/kg of plum pulp) to assess their impact on the profile of volatile compounds and the organoleptic characteristics of the plum distillate; and (c) dry wine yeast BC S103 (Saccharomyces bayanus; Fermentis, Div. of S.I. Lesaffre, France) in the amount of 0.3 g/kg of plum pulp, with an alcohol tolerance of up to 18% v/v and a wide fermentation temperature spectrum (10 35 C). Fermentation Alcoholic fermentations of plum mashes were carried out in 4 L flat-bottomed flasks, each containing 3.5 kg of inoculated pulp. The flasks were closed with stoppers, equipped with fermentation pipes filled with glycerol. The process was conducted at 18 or at 30 C, with occasional stirring and measurement of apparent extract (extract of filtered mash containing ethanol) being treated as the index of fermentation dynamics. This parameter was measured using a hydrometer indicating the concentration of dissolved solids, mostly sugar, calibrated in grams of saccharose per kilogram of aqueous solution. The process was continued until the apparent extract measured at 3 h intervals was not subject to change. Accordingly, the following batches were prepared: fermentation with indigenous microflora of plums, 18 C S-18; fermentation with indigenous microflora of plums, 30 C S-30; fermentation with indigenous microflora of plums and raisins, 18 C R-18; Table 1. Chemical composition of plum var. Węgierka Zwykła Parameter Content Total soluble solids ( g/kg) 190.8 ± 6.3 Reducing sugars ( g/kg) 111.3 ± 3.5 Total sugars ( g/kg) 119.5 ± 5.5 ph 3.62 ± 0.2 Acidity (% malic acid) 1.05 ± 0.1 Results expressed as average values ± SE (n =3). fermentation with indigenous microflora of plums and raisins, 30 C R-30; fermentation with S. bayanus wine yeast, 18 C Sb-18; fermentation with S. bayanus wine yeast, 30 C Sb-30. Fermentations of each type were performed in triplicate. Distillation After the completion of the fermentations, volatile compounds (i.e. ethanol and others) were distilled from the mashes using a laboratory distillation unit consisting of a 6 L distillation flask, a Liebig cooler, a receiver of distillate and a thermometer. Distillates containing 14 23% (v/v) ethanol were concentrated (without fractionation) to ~43 ± 1% (v/v) in a glass distillation apparatus consisting of a 2 L distillation flask, a glass distilling column connected to a dephlegmator (according to Golodetz), a Liebig cooler, a receiver of distillate and a thermometer (14), and subjected to chemical analyses. Analytical methods Fermentable sugars, i.e. glucose, fructose and sucrose (after inversion), in the plum mashes were determined by HPLC (Agilent 1260 Infinity, USA) with a Hi-Plex H column (7.7 300 mm, 8 μm; Agilent Technologies, USA), equipped with a refractive index detector at 55 C. Column temperature was maintained at 60 C and 5 mm H 2 SO 4 was used as a mobile phase at a flow rate of 0.7 ml/min with a sample volume of 20 μl. Preparation of liquid samples was performed by filtration through a 0.45 μm polyethersulfone membrane prior to analysis. To perform isolation and identification of fungi present on the surface of plums and raisins, the samples of raw materials were prepared according to ISO 6887 (15). The total fungi counts were determined on DRBC agar (incubation at 25 C, aerobically). The colonies were counted as colony forming units (CFU) per gram of sample. Predominant yeasts were identified using API 20C AUX system (biomérieux, France); also sexual reproduction, hyphae and pseudohyphae formation, fermentation of glucose, raffinose, maltose and assimilation of nitrate and ethanol, were examined. Fungal colonies were isolated and cultured on Czapek Dox medium (16). After incubation, mould strains were microscopically visualized and identified by morphological traits. On completion of fermentation, mashes were analysed for their ethanol concentrations, after distillation in a Digital distilling unit mod named Super Dee (Gibertini, Italy). The ethanol concentration in the distillates was measured using a hydrometer with the scale provided in percentage v/v of ethanol. In mashes after distillation of ethanol, the concentration of the residual sugars was assayed by HPLC. The fermentation efficiency was calculated according to the stoichiometric Gay Lussac equation in relation to total sugars and expressed as a percentage of the theoretical yield (17). Ethanol yield was expressed as the amount of absolute ethanol (A 100 ) obtained from 100 kg of plums (L of A 100 /100 kg plum). Chromatographic analysis of volatile compounds in the obtained distillates was carried out using a gas chromatography (GC) apparatus (Agilent 7890 A, USA) with mass spectrometer (Agilent MSD 5975C, USA). A VF-WAX MS capillary column (60 m length, 0.50 μm film thickness and 0.32 mm i.d.) was used to separate compounds. The GC oven temperature was programmed from 40 C (6 min) to 80 C at a rate of 2 C/min, and then increased to 220 C at a rate of 10 C/min (hold time: 5 min). The flow rate of the carrier gas (helium) through the column was 2.0 ml/min. The temperature of injector (split/splitless) was kept at 250 C. Direct injections of tested distillates (1 μl) were made in split mode (1:40). Each sampling was performed in triplicate. MS conditions were as follows: ion source temperature, 230 C; transfer line temperature, 250 C; and quadrupole temperature, 150 C. The ionization energy was 70 ev. Identification of the volatile components was based on the comparison of their mass spectra with the mass spectra from the NIST MS library (NIST 98.1 and the Wiley Registry of Mass Spectral Data, 8th edition) (18). The comparison of retention indices with reference compounds and literature 613 J. Inst. Brew. 2016; 122: 612 623 Copyright 2016 The wileyonlinelibrary.com/journal/jib

data was also carried out (18 21). The retention indices were calculated according to the formula of van den Dool and Kratz (22) relative to a homologous series of n-alkanes from pentane to octadecane. Quantification of the volatile compounds was carried by calibration curves in the selected ion monitoring mode. Six calibration samples containing different concentrations of each standard compound were prepared. A 4-heptanone sample with a concentration of 45 mg/ml of A 100 in the analysed samples was used as internal standard to monitor the instrument response and retention time stability. Quantitative analysis was performed using software Agilent MassHunter (USA). All gas chromatography standards were purchased from Sigma Aldrich (St Louis, MO, USA) and all were of GC purity. Standard solutions were prepared using anhydrous ethanol (Sigma-Aldrich) as a solvent and refrigerated at 4 C for storage. Free hydrocyanic acid (HCN) content in the tested distillates was determined spectrophotometrically using pyridine pyrazolone reagents (23). The method involves the conversion of HCN to cyanogen chloride with chloramine T solution. As a result of the reaction of this compound with a mixture of pyridine containing 1-phenyl-3-methyl-5-pyrazolone and 4.4-bis(1-phenyl-3-methyl-5-pyrazolone), a coloured complex was formed and measured spectrophotometrically at a wavelength of 490 nm. The amount of hydrocyanic acid in the samples was quantified using the standard solutions prepared from NaCN, ranging from 0 to 1 mg HCN equivalents/l. The content of HCN was determined using a cyanide test kit purchased from Hach Company (Loveland, CO, USA). All other reagents used were of analytical reagent grade. Determination of odour thresholds Odour thresholds of aroma compounds present in tested plum distillates were determined using the gas chromatography olfactory technique. For each volatile compound identified in studied plum distillates, a series of dilutions in 40% (v/v) ethanol water solution were prepared and then evaluated using a sniff detector (ODP-3, Gerstel, Mülheim an der Ruhr, Germany). The GC conditions were the same as those for GC MS analysis. The injection (2 μl) was performed in split mode (1:5). The split ratio of 1:1 was set between the MS and the sniff port. Humidified air was added to the column effluent inside a heated mixing chamber of the sniff port to avoid condensation of high-boiling compounds. Transfer line temperature was set at 155 C. Identification of the components was based on the comparison of their mass spectra with the mass spectra from the NIST MS library, as well as on the retention indices and sensory descriptors. Twelve panellists evaluated the odour of each compound. The samples were repeatedly diluted and sniffed as long as the threshold was not reached. All measurements were carried out in triplicate. Odour activity values of tested aroma compounds were calculated by dividing the concentrations of these compounds in plum distillates with their determined odour thresholds. followed by Tukey s post-hoc test to verify statistical differences at the 0.05 significance level. Results and discussion K. Pielech-Przybylska et al. Before carrying out experiments concerning the effect of microorganisms used for fermentation of plum mashes on the process efficiency, the volatile profile and the taste and flavour of distillates obtained, the isolation and identification of fungi present on the surface of plums and raisins was performed. Results are shown in Figs. 1 and 2. In regard to the microflora present on the surface of plums and raisins used in this study, the sample of plums was found to be contaminated with fungal cells at the level 2.7 10 3 CFU/g. For raisins it reached the level 1.3 10 5 CFU/g. The predominant microflora of both raw materials consisted of moulds. Moreover, relatively high numbers of yeasts were observed. Candida famata and Rhodotorula mucilaginosa were isolated from plums (Fig. 1) and Hanseniaspora guilliermondi and Kloeckera apiculata from raisin samples (Fig. 2). Saccharomyces cerevisiae yeasts were also present on the surface of both raw materials. There is little information in the literature concerning the indigenous yeast microflora of raisins. We suggest that its composition is similar to that occurring on grape surfaces (27). Kloeckera and Hanseniaspora are the predominant species on the surface of grape berries, accounting for roughly 50 ± 75% of the total yeast population (28). The yeast belonging to these species predominate in the early stages of uninoculated fermentation. The latter stages of natural fermentations are invariably dominated by the alcohol-tolerant strains of S. cerevisiae. The microflora isolated during our investigation from plum samples was similar to what was isolated by Satora and Tuszyński (5). The fermentation indices of plum mashes are shown in Table 2. The yield of the plum mash fermentations was calculated relative to the total sugars present in the fresh pulp. The high content of total sugar in the plums used in the study (119.5 ± 5.5 g/kg; Table 1) was advantageous from a technological point of view because it suggests that a high yield of ethanol would be produced from the raw material. When comparing the microflora used and abilities for fermentation, it can be seen that, regardless of whether the fermentation was carried out using the wine yeast S. bayanus or by the native microflora present on the plums or introduced with the raisins, the efficiency of the process was high and ranged from Sensory analysis Samples of plum distillates diluted to 40% v/v were subjected to sensory evaluation by a panel of six qualified assessors, who possess knowledge of spirits, and their quality requirements, pursuant to Polish Standards (24 26). Sensory assessment of obtained plum distillates was performed using the Buxbaum model of positive ranking (6). This model is based on four sensory experiences rated by a maximum of 20 points in overall. In such a test, the judge gives scores for colour, 0 2, clearness, 0 2, aroma (odour), 0 4, and taste, 0 12. Statistical analysis 614 All samples were prepared and analysed in triplicate. Statistical analyses were performed using Statistica 10 software (Statsoft, USA). The results obtained were evaluated using two-way analysis of variance (ANOVA) Figure 1. Fungal microflora of plums. wileyonlinelibrary.com/journal/jib Copyright 2016 The J. Inst. Brew. 2016; 122: 612 623

Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates Figure 2. Fungal microflora of raisins. compared with the mashes fermented with the wine yeast S. bayanus (Table 2). Satora and Tuszyński (29) also stated that the most rapid fermentation occurred in musts inoculated with S. cerevisiae yeast, while the samples fermented by Aureobasidium sp. and K. apiculata strains were characterized by a slower fermentation rate. This was probably connected with a low level of epiphytic microbiota at the beginning of the fermentation as well as successive changes in the yeast population composition owing to the increase in ethanol concentration and the exhaustion of nutrients (30,31). Asimilar effect was found by Herraiz et al. (32), who showed that the presence of K. apiculata strains during fermentation delayed the growth of S. cerevisiae yeast as well as weakening the fermentation kinetics. Strains of the species S. cerevisiae are characterized by greater tolerance to alcohol and simultaneously their ability for ethanol production (33,34). 91.9 ± 2.3 to 96.9 ± 3.1% of the theoretical efficiency (Table 2; p > 0.05). On the other hand, the results of two-way ANOVA showed a significant temperature fermentation effect on the efficiency of the process (p =0.022). For all batches fermented at 18 C, the efficiency of process was >30 C. Furthermore, no significant effect of microflora was observed. Taking into account the intake of the sugars, there were no significant differences amongst values for different temperatures of fermentation (p =0.519), as well as for applied microorganisms (p=0.942). Moreover, no significant effect of a combined action of these factors on the consumption of sugars was detected (p = 0.937). The high fermentation efficiency and sugar intake confirm the correct course of the process, which is also a consequence of the high quality of the processed fruit. Differences were observed in terms of the time necessary for completion of fermentation. At a temperature of 30 C, fermentation time for all of the mashes was shorter. The shortest time of process, lasting ~7 days, was observed for plum mashes fermented with wine yeast S. bayanus at 30 C. The spontaneous fermentation of plum pulp with the contribution of native microflora of fruits as well as of raisins, especially at 18 C, resulted in an elongation of fermentation and the overall process lasted for 11 12 days Chemical composition of the distillates obtained During fermentation, yeast converts sugars into ethanol and carbon dioxide as well as into volatile and non-volatile by-products. The volatile compounds such as carbonyl compounds, organic acids, higher alcohols, esters and others determine the flavour and aroma of alcoholic beverages (35,36). The types and concentrations of these compounds depend, amongst other factors, on the microorganisms activities during the fermentation process (37). Concentrations of volatile aroma compounds in the obtained plum distillates are presented in Table 3. Of the volatile compounds contained in alcoholic beverages are the aldehydes, amongst others acetaldehyde (38 40). There are large species and strain differences in acetaldehyde production by yeasts, for instance, 0.5 286 mg/l for S. cerevisiae and 9.5 66 mg/l for K. apiculata (41). Organoleptic properties of acetaldehyde vary, depending on its concentration. In general, aldehydes can be related to green, grassy and herbaceous notes (42). A high concentration of acetaldehyde, exceeding 125 mg/l, can negatively influence the sensory profile of spirits and other alcoholic beverages (43). In the opinion of Romano et al. (44), the Table 2. Fermentation results of plum mashes Fermentation factor Batch p-value S-18 S-30 R-18 R-30 Sb-18 Sb-30 T* M* T M* Intake of sugars (%) 98.9 A ± 4.6 96.0 A ±4.8 98.1 A ±4.9 96.9 A ± 4.8 98.9 A ± 4.7 97.9 A ± 4.7 0.519 0.942 0.937 Fermentation efficiency 95.6 B ± 3.2 91.9 A ± 2.3 95.6 B ±2.9 93.0 A ± 2.1 96.9 B ± 3.1 93.2 A ± 2.4 0.022 0.767 0.943 (percentage of the theoretical) Ethanol yield 7.4 A ±0.3 7.1 A ±0.3 7.4 A ± 0.3 7.2 A ± 0.3 7.5 A ±0.3 7.2 A ± 0.3 0.084 0.848 0.946 (L A 100 /100 kg plum) Time of fermentation (days) 12 10 11 10 9 7 Results expressed as average values ± SE (n =3). Means in a row with a different superscript letter are significantly different (p < 0.05) as analysed by two-way ANOVA. * T, Temperature effect; M, microflora effect; T M, temperature microflora interaction effect (two-way ANOVA; p < 0.05). Designation of the batches: S-18, fermentation with indigenous microflora of plums, 18 C; S-30, fermentation with indigenous microflora of plums, 30 C; R-18, fermentation with indigenous microflora of plums and raisins, 18 C; R-30, fermentation with indigenous microflora of plums and raisins, 30 C; Sb-18, fermentation with Saccharomyces bayanus wine yeast, 18 C; Sb-30, fermentation with S. bayanus wine yeast, 30 C. 615 J. Inst. Brew. 2016; 122: 612 623 Copyright 2016 The wileyonlinelibrary.com/journal/jib

K. Pielech-Przybylska et al. 616 Table 3. Concentrations of volatile aroma compounds (mg/l alcohol 40% v/v) in the tested plum distillates Chemical compound Retention index Plum distillate/batch p-value S-18 S-30 R-18 R-30 Sb-18 Sb-30 T M T M Aldehydes Acetaldehyde 712 53.94 a ±5.98 59.62 a ±5.49 61.27 a ±4.84 53.36 a ±6.25 62.98 a ±7.32 61.07 a ± 5.45 0.631 0.280 0.182 Propionaldehyde 800 0.45 b ±0.02 0.90 d ± 0.03 0.75 a ± 0.02 0.64 c ±0.01 0.81 a ± 0.03 1.08 e ±0.02 <0.001 <0.001 <0.001 Isovaleraldehyde 921 5.08 a ± 0.22 5.75 BC ±0.15 6.13 cd ±0.18 6.64 de ± 0.24 5.14 ab ± 0.26 7.16 e ±0.22 <0.001 <0.001 <0.001 Hexanal 1080 4.39 c ± 0.06 3.14 a ± 0.03 5.00 b ±0.18 3.53 a ±0.14 5.15 b ±0.34 3.22 a ±0.02 <0.001 <0.001 0.014 Furfural 1470 25.64 BC ± 6.77 26.06 b ±8.96 18.77 a ± 5.96 19.42 ac ±8.74 26.93 b ±5.91 18.52 a ± 1.76 0.052 0.002 0.010 Benzaldehyde 1522 1.86 b ± 0.34 0.89 a ±0.14 3.18 c ±1.42 0.98 a ±0.13 1.82 b ±0.15 0.89 a ±0.20 <0.001 <0.001 0.001 Esters Ethyl acetate 892 1032.16 b ± 81.04 471.69 a ± 9.82 1750.64 c ± 157.59 585.84 a ± 84.88 1029.82 b ± 309.11 440.25 a ± 33.00 <0.001 <0.001 0.007 Isoamyl acetate 1124 3.78 b ± 0.06 1.68 a ±0.06 5.32 c ±1.18 1.34 a ±0.09 3.37 b ±0.47 1.16 a ±0.10 <0.001 0.014 0.015 Hexyl acetate 1269 0.38 b ±0.01 0.14 ac ± 0.01 0.38 b ±0.11 0.10 a ±0.01 0.26 BC ±0.04 0.09 a ±0.01 <0.001 0.026 0.181 Ethyl butanoate 1036 0.36 ab ± 0.01 0.34 a ± 0.01 0.38 b ±0.02 0.34 a ±0.01 0.38 b ± 0.02 0.37 ab ± 0.01 0.004 0.030 0.215 Methyl benzoate 1614 0.18 b ± 0.01 0.13 a ± 0.01 0.10 a ±0.01 0.11 a ±0.02 0.13 a ±0.01 0.13 a ± 0.01 0.040 <0.001 0.003 Ethyl benzoate 1656 6.08 e ±0.24 4.86 d ± 0.08 3.56 a ± 0.03 3.92 ab ±0.13 4.30 BC ± 0.20 4.72 cd ± 0.30 0.124 <0.001 <0.001 Ethyl hexanoate 1240 1.88 a ± 0.08 1.89 a ± 0.17 1.48 a ±0.23 1.42 a ±0.13 1.91 a ±0.28 1.60 a ± 0.20 0.212 0.006 0.352 Ethyl octanoate 1440 6.65 c ±0.24 6.48 cd ±0.29 4.66 ab ±0.72 4.46 a ±0.47 5.41 abc ±0.22 5.69 bcd ± 0.16 0.876 <0.001 0.525 Alcohols Methanol 908 5856.44 a ± 383.24 6305.42 a ± 310.11 5682.50 a ± 675.06 6068.57 a ± 515.50 6355.32 a ± 315.73 6814.19 a ± 335.96 0.170 0.207 0.585 1-Propanol 1040 759.97 b ± 26.91 932.37 ab ± 99.99 882.26 ab ± 33.21 1060.50 a ± 104.53 985.07 a ± 50.84 998.68 a ± 56.72 0.003 0.007 0.104 2-Methyl-1-propanol 1090 243.94 c ± 24.37 559.64 c ± 22.96 328.10 ac ± 16.92 382.46 ab ± 41.90 364.99 ab ± 33.49 456.73 b ± 56.92 <0.001 0.040 <0.001 1-Butanol 1140 4.48 a ±0.13 6.51 cd ±1.14 5.00 ab ±0.25 6.82 d ±0.04 5.48 abc ±0.07 5.86 bcd ±0.06 <0.001 0.357 0.023 2-Methyl-1-butanol 1204 205.47 c ± 8.68 221.69 c ± 2.12 158.96 a ± 3.77 163.32 ab ± 7.15 160.67 a ± 6.38 178.13 b ± 7.17 0.001 <0.001 0.180 3-Methyl-1-butanol 1207 685.02 c ± 35.95 577.80 b ± 25.10 381.68 a ±4.11 419.19 a ± 23.88 381.32 a ± 16.66 375.78 a ± 2.80 0.030 <0.001 <0.001 1-Hexanol 1355 28.69 c ± 0.35 26.04 ab ±0.14 25.03 a ± 1.20 21.90 e ±1.03 27.33 BC ±0.67 19.85 d ±0.26 <0.001 <0.001 <0.001 Benzyl alcohol 1867 5.42 c ± 0.49 7.42 BC ±0.17 12.06 a ±2.86 12.87 a ± 0.59 9.39 ab ± 1.13 9.78 ab ± 0.60 0.114 <0.001 0.585 Others Hydrocyanic acid 0.45 BC ± 0.05 0.55 c ± 0.02 0.25 a ±0.05 0.50 BC ±0.08 0.20 a ± 0.05 0.35 ab ±0.07 <0.001 <0.001 0.105 For designation of the batches: see Table 2. Means in a row with a different superscript letters are significantly different (p < 0.05) as analysed by two-way ANOVA and the Tukey test. wileyonlinelibrary.com/journal/jib Copyright 2016 The J. Inst. Brew. 2016; 122: 612 623

Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates concentration of this compound varies with the type of yeast species/strain involved in the fermentation process. Concentrations of acetaldehyde in the tested plum distillates (Table 3), regardless of the microorganisms used for fermentation and temperature of process, were all similar and ranged between 53.36 ± 6.25 and 62.98 ± 7.32 mg/l alcohol 40% v/v, (p > 0.05). The results of our study are in agreement with the findings of Amerine and Ough (45), who reported that fermentation temperature did not influence the total acetaldehyde content. As regards the lack of significant differences in the acetaldehyde content irrespective of microorganisms used to start the plum mash fermentation (p=0.280), literature data (46) suggest that the yeast response to acetaldehyde employs the same mechanisms that participate in the response to other forms of stress. Thus, acetaldehyde exchange between strains could inhibit the growth of some yeast strains while encouraging the growth of others. This phenomenon could be particularly connected to the colonization of complex fermentation media by S. cerevisiae after the elimination of non-saccharomyces yeasts. During spontaneous fermentations, a succession of different indigenous S. cerevisiae yeasts could be observed throughout the stationary phase of fermentation (47,48). When fermentation is inoculated with pure Saccharomyces starter cultures, the persistence of several indigenous S. cerevisiae strains during fermentation may also be observed (47). There is great variation in metabolic capability among isolates of naturally occurring S. cerevisiae, which includes significant heterogeneity among strains in the production of ethanol, and other products of metabolism (49). In addition to acetaldehyde, the presence of propionaldehyde, isovaleraldehyde, hexanal, furfural and benzaldehyde was determined in the studied plum distillates. The concentrations of isovaleraldehyde were higher (p < 0.05) in the distillates obtained after fermentation at the temperature of 30 C rather than 18 C, whereas formation of hexanal showed the opposite tendency for synthesis (p < 0.05;,Table 3). Also the type of microorganisms used for the plum mash fermentations affected the concentration of isovaleraldehyde and hexanal in the obtained distillates (p < 0.001), and furthermore a significant temperature microflora interaction was recorded for synthesis of these compounds (p < 0.001 and p=0.014, respectively). Furfural generated by the dehydration of pentoses is typical for fruit distillates (40). Its concentrations in tested plum distillates were between 18.52 ± 1.76 and 26.93 ± 5.91 mg/l alcohol 40% v/v and there was no significant effect (p > 0.05) of processing temperature on the furfural content observed. However, significant differences in the concentration of this compound were detected in the distillates obtained after fermentation with the use of various microorganisms (p =0.002). Different concentrations of furfural in spirit beverages can result from different conditions of distillation and equipment used in this process (50). Distillation in a column apparatus results in lower concentrations of furfural in spirits than processes conducted in a co-current distillation apparatus equipped with a helmet. Therefore cognacs derived by the latter method contain more furfural, above 30 mg/l alcohol 100% v/v (51). The flavour of stone fruit spirits is affected by the aroma compound benzaldehyde, which originates from the enzymatic degradation of amygdalin present in the stones of the fruits, passing into the mash during fermentation and later into the distillate (52). Higher concentrations of benzaldehyde (p < 0.05) were found in the plum distillates derived from mashes fermented at a lower temperature (18 C). An increase in the fermentation temperature to 30 C resulted in a decrease in the concentration of benzaldehyde in the obtained distillates (Table 3). The probable cause of this can be the accelerated conversion of this aldehyde to benzyl alcohol and/or 1-phenyl-1,2-propanediol (PAC-diol), catalysed by elevated temperature (53). Significantly different concentrations of this compound by microflora interaction (p < 0.001) were also detected. When comparing the concentration of benzaldehyde in the distillates obtained from plum mashes fermented at 18 C, a 2-fold higher level of this compound was shown in the sample fermented spontaneously with the raisin microflora compared with the others (Table 3). According to the literature data (54), K. apiculata yeast (19 strains assayed) exhibit a high ability for the biosynthesis of benzaldehyde. The increased level of benzaldehyde in the distillate originating from spontaneous fermentations with microflora of raisins can be explained by the presence of the K. apiculata yeast on their surface (Fig. 2). During alcoholic fermentation, many esters can be formed in the reaction between alcohols and acetyl-coa catalysed, among others, by acetyltransferases. The predominant ester synthesized by yeast is ethyl acetate formed from ethanol and acetyl-coa (31). In our studies its concentrations were higher in trials obtained after fermentation at 18 C than at 30 C (p < 0.05; Table 3). Especially rich in ethyl acetate and isoamyl acetate was the distillate derived after fermentation with the indigenous microflora of plums and raisins, at 18 C, which confirm the results of study of Romano et al. (55) that esters, especially ethyl acetate, are mainly produced by Kloeckera yeast. Moreover the concentrations of these esters as well as hexyl acetate, ethyl butanoate and methyl benzoate were higher in the distillates obtained from plum mashes fermented at 18 C than at 30 C (p < 0.05), while the concentrations of ethyl benzoate, as well as ethyl hexanoate and ethyl octanoate, were not strictly correlated with temperature of fermentation (p > 0.05). Taking into account the type of microorganisms used for fermentation, there were statistically significant differences in the concentration of all of detected esters between samples fermented spontaneously and with the participation of the S. bayanus wine yeast (p < 0.05). This indicates that there was no microbial contamination present during spontaneous fermentation, which resulted in the growth of S. cerevisiae yeast and the proper duration of the alcoholic fermentation (5). It is interesting that distillates obtained from plum mashes fermented by only native microbiota of plums (spontaneous fermentation) were characterized by a similar content of esters such as ethyl acetate, isoamyl acetate, ethyl hexanoate and ethyl octanoate to that indicated in the trials obtained after fermentation with wine yeast S. bayanus (p > 0.05; Table 3). This might mean a dominance of Saccharomyces sp. strains during the turbulent stage of the plum mash fermentation, both with epiphytic microbiota of fruits and inoculated with wine yeast (S. bayanus). In a study carried out by Satora and Tuszyński (29), the highest concentration of total esters was determined in plum brandies obtained after spontaneous fermentation (2470 mg/l alcohol 100% v/v) and by K. apiculata yeast (1829 mg/l alcohol 100% v/v). The amount of total esters determined in these spirits exceeded over three times the amount detected in plum brandies produced using distillery and wine strains of S. cerevisiae. Very intensive formation of esters is a characteristic of Kloeckera/Hanseniaspora genus representatives. The dominance of non-saccharomyces yeast during the turbulent stage 617 J. Inst. Brew. 2016; 122: 612 623 Copyright 2016 The wileyonlinelibrary.com/journal/jib

618 of spontaneous fermentation of plum musts undoubtedly contributed to the high level of esters in the obtained plum brandies (5). From a quantitative viewpoint, the most important volatile compounds in fruit spirits are methanol and higher alcohols also known as fusel alcohols (56 59). Methanol is a natural component of plants and fruits and is liberated from pectic substances by enzymatic degradation under the influence of a specific pectolytic enzyme, pectin methylesterase, particularly during ripening and fermentation processes. While methanol does not directly affect the flavour of the distillate, it is subjected to restrictive controls owing to its high toxicity (60,61). According to EU Regulation (EC) no. 110/2008 (11) for plum brandies, the concentration of methanol should not exceed 12 g/l alcohol 100% v/v (i.e. 4.8 g/l alcohol 40% v/v). Unfortunately, all tested plum distillates were characterized by higher concentrations of this compound (Table 3), which indicates the necessity of conducting corrective distillation to reduce the content of methanol. It is commonly known that methanol forms azeotropes and also transfers to the main fraction as well as to its tails (3). In contrast to methanol, higher alcohols have a significant effect on both the sensory characteristics and quality of fruit (plum) distillates (2,29). 3-Methyl-1-butanol (isoamyl alcohol), 2-methyl- 1-butanol (active amyl alcohol), 2-methyl-1-propanol (isobutanol) and n-propanol (1-propanol) are the principal constituents of higher alcohols (62,63). All of the obtained plum distillates were rich in higher alcohols irrespective of fermentation variant. Among the identified and determined higher alcohols, the highest concentrations were observed in the case of 1-propanol and then 3-methyl-1-butanol as well as 2-methyl-1-propanol. The effect of fermentation temperature on the higher alcohol biosyntheses (with the exception of benzyl alcohol) was shown (p < 0.05). The higher concentrations of 1-propanol and 1-butanol were observed in the distillates originating from the plum mashes fermented spontaneously at 30 C (with and without raisin additions) than at 18 C (p < 0.05), whereas the 2-methyl-1-propanol content was higher in all distillates obtained after fermentation at 30 C (p < 0.001). On the other hand, concerning the effect of the microflora used for fermentation, the spontaneous fermentation of the plum mashes (S-18 and S-30) resulted in higher amounts of 2-methyl-1-butanol and 3-methyl-1-butanol in the distillates in comparison with the samples in which the fermentation was carried out spontaneously with addition of raisins or with S. bayanus yeast (Table 3; p < 0.05). The distillates obtained in our previous studies concerning the application of intermediate products of plum processing for alcoholic fermentation (3) contained higher concentrations of 3-methyl-1-butanol and lower of 1-propanol and 2-methyl-1- propanol, compared with the results in this study. The reasons for the differences in the amounts of higher alcohols can result from differences in the quantitative composition of amino acids present in plums and plum processing intermediates, content of sugar in the fermentation worts and the fermentation method (using single S. cerevisiae culture, epiphytic microflora of plums and of raisins). Satora and Tuszyński (29) found that spirits obtained after spontaneous fermentation were characterized by less than half the concentration of fusel alcohols compared with samples obtained after fermentation by distillery as well as wine strains of S. cerevisiae. It is known from literature data (64,65) that K. apiculata yeast which are dominant on the fresh var. Węgierka Zwykła plum fruits, as well as during the K. Pielech-Przybylska et al. initial and final stages of the plums spontaneous fermentation, are distinguished by relatively poor synthesis of fusel alcohols. Simultaneously, indigenous strains of S. cerevisiae also produced significantly lower amounts of higher alcohols than commercial cultures (29). In the tested plum distillates the relatively small amounts of 1-butanol, benzyl alcohol and 1-hexanol were marked (Table 3). Benzaldehyde and benzyl alcohol are formed during hydrolysis of amygdalin in stones and are present in fruit spirits in much higher concentrations if the mash is fermented with the stones (66). In turn, hexanol is formed in plant tissues by the activity of the enzyme alcohol oxidoreductase on hexanal (67) and is considered an important contributor to the aroma of fresh plums (1). Also 1-hexanol could be produced by a non-saccharomyces yeast (68) and in concentrations above 100 mg/l 100% v/v (i.e. 40 mg/l alcohol 40% v/v) usually negatively affects the flavour and aroma of brandies (29). Plum distillates derived in our experiments contained this compound in concentrations ranging from 19.85 ± 0.26 to 28.69 ± 0.35 mg/l alcohol 40% v/v. Its higher amounts were determined in the distillates originating from plum pulp fermented at 18 C rather than at 30 C (p < 0.001), especially in trials fermented with wine yeast S. bayanus (Table 3). The amyl alcohols/1-propanol ratio may be used as an index to distinguish spontaneously fermented samples from those produced by monoculture (69). Its value is <1 or close to 1 for the former and it is >1 for the later. Whisky quality is evaluated by, among others, methods calculating the amyl alcohols isobutanol and isobutanol 1-propanol ratios, which should be >1 (70). All plum distillates obtained during our experiments were characterized by a > 1 values of amyl alcohols 2- methyl-1-propanol. As regards the amyl alcohols 1-propanol ratio, its values were <1 for almost all distillates, while one derived after spontaneous fermentation at 18 C showed this ratio value on a level > 1 (Table 4). The relatively high concentrations of 1-propanol had a significant influence on the calculated indices and as a result the 2-methyl-1-propanol 1-propanol ratio in all tested spirits was <1 and comparable to the one in slivovitz tested by Satora and Tuszyński (2). Degradation of the commonly occurring glycosides from Prunus sp., prunasin and amygdalin, leads to the liberation of benzaldehyde and hydrocyanic acid (HCN). The maximum level of HCN in spirit beverages is limited. Regulation on the definition, description, presentation, labelling and the protection of geographical indications of spirit drinks (11), stipulates that the maximum HCN content in stone fruit spirits and stone fruit marc spirits shall amount to 7 g/hl of 100% v/v alcohol (70 mg/l). The tested plum distillates contained small amounts of free HCN, ranging from 0.20 ± 0.05 to 0.50 ± 0.08 mg/l alcohol 40% v/v (Table 3). It was found that the plum spirits obtained after fermentation at a temperature of 18 C contained lower concentrations of this compound than distillates manufactured from mashes fermented at 30 C (p < 0.001). In addition, an effect of the microorganisms used for fermentation on the concentration of hydrocyanic acid was observed (p < 0.001). The most rapid fermentation occurred in plum mashes inoculated with S. bayanus yeast and higher ethanol concentration in the initial phase of process than in mashes fermented with epiphytic microbiota is a probable cause of inhibition of enzymes catalysing the hydrolysis of cyanogenic glycoside activity, which results in lower concentrations of hydrocyanic acid in the obtained distillates. wileyonlinelibrary.com/journal/jib Copyright 2016 The J. Inst. Brew. 2016; 122: 612 623

Influence of yeast on the yield of fermentation and volatile profile of Węgierka Zwykła plum distillates Table 4. Selected indices for evaluation of the tested distillates. Ratio Batch p-value S-18 S-30 R-18 R-30 Sb-18 Sb-30 T M T M Amyl alcohols/ 2-methyl-1-propanol 3.65 b ± 0.55 1.43 a ± 0.11 1.65 a ± 0.11 1.54 a ± 0.25 1.48 a ± 0.19 1.31 a ±0.10 <0.001 <0.001 <0.001 Amyl alcohols/ n-propanol 1.17 c ± 0.10 0.86 b ± 0.20 0.61 ab ± 0.04 0.55 a ± 0.08 0.55 a ± 0,05 0.55 a ± 0.04 0.024 <0.001 0.049 2-methyl-1-propanol/ n-propanol 0.32 a ± 0.05 0.60 b ± 0.09 0.37 a ± 0.04 0.36 a ± 0.07 0.37 a ± 0.06 0.48 ab ± 0.06 0.001 0.067 0.007 For designation of the batches, see Table 2. Means in a row with a different superscript letters are significantly different (p < 0.05) as analysed by two-way ANOVA and the Tukey test. Aromatic profile and sensory analysis of plum distillates The OAV index permits the evaluation of contribution degree of each compound to the final aroma of product. In this sense, only compounds with OAV >1 are considered to contribute to the characteristic aroma in the food and beverages (71 73).Asshown in Table 5, among volatile compounds identified and quantitatively determined in the studied plum distillates, only 11 compounds were detected at levels above their odour thresholds (OAVs >1). On the other hand, the studies of Rocha et al. (74) showed that the substances whose OAVs are >0.2 can also affect the flavour of a product. Thus, five more volatile compounds (acetaldehyde, hexyl acetate, ethyl butanoate, ethyl octanoate and 2-methyl-1-propanol) were also considered as contributors to the aroma of these distillates. The highest OAVs in all tested distillates were shown by isovaleraldehyde (50.80 ± 2.50 to 71.60 ± 3.30) and ethyl acetate (12.95 ± 0.95 to 51.49 ± 2.35) (Table 5). Levels of these compounds in the studied spirits ranged from ca. 50 to ca. 72 times higher in the case of isovaleraldehyde and from ca. 14 to ca. 51 times higher in the case of ethyl acetate than their odour thresholds. Another important aroma compound in tested distillates was benzaldehyde, with OAVs >1 and with the highest obtained value in the distillate from plum pulp fermented spontaneously with the addition of raisins, at 18 C compared with other batches fermented at 18 C (p < 0.05). Moreover, hexanal can be considered as a distinctive compound among aldehydes with OAVs close to 1 in the distillates derived after fermentation at 30 C and >1 in samples obtained after fermentation at 18 C (p < 0.001). Benzaldehyde, as well as the C6 compounds (among others hexanal) and related esters, are considered as important contributors to the aroma of fresh plums (75,76). Hexanal is the compound associated with the smell of foliage and grass and has been described as having a plum-like aroma so it might add a pleasant flavour to plum purée (75,77) despite the fact that Pino and Quijano (1) have reported that OAV for hexanal in fresh plums was <1. The significant aroma compounds in alcoholic beverages are acetate esters of higher alcohols (among others ethyl acetate, isoamyl acetate, isobutyl acetate) and ethyl esters of fatty acids (among others ethyl butanoate, ethyl lactate, ethyl decanoate, ethyl octanoate and ethyl hexanoate). These compounds may contribute a pleasant fruity fragrance to the general aroma of fruit distillates (6,78). In the studied plum distillates, apart from ethyl acetate, another odour active ester is also isoamyl acetate, the OAVs of which ranged from ca. 1 to ca. 5 units, and higher values were observed in spirits samples obtained after fermentation at 18 C rather than at 30 C (p < 0.001). A typical description of isoamyl acetate is sweet, fruity and banana-like at levels >1 mg/l flavour threshold (63). Moreover ethyl benzoate and ethyl hexanoate also showed OAVs >1 in the studied plum spirits. The study of Pino and Quijano (1) showed that esters were the dominant volatile compounds in plums and the highest OAVs were calculated for ethyl 2-methylbutanoate (1837) and hexyl acetate (1702). Ethyl 2-methylbutanoate has a green-fruity odour, reminiscent of the peels of unripe plums, while hexyl acetate has a sweet-fruity berry and pear-like odour. Relatively high OAVs were also calculated by the authors for ethyl butanoate (847), and ethyl hexanoate (323). The OAVs of ethyl hexanoate in plum fruit were higher than in the plum distillates obtained in our studies (from 1.42 to 1.91). In turn, the OAV of ethyl octanoate in fresh plums was <1 (1), which is in agreement with the data concerning plum distillates tested by us. Almost all of the higher alcohols identified and determined in plum spirits (except 1-butanol and benzyl alcohol) were characterized with OAVs >1 or 0.2, which suggests their role in the creation of the aroma. 3-Methyl-1-butanol, followed by 2-methyl-1-butanol, showed the highest OAVs compared with the rest of the higher alcohols. Also 1-propanol was characterized by with OAVs >1, while 2-methyl-1-propanol was marked by OAVs between 0.41 and 0.93. Higher alcohols are reported to contribute more to the intensity of the odour of the wine than to its quality (40), while alcohols with six carbon-atoms constitute a defect by depending on their concentration (79). In this sense, only 1-hexanol exceeded its odour threshold value (OAVs ranged from 1.32 to 1.91; Table 5). Cacho et al. (80), who tested Peruvian piscos (beverages produced in Peru by distilling wine made from several varieties of grapes), also reported aroma values >1 for higher alcohols such as 2-methyl-1- butanol and 3-methyl-1-butanol. The authors suggest that these compounds should be considered as having a possible effect on the overall aroma of pisco. The results of sensory evaluation of the tested plum distillates are presented in Table 6. Their total sensory quality varied between 13.0 ± 0.5 and 15.2 ± 0.5 points (on the 20-point scale). According to the results of the performed sensory ranking, the best rated distillate was the one obtained after the spontaneous fermentation with indigenous microflora of plums and raisins, at 18 C (p < 0.05). It was characterized by a pleasant soft plum aroma (odour) and a well-harmonized taste, characteristic of slivovitz. It can be assumed that a significant impact on the quality of this 619 J. Inst. Brew. 2016; 122: 612 623 Copyright 2016 The wileyonlinelibrary.com/journal/jib

K. Pielech-Przybylska et al. 620 Table 5. Odour activity values of aroma compounds in the tested plum distillates Chemical compound Odour threshold (mg/l alcohol 40% v/v) Odour activity value p-value S-18 S-30 R-18 R-30 Sb-18 Sb-30 T M T M Aldehydes Acetaldehyde 72.0 ± 5.0 0.80 ab ±0.05 0.83 ab ±0.05 0.85 a ±0.05 0.70 b ±0.05 0.87 a ±0.06 0.85 ab ± 0.05 0.043 0.057 0.024 Propionaldehyde 10.0 ± 2.0 0.05 c ± 0.01 0.09 a ± 0.01 0.08 ab ±0.01 0.06 BC ± 0.01 0.08 ab ± 0.01 0.10 a ± 0.01 0.015 0.006 <0.001 Isovaleraldehyde 0.1 ± 0.02 50.80 a ± 2.5 57.50 ab ±3.5 61.30 b ±2.35 66.40 BC ±5.25 51.40 a ± 3.25 71.60 c ±3.3 <0.001 0.001 0.005 Hexanal 4.0 ± 0.5 1.10 BC ± 0.05 0.78 a ±0.06 1.25 b ± 0.15 0.88 ac ±0.06 1.29 b ± 0.15 0.81 a ±0.07 <0.001 0.099 0.393 Furfural 215.0 ± 15.0 0.12 ab ±0.01 0.12 ab ±0.01 0.09 a ±0.01 0.09 a ±0.01 0.13 b ± 0.02 0.09 a ± 0.01 0.040 0.004 0.022 Benzaldehyde 0.8 ± 0.2 2.33 b ± 0.10 1.11 a ± 0.15 3.97 c ±0.45 1.23 a ±0.12 2.28 b ± 0.35 1.11 a ±0.08 <0.001 <0.001 <0.001 Esters Ethyl acetate 34.0 ± 5.0 30.36 b ±2.25 13.87 a ± 1.23 51.49 c ±2.35 17.23 a ±1.2 30.29 b ±2.25 12.95 a ±0.95 <0.001 <0.001 <0.001 Isoamyl acetate 1.0 ± 0.2 3.78 b ± 0.45 1.68 a ± 0.15 5.32 c ±0.45 1.34 a ±0.05 3.37 b ± 0.35 1.16 a ±0.03 <0.001 <0.001 <0.001 Hexyl acetate 2.0 ± 0.5 0.19 b ±0.05 0.07 ac ±0.01 0.19 b ±0.01 0.05 a ±0.01 0.13 BC ± 0.01 0.05 a ±0.01 <0.001 0.017 0.121 Ethyl butanoate 2.0 ± 0.5 0.18 a ± 0.03 0.17 a ±0.01 0.19 a ±0.01 0.17 a ±0.01 0.19 a ± 0.01 0.19 a ± 0.01 0.190 0.262 0.543 Methyl benzoate 7.0 ± 1.0 0.03 b ± 0.003 0.02 ab ± 0.005 0.01 a ± 0.005 0.02 ab ± 0.005 0.02 ab ± 0.002 0.02 ab ± 0.002 1 0.003 0.003 Ethyl benzoate 3.0 ± 0.5 2.03 b ±0.24 1.62 ab ±0.12 1.19 a ±0.10 1.31 a ±0.12 1.43 a ±0.18 1.57 ab ± 0.25 0.571 <0.001 0.032 Ethyl hexanoate 1.0 ± 0.2 1.88 ab ±0.20 1.89 ab ± 0.15 1.48 ab ±0.10 1.42 a ±0.12 1.91 b ±0.20 1.60 ab ± 0.25 0.178 0.003 0.297 Ethyl octanoate 10.0 ± 1.5 0.67 c ± 0.05 0.65 BC ±0.05 0.47 a ±0.03 0.45 a ± 0.05 0.54 ab ± 0.06 0.57 abc ± 0.03 0.881 <0.001 0.574 Alcohols Methanol 49500 ± 1500 0.12 a ± 0.02 0.13 a ±0.03 0.11 a ±0.02 0.12 a ±0.02 0.13 a ± 0.05 0.14 a ± 0.02 0.477 0.507 1 1-Propanol 400 ± 50 1.90 a ± 0.30 2.33 a ±0.25 2.21 a ±0.52 2.65 a ±0.35 2.46 a ± 0.25 2.50 a ± 0.32 0.086 0.180 0.535 2-Methyl-1-propanol 600 ± 50 0.41 c ± 0.02 0.93 d ± 0.05 0.55 ac ± 0.04 0.64 ab ± 0.05 0.61 ab ± 0.05 0.76 b ±0.10 <0.001 0.039 <0.001 1-Butanol 300 ± 30 0.01 a ± 0.002 0.02 a ± 0.005 0.02 a ± 0.005 0.02 a ± 0.002 0.02 a ± 0.002 0.02 a ± 0.005 0.088 0.066 0.066 2-Methyl-1-butanol 40 ± 5 5.15 ab ± 0.55 5.54 b ±0.48 3.97 a ±0.45 4.08 a ±0.39 4.02 a ±0.45 4.45 ab ± 0.55 0.197 <0.001 0.824 3-Methyl-1-butanol 40 ± 5 17.13 c ±0.55 14.45 b ±0.35 9.54 a ±0.45 10.48 a ±0.62 9.53 a ± 0.35 9.39 a ± 0.42 0.015 <0.001 <0.001 1-Hexanol 15 ± 3 1.91 a ±0.05 1.74 ab ± 0.06 1.67 ab ±0.15 1.46 BC ±0.12 1.82 a ±0.15 1.32 c ±0.15 <0.001 0.004 0.071 Benzyl alcohol 181 ± 10 0.03 a ± 0.01 0.04 a ±0.1 0.07 b ±0.01 0.07 b ± 0.01 0.05 ab ±0.01 0.05 ab ± 0.01 0.493 <0.001 0.619 Designation of the batches: see Table 2. Means in a row with a different superscript letters are significantly different (p < 0.05) as analyzed by two-way ANOVA and the Tukey test. *T Temperature effect; M Microflora effect; T M = Temperature Microflora interaction effect (two-way ANOVA; p < 0.05) wileyonlinelibrary.com/journal/jib Copyright 2016 The J. Inst. Brew. 2016; 122: 612 623