Lactiplantibacillus plantarum-inoculated silage improves milk fat in dairy goats by reprogramming rumen microbiota to promote pyruvate metabolism

Jing Ma; Qiang Li; Guanghao Xia; Jiayao Zhang; Samaila Usman; Xusheng Guo

doi:10.1038/s41522-026-01031-6

. 2026 Jun 6;12:170. doi: 10.1038/s41522-026-01031-6

Lactiplantibacillus plantarum-inoculated silage improves milk fat in dairy goats by reprogramming rumen microbiota to promote pyruvate metabolism

Jing Ma ^1,^2,^#, Qiang Li ^1,^2,^#, Guanghao Xia ^1,², Jiayao Zhang ^1,², Samaila Usman ^1,², Xusheng Guo ^1,^2,^✉

PMCID: PMC13562649 PMID: 42248895

Abstract

Silage inoculated with homofermentative lactic acid bacteria exhibits superior nutritional quality and improves ruminant production performance. However, the mechanisms by which the inoculated silage modulates microbial and metabolic alterations along the rumen-mammary gland axis remain unclear. Here, Lactiplantibacillus plantarum BX62 inoculation enhanced silage quality and flavonoid content through reshaping the microbiome, which promotes carbohydrate-active enzyme secretion. Feeding the silage to dairy goats modulated ruminal microbiota through changes in silage flavonoid, fiber, and crude protein contents. Notably, increased flavonoid intake enriched ruminal flavonoid-degrading bacteria, leading to enhanced flavonoid transformation, fiber degradation, and pyruvate-centered carbohydrate metabolism, promoting acetate production. The increased ruminal acetate and upregulated expression of mammary lipogenic genes ultimately improved milk fat synthesis. Our findings reveal how L. plantarum links silage quality to rumen function and mammary gland metabolism, offering a mechanistic basis for improving lactation performance of dairy goats via modulating silage fermentation.

Subject terms: Biotechnology, Microbiology

Introduction

Ensiling represents a cornerstone technology in sustainable livestock production, which preserves the nutritional quality of fresh forages via anaerobic microbial fermentation primarily driven by lactic acid bacteria (LAB)¹. By mitigating seasonal fluctuations in feed supply, ensiled forages constitute a major dietary component for ruminants, typically contributing 50−80% of the ration for dairy cows during winter². With more than 133 million dairy cows globally, annual silage consumption exceeds 665 million metric tons³, underscoring its pivotal role in supporting animal productivity and the economic and environmental sustainability of intensive farming systems.

Alfalfa is widely recognized as the “King of Forages” due to its high protein, mineral, and vitamin content⁴. However, it poses challenges for ensiling successfully, owing to its high buffering capacity and low water-soluble carbohydrate (WSC) content⁴. These characteristics result in a slow pH decline during ensiling, which favors the proliferation of undesirable and spoilage microorganisms such as clostridia, enterobacteria, yeasts, and molds, ultimately impairing silage fermentation and nutritional quality. Inoculation with LAB, especially homofermentative strains such as Lactiplantibacillus plantarum, has emerged as an effective and practical strategy to improve alfalfa silage quality. These inoculants rapidly ferment available WSC into lactic acid, accelerating acidification and suppressing the growth of harmful microorganisms, thereby preserving nutrients and ensuring stable fermentation^1,5.

Beyond improving silage fermentation quality, mounting evidence indicates that homofermentative LAB inoculants also shape the rumen microecosystem and enhance ruminant production performance. Firstly, inoculation may enhance silage degradability and nutrient utilization by ruminal microbes⁶, or modulate the rumen microbiome through bioactive compounds produced during ensiling⁷. In vitro fermentation studies have confirmed that L. plantarum-inoculated silage reshapes rumen microbial composition, enriching fiber-degrading and acid-producing bacteria (e.g., Butyrivibrio fibrisolvens, Ruminococcus flavefaciens, Tepidimicrobium, Ruminiclostridium, Rikenellaceae RC9 gut group, and Treponema 2) while suppressing methanogens and eukaryotic microorganisms^8,9. Subsequently, these microbial shifts are accompanied by reduced methane emissions and elevated volatile fatty acid (VFA) concentrations. This indicates enhanced metabolic energy availability for animal production, which may further influence milk yield and quality. A meta-analysis of 43 studies on LAB-inoculated silage (with L. plantarum dominating the majority) reported a significant increase in milk yield and tendencies for improved milk fat and protein content in dairy cows¹⁰. In dairy goats, alfalfa silage inoculated with L. plantarum significantly increased the milk fat, protein, and total solids^6,11.

Although the beneficial effects of LAB-inoculated silage on milk yield and quality are well established, the integrated mechanisms underlying multi-level regulation along the silage-rumen-mammary gland axis remain largely unexplored. To address this critical knowledge gap, we conducted a feeding trial with dairy goats fed silage inoculated with L. plantarum BX62. We employed an integrated multi-omics strategy to systematically elucidate how this inoculant modulates microbial community shifts, associated nutrient flow, and metabolic crosstalk along the rumen-mammary gland axis. This strategy included metagenomic and untargeted metabolomic analyses of silage, as well as metagenomic, untargeted metabolomic, and targeted metabolomic analyses of rumen contents. Our findings establish a comprehensive mechanistic framework for elucidating the central roles of silage and rumen microbiomes in ruminant nutrition, and provide a theoretical basis for developing a precise strategy of feeding silage to enhance productivity and sustainability in dairy animal production systems.

Results

Fermentation characteristics and chemical composition of alfalfa silage

Following inoculation with L. plantarum BX62, the pH of alfalfa silage decreased significantly (P < 0.001, Table 1). Butyrate was not detected in either group of silage. Compared with control silage, BX62 silage exhibited a higher concentration of acetate (P < 0.001) and a lower concentration of propionate (P < 0.05). The dry matter (DM) content did not differ between groups (P > 0.05). However, fiber fractions (aNDF and ADF) were reduced in BX62 silage (P < 0.05). This resulted in the production of more WSC (P < 0.05). In addition, compared to the silage in the control group, the CP content was higher in the BX62 group (P < 0.05).

Table 1.

Fermentation characteristics and chemical composition of ensiled alfalfa

Items^a	Treatment^b		P-value
Items^a	Control	BX62	P-value
pH	4.90 ± 0.01	4.72 ± 0.01	<0.001
Organic acid profile (g/kg, DM)
LA	38.19 ± 0.167	38.49 ± 0.164	>0.05
AA	21.71 ± 0.063	26.16 ± 0.226	<0.001
PA	6.12 ± 0.139	5.12 ± 0.150	0.003
BA	NA	NA	-
Chemical components
DM (g/kg, FW)	345.80 ± 14.99	339.62 ± 5.42	>0.05
aNDF (g/kg, DM)	380.74 ± 3.812	366.27 ± 2.211	0.017
ADF (g/kg, DM)	278.31 ± 3.176	264.25 ± 0.827	0.005
ADL (g/kg, DM)	7.91 ± 1.555	8.46 ± 0.440	>0.05
CP (g/kg, DM)	177.24 ± 0.248	181.61 ± 1.39	0.021
NH₃-H (g/kg TN)	24.35 ± 0.751	23.64 ± 0.659	>0.05
WSC (g/kg, DM)	8.98 ± 0.441	12.44 ± 0.646	0.004

Open in a new tab

^aLA Lactate, AA Acetate, PA Propionate, BA Butyrate, DM Dry matter, FW Fresh weight, WSC Water-soluble carbohydrates, CP Crude protein, NH₃-N Ammonia nitrogen, aNDF Neutral detergent fiber treated with a heat-stable amylase, ADF Acid detergent fiber, ADL Acid detergent fiber lignin.

^bControl alfalfa silage inoculated with no strains; BX62 alfalfa silage inoculated with L. plantarum BX62.

Data are presented with mean ± standard error of the mean (SEM) from four independent experiments. Statistical significance was determined using Student’s t test.

Metagenomic and metabolomic changes during silage fermentation

From metagenomic perspective, Shannon index did not differ between two silage groups (P > 0.05, Fig. 1a) whereas β-diversity revealed distinct community structures (Fig. 1b). Taxonomic profiling showed that control silage was dominated by Weissella cibaria, while BX62 silage was enriched in L. plantarum and Lentilactobacillus buchneri (Fig. 1c). LEfSe analysis further identified differential bacterial taxa, with higher abundances of L. buchneri and L. plantarum, but lower abundances of several species of Weissella, Pediococcus, Lactobacillus, Levilactobacillus, Loigolactobacillus, Enterococcus, and Leuconostoc in the BX62 group (Fig. 1d). Functional profiling revealed significant differences in Carbohydrate-active enzymes (CAZyme) families. The abundance of carbohydrate esterase (CE) and glycoside hydrolase (GH) families was increased in the BX62 group, while glycosyl transferase (GT) and polysaccharide lyase (PL) families were increased in the control group (P < 0.05, Supplementary Fig. 1a, b). A larger number of individual subfamilies were upregulated in BX62 silage (P < 0.05; Supplementary Fig. 1c). The KEGG annotation further indicated that most of the metabolism pathways were upregulated in the BX62 group (P < 0.05, Supplementary Fig. 1d).

A total of 1798 metabolites were identified in the silage metabolome. OPLS-DA showed a clear separation between BX62 and control silages (Fig. 1e). In total, 567 metabolites differed significantly, 284 of which were up-regulated in BX62 (Supplementary Fig. 2). These differential metabolites were assigned to seven superclasses, including alkaloids, amino acids/peptides, carbohydrates, fatty acids, polyketides, shikimates/phenylpropanoids, and terpenoids. Notably, metabolites within the shikimates and phenylpropanoids superclass predominantly enriched in BX62 silage (Fig. 1f). Detailed class analysis further indicated that flavonoids, phenolic acid (C6-C1), phenylpropanoids (C6-C3), and isoflavonoids were markedly accumulated in the BX62 silage (Fig. 1g). Functional annotation of KEGG pathways screened 15 pathways with significant association to differential metabolites (Fig. 1h). Particularly, flavonoid and flavonol biosynthesis were dramatically activated, and all metabolites annotated to these pathways exhibited consistent upregulation in BX62 silage (Fig. 1i). Quantitative analysis further confirmed that ensiling increased total flavonoid levels relative to fresh samples, and BX62 inoculation further increased concentration relative to the control (P < 0.05, Fig. 1j).

Silage microbiome explained the mechanism underlying the increase in flavonoids during ensiling. Although the overall functional abundance of flavonoid biosynthesis remained markedly low, distinct variations were observed in the canonical pathways ko00941 (flavonoid biosynthesis) and ko00944 (flavone and flavonol biosynthesis). In addition, the flavonoid degradation pathway (ko00946) was upregulated in the BX62 silage, with α-L-rhamnosidase (K05989)−an enzyme involved in rutin hydrolysis−being the only significantly altered enzyme in the pathway (Supplementary Fig. 3). To unravel the key drivers of flavonoid accumulation, we constructed a correlation network between differential CAZymes and flavonoid metabolites. The network displayed extensive positive correlations (Fig. 2a), suggesting that CAZymes facilitate flavonoid release and biotransformation during ensiling. Notably, most CAZymes involved in these regulatory relationships were taxonomically derived from L. plantarum (Fig. 2b).

Fig. 2 — a Spearman’s correlation analysis between differencial CAZymes with differencial flavonoid compounds in silage. Strong correlations (|R| > 0.8, P < 0.05) were shown. b Taxonomic attribution of differential CAZymes. The figure displays traceable CAZymes, while some individuals lack source-tracking information. Control: alfalfa silage inoculated with no strains; BX62: alfalfa silage inoculated with *L. plantarum* BX62.

Effects of L. plantarum-inoculated silage on rumen fermentation and milk production in dairy goats

An initial in vitro fermentation experiment was conducted using diets supplemented with silages from two groups. The inoculant did not affect pH and total gas production (P > 0.05), but significantly increased in vitro dry matter digestibility (IVDMD) (P < 0.05, Supplementary Table 3). Although total VFA concentrations did not differ between groups (P > 0.05), the BX62 silage increased the molar proportions of propionate and valerate and reduced the acetate-to-propionate ratio (A:P) (P < 0.05). Moreover, CH₄ production and proportion were decreased in the BX62 group (P < 0.05).

Based on these observations, an in vivo feeding trial was performed in dairy goats. Rumen fermentation parameters showed no differences in A:P ratio (P > 0.05, Supplementary Fig. 4). Nevertheless, the total VFA production was higher in goats receiving BX62 silage, accompanied by increased acetate, propionate, and branched-chain fatty acids (P < 0.05, Fig. 3a). In addition, milk fat content was significantly elevated in the BX62 group (P < 0.05, Fig. 3b).

To elucidate the mechanisms underlying enhanced milk fat synthesis, protein expression in mammary tissue was assessed (Fig. 3c). Western blot analysis revealed higher levels of acetyl-CoA carboxylase 1 (ACC1) and fatty acid synthase (FASN), two key enzymes involved in de novo fatty acid synthesis, in the BX62 group (P < 0.05, Fig. 3c). Acyl-CoA synthetase short-chain family member 2 (ACSS2), responsible for converting acetate to acetyl-CoA, was also upregulated (P < 0.05, Fig. 3c), consistent with enhanced ruminal acetate availability in the BX62 group. Furthermore, phosphorylation of ACC1 at Ser-79 (P-ACC1), an inhibitory modification, was reduced in the BX62 group (Fig. 3c), indicating enhanced ACC1 activity. In contrast, β-hydroxybutyrate dehydrogenase 1 (BDH1) expression did not differ between groups (P > 0.05, Fig. 3c), consistent with unaffected ruminal butyrate concentrations (Fig. 3a).

Specific rumen microbial taxa mediated the link between silage and ruminal acetate synthesis

Feeding with BX62 silage led to significant changes in the ruminal microbial community structure. Although no differences in Shannon index were observed (Fig. 4a), β-diversity revealed clear separation of microbial communities in two groups (Fig. 4b). At the kingdom level, bacterial abundance was higher in the BX62 group, while archaea and eukaryotes showed substantial decreases (P < 0.05, Fig. 4c). WGCNA clustered the bacteria genera into eight modules associated with ruminal fermentation parameters and milk quality traits. Among these, the MEyellow and MEbrown modules were positively correlated with key indicators such as ruminal acetate and milk fat (P < 0.05, Fig. 4d). To identify potential drivers of ruminal acetate, we constructed a co-occurrence network based on MEyellow and MEbrown bacterial modules. This analysis pinpointed several hub bacteria (e.g., Butyrivibrio, Pseudobutyrivibrio, Blautia, and Faecalibacterium), which were considered keystone species due to their high connectivity (Fig. 4e). Differential abundance analysis revealed that 18 bacterial genera in the two modules were significantly altered by silage feed, with most being enriched in the BX62 group (Fig. 4f). Notably, these differentially abundant taxa overlapped substantially with the hub bacteria, including Butyrivibrio, Pseudobutyrivibrio, Acetivibrio, Collinsella, Enterocloster, Dorea, Mediterraneibacter, and Coprococcus identified in the MEbrown module, and Paenibacillus, Clostridioides, Blautia, Faecalibacterium, Pantoea, and Candidatus Limivicinus identified in the MEyellow module (Fig. 4f). Correlation analysis revealed that, with the exception of Candidatus Saccharimonas and Fibrobacter, all other differential bacterial genera were significantly positively correlated with ruminal acetate (Fig. 4g).

Fig. 4 — a Differential analysis of alpha diversity (Shannon index) in rumen microbes. b The community dissimilarity analysis of rumen microbes using PCoA. c Differential analysis of rumen microbial community of two groups at the kingdom level. d WGCNA identification of rumen bacteria modules correlated with dairy goat phenotypes. e Correlation network analysis of bacteria in the MEbrown and MEyellow modules. Node size corresponds to the Module Membership value. Hub genera are colored brown and yellow, respectively. Red lines indicate positive correlations, and blue lines indicate negative correlations. f Differential bacterial genera in the MEbrown and MEyellow modules between the two groups. g Spearman correlation analysis of differential bacterial genera with dairy goat phenotypes (rumen VFAs and milk quality) (P < 0.05, |r| > 0.5). *: 0.01 < P ≤ 0.05; **: 0.001 < P ≤ 0.01; ***: P ≤ 0.001. Control: alfalfa silage inoculated with no strains; BX62: alfalfa silage inoculated with *L. plantarum* BX62.

Mantel-test was employed to screen for silage factors that significantly explained the variation in the ruminal fermentation and milk quality. Acetate exhibited the strongest positive correlation with milk fat (Fig. 5a). The silage fermentation quality influenced ruminal acetate production (P < 0.05, Fig. 5a). Although silage total metabolites had no significant impact on ruminal acetate, silage flavonoids displayed a strong correlation with ruminal acetate (P < 0.05, Fig. 5a). Mediation analysis revealed that silage aNDF, ADF, CP, and flavonoids influenced ruminal acetate production by modulating the microbial community in the MEbrown module (Fig. 5b). Additionally, aNDF, ADF, and flavonoids also regulated microorganisms in the MEyellow module (Fig. 5c). Redundancy analysis (RDA) was performed to explore the relationships between silage factors and rumen microbial community composition. Two canonical axes explained 84.32% of the total variation in rumen microbial community, with RDA1 accounting for 71.15% and RDA2 accounting for 13.17% (Fig. 5d). CP and flavonoids displayed longer vectors in the ordination than aNDF and ADF. Specifically, genera such as Butyrivibrio and Pseudobutyrivibrio clustered along the direction of the CP and flavonoids vectors, while Ruminococcus and Fibrobacter were associated with aNDF and ADF. To further evaluate the unique contributions of individual silage parameters to variation of rumen microbial community, variation partitioning analysis (VPA) was performed. The results indicated that flavonoids and CP explained the highest proportions of the microbial community variation, accounting for 53.5% and 49.9% (adjusted R², P < 0.05), respectively. These values were substantially higher than those observed for fiber-related indicators (aNDF: 36.3%, ADF: 39.3%, P < 0.05, Fig. 5e). This finding aligns with the trends observed in the RDA ordination plot (Fig. 5d).

Besides bacterial genera, no significant differences were observed among the top 10 archaeal genera (P > 0.05; Supplementary Fig. 5a). Additionally, all differentially abundant protozoal and fungal taxa within the Top30 eukaryotic genera were significantly depleted in the BX62 group (P < 0.05; Supplementary Fig. 5b).

Flavonoid-degrading bacteria in the rumen responded to higher intake of silage flavonoids

Targeted metabolomics identified differential accumulation of flavonoids and their metabolites in the rumen between groups (Fig. 6a). Seven compounds were significantly enriched in the control group, whereas nine compounds were elevated in the BX62 group (P < 0.05, VIP > 1), indicating different ruminal flavonoid metabolism. Metagenomic analysis revealed that the flavonoid degradation pathway (ko00946) was significantly enriched in the BX62 group (P < 0.05, Fig. 6b). Most enzymes in this pathway exhibited increased abundance, with flavanone/flavanol-cleaving reductase (K26178), alpha-L-rhamnosidase (K05989), puerarin oxidase (K26172), naringenin degradation protein FdeE (K26180), deglycosylation enzyme subunit DgpC (K26174), phloretin hydrolase (K22906), and chalcone isomerase (K26177) significantly increased (P < 0.05, Fig. 6c). The representative strain Clostridioides difficile 630, which harbors flavonoid-degrading enzymes (flavone/flavonol reductase, chalcone isomerase), was subjected to flavonoid degradation assays. In vitro experiments demonstrated that this strain exhibited robust growth across quercetin concentrations ranging from 10 to 200 μM. Furthermore, it was capable of nearly completely depleting 200 μM quercetin within 6 h (Fig. 6d).

To connect these functional patterns to specific microbial populations, we reconstructed 1347 medium- and high-quality rumen MAGs. Twenty-nine MAGs encoded flavonoid-degrading enzymes, including daidzein reductase (K26069), quercetin 2,3-dioxygenase (K07155), flavanone/flavanol-cleaving reductase (K26178), phloretin hydrolase (K22906), and flavone/flavonol reductase (K26176). Flavanone/flavanol-cleaving reductase and flavone/flavonol reductase represented the most abundant enzyme types. These MAGs were distributed across 10 genera. Prevotella represented the most abundant carrier, while Pseudobutyrivibrio exhibited the broadest enzyme repertoire (Supplementary Fig. 6).

Recognizing the limitations of MAG recovery from our dataset, we further curated four large rumen MAG reference databases to identify strains carrying flavonoid-degrading genes. A total of 32,042 rumen MAGs were included, of which 1608 (5.02%) harbored flavonoid-degrading enzymes, spanning 9.45% of species, 20.27% of genera, and 29.95% of families (Fig. 6e). Phylogenetic profiling illustrated the evolutionary relationships of the flavonoid-degrading bacterial MAGs and their family-level classification. Lachnospiraceae and Oscillospiraceae dominated the tree, followed by representatives from Erysipelotrichaceae, Anaerovoracaceae, Ruminococcaceae, Bacteroidaceae, Methanomethylophilaceae, UBA932, Atopobiaceae, Selenomonadaceae, Eggerthellaceae, and Aristaeellaceae (Fig. 6f). A substantial proportion of the Top50 genera containing flavonoid-degrading enzymes still originated from Lachnospiraceae and Oscillospiraceae (Fig. 6g). Notably, among the two groups of dairy goats, 18 differential bacterial genera were identified in the rumen (Fig. 4f). Of these, 12 possessed flavonoid-degrading enzymes and were all enriched in the BX62 group (Fig. 6f). Moreover, among the Top50 bacterial genera harboring flavonoid-degrading enzymes, five also belonged to the differential genera and were enriched in the BX62 group (Fig. 6g). To determine whether flavonoid-degrading bacteria responded to dietary flavonoid exposure, we aligned clean metagenomic reads from both groups to a reference rumen MAG collection. Among differential MAGs carrying flavonoid-degrading enzymes, 37 were upregulated in the BX62 group compared to only 9 in the control group (Fig. 6h). These results suggest that rumen bacteria equipped with flavonoid-degrading enzymes exhibit adaptive growth in response to flavonoid exposure.

The reshaped ruminal microbial community enhanced carbohydrate metabolism and VFA production

Silage quality and flavonoids shaped the ruminal microbiota, with consequent effects on its function (Fig. 7a and Supplementary Fig. 7). The BX62 group exhibited a significant upregulation in most KEGG metabolism pathways, including carbon metabolism, biosynthesis of cofactors, and biosynthesis of amino acids, among others (Fig. 7a). CAZymes profiling further supported these findings. There were 26 enzymes (16 GHs, 6 GTs, 3 CEs, and 1 PL) that were significantly upregulated in the BX62 group among the top 30 differentially abundant CAZymes. Only GH13, GH26, GT92, and GT20 increased significantly in the control group (P < 0.05, Fig. 7b).

Given that milk fat precursors (acetate and butyrate) formation mainly depends on rumen microbial carbohydrate degradation, we analyzed feed transformation processes mediated by differential CAZymes and KEGG pathways. The reshaped ruminal microbial community enhanced carbohydrate metabolism. Specifically, three cellulose-degrading enzymes (GH94, GH116, GH124), three hemicellulose-degrading enzymes (GH26, GH42, CE3), and one starch-degrading enzyme (GH176) were significantly upregulated in the BX62 group, whereas only one starch-degrading enzyme (GH13) was upregulated in control group (Fig. 7c). Accordingly, 15 KEGG pathways involved in carbohydrate degradation were significantly enriched in the BX62 group (Fig. 7c). These upregulated CAZymes and pathways enhanced the pyruvate-centered carbohydrate fermentation capacity of rumen microorganisms, resulting in elevated acetate concentrations in the rumen of the BX62 group (Fig. 7c).

Silage-driven alterations in the rumen metabolome and their association with milk fat

Comparative analysis of rumen metabolomic profiles revealed distinct clustering patterns between the two dairy goat groups (Fig. 8a, b). 601 metabolites enriched in the control group and 424 enriched in the BX62 group (Fig. 8c). Correlation analysis showed that 133 metabolites were positively associated with milk fat content, whereas 42 were negatively associated (Fig. 8d). KEGG enrichment of milk fat-associated metabolites identified 26 significantly enriched pathways (Fig. 8e). We focused on the differential metabolites linked to pyruvate-centered carbohydrate metabolism due to its established links to ruminal VFA production. Integrating differential metabolite abundance with pathway annotations allowed the construction of a metabolite flow network (Fig. 8f). Most elevated metabolites in the BX62 group goats showed positive correlations with milk fat content and converged toward pyruvate biosynthesis, supporting enhanced substrate flow for VFA generation. In contrast, metabolites enriched in the control group exhibited inverse metabolic flux patterns (Fig. 8f).

Discussion

While homofermentative LAB inoculants are known to improve silage quality and ruminant production performance, their regulatory mechanisms along the feed-rumen-host axis remain unclear. Here, we systematically investigated the mechanisms by which inoculation with L. plantarum BX62 enhanced alfalfa silage quality and modulated rumen metabolism to improve milk quality in dairy goats. Multi-omics integration revealed that enhanced nutritional properties and flavonoid levels in inoculated silage collectively reshaped the rumen microbial community. This thus enhanced ruminal pyruvate-centered carbohydrate metabolic flux, stimulating acetate biosynthesis and consequently improving mammary lipogenic efficiency via the rumen-mammary gland axis. This study provides a theoretical basis for efficient ruminant production via inoculant-modulated silage fermentation.

Inoculation with L. plantarum BX62 improved alfalfa silage fermentation quality, with elevated acetate, CP, and WSC, and reduced aNDF and ADF. Untargeted metabolomics and flavonoid quantification further confirmed significant flavonoid enrichment in the BX62 group silage. Metagenomic analysis showed that strain L. plantarum BX62 outperforms native epiphytic flora on alfalfa, and its inoculation increased the abundance of L. buchneri while reducing the abundances of Weissella, Pediococcus, and Enterococcus. This is attributed to the strong adaptability and superior environmental fitness of BX62 to alfalfa raw materials and silage conditions. The rapid proliferation of BX62 during ensiling drove massive production of lactic acid and other organic acids, suppressing acid-intolerant epiphytic microbes^12,13. Notably, the dominant L. plantarum and associated microbes in the BX62 silage expressed significantly higher levels of CAZymes, whereas the native epiphytic microbiome in the control group showed far lower abundance. Rapid acidification and enzymolysis not only accelerate plant fiber breakdown and the release of cell wall-immobilized flavonoids^14,15, but also inhibit acid-intolerant detrimental microbes. This suppression effectively reduces non-target microbial hydrolytic degradation of proteins and flavonoids¹³. This may account for the quality changes and flavonoid enrichment in the BX62 silage.

Based on these silage characteristics, we further conducted an in vivo feeding trial using dairy goats fed either control or L. plantarum BX62-inoculated alfalfa silage. Feeding BX62 silage significantly increased ruminal VFA production and milk fat synthesis. Milk fat synthesis is highly regulated by the rumen-mammary gland axis, where ruminal acetate and butyrate act as major precursors for mammary de novo fatty acid synthesis¹⁶. Acetate provides most carbon sources, and nearly half the NADPH required for lipogenesis via the isocitrate pathway¹⁷, whereas butyrate contributes ~15% of carbon following conversion to β-hydroxybutyrate¹⁸. In our study, the BX62 group exhibited elevated ruminal acetate concentration, along with upregulated mammary expression of lipogenic proteins ACC1, FASN, and ACSS2, and decreased level of phosphorylated ACC1. These findings indicate that BX62 silage promoted milk fat synthesis by altering rumen microbial metabolism, with acetate acting as a key mediator.

Variations in the ruminal VFA profile are derived from shifts in the microbial community¹⁹, which are regulated by dietary components. Our study revealed that shifts in key acid-producing genera (Butyrivibrio, Pseudobutyrivibrio, Collinsella, etc.) were closely correlated with silage aNDF, ADF, CP, and flavonoid contents. Inoculation with LAB during ensiling promotes the degradation of structural carbohydrates^15,20. In this study, L. plantarum BX62 inoculation significantly reduced silage aNDF and ADF contents, which will facilitate substrate utilization by ruminal microorganisms and enhance fermentation efficiency. Lower aNDF and ADF shifted the fibrolytic community: diets with reduced NDF typically decreased Fibrobacter abundance while enriching acetate-producing bacteria such as Coprococcus and Blautia^21,22, consistent with our observations. Increased silage CP also contributed to rumen microbial restructuring. The enhanced abundance of rumen Butyrivibrio and Pseudobutyrivibrio in the BX62 group may be linked to higher CP in the inoculated silage. CP hydrolysis releases small peptides, which have been shown to enrich these genera^23,24. They are known to be heavily involved in the metabolism of polysaccharides and proteins²⁵. Thus, the elevated abundance of these bacteria contributes to improved ruminal carbohydrate and nitrogen metabolism.

In addition to nutritional components, increased silage flavonoids significantly drove rumen microbiota shifts. In this study, the flavonoid degradation pathway was significantly upregulated in the BX62 group, indicating active rumen microbial responses to silage flavonoids. We obtained 1,341 rumen MAGs from metagenomic data, and annotated flavonoid-degrading enzymes, including daidzein reductase, quercetin 2,3-dioxygenase, flavanone/flavanol-cleaving reductase, phloretin hydrolase, and flavone/flavonol reductase, with Pseudobutyrivibrio genomes harboring the highest number of these enzymes. Flavonoid-degrading enzymes in 32,045 MAGs from public rumen databases corroborated our findings, confirming that these enzymes are present in a limited set of rumen microorganisms. Notably, 12 of the 17 significantly upregulated genera in the BX62 group were flavonoid-degrading taxa, suggesting these bacteria gain a competitive advantage under high-flavonoid silage feeding. Interestingly, microorganisms responsive to dietary flavonoids have previously been demonstrated to be associated with organic acid production and carbohydrate degradation, including Butyrivibrio, Pseudobutyrivibrio, Blautia, Coprococcus, and Collinsella^26–29. Beyond these genera, multiple other microbial taxa encoding flavonoid-degrading enzymes were enriched in the BX62 group, likely attributable to their flavonoid metabolic capability. Our in vitro experiments verified that C. difficile 630, harboring flavone/flavonol reductase and chalcone isomerase, can directly metabolize flavonoids such as quercetin, which corroborated the MAG functional annotation. Moreover, several genera, including Butyrivibrio³⁰, Dorea³¹, Paenibacillus³², Blautia³³, Coprococcus³⁴, and Collinsella³⁵, have been reported to exhibit flavonoid-transforming and degrading capacities, in line with our results. The degradation intermediates derived from flavonoid catabolism may modulate the growth and metabolic performance of these key rumen bacteria. In this study, high levels of vanillic acid and hydrocinnamic acid were detected in the rumen. Previous studies have confirmed that Butyrivibrio can modify vanillin and hydrocinnamic acid³⁶. Both compounds are typical phenolic intermediates derived from the microbial breakdown of ruminal flavonoids. Vanillic acid, as the key downstream metabolite of vanillin, is also an important degradation intermediate of flavonoids. Butyrivibrio may detoxify and utilize these phenolic intermediates to support its growth and metabolic activity.

Flavonoids also markedly suppress rumen protozoa and fungi^29,37, lowering interspecies hydrogen transfer with methanogens^38,39 and thus reducing methane production while enhancing energy utilization. The downregulated rumen eukaryotes in the BX62 group further suggested that silage flavonoids may redirect carbon flux toward VFA production by modulating methanogenesis. Although reduced eukaryotes may slightly weaken fiber degradation capacity, the improved bacterial carbohydrate-degrading function in the BX62 group could compensate for this deficit, maintaining stable rumen fiber digestibility in the long term.

Optimized feed components promote microbial pyruvate metabolism. Previous studies confirm that dietary small peptides and easily fermentable, low-fiber substrates upregulate key rumen microbial enzymes (e.g., pyruvate:ferredoxin oxidoreductase and pyruvate phosphate dikinase) and pyruvate-related pathways^24,40,41. Meanwhile, feed flavonoids can be degraded into phenolic acids and non-aromatic fermentation products, which further break down to pyruvate as alternative carbon sources, thereby enhancing their metabolic flux^42,43. In our study, the altered microbiome drove a “pyruvate-centered” metabolism through stronger substrate degradation and carbon flux redistribution. First, the upregulated cellulases, hemicellulases, and amylases in the BX62 group facilitated the breakdown of carbohydrates into pentoses and hexoses. These monosaccharides were then processed through multiple sugar metabolic pathways, including the pentose phosphate pathway (ko00030), fructose and mannose metabolism (ko00051), amino sugar and nucleotide sugar metabolism (ko00520), pentose and glucuronate interconversions (ko00040), and glycolysis/gluconeogenesis (ko00010), ultimately increasing the pool of glyceraldehyde-3P and driving carbon flux toward pyruvate. This redirection was further reinforced via amino acid catabolism (e.g., alanine, serine, cysteine via ko00470, and arginine, proline via ko00330), which expanded the pyruvate hub. Then, the enriched pyruvate in the BX62 group was further metabolized to generate sufficient intermediate substrates for downstream VFA synthesis. Consequently, the increased acetate served as the primary precursor and markedly promoted milk fat synthesis in the mammary gland. Rumen metabolomics further supported this metabolic remodeling: differential metabolites associated with milk fat (e.g., fumarate, O-acetylserine) in the BX62 group showed enrichment patterns linked to pyruvate biosynthesis. Together, these results indicate that nutrients and flavonoids in BX62 silage collectively shape rumen microbial structure and function, enhance pyruvate-centered metabolism, increase VFA production, and ultimately promote milk fat synthesis.

Several limitations should be noted. First, conclusions were mainly based on multi-omics and correlation analyses. Thus, microbes associated with ruminal flavonoid and pyruvate metabolism remain to be verified by in vitro pure culture or synthetic microbial consortia. Second, the current study lacks peripheral blood metabolomic data, which limits in-depth dissection of the transmission and functional mechanism of blood fatty acids and flavonoids along the rumen-mammary gland axis. Further studies integrating serum metabolomics and mammary epithelial cell culture are needed to fully clarify how silage nutrients and bioactive compounds regulate ruminant production performance.

In summary, this study illustrated the mechanism by which L. plantarum BX62 improved alfalfa silage nutritional quality and flavonoid enrichment, thereby reprogramming the rumen microbiome to enhance pyruvate metabolism. This promotes acetate production and subsequently boosts milk fat synthesis via the rumen-mammary gland axis (Fig. 9). These findings offer a sustainable strategy to enhance ruminant production performance through targeted microbiome engineering via manipulation of silage fermentation.

Fig. 9 — (1) Silage phase: Inoculation with *L. plantarum* reshaped the silage microbiome. The secreted CAZymes (e.g., GH/CE families) degrade plant cell walls, liberating plant flavonoids and driving their biotransformation. (2) Rumen intake phase: Dietary intake of *L. plantarum*-inoculated silage markedly suppresses rumen eukaryotes (e.g., fungi, protozoa) and methane production, while enriching acid-producing microbes and flavonoid-degrading bacteria (e.g., *Butyrivibrio and Pseudobutyrivibrio*). (3) Microbial functional remodeling: Structural shifts in the microbial community drive functional adaptation, with activated fiber degradation and pyruvate-centered carbohydrate metabolic pathways. This significantly enhances VFA production in the rumen. (4) Host metabolic response: Microbial-derived VFAs (particularly acetate) are transported via circulation to the mammary gland, serving as precursors for de novo fatty acid synthesis, ultimately leading to significantly increased milk fat content. The figure was created with the assistance of the Figdraw online drawing tool (https://www.figdraw.com/static/index.html#/), with permission and export ID: YPRRTfaa7a.

Methods

Alfalfa silage preparation

Alfalfa (Medicago sativa L.) was mowed at full-flowering stage from an alfalfa field located in Dingxi, Gansu Province, China, and was chopped to 2–4 cm lengths using a forage harvester (9Z-3.0, Zhengzhou Darui Machinery Equipment Co., Ltd., Zhengzhou, China). The DM content of the fresh alfalfa was 330 g/kg of fresh weight (FW), and the chemical composition of the fresh forage was presented in Supplementary Table 1. The forage was divided into two groups for different silage group. L. plantarum BX62 (CGMCC No. 15779), a laboratory-preserved strain with high antioxidant activity and proven efficacy as an effective silage inoculant⁴⁴, was used to prepare the experimental silage. The inoculant was diluted in non-chloride water and sprayed evenly with an application rate of 1 × 10⁵ colony forming unit (CFU)/g fresh matter (FM). The control silage was sprayed with the same water source with no inoculant. Evenly-sprayed alfalfa was tightly wrapped with 4–6 layers of 25 µm thick polypropylene films by a bale wrapper (PT-DB, Jining Pangtai Machinery Co., Ltd., Jining, China). The wrapped round bales were 65 cm in height and 55 cm in diameter, with an average fresh weight of 60 kg per bale. The calculated bale density was 388.53 kg/m³. All bales were transferred to a dry and well-ventilated storage area, and the feeding trial commenced after 30 days of ensiling.

Chemical composition and fermentation quality analysis of silage

Silage sample was dried in a 65 °C oven for 72 h to calculate DM content. The ground dry sample passed through a 1 mm sieve was used to measure neutral detergent fiber treated with a heat-stable amylase (aNDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) contents according to a previous procedure⁴⁵. Crude protein (CP) was measured by the AOAC methods⁴⁶. 20 g fresh silage sample was mixed and homogenized with 180 mL of distilled water, and filtered through 4 layers of sterile gauze. The pH of the filtrate was immediately measured. Five milliliters of filtrate was acidized with H₂SO₄ and the subsequent analysis of organic acids was conducted according to the method of Li et al.⁸. The determination of NH₃-N and WSC concentrations in unacidified filtrate was performed according to the methods of Broderick and Kang⁴⁷, and Murphy⁴⁸, respectively.

Assay of total flavonoids in alfalfa samples

Raw and ensiled alfalfa samples were air-dried at 60 °C, smashed, and then sieved through a 40-mesh sieve. The total flavonoid content was assayed using the total flavonoid kit (Jianglai Biotechnology Co., Ltd., Shanghai, China) following the manufacturer’s protocol with the NaNO₂-Al(NO₃)₃-NaOH chromogenic method.

Animals, management, and experimental design

The animal experiment was conducted on a farm of Dingxi Jupencao Husbandry Co., LTD (Dingxi, Gansu, China) from July to September in 2024. A total of twenty-two healthy mid-lactation Guanzhong dairy goats were used for the experiment, who were 2–3 years old with a mean body weight (BW) of 37.91 ± 0.77 kg and milk yield of 0.61 ± 0.01 kg/d. They were randomly assigned to 2 experimental groups and were offered diets containing 40% concentrate and 60% alfalfa silage inoculated with nothing (control group) or inoculated with L. plantarum BX62 (BX62 group). The specific ingredients and chemical compositions of the prepared total mixed ration (TMR) were presented in Supplementary Table 2. The goats were fed and milked twice a day at 06:00, 18:00, and received free access to water. Milking and recording daily milk yield before feeding. Silage samples were collected after opening, and the concentrate was collected weekly, both of which were immediately frozen at −20 °C.

In vitro rumen fermentation and measurement of fermentation parameters

The substrate for in vitro rumen fermentation was the mixture of the concentrate and alfalfa silage. After freeze-drying for 72 h, the samples were smashed, and the fragments sieved through a 40-mesh sieve were selected as the fermentation substrates. Four rumen-fistulated Hu sheep, provided by a commercial farm in Lanzhou City, Gansu Province, were used as rumen fluid donors. Fresh rumen fluid from each sheep was collected before morning feeding, filtered through 4 layers of sterile gauze, and mixed in equal volumes. The well-mixed rumen fluid was kept warm (39 °C) in a thermos pot and continuously infused with CO₂ until mixed with artificial buffer solution in a ratio of 1:5 (v/v) to prepare the buffered rumen inoculum. The preparation process was conducted following the description of Menke et al.⁴⁹. The fermentation was conducted using the ANKOM RF Gas Production Measurement System (ANKOM Technology, New York, USA). All bottles were incubated at 39 °C and were linked to a gas-tight aluminum bag (Dalian Delin Gaspacking Co., Ltd., Dalian, China) for automated gas collection for subsequent CH₄ production analysis. There were four replicates for each sample. Three more incubation without substrate were used as blanks. After 48 h, the fermentation was ceased on ice. The pH was immediately measured and an aliquot of about 3 mL supernatant from each bottle was transferred into four 5 mL sterile tubes and stored at −80 °C for the analysis of fermentation parameters. The remaining residues were cleaned with water 3 times and dried at 65 °C until the weight remained constant, to calculate the in vitro digestibility of dry matter (IVDMD). The VFAs and CH₄ were determined with reference to the methods described by Li et al.⁸. The NH₃-N concentration was assayed according to the method described by Broderick and Kang⁴⁷.

Sample collection

Milk samples were mixed daily during the last week of the experiment (Day 52 to 59) at a ratio of 6:4 (v/v, morning:afternoon) for analysis of milk quality using an Automatic milk composition analyzer (MilkoScan FT1, Foss, Hillerød, Denmark). On Day 60 of the formal experiment, all experimental dairy goats were subjected to 12 h of fasting treatment with free access to water. Subsequently, all goats were transported to the designated standardized slaughterhouse. All individuals were first treated with electrical stunning to achieve complete unconsciousness, and then humanely sacrificed by exsanguination. This slaughter method is widely used in ruminant production research, which can effectively minimize animal pain and stress response. The rumen content was immediately mixed, and then the ruminal fluid sample was strained through four layers of sterile gauze and transferred to liquid nitrogen for the determination of rumen fermentation parameters, sequencing, and analysis of microbiome and metabolome. The breast tissue was taken into a sterile cryopreserved tube, stored at −80 °C for protein extraction and Western blot (WB) analysis.

Metagenome sequencing and data processing

The total microbial DNA was extracted using the TIANamp stool DNA kit (DP328, TIANGEN Biotechnology, Beijing, China). The quality and concentration of DNA were determined through 1.0% agarose gel electrophoresis and NanoDrop 2000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), respectively. The DNA was randomly fragmented by sonication for paired-end library construction. The qualified metagenome libraries were then subjected to high-throughput sequencing using Illumina NovaSeq6000 (Illumina Inc, San Diego, CA, USA) PE150 platform. Rawdata was preprocessed using the fastp v0.23.4⁵⁰ (parameters: -q 30 –average_qual 20 -l 50) to remove sequencing adapters and reads with low quality. The BWA-MEM2 (v2.2.1)⁵¹ (default parameters) were used to sequentially align the reads to the goat reference genome (GCF_001704415.2), human reference genome (GCF_000001405.40), and Medicago sativa reference genome (GCA_048418095.1). Reads that aligned to these reference genomes were removed. The clean reads were de novo assembled with MEGAHIT (v1.2.9)⁵². Contigs with length <500 bp were filtered out, and CDS prediction was performed using MetaGeneMark (v3.38)⁵³. The non-redundant Unigenes set was then obtained by sequence clustering using CD-HIT (v4.8.1)⁵⁴ (parameters: -n 9 -g 1 -c 0.95 -G 0 -M 0 -d 0 -aS 0.9 -T 45).

Taxonomic classification and functional annotation

The Unigene set was aligned with the NR database (downloaded on April 23, 2024) using diamond blastp⁵⁵ to obtain species taxonomic information. The encoded protein sequences of Unigenes were annotated using Prodigal (v2.6.3)⁵⁶, followed by functional prediction with eggNOG-mapper (v2.1.12, database v5.0.2)⁵⁷ and CAZymes identification via run_dbcan (v.4.1.4) (-t hmmer –hmm_eval 1e-5)⁵⁸. For flavonoid degradation genes, KofamScan⁵⁹ was utilized to perform individual annotation of KO hmm under the pathway ko00946. The abundance profiles of genes were normalized into transcripts per million (TPM) for downstream analysis. For taxonomic profiles, each annotated microbial gene was assigned to a phylogenetic group, and the abundances of genes belonging to the same category levels were aggregated to estimate the abundance of each category level. The same process was applied to calculate the abundance of KEGG Orthology (KOs), KEGG pathways, and CAZymes. The taxonomic origins of functional genes were analyzed using ggplot2 (v3.3.5) in R.

Clustering of rumen microbes via WGCNA analysis

The R software package WGCNA⁶⁰ was used to identify phenotype-related microbial modules from rumen bacterial genera. Microbial abundance data were log₂-transformed (log₂(x + 1)) and standardized, with samples aligned and low-quality data filtered prior to analysis. The Pearson correlation matrix was transformed into an adjacency matrix using a soft thresholding power of 6, and a dynamic tree cut algorithm detected modules of highly correlated genera (minimum module size = 5, merge threshold = 0.25). Each module was assigned a unique color, and grey modules were excluded. Module-trait relationships with dairy goat phenotypes (rumen VFAs and milk quality traits) were quantified by Pearson correlation, with significant modules defined as |r| ≥ 0.4 (P < 0.05). Hub genera within significant modules were identified as those with module membership ≥0.7 and gene significance ≥0.4.

Metagenome-assembled genomes (MAGs) analysis

To comprehensively characterize the rumen microbiome of dairy goats, we performed metagenomic binning to reconstruct MAGs. Contigs from individual assemblies were binned using BASALT v1.1.0 (parameters: –autopara sensitive)⁶¹, a tool that integrates multiple binning algorithms and refines MAGs using neural network-based optimization. Redundancy removal was conducted with dRep v2.5.4⁶² (parameters: -p 72 –ignoreGenomeQuality -pa 0.95 -sa 0.99 -cm larger), applying a 99% average nucleotide identity (ANI) threshold, yielding 2,085 non-redundant bins. Quality assessment was performed using CheckM2 v1.0.2⁶³, identifying 1,347 medium- and high-quality MAGs (≥50% completeness and <10% contamination), including 363 near-complete genomes (>90% completeness and <5% contamination). Open reading frames (ORFs) were predicted with Prodigal⁵⁶. Taxonomic annotation of MAGs was achieved using GTDB-tk v2.4.1(database: release220)⁶⁴. A phylogenetic tree was constructed based on GTDB-tk-derived marker gene alignments, inferred with FastTree (default parameters)⁶⁵, and visualized using iTOL v7.2.1⁶⁶. Functional annotation of flavonoid-degrading genes of MAGs was performed with CAMPER⁶⁷ to identify microbial taxa with potential flavonoid metabolic capabilities.

To establish response patterns of MAGs carrying flavonoid-degrading genes to dietary flavonoid concentrations, differential abundance analysis was performed using DESeq2 (v.1.28.1)⁶⁸. To address dataset size limitations, clean reads were subsequently mapped to a curated ruminant gastrointestinal microbiome database comprising 8745 non-redundant MAGs for enhanced representation⁶⁹. Specifically, clean reads from both BX62 and control groups were initially aligned to the MAGs-derived index database using BWA-MEM2⁵¹. Read counts were generated using SAMtools (v.1.9)⁷⁰ with default parameters, and DESeq2 (v.1.28.1)⁶⁸ was employed for differential analysis, with significance defined as |log2FoldChange| > 1 and p-value < 0.05.

Untargeted metabolome sequencing and bioinformatics analysis

Twenty-five milligrams of freeze-dried silage sample was weighed into an EP tube, homogenizing beads and 1000 μL of extraction solution (methanol: acetonitrile: water = 2:2:1 (v/v/v)) containing isotope-labeled internal standard were added. The vortex-mixed samples were homogenized and then transferred to an ice-water bath for sonication for 5 min, and this step was repeated three times. After standing at −40 °C for 1 h, the samples were centrifuged at 4 °C, 12,000 rpm for 15 min. The supernatant was filtered through 0.22 μm membrane filters and then taken into the injection bottle for assaying on the machine. The target compounds were separated and analyzed via a Vanquish ultra-performance liquid chromatograph (UPLC, Thermo Fisher Scientific) coupled with an Orbitrap Exploris 120 mass spectrometer⁷¹.

Rumen fluid samples were thawed on ice and put into the Starlid™ automated workstation for metabolite extraction. Hundred microliters of sample and 400 µL of extraction solution (methanol: acetonitrile = 1:1 (v/v), including isotope internal standard substance) were pipetted into 96-well protein precipitation plate. The plate was shaken and then let stand for 5 min. The filtered solution was collected after filtering. An aliquot of supernatant was taken and mixed into a QC sample for assaying on the machine. Non-targeted metabolome analysis was performed using a Vanquish ultra-high-performance liquid chromatograph (UHPLC, Thermo Fisher Scientific) in tandem with an Orbitrap Exploris 120 mass spectrometer (Orbitrap MS, Thermo), according to the methods previously described by Yang et al.⁷².

Raw data were converted to mzXML format using ProteoWizard software, and an in-house R package based on XCMS and the BiotreeDB (v3.0) was applied in metabolite identification⁷³. Features and metabolites were filtered based on relative standard deviation de-noising. Then, the missing values were filled with half of the minimum value. Also, internal standard normalization methods were employed in this data analysis. The final dataset containing the information of feature number, sample name, and normalized feature area was imported to SIMCA18.0.1 software package (Sartorius Stedim Data Analytics AB, Umea, Sweden) for multivariate analysis. Data was scaled and logarithmically transformed. In addition, databases including KEGG and MetaboAnalyst (http://www.metaboanalyst.ca/) were used for pathway enrichment analysis.

Targeted metabolomic analysis of flavonoids in the rumen

Two microliters of rumen fluid sample was evaporated to dryness in a grinding tube, then mixed with pre-cooled (−20 °C) 70% (v/v) methanol aqueous solution at a 2:1 ratio, followed by homogenization at 60 Hz for 90 s. After 30 min of incubation on ice, the samples were centrifuged at 12,000 rpm for 3 min at 4 °C. A 300 µL aliquot of the supernatant was transferred to a new EP tube and dried using a vacuum centrifugal concentrator. The residue was reconstituted in 50 µL of 70% methanol aqueous solution, and following centrifugation, 30 µL of the supernatant was analyzed by HPLC (Vanquish Core HPLC, Thermo Scientific, USA) coupled with tandem mass spectrometry (Q Exactive HF-X MS Orbitrap, Thermo Scientific, USA). HPLC analysis was performed on a CSH C18 column (50 mm × 2.1 mm, 1.7 μm; Waters, USA) maintained at 40 °C. The mobile phase consisted of 10 mM ammonium acetate in water (A) and acetonitrile (B), delivered at a flow rate of 0.4 mL/min with a 1 μL injection volume. The gradient elution program was: 10–50% B (0–8 min), 50–80% B (8–9 min), 80–95% B (9–11 min), 95-95% B (11–12 min), 95-10% B (12–12.01 min), and re-equilibration at 10% B (12.01–14 min). Mass spectrometry ion source parameters were set as follows: spray voltage at 3200 V, capillary temperature at 320 °C, sheath gas flow rate at 30 arbitrary units (arb), auxiliary gas flow at 10 arb, and auxiliary gas heater temperature at 400 °C. All target compounds were qualitatively and quantitatively analyzed using Thermo Xcalibur Qual Browser software in this project.

A mixed standard solution containing 231 flavonoid compounds at known concentrations was prepared in 70% methanol to establish calibration curves. The concentrations of target flavonoids in rumen fluid extracts were quantified by comparing their peak areas with the corresponding standard curves.

Western blot analysis

Each 100 mg of mammary gland tissue sample was treated with 1 mL of RIPA lysis buffer (KGB5203, KeyGEN Bio TECH, Nanjing, China), in which PMSF (97064-898, Amresco, VWR International, OH, USA) was added. Under low-temperature conditions, the tissue was fully lysed using a homogenizer. After centrifugation (12,000 × g, 5 min, 4 °C), the supernatant was immediately drawn into a pre-cooled Eppendorf tube, which contained the extracted cell protein and was quantificated using a BCA assay (KGB2101, KeyGEN Bio TECH, Nanjing, China). Samples were separated by 10% SDS-PAGE with the loading amount was 60 μg. After electrophoresis, the separated proteins are transferred onto a PVDF membrane. The membrane was shaken and blocked with 5% skim milk in TBST for 1 h, and then incubated overnight at 4 °C with the TBST-diluted primary antibody, including Anti-ACC1 antibody (3662, Cell Signaling Technology, Boston, USA), Anti-p-ACC1 (Ser79) antibody (3661, Cell Signaling Technology, Boston, USA), Anti-ACSS2 antibody (ABIN2788691, antibodies online, Limerick PA, US), Anti-FASN antibody (10624-2-AP, Proteintech, Wuhan, China), Anti-BDH1 antibody (67448-1-Ig, Proteintech, Wuhan, China) and Anti-β-actin antibody (ab170325, Abcam, Cambridge, UK). After washing three times with TBST, the membranes were probed with HRP-conjugated secondary antibodies, including Donkey Anti-Rabbit IgG H&L (HRP) (ab205722, Abcam, Cambridge, UK), Rabbit Anti-Mouse IgG H&L (HRP) (ab6728, Abcam, Cambridge, UK). Then, the membranes were treated with chemiluminescence ECL kit (32209, Thermo Fisher Scientific, Pittsburgh, PA, USA) and signals were detected with the Tanon 5200 luminescence imaging system. Protein expression levels were quantified by densitometric analysis using Image Pro Plus 6.0 software, with band intensities normalized to β-actin as the loading control.

Microbial culture experiment

The strain C. difficile 630 (ATCC BAA-1382) was obtained from Wuhan Huizao Biotechnology Co., LTD (Hubei, China), which sourced it from the ATCC (Manassas, VA, USA). The bacterial strain was cultured under strictly anaerobic conditions in 30 mL Hungate tubes containing DSMZ Medium 110. The tubes were equipped with butyl rubber stoppers and aluminum seals and incubated at 37 °C. Prior to experimentation, C. difficile 630 was subcultured twice in succession. Then, 150 μL aliquots of quercetin (Cool Biotechnology Co., Ltd., Anhui, China) stock solutions, prepared in DMSO at concentrations of 0, 1, 2.5, 5, 10, and 20 mM, were aseptically added to 14.3 mL of sterile culture medium, followed by inoculation with 550 μL of log-phase bacterial culture. The test cultures were then incubated at 37 °C for 48 h. At designated time points, aliquots of bacterial suspension were aseptically collected using sterile syringes. The optical density at 600 nm (OD600) was measured, while 2 mL samples were cryopreserved at −80 °C for subsequent HPLC analysis of quercetin.

Statistical analysis

Independent sample T-test was performed to assess differences between groups in silage fermentation quality, the total flavonoid content in alfalfa samples, rumen fermentation parameters, dairy goat performance, and relative expression levels from WB analysis. Statistical significance was set at P < 0.05. For alfalfa silage microbial metagenomics, Linear Discriminant Analysis Effect Size (LEfSe) was employed to identify differentially abundant taxa across taxonomic levels (LDA score > 2, P < 0.05). Wilcoxon rank-sum tests were used to compare α-diversity, KEGG pathways, and CAZyme profiles, with significance determined by Benjamini-Hochberg false discovery rate (FDR) -adjusted P-values (P < 0.05). For rumen microbial metagenomics, microbial α-diversity, domain-level taxonomy, and CAZyme abundances were compared between groups using Wilcoxon rank-sum tests (FDR-adjusted P < 0.05). KEGG pathway enrichment analysis was performed using the clusterProfiler R package⁷⁴. For all metagenomic data, Principal Coordinates Analysis (PCoA) based on Bray-Curtis dissimilarity was performed to assess β-diversity between groups. For metabolomes, metabolites with P < 0.05 in Students’ t test and VIP (Variable Importance in the Projection) > 1 in the first principal component of the OPLS-DA model were identified as differential metabolites.

Mediation analysis was conducted using the mediation package in R, and mediation path diagrams with labeled coefficients and significance levels were generated using the DiagrammeR package. Variance Partitioning Analysis (VPA) was performed using the vegan package in R, followed by single-factor redundancy analysis (RDA) with significance tested by 999 permutations and adjusted R² calculated. Mantel test was performed using the vegan package in R. All visualizations were generated using the ggplot2 and linkET packages. Correlation analysis was conducted using Spearman’s rank correlation. Visualize the multiplexing network using Gephi software v 0.10. Further, enrichment analysis of metabolites related to milk fat was conducted using the online software TUTU analysis platform (https://www.cloudtutu.com/).

Ethics approval and consent to participate

All animal experiments were performed in strict accordance with the national guidelines for laboratory animal welfare and ethical standards. All experimental protocols were approved by the Animal Ethics Committee of Lanzhou University (Approval No. EAF2025017). All animal handling, feeding management, and slaughter procedures in this study complied with the ARRIVE guidelines.

Supplementary information

Supplementary Table^{(3.8MB, pdf)}

WB_origin^{(1.3MB, zip)}

Acknowledgements

This work was supported by the National Natural Science Foundation of China (32561143286). And, Fig. 9 was created with the assistance of the Figdraw network tool (https://www.figdraw.com/static/index.html#/), with permission and export ID: YPRRTfaa7a.

Author contributions

J.M. and Q.L. conceptualized the study and conducted the majority of the research. J.M. drafted the main manuscript. J.M. and Q.L. visualized the figures and tables, developed custom scripts/code, and conducted bioinformatics analysis. G.H.X. and S.U. are involved in the determination of phenotypic indicators. J.M., G.H.X., and J.Y.Z. developed the methodology, managed data curation, and conducted data analysis. Q.L., G.H.X., J.Y.Z., and S.U. were involved in sample collection and revised the manuscript. X.S.G. revised the manuscript. X.S.G. acquired financial support and supervised the experiments. All listed authors approved the final version of the manuscript.

Data availability

The metagenome sequence data that support this study have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession numbers PRJNA1309858 and PRJNA1309856, respectively. The rumen non-targeted metabolomics data and rumen flavonoid-targeted metabolomics data have been submitted to the MetaboLights repository with the accession numbers MTBLS13648 and MTBLS13651, respectively. The silage non-targeted metabolomics data have been submitted to Metabolomics Workbench with the https://doi.org/10.21228/M8Z857.

Code availability

All analyses were conducted using open-source software, which has been explicitly stated in the method section. No custom code or mathematical algorithm was utilized in this study.

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.

These authors contributed equally: Jing Ma, Qiang Li.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41522-026-01031-6.

References

1.Xu, D. et al. Modulation of metabolome and bacterial community in whole crop corn silage by inoculating homofermentative Lactobacillus plantarum and heterofermentative Lactobacillus buchneri. Front. Microbiol.9, 3299 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Vaičiulienė, G. et al. Origanum vulgare and Thymus vulgaris extract usability to improve silage hygienic quality and reduce mycotoxin concentrations. J. Microbiol. Biotechnol.30, 1149–1155 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Xu, D. et al. The bacterial community and metabolome dynamics and their interactions modulate fermentation process of whole crop corn silage prepared with or without inoculants. Micro. Biotechnol.14, 561–576 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Zhao, Y. et al. Enhancing alfalfa and sorghum silage quality using agricultural wastes: fermentation dynamics, microbial communities, and functional insights. BMC Plant Biol.25, 728 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Okoye, C. O. et al. Metagenomics analysis reveals the performance of homo- and heterofermentative lactic acid bacteria in alfalfa silage fermentation, bacterial community, and functional profiles. J. Anim. Sci.101, skad163 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Li, F. et al. Probiotic effect of ferulic acid esterase-producing Lactobacillus plantarum inoculated alfalfa silage on digestion, antioxidant, and immunity status of lactating dairy goats. Anim. Nutr.11, 38–47 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Contreras-Govea, F. E., Muck, R. E., Broderick, G. A. & Weimer, P. J. Lactobacillus plantarum effects on silage fermentation and in vitro microbial yield. Anim. Feed Sci. Technol.179, 61–68 (2013). [Google Scholar]
8.Li, Z. et al. Effects of bacteriocin-producing Lactiplantibacillus plantarum on bacterial community and fermentation profile of whole-plant corn silage and its in vitro ruminal fermentation, microbiota, and CH(4) emissions. J. Anim. Sci. Biotechnol.15, 107 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Oskoueian, E. et al. Manipulation of rice straw silage fermentation with different types of lactic acid bacteria inoculant affects rumen microbial fermentation characteristics and methane production. Vet. Sci.8, 100 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Oliveira, A. S. et al. Meta-analysis of effects of inoculation with homofermentative and facultative heterofermentative lactic acid bacteria on silage fermentation, aerobic stability, and the performance of dairy cows. J. Dairy Sci.100, 4587–4603 (2017). [DOI] [PubMed] [Google Scholar]
11.Zhang, Y. et al. Effects of antioxidant-rich Lactiplantibacillus plantarum inoculated alfalfa silage on rumen fermentation, antioxidant and immunity status, and mammary gland gene expression in dairy goats. J. Anim. Sci. Biotechnol.15, 9 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Liu, Y. et al. Volatile metabolomics and metagenomics reveal the effects of lactic acid bacteria on alfalfa silage quality, microbial communities, and volatile organic compounds. Commun. Biol.7, 1565 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Zhao, M. et al. Lactobacillus plantarum and propionic acid improve the fermentation quality of high-moisture amaranth silage by altering the microbial community composition. Front. Microbiol.13, 1066641 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Chen, L. et al. Effects of applying lactic acid bacteria and propionic acid on fermentation quality, aerobic stability and in vitro gas production of forage-based total mixed ration silage in Tibet. Anim. Prod. Sci.59, 376–383 (2018). [Google Scholar]
15.Xu, Y. et al. Ensiling of rice straw enhances the nutritive quality, improves average daily gain, reduces in vitro methane production and increases ruminal bacterial diversity in growing Hu lambs. Anim. Feed Sci. Technol.295, 115513 (2023). [Google Scholar]
16.Tian, Z. et al. Transcriptional regulation of milk fat synthesis in dairy cattle. J. Funct. Foods96, 105208 (2022). [Google Scholar]
17.Urrutia, N. L. & Harvatine, K. J. Acetate dose-dependently stimulates milk fat synthesis in lactating dairy Cows1334. J. Nutr.147, 763–769 (2017). [DOI] [PubMed] [Google Scholar]
18.Chilliard, Y., Ferlay, A., Mansbridge, R. M. & Doreau, M. Ruminant milk fat plasticity: nutritional control of saturated, polyunsaturated, trans and conjugated fatty acids. Ann. Zootech.49, 181–205 (2000). [Google Scholar]
19.Zhong, P. et al. Precision microbial regulation: strategies for modulating GIT microbiota for host health. iMetaOmics2, e54 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Li, Z., Li, F., Kharazian, Z. A. & Guo, X. Effect of inoculating two bacteriocin-producing Lactiplantibacillus plantarum strains at ensiling on in vitro rumen fermentation and methane emissions of alfalfa silage with two dry matter contents. Animals13, 384 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Pickett, A. T. et al. Shifts in bacterial communities in the rumen, vagina, and uterus of beef heifers receiving different levels of concentrate. J. Anim. Sci.100, skac338 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Zhang, Z. et al. Effects of dietary forage neutral detergent fiber and rumen degradable starch ratios on chewing activity, ruminal fermentation, ruminal microbes and nutrient digestibility of Hu sheep fed a pelleted total mixed ration. J. Anim. Sci.102, skae100 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Ma, P. et al. Rumen microbiota responses to the enzymatic hydrolyzed cottonseed peptide supplement under high-concentrate diet feeding process. Front. Vet. Sci.9, 984634 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Liu, E. et al. Reciprocal regulation of rumen microbiota and epithelial genes in response to small peptide supplementation for feed efficiency in beef cattle. Front. Vet. Sci.12, 1714827 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Pidcock, S. E. et al. Phylogenetic systematics of Butyrivibrio and Pseudobutyrivibrio genomes illustrate vast taxonomic diversity, open genomes and an abundance of carbohydrate-active enzyme family isoforms. Microb. Genom.7, 000638 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Li, M. et al. Mulberry flavonoids modulate rumen bacteria to alter fermentation kinetics in water buffalo. PeerJ10, e14309 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Fang, W. et al. Theaflavins in black tea ameliorate high-fat diet-induced obesity and inflammation via gut microbiota, AMPK-mediated metabolism, and NF-κB pathway. Phytomedicine148, 157314 (2025). [DOI] [PubMed] [Google Scholar]
28.Zhou, Y. et al. Taxifolin increased semen quality of Duroc boars by improving gut microbes and blood metabolites. Front. Microbiol.13, 1020628 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Ma, T. et al. Dietary supplementation with mulberry leaf flavonoids inhibits methanogenesis in sheep. Anim. Sci. J.88, 72–78 (2017). [DOI] [PubMed] [Google Scholar]
30.Krishnamurty, H. G., Cheng, K.-J., Jones, G. A., Simpson, F. J. & Watkin, J. E. Identification of products produced by the anaerobic degradation of rutin and related flavonoids by Butyrivibrio sp. C3. Can. J. Microbiol.16, 759–767 (1970). [DOI] [PubMed] [Google Scholar]
31.Mi, H. T. N., Kim, H., Lee, J. S., Eser, B. E. & Han, J. Flavonoids biotransformation by human gut bacterium Dorea sp. MRG-IFC3 cell-free extract. J. Microbiol. Biotechnol.34, 1270–1275 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Park, D.-J., Lee, Y.-S. & Choi, Y.-L. Characterization of a cold-active β-glucosidase from Paenibacillus xylanilyticus KJ-03 capable of hydrolyzing isoflavones daidzin and genistin. Protein J.32, 579–584 (2013). [DOI] [PubMed] [Google Scholar]
33.Kim, M., Kim, N. & Han, J. Metabolism of Kaempferia parviflora polymethoxyflavones by human intestinal bacterium Bautia sp. MRG-PMF1. J. Agric. Food Chem.62, 12377–12383 (2014). [DOI] [PubMed] [Google Scholar]
34.Tsai, C. G., Gates, D. M., Ingledew, W. M. & Jones, G. A. Products of anaerobic phloroglucinol degradation by Coprococcus sp. Pe15. Can. J. Microbiol.22, 159–164 (1976). [DOI] [PubMed] [Google Scholar]
35.Li, Y. et al. An aldolase-dependent phloroglucinol degradation pathway in Collinsella sp. zg1085. Appl. Environ. Microbiol.90, e0104724 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Varel Vincent, H. & Jung Hans-Joachim, G. Influence of forage phenolics on ruminal fibrolytic bacteria and in vitro fiber degradation. Appl. Environ. Microbiol.52, 275–280 (1986). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Yu, S. et al. Dietary citrus flavonoid extract improves lactational performance through modulating rumen microbiome and metabolites in dairy cows. Food Funct.14, 94–111 (2023). [DOI] [PubMed] [Google Scholar]
38.Liang, J. et al. Rumen microbes, enzymes, metabolisms, and application in lignocellulosic waste conversion - A comprehensive review. Biotechnol. Adv.71, 108308 (2024). [DOI] [PubMed] [Google Scholar]
39.Li, Y. et al. Interactions between anaerobic fungi and methanogens in the rumen and their biotechnological potential in biogas production from lignocellulosic materials. Microorganisms9, 190 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Mulakala, B. K. et al. Influence of dietary carbohydrate profile on the dairy cow rumen meta-proteome. J. Dairy Sci.105, 8485–8496 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Mousa, A. A., Zhang, H., Duan, H., Zhang, J. & Mao, S. Metagenomic analysis reveals rumen microbiome enrichment and functional genes adjustment in carbohydrate metabolism induced by different sorting behavior in mid-lactation dairy cows. Anim. Microbiome7, 82 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Zhang, X., Du, W., Xie, K. & Hou, F. Lycium barbarum byproducts modulate rumen fermentation, enhance digestive enzyme activity, and improve immune and antioxidant status in grazing sheep. Food Sci. Nutr.14, e71509 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kong, L. et al. Rumen fermentation characteristics in pre- and post-weaning calves upon feeding with mulberry leaf flavonoids and Candida tropicalis individually or in combination as a supplement. Animals9, 990 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Zhang, X. et al. Antioxidant, flavonoid, α-tocopherol, β-carotene, fatty acids, and fermentation profiles of alfalfa silage inoculated with novel Lactiplantibacillus plantarum and Pediococcus acidilactici strains with high-antioxidant activity. Anim. Feed Sci. Technol.288, 115301 (2022). [Google Scholar]
45.Van Soest, P. J., Robertson, J. B. & Lewis, B. A. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci.74, 3583–3597 (1991). [DOI] [PubMed] [Google Scholar]
46.Feldsine, P., Abeyta, C. & Andrews, W. H. AOAC INTERNATIONAL Methods Committee guidelines for validation of qualitative and quantitative food microbiological official methods of analysis. J. AOAC Int.85, 1187–1200 (2002). [PubMed] [Google Scholar]
47.Broderick, G. A. & Kang, J. H. Automated simultaneous determination of ammonia and total amino acids in ruminal fluid and in vitro media. J. Dairy Sci.63, 64–75 (1980). [DOI] [PubMed] [Google Scholar]
48.Murphy, R. P. A method for the extraction of plant samples and the determination of total soluble carbohydrates. J. Sci. Food Agric.9, 714–717 (1958). [Google Scholar]
49.Menke, K. H. Estimation of the energetic feed value obtained from chemical analysis and in vitro gas production using rumen fluid. Anim. Res. Dev.28, 7–55 (1988). [Google Scholar]
50.Chen, S. Ultrafast one-pass FASTQ data preprocessing, quality control, and deduplication using fastp. iMeta2, e107 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Vasimuddin, M., Misra, S., Li, H. & Aluru, S. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. IEEE Int. Parallel Distrib. Process. Symp.2019, 314–324 (2019). [Google Scholar]
52.Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics31, 1674–1676 (2015). [DOI] [PubMed] [Google Scholar]
53.Ismail, W. M., Ye, Y. & Tang, H. Gene finding in metatranscriptomic sequences. BMC Bioinform.15, S8 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Li, W. & Godzik, A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics22, 1658–1659 (2006). [DOI] [PubMed] [Google Scholar]
55.Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using DIAMOND. Nat. Methods12, 59–60 (2015). [DOI] [PubMed] [Google Scholar]
56.Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinform.11, 119 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Huerta-Cepas, J. et al. Fast genome-wide functional annotation through orthology assignment by eggNOG-Mapper. Mol. Biol. Evol.34, 2115–2122 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Ausland, C. et al. dbCAN-PUL: a database of experimentally characterized CAZyme gene clusters and their substrates. Nucleic Acids Res.49, D523–d528 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Aramaki, T. et al. KofamKOALA: KEGG Ortholog assignment based on profile HMM and adaptive score threshold. Bioinformatics36, 2251–2252 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinform.9, 559 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Qiu, Z. et al. BASALT refines binning from metagenomic data and increases resolution of genome-resolved metagenomic analysis. Nat. Commun.15, 2179 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J.11, 2864–2868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Chklovski, A., Parks, D. H., Woodcroft, B. J. & Tyson, G. W. CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning. Nat. Methods20, 1203–1212 (2023). [DOI] [PubMed] [Google Scholar]
64.Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk v2: memory friendly classification with the genome taxonomy database. Bioinformatics38, 5315–5316 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol. Biol. Evol.26, 1641–1650 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res.52, W78–W82 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.McGivern, B. B. et al. Microbial polyphenol metabolism is part of the thawing permafrost carbon cycle. Nat. Microbiol.9, 1454–1466 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.15, 550 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Xie, F. et al. An integrated gene catalog and over 10,000 metagenome-assembled genomes from the gastrointestinal microbiome of ruminants. Microbiome9, 137 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics25, 2078–2079 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Cui, Z. et al. Integrated physiological, transcriptomic and metabolomic analyses reveal ROS regulatory mechanisms in two castor bean varieties under alkaline stress. Plant Physiol. Biochem.220, 109518 (2025). [DOI] [PubMed] [Google Scholar]
72.Yang, S.-L. et al. Multi-omics reveals the molecular mechanism of the combined toxic effects of PFOA and 4-HBP exposure in MCF-7 cells and the key player: mTORC1. Environ. Int.188, 108778 (2024). [DOI] [PubMed] [Google Scholar]
73.Zhou, Z. et al. Metabolite annotation from knowns to unknowns through knowledge-guided multi-layer metabolic networking. Nat. Commun.13, 6656 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Xu, S. et al. Using clusterProfiler to characterize multiomics data. Nat. Protoc.19, 3292–3320 (2024). [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Table^{(3.8MB, pdf)}

WB_origin^{(1.3MB, zip)}

Data Availability Statement

All analyses were conducted using open-source software, which has been explicitly stated in the method section. No custom code or mathematical algorithm was utilized in this study.

[CR1] 1.Xu, D. et al. Modulation of metabolome and bacterial community in whole crop corn silage by inoculating homofermentative Lactobacillus plantarum and heterofermentative Lactobacillus buchneri. Front. Microbiol.9, 3299 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Vaičiulienė, G. et al. Origanum vulgare and Thymus vulgaris extract usability to improve silage hygienic quality and reduce mycotoxin concentrations. J. Microbiol. Biotechnol.30, 1149–1155 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Xu, D. et al. The bacterial community and metabolome dynamics and their interactions modulate fermentation process of whole crop corn silage prepared with or without inoculants. Micro. Biotechnol.14, 561–576 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Zhao, Y. et al. Enhancing alfalfa and sorghum silage quality using agricultural wastes: fermentation dynamics, microbial communities, and functional insights. BMC Plant Biol.25, 728 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Okoye, C. O. et al. Metagenomics analysis reveals the performance of homo- and heterofermentative lactic acid bacteria in alfalfa silage fermentation, bacterial community, and functional profiles. J. Anim. Sci.101, skad163 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Li, F. et al. Probiotic effect of ferulic acid esterase-producing Lactobacillus plantarum inoculated alfalfa silage on digestion, antioxidant, and immunity status of lactating dairy goats. Anim. Nutr.11, 38–47 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Contreras-Govea, F. E., Muck, R. E., Broderick, G. A. & Weimer, P. J. Lactobacillus plantarum effects on silage fermentation and in vitro microbial yield. Anim. Feed Sci. Technol.179, 61–68 (2013). [Google Scholar]

[CR8] 8.Li, Z. et al. Effects of bacteriocin-producing Lactiplantibacillus plantarum on bacterial community and fermentation profile of whole-plant corn silage and its in vitro ruminal fermentation, microbiota, and CH(4) emissions. J. Anim. Sci. Biotechnol.15, 107 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Oskoueian, E. et al. Manipulation of rice straw silage fermentation with different types of lactic acid bacteria inoculant affects rumen microbial fermentation characteristics and methane production. Vet. Sci.8, 100 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Oliveira, A. S. et al. Meta-analysis of effects of inoculation with homofermentative and facultative heterofermentative lactic acid bacteria on silage fermentation, aerobic stability, and the performance of dairy cows. J. Dairy Sci.100, 4587–4603 (2017). [DOI] [PubMed] [Google Scholar]

[CR11] 11.Zhang, Y. et al. Effects of antioxidant-rich Lactiplantibacillus plantarum inoculated alfalfa silage on rumen fermentation, antioxidant and immunity status, and mammary gland gene expression in dairy goats. J. Anim. Sci. Biotechnol.15, 9 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Liu, Y. et al. Volatile metabolomics and metagenomics reveal the effects of lactic acid bacteria on alfalfa silage quality, microbial communities, and volatile organic compounds. Commun. Biol.7, 1565 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Zhao, M. et al. Lactobacillus plantarum and propionic acid improve the fermentation quality of high-moisture amaranth silage by altering the microbial community composition. Front. Microbiol.13, 1066641 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Chen, L. et al. Effects of applying lactic acid bacteria and propionic acid on fermentation quality, aerobic stability and in vitro gas production of forage-based total mixed ration silage in Tibet. Anim. Prod. Sci.59, 376–383 (2018). [Google Scholar]

[CR15] 15.Xu, Y. et al. Ensiling of rice straw enhances the nutritive quality, improves average daily gain, reduces in vitro methane production and increases ruminal bacterial diversity in growing Hu lambs. Anim. Feed Sci. Technol.295, 115513 (2023). [Google Scholar]

[CR16] 16.Tian, Z. et al. Transcriptional regulation of milk fat synthesis in dairy cattle. J. Funct. Foods96, 105208 (2022). [Google Scholar]

[CR17] 17.Urrutia, N. L. & Harvatine, K. J. Acetate dose-dependently stimulates milk fat synthesis in lactating dairy Cows1334. J. Nutr.147, 763–769 (2017). [DOI] [PubMed] [Google Scholar]

[CR18] 18.Chilliard, Y., Ferlay, A., Mansbridge, R. M. & Doreau, M. Ruminant milk fat plasticity: nutritional control of saturated, polyunsaturated, trans and conjugated fatty acids. Ann. Zootech.49, 181–205 (2000). [Google Scholar]

[CR19] 19.Zhong, P. et al. Precision microbial regulation: strategies for modulating GIT microbiota for host health. iMetaOmics2, e54 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Li, Z., Li, F., Kharazian, Z. A. & Guo, X. Effect of inoculating two bacteriocin-producing Lactiplantibacillus plantarum strains at ensiling on in vitro rumen fermentation and methane emissions of alfalfa silage with two dry matter contents. Animals13, 384 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Pickett, A. T. et al. Shifts in bacterial communities in the rumen, vagina, and uterus of beef heifers receiving different levels of concentrate. J. Anim. Sci.100, skac338 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Zhang, Z. et al. Effects of dietary forage neutral detergent fiber and rumen degradable starch ratios on chewing activity, ruminal fermentation, ruminal microbes and nutrient digestibility of Hu sheep fed a pelleted total mixed ration. J. Anim. Sci.102, skae100 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Ma, P. et al. Rumen microbiota responses to the enzymatic hydrolyzed cottonseed peptide supplement under high-concentrate diet feeding process. Front. Vet. Sci.9, 984634 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Liu, E. et al. Reciprocal regulation of rumen microbiota and epithelial genes in response to small peptide supplementation for feed efficiency in beef cattle. Front. Vet. Sci.12, 1714827 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Pidcock, S. E. et al. Phylogenetic systematics of Butyrivibrio and Pseudobutyrivibrio genomes illustrate vast taxonomic diversity, open genomes and an abundance of carbohydrate-active enzyme family isoforms. Microb. Genom.7, 000638 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Li, M. et al. Mulberry flavonoids modulate rumen bacteria to alter fermentation kinetics in water buffalo. PeerJ10, e14309 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Fang, W. et al. Theaflavins in black tea ameliorate high-fat diet-induced obesity and inflammation via gut microbiota, AMPK-mediated metabolism, and NF-κB pathway. Phytomedicine148, 157314 (2025). [DOI] [PubMed] [Google Scholar]

[CR28] 28.Zhou, Y. et al. Taxifolin increased semen quality of Duroc boars by improving gut microbes and blood metabolites. Front. Microbiol.13, 1020628 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Ma, T. et al. Dietary supplementation with mulberry leaf flavonoids inhibits methanogenesis in sheep. Anim. Sci. J.88, 72–78 (2017). [DOI] [PubMed] [Google Scholar]

[CR30] 30.Krishnamurty, H. G., Cheng, K.-J., Jones, G. A., Simpson, F. J. & Watkin, J. E. Identification of products produced by the anaerobic degradation of rutin and related flavonoids by Butyrivibrio sp. C3. Can. J. Microbiol.16, 759–767 (1970). [DOI] [PubMed] [Google Scholar]

[CR31] 31.Mi, H. T. N., Kim, H., Lee, J. S., Eser, B. E. & Han, J. Flavonoids biotransformation by human gut bacterium Dorea sp. MRG-IFC3 cell-free extract. J. Microbiol. Biotechnol.34, 1270–1275 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Park, D.-J., Lee, Y.-S. & Choi, Y.-L. Characterization of a cold-active β-glucosidase from Paenibacillus xylanilyticus KJ-03 capable of hydrolyzing isoflavones daidzin and genistin. Protein J.32, 579–584 (2013). [DOI] [PubMed] [Google Scholar]

[CR33] 33.Kim, M., Kim, N. & Han, J. Metabolism of Kaempferia parviflora polymethoxyflavones by human intestinal bacterium Bautia sp. MRG-PMF1. J. Agric. Food Chem.62, 12377–12383 (2014). [DOI] [PubMed] [Google Scholar]

[CR34] 34.Tsai, C. G., Gates, D. M., Ingledew, W. M. & Jones, G. A. Products of anaerobic phloroglucinol degradation by Coprococcus sp. Pe15. Can. J. Microbiol.22, 159–164 (1976). [DOI] [PubMed] [Google Scholar]

[CR35] 35.Li, Y. et al. An aldolase-dependent phloroglucinol degradation pathway in Collinsella sp. zg1085. Appl. Environ. Microbiol.90, e0104724 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Varel Vincent, H. & Jung Hans-Joachim, G. Influence of forage phenolics on ruminal fibrolytic bacteria and in vitro fiber degradation. Appl. Environ. Microbiol.52, 275–280 (1986). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Yu, S. et al. Dietary citrus flavonoid extract improves lactational performance through modulating rumen microbiome and metabolites in dairy cows. Food Funct.14, 94–111 (2023). [DOI] [PubMed] [Google Scholar]

[CR38] 38.Liang, J. et al. Rumen microbes, enzymes, metabolisms, and application in lignocellulosic waste conversion - A comprehensive review. Biotechnol. Adv.71, 108308 (2024). [DOI] [PubMed] [Google Scholar]

[CR39] 39.Li, Y. et al. Interactions between anaerobic fungi and methanogens in the rumen and their biotechnological potential in biogas production from lignocellulosic materials. Microorganisms9, 190 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Mulakala, B. K. et al. Influence of dietary carbohydrate profile on the dairy cow rumen meta-proteome. J. Dairy Sci.105, 8485–8496 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Mousa, A. A., Zhang, H., Duan, H., Zhang, J. & Mao, S. Metagenomic analysis reveals rumen microbiome enrichment and functional genes adjustment in carbohydrate metabolism induced by different sorting behavior in mid-lactation dairy cows. Anim. Microbiome7, 82 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Zhang, X., Du, W., Xie, K. & Hou, F. Lycium barbarum byproducts modulate rumen fermentation, enhance digestive enzyme activity, and improve immune and antioxidant status in grazing sheep. Food Sci. Nutr.14, e71509 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Kong, L. et al. Rumen fermentation characteristics in pre- and post-weaning calves upon feeding with mulberry leaf flavonoids and Candida tropicalis individually or in combination as a supplement. Animals9, 990 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Zhang, X. et al. Antioxidant, flavonoid, α-tocopherol, β-carotene, fatty acids, and fermentation profiles of alfalfa silage inoculated with novel Lactiplantibacillus plantarum and Pediococcus acidilactici strains with high-antioxidant activity. Anim. Feed Sci. Technol.288, 115301 (2022). [Google Scholar]

[CR45] 45.Van Soest, P. J., Robertson, J. B. & Lewis, B. A. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci.74, 3583–3597 (1991). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Feldsine, P., Abeyta, C. & Andrews, W. H. AOAC INTERNATIONAL Methods Committee guidelines for validation of qualitative and quantitative food microbiological official methods of analysis. J. AOAC Int.85, 1187–1200 (2002). [PubMed] [Google Scholar]

[CR47] 47.Broderick, G. A. & Kang, J. H. Automated simultaneous determination of ammonia and total amino acids in ruminal fluid and in vitro media. J. Dairy Sci.63, 64–75 (1980). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Murphy, R. P. A method for the extraction of plant samples and the determination of total soluble carbohydrates. J. Sci. Food Agric.9, 714–717 (1958). [Google Scholar]

[CR49] 49.Menke, K. H. Estimation of the energetic feed value obtained from chemical analysis and in vitro gas production using rumen fluid. Anim. Res. Dev.28, 7–55 (1988). [Google Scholar]

[CR50] 50.Chen, S. Ultrafast one-pass FASTQ data preprocessing, quality control, and deduplication using fastp. iMeta2, e107 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Vasimuddin, M., Misra, S., Li, H. & Aluru, S. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. IEEE Int. Parallel Distrib. Process. Symp.2019, 314–324 (2019). [Google Scholar]

[CR52] 52.Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics31, 1674–1676 (2015). [DOI] [PubMed] [Google Scholar]

[CR53] 53.Ismail, W. M., Ye, Y. & Tang, H. Gene finding in metatranscriptomic sequences. BMC Bioinform.15, S8 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Li, W. & Godzik, A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics22, 1658–1659 (2006). [DOI] [PubMed] [Google Scholar]

[CR55] 55.Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using DIAMOND. Nat. Methods12, 59–60 (2015). [DOI] [PubMed] [Google Scholar]

[CR56] 56.Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinform.11, 119 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Huerta-Cepas, J. et al. Fast genome-wide functional annotation through orthology assignment by eggNOG-Mapper. Mol. Biol. Evol.34, 2115–2122 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Ausland, C. et al. dbCAN-PUL: a database of experimentally characterized CAZyme gene clusters and their substrates. Nucleic Acids Res.49, D523–d528 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Aramaki, T. et al. KofamKOALA: KEGG Ortholog assignment based on profile HMM and adaptive score threshold. Bioinformatics36, 2251–2252 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinform.9, 559 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Qiu, Z. et al. BASALT refines binning from metagenomic data and increases resolution of genome-resolved metagenomic analysis. Nat. Commun.15, 2179 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J.11, 2864–2868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Chklovski, A., Parks, D. H., Woodcroft, B. J. & Tyson, G. W. CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning. Nat. Methods20, 1203–1212 (2023). [DOI] [PubMed] [Google Scholar]

[CR64] 64.Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk v2: memory friendly classification with the genome taxonomy database. Bioinformatics38, 5315–5316 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol. Biol. Evol.26, 1641–1650 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res.52, W78–W82 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.McGivern, B. B. et al. Microbial polyphenol metabolism is part of the thawing permafrost carbon cycle. Nat. Microbiol.9, 1454–1466 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.15, 550 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Xie, F. et al. An integrated gene catalog and over 10,000 metagenome-assembled genomes from the gastrointestinal microbiome of ruminants. Microbiome9, 137 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics25, 2078–2079 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] 71.Cui, Z. et al. Integrated physiological, transcriptomic and metabolomic analyses reveal ROS regulatory mechanisms in two castor bean varieties under alkaline stress. Plant Physiol. Biochem.220, 109518 (2025). [DOI] [PubMed] [Google Scholar]

[CR72] 72.Yang, S.-L. et al. Multi-omics reveals the molecular mechanism of the combined toxic effects of PFOA and 4-HBP exposure in MCF-7 cells and the key player: mTORC1. Environ. Int.188, 108778 (2024). [DOI] [PubMed] [Google Scholar]

[CR73] 73.Zhou, Z. et al. Metabolite annotation from knowns to unknowns through knowledge-guided multi-layer metabolic networking. Nat. Commun.13, 6656 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR74] 74.Xu, S. et al. Using clusterProfiler to characterize multiomics data. Nat. Protoc.19, 3292–3320 (2024). [DOI] [PubMed] [Google Scholar]

PERMALINK

Lactiplantibacillus plantarum-inoculated silage improves milk fat in dairy goats by reprogramming rumen microbiota to promote pyruvate metabolism

Jing Ma

Qiang Li

Guanghao Xia

Jiayao Zhang

Samaila Usman

Xusheng Guo

Abstract

Introduction

Results

Fermentation characteristics and chemical composition of alfalfa silage

Table 1.

Metagenomic and metabolomic changes during silage fermentation

Fig. 1. Metagenomic and metabolomic analyses of two groups of silage.

Fig. 2. Associations between differential CAZymes and flavonoid compounds in silage, and taxonomic attribution of these CAZymes.

Effects of L. plantarum-inoculated silage on rumen fermentation and milk production in dairy goats

Fig. 3. Effects of BX62 silage on ruminal fermentation, milk composition, and mammary lipogenic pathways in dairy goats.

Specific rumen microbial taxa mediated the link between silage and ruminal acetate synthesis

Fig. 4. Effect of feeding different silage on rumen microbial community structure in dairy goat.

Fig. 5. Associations of silage-related indicators with rumen microbiota and dairy goat phenotypes.

Flavonoid-degrading bacteria in the rumen responded to higher intake of silage flavonoids

Fig. 6. Profiling of flavonoid degradation and associated bacterial communities in the rumen.

The reshaped ruminal microbial community enhanced carbohydrate metabolism and VFA production

Fig. 7. Effect of feeding different silage on rumen microbial function profiles in dairy goat.

Silage-driven alterations in the rumen metabolome and their association with milk fat

Fig. 8. Integrated analysis of milk fat-associated metabolites and microbiota in dairy goats.

Discussion

Fig. 9. Schematic diagram of the silage-rumen microbiota-milk fat synthesis regulatory mechanism.

Methods

Alfalfa silage preparation

Chemical composition and fermentation quality analysis of silage

Assay of total flavonoids in alfalfa samples

Animals, management, and experimental design

In vitro rumen fermentation and measurement of fermentation parameters

Sample collection

Metagenome sequencing and data processing

Taxonomic classification and functional annotation

Clustering of rumen microbes via WGCNA analysis

Metagenome-assembled genomes (MAGs) analysis

Untargeted metabolome sequencing and bioinformatics analysis

Targeted metabolomic analysis of flavonoids in the rumen

Western blot analysis

Microbial culture experiment

Statistical analysis

Ethics approval and consent to participate

Supplementary information

Acknowledgements

Author contributions

Data availability

Code availability

Competing interests

Footnotes

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases