International Journal of Food Microbiology

International Journal of Food Microbiology 166 (13) 323 3 Contents lists available at ScienceDirect International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro Yeast community associated with the solid state fermentation of traditional Chinese Maotai-flavor liquor Qun Wu, Liangqiang Chen, Yan Xu State Key Laboratory of Food Science and Technology, Key Laboratory of Industrial Biotechnology of the Ministry of Education, School of Biotechnology, Jiangnan University, Wuxi, Jiangsu 214122, China article info abstract Article history: Received 28 March 13 Received in revised form 19 June 13 Accepted 7 July 13 Available online 12 July 13 Keywords: Chinese liquor Diversity Solid-state fermentation Yeast community Yeasts are the most important group of microorganisms contributing to liquor quality in the solid-state fermentation process of Chinese Maotai-flavor liquor. There occurred a complex yeast community structure during this process, including stages of Daqu (the starter) making, stacking fermentation on the ground and liquor fermentation in the pits. In the Daqu making stage, few yeast strains accumulated. However, the stacking fermentation stage accumulated nine yeast species with different physio-biochemical characteristics. But only four species kept dominant until liquor fermentation, which were Zygosaccharomyces bailii, Saccharomyces cerevisiae, Pichia membranifaciens, and Schizosaccharomyces pombe, implying their important functions in liquor making. The four species tended to inhabit in different locations of the stack and pits during stacking and liquor fermentation, due to the condition heterogeneity of the solid-state fermentation, including the different fermentation temperature profiles and oxygen density in different locations. Moreover, yeast population was much larger in the upper layer than that in the middle and bottom layers in liquor fermentation, which was in accordance with the profile of reducing sugar consumption and ethanol production. This was a systematical investigation of yeast community structure dynamics in the Maotai-flavor liquor fermentation process. It would be of help to understand the fermentative mechanism in solid-state fermentation for Maotai-flavor liquor. 13 Elsevier B.V. All rights reserved. 1. Introduction Maotai-flavor liquor is a symbolic drink in China just as whisky in Scotland, and brandy in France (Fan et al., 11; Xu and Ji, 12). It is famous for its soy sauce-like and roasted aroma style. This is due to a unique and complicated spontaneous fermentation process, which includes Daqu (the starter) making, stacking fermentation and liquor fermentation stages (Fig. 1). In the Daqu making stage, the temperature of Daqu increases as the fermentation processes, and the maximum temperature reaches about 6 C. In stacking fermentation, the mixtures of ground Daqu and steamed sorghum are piled up as a cone on the ground for stacking fermentation, and are terminated when the temperature on the top of the stack reaches about C. Then the fermented grain (Zaopei) is put into the underground cubic pits, and sealed for liquor fermentation. After days of fermentation, the ethanol content in Zaopei reaches about 4% 6% (w/w). Then Zaopei is distilled and liquor is collected, and the yield of liquor is about 13.4% (v/w) with the average alcohol content of 6% (v/v). It is a statically solid-state fermentation, and qualities of liquors distilled from Corresponding author at: State Key Laboratory of Food Science and Technology, Key Laboratory of Industrial Biotechnology of the Ministry of Education, School of Biotechnology, Jiangnan University, No. 18, Lihu Avenue, Wuxi, Jiangsu 214122, China. Tel./fax: +86 1 8864112. E-mail address: yxu@jiangnan.edu.cn (Y. Xu). different layers of the pit are different, which are called top (soy sauce flavor), middle (ethanol-sweet) and bottom (ester fragrant) liquors, separately. Such a solid state and spontaneous fermentation process accumulates a specific microorganism's community, including fungi, yeasts and bacteria. Previous studies have shown that the stacking and liquor fermentation processes are microbiologically and biochemically complex, but they have not been studied in depth (Wang et al., 11; Wu et al., 9, 12). Fungal species are significant in the stacking fermentation, where they produce the amylolytic enzymes to degrade the starch material in the raw material, sorghum, to fermentable sugars such as glucose and maltose (Wang et al., 8; Wu et al., 9). These then become substrates for alcoholic fermentation and flavor production by yeasts and bacteria. Lactic acid bacteria are the main producers of lactic acid and provide substrates for esterification of yeasts (Wang et al., 11; Wu et al., 9). Yeasts represent the most important group of microorganisms in the fermentation process, since they contribute significantly to the fermentation rate, product flavor and quality. Therefore, learning yeast community structures and dynamics during the fermentation process would shed light on the yeast function and the fermentative mechanism. So far, yeast community structures in many fermentation alcoholic beverages have been investigated. For instance, yeast community structure and dynamics in wine fermentation have been clearly described. Generally, Hanseniaspora (Kloeckera), Candida and Metschnikowia are 168-16/$ see front matter 13 Elsevier B.V. All rights reserved. http://dx.doi.org/1.116/j.ijfoodmicro.13.7.3

324 Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 Fig. 1. Flow sheet for making Maotai-flavor liquor. the main species initiating the wine fermentation, but they begin to decline and die off after mid-fermentation, and Saccharomyces cerevisiae becomes predominant and continues the fermentation until its completion (Fleet, 3). However, for Chinese Maotai-flavor liquor, the longperiod (2 7 days of stacking fermentation and days of liquor fermentation) and unique simultaneous saccharification and fermentation with the solid-state fermentation process give birth to a more complex yeast community structure in this liquor, which is also significant to be revealed. The yeast communities of Chinese Maotai-flavor liquor have been investigated since the 196's (Xu and Ji, 12; Wang et al., 11). Five yeast genera were revealed during the ripening of Daqu and these were Saccharomyces, Hansenula, Candida, Pichia, and Torulaspora (Wang et al., 8). Using a nested PCR-denaturing gradient gel electrophoresis technique, Liu et al. (12) reported contributions from Hanseniaspora uvarum and Candida allociferrii during the shaping, ripening and drying of Daqu. Nine yeast species were isolated from the stacking and liquor fermentation stages and identified by 26S rdna sequencing as S. cerevisiae, Zygosaccharomyces bailii, Pichia membranifaciens, Schizosaccharomyces pombe, Rhodotorula mucilaginosa, Kazachstania exigua, Debaryomyces hansenii, Issatchenkia orientalis (Pichia kudriavzevii) and Galactomyces geotrichum (Wu et al., 12). However, the yeast community structure and dynamics, as well as the relationship between yeast and fermentation parameters, have not been systematically investigated. To improve the quality of Maotai-flavor liquor and the efficiency of its production, a more detailed study of the association of yeasts with the production process is required. In this study we reported a detailed investigation of the yeast species associated with the Daqu, stacking and liquor stages of the fermentation process. 2. Materials and methods 2.1. Sampling Sampling was carried out in Guizhou Province, China. Within this factory, two separate liquor making processes were investigated with samples taken for each process from the Daqu, stacking fermentation and liquor fermentation stages. The results were the average of the data sets from the two separate fermentation processes. Samples of Daqu were collected from Daqu powders before they were used for stacking fermentation. As shown in Fig. 2, samples of Zaopei in stacking fermentation were collected from different positions of the stack, including its top (a), center (b), surface (c), and bottom (d). The pit was a cubic-shaped underground cellar (about 3.2 m 2.4 m 2.9 m). The samples were collected from three layers: the upper (e), the middle (f) and the bottom (g) layers. The yeast community structures in different points (e 1,e 2,e 3 ) of the same layer were nearly the same in preliminary tests, so we collected and mixed samples from the three different points in each layer before analysis. Sampling in pits was taken by a steel tube inserted into the Zaopei. The sealed end of the tube was kept open only for sampling in the sampling location, which would avoid cross-contamination among different layers. 2.2. Enumeration and isolation of yeasts Samples (1 g) were mixed with 9 ml sterile saline (8. g/l NaCl), soaked at 4 C for min. For the determination of colony-forming units (CFUs), 1 μl of each dilution was spread on WLN medium (Pallmann et al., 1) in triplicate. In order to inhibit bacterial growth, all media were supplemented with 1 mg/l chloramphenicol and penicillin. Cultures were incubated at C for days. Colonies of different types on WLN medium were counted separately according to the macroscopic properties (texture, surface, margin, elevation and color). All the samples were analyzed in triplicate. The representatives isolated from different samples were purified by repetitively streaking on YPD agar (1 g/l yeast extract, g/l peptone, g/l glucose and g/l agar). 2.3. DNA extraction and PCR amplification The genomic DNA of yeast was isolated according to the method reported (Makimura et al., 1994). The.8S ITS rdna region of yeast isolates was amplified by using the primers ITS1 ( -TCCGTAGGTGAA CCTGCGG-3 ) and ITS4 ( -TCCTCCGCTTATTGATATGC-3 ), and the D1/

Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 3 Fig. 2. Sampling locations of Zaopei in stacking fermentation and liquor fermentation in pits. A, in stacking fermentation, a, b, c, d represented the top, center, surface, and bottom positions of the stack. B, in liquor fermentation in pits. e, f, g represented the upper, middle and bottom layers of the pit. In each layer, samples of three different points (such as e 1,e 2,e 3 ) were collected and mixed before analysis. D2 domain of the 26S rdna gene was amplified by using the primer pair NL1 ( -GCATATCAATAAGCGGAGGAAAAG-3 ) and NL4 ( -GGTCCG TGTTTCAAGACGG-3 ). Each μl PCR reaction contained a final concentration of the following reagents: 1 ng of genomic DNA,. μm of each of the primers, μm dntps, 1 Taq reaction buffer, 2 mm MgCl 2, and 1 U Taq polymerase (Takara, Japan). PCR reactions were performed as previously reported (Baleiras Couto et al., ; Esteve-Zarzoso et al., 1999). 2.4. PCR RFLP analysis Up to twenty five colonies of each representative type classified by WLN medium were obtained for PCR RFLP analysis (if the isolate number of certain types was less than, we analyzed all the colonies of this group, but if the isolate number of certain types was more than, then we randomly chose colonies of this type for analysis.). The obtained.8s ITS rdna region of yeast isolate (1 μl) was digested with appropriate restriction enzymes, Hha I, Hae III, Hinf I (Takara, Japan), in accordance with the manufacturer's instructions. PCR products and restriction fragments were separated by gel electrophoresis in 14 g/l and g/l agarose gel, detected by ethidium bromide staining and photographed. Sizes of PCR and fragments were estimated by using standard molecular weight markers (1-bp ladder, Sangon Biotech, China; DS DNA marker, Dongshen Biotech, China) and Quantity One software (Bio-Rad Laboratories). For yeast species assignment, comparisons were conducted among restriction profiles of isolates, A C 1 1 4 ( o C) 1 2 3 4 6 7 1 2 3 4 6 7 ( o C) 8 7 6 4 3 2 1 1 2 3 4 6 7 1 2 3 4 6 7 B D 9 8 7 6 4 3 2 1 4 ( o C) ( o C) Fig. 3. Time profiles of yeast population and temperature of Zaopei during stacking fermentation. A, the top of the stack; B, the center of the stack; C, the surface of the stack; D, the bottom of the stack. Error bars represented the standard deviation of the triplicate experimental data of the same position from two representative fermentation processes.

326 Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 Table 1 Counts of different yeast species during stacking fermentation (1 CFU/g). Yeast species Top Center day 1 day 2 days 3 days 4 days days 6 days 7 days 1 day 2 days 3 days 4 days days 6 days 7 days S. cerevisiae a..77 1.4 2.19 1.86 1.34.11.28.34.7.29.6 Z. bailii.37 2.7 7.23 16.8 18.32 22.7 14.47.98.62.8 2.12 2.16 4.37 1.98 2.6 P. membranifaciens *..4.88.76.9.1.1.33.32. S. pombe 1.7.9.6..1.16.44.22.44 Total b.37 2.7 7.38 18.81 21.31 26.71 18.4 7.97.63.92 2.74 3.8.93 2.49 3.6 Yeast species Surface Bottom 1 day 2 days 3 days 4 days days 6 days 7 days 1 day 2 days 3 days 4 days days 6 days 7 days S. cerevisiae.27...83.4.14 2.13 1.7.89.82.71 Z. bailii 6.11.43.47.4 12.4 11.61 3.43 6.6 4.76 3.94 4.24 2.76 2.13 P. membranifaciens.44.86 1.9 1.19 1.82 9.97.12.31.38.32 S. pombe.84.49 Total b 6..44 17.13 22.32 26.99 23.91 22.41 3.9 6.1 7.27 6.1.13 3.8 2.84 a b Yeast species was undetectable. The other yeast species were sporadically found, and were not listed here. The total counts were obtained from the counts of all the yeast species. reference strains and other published profiles (Esteve-Zarzoso et al., 1999; Heras-Vazquez et al., 3; Nisiotou et al., 7). 2.. Sequencing the D1/D2 domains of the 26S rdna PCR products of the D1/D2 domain of the 26S rdna of 8 randomly selected isolates per distinct restriction pattern with PCR RFLP analysis were obtained (Prakitchaiwattana et al., 4). The sequences were compared with sequences available in GenBank database by using the basic local alignment search tool (BLAST) available in the 26S rdna region. 2.6. Analysis of chemical properties of Zaopei Zaopei temperature was measured with a sensing probe embedded in locations of a, b, c, d, e 1,f 1 and g 1 in stacking and liquor fermentations (Fig. 2A and B). Zaopei (1 g) was mixed with 9 ml distilled water, ultrasonically treated at C for min, then centrifuged at 4 C for min, the obtained supernatant was used to determine the content of reducing sugar, ethanol and acid. Reducing sugar was analyzed using the DNS method (Miller, 199). The ethanol content was determined by HPLC (Agilent) using a column Aminex HPX-87H (Bio-Rad). The column was eluted at 6 C with a degassed mobile phase containing 3mMH 2 SO 4 at a flow rate of.6 ml/min. All the compounds were determined with a RI detector (SFD). The identification and quantification of compounds were carried out by comparing retention time and concentration with standard solution. Acid was titrated by utilizing sodium hydroxide standard solution with phenothalin as an indicator, and the acidity was defined as the amount of consumed sodium hydroxide (mmol) per gram Zaopei. Starch was extracted from Zaopei by acid hydrolysis (% HCl, v/v) for min. After the ph value of the hydrolysate was adjust to 7. with % (w/v) NaOH, the total reducing sugar was determined. Starch content was estimated by calculating the difference between total reducing sugar and the original reducing sugar. All the samples were analyzed in triplicate. 3. Results and discussion 3.1. Yeast community structure and dynamic analyses in Daqu and stacking fermentation All the isolated yeast species were divided into 9 types by WLN medium and PCR RFLP analysis of the.8s ITS rdna region. They were assigned to the species S. cerevisiae, Z. bailii, P. membranifaciens, S. pombe, R. mucilaginosa, K. exigua, D. hansenii, P. kudriavzevii, and G. geotrichum based on the sequencing of D1/D2 domains of the 26S rdna genes (Wu et al., 12). With the unique colony properties (texture, surface, margin, elevation and color) of different yeast species on WLN medium, yeast community structures and dynamics were investigated in different positions (Fig. 2A and B) during the whole liquor making process. In mature Maotai-flavor Daqu, the population of yeasts was less than 1 CFU/g. Only Z. bailii and S. cerevisiae were isolated, whose populations were 87 and 13 CFU/g, respectively. The number of species in mature Daqu was less than that during the ripening period of Daqu (Wang et al., 8). This was because that most yeasts had died after the high temperature ripening process of Daqu. Fig. 3 presented the sequential development of total yeast population during stacking fermentation lasting for 7 days. It exhibited different population dynamics in different positions of the stack and reflected the impact of temperature on their development. The yeast counts all increased at the initial stage of fermentation. It reached the maximum level at the th day before decreasing at the top, center and surface of the fermentative stack, while it began to decrease after 3 day in the bottom positions. Table 1 showed the counts of different yeast species during stacking fermentation. Nine yeast species existed in this stage. However, except the four dominant species S. cerevisiae, Z. bailii, P. membranifaciens and S. pombe, the other five species were occasionally found in the stacking fermentation process. In addition, the community of the four dominant species differed in different positions. Z. bailii was the most predominant species in the stacking fermentation stage. It accounted for more than % of the total population in each studied position during this stage. For S. cerevisiae,its population came right after that of Z. bailii at the top and bottom positions of the fermentation stack. S. pombe was the species occurred at the end of fermentation, and was mainly inhabited at the top and center positions. At the later stage of stacking fermentation, P. membranifaciens was mainly found on the surface positions, whose population was nearly equivalent to Z. bailii at the end of fermentation. This implied the high oxygen requirement of P. membranifaciens. In the end of stacking fermentation, a large amount of white mycoderm would come into being on the surface of the stack, which was resulted from the growth of P. membranifaciens. The fermentation conditions differed from one position of stacking fermentation to another, including the differences in oxygen density and temperature. The heterogeneous conditions in different positions of the stacking led to the different physio-biochemical characteristics of yeast species. This presented the special function of stacking

Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 327 fermentation. As an indispensable stage of making Maotai-flavor liquor, it provided different yeasts for liquor fermentation in pits. 3.2. Yeast community structure and dynamic analyses in liquor fermentation in pits The yeast population during liquor fermentation in pits over days was shown in Fig. 4. The population of yeasts was much larger in the upper layer than in the middle and bottom layers. In the upper layer, the total population of yeast kept stable at the first 1 days, and then it quickly decreased. However, quite a few yeasts (about 7 1 4 CFU/g) remained in this layer in the end. In the middle and bottom layers, the yeast population decreased sharply on end till the th day, and then slowed its pace until the end of fermentation. There were only 19 and 1 CFU/g yeasts in the middle and bottom layers respectively in the end. The counts of different yeast species during liquor fermentation were shown in Table 2. Only S. cerevisiae, Z. bailii, P. membranifaciens and S. pombe, namely the dominant species in the stacking, survived at the stage of liquor fermentation. During this stage, Z. bailii was still the most predominant species in all layers, and it became the only surviving species in the middle and bottom layers after days. S. cerevisiae seemed to prefer to inhabit in the upper layer. It increased and was nearly equivalent to Z. bailii at the end of fermentation, while it decreased and was undetectable in the middle and bottom layers after days. P. membranifaciens also tended to inhabit in the upper layer, but it was less viable than S. cerevisiae. It decreased and was undetectable in the middle layer after days, and could not be detected in the bottom layer. S. pombe had almost the same population in the middle and bottom layers, which was slightly less than that in the upper layer. It decreased and could not be detected in all the three layers after days. Z. bailii was an important ethanol and higher alcohol producer in liquor making (Wu et al., 12). Although, it had long been defined to be notorious since it caused the spoilage of many fermented foods (Fleet, 3; Sousa et al., 1996), it might be a useful species in the Maotai-flavor liquor making process. S. cerevisiae could not only produce ethanol, but also produce many types of higher alcohols and esters (Wu et al., 12). S. pombe was found to be one of the largest contributors for volatiles, and was associated with many types of esters, alcohols and acids. However, P. membranifaciens produced minimal levels of volatiles (Wu et al., 12). The different metabolic characteristics of different yeast species and the heterogeneity of yeast community in different layers might be important reasons why liquors obtained from different layers in pits exhibited different flavors. 3.3. profile of Zaopei during stacking and liquor fermentation stages Zaopei temperature varied in different positions during the solid-state liquor making process. It was mainly associated with microorganism growth. Therefore, it has been accepted as an important indicative parameter for determining the development of liquor making (Xu and Ji, 12). profile of Zaopei during the Maotai-flavor liquor making process was investigated. As shown in Fig. 3, the temperature profile of different locations was different in stacking fermentation. in the top and center positions continuously increased during this stage, and the final temperature reached about C and 42 C respectively. in the surface and bottom positions decreased at first, and began to increase after 3 days. The final temperature in the surface and bottom was the same, about C. The heat accumulation related to yeast growth was the reason of the increase of Zaopei temperature. variation of Zaopei was nearly in consistency with yeast biomass profile at the top of the stack before days. However, as the temperature quickly increased to 4 C at days, the generated high temperature inhibited yeast growth, which resulted in yeast death in part after days, despite sufficient oxygen for cell growth at the top. On the contrary, the temperature condition would not inhibit yeast growth in the center and bottom of the stack, but the insufficient oxygen supply would be a main cause for lower biomass of yeast and the yeast death at the end of fermentation in the center and bottom positions. While in the surface of the stack, yeast grew quickly since neither oxygen supply nor temperature condition would inhibit yeast growth on the surface. Therefore, the heterogeneous conditions in different positions of the stack led to the accumulation of yeast species with different physio-biochemical characteristics, for instance, aerobic and facultative aerobic yeasts, thermophilic and non-thermophilic yeasts. This was also the superiority of solid-state stacking fermentation. As a particular stage for making Maotai-flavor liquor, it provided rich yeast community for liquor fermentation in pits. In addition, since the acidity was around..22 mmol NaOH/g, and only a trace amount of ethanol was produced in Zaopei during stacking fermentation, the ethanol and acid showed little effect on yeast growth in stacking fermentation. Therefore, temperature and oxygen might be the main parameters influencing yeast growth during stacking fermentation. In the liquor fermentation stage, the temperature profile in three layers was also different (Fig. 4). Although the yeast population was the highest in the upper layer, Zaopei temperature of this layer was the lowest among all the three, the inconsistency of yeast growth and Zaopei temperature might be attributed to the cooling of the environment, since the room temperature was only about 13 18 C. However, since Zaopei was sealed in pits for liquor fermentation, the poor heat dissipation led to the increase in temperature in the middle and bottom layers, which resulted from cell metabolism at the first 1 days. But temperatures in the two layers decreased gradually after then, and the bottom layer temperature decreased more quickly than that of the middle layer, due to the complete absence of oxygen in the bottom layer of the sealed underground pits. In Maotai-flavor liquor fermentation, the variation of temperature consequently influenced the yeast growth and metabolism, which finally influenced metabolic compositions and caused flavor differences of liquors from three layers. 3.4. Chemical properties of Zaopei in the liquor fermentation stage in pits Liquor fermentation in pits was a main fermentation stage in Maotai-flavor liquor making. It was characterized by the spontaneous fermentation process, which accumulated a complex microbial community, including bacterial, yeast and fungi (Wang et al., 8; Wu et al., 12). In this process, fungi mainly produced amylase and glucoamylase for starch saccharification (Wang et al., 8; Wu et al., 9), and yeasts and bacteria transformed the reducing sugar to ethanol and flavor compounds (Wang et al., 11; Wu et al., 12; Zhang et al., 13). Then, reducing sugar content represented the balance between the starch saccharification and sugar consumption. Therefore, we determined the related parameters in this process in the locations shown in Fig. 2B. At the end of fermentation, starch contents were lower in the upper layer than that in the middle and bottom layers (Fig. A). The reducing sugar contents were 8.4, 33. and.9 mg/g in the upper, middle and bottom layers, respectively (Fig. B), indicating that starch hydrolysis and sugar consumption in the upper layer were much quicker than those in the other two layers. Especially, in the upper layer, sugar consumption rate became quicker than starch saccharification rate. These indicated that microorganisms grew more vigorously in the upper layer than those in the other layers, which might be due to the limit of oxygen in the middle and bottom layers. Ethanol content is also an important parameter for liquor fermentation. As shown in Fig. C, ethanol content in the upper layer increased

328 Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 A B C 4 3 2 1. 4. 4. 3. 3..4.2.. 4. 4. 3. 3..4.2. 1 1 1 Fig. 4. Time profiles of yeast population and temperature of Zaopei during liquor fermentation in pits. A, in the upper layer; B, in the middle layer; C, in the bottom layer. Error bars represented the standard deviation of the triplicate experimental data of the same position from two representative fermentation processes. much faster than that in the middle and bottom layers of the pit. It reached the maximum value of 44. mg/g Zaopei at days in the upper layer, which was 2. and 2.4 times more than that in the middle and bottom layers, respectively. Since ethanol was mainly produced by yeast, the difference of ethanol indicated that yeast was more vigorous in the upper layer, which was in accordance with the yeast amounts in different layers. In addition, ethanol content all decreased in the later period of liquor fermentation, this was due to the transformation of ethanol to other flavor compounds, including esters, aldehydes and acids by microorganisms and enzymes, as well as its utilization as a carbon source by the yeasts. ( o C) ( o C) ( o C) Table 2 Counts of different yeast species during liquor fermentation in pits (1 4 CFU/g). Yeast species Upper layer Middle layer Bottom layer day days 1 days days days days days days 1 days days days days days days 1 days days days days days S. cerevisiae 3.78 4.79.6 4.7 4.17 3.64 3.9.44.27.22. a.41.7.7. Z. bailii.26 27.74.76 8.1.27 3.96 3.22 1.71 1.64 1.12.98.76.19 1.88.9.94.84.68.14 P. membranifaciens.24 2.8 7.92 1.73 1.68.29..2.2.1 S. pombe.8 2.62.1.76.4.3..2 Total 39.36 38. 39..7 11.12 7.89 6.86 2.21 1.96 1. 1.3.76.19 2.34 1.4 1.1.89.68.14 Yeast species was undetectable. a

Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 329 A Starch (mg/g) B Reducing sugar (mg/g) C Ethanol (mg/g) 29 28 27 26 2 2 1 1 4. Conclusion upper layer middle layer bottom layer upper layer middle layer bottom layer 1 1 upper layer middle layer bottom layer 1 Fig.. Chemical properties of Zaopei in the liquor fermentation process in pits. A, starch content; B, reducing sugar content; C, ethanol content. The data were obtained from two representative fermentation processes. Error bars represented the standard deviation of the triplicate experimental data of the same position from two representative fermentation processes. It is a unique yeast community structure in the Chinese Maotaiflavor liquor making process. A total of 9 yeast species were discovered in this process. There were only four dominant yeast species in both stacking and liquor fermentation, including S. cerevisiae, Z. bailii, P. membranifaciens and S. pombe. Their dominance, indicating that they might play important roles in liquor fermentation, should be further investigated. The yeast community in different fermentation stages played different roles for liquor making. Since the starter Daqu could not inoculate yeast in liquor making, the stacking fermentation could function as the yeast starter incubation stage in the liquor making process. It played an important role not only for multiplication of yeast cells, but also for regulation of yeast community structure. The conditional heterogeneity of the stack led to the accumulation of yeast species with different physio-biochemical characteristics. In the liquor fermentation stage, it also showed the structural diversity of yeasts in different layers. It resulted from the trace oxygen limit in the middle and bottom layers. The heterogeneity of yeast community in different layers might be an important reason of flavor difference of liquors from different layers of the pit. The present study seemed to take the lead in providing an integrated profile of the yeast community for making Maotai-flavor liquor. It showed that the spontaneous Maotai-flavor liquor fermentation was driven by a diverse and complex yeast community, which was responsible for the development of Maotai-flavor liquor. Therefore, further understanding was essential for the function of yeasts in liquor making, and the fermentative mechanism of Maotai-flavor liquor, which would also be useful for the improvement of liquor making technology and liquor quality. Acknowledgments This work was supported by the National High Technology Research and Development Program of China (12AA211, 13AA1218), the National Natural Science Foundation of China (3186), the Cooperation Project of Jiangsu Province Among Industries, Universities and Institutes (BY1116), and the Program of Introducing Talents of Discipline to Universities (111 Project) (111-2-6). References Baleiras Couto, M., Reizinho, R., Duarte, F.,. Partial 26 s rdna restriction analysis as a tool to characterise non-saccharomyces yeasts present during red wine fermentations. International Journal of Food Microbiology 12, 49 6. Esteve-Zarzoso, B., Belloch, C., Uruburu, F., Querol, A., 1999. Identification of yeasts by RFLP analysis of the.8 s rrna gene and the two ribosomal internal transcribed spacers. International Journal of Systematic and Evolutionary Microbiology 49, 329 337. Fan, W., Shen, H., Xu, Y., 11. Quantification of volatile compounds in Chinese soy sauce aroma type liquor by stir bar sorptive extraction and gas chromatography mass spectrometry. Journal of the Science of Food and Agriculture 91, 1187 1198. Fleet, G., 3. Yeast interactions and wine flavour. International Journal of Food Microbiology 86, 11 22. Heras-Vazquez, F., Mingorance-Cazorla, L., Clemente-Jimenez, J.M., Rodriguez-Vico, F., 3. Identification of yeast species from orange fruit and juice by RFLP and sequence analysis of the.8 s rrna gene and the two internal transcribed spacers. FEMS Yeast Research 3, 3 9. Liu, Xiu, Guo, K., Zhang, H., 12. Determination of microbial diversity in Daqu, a fermentation starter culture of Maotai liquor, using nested PCR-denaturing gradient gel electrophoresis. World Journal of Microbiology and Biotechnology 28, 237 2381. Makimura, K., Murayama, S., Yamaguchi, H., 1994. Detection of a wide range of medically important fungi by the polymerase chain reaction. Journal of Medical Microbiology, 8 364. Miller, G.L., 199. Use of dinitrosalicylic acid reagent for determination of reducing sugars. Analytical Chemistry 31, 426 428. Nisiotou, A., Spiropoulos, A., Nychas, G., 7. Yeast community structures and dynamics in healthy and botrytis-affected grape must fermentations. Applied and Environmental Microbiology 73, 67 6713. Pallmann, C.L., Brown, J.A., Olineka, T.L., Cocolin, L., Mills, D.A., Bisson, L.F., 1. Use of WL medium to profile native flora fermentations. American Journal of Enology and Viticulture 2, 198 3. Prakitchaiwattana, C.J., Fleet, G.H., Heard, G.M., 4. Application and evaluation of denaturing gradient gel electrophoresis to analyse the yeast ecology of wine grapes. FEMS Yeast Research 4, 86 877. Sousa, M., Miranda, L., Corte-Real, M., Leao, C., 1996. Transport of acetic acid in Zygosaccharomyces bailii: effects of ethanol and their implications on the resistance of the yeast to acidic environments. Applied and Environmental Microbiology 62, 32 37. Wang, C., Shi, D., Gong, G., 8. Microorganisms in Daqu: a starter culture of Chinese Maotai-flavor liquor. World Journal of Microbiology and Biotechnology 24, 2183 219. Wang,H.,Gao,Y.,Fan,Q.,Xu,Y.,11.Characterization and comparison of microbial community of different typical Chinese liquor Daqus by PCR DGGE. Letters in Applied Microbiology 3, 134 1.

3 Q. Wu et al. / International Journal of Food Microbiology 166 (13) 323 3 Wu, Z., Zhang, W., Zhang, Q., Hu, C., Wang, R., Liu, Z., 9. Developing new sacchariferous starters for liquor production based on functional strains isolated from the pits of several famous Luzhou-flavor liquor brewers. Journal of the Institute of Brewing 1, 111 1. Wu, Q., Xu, Y., Chen, L., 12. Diversity of yeast species during fermentative process contributing to Chinese Maotai-flavour liquor making. Letters in Applied Microbiology, 1 7. Xu, Y., Ji, K., 12. Moutai (Maotai): production and sensory properties. In: Piggott, J. (Ed.), Alcoholic Beverages: Sensory Evaluation and Consumer Research. Woodhead, Cambridge, pp. 3 3. Zhang, R., Wu, Q., Xu, Y., 13. Aroma characteristics of Moutai-flavourliquorproduced with Bacillus licheniformis by solid-state fermentation. Letters in Applied Microbiology 7, 11 18.