Skip to main content
NCBI home page
BMC Microbiology logoLink to BMC Microbiology
. 2025 Jul 2;25:374. doi: 10.1186/s12866-025-04053-w

Metagenomic insights into microbial community succession and its functional changes during the stage of acetic acid fermentation of shanxi aged vinegar

Lingling Yang 1,2,#, Yufeng Yan 1,#, Jin Shen 1, Yaoyao Xia 1, Fanfan Lang 1, Cong Chen 2, Wei Zou 1,2,✉
PMCID: PMC12220366  PMID: 40597620

Abstract

Traditional fermentation of Shanxi aged vinegar involves complex microbial interactions driving flavor synthesis, but the mechanisms underpinning metabolic adaptation and community succession remain poorly characterized. This study aimed to unravel stage-specific microbial dynamics and their functional contributions to flavor formation during Cupei fermentation. Metagenomic sequencing analyzed microbial communities and metabolic pathways at three fermentation stages (D3, D6, D9). Functional annotation (KEGG, CAZy) and species-level contribution assessments identified key taxa and genes linked to flavor biosynthesis. Microbial succession shifted from Lactobacillus dominance (64.68% at D3) to Acetobacter prevalence (48.04% at D9), with Lactobacillus acetotolerans persisting throughout (17.15–26.23%). Early-stage carbohydrate metabolism (GHs-driven: 60.38% at D3) transitioned to late-stage amino acid (15.62%) and cofactor synthesis (12.17%), activating valine, leucine, and histidine pathways critical for flavor compounds. Acetobacter oryzoeni and Acetobacter pomorum drove acetate (ALDH: 27.07–41.52%), valine (ilvE: 53.21–20.22%), and histidine (hisD: 41.83–33.30%) metabolism at D9. Low abundance species (Weissella confusa, 0.51%) and uncultured Limosilactobacillus sp. contributed via multi-gene networks (e.g., dat, ldh), which revealed an important functional contribution by overlooked low-abundance species. The study uncovers ecological coupling between microbial succession and metabolic adaptation, where dominant taxa and rare species synergistically govern flavor formation. These insights enable targeted microbial community design for flavor optimization in traditional fermented foods.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-025-04053-w.

Keywords: Shanxi aged vinegar, Metagenomics, Metabolic function dynamics, Functional redundancy, Uncultured microorganisms

Introduction

The flavor formation of traditional fermented foods is closely related to the dynamic succession of microbial communities and their metabolic functions. This process involves a complex “species-function-environment” interaction network [1]. In recent years, the rapid development of metagenomics technology has provided a new perspective for deciphering the composition and functions of microbial communities in fermentation systems [2], and it has increasingly been used to study the microorganisms of fermented foods [3, 4]. In the study of vinegar, metagenomics is also widely applied. Sichuan sun vinegar, Zhanjiang aged vinegar, and other grain vinegars have further explored their microbial communities using this method [5–7].

Chinese vinegar brewing has a history spanning more than 3,000 years [8], with distinct brewing techniques and characteristics varying across different regions. Among the most renowned vinegars from Northern China is Shanxi aged vinegar (SAV), one of the four famous vinegars in the country [9]. SAV is a type of cereal-based vinegar, typically made from sorghum, wheat bran, and rice hulls through solid-state fermentation [10]. The fermentation process consists of three main steps: starch saccharification (SS), alcoholic fermentation (AF), and acetic acid fermentation (AAF) [11]. SS is accomplished using daqu, a ferment rich in microorganisms and enzymes. This step often overlaps with AF, as steamed sorghum is mixed with daqu powder and water [10]. Once AF is complete, AAF begins, using Cupei, a solid mixture of materials including alcoholic residues from AF, wheat bran, rice hulls, and pei (vinegar seeds) from previous batches [12]. The active participation of diverse microorganisms during AAF plays a crucial role in shaping the final flavor, quality, and functional characteristics of SAV [13]. This stage is widely regarded as the most critical step in the vinegar fermentation process [14]. However, research on this stage mostly focuses on the exploration of microbial community structures and flavor compounds [6, 15]. There is a lack of systematic discussions on the functional analysis of microorganisms, such as the metabolic contributions of microorganisms at different fermentation stages. Moreover, the ecological mechanisms of microbial functional redundancy and metabolic complementarity remain unclear.

Based on this, this study took the Cupei of Shanxi aged vinegar as the research object, by analyzing the results of metagenomic sequencing and conducting functional gene annotation and species contribution analysis to reveal the ecological mechanism of the dynamic coupling between microbial community succession and metabolic functions, the species sources of key flavor synthesis genes and their cross—stage collaboration patterns, as well as the functional redundancy strategies of low—abundance microbiota and uncultured microorganisms. The research results provide a theoretical basis for the process optimization of traditional fermented foods and advance the understanding of the “function-species decoupling” phenomenon in microbial ecology.

Materials and methods

Sample collection

The samples were collected from Shanxi Zilin Vinegar Industry Co., Ltd., located in Qingxu County, Shanxi Province, China (112.35°E, 37.6°N), which has a temperate continental climate. After 3 days of fermentation, Cupei samples (300 g) were obtained from three AAF tanks (100 g were taken from each AAF tank) that had recently undergone a Cupei-turnover process (Fig. S1). In each tank, four sampling points were selected 30 cm below the Cupei surface, spaced 5 m apart. The samples from the three tanks were pooled into a single 300 g sample, transported to the laboratory in sterile, self-sealing bags, and stored at − 80℃ for metagenome sequencing. The same sampling method was used for Cupei samples fermented for 6 days and 9 days. A total of three groups of Cupei samples were collected: group D3 (3 days), group D6 (6 days), and group D9 (9 days). For metagenomic sequencing, each group had three parallels, resulting in nine samples.

Metagenome sequencing

The extracted microbial DNA was processed to construct metagenome shotgun sequencing libraries with insert sizes of ~ 400 bp by using the Illumina TruSeq Nano DNA LT Library Preparation Kit (Illumina, USA). Each library was sequenced by Illumina NovaSeq platform (Illumina, USA) with PE150 strategy at Personal Biotechnology Co., Ltd. (Shanghai, China).

Metagenomic data analysis

Raw data from sequencing is screened and filtered by fastp 0.23.2 [16] to remove sequences less than 50 bp in length and those containing ambiguous bases. Kaiju 1.9.0 [17] was used for species annotation of the reads; each read was annotated by comparing it against the GTDB database [18]. Reads were corrected using BBCMS, then assembly was performed using MEGAHIT [19]. Kyoto Encyclopedia of Genes and Genomes (KEGG) [20] was used for gene functional annotation. Flavor-related genes were identified through a two-step approach: (1) manual curation of known flavorant biosynthesis pathways;(2) the KEGG annotation results at the KO (KEGG Orthology) level were compared with genes involved in flavor compound biosynthesis pathways to identify flavor metabolism-related genes in Cupei microbiota. The CAZy (Carbohydrate-Active Enzymes) database was used for the annotation of carbohydrate-active enzyme genes [21].

Statistical analysis and visualization

Differential analysis of abundance and functional pathways across fermentation stages was performed using the Python LEfSe package (LDA scores > 3.0, p < 0.05), which was then visualized using the R language. After obtaining the annotated genes in the KEGG database, the metabolic pathways of flavorants were mapped using Adobe Illustrator (https://www.adobe.com/products/illustrator.html). Species-function analysis was achieved using HUMAnN2 [22], and finally the results were imported into Cytoscape 3.9.1 (https://cytoscape.org/) for visualization of gene-species contribution.

Results

Sequence quality control

Metagenomic sequencing of Cupei samples fermented for 3, 6, and 9 days was performed on the Illumina NovaSeq platform to characterize microbial diversity. Raw and valid data can be seen in Table 1. The average quality distribution of sequencing data is evaluated, and the overall distribution of sequence sequencing quality is captured by counting the average quality distribution of all sequences. The sequencing quality was mainly above Q20 (bases with less than 1% sequencing error rate) (Fig S2), which provided a guarantee for the subsequent data analysis.

Table 1.

Summary of metagenomic quality control data

Samples Number of reads Number of bases N (%) GC (%) Q20(%) Q30(%) Valid sequences Count (%) Valid base sequences Count (%)
D3-1 72,100,092 10,887,113,892 0.0042 40.73 99.01 96.99 71,021,862 98.5 10,718,478,525 98.45
D3-2 72,519,856 10,950,498,256 0.0043 40.38 98.97 96.85 71,438,118 98.51 10,783,681,437 98.48
D3-3 75,445,618 11,392,288,318 0.0040 40.45 99.08 97.14 74,453,100 98.68 11,238,224,794 98.65
D6-1 67,485,844 10,190,362,444 0.0037 43.49 98.86 96.60 66,226,088 98.13 9,993,207,353 98.07
D6-2 73,979,414 11,170,891,514 0.0042 43.99 98.95 96.87 72,709,784 98.28 10,973,195,229 98.23
D6-3 89,999,720 13,589,957,720 0.0037 42.98 98.72 96.23 87,214,484 96.91 13,101,508,553 96.41
D9-1 73,302,282 11,068,644,582 0.0041 43.33 98.93 96.76 72,120,366 98.39 10,885,065,224 98.34
D9-2 75,901,120 11,461,069,120 0.0042 45.42 98.87 96.64 74,469,102 98.11 11,238,562,853 98.06
D9-3 73,910,370 11,160,465,870 0.0042 42.98 99.04 97.10 72,861,548 98.58 10,995,444,107 98.52
Open in a new tab

Microbial community analysis of Cupei with different fermentation days

We compared the metagenomic sequences with the GTDB database and found that bacteria accounted for more than 98% of the total microorganisms in samples D3, D6, and D9 (Fig. S3). On the ninth day of Cupei fermentation, this percentage reached 99.65%. At the phylum level (Fig. 1a), Firmicutes and Proteobacteria were the dominant phyla in the microbial community across all fermentation stages; however, their proportions varied significantly. The relative abundance of Firmicutes decreased from 82.08% at D3 to 35.49% at D9, while Proteobacteria showed an increasing trend, rising from 15.13% to 62.14%. The relative abundance of other annotated phyla, such as Actinobacteriota and Ascomycota, was low (< 1%) and showed no significant trends. At the genus level, Lactobacillus, Acetobacter, Limosilactobacillus, and Komagataeibacter were prominent in samples D3, D6, and D9 (Fig. 1b). Specifically, Lactobacillus decreased 64.68% at D3 to 18.56% at D9, Limosilactobacillus had its lowest relative abundance at D6 (15.32% to 9.85% to 10.48%), Acetobacter increased significantly, from 10.15% at D3 to 48.04% at D9, and Komagataeibacter increased from 0.82% to 5.34%. Other genera, including Agrobacterium (0.78% to 2.11%), Microbacterium (0.76% to 1.87%), and Acetilactobacillus increased in relative abundance with Cupei fermentation. Notably, Acetilactobacillus increased 79.31-fold from 0.09% at D3 to 7.26% at D9. At the species level, 2,094 species were annotated (Table S1). Figure 1c highlights the top 20 species with the highest relative abundance. Lactobacillus acetotolerans remained the most abundant species across all samples, despite a decrease in relative abundance (26.23% to 17.15%). Several species, including Acetobacter pomorum (1.46% to 15.26%), Acetobacter oryzoeni (3.22% to 9.49%), and Acetilactobacillus jinshanensis (0.13% to 9.61%) showed a significant increase in relative abundance. Conversely, some species showed a clear decline: Lactobacillus xujianguonis (7.67% to 0.59%), Lactobacillus helveticus (6.81% to 0.25%), Lactobacillus amylovorus (5.47% to 0.55%), and Lactobacillus kefiranofaciens (5.72% to 0.58%).

Fig. 1.

Fig. 1

Open in a new tab

Evolution of microbial community structure during Cupei fermentation. a Microbial composition at the phylum level (p: phylum); b microbial composition at the genus level (g: genus); c microbial composition at the species level (s: species)

Functional profile of microbial communities

Analysis of the metagenomic data using the KEGG database annotated most of the genes to metabolism (72.75% to 76.26%), including carbohydrate metabolism (19.06% to 16.36%), amino acid metabolism (9.21% to 15.62%), metabolism of cofactors and vitamins (9.40% to 12.17%), glycan biosynthesis and metabolism (5.98% to 5.02%), energy metabolism (5.61% to 4.92% to 5.97%), biosynthesis of other secondary metabolites (6.88% to 2.93%), lipid metabolism (4.68% to 3.49%), nucleotide metabolism (3.02% to 4.67%), metabolism of other amino acids (2.79% to 4.93%) and “other” metabolisms, suggesting that the metabolic potential within the Cupei microbiome is high and diverse. The metabolic pathways were annotated and a total of 97 metabolic pathways were identified. Among these, pathways such as Glycolysis/Gluconeogenesis (ko00010, 3.24% to 2.46% to 2.49%), Pentose phosphate pathway (ko00030, 1.67% to 1.40% to 1.47%), Alanine, aspartate and glutamate metabolism (ko00250, 1.45% to 2.03%), Valine, leucine and isoleucine biosynthesis (ko00290, 0.75% to 2.24%), Starch and sucrose metabolism (ko00500, 3.19% to 1.48%), and Pyruvate metabolism (ko00620, 1.72% to 1.84%) exhibited relatively high abundances across D3, D6, and D9, despite variations in their relative proportions at each time point (Fig. 2a). This indicates that these metabolic pathways are central to the biological processes being studied and that their relative abundances provide insights into the metabolic dynamics and adaptive strategies of the organisms. To effectively highlight the significant differences in various metabolic pathways across different time points, the LEfSe (Linear Discriminant Analysis Effect Size) analysis results were utilized with a threshold of 3 (Fig. 2b). At D3, pathways such as starch and sucrose metabolism (LDA = 3.94) and galactose metabolism (LDA = 3.73) had higher LDA scores, indicating prominent activity at this time point. During the D6 stage, the fatty acid biosynthesis pathway was relatively significant (LDA = 3.83). By D9, multiple amino acid metabolism pathways, including valine, leucine, and isoleucine biosynthesis (LDA = 3.80), arginine biosynthesis (LDA = 3.45), lysine biosynthesis (LDA = 3.45), as well as histidine metabolism (LDA = 3.42) had higher LDA scores, showing an active state. These suggest that the activity of different metabolic pathways in the samples changed dynamically over time, reflecting shifts in metabolic demands and patterns at various stages.

Fig. 2.

Fig. 2

Open in a new tab

Dynamic changes of microbial functions during Cupei fermentation process. a Changes in abundance of metabolism-related pathways (ko00020: Citrate cycle; ko00520: Amino sugar and nucleotide sugar metabolism; ko00552: Teichoic acid biosynthesis; ko00450: Selenocompound metabolism; ko00030: Pentose phosphate pathway; ko00290: Valine, leucine and isoleucine biosynthesis; ko00770: Pantothenate and CoA biosynthesis; ko00983: Drug metabolism—other enzymes; ko00300: Lysine biosynthesis; ko00670: One carbon pool by folate; ko00480: Glutathione metabolism; ko00620: Pyruvate metabolism; ko00550: Peptidoglycan biosynthesis; ko00250: Alanine, aspartate and glutamate metabolism; ko00730: Thiamine metabolism; ko00500: Starch and sucrose metabolism; ko00240: Pyrimidine metabolism; ko00521: Streptomycin biosynthesis; ko00010: Glycolysis/Gluconeogenesis; ko00710: Carbon fixation in photosynthetic organisms); b Linear discriminant analysis effect size analysis of microbial metabolic pathways during Cupei fermentation, with longer bar lengths indicating more significant differences for this taxonomic unit; c Changes in abundance of carbohydrate-active enzymes (GH: Glycoside Hydrolases; GT: Glycosyl Transferases; CBM: Carbohydrate-Binding Modules; CE: Carbohydrate Esterases; PL: Polysaccharide Lyases; AA: Auxiliary Activities)

While annotating through the CAZy database, we found the most genes associated with glycoside hydrolases (GHs) and glycosyl transferases (GTs). The relative abundance of GHs decreased (60.38% to 49.23%) and that of GTs increased (27.15% to 37.82%) during the fermentation of Cupei (Fig. 2c). In the GH family, GH36, GH1, and GH4 had relatively high abundances (Table S2), but their relative abundances all decreased over time. In the GT family, GT2 and GT4 had the highest relative abundances, and their relative abundances increased as the fermentation of Cupei progressed. The relative abundance of carbohydrate-binding modules (CBMs) showed a slight increase from D3 to D6 (7.66% to 7.82%), followed by a decrease at D9 (6.68%). Among the annotated enzymes in the CBM family, the relative abundance of CBM50 was significantly higher than that of other enzymes. Carbohydrate esterases (CEs) showed relatively stable abundance across the time points, with minor fluctuations (3.83% to 4.92%). Among the members of the CE family, CE4 has the largest proportion. The overall abundance of auxiliary activities (AAs) and polysaccharide lyases (PLs) was very small, with the average relative abundance of GHs being 89.16 and 95.42 times higher than them, respectively. Overall, GHs and GTs were abundant, while AAs and PLs were less abundant.

Metabolism related to primary flavor compounds

Previous studies have indicated that acetate and lactate are the primary acids in SAV, and alanine, glutamate, valine, and histidine make significant contributions to the flavor of SAV [23]. These five substances are primarily produced through seven pathways: Glycolysis/Gluconeogenesis (ko00010), Pentose phosphate pathway (ko00030), Starch and sucrose metabolism (ko00500), Pyruvate metabolism (ko00620), Alanine, aspartate and glutamate metabolism (ko00250), Valine, leucine and isoleucine biosynthesis (ko00290), and Histidine metabolism (ko00340, 0.20% to 0.63%). Through gene annotation, we have identified the metabolic pathways for the formation of these five flavor substances during the fermentation process of Cupei (Fig. 3a).

Fig. 3.

Fig. 3

Open in a new tab

Metabolic pathways of flavor compounds and the associated genes and enzymes. a Metabolic pathways of major flavor substances (AscF: PTS system, beta-glucoside (arbutin/salicin/cellobiose)-specific IIC component; Crr: PTS system, sugar-specific IIA component; PtsG: PTS system, glucose-specific IIC component; ScrA: PTS system, sucrose-specific IIC component). b Hetmap of the relative abundance of key flavor—related enzymes

A variety of sugars, including arbutin, salicin, glucose, and sucrose, are transported into the cell via different transport proteins (e.g., AscF, PtsG, ScrA, etc.) and are phosphorylated to form the corresponding sugar-6-phosphate. Among these, the number of mapped reads for Crr (D3: 9; D6: 30; D9: 14), AscF (D3: 1; D6: 17; D9: 9), and PtsG (D3: 3; D6: 8; D9: 4) remained relatively low. In contrast, ScrA exhibited the highest mapped reads across all stages (D3: 898; D6: 979; D9: 300), which partially reflects the high sucrose utilization efficiency of the Cupei microbial community. These phosphorylated sugars enter the glycolysis pathway through enzymatic reactions mediated by glucose-6-phosphate isomerase, fructokinase, and other key enzymes. Pyruvate, a key intermediate product generated by glycolysis, can be converted into lactate under the action of L-lactate dehydrogenase (EC:1.1.1.27, peak in D3, Glycolysis/Gluconeogenesis, Fig. 3b), imparting a unique sour flavor to fermented foods. Additionally, pyruvate can be catalyzed by pyruvate dehydrogenase E1 component (EC:1.2.4.1, peak in D6, Pyruvate metabolism) and pyruvate decarboxylase (EC:4.1.1.1, peak in D3, Pyruvate metabolism) to form acetaldehyde, which is then converted into acetate through aldehyde dehydrogenase (NAD +) (EC:1.2.1.3, Pyruvate metabolism, peak in D6), contributing a distinctive flavor to the product. Valine can be synthesized from pyruvate through a series of enzymatic reactions involving dihydroxy-acid dehydratase (EC:4.2.1.9, peak in D9, Valine, leucine and isoleucine biosynthesis), branched-chain amino acid aminotransferase (EC:2.6.1.42, peak in D6, Valine, leucine and isoleucine biosynthesis), and other enzymes, while alanine can be synthesized through the action of alanine transaminase (EC:2.6.1.2, peak in D3, Alanine, aspartate and glutamate metabolism). Glutamate can be synthesized from TCA cycle intermediates such as oxaloacetate and 2-oxoglutarate by the action of enzymes like glutamate synthase (NADPH/NADH) large chain (EC:1.4.1.13 1.4.1.14, peak in D9, Alanine, aspartate and glutamate metabolism) and glutamate synthase (NADPH/NADH) small chain (EC:1.4.1.13 1.4.1.14, peak in D9, Alanine, aspartate and glutamate metabolism). Furthermore, histidine can be synthesized starting from ribose 5-phosphate generated by the pentose phosphate pathway through reactions involving several enzymes such as phosphoribosyl-ATP pyrophosphohydrolase (EC:3.6.1.31, peak in D9, Histidine metabolism), ATP phosphoribosyltransferase (EC:2.4.2.17, peak in D9, Histidine metabolism), and histidinol dehydrogenase (EC:1.1.1.23, peak in D9, Histidine metabolism). The elucidation of the metabolic pathways of these key flavor compounds and amino acids reveals how microorganisms utilize substrates such as sugars during metabolism to generate substances that are essential to the flavor and quality of the product.

Dynamics of species contribution to key flavor synthesis genes

During the fermentation of Shanxi aged vinegar, the microbial communities and their contributions to key flavor synthesis genes exhibited significant dynamic variations across different stages (Fig. 4). For the ALDH gene (acetic acid synthesis), A. oryzoeni (24.42%) and an uncultured yeast (Saccharomyces sp., 19.37%) synergistically drove ethanol oxidation at the early stage (D3). By the mid-stage (D6), the contribution of A. oryzoeni surged to 50.05%, dominating the process exclusively. In the late stage (D9), A. pomorum (41.52%) became the primary contributor to ALDH, with A. oryzoeni (27.07%) and Agrobacterium cavarae (15.17%) collectively accounting for 56.69% of the total contribution, indicating a collaborative role between Acetobacter and Agrobacterium in late-stage acetic acid synthesis. For the ldh gene (lactate synthesis), the microbial community composition displayed strong taxon-specific dependency over fermentation stages. At D3, uncultured Limosilactobacillus sp905214235 (29.13%) and Limosilactobacillus panis (32.00%) co-dominated lactate synthesis, with supplementary contributions from other Limosilactobacillus members (e.g., L. fermentum, 0.88%; L. pontis, 2.46%), forming an early-stage collaborative network. By D6, L. panis (21.74%) and Limosilactobacillus sp905214235 (13.35%) remained core contributors, while transient increases in contributions from L. fermentum (12.12%) and L. pontis (5.46%) were observed. At D9, L. panis (23.32%) and Limosilactobacillus sp905214235 (20.65%) retained the highest contributions (43.97% combined), whereas other species, such as L. pontis (9.81%) and Limosilactobacillus reuteri (3.22%), showed reduced roles.

Fig. 4.

Fig. 4

Open in a new tab

Species contribution network of flavor-related genes, with thicker lines representing higher contribution values. Red ellipses represent species, and blue ellipses represent metabolic genes. The contribution level of species to genes is positively correlated with the thickness of the connecting lines, where thicker lines indicate higher contribution levels. Numerical labels (> 10%) on the lines denote significant contribution relationships

The ilvE gene (valine biosynthesis) exhibited distinct stage-specific succession patterns. At D3, an uncultured Lactobacillus sp. (48.60%) and Lactobacillus ultunensis (16.54%) dominated branched-chain amino acid metabolism, reflecting the central role of lactic acid bacteria in early nitrogen metabolism. By D6, A. oryzoeni (53.56%) replaced lactic acid bacteria as the dominant contributor, with the overall contribution of Lactobacillus spp. declining to 5.00%. At D9, A. oryzoeni (53.21%) maintained dominance, while A. pomorum (20.22%) significantly increased its contribution, collectively accounting for 73.43% of valine synthesis. Minor contributions from L. fermentum (3.43%), Streptococcus macedonicus (2.94%), and Microbacterium testaceum (2.00%) suggested auxiliary roles in supporting metabolic flux during later stages. For the hisD gene (histidine metabolism), species diversity decreased progressively. At D3, A. oryzoeni (26.08%), A. cavarae (17.03%), and uncultured Limosilactobacillus spp. (16.01%) jointly contributed to histidine synthesis. By D9, A. oryzoeni (41.83%) and A. pomorum (33.30%) dominated (75.13% combined), while contributions from L. fermentum (declining from 12.00% to 1.05%) and Liquorilactobacillus sp. (16.02% to 1.84%) diminished markedly, reflecting streamlined nitrogen metabolism in late fermentation. The gltB and gltD genes (glutamate synthesis) revealed persistent involvement of uncultured microorganisms. For gltB, A. cavarae (23.12%) and A. pomorum (14.03%) were primary contributors at D3, while uncultured Microbacterium sp. (5.37%) and Acetobacter sp. (6.35%) collectively accounted for 11.72%. By D9, contributions from uncultured microorganisms expanded, including Xanthomonas A sp. (4.78%), Acetobacter sp. (6.35%), and Microbacterium sp. (5.37%), totaling 16.5%, underscoring their sustained role in glutamate synthesis. For gltD, A. pomorum remained dominant (67.50% at D3 to 69.50% at D9), while Komagataeibacter oboediens emerged as a key secondary contributor (9.15% at D9). Contributions from uncultured Liquorilactobacillus sp. (4.97% at D3) and M. testaceum (3.98% at D3) declined over time. The dat gene (alanine synthesis) transitioned from single-species dominance to multi-taxon collaboration. At D3, A. cavarae exclusively contributed (100%). By D6, W. confusa (31.13%) and uncultured Kroppenstedtia sp. (13.53%) joined A. cavarae (55.34%) in a collaborative network. At D9, A. cavarae (56.32%), W. confusa (20.64%), and Kroppenstedtia sp. (23.05%) collectively accounted for 99.01%, demonstrating adaptive diversification of substrate utilization strategies to enhance environmental resilience.

Discussion

Changes in the microbial community

The vinegar fermentation system exhibited significant microbial community succession characteristics. The relative abundance of Firmicutes sharply decreased from 82.08% at D3 to 35.49% at D9, while Proteobacteria surged from 15.13% to 62.14%. This phylum-level trend aligns with previous studies on Shanxi aged vinegar [12] and mirrors microbial succession patterns observed in other vinegar fermentation processes [24, 25]. Genus-level dynamics closely mirrored phylum-level restructuring: acid-producing bacteria represented by Lactobacillus dominated the early fermentation stage (D3, 64.68%), but were later supplanted by acid- and ethanol-tolerant strains such as Acetobacter (48.04%) and Acetilactobacillus (7.26%) at D9. This succession pattern reflects carbon metabolism coupled with environmental adaptation strategies in fermented foods [26, 27]. At the species level, the succession displayed three distinct patterns: (1) Early-stage specialists: including L. xujianguonis (7.67% to 0.59%), L. helveticus (6.81% to 0.25%) and L. amylovorus (5.47% to 0.55%), which thrived in initial high-sugar conditions but rapidly declined as acidity increased; (2) Persistent generalists: represented by L. acetotolerans that maintained stable abundance (26.23% to 17.15%) throughout fermentation, likely due to its exceptional acid tolerance and versatile metabolic capabilities [28, 29]; (3) Late-stage specialists: notably A. pomorum (1.46% to 15.26%) and A. jinshanensis (0.13% to 9.61%) that dominated the final stages, correlating with their strong acetate resistance [30].

Functional changes of the microbial community

KEGG and CAZy analyses revealed stage-specific metabolic restructuring. Early fermentation (D3) prioritized carbohydrate metabolism (19.06%), with elevated starch/galactose metabolism (LDA = 3.94/3.73) aligning with Lactobacillus's preferential utilization of soluble sugars [31]. At the mid-fermentation stage (D6), fatty acid biosynthesis emerged as a prominent pathway (LDA = 3.83), likely driven by microbial adaptation to environmental stressors such as pH decline [32]. Fatty acids are essential for maintaining membrane fluidity and integrity under acidic conditions [33], and their synthesis may also serve as a precursor for ester formation, which contributes to the fruity aroma in vinegar [34]. As carbon depletion and acidification progressed, metabolic emphasis shifted to nitrogen metabolism. By D9, amino acid (15.62%) and cofactor metabolism (12.17%) intensified, peaking in branched-chain amino acids (valine, leucine, isoleucine; LDA = 3.80) and histidine biosynthesis (LDA = 3.90), reflecting microbial nitrogen compensation under carbon limitation [35] to enhance valine and histidine synthesis—key contributors to vinegar flavor [23]. Dynamic glycoside hydrolase (GHs decreased by 11.15%) and glycosyltransferase (GTs up 10.67%) profiles confirmed adaptive metabolic shifts. Early GH dominance (60.38%) facilitated polysaccharide decomposition [36], while later GT enrichment (37.82%) likely supported oligosaccharide reconstruction or flavor precursor biosynthesis [37], marking a transition from substrate breakdown to product synthesis. This shift aligns with carbon source evolution (complex polysaccharides to monosaccharides/organic acids) and redox balance requirements.

Analysis of the dynamic contributions of species to flavor-related genes

Flavor-related gene contribution dynamics unveiled core species'functional diversity and ecological adaptation, revealing a sophisticated temporal partitioning of metabolic roles among microbial taxa throughout fermentation stages. Acetobacter members played pivotal roles: A. oryzoeni demonstrated high contributions to acetate (ALDH), valine (ilvE), and histidine (hisD) synthesis (24.42%−53.56% from D3-D9), likely due to efficient ethanol oxidation and nitrogen assimilation [38], while A. pomorum dominated late-stage acetate (41.52%) and glutamate (gltD, 69.50%) synthesis, showcasing cross-pathway adaptability. Environmental selection pressures drove community restructuring [39], with functional redundancy maintaining metabolic stability under industrial fermentation stress [40]. ALDH contributors shifted from early A. oryzoeni-yeast synergy (D3: 43.79%) to late A. pomorum-Agrobacterium cavarae collaboration (D9: 56.69%), reflecting ethanol-driven redundancy to stabilize acetate production. For lactate synthesis (ldh), Limosilactobacillus species (L. panis and uncultured Limosilactobacillus sp905214235) maintained persistent contributions despite proportional fluctuations, confirming their acidogenic persistence [41]. ilvE contributors transitioned from lactobacilli to Acetobacter dominance, correlating with nitrogen utilization efficiency. A. oryzoeni (relative abundance: 3.22% → 9.27%) emerged as the dominant ilvE contributor via enhanced transaminase activity [38]. hisD diversity decreased progressively, with A. oryzoeni and A. pomorum dominating late-stage histidine synthesis (75.13%), attributable to their environmental resilience [42]. gltB analysis revealed sustained involvement of uncultured microbes, suggesting unexplored"metabolic dark matter"potentially supporting core functions through alternative pathways or novel enzymatic activities [43]. Notably, L. acetotolerans maintained the highest relative abundance across all annotated species, yet exhibited limited contributions to key flavor synthesis genes, primarily influencing ldh (declining from 3.49% to 0.89%). However, it actively participated in substrate degradation [44]. In contrast, low-abundance species demonstrated significant functional roles. For instance, W. confusa (0.51% at D9) contributed to alanine, lactate, and acetate synthesis through coordinated dat (20.64%), ldh (1.03%), and ALDH (1.34%) gene activities, reinforcing Weissella's importance in food fermentation [45] Additionally, uncultured Saccharomyces sp. contributed 19.37% to ALDH in early fermentation, while Lactobacillus sp. provided 48.60% to ilvE, highlighting unexplored functional microbial resources in traditional fermentation systems.

While this study provides insights into microbial functional dynamics through metagenomic profiling, its limitations include the lack of direct flavor compound quantification to validate gene-metabolite linkages and reliance on database-dependent functional predictions. Notably, its strengths lie in deciphering niche differentiation mechanisms and identifying low-abundance functional coordinators. The findings demonstrate non-linear abundance-function relationships, offering actionable strategies for industrial optimization—such as engineering synthetic consortia. Future studies should prioritize: (1) Targeted metabolomics (GC–MS/MS) to map microbial gene activity to flavor compound dynamics, (2) Culturomic validation of uncultured taxa’s metabolic roles, and (3) Mechanistic modeling of functional redundancy thresholds in synthetic communities. This integrated approach will bridge the gap between ecological theory and precision fermentation applications.

Conclusion

The dynamic succession of microbial communities and their metabolic functions during Shanxi aged vinegar fermentation revealed intricate function-environment-species interactions. While Lactobacillus dominated early stages with L. acetotolerans maintaining high abundance throughout the fermentation process, its direct flavor contributions were limited, contrasting sharply with Acetobacter’s disproportionate role in synthesizing key flavor compounds through cross-stage gene collaboration. This highlights niche differentiation driven by metabolic plasticity rather than abundance alone. Metabolic strategies transitioned from polysaccharide degradation to nitrogen-dependent refinement, reflecting adaptive responses to substrate depletion. Low-abundance taxa and uncultured microorganisms further disrupted abundance-function linearity via multi-gene coordination, suggesting unresolved modular metabolic division critical for community robustness. Future research should integrate multi-omics data (e.g., targeted metabolomics, culturomics) to validate the metagenomics—based conclusions of this study, and explore the feasibility of simulating natural fermentation via synthetic microbial consortia. This work offering actionable insights for optimizing microbial community design in fermented foods.

Supplementary Information

Supplementary Material 1. (2.4MB, zip)

Acknowledgements

Not applicable.

Authors’ contributions

Lingling Yang: Formal analysis, Investigation, visualization, writing—original draft, Writing—review & editing. Yufeng Yan: Formal analysis, Resources, Validation, Writing—review & editing. Jin Shen: Data curation, Conceptualization. Yaoyao Xia: visualization. Fanfan Lang: Conceptualization. Cong Chen: Methodology. Wei Zou: Methodology, Resources, Supervision, Writing—review & editing.

Funding

This study was supported by the Open Project Program of Shanxi Provincial Key Laboratory for Vinegar Fermentation Science and Engineering (20220401931002), Key Research and Development Program Projects in Shanxi Province (202202140601018; 202102130501008), and Sichuan Natural Science Foundation General Project (2023 NSFSC0184).

Data availability

Sequence data that support the findings of this study have been deposited in the National Center for Biotechnology Information database with the accession code SRP584453.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Lingling Yang and Yufeng Yan contributed equally to this work.

References

  • 1.Wu W, Wang Z, Xu B, Cai J, Cheng J, Mu D, et al. Exploring core microbiota based on characteristic flavor compounds in different fermentation phases of Sufu. Molecules. 2022;27(15):4933. 10.3390/molecules27154933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Alves LDF, Westmann CA, Lovate GL, de Siqueira GMV, Borelli TC, Guazzaroni M-E. Metagenomic approaches for understanding new concepts in microbial science. Int J Gen. 2018;2018(1):2312987. 10.1155/2018/2312987. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Chen C, Yang H, Zhang K, Ye G, Luo H, Zou W. Revealing microbiota characteristics and predicting flavor-producing sub-communities in Nongxiangxing baijiu pit mud through metagenomic analysis and metabolic modeling. Food Res Int. 2024;188:114507.https://doi.org/10.1016/j.foodres.2024.114507. [DOI] [PubMed] [Google Scholar]
  • 4.Lee SH, Whon TW, Roh SW, Jeon CO. Unraveling microbial fermentation features in kimchi: from classical to meta-omics approaches. Appl Microbiol Biotechnol. 2020;104(18):7731–44. 10.1007/s00253-020-10804-8. [DOI] [PubMed] [Google Scholar]
  • 5.Li N, Fu J, Zhang G, Liu J, Li Z, Luo R, et al. Investigating the mechanism of the flavor formation in Sichuan sun vinegar based on flavor-orientation and metagenomics. Curr Res Food Sci. 2023;6:100460. 10.1016/j.crfs.2023.100460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Wu L, Lu Z, Zhang X, Wang Z, Yu Y, Shi J, et al. Metagenomics reveals flavour metabolic network of cereal vinegar microbiota. Food Microbiol. 2017;62:23–31. 10.1016/j.fm.2016.09.010. [DOI] [PubMed] [Google Scholar]
  • 7.Yu Z, Ma Y, Guan Y, Zhu Y, Wang K, Wang Y, et al. Metagenomics of virus diversities in solid-state brewing process of traditional Chinese vinegar. Foods. 2022;11(20):3296.https://doi.org/10.3390/foods11203296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Zhang X, Wang P, Xu D, Wang W, Zhao Y. Corrigendum to aroma patterns of Beijing rice vinegar and their potential biomarker for traditional Chinese cereal vinegars. Food Res Int. 2020;119:398–410. 10.1016/j.foodres.2019.108830. [DOI] [PubMed] [Google Scholar]
  • 9.Al-Dalali S, Zheng F, Xu B, Abughoush M, Li L, Sun B. Processing technologies and flavor analysis of Chinese cereal vinegar: a comprehensive review. Food Analytical Methods. 2022;16(1):1–28. 10.1007/s12161-022-02328-w. [Google Scholar]
  • 10.Nie Z, Zheng Y, Du H, Xie S, Wang M. Dynamics and diversity of microbial community succession in traditional fermentation of Shanxi aged vinegar. Food Microbiology. 2015;47:62–8. 10.1016/j.fm.2014.11.006. [DOI] [PubMed] [Google Scholar]
  • 11.Lu ZM, Wang ZM, Zhang XJ, Mao J, Shi JS, Xu ZH. Microbial ecology of cereal vinegar fermentation: insights for driving the ecosystem function. Curr Opinion Biotechnol. 2018;49:88–93. 10.1016/j.copbio.2017.07.006. [DOI] [PubMed] [Google Scholar]
  • 12.Zhu Y, Zhang F, Zhang C, Yang L, Fan G, Xu Y, et al. Dynamic microbial succession of Shanxi aged vinegar and its correlation with flavor metabolites during different stages of acetic acid fermentation. Sci Rep. 2018;8(1):8612. 10.1038/s41598-018-26787-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wang ZM, Lu ZM, Shi JS, Xu ZH. Exploring flavour-producing core microbiota in multispecies solid-state fermentation of traditional Chinese vinegar. Scientific Reports. 2016;6(1):26818. 10.1038/srep26818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Du P, Li Y, Zhen C, Song J, Hou J, Gou J, et al. Effect of microbial reinforcement on polyphenols in the acetic acid fermentation of Shanxi Aged Vinegar. Fermentation. 2023;9(8):756. 10.3390/fermentation9080756. [Google Scholar]
  • 15.Han D, Yang Y, Guo Z, Chen K, Dai S, Zhu Y, et al. Metagenomics profiling of the microbial community and functional differences in solid-state fermentation vinegar starter (seed Pei) from different Chinese regions. Front Microbiol. 2024;15:1389737. 10.3389/fmicb.2024.1389737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–90. 10.1093/bioinformatics/bty560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Menzel P, Ng KL, Krogh A. Fast and sensitive taxonomic classification for metagenomics with Kaiju. Nat Commun. 2016;7(1):11257. 10.1038/ncomms11257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Parks DH, Chuvochina M, Rinke C, Mussig AJ, Chaumeil PA, Hugenholtz P. GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy. Nucleic Acids Res. 2022;50(D1), D785-D794. 10.1093/nar/gkab776 [DOI] [PMC free article] [PubMed]
  • 19.Li D, Liu CM, Luo R, Sadakane K, Lam TW. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics. 2015;31(10):1674–6. 10.1093/bioinformatics/btv033. [DOI] [PubMed] [Google Scholar]
  • 20.Ogata H, Goto S, Sato K, Fujibuchi W, Bono H, Kanehisa M. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 1999;27(1):29–34. 10.1093/nar/27.1.29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Lombard V, Golaconda Ramulu H, Drula E, Coutinho PM, Henrissat B. The carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res. 2014;42(D1):D490–5. 10.1093/nar/gkt1178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Franzosa EA, McIver LJ, Rahnavard G, Thompson LR, Schirmer M, Weingart G, et al. Species-level functional profiling of metagenomes and metatranscriptomes. Nature Methods. 2018;15(11):962–8. 10.1038/s41592-018-0176-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zheng Y, Zhao C, Wu Y, Liu J, Cheng C, Shi L, et al. Composition of characteristic flavor components and analysis of flavor wheel for shanxi aged vinegar. Journal of Food Science and Technology. 2019;37(04):24–34. 10.3969/j.issn.2095-6002.2019.04.004. [Google Scholar]
  • 24.Li X, Rao W, Hu S, Zhu S, Ouyang L, Zhou J. Study on the relationship between microbial community succession and physicochemical factors in the fermentation of rice-flavor Baijiu based on high-throughput and redundancy analysis techniques. Lwt. 2024;213:117031. 10.1016/j.lwt.2024.117031. [Google Scholar]
  • 25.Luo A, Cheng Z, Zhao J, Hao J, Shi S, Hu B. The Relationship between microbial community succession and flavor formation during the natural fermentation of Hongqu sufu. Foods. 2023;12(14):2800. 10.3390/foods12142800. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Li L, Li N, Fu J, Liu J, Ping Wen X, Cao H, et al. Synthesis of an autochthonous microbial community by analyzing the core microorganisms responsible for the critical flavor of bran vinegar. Food Res Int. 2024;175:113742.https://doi.org/10.1016/j.foodres.2023.113742. [DOI] [PubMed] [Google Scholar]
  • 27.Liu M, Lin X, Yang J, Yan X, Ji C, Liang H, et al. Effects of salt and rice flour concentration on microbial diversity and the quality of sour meat, a Chinese traditional meat. Food Sci Human Wellness. 2024;13(5):2790–8. 10.26599/fshw.2022.9250226. [Google Scholar]
  • 28.Yang X, Teng K, Zhang J, Wang F, Zhang T, Ai G, et al. Transcriptome responses of Lactobacillus acetotolerans F28 to a short and long term ethanol stress. Scientific Reports. 2017;7(1):2650. 10.1038/s41598-017-02975-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Zhao H, Feng J, Wang X, Zhou H, Luo X, Qiu S, et al. Screening, identification and characterization of acid tolerant Lactobacillus acetotolerans FBKL1.0204. Food Ferment Indust. 2022;48(19):50–7. 10.13995/j.cnki.11-1802/ts.029928. [Google Scholar]
  • 30.Qiu X, Zhang Y, Hong H. Classification of acetic acid bacteria and their acid resistant mechanism. AMB Express. 2021;11(1):29. 10.1186/s13568-021-01189-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Srinivas D, Mital BK, Garg SK. Utilization of sugars by Lactobacillus acidophilus strains. Int J Food Microbiol. 1990;10(1):51–7. 10.1016/0168-1605(90)90007-r. [DOI] [PubMed] [Google Scholar]
  • 32.Li Y, Yan P, Lei Q, Li B, Sun Y, Li S, et al. Metabolic adaptability shifts of cell membrane fatty acids of Komagataeibacter hansenii HDM1–3 improve acid stress resistance and survival in acidic environments. J Industrial Microbiol Biotechnol. 2019;46(11):1491–503. 10.1007/s10295-019-02225-y. [DOI] [PubMed] [Google Scholar]
  • 33.Maulucci G, Cohen O, Daniel B, Sansone A, Petropoulou PI, Filou S, et al. Fatty acid-related modulations of membrane fluidity in cells: detection and implications. Free Rad Res. 2016;50(sup1):S40–50. 10.1080/10715762.2016.1231403. [DOI] [PubMed] [Google Scholar]
  • 34.Liu SQ, Holland R, Crow VL. Esters and their biosynthesis in fermented dairy products: a review. Int Dairy J. 2004;14(11):923–45. 10.1016/j.idairyj.2004.02.010. [Google Scholar]
  • 35.Lu, P., Yan, H., Ding, A., Li, C., and Quan, L. (2023). Metabolic regulation of bacteria with limited carbon and nitrogen sources. 63(03), 946–962. 10.13343/j.cnki.wsxb.20220549.
  • 36.Baldrian, P., Berlemont, R., & Martiny, A. C. (2016). Glycoside hydrolases across environmental microbial communities. PLOS Computational Biology, 12(12), e1005300.10.1371/journal.pcbi.1005300 [DOI] [PMC free article] [PubMed]
  • 37.Liang DM, Liu JH, Wu H, Wang BB, Zhu HJ, Qiao JJ. Glycosyltransferases: mechanisms and applications in natural product development. Chem Soc Rev. 2015;44(22):8350–74. 10.1039/c5cs00600g. [DOI] [PubMed] [Google Scholar]
  • 38.Baek JH, Kim KH, Moon JY, Yeo SH, Jeon CO. Acetobacter oryzoeni sp. nov., isolated from Korean rice wine vinegar. Int J Syst Evol Microbiol. 2020;70(3):2026–33. 10.1099/ijsem.0.004008. [DOI] [PubMed] [Google Scholar]
  • 39.Louca S, Polz MF, Mazel F, Albright MBN, Huber JA, O’Connor MI, et al. (2018). Function and functional redundancy in microbial systems. Nat Ecol Evol. 2(6), 936–943.10.1038/s41559-018-0519-1 [DOI] [PubMed]
  • 40.Condic N, Amiji H, Patel D, Shropshire WC, Lermi NO, Sabha Y, et al. Selection for robust metabolism in domesticated yeasts is driven by adaptation to Hsp90 stress. Science. 2024;385(6707):eadi3048. 10.1126/science.adi3048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Gänzle MG. Lactic metabolism revisited: metabolism of lactic acid bacteria in food fermentations and food spoilage. Curr Opinion Food Sci. 2015;2:106–17. 10.1016/j.cofs.2015.03.001. [Google Scholar]
  • 42.Sokollek SJ, Hertel C, Hammes WP. Description of Acetobacter oboediens sp. nov. and Acetobacter pomorum sp nov., two new species isolated from industrial vinegar fermentations. International Journal of Systematic Bacteriology. 1998;48(3):935–40. 10.1099/00207713-48-3-935. [DOI] [PubMed] [Google Scholar]
  • 43.Lewis K, Epstein S, D’Onofrio A, Ling LL. Uncultured microorganisms as a source of secondary metabolites. J Antibiotics. 2010;63(8):468–76. 10.1038/ja.2010.87. [DOI] [PubMed] [Google Scholar]
  • 44.Liu J, Li L, Peters BM, Li B, Deng Y, Xu Z, et al. Draft genome sequence and annotation ofLactobacillus acetotoleransBM-LA14527, a beer-spoilage bacteria. FEMS Microbiol Lett. 2016;363(18):fnw201. 10.1093/femsle/fnw201. [DOI] [PubMed] [Google Scholar]
  • 45.Mandhania MH, Paul D, Suryavanshi MV, Sharma L, Chowdhury S, Diwanay SS, et al. Diversity and succession of microbiota during fermentation of the traditional Indian food Idli. Appl Environ Microbiol. 2019;85(13):e00368-19. 10.1128/aem.00368-19. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1. (2.4MB, zip)

Data Availability Statement

Sequence data that support the findings of this study have been deposited in the National Center for Biotechnology Information database with the accession code SRP584453.

Articles from BMC Microbiology are provided here courtesy of BMC

RESOURCES