Yeast diversity in pit mud and related volatile compounds in fermented grains of chinese strong-flavour liquor

Abstract

Chinese strong-flavour liquor is produced via a traditional solid-state fermentation strategy facilitated by live microorganisms in pit mud-based cellars. For the present analysis, pit mud samples from different spatial locations within fermentation cellars were collected, and the yeast communities therein were assessed via culture-based and denaturing gradient gel electrophoresis (DGGE) approaches. These analyses revealed significant differences in the composition of yeast communities present in different layers of pit mud. In total, 29 different yeast species were detected, and principal component analyses revealed clear differences in microbial diversity in pit mud samples taken from different cellar locations. Culture-dependent strategies similarly detected 20 different yeast species in these samples. However, while Geotrichum silvicola, Torulaspora delbrueckii, Hanseniaspora uvarum, Saturnispora silvae, Issatchenkia orientalis, Candida mucifera, Kazachstania barnettii, Cyberlindnera jadinii, Hanseniaspora spp., Alternaria tenuissima, Cryptococcus laurentii, Metschnikowia spp., and Rhodotorula dairenensis were detected via a PCR-DGGE approach, they were not detectable in culture-dependent analyses. In contrast, culture-based approaches led to the identification of Schizosaccharomyces pombe and Debaryomyces hansenii in these pit mud samples, whereas they were not detected using DGGE fingerprints profiles. An additional HS-SPME-GC-MS-based analysis of the volatile compounds present in fermented grains samples led to the identification of 66 such compounds, with the highest levels of volatile acids, esters, and alcohols being detected in fermented grains from lower layer samples. A canonical correspondence analysis (CCA) suggested they were significant correlations between pit mud yeast communities and associated volatile compounds in fermented grains.

Introduction

Chinese liquor is a traditional fermented distilled spirit that is widely consumed in China and plays an important role in Chinese culture. Owing to its unique flavor profile, it is also becoming increasingly popular in other areas in East Asia. The flavor characteristics of different liquor preparations allow them to be classified into 12 different categories, including soy sauce flavour, strong flavour, light flavour, and miscellaneous flavour types (Xu et al. 2017). Of these, Chinese strong-flavour liquor is the most popular owing to its strong aroma and sweet flavour, accounting for roughly 70% of total liquor consumption in China (Pu and Yan 2022). Chinese strong-flavour liquor is produced via the distillation of a mixture of fermented grains including what, sorghum, and rice in a specialized fermentation pit (about 3.4 m long, 1.8 m wide, and 2.0 m deep) containing bacteria, archaea, and fungi. The walls and bottom of this fermentation pit are covered with pit mud, which is a type of fermented clay containing an array of anaerobic microbes. During the fermentation process, this pit mud supports the growth of microbes responsible for generating the volatile compounds that give Chinese strong-flavour liquor its unique taste (Tao et al. 2017). The composition of pit mud microbial communities thus determines the quality and flavour of the resultant liquor. Individual fermentation cellars are generally used for many years, and the fermented grains placed in the lower portion of the cellar can help to produce high-quality Chinese strong-flavour liquor. Prior studies have shown that microbial diversity is significantly increased in pit mud samples from the bottom of these fermentation cellars relative to samples from the upper wall pit layer (Ding et al. 2016). It is generally understood that the best Chinese strong-flavour liquor is also generated in the lower portion of the pit closer to the fermented grains, emphasizing the importance of the composition of pit mud along the lower walls and bottom of the cellar on Chinese strong-flavour liquor fermentation. Location-dependent effects on the production of Chinese strong-flavour liquor are thought to be attributable to the microbial domestication that occurs within a given fermentation pit during the process of recycling fermentation (Zhang et al. 2017), therefore, it is necessary to clarify the mechanisms underlying these effects and to investigate pit mud microbial composition.

Both culture-dependent and -independent strategies have previously been employed to study pit mud microbial communities. An early culture-based study identified Clostridium sp. W1 as the primary caproic acid-producing bacteria presented in Wuliangye liquor pit mud (Xue et al. 1988), while pit mud samples associated with the production of Luzhou Laojia liquor were dominated by by Hydrogenispora (57.2%), Sedimentibacter (5.4%), and Caproiciproducens (4.9%) (Qian et al. 2020). A range of bacteria, fungi, and archaea have been detected in pit mud samples (Xiao et al. 2023). In an effort to better understand time-dependent changes in these pit mud microbial communities, Tao et al. (2014) studied pit mud samples from pits that were 1, 10, 25, and 50 years old, revealing an upward trend in microbial diversity with pit age that plateaued after 25 years. Zhang et al. (2020) similarly conducted a multidimensional analysis of microbial communities in older and younger pit mud samples, and found that microbial diversity varied significantly as a function of vertical depth but not horizontal position within a given pit. Specifically, they found pit mud samples from the center of the pit were dominated by Lactobacillus species (12.80-42.72%), whereas those from the corner were dominated by caproiciproducens species (17.85-64.45%). These researchers ultimately determined that the factors most important for regulating pit mud microbial growth were pH, lactic acid, and soluble Ca²⁺ concentrations. Zhang et al. (2015) utilized culture-independent strategies including nested PCR-denaturing gradient gel electrophoresis (PCR-DGGE), phospholipid fatty acid (PLFA), phospholipid ether lipids (PLEL), and fluorescence in situ hybridization (FISH) analyses to characterize microbial communities in samples of artificial pit mud (APM) used to brew Chinese strong-flavour liquor. dominant bacteria in these samples included Clostridiales, Lactobacillales, Bacteroidales, and Rhizobiales species, while archaea present therein included Methanomicrobiales and Methanosarcinales species, and fungi included Saccharomycetales and Eurotiales species. They additionally determined that the pattern of APM piling influenced the consequent microbial community structure in a given sample. While many prior studies have explored bacterial community structures and functional properties in pit mud samples, there have been fewer analyses to date of pit mud yeast communities or the impact of cellar spatial locations on these community structures.

Yeast plays an important role in the preparation of Chinese liquor, controlling both the fermentation rate and the flavour profile of the resultant brew through the metabolic processing of different nutrients into volatile compounds (Wang et al. 2019). However, pit mud yeast diversity in the context of strong-flavour liquor production is poorly understood, as are the yeast-derived volatile compounds that ultimately contribute to liquor flavour.

In the present study, we employed a PCR-DGGE approach to study the structures of yeast communities in pit mud samples from different fermentation cellar depths. In addition, a head space-solid phase micro-extraction combined gas chromatography-mass spectrometry (HS-SPME-GC-MS) approach was additionally used to identify volatile compounds in liquor samples from upper, middle, and lower layers of fermented grains fermentation. Correlations between identified yeast communities and liquor flavour compounds were additionally assessed. Overall, the results of this study will offer new insights regarding the role of pit mud yeast communities in Chinese strong-flavour liquor production.

Materials and methods

Samples of pit mud and fermented grains

Pit mud samples were collected from a famous Chinese strong-flavour liquor distilleries located in Anhui provinces, China, and the pit ages was about 20 years. Samples were taken from the wall or bottom of the pits. The source, cellar age and sampling location of the pit muds are shown in Fig. S1. Each sample plot was divided into eight subplots (centre and edges) except bottom with nine subplots (side centre, side edges and bottom middle), and about 100 g of pit mud was collected from each subplot, then eight or nine subsamples were sufficiently mixed. The sampling depth of each subplot was about 5 cm.

Additionally, the fermented grains samples were taken respectively from the center of the top, middle and bottom layer of the fermentation pit filled with multiple-grains at the end of the fermentation. Finally, all samples were transferred to sterile polyethylene bags without air, sealed and stored at -20 °C until used.

Examination of yeast community

DNA extraction

Extraction total DNA from pit mud was performed by modified methods of Tan et al. (2020). Briefly, pit mud (5 g) was mixed with 15 mL CTAB solution and 100 µL protease K (10 mg/mL) and shaken horizontally at 225 rpm at 30 °C for 30 min. After the shaking, 1.5 mL 20% SDS was added and the mixture was incubated at 65 °C for 120 min, and then was inverted gently every 15 min. After centrifugation at 8,000 × g for 5 min at room temperature, the supernatant was mixed with an equal volume of chloroform/isoamyl/alcoholsolution (25: 24: 1). The mixture was centrifuged at 8,000 × g at room temperature for 5 min. Isopropanol (0.6-1.0× supernatant volume) and the mixture were incubated for 60 min at room temperature. Precipitates were collected by centrifugation at 20,000 x g for 20 min at room temperature, washed twice with 70% (v/v) ethanol and resuspended in sterile deionised water to a final volume of 200 µL. The DNA was purified using Universal UNlQ-10 Column DNA Purification Kit (Sangon, Shanghai, China) and quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Carlsbad, CA, USA).

PCR amplification

For yeast diversity analysis, the D1/D2 domain of the 26 S rRNA gene was amplified using universal primers NL1 (5′-GCGATATCAATAAGCGGAGGAAAAG-3′) and NL4 (5′-GGTCCGTGTTTCAAGACGG-3′) in the first round of the nested PCR approach according to Yan et al. (2019). Subsequently, this initial PCR product was diluted and used as a template for a nested PCR with primers NL1 containing a GC-clamp (5′-CGCCCGGGGCGCGCCCCGGGGCGGGGCGGGGGCGCGGGGGG-3′) at the 5′ end and LS-2 (5′-ATTCCCAAACAACTCGACTC-3′) (Nielsen et al. 2007). All reactions were carried out in a 50 µL volume containing 5 µL 10× PCR reaction buffer, 3.2 µL dNTP Mixture (2.5 mM), 0.4 µL ExTaq (5 U/µL), 50 ng DNA template, 1 µL of each primer (20 µM), and double deionized wate for adjustment of the volume to 50 µL. The first PCR amplification conditions was performed as follows: initial denaturation at 94 °C for 3 min, then 35 cycles of denaturation at 94 °C for 35 s, annealing at 50 °C for 35 s, extension at 72 °C for 1 min and 10 s; extension at 70 °C for 10 min. The second PCR amplification conditions was the same with the first PCR process except that the conditions of annealing at 60 − 55 °C for 35 s. The PCR products were then purified using a SanPrep Column PCR Product Purification Kit (Sangon, Shanghai, China). Before applied to DGGE analysis, all the PCR products were examined by electrophoresis on 1% agarose gels with ethidium bromide.

DGGE analysis

DGGE analysis of the PCR products was performed on a DCode Universal Mutation Detection System (Bio-Rad, Hercules, CA, USA). Polyacrylamide gels (7% w/v acrylamide–bisarylamide) were prepared with a Bio-Rad Gradient Delivery System (Model 475, Bio-Rad) using solutions containing 40% and 60% denaturant (100% denaturant corresponds to 7 M urea and 40% v/v formamide). Gels were run at 60 °C for 5 h at 150 V. The amplified fragments were visualized by AgNO₃ solution staining and UV transillumination (Yan et al. 2019). The yeast fingerprint on the DGGE gel was analyzed using the Quantity one software (Bio-Rad).

Excision of DGGE bands and sequencing

The predominant DGGE bands observed in the DGGE profiles were excised and eluted in ultrapure water at 4 °C overnight, and the eluted DNA was re-amplified using the second round primers mentioned in 2.2.2 without GC clamp. The PCR products were purified with a universal PCR purification kit (Tiangen, Beijing, China) Then the purified DNA was ligated into a pGEM-T easy vector and transformed into competent Escherichia coli DH5a cells according to the manufacturer’s instructions and the laboratory manual. Inserts from white colonies were amplified by adding whole cells directly to PCR reactions using the primer set M13F and M13 R (Sangon, Shanghai, China) as described by Liu et al. (2012). All positive colonies extracted from white colonies were sequenced by an automated DNA sequencer (Sangon, Shanghai, China). Subsequently, GenBank BLAST (http://www.ncbi.nlm.nih.gov/BLAST) was performed to identify the closest phylogenetic relatives of the partial rDNA sequences tested above.

Data analysis

The DGGE bands intensity and similarity matrix of DGGE profiles were calculated and exported out using Quantity one software (Bio-Rad). The community diversity indices including Shannon–Wiener index of general diversity (H), the Evenness (E), and the species richness (S) were calculated according to previous protocols (Yan et al. 2019). The dendrograms were calculated on the basis of Dice’s coefficient of similarity (weighted data), using the unweighted pair group method with arithmetic averages clustering algorithm (UPGMA).

Enumeration and isolation of yeasts

Yeasts were isolated and quantified using spread plates. Ten grams of pit mud sample were homogenized with 90mL sterile distilled water and the mixture was incubated at 25 °C for 30 min with shaking at 180 rpm. Diluted suspension (100 µL) was plated on YPD agar (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose and 20 g/L agar) supplemented with 100 µ g/mL ampicillin for yeasts. All assays were in triplicate. The yeasts were incubated at 30 °C for 2 days. Colonies were identified by their morphology and by performing PCR with primer pairs ITS1/ITS4 (ITS1: TCCGTAGGTGAACCTGCGG, ITS4: TCCTCCGCTTATTGATATGC) for yeast (Li et al. 2021). Sequence identity was analyzed with a GenBank search (http://www.ncbi.nlm.nih.gov/BLAST/).

HS-SPME-GC-MS analysis of fermented grains

The liquor samples, respectively collected from the distillation of the up, middle, and bottom layer of fermented grains, were detected via headspace solid-phase micro-extraction (HS-SPME) combined with gas chromatography mass spectrometry (GC-MS). HS-SPME was performed under previously reported conditions with slight modifications (Yan et al. 2019). A 5.0mL liquor sample diluted to 10% ethanol by volume, was transferred to a 20.0mL conical bottomed glass vial, then saturated with NaCl (1.5 g). After 100µL 2-octanol (70 mg/L, internal standard) solution was was injected into the the vial, the mixture were equilibrated by ultrasonic vibration in 50 °C constant temperature water bath for 10 min. After that, the extraction head was then inserted into each vial, and the sample was extracted at 60 °C for 30 min.

After HS-SPME, the extraction head was inserted into the injection port of the GC-MS system (Agilent 6890 GC and Agilent 5975 mass selective detector (MSD); Agilent, San Diego, USA) to separate and analyze the different compounds in the extracts. GC-MS was performed as previously reported with slight modifications (Yan et al. 2020). The samples were separated through a DB-Wax column (60 m length, 0.25 mm internal diameter, 0.25 μm film thickness) using helium as the carrier gas at a constant flow rate of 1 mL/min. The column temperature was programmed as follows: 40 °C for 2 min, increased by 5 °C/min to 80 °C for 2 min, and again increased by 8 °C/min to 230 °C for 7 min. High-purity nitrogen was applied as eluant gas to split sampling with a split ratio of 30: 1. The ionization energy was set equal to 70 eV, and the ion source and quadruple temperatures were set at 200 and 250 °C respectively. MS spectra were performed in scan mode (33–450 amu). Each sample was analyzed in triplicate.

Results

DGGE-based yeast community detection

To gain comprehensive insights regarding yeast spatial distributions, we next analyzed yeast community structures in pit mud samples from the upper, middle, lower, and bottom cellar layers via a PCR-DGGE approach which enabled us to calculate yeast diversity indices associated with these different spatial distributions (Table 1). We found that species richness was highest for samples from the upper pit mud layer, followed by that of samples from the bottom layer. Samples from the upper and bottom laters also exhibited higher levels of evenness relative to samples from the middle and lower levels (Table 1). Samples from the upper and bottom pit mud layers also had higher Shannon–Wiener index values than middle and lower layer samples, with samples from the Shannon-Wiener index value (3.03).

Table 1 Indices of yeast community in the samples collected from different spatial positions of cellar according to quantified bands from Fig. 1

Fig. 1

Denaturing gradient gel electrophoresis (DGGE) pattern of yeast 26 S rRNA in the pit mud samples collected from different spatial positions of cellar. Lanes U, M, D, and B represent samples collected from up wall layer of cellar, middle wall layer of cellar, down wall layer of cellar, and bottom layer of cellar, respectively. The bands indicated with numbers were excised and sequenced and the alignment results are listed in Table 2

Table 2 Identities of 26 S rRNA sequences of DGGE bands via BLAST

In total, 36 dominant bands were identified in DGGE profiles (labeled from 1 to 36 in Fig. 2). These bands were then sequenced and compared to the GenBank database (Table 2, supplementary materials 2). This revealed that the upper pit mud samples contained high levels of Saturnispora silvae (band 7), Geotrichum bryndzae (band 8), Pichia farinosa (bands 12), Candida intermedia (band 18), Pichia kudriavzevii (band 19), Kazachstania barnettii (band 24), Pichia guilliermondii (band 25), Hanseniaspora spp. (band 26), Candida humilis (band 27), Cyberlindnera jadinii (band 29), and Cryptococcus laurentii (band 31), whereas they were present at low levels or were absent in other layers. In the middle layer of pit mud, Hanseniaspora uvarum (band 6), Saccharomycopsis fibuligera (band 10), Candida tropicalis (band 28), Hanseniaspora vineae (band 30), and Rhodotorula dairenensis (band 34) were present at higher layers than in other samples with the exception of Pichia kudriavzevii (band 19). In lower layer samples, Wickerhamomyces anomalus (band 17), Pichia kudriavzevii (band 19), and Pichia kudriavzevii (band 20) were dominant, with Pichia kudriavzevii (band 19) being present at higher levels than in other samples. Pichia kudriavzevii (band 19) were also present at high levels in bottom layer pit mud samples. As Pichia kudriavzevii (band 19) was present in all samples other than the middle layer, suggesting they may be a key member of the yeast pit mud flora.

Fig. 2

Principal component analysis of yeast communities on three layers of pit mud samples. The first principal component (X axis) explains 45.9% of the total variance of the dataset, while the second principal component (Y axis) explains 30.2% of the total variance of the dataset. Yeasts are numbered as indicated on the DGGE fingerprint files shown in Fig. 2; Table 2; U, M, D, and B, represent up, middle, down, and bottom layer of pit mud, respectively

We next performed a PCA analysis of the data in Fig. 2, revealing clear microbial community-dependent discrimination between pit mud samples from different physical locations within the fermentation cellar (Fig. 1). Yeast composition profiles separated these pit mud samples into these three groups, each exhibiting unique microbial diversity.

In total, 20 yeast species were detected in pit mud via our culture-dependent approach (Table 3). However, certain species (Geotrichum silvicola, Torulaspora delbrueckii, Hanseniaspora uvarum, Saturnispora silvae, Issatchenkia orientalis, Candida mucifera, Kazachstania barnettii, Cyberlindnera jadinii, Hanseniaspora spp. Alternaria tenuissima, Cryptococcus laurentii, Metschnikowia spp., and Rhodotorula dairenensis) that we detected in our initial DGGE analysis were not isolated via the present culture-bassed method. This may suggest that the utilized culture medium was not appropriate for these yeast species, or that they were no longer viable in analyzed samples. Future studies of culture media selectivity will be necessary to more fully understand pit mud microecology.

Table 3 Isolated yeast strains identities following purification

We additionally noted that certain species detected via our culture-dependent approach (Schizosaccharomyces pombe and Debaryomyces hansenii) were not evident in the above DGGE fingerprints profiles. This may be a consequence of differences in sample handling protocols that impacted microbial growth or viability, such as variations in sample temperature or aerobic/anaerobic storage (Zhang et al. 2016). The PCR-DGG approach also has a detection limit of 10⁴-10⁸ cfu/mL (Ercolini 2004). As such, microbe concentrations and numbers and pit mud may limit our ability to detect less abundant species via DGGE as a consequence of changes in DNA extraction and PCR amplification efficiency.

Many of the yeast species identified in the present analysis were also detected in our prior analysis of the microbial communities in Daqu-starter samples (Yan et al. 2019). Daqu-starter contains large quantities of yeast, making it a valuable crude microorganism source accounting for 10–20% of the raw material used in liquor production. We therefore speculate that pit mud microbial communities are derived in large part from the initial Daqu-starter.

Assessment of spatial volatile compound profiles in fermented grains samples

In total, 66 different volatile compounds were detected via HS-SPME-GC-MS in analyzed samples collected from the upper, middle, and bottom layers of fermented grains, including 14 acids, 19 esters, 18 alcohols, 6 aldehydes, 2 ketones, 5 alkanes, and 2 volatile phenols (Table 4).

Table 4 The volatile aroma compounds detected and measured in the samples collected from different spatial positions of fermented grain

Of the 14 acids detected in the middle and bottom fermented grains layers, the levels of acetic acid were highest in all three layers, while 2-methyl-butanoic acid and 3-methyl-pentanoic acid were present only in the middle and bottom layers and not in the upper layer.

Esters were the most abundant and important aroma compounds in these fermented grains samples. We found that levels of ethyl acetate, ethyl hexanoate, ethyl butanoate, ethyl hexanoate, ethyl oenanthate, ethyl 2-methylbutanoate, ethyl 3-methylbutanoate, nonanoic acid ethyl ester, ethyl heptanoate, ethyl laurate, γ-nonylactone, and ethyl octadecanoate were highest in samples collected from the bottom layer of fermented grains, followed by levels the middle layer. Ethyl isobutanoat levels were highest in the middle layer of fermented grains, while benzeneacetic acid ethyl ester and ethyl pentadecanoate were present at the highest levels in the upper layer. Levels of ethyl decanoate, ethyl oleate, ethyl 9-hexadecenoate, ethyl palmitate, and ethyl linoleate did not differ significantly among fermented grains layers.

Alcohols were also present at high levels in fermented grains samples, as shown in Table 4. Levels of 3-methyl-butanol, 2-methyl-1-propanol, isobutanol, 1-butanol, 2,3-butanediol, 2-methylbutanol, 1-pentanol, 2-methylbutanol, 1-pentanol, 2-heptanol, and phenylethyl alcohol in the bottom fermented grains layer were significantly higher than those in other layers, while the middle layer contained the highest levels of 1-hexanol, and the upper layer contained the highest levels of isoamyl alcohol, 1-octen-3-ol, isooctanol, octanol, isopentanol, 1-nonanol, and benzyl alcohol. Ethanol levels did not differ significantly among fermented grains layers.

The highest total levels of other volatile compounds such as aldehydes, ketones, alkanes, and volatile phenols were detected in the middle layer of fermented grains, with the second highest levels being detected in the bottom fermented grains layer, whereas these levels were lowest in the upper fermented grains layer.

A PCA approach was next used to assess the distributions of these 66 volatile compounds in different fermented grains sample layers (Fig. 3). Samples from these three layers clearly separated into three clusters based upon the volatile compounds detected therein. The bottom layer of fermented grains contained relatively high levels of volatile acids and esters including acetic acid (AC1), propionic acid (AC2), butyric acid (AC3), caproic acid (AC4), 3-methyl-pentanoic acid (AC5), 2-methyl-butanoic acid (AC6), 2-methyl butanoic acid (AC8), pentanoic acid (AC9), nonanoic acid (AC10), palmitic acid (AC12), octanoic acid (AC13), decanoic acid (AC14), ethyl acetate (ES1), ethyl butanoate (ES3), ethyl hexanoate (ES4), ethyl oenanthate (ES5), ethyl 2-methylbutanoate (ES6), ethyl 3-methylbutanoate (ES7), nonanoic acid ethyl ester (ES8), ethyl heptanoate (ES10), ethyl laurate (ES12), ethyl octadecanoate (ES19), 2-methyl-1-propanol (AL4), 2,3-butanediol (AL10), 2-methylbutanol (AL12), 1-pentanol (AL14), 2-heptanol (AL16), phenylethyl alcohol (AL18), and 2-methoxy-4-vinylphenol (VP2), consisteint with our previous studies demonstrating high levels of esters in this lower fermented grains layer (Yan et al. 2015). In the present analyses, we found that the fusel alcohols isoamyl alcohol (AL2), 1-octen-3-ol (AL5), isooctanol (AL8), octanol (AL11), isopentanol (AL13), 1-nonanol (AL15), and benzyl alcohol (AL17) were primarily concentrated in the upper layer of fermented grains, while the middle fermented grains layer contained high levels of tetradecane (AK2), hexadecane (AK5), ethyl isobutanoat (ES2), ethyl isobutanoat (KE2), phenol (VP1), caryophyllene (AK4), and 2-undecenal (AD4).

Fig. 3

Principal component analysis of volatile compounds on three layers of fermented Zaopei samples. The first principal component (X axis) explains 78.3% of the total variance of the dataset, while the second principal component (Y axis) explains 18.7% of the total variance of the dataset. UZ, MZ, and DZ, represent the samples collected from up, middle, and down layer of fermented Zaopei, respectively

Correlations between yeast communities and volatile compounds

We next conducted a canonical correspondence analysis (CCA) to evaluate correlations between pit mud yeast communities and volatile compounds present in fermented grains. As shown in Fig. 4, the first two component axes in this analysis explained 76.1% of the variation in community composition. Torulaspora delbrueckii (4), Hanseniaspora uvarum (6), Saturnispora silvae (7), Geotrichum bryndzae (8), and Pichia farinosa (12) were positively correlated with levels of caproic acid (AC4), 2-methyl-butanoic acid (AC6), octanol acid (AC7), 2-methyl butanoic acid (AC8), and palmitic acid (AC12), while Pichia anomala (11), Issatchenkia orientalis (13), Yarrowia lipolytica (16), Wickerhamomyces anomalus (17), Candida intermedia (18), Trichosporon asahii (22), Pichia guilliermondii (25), Candida humilis (27), Candida tropicalis (28), Cyberlindnera jadinii (29), Hanseniaspora vineae (30), Metschnikowia spp. (33), and Saccharomyces cerevisiae (36) were positively correlated with levels of hexanoic acid (AC11), octanoic acid (AC13), 1-hexanol (AL3), ethyl butanoate (ES3), ethyl hexanoate (ES4), nonanoic acid ethyl ester (ES8), benzeneacetic acid ethyl ester (ES11), γ-nonylactone (ES13), ethyl oleate (ES14), ethyl pentadecanoate (ES15), ethyl 9-hexadecenoate (ES16), and ethyl octadecanoate (ES19). Geotrichum silvicola (2), Geotrichum silvicola (3), Geotrichum bryndzae (9), Saccharomycopsis fibuligera (10), Alternaria tenuissima (14), Pichia kudriavzevii (19), Pichia kudriavzevii (20), Pichia occidentalis (21), Kazachstania barnettii (24), and Cryptococcus laurentii (32) were closely associated with levels of propionic acid (AC2), butyric acid (AC3), pentanoic acid (AC9), nonanoic acid (AC10), decanoic acid (AC14), ethyl oenanthate (ES5), ethyl 2-methylbutanoate (ES6), ethyl 3-methylbutanoate (ES7), ethyl heptanoate (ES10), and ethyl linoleate (ES18). Geotrichum silvicola (1), Candida mucifera (15), Trichosporon asahii (23), Cryptococcus laurentii (31), Rhodotorula dairenensis (34), and Saccharomyces cerevisiae (35) were positively correlated with 3-methyl-butanol (AL1), isoamyl alcohol (AL2), ethyl acetate (ES1), ethyl decanoate (ES9), and ethyl palmitate (ES17) levels.

Fig. 4

Canonical correspondence analysis (CCA) of yeast community and volatile compounds. Yeasts are numbered as indicated on the DGGE fingerprint files shown in Fig. 2; Table 2

Discussion

Distilled liquors contain a high ethanol content, and Chinese liquors are those to be among the oldest distillates in the world (Pu and Yan 2022). Chinese liquors are broadly classified into 12 different flavour types. Of these, strong-flavour liquor is the most popular in China. This liquor is prepared via fermentation in specialized rectangular pit mud cellars (Tan et al. 2020). This pit mud provides an effective habitat for microbial growth and metabolism during liquor distillation, with the microbes therein serving as important determinants of the flavour of the resultant alcohol (Wu et al. 2009). Pit mud composition is thus a key factor influencing Chinese strong-flavour liquor quality (Xu et al. 2017). Pit mud can provide an environment conducive to fermentation, with the filtration and heat retention properties of this material having a pronounced impact on this process. In addition, pit mud can serve as an environment for microbial growth, and the aromatic compounds derived from these microbes can ensure liquor quality. Many microbes are present within pit mud, including bacteria and archaea, and their metabolic byproducts are a primary source of aroma-related compounds (Zhao et al. 2012). As such, most studies of pit mud to date have focused on bacteria.

Although the yeast are an essential part of pit mud microorganisms (Zhao et al. 2012), the diversity of yeast was low, with only the genera Wickerhamomyces, Kluyveromyces, Pichia, Candida, Zygosaccharo-myces, and Geotrichum was reported in previous investigation (Wang et al. 2017). In the present study, we employed both culture-dependent and PCR-DGGE approaches to facilitate multidimensional analyses the yeast communities of pit mud. Our data suggested that there were significant differences in yeast communities in different pit mud layers. Geotrichum silvicola (band 1), Pichia farinosa (band 12), Kazachstania barnettii (bands 24), Pichia guilliermondii (band 25), Hanseniaspora spp. (band 26), Candida humilis (band 27), Cyberlindnera jadinii (band 29), and Cryptococcus laurentii (band 32) were only detected in the middle pit mud layer, whereas Torulaspora delbrueckii (band 4), Hanseniaspora uvarum (band 6), Candida tropicalis (band 28), Hanseniaspora vineae (band 30), and Rhodotorula dairenensis (band 34) were only present within the bottom layer. In addition, Geotrichum bryndzae (band 9) and Issatchenkia orientalis (band 13) were only present in the bottom pit mud layer. PCA analyses revealed clear differences in the microbial profiles of pit mud samples from different cellar locations (Fig. 3).

In our culture-dependent analysis, we did not detect the presence of several yeast species (Geotrichum silvicola, Torulaspora delbrueckii, Hanseniaspora uvarum, Saturnispora silvae, Issatchenkia orientalis, Candida mucifera, Kazachstania barnettii, Cyberlindnera jadinii, Hanseniaspora spp. Alternaria tenuissima, Cryptococcus laurentii, Metschnikowia spp., and Rhodotorula dairenensis) that were observed via PCR-DGGE. In contrast, other species (Schizosaccharomyces pombe and Debaryomyces hansenii) were detected only in culture-dependent analyses and not in DGGE fingerprint profiles. These findings emphasize the value of simultaneously conducting both culture-dependent and -independent assays in order to fully characterize pit mud yeast communities. Interestingly, many of the species detected in the present analysis were similar to those detected in our prior study of the microbial communities associated with Daqu-starter samples (Yan et al. 2019). Indeed, Daqu-starter is generally utilized as a crude microorganism source containing high levels of yeast. Daqu-starter accounts for 10–20% of the total raw material used in the liquor production process, suggesting that the microbial community of pit mud is largely influenced by the Daqu-starter.

The contents of flavour compounds in fermented grains displayed striking changes associated with the spatial locations of the cellar. And the microbiotas also showed striking changes associated with spatial location. Because the various flavour components are produced by the diversity of micro-organisms in the pit-mud. In our multidimensional HS-SPME-GC-MS analysis, we detected 66 volatile compounds in analyzed fermented grains samples, revealing the highest levels of these volatile acids, esters, and alcohols in the bottom layer of fermented grains, in line with prior studies (Zhang et al. 2020). The middle fermented grains layer contained the highest levels of aldehydes, ketones, alkanes, and volatile phenols, followed by the bottom layer. A CCA approach further revealed strong correlations between pit mud yeast community composition and the volatile flavour compounds detected in fermented grains samples, suggesting that yeast species are likely to have a profound impact on the flavour of Chinese strong-flavour liquor even though they are present at relatively low levels in pit mud as compared to bacterial species (Zhang et al. 2015). It was previously reported that Saccharomyces cerevisiae have the ability to ferment saccharides from the fermentative raw materials (cereals) of Chinese strong-flavour liquor to obtain ethanol (Wu et al. 2015). Candida and Pichia have the ability to metabolize esterases for the biocatalytic synthesis of flavour esters in the liquor (Raghavendra et al. 2014). Saccharomycopsis fibuligera produces ethanol and higher alcohols as well as substantial levels of esters and volatile acids(Liu et al. 2017). These results can be also confirmed by our present study.

This study is the first we are aware of to have assessed pit mud yeast community composition via both culture-dependent and –independent approaches. By highlighting the potential importance of yeast as determinants of fermented grains flavour, our results provide a strong foundation for the study and improvement of pit mud composition during Chinese strong-flavour liquor fermentation.