Introduction

Chinese liquor and vinegar have a long history of production and consumption and are produced through unique brewing processes. They are typically produced from cereals, such as sorghum, by solid-state fermentation using a natural fermentation starter termed Daqu. Daqu comprises a microbial community and is rich in enzymes. Daqu is made by a natural fermentation process running for a few weeks such that a microbial succession occurs (Zheng et al. 2011, 2014). Being a major source of microorganisms and enzymes, Daqu is crucial for the quality, safety, and flavor of its derived products, such as liquor and vinegar. Different types of Daqu can be distinguished (Zheng et al. 2011), for instance, according to the maximum incubation temperature during the fermentation. Daqu can be grouped into three classes based on the production temperature: (1) high-temperature Daqu (60 and 70 °C), (2) medium-temperature Daqu (50–60 °C), and (3) low-temperature Daqu (40–50 °C). According to the raw materials used for production, Daqu can be classified as single-grain Daqu or multi-grain Daqu. Daqu can also be classified geographically into southern and northern Daqu. Generally, southern Daqu is classified as a single-grain product produced at medium to high temperature. The northern variant is commonly a multi-grain, low to medium temperature Daqu (Shen 2001). Several studies have shown the diversity of the microbial community in Daqu (Wang et al. 2011a). We hypothesized that the microbial composition of Daqu correlates with environmental factors prevailing during the fermentation process. Thus, the microbial community in similarly classified Daqu is predicted to harbor common species or similar dominant groups of microorganisms.

The microbial community of Daqu has been analyzed in previous studies using culture-dependent methods, such as isolation and enumeration on selective media (Li et al. 2009; Ma et al. 2011; Zheng et al. 2012), as well as by culture-independent methods, such as polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE), and amplified fragment length polymorphism (Gao et al. 2010; Meng et al. 2010; Yan et al. 2012). In the present study, a semi-quantitative culture-independent cloning method was used for the analysis of microbial communities in Daqu. By comparing clone libraries, not only qualitative information on the composition of the microbial community is obtained, but also quantitative information of the relative abundance of the identified species. The main objective of this study was to obtain an overview of the composition of the microbial communities in different types of Daqu. This analysis is expected to deliver potential biomarkers for fast and reliable verification of the authenticity of Daqu types.

Materials and methods

Sampling

Eight types of brick-shaped Daqu were obtained from five commercial distilleries located in northern and southwestern China. Daqu was produced and matured according to the procedures of the different distilleries. An overview of the types of Daqu and their technological parameters is presented in Table 1. In order to obtain adequate repetition, three blocks of each type of Daqu were randomly selected from each of the upper, middle, and lower stacked layers, and ground together. About 100 g of these Daqu powders was used as an experimental Daqu powder sample. Samples were then collected in sterile Stomacher® bags (Seward Laboratory Systems Inc., London, UK), transported to the laboratory in a cool box, and stored at −20 °C until analysis.

Table 1 Daqu samples investigated
Full size table

DNA extraction and PCR amplification

DNA from seven powdered Daqu samples was extracted according to the method of Wang et al. (Wang et al. 2008) and diluted to a DNA concentration of 50 ng/μL. The 16S rRNA bacterial gene was amplified using universal primers “B-for” (5′-AGAGTTTGATCCTGGCTCAG-3′) and “B-rev” (5′-AAGGAGGTGATCCAGCCGCA-3′) (Edwards et al. 1989). The D1/D2 domain of the 26S rRNA fungal gene was amplified using universal primer “NL1” (5′-TGCTGGAGCCATGGATC-3′) and reverse primer “RLR3R” (5′-GGTCCGTGTTTCAAGAC-3′) (Okoli et al. 2007). PCR was performed in a total reaction volume of 50 μl containing 26.6 μl ddH2O, 5 μl PCR buffer, 3 μl MgCl2 (25 mM), 10 μl dNTP (2 mM), 2 μl of each primer (10 μM), 1 μl DNA template (approximately 50 ng), and 0.4 μl Taq DNA polymerase (5 U/μl) (Fermentas, USA). PCR was performed using a GeneAmp PCR system 9700 (Applied Biosystems, USA) with the following PCR conditions: initial denaturation for 5 min at 94 °C; 35 cycles each consisting of 30 s at 94 °C, 20 s at 56 °C (bacteria) or 52 °C (fungi), and 1 min at 72 °C; and extension of incomplete products for 7 min at 72 °C, followed by cooling at 4 °C. The sizes and quantities of the PCR products were determined using 1.5 % (wt/vol) agarose gel electrophoresis. The PCR products were analyzed by electrophoresis and then stored at −20 °C for future experiments.

DNA clone library construction

Clone libraries of 16S rDNA and 26S rDNA amplicons from Daqu samples were constructed. Amplicons derived from PCR products were purified with a QIAquick PCR purification Kit (Qiagen, Hilden, Germany), cloned using a pGEM-T Easy Cloning Kit (Promega, Madison, WI, USA) and transformed into Escherichia coli JM109 High Efficiency Competent Cells (Promega), following the manufacturer’s instructions. Around 90 positive clones (white colonies) were randomly picked from the plates of each sample. These plasmid-harboring clones were transferred with a sterile toothpick into 50 μl of Tris–EDTA buffer, lysed, and amplified with “T7” and “Sp6 pGem-T”-specific primers to confirm the appropriate size of the insert (approximately 1,500 bp for bacteria and 700 bp for fungi). Clones containing the plasmid with an insert were sent for sequencing at the Beijing Genomics Institute (Beijing, China). Sequences were assembled and edited with Seqman II software (DNAStar Inc., Madison, WI, USA) and aligned with Megalign (DNAStar Inc.). Chimeric sequences in the clone library were identified and discarded using the software package Chromas v.2.31 (Technelysium Pty Ltd.). The nucleotide sequences obtained were identified in GenBank using BLAST (http://blast.ncbi.nlm.nih.gov/) to determine the closest known relatives of the partial ribosomal DNA sequences obtained.

Calculation of species diversity indices

To determine the diversity of species in Daqu samples (as revealed by cloning), Shannon’s diversity index (H′ = −∑p i ln(p i ), where p i is the proportion of taxon i), was calculated.

Statistical analysis

The composition of microbiological communities in all Daqu samples was analyzed by principal component analysis (PCA) using the software package SIMCA-P 12.0 (Umetrics, Umeå, Sweden) to cluster the samples into different groups. Samples were plotted in two dimensions based on scores for the first two principal components to evaluate relationships among samples. The proportion of variance explained by each principal component was calculated.

Results

Composition of microbial communities in Daqu

The composition of the microbial communities in Daqu, representing three temperature types obtained from five factories, is shown in Table 2. About 69 bacterial species and 19 fungal species were detected by the cloning method. Only three species (Bacillus licheniformis, Saccharomycoposis fibuligera and one uncultured bacterium) were detected in all types of Daqu. Twenty bacterial species and 11 fungal species were found in high-temperature Daqu, but only five species, i.e. B. licheniformis, Enterobacter sp., Pichia kudriavzevii, S. fibuligera, and Thermomyces lanuginosus were found in all high-temperature Daqu samples. Forty-three bacterial species and 10 fungal species were found in medium-temperature Daqu, but only four species, i.e. B. licheniformis, Bacillus sp., S. fibuligera, and Lichtheimia ramosa were found in all medium-temperature Daqu. Twenty-nine bacterial species and four fungal species were found in low-temperature Daqu, with B. licheniformis, S. fibuligera, Lichtheimia corymbifera, and Pichia kudriavzevii as common species occurring in all low-temperature Daqu. The lowest bacterial diversity (H′ = 1.19) was found in sample 9-H–S–W, and the highest bacterial diversity (H′ = 3.40) was found in sample 4-M–S–W. For fungi, the lowest diversity (H′ = 0.49) was in sample 8-M–N–BP, and the highest (H′ = 1.59) in sample 5-H–S–W (Table 2).

Table 2 Microbial composition of three types of Daqu
Full size table

Characteristics of different types of Daqu

Group-wise PCA comparisons of the bacterial and fungal composition of the different types of Daqu were constructed (Figs. 1, 2). The loading plots indicate the species that are responsible for the separation of the clusters (Figs. 1b, 2b).

Fig. 1
figure 1

PCA of Daqu extracts on bacterial composition a Score plots of three temperature types of Daqu b Loading plots of PC1 and PC2. Traingle low-temperature Daqu; circle medium-temperature Daqu; square high-temperature Daqu

Full size image
Fig. 2
figure 2

PCA of Daqu extracts on fungal composition a Score plots of three temperature types of Daqu b Loading plots of PC1 and PC2. Traingle low-temperature Daqu circle medium-temperature Daqu; square high-temperature Daqu

Full size image

Based on bacteria detected, five of the eight samples (8-L–N–BP, 4-M–S–W, 5-H–S–W, 5-M–S–W and 9-H–S–W) clustered together (cluster 1). Furthermore, samples 7-L–N–BP and 7-L–N–BP′ clustered together (cluster 2) and sample 8-M–N-BP was separated from all other samples (Fig. 1a). With one exception (i.e., sample 8-L–N–BP), all Daqu samples in the main cluster 1 are from southern China, while the Daqu samples 7-L–N–BP, 7-L–N–BP′ and 8-M–N–BP are from northern China. The loading plot (Fig. 1b) indicates the bacterial species that contributed to this discrimination. The microbial species that most significantly characterized different types by their increased relative abundance were in cluster 1 are: Saccharypolyspora rosea, Streptomyces albus, Thermomonospora chromogena, Staphylococcus gallinarum, Staphylococus sp., Staphylococcus saprophyticus, Bacillus sp., Ent. cowanii, and E. hermannii. In cluster 2, W. confusa showed a marked increase in relative abundance and in cluster 3 the species Thermoactinomyces sanguinis, Saccharopolyspora sp., Saccharopolyspora rectivirgula were detected with an increased abundance in the microbial population.

The PCA of three temperature types of Daqu based on fungal composition (see Fig. 2), showed three groups (Fig. 2a). The species that most significantly characterized the different clusters by their fungal composition were in cluster 1 Rhizomucor pusillus, Absidia idahoenis and L. corymbifera; in cluster 2 S. fibuligera and L. ramosa and in cluster 3 Thermomyces lanuginosus and A. flavus (Fig. 2b).

Discussion

Knowledge of the microbiota of Daqu is still far from complete. Therefore, this study was initiated to understand the composition of the microbial community in three temperature types of this saccharafication agent. Daqu is made from different ingredients (barley, pea, or wheat), and is produced in different locations in China, each location applying different fermentation conditions. The most variable parameter is the maximum temperature of fermentation. Three arbitrary classes can be distinguished with regard to the latter parameter: (1) high-temperature-, (2) medium-temperature-, and (3) low-temperature processes for Daqu production. It is expected that the relative abundance of several of the identified microorganisms correlates with specific environmental conditions. For instance, the prevailing temperature is expected to have a major selective effect on the microbiota. The presence or relative abundance of other microorganisms could be associated with available substrates for fermentation, the location of production facility, and unique factory conditions.

The lowest bacterial diversity as measured by the Shannon index (H′ = 1.19) was found in a high-temperature Daqu (9-H–S–W). Temperatures higher than 65 °C occurred during the production of high-temperature Daqu, and such temperatures only permit the survival and growth of thermophilic or thermotolerant bacteria and fungi, such as Bacillus spp. and Thermomyces spp., respectively (Moretti et al. 2012). This explanation is in line with our observations of the high abundance of B. licheniformis, and Thermomyces lanuginosus in high-temperature Daqu (samples 9-H–S–W and 5-H–S–W). Samples 5-H–S–W and 5-M–S–W were made from the same raw materials (wheat) and produced in the same factory; they only varied in their fermentation temperatures (about 10 °C differences). The comparison of the microbial diversity between samples 5-H–S-W and 5-M–S–W revealed a reduction in fungal diversity upon elevation of the fermentation temperature: sample 5-H–S–W had a lower number of species and a lower value of Shannon’s diversity index compared to sample 5-M–S–W. Two samples obtained from factory 8 (i.e., 8-L–N–BP and 8-M–N–BP) were also produced in the same factory and they revealed the same trend: the higher the temperature, the lower the diversity in fungal composition (Table 2). However, the bacterial composition revealed a opposite trend. This indicates that the bacterial composition, in comparison to the fungal composition, is affected more by other factors such as moisture content and oxygen condition. The production technique used could be another factor affecting the bacterial community in Daqu. One specific technique called “back-slopping” was used in factory 9. The Daqu (4–8 %) that was produced 1 year ago (named “mother Daqu”) was added to the raw materials, and the mixture was used to carry out the Daqu fermentation. On the one hand, the mild acidity of the “mother Daqu” could inhibit the growth of fungi (Li 2013), on the other hand however, the dominant microorganisms in the “mother Daqu” could dominate the Daqu fermentation, thereby suppressing the less prevalent microorganisms. This presumably explains why only seven bacterial species were detected in the sample 9-H–S–W.

Bacillus licheniformis, S. fibuligera and one uncultured bacterium were detected in all tested Daqu samples. This result is in agreement with the study of Wang et al. (2011b). B. licheniformis is a ubiquitous spore-forming bacterium associated with a variety of fermented food products (Lima et al. 2012; Ramos et al. 2010; Wang et al. 2011a.), and it is a well-known producer of proteases and amylases (Karataş et al. 2013). The high relative abundance of B. licheniformis in Daqu suggests that it plays an important role in flavor formation in products such as Chinese liquor and vinegar by hydrolysis of complex carbohydrates and proteins during fermentation. B. licheniformis was found to produce more than 70 metabolites, most of which are flavor compounds and flavor precursors important for the aroma of fermented products (Yan et al. 2013). Yan et al. (2013) reported high levels of acetic acid and lactic acid produced by B. licheniformis. These organic acids may give rise to a variety of aroma compounds by esterification with ethanol. This corresponds well with the fact that the key aroma compounds in light-flavor liquor, such as Fen liquor (factory 7), are mainly ethyl acetate and ethyl lactate. An abundance of B. licheniformis, B. subtilis, and non-specified Bacillus sp. was found in high-temperature Daqu, as has been observed elsewhere (Huang et al. 2006; Yan et al. 2007). Wu et al. (2009) analyzed the metabolite composition in different types of Daqu, and found higher concentrations of amino acids, such as isoleucine and leucine, in high-temperature Daqu. This correlates with the abundance of Bacillus spp. in high-temperature Daqu, since Bacillus spp. were shown to be important thermophilic protease producers (Zhang et al. 2007).

Saccharomycoposis fibuligera was encountered in different types of Daqu (Wang et al. 2011b). The role of S. fibuligera in Daqu production may be the secretion of amylases, acid proteases, and β-glucosidases, which have high potential application in the fermentation industry (Chi et al. 2009). S. fibuligera also has been reported to degrade and assimilate raw starch as a carbon source (Chi et al. 2009); thus, it may contribute to the formation of fermentable carbohydrates for subsequent alcoholic fermentation. In the current study, various genera of lactic acid bacteria (LAB) were identified in Daqu samples, including Enterococcus, Lactobacillus, Leuconostoc, Pediococcus, Streptococcus and Weisella. In general, LAB were found in low abundance, except W. confusa which was found at high abundance in two Daqu samples (i.e., 7-L–N–BP and 7-L–N–BP’). Based on this, the high abundance of W. confusa can be used potentially to distinguish Daqu from factory 7 from Daqu samples originating from other production locations. Among the LAB species, L. fermentum, L. citreum, P. acidilactici, W. confusa, and W. cibaria were reported in earlier studies on other types of Daqu (Wang et al. 2011a; Zheng et al. 2012). Several studies mentioned the importance of LAB during the production of Daqu, but it was only found at high abundance during the beginning of the Daqu production process (Lei 2011). The study of Katina et al. (2002) indicated that some species of Lactobacillus inhibit the growth of Bacillus spp., especially B. subtilis and B. licheniformis. This might explain that a high abundance of LAB was present at early stages of Daqu production. However, the increase in temperature throughout the fermentation process results in the fast growth of thermophilic bacteria such as Bacillus spp. in Daqu and slower growth of mesophilic LAB (Lei 2011). This might explain the low abundance of LAB in the final Daqu products.

Thirteen species of actinomycetes were detected in Daqu, i.e., Actinopolyspora salina, Saccharopolyspora hordei, Saccharopolyspora rectivirgula, Saccharopolyspora rosea, Saccharopolyspora spinosa, Saccharopolyspora sp., Streptomyces cacaoi, Streptomyces albus, Streptomyces sp., T. sanguinis, T. chromogena, Thermoactinomyces bacterium and Thermobispora bispora (Table 2). Of these, Saccharopolyspora rectivirgula, Saccharopolyspora sp. and Streptomyces albus were detected particularly in 8-M–N–BP and 8-L–N–BP, which both originate from the same factory (i.e., factory 8), but were processed at different fermentation temperatures. Since Saccharopolyspora rectivirgula, Saccharopolyspora sp. and Streptomyces albus were not present in other Daqu samples, these three actinomycetes may represent the “house microbiota” of the factory 8. Wang et al. (2012a) monitored the presence of actinomycetes during the production of liquor and observed that >80 % of all the identified actinomycetes (especially Streptomyces spp.) originated from the air in the production room. This result was in line with our hypothesis that Saccharopolyspora rectivirgula, Saccharopolyspora sp., and Streptomyces albus belong to the “house microbiota”. To date, no studies have been published on the role of actinomycetes in the production of Daqu, even though they commonly occur in Daqu. However, other studies have reported the ability of Thermoactinomycetes sp. and Streptomyces spp. to secrete alkaline phosphatase, esterase, lipid esterase, and phosphate hydrolase (Wang et al. 2012a; Yang et al. 2012), which might play important roles in the formation of the flavor compounds or flavor precursors during Daqu fermentation processes.

Analysis of different samples of southern Daqu revealed that Staphylococcus spp., especially Sta. gallinarum and Sta. saprophyticus could be considered as biomarkers of southern Daqu (Gao et al. 2010; Wang et al. 2012b), since these two bacteria were absent in the northern Daqu samples (8-M–N–BP, 8-L–N–BP, 7-L–N–BP and 7-L–N–BP’). This indicates that the bacterial community of Daqu is highly dependent on locations. In addition, the selection of raw materials and the environmental conditions (soil, air, etc.) could also influence the bacterial community in Daqu (Gao 2010; Xu et al. 2004). Klebsiella was found to be heavily associated with the soil used for planting wheat. Only one Daqu sample (4-M–S–W) contained Klebsiella (including K. pneumonia, K. variicola, and Klebsiella sp.), and its presence probably indicates soil contamination. Another study reports that bacteria belonging to the genera Bacillus and Micrococcus were the only dominant bacterial species in wheat (Xu et al. 2004). The high number of Bacillus sp. in the Daqu samples, 5-H–S–W and 5-M–S–W, may be attributed to the dominance of Bacillus in the wheat samples originating from factory 5.

Temperature is an important environmental parameter that affects the growth and survival of microorganisms and, consequently, largely contributes to the microbial community structure of Daqu (Wang et al. 2011b). In general, yeasts and molds are more sensitive to heat than bacteria (Wang et al. 2011a). PCA confirmed grouping of the composition of the fungal communities of Daqu according to the fermentation temperature (Fig. 2). Thermomyces lanuginosus, a thermophilic fungus that survives at temperatures higher than 60 °C (Singh et al. 2003), is a candidate biomarker for high-temperature Daqu. Tm. lanuginosus has been reported to be an efficient xylanase producer, and the xylanase from this fungus is active over a wide pH range (Singh et al. 2003). This might imply that Tm. lanuginosus also plays a role in degrading xylan, as reported previously (Archana and Satyanarayana 1997). S. fibuligera occurred in all tested types of Daqu and represented about 50 % of the total fungal community in low/medium-temperature Daqu. This observation suggests an important role for this species in Daqu. L. ramosa is known to occur on wheat (Shang et al. 2012). Liu et al. (2010) compared the microbial diversity on wheat and its derived Daqu, and observed that L. ramosa and Rhizomucor pusillus occurred in both Daqu and wheat. In our study, Rhizomucor pusillus was present in relatively high numbers in the Daqu sample 4-M–S–W, and this species probably originated from the wheat used in factory 4 (Xu et al. 2004). A comparison of dominant microorganisms in different wheat varieties (Xu et al. 2004) revealed two dominating fungal genera (Rhizopus and Aspergillus) in wheat. Therefore, the Aspergillus flavus and Aspergillus fumigatus species detected in the southern Daqu samples (9-H–S–W and 5-M–S–W) could possibly be associated with the wheat used. All these findings indicate that the fungal communities in Daqu also depend on the raw materials formulation used during production.

Sample 9-H–S–W showed a relatively high abundance of Enterobacteriaceae (Ent. cowanii and E. hermannii) and this may indicate problems with hygienic processing in factory 9. Also, 9-H–S–W was the only sample with Aspergillus flavus (but at low abundance). Although we did not study the effect of A. flavus on the quality and safety of Daqu, the fact that this species is potentially able to produce aflatoxins, indicates that factory 9 requires a more strict quality control than other factories during the whole Daqu production process. Fortunately, this fungus was not observed in other types of Daqu, and therefore we do not regard this as a potential safety risk during Daqu production in general.

Until now, little attempt has been made to compare the microbial community structures of different types of Daqu. We have demonstrated that the fungal diversity in Daqu is highly influenced by fermentation temperature and raw materials, and that the bacterial diversity is influenced by fermentation temperature and geographic environment (i.e. climate, water, and air). The microbial communities of different types of Daqu samples differed significantly from each other. However, the relative abundances of species belonging to the genus Bacillus was higher than that of species of other bacterial genera. Among the Bacillus species, B. licheniformis was predominant and found in all Daqu samples, consistent with previous studies (Yiao et al. 2005; Zheng et al. 2011). On the other hand, each type of Daqu contained a high proportion of sample-specific bacteria. These bacteria and fungi are regarded as candidate biomarkers to distinguish different types of Daqu.

Differences in abundance of specific microorganisms present in Daqu samples as a function of regional origin potentially facilitate the selection of starters for creation of unique, region-specific flavors. Further research is required to establish the impact of Daqu composition on other quality aspects such as health effect of its derived product. This work may help liquor and vinegar industries to understand the microbial ecology of Daqu, and this enables further optimization of using different types of Daqu for liquor and vinegar production.