Workshop environment heterogeneity shaped the microbiome and metabolome profiles during Xiasha round of Jiangxiangxing Baijiu

Highlights

• •Microbials of environment and fermentation process among workshop N and O varied.
• •Dominant environment microbiota from O were migrated into fermented grains.
• •Environment is the most crucial factor that affects the fungal community structure.
• •Flavor compounds were mainly formed at pit fermentation of O.

Abstract

Workshop with different fermentation years plays an essential role in the yield and quality of Baijiu. In actual production, the quality of base Baijiu in newly built workshop is inferior to the older one. In this study, the microbiota of workshop environment and fermentation process from two workshops namely N (ferment 2 years) and O (ferment 20 years) and flavor compounds were studied during Xiasha round. Results showed workshop O accumulated more environmental microorganisms and fungi including P. kudriavzevii, Wickerhamomyces anomalus and Saccharomyces sp mainly came from ground. Yeasts including Pichia, Cyberlindnera, Wickerhamomyces and Candida were responsible for flavor substances formation in O while Saccharopolyspora was in N. This study for the first time explored the reasons for the brewing differences among N and O workshop from perspectives of workshop environment, microbial community and flavor substances, providing new ideas for guiding production as well as improvement of Baijiu quality.

Introduction

Jiangxiangxing Baijiu (also named as Moutai-aroma or sauce-aroma baijiu), mainly produced in the southern part of China, is popular for its soy sauce-like roasted aroma, endowed by its unique brewing technique and complex microbial consortia (Jin, Zhu, & Xu, 2017). Its production process takes about a whole year with 7 rounds, comprising of seven repetitions: stacking fermentation on the ground, and alcoholic fermentation in the pit (Wang et al., 2018, Wang et al., 2018). Stacking fermentation was conducted in an open or semi-open environment so that specific microbiota have been well enriched with the unique ecological environments and manufacturing procedures through repeated practices for a long time (P. Li, Lin, Liu, Wang, & Luo, 2016). Stacking fermentation is an important process that are critical for the yield and quality of raw Baijiu because complex microbiome is involved, especially the enrichment of fermentation microorganisms (Yang et al., 2023). As heap fermentation progressed, the microbial communities succession driven by environmental conditions and microbial communities triggers bioheat and contributes to a rapid increase in temperature of fermented grains (Wang et al., 2018, Wang et al., 2018). When the top of the heap reach a certain temperature, the fermented grains are transferred to a pit and then are sprinkled evenly with tail baijiu (the baijiu obtained from the final stage of distillation)(P. Li et al., 2016). Pit fermentation is to create an environment for microorganisms decomposing starch and protein, producing alcohol, aromatic substances as well as organic acids.

With the rapid development of modern Baijiu enterprises, the improvement of people's living standard and the growing demand of consumers for Baijiu, many liquor enterprises begin to expand their production scales and build some new factories and workshops to ensure the production of Baijiu. An increasing number of liquor companies are paying high attention to improving the quality of Baijiu, constantly pursuing excellence to ensure the sustainable development of the enterprise. However, in empirical practice, there exists a phenomenon that the main aroma of body liquor from new workshop is not prominent with poor mellowness and lacking of coordination. In addition, some flavor components may be high or low, resulting in poor taste. By contrast, the older workshop has experienced mature domestication and formed stable microbial consortia, ensuring suitable environment and stable liquor quality (Xiao et al., 2021). Xiasha round is the beginning stage of the whole fermentation process and fresh sorghum is infiltrated with water, steamed and gelatinized to promote microbial metabolism, producing flavor substances and their precursors for the first time (Hao et al., 2021), which lays foundation for the latter rounds.

Previous studies have proved microbiota from the environment is participated in the fermentation process of Baijiu and could drive both microbial succession and metabolic profiles (Lu et al., 2022, Xu et al., 2024). However, relatively rare research was conducted on comparing different brewing workshops in the Xiasha round. High-throughput sequencing based on the MiSeq platform was applied to reveal the composition structure of microorganism from workshop environment and fermentation stages during Xiasha round, and combined with HS-SPME-GC–MS to detect main flavor compounds in fermented grains. Plus, the potential correlation between the microbial communities and physicochemical index as well as volatile compounds were also analyzed, in the hope of providing some basis for exploring the impact of the environment during the course of Jiangxiangxing baijiu fermentation process.

Material and methods

Sample collection

Jiangxiangxing Baijiu production is a year-long process that starts in autumn and ends in summer of the next year. Samples were collected in two workshops namely N, fermentation for about 2 years, as well as O, fermentation for approximately 20 years, from the same liquor distillery from October 2021 to December in Zunyi city (27.13 N,106.17 E), Guizhou Province, China. Environment samples were collected from air, ground and the surface of cellar walls from two workshops. Fermented grains samples were collected at day 0, 2, 4 of stacking fermentation and day 3, 7, and 25 of pit fermentation. The diagrams of stacking fermentation and pit fermentation process were shown in Fig. S1. During stacking fermentation, samples were collected separately using five-point sampling method (four corners and center of middle layer), while pit fermentation were six-point sampling method (upper layer (0.5 m), middle layer (1.5 m), bottom layer (2.2 m) from two positions), and then the samples from each point were mixed as one sample to eliminate sampling errors. Two pits from workshop N and O were tracked during Xiasha round. That is to say, a total of 12 fermented grains samples were obtained. Ten grams evenly mixed samples were dissolved in 100 mL PBS buffer, shaken well and centrifuged to obtain precipitate for sequencing. The remaining samples were sealed in sterile sampling bags and stored in −20 °C refrigerator for identification of flavor substances and physicochemical analysis.

High throughput sequencing

Genomic DNA was extracted from samples using the Powersoil DNA extraction kit (Mobio Laboratories). For bacteria, the V3-V4 hypervariable regions of the 16S rRNA gene were amplified using the primers 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′- GGACTACHVGGGTWTCTAAT-3′)(Soergel, Dey, Knight, & Brenner, 2012). For fungi, the internal transcribed spacer regions were amplified using the primers ITS3(5′ GCATCGATGAAGAACGCAGC-3′) and ITS4(5′-TCCTCCGCTTATTGATATGC-3′)(Ihrmark et al., 2012). The amplification conditions were listed as follows: 95 °C for 3 min; 27 cycles of 95 °C for 30 s, 55 °C for 30 s, 72 °C for 45 s and 72 °C for 10 min. The amplicons and DNA samples were subjected to high throughput sequencing (Illumina Miseq), performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd.

Determination of physicochemical properties

The moisture content of samples was measured using the gravimetric method by drying samples to a constant weight at 105 ℃. The total titratable acidity, starch and reducing sugar content were determined as previously described (M.-Y. Wang, Zhao, Chang, & Yang, 2019).

Analysis of volatile flavor compounds

Headspace solid-phase microextraction combined with gas chromatography-mass spectrometry (HS-SPME-GC–MS) was used to detect volatile flavor compounds in fermented grains, including a 7890B gas chromatography-mass spectrometer coupled to a 5977A mass selective detector (Agilent) with a CP-WAX57CB polar capillary column (60.0 m × 0.25 mm × 0.4 μm; Agilent). Fermented grains (5 g) were added to 20 mL sterile saline (0.85 % NaCl, 1 % CaCl2), ultrasonically treated for 30 min (in an ice-water mixture as 0 ^◦C), and then centrifuged at 8,000 × g for 5 min (4 ◦C). Eight milliliters supernatant and 10 µL internal standard (Menthyl Acetate, 0.1 g/L, Sigma-Aldrich, St. Louis, MO, United States) were placed into a 20-mL headspace vial containing 3 g sodium chloride. The SPME needle (50/30 µm DVB/CAR/PDMS fiber (Supelco, Bellefonte, PA, United States) preconditioned in the GC injection port at 250 ^◦C for 30 min, being inserted into the headspace of the vial for 60 min at 60 ^◦C to adsorb volatile substances. Then, desorption was conducted immediately for 5 min at 250 ^◦C. The GC conditions were: initial temperature 55 ℃, which was increased at 3 ^◦ C/min to 85 ℃ and maintained for 5 min, and then elevated to 150 ^◦C at 5 ^◦C/min and maintained for 4 min; The flow rate of a helium (99.999 %) carrier gas was 1 mL/min. The front inlet temperature was 250 ℃ in the splitless mode. The operation of the MS was: temperature of the ion source at 230 ^◦C, ionization energy at 70 eV, acquisition mode in full scanning mode, and data collection over the range of 30–400 m/z.

Data analysis

Redundancy analysis (RDA) was performed to reveal the potential correlations between microbiota and fermentation parameters with Canoco 5.0 software. Principal coordinates analysis (PCoA) was applied to explore the possible differences between the samples based on Bray-Curtis distances via ANOSIM using R (version 3.3.1). Volatile flavor compounds were visualized in heatmap via TBtools. Spearman correlation analysis between volatile compounds and the dominant genera in fermented grains were calculated via Cytoscape (v.3.4.0).

Results and discussion

Analysis of physicochemical factors of fermented grains from different workshops

Fermentation parameters were important factors regulating the microbial successions and metabolism (Bian et al., 2022). These physicochemical indicators are usually applied to determine the state and degree of fermentation and can also reflect the microbial state during fermentation (Zhao, Su, Mu, Jiang, & Mu, 2020). The dynamics of fermentation parameters namely moisture, acidity, starch and reducing sugar among workshop N and O were shown in Fig. 2. The moisture content was higher in O through entire process and had less fluctuation, yet it present a stable upward trend in N during period of pit fermentation (Fig. 2A). Acidity accumulated throughout the whole fermentation process among two workshops and there was no significant difference among N and O during stacking fermentation. However, in the course of pit fermentation, the acidity content was significantly higher in O particularly at the end of pit fermentation (Fig. 2B), which may be attributed to the pit environment, continuous fermentation and microbial interactions in the old pit endow it with higher acidity (S. Xu et al., 2022). The initial starch content was higher in N and gradually declined on day 0 from 43.25 % to 38.5 %, with a decreasing trend in the stacking fermentation (SF) stage. By contrast, the content increased on day 0 in O from 34.2 % to 37.7 %, with an increasing trend in SF stage (Fig. 1D). Till the end of pit fermentation, the two workshops were going to be similar. The reducing sugar content increased significantly on day 0 in N from 0.67 to 2.13 g and gradually declined with the process of stacking fermentation, by contrast, it also showed decreased trend during initial stage of SF in O but an elevating trend during stage SF to PF and was higher in O at the end of pit fermentation (Fig. 1C), which may attributed to divergence of active microbial metabolism among two workshops(Guan, Lin, Chen, Ou, & Zhang, 2020).

Fig. 2.

Microbial communities compositions of the workshops at the genus level. (A), (C) Prokaryotic community composition of the O and N; (B), (D) Fungal community composition of the O and N workshop. Taxa compromising < 1 % of the total relative abundance across all samples are grouped as others.

Fig. 1.

Dynamics of physicochemical indexes of Moisture (A), Acidity(B), Reducing sugar(C) and Starch (D) in fermented grains samples from workshop N and O throughout fermentation. S0, S4 means stacking fermentation for 0d, 2d, 4d; P3, P7, P25 means pit fermentation for 3d, 7d, 25d.

Microbial communities landscape in environment and fermentation process among N and O workshops

Environment is a crucial factor affecting the microbial community succession in the naturally occurring fermentation ecosystem, understanding how the environmental microbiota participates in fermentation is helpful to clarify the importance of environmental microbiota and provide controllable management strategies (Yang et al., 2023). Environmental microbes are also one of the most important factors affecting the quality, quantity, and favors of Baijiu(Zheng et al., 2014). In order to explore microbial ecosystems in workshop of N and O, samples were collected from air, ground, cellar walls and fermentation process for high throughput sequencing. According to the microbial taxonomic profile, as for bacteria, air samples from different workshops showed different compositional patterns. The main bacterial genera included Saccharopolyspora (27.47 %), Kroppenstedtia (26.95 %) and Thermoactinomyces (15.23 %) from O, Kroppenstedtia (23.78 %), Virgibacillus (13.4 %) and Saccharopolyspora (12.38 %) from N were the dominant genera of air samples, respectively. For the ground samples, the two workshops also showed big difference in bacterial compositions. Microbial communities were more diverse in O while relatively single in N. Pediococcus (37.18 %), Lactococcus (22.79 %), Staphylococcus (13.68 %) and Leuconostoc (13.58 %) from O were dominant genera, by contrast, Pantoea (90.7 %) showed higher relative abundance in N. For the contact surface of CW samples, Lactobacillus (90.9 %) and Leuconostoc (83.4 %) showed higher relative abundances in O and N, respectively. Thus, the dominant microbial communities in environmental samples among O and N showed obvious distinction(H.-Y. Wang, Zhang, Zhao, & Xu, 2008). In addition, bacterial compositions during fermentation process were also analyzed. Bacterial communities in fermented grains samples were mainly composed of Weissella, Lactobacillus, Pediococcus, Acinetobacter, Kroppenstedtia, Saccharopolyspora, Virgibacillus and Bacillus. Weissella and Lactobacillus were dominant in both workshop among entire fermentation process. During stacking fermentation, Pediococcus (29.55 %) and Acinetobacter (30.58 %) were abundant at initial stage of SF in O, with the progress of stacking fermentation, the abundance of Weissella increased while Pediococcus decreased. In the course of pit fermentation, Weissella and Lactobacillus compete with each other, till the end of pit fermentation, Lactobacillus occupied absolute advantage position with an abundance of 70.34 %, which might also account for higher content of acid in O. Generally, cellars with longer fermentation years would accumulate more Lactobacillus (Song, Du, & Zhang, 2017). Lactobacillus, as the core functional microorganism, is able to produce lactic acid, ethanol, and acetic acid through heterolactic fermentation to increase acid (Song et al., 2017). Also, Lactobacillus regulates the acidity of fermented grains through lactic acid metabolism, thus inhibiting the growth of other miscellaneous bacteria(Katina, Sauri, Alakomi, & Mattila-Sandholm, 2002). For fermentation samples of N, Weissella was dominant among entire fermentation process, the abundance ranging from 33.69 % to 94.21 %, followed by Lactobacillus (20.33–44.79 %) and Pediococcus (7.63 %-35.87 %). The above analysis showed that dominant genera showed obvious abundant difference due to distinction of workshop environment.

Fungi usually play an important role in degrading of biopolymers, secreting various enzymes, producing alcohol and forming aromatic compounds(B. Chen, Wu, & Xu, 2014; Y. Li, Cheng, Wang, Hu, Wang, & Huang, 2022). Starch in sorghum cannot be directly utilized by most of the yeasts and bacteria, so it needs to be hydrolyzed into fermentable sugars through α-amylase and glucoamylase, which are produced by filamentous fungi(B. Chen et al., 2014). As for environment samples in air, as shown in Fig. 2B, the main fungal genera included Thermomyces (64.41 %) and Thermoascus (15.11 %) from O, Aspergillus (37 %) and Talaromyces (34 %) from N were dominant genus. For the ground sample, the main fungal genera included Torulaspora (82.41 %) from O, Penicillium (17.33 %) and Saccharomyces (16.58 %) from N, were dominant genus. Cystofilobasidium (37.07 %) was dominant on surface of CW. For the surface of CW samples, Penicillium (71.57 %) and Cystofilobasidium (37.07 %) were respectively dominant fungal genus among O and N. The fungal communities in fermented grains were mainly composed of Pichia, Thermomyces, Thermoascus, Cyberlindnera, Saccharomyces, Saccharomycopsis, Aspergillus, Candida, Byssochlamys, Monascus, Torulaspora and Wickerhamomyces. During the fermentation process in O, Pichia possessed higher abundance over 70 %, at the end of pit fermentation, Saccharomyces replaced Pichia being the dominant genus with an abundance of 81.44 %. In N, species were more diverse. During stacking fermentation process, Candida and Cyberlindnera were dominant, Saccharomycopsis and Pichia occupied a superior position in process of PF. At the end of PF, Pichia hold an abundance of 51 %, while Candida and Saccharomyces only occupied 14 %, forming a significant difference with that of O. Both of Saccharomyces and Pichia were mainly participated in ethanol fermentation, producing main odor substances and metabolizing a variety of esters and aromatic compounds, possessing a remarkable ability to ferment sugars into lactic acid, acetic acid and ethanol(C. Chen et al., 2022). Saccharomycopsis (47.38 %) and Candida (22.12 %) mainly existed in N, Candida is an aromatic yeast that can produce esterase and promote the formation of esters (Perrusquía-Luévano, Cano-Herrera, Guigón-López, Avitia-Talamantes, Torres-Torres, & Villalpando, 2018). Saccharomycopsis can actively participate in the saccharification by secreting abundant enzymes such as ɑ-amylase, glucoamylase and xylanase with high activity(Xie, Zhang, Kang, & Yang, 2021).

Stacking fermentation is a solid-state open fermentation, which causes differences in spatial microecology of the fermented grains(Yang et al., 2023). The main purpose of pit fermentation is to boost the relative abundance of functional microorganisms and enrich aromatic compounds(H. Liu & Sun, 2018). To reveal the similarities and dissimilarities of microbial communities during different fermentation stages among two workshops, groups were retrieved from result of PCoA. The cumulative explained variance ratios of the first two PCoA based on bacterial (Fig. 3A) and fungal (Fig. 3B) communities were as high as 73.87 % and 67.13 %, respectively, indicating that the PCoA could comprehensively reflect the profiles of samples. As shown in Fig. 3A, the bacterial communities were significantly differentiated during stacking fermentation (SF) and pit fermentation (PF) among both workshops (p < 0.005). Nevertheless, the communities were similar based on the OTU (operational taxonomic unit) levels at the same stage of SF or PF among N and O. However, it is not true for fungi communities. As shown in Fig. 3B, the samples points were clearly separated in fungal genus among two workshops based on SF stage, illustrating that fungal communities were more susceptible to workshop environment than bacterial communities. Therefore, fungal communities in the fermentation process may primarily originate from the environment and this accords with previous studies(Y. Li et al., 2022). Further, LEfSe analysis was used to asses multi-level differential microorganisms among two workshop, and LDA scores were used to measure the impact of species on differential effects. As shown in Fig. 3C and 3E, whether in SF stage or the PF stage, Weissella cibari was differential bacterial species of N, while Lactobacillus brevis, Lactobacillus dextricicus and Lactobacillus vaccinosterus were differential fungi species of O. Specially, Lactobacillus pontis and Lactobacillus buchneri were only differential microorganisms in the PF stage of O, which may be potential important biomarkers. Candida tropicals from N, Torulaspora delbrueckii, Lodderomyces elongosporus, Cyberlindnera rhodanensis from O, were differential fungal species during different fermentation stages (Fig. 3D and 3F). Based on these results, it can be concluded that there are significant differences in microorganisms during different fermentation stages between N and O workshops.

Fig. 3.

PCoA analysis and Linear discriminant effect size analysis of the bacteria and fungi compositions in two workshops and different fermentation stages (LDA greater than 3, p < 0.05). Histogram of LDA scores calculated for features differentially abundant between groups. (A), (B) PCoA analysis of bacteria and fungi communities. (C), (D) differential bacteria and fungi microbial during stacking fermentation. (E), (F) differential bacteria and fungi microbial during pit fermentation.

The microorganisms transition from environment to fermentation process

Xiasha round is initial stage among entire Jiangxiangxing baijiu fermentation process, from an ecological perspective, the microorganisms in the fermented grains samples mainly come from the air, ground, instruments, sorghum and Daqu (H. Liu et al., 2018; Zhang et al., 2021). With the repeated mass production of Jiangxiangxing baijiu, the advantageous microorganisms in the fermented grains with stronger environmental tolerance gradually established stable microbial communities in this specific environment (Hu, Du, Ren, Xu, & Björkroth, 2016). These environmental microorganisms interact with the microorganisms in the fermented grains, profoundly influencing the composition and metabolism of the microbial community (Z.-M. Wang et al., 2016, Wolfe and Dutton, 2015). In order to clarify how differences in workshop environment affect the composition of microbial communities during the fermentation process, microorganisms with relative abundance greater than 1 % in species level were selected for analysis. As shown in Fig. 4, Thermoactinomyces vulgaris, Saccharopolyspora rectivirgula and Bacillus thermolactis, dominant genera in air of O, were also dominant at early stage of accumulation; Pediococcus pentosaceus, as dominant genus on ground, was also dominant in fermentation process. However, unclassified Lactobacillus, Leuconostoc lactis and Staphylococcus equorum, accounting for the largest proportion on the ground, did not appear in fermentation process. unclassified Acinetobacter, Lactobacillus acetotolerans and Lactobacillus parafarraginis, as dominant microorganism on the surface of the CW, have also not been detected during the fermentation process. Weissella paramesenteroides, Lactobacillus plantarum, uncultured Lactobacillus and Lactobacillus brevis were not detected in the environment sample, but holding dominant position in the fermentation process, which are indispensable dominant microbiota in the brewing process. As for species in N, as shown in Fig. 5, the dominant microorganisms in air were Paraburkholderia fungorum, Sphingobacterium daejeonense, Mucilaginibacter daejeonensis, Prauserella rugosa and Cutibacterium acnes. Microbials including unclassified Pantoea and Weissella viridescens on ground, Leuconostoc lactis and unclassified Lactococcus on the surface of CW, dominant in N, did not existed in fermentation process. Instead, unclassified Nocardiopsis was detected in large quantities during fermentation process. For fungal species of O, Thermomyces sp, Thermoascus crustaceus, Byssochlamys spectabilis, Saccharomycopsis fibuligera, Thermoascus aurantiacus and Aspergillus cristatus, mainly existed in air, were also found at initial stage of SF, indicating the initial stage of accumulation is a process of enriching air microorganisms. Microbiomes like P. kudriavzevii, Wickerhamomyces anomalus and Saccharomyces sp, abundant on the ground, were also rich in fermentation process(Song, Du, Zhang, & Xu, 2017). Penicillium carneum and Torulaspora delbrueckii, remarkably dominant on the surface of CW, yet have not appeared in fermentation process. Cyberlindnera fabianii, with higher abundance particularly at SF stage, was not found in environment. As for fungi in N, species were more diverse(Fig. 5). Aspergillus cristatus and Thermomyces sp, mainly abundant in air, were also abundant at initial stage of SF, which also proved microorganisms during the initial accumulation mainly originated from the air. Specially, Talaromyces wortmannii was also abundant in air, but did not exist in fermentation process. Dominant ground microbials like P. kudriavzevii and Saccharomyces sp were also a crucial source of fermentation process. However, Debaryomyces sp, Torulaspora delbrueckii, Rasamsonia composticola and Cystofilobasidium infirmominiatum, both abundant on ground and CW, did not appear in fermentation process. Candida tropicalis, Saccharomycopsis fibuligera and Cyberlindnera fabianii were abundant during fermentation process, yet have not traced back to environment. Based on these results, we concluded that workshop O was more likely to enrich environmental microorganisms forming a stable microecology through long-term domestication. By contrast, environmental microorganisms in workshop N may be not suitable for brewing environment and have not penetrated in fermented grains sample.

Fig. 4.

Bacterial species in the O and N workshop. Only showing genera with relative content ≥ 1 % in at least one type of sample.

Fig. 5.

Fungal species in the O and N workshop. Only showing genera with relative content ≥ 1 % in at least one type of sample.

Correlation between microbial communities and physicochemical parameters

Redundancy analysis (RDA) and canonical correlation analysis (CCA) were used to clarify key environmental factors impacting bacterial (Fig. 6a) and fungal (Fig. 6b) microbial composition and diversity in entire fermentation process of two workshops. Results indicated that microbial community structure showed a strong response to dynamics of fermentation parameters. Acidity(r² > 0.9) and starch (r² > 0.6) were main driving forces of both bacterial and fungi communities succession in two workshops. Particularly, the two fermentation indexes showed consistency among two workshops in terms of bacterial communities. By contrast, moisture mainly drove the succession of fungal communities in O while starch did in N.

Fig. 6.

RDA analysis of the physicochemical properties with microbiota. Physicochemical properties with bacterial (a) and fungal genera (b); The heatmap of the correlation index between the top 20 bacterial (c) and fungal (d) genera in relative abundance and physicochemical properties. (*p < 0.5 **p < 0.1).

In order to gain a deeper understanding of the relationship between dynamics of fermentation parameters and microbial community succession, we selected the top 20 bacterial and fungal genera in relative abundance and analyzed their correlation with fermentation parameters (Fig. 6c and 6d). Acidity present stronger positive correlation with Lactobacillus and Saccharomyces, while negatively correlated with Staphylococcus and Thermomyces. Previous study also highlighted the importance of acidity on the composition of the microbial community, indicating the increase of acidity enhanced abundance of Lactobacillus(Zheng et al., 2014). In addition, yeast enhanced the tolerance to acetic acid by some transcription factor(Swinnen, Henriques, Shrestha, Ho, Sá-Correia, & Nevoigt, 2017). Reducing sugar was negative with fungal genus like Thermomyces, Byssochlamys and Wickerhamomyces. Moisture was positive with Saccharomyces, negatively with Staphylococcus. Based on these results, fermentation paramenters exert significant effect on microbial communities, among which acidity was important factors driving the succession of both bacterial and fungal communities. Although moisture and reducing sugar were not main driving forces, they also exerted crucial influence on fungal communities to some extent. To our knowledge, most fungi are more sensitive to heat and moisture than bacteria, in particular yeasts (Wu, Chen, & Xu, 2013).

Volatile flavor compounds and correlations with microbiota

Xiasha round is a non-liquor-producing round in the production cycle of Jiangxiangxing baijiu (Ren, Su, Mu, Qi, & Zhang, 2023). In this round, the role of microorganisms primarily lies in the production of flavor or precursor substances (Liu et al., 2022, Wang et al., 2018). The metabolism of microorganisms in complex fermentation systems is a significant area of interest and a bottleneck in industrial research. Given the substantial disparity in microorganism composition between two workshops, it is important to examine the specific aspects in which this difference manifests. Therefore, we analyzed the volatile flavor components of fermented grains using HS-SPME-GC–MS during fermentation process. In Fig. 7A, it is evident that a total of 26 skeleton flavor compounds were quantified in the samples, incorporating 7 alcohols, 5 esters, 1 pyrazine, 4 aldehydes, 3 phenols, 3 alkanes and 3 terpenes. These volatile flavor substances can contribute to floral and fruity aromas in Baijiu (Cai et al., 2019, Duan et al., 2022, Xu et al., 2022). Notably, the volatile compounds exhibited significantly higher concentrations in O, particularly during the stage of pit fermentation, indicating that more abundant compounds were produced at the stage of pit fermentation in O (Fig. 7A). Noteworthy flavor compounds like A1(Ethanol), A3(1-Butanol,3-methyl-), A5(Phenylethyl Alcohol), A6(1-Butanol), A7(2,3-Butanediol, [S-(R*,R*)]), E3(Ethyl lactate), E5(Decanoic acid, ethyl ester), PY1(Pyrazine, trimethyl-), AD2(Benzeneacetaldehyde), AD4(benzaldehyde,3,4-dimethyl), PH2(phenol, 4-ethyl-2-methoxy-), AK1(Cyclopentasiloxane, decamethyl-), AK3(Octadecane) and T1(menthol)exhibited higher concentrations in O, with A7(2,3-Butanediol, [S-(R*,R*)]), E5(Decanoic acid, ethyl ester), AD2(Benzeneacetaldehyde), and alkanes such as AK1(Cyclopentasiloxane, decamethyl-) and AK3(Octadecane) predominantly formed during the initial stage of stacking fermentation. A3(1-Butanol,3-methyl-), A5(Phenylethyl alcohol), A6(1-Butanol), E3(Ethyl lactate), PY1(Pyrazine, trimethyl-), AD4(Benzaldehyde,3,4-dimethyl), PH2(Phenol,4-ethyl-2-methoxy-) and T1(Menthol) were observed to be formed at the end of pit fermentation. Additionally, AK2(Cyclotrisiloxane, hexamethyl-), A4(1-Pentanol), PH1(Phenol), PH3(Phenol,4-ethyl-), AD1 (Benzaldehyde) were mainly formed in N, among which AK2 (Cyclotrisiloxane, hexamethyl-) were generated at the end of stacking fermentation, A4(1-Pentanol), PH1 (Phenol) and PH3 were produced at the end of pit fermentation(Fig.7B).

Fig. 7.

The heatmap of flavor compounds and correlation with microbiota. Heatmap analysis of volatile flavors during fermentation among O(A)and N(B). Correlation network between microbial genera and flavors in O(C) and N(D).

Microbial metabolism during the fermentation process lays an important foundation for shaping the flavor profile of the base Baijiu (Tan, Zhong, Zhao, Du, & Xu, 2019). To further investigate which microorganisms from 2 workshops had made a difference on the formation of flavor substances, based on Spearman correlation analysis between the dominant microbial genera and the significantly different volatile metabolites were visualized. Results showed that these dominant microorganisms were important contributors to the formation of volatile compounds. Due to the abundance differences in microbial composition among two workshops, their contributions to flavor substances formation also differ. In the two networks, we obtained 22 nodes and 25 edges for the O group, as well as 31 nodes and 32 edges for the N group. In terms of correlation analysis in O (Fig. 7C), for the flavor substances higher in O, A3 was significantly positive correlated with Weissella, Pichia and Cyberlindnera, negatively with Acinetobacter, Oceanobacillus and Aspergillus. A6 was positively correlated with yeasts like Wickerhamomyces and Candida, negatively correlated with Enterococcus and Saccharopolyspora. PH2 present strong positive correlation with Acetobacter and E5 was negative with Saccharopolyspora. Thus, main alcohols were produced by yeasts (De Vuyst, Harth, Van Kerrebroeck, & Leroy, 2016). In terms of correlation analysis in N (Fig. 7D), higher contents compounds in N like PH3 was positive with Staphylococcus and Saccharopolyspora; PH1 and AK2 were both positive with Saccharopolyspora; A4 was positive with Saccharomycopsis and Pediococcus, negative with Saccharopolyspora. However, we also observed that A3 was positively correlated with Weissella in O wherease negatively correlated in N. E4 was negative with Saccharopolyspora in O while positive in N. In summary, Weissella and yeasts like Pichia, Cyberlindnera, Wickerhamomyces and Candida were mainly responsible for the flavor formation in O while bacterial genus namely Saccharomycopsis, Pediococcus and Saccharopolyspora were mainly responsible for the flavor formation in N.

Baijiu fermentation, as a traditional SSF process, relies on individual operation skills and experiences. Expanding production scale and building more and more new workshop are the trend of Baijiu industry development. However, the quality of base liquor in newly built workshop may exist some shortcomings compared with the older one. The difference could be caused by primitive operation environment, which in turn affects the microbial succession and ultimately alter the quality of product (Du, Wang, Zhang, & Xu, 2019). In this study, we investigated workshop environment heterogeneity lead to the difference in microbial community succession and metabolisms during Baijiu fermentation, and estimated the correlation among dominant microbial genus and metabolites via spearman network analysis. Our results revealed that more flavor substances were formed in O, especially at the end of pit fermentation. This strongly indicated that the composition of domesticated microbial communities in O played a crucial role in the formation of flavor compounds. Further, it may explain the reasons for the brewing differences between new and old workshops.

Conclusion

In summary, this study compared the difference of microbials and their metabolites among workshop N and O during Xiasha round of Jiangxiangxing baijiu for the first time. Our research highlighted the importance of fermentation environment on the microbiome during fermentation process, further lead to difference among metabolites. Results indicated that the workshop with longer fermentation years tended to accumulate more environmental microorganisms with fungi mainly originating from the environment. This study also elucidated the contribution differences of dominant microbial genera to the formation of key flavor compounds. Yeasts like Pichia, Cyberlindnera, Wickerhamomyces as well as Candida were responsible for the formation of alcohols in O. By contrast, dominant bacterial genera like Saccharopolyspora was responsible for the formation of phenols and alkanes in N. This study can provide better understanding of connection in terms of microbials among environment and fermentation process. At the same time, it can reveal the differences between new and old workshop in various aspects, which will help develop a potential strategy to improve the quality of Jiangxiangxing Baijiu in new built workshop.

CRediT authorship contribution statement

Cailing Wang: Writing – review & editing, Writing – original draft. Chenyao Li: Writing – review & editing, Validation. Zhiqiang Bin: Software, Resources. Guojun Zhu: Conceptualization. Shaopei Tang: Methodology, Investigation. Jinyu Zhang: Software, Investigation. Yefu Chen: Investigation, Formal analysis. Dongguang Xiao: Investigation, Funding acquisition. Xuewu Guo: Writing – review & editing, Resources, Investigation, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Data will be made available on request.