Abstract
Background
In high-producing dairy systems, the average productive lifespan of cows is around 2.5–4 years. Persistent productivity and longevity are key determinants of dairy cow production performance and herd profitability. Although gastrointestinal microbiota influences dairy cow productivity, the mechanisms by which host-microbiome interactions support sustained productivity in long-lived dairy cows remain unclear. Therefore, this study integrated the metagenomics and metabolomics of the rumen and rectum, along with serum and milk metabolomics, to elucidate the potential impact of the rumen and rectum microbiota on the productivity of long-lived dairy cows.
Results
Serum alanine aminotransferase (ALT), alkaline phosphatase (ALP), total cholesterol (TC), and high-density and low-density lipoprotein cholesterol (HDL-C and LDL-C) levels in long-lived dairy cows were positively correlated with milk yield (MY) and elevated in long-lived high-yielding (LH) dairy cows, whereas insulin (INS) and glucagon (GCG) were negatively correlated with MY and higher in long-lived low-yielding (LL) dairy cows. Rumen propionate level was elevated in LH group and positively correlated with MY. The rumen microbiome, in LH cows upregulated pathways involved in amino acid, cofactor, and vitamin metabolism. LH cows’ rumen and rectum microbial networks had cohesion and vulnerability levels similar to those of LL cows and exhibited dependence on key nodes. The rumen and rectum MY-associated purine metabolites, guanosine and D-ribose-1-phosphate, mediated 65.56% and 67.55% of the significant positive effects of Acidaminococcaceae bacterium and Parabacteroides sp. on MY, respectively. Furthermore, the specific lipid metabolism-associated rumen microbiota module enhanced serum eicosapentaenoic acid (EPA) levels by modulating rumen α-linolenic acid metabolism, thereby promoting the synthesis of Pe(20:5/0:0) in milk, which positively contributed to MY.
Conclusions
This study revealed the potential contributions of the rumen and rectum microbiota to the productivity of long-lived dairy cows via purine metabolites, as well as the potential role of the rumen microbial network module in influencing productivity through α-linolenic acid metabolism, providing new insights for nutritional management strategies aimed at improving the persistent production capacity of dairy cows.
Video Abstract
Similar content being viewed by others
Background
Environmental sustainability and ethical food consumption, along with animal welfare in food production, are now major global concerns [1, 2]. Moreover, consumers and the market increasingly favor animal welfare-friendly products [3]. Productive lifespan in dairy cows refers to the length of time an animal remains in the herd which is determined primarily by involuntary culling, and is influenced by an integrated metric of production performance and health status. The average productive lifespan of dairy cows ranges from 2.5 to 4 years, which is significantly shorter than their natural life expectancy of approximately 20 years [4]. Cows with extended productive lifespans typically exhibit greater persistence in milk production [5, 6]. The primary factors that influence involuntary culling rates in dairy cow populations include reproductive performance, injuries and related factors, low milk yield (MY), and mastitis. Health-related issues account for up to 80% of culling cases [7]. Long-lived dairy cows often reflect good animal welfare standards, high production levels, and robust health status [8]. Additionally, extending productivity reduces methane emissions per kilogram of milk [9]. A longer productive lifespan also correlates with higher lifetime output and lower production and management costs. Furthermore, the average herd lifespan correlates positively with overall economic performance at the farm level [10, 11]. Extending productive lifespan leads to more environmentally friendly, sustainable, and socially acceptable dairy production. Increased lifespan also improves animal welfare and farm economic performance. The optimal productive lifespan is influenced not only by genetic selection for longevity traits but also by the interaction between genetics and the environment [12].
The rumen microbiota of dairy cows ferments carbohydrates and endogenous substrates to generate short-chain fatty acids (SCFAs)—primarily acetate, propionate, and butyrate—that meet approximately 70% of the energy requirements of ruminants [13]. Rumen microbial metabolic activity directly influences lactation performance in dairy cows, with acetate and propionate produced by microbial fermentation supplying the precursors for milk fat and lactose synthesis [14, 15]. Furthermore, rumen microbial protein serves as high-quality dietary protein that is digested and absorbed in the small intestine of dairy cows and supplies amino acids for milk protein synthesis [16]. The composition, function, and metabolism of the rumen microbiota differ significantly between high- and low-MY dairy cows [17]. Previous studies have demonstrated that the rumen microbiota, its functions, and rumen metabolites collectively account for approximately 70% of the variation in milk protein yield (MPY) [18]. Studies have shown that host-genetically influenced heritable bacteria, such as Pseudoruminococcus massiliensis and uncultured Ruminococcus spp., are associated with MY [19]. Heritable microbes enhance the abundance of Pseudobutyrivibrio via lysozymes, increasing levels of the linoleic acid-derived metabolite 9,10,13-TRIHOME and thereby improving lactation performance [20]. In addition to a role in milk production, the rumen microbiota influences mammary gland health. Rumen microbial dysbiosis triggers excessive release of lipopolysaccharide (LPS) that results in elevated circulating LPS levels, systemic inflammation, and compromised mammary blood-milk barrier function which precipitates mastitis [21]. Dysbiosis of the rumen microbiota exacerbates mammary inflammation through LPS-induced neuraminidase expression which inhibits alkaline phosphatase activity, thereby diminishing the anti-inflammatory effect of the latter. Furthermore, LPS-induced mucosal inflammation stimulates sialic acid release which pathogens, including Escherichia coli, exploit to expand the population of opportunistic pathogens which further exacerbates microbial dysbiosis and subsequent mastitis [22, 23].
Microbially mediated hindgut fermentation contributes 5–10% of total carbohydrate digestion in dairy cows and plays a pivotal role in nutrient utilization [24]. As the hindgut lacks both the keratinized rumen epithelium that confers additional impermeability and antimicrobial protection and saliva-derived bicarbonate that is necessary for pH stabilization, this region is particularly susceptible to damage from opportunistic pathogens and metabolites that arise as a consequence of microbial dysbiosis [25, 26]. Reduced abundance of beneficial hindgut bacteria, including Marasmitruncus and Butyricicoccus, impairs glutathione synthesis which leads to elevated oxidative stress in dairy cows [27]. In addition, Paraprevotella clara, P. xylaniphila and other hindgut bacteria utilize bile acids to generate secondary acids that reduce bile acid-mediated activation of circulating immune cells. This effect triggers immunosuppression during excessive lipolysis which potentially exacerbates metabolic stress and inflammation in the transition period [28]. Hindgut microbiota-induced stress and immunosuppression also may indirectly impair productive performance in dairy cows. Periparturient haptoglobin (HPT), which is a biomarker of stress and inflammation, is negatively correlated with MY, whereas postpartum HPT and β-hydroxybutyrate (BHB) levels correlate negatively with reproductive performance [29]. Moreover, pathogenic taxa within the hindgut microbiome, including Bacteroides, Porphyromonas, and Fusobacterium, may disseminate hematogenously to the uterus [30]. Dysbiosis of the hindgut microbiota which leads to compromised intestinal and mammary blood barriers, along with the amplifying effect of microbial imbalance on disease, also contribute to the development of mastitis [31, 32].
Long-lived dairy cows tend to exhibit better productivity and health status [33], which may be influenced by the gastrointestinal microbiota of these animals. Thus, long-lived cows may have a unique microbiota composition, functionality, and metabolism that affect host metabolism and phenotype. However, previous studies focusing on the relationship between the gastrointestinal microbiome of dairy cows and productivity rarely address the potential effects of microbial communities on the productivity of long-lived dairy cows with persistent productivity [34, 35]. To address this gap, we selected long-lived dairy cows with varying levels of productivity under the same management conditions and performed integrated analyses of serum parameters, rumen and rectum fermentation parameters, rumen and rectum metagenomes, as well as rumen, rectum, blood, and milk metabolomes, to compare and identify the unique production adaptation traits of long-lived cows with sustained productivity mediated by the microbiome. This study provides new insights for nutritional management strategies aimed at enhancing the lifetime productive performance of dairy cows.
Materials and methods
Animal, study design, and sample collection
The study was carried out from April to May 2024 at an intensive commercial dairy farm (37°49′ N, 106°21′ E; Lingwu, Ningxia, China). Experimental design, sampling procedures, and the measurement of body condition score (BCS) followed those described previously [36, 37]. To support a practical breeding program for long-lived dairy cows, the farm specifically maintained a cohort of long-lived individuals: Group 219, parity 5.42 ± 0.77, days in milk (DIM) 148.19 ± 41.09, and BCS 2.5–4.0. These cows were all born and raised on the same farm. All cows were provided with uniform feed (Table S1) and free access to water under identical management conditions. To minimize confounding factors affecting the gastrointestinal microbiota, cows that had received antibiotic treatment within the past 3 months were excluded. Only cows with a good reproductive history and no recorded reproductive disorders were included in this study. Among them, cows with similar parity, DIM, and BCS were further selected (Group 54; Parity 6.13 ± 0.14; DIM 130.06 ± 4.74; BCS 2.5–3.5). Ten cows with the highest and 10 cows with the lowest historical average 305-day MY were selected from this herd. Then, four cows at risk of subclinical ketosis were excluded. Rumen contents were collected using an oral stomach tube; rectum contents were sampled manually using sterile gloves; blood samples were drawn from the coccygeal vein; and milk samples were collected in a 4:3:3 volume ratio corresponding to each milking time. This study was performed in accordance with the guidelines recommended by the Administration of Affairs Concerning Experimental Animals (Ministry of Science and Technology, China, revised 2004). The protocol was approved by the Institutional Animal Care and Use Committee at Northwest A&F University.
Measurement of SCFAs and serum biomarkers concentrations
Samples were processed according to previous studies [37], and SCFAs concentration in rumen fluid and rectum contents were quantified, SCFAs were quantified by gas chromatography using an Agilent 7820A system (Agilent Technologies, Santa Clara, CA, USA) with a capillary column (AE-FFAP, 30 m × 0.25 mm × 0.33 μm; ATECH Technologies Co., Lanzhou, China).
The concentrations of serum total protein (TP), albumin (ALB), globulin (GLB), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bile acids (TBA), gamma-glutamyl transferase (γ-GGT), blood urea nitrogen (BUN), creatinine (CRE), lactate dehydrogenase (LDH), glucose (GLU), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), immunoglobulin G (IgG), immunoglobulin A (IgA), superoxide dismutase (SOD), glutathione peroxidase (GSH-PX), total antioxidant capacity (T-AOC), total glutathione (T-GSH), and catalase (CAT) in plasma were measured using an AutoAnalyzer KHB-1280 instrument (Shanghai Kehua Bio-Engineering Co., Ltd., Shanghai, China) with commercial kits (Beijing Jinhai Keyu Biotechnology Development Co., Ltd., Beijing, China). In addition, non-esterified fatty acids (NEFA), beta-hydroxybutyrate (BHB), haptoglobin (HPT), serum amyloid A (SAA), prolactin (PRL), insulin (INS), glucagon (GCG), insulin-like growth factor 1 (IGF-1), growth hormone (GH), cortisol (COR), triiodothyronine (T3), and thyroxine (T4) were analyzed using commercial enzyme-linked immunosorbent assay (ELISA) kits from Shanghai Keshun Bioengineering Institute (Shanghai, China). Tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), interleukin-6 (IL-6), interleukin-10 (IL-10), and malondialdehyde (MDA) were analyzed using commercial ELISA kits from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Protein carbonyl (PC) was measured using commercial ELISA kits from Solarbio Science & Technology Co., Ltd. (Beijing, China). All assays were conducted according to the manufacturers' instructions. All samples were measured in triplicate, and the average value was calculated.
DNA extraction, metagenome sequencing, and metagenomics data processing
Genomic DNA was extracted from each sample using the E.Z.N.A.® Soil DNA Kit (Omega Bio-tek, USA) following the manufacturer’s instructions. Metagenomic libraries with approximately 400 bp insert sizes were constructed using the NEXTFLEX Rapid DNA-Seq Kit (Bioo Scientific, USA) and sequenced on an Illumina NovaSeq 6000 platform (150 bp paired-end reads, ~ 500 bp insert size) at Majorbio Biopharm Technology Co., Ltd. (Shanghai, China). Quality control of raw Illumina reads was performed using fastp (https://github.com/OpenGene/fastp, version 0.20.0) by trimming reads shorter than 50 bp, reads with an average quality score below 20, and reads containing ambiguous (N) bases [38]. To remove host-derived sequences, high-quality reads were aligned to the bovine reference genome (UCSC bosTau8 3.7; https://doi.org/10.18129/B9.bioc. BSgenome.Btaurus.UCSC.bosTau8) using Burrows-Wheeler Aligner (BWA) (http://bio-bwa.sourceforge.net, version 0.7.17) [39]. Quality-filtered sequences were assembled using MEGAHIT (https://github.com/voutcn/megahit, version 1.1.2) [40].
Open reading frames (ORFs) of each assembled contig cluster were predicted using Prodigal (https://github.com/hyattpd/Prodigal, version 2.6.3) [41]. A non-redundant gene catalog was constructed using CD-HIT (http://www.bioinformatics.org/cd-hit/, version 4.7) [42]. High-quality reads were mapped to the non-redundant gene set using SOAPaligner (http://soap.genomics.org.cn/, version 2.21) to estimate gene abundances [43]. Taxonomic annotations were obtained by aligning predicted amino acid sequences against the National Center for Biotechnology Information (NCBI) NR database. Functional annotations were assigned based on sequence alignments against the Kyoto Encyclopedia of Genes and Genomes (KEGG) database using DIAMOND (https://github.com/bbuchfink/diamond, version 2.0.13) [44] and against the Carbohydrate-Active enZymes (CAZymes) database (http://www.cazy.org/) using HMMER (http://hmmer.org/, version 3.1b2). The abundances of taxa and functions were normalized to transcripts per million (TPM) for downstream analyses. Detailed experimental conditions and parameter settings are provided in Table S2.
Metabolomic analysis of rumen fluid, rectum contents, serum, and milk samples
Untargeted metabolomic profiling of rumen, rectum, serum, and milk samples was performed by LC–MS/MS on a Thermo UHPLC-Q Exactive HF-X system (Thermo Fisher Scientific, USA) at Shanghai Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). Metabolites from both liquid (rumen fluid, serum, milk) and solid (rectum content) matrices were extracted using optimized methanol/acetonitrile- or methanol/water-based solvent systems with internal standard correction, followed by homogenization, low-temperature ultrasonic extraction, centrifugation, solvent removal (where applicable) and reconstitution. Chromatographic separation was performed on an ACQUITY UPLC HSS T3 column and mass spectrometric detection was acquired in both positive and negative electrospray ionization modes (ESI ±) using data-dependent acquisition (DDA).
Raw LC–MS/MS data were processed using Progenesis QI (Waters Corporation, USA) for peak alignment, deconvolution, and normalization. Metabolite annotation was carried out by matching accurate mass and, when available, MS/MS fragmentation patterns against multiple public databases, including the Human Metabolome Database (HMDB; http://www.hmdb.ca/), the KEGG (https://www.genome.jp/kegg/), METLIN (https://metlin.scripps.edu/) and the Majorbio database. Unsupervised principal component analysis (PCoA) was performed to visualize overall clustering and detect potential outliers. Orthogonal partial least squares–discriminant analysis (OPLS-DA) was then used to maximize the separation between experimental groups. A sevenfold cross-validation procedure was applied to evaluate the robustness and predictive ability of each OPLS-DA model. Variables contributing most strongly to group discrimination were identified by their variable importance in projection (VIP) scores. Metabolite features with VIP > 1 and Wilcoxon rank-sum test P-value < 0.05 were considered significant and thus retained as differential metabolites. PCoA and OPLS-DA were visualized using Majorbio Cloud [45]. Metabolite Set Enrichment Analysis (MSEA) was performed using MetaboAnalyst 6.0, based on the pathway-associated metabolite sets library (https://www.metaboanalyst.ca/) [46]. The metabolite origin tracing and enrichment analysis were performed using MetOrigin 2.0 (https://metorigin.met-bioinformatics.cn/) [47]. Detailed experimental conditions and parameter settings are provided in Table S2.
Statistical analysis
Statistical analysis of production phenotypes, serum factors, and fermentation parameters was performed using the independent samples t-test in IBM SPSS Statistics 29 software, with a significance level of P < 0.05. The Wilcoxon rank-sum test was used to assess microbial α-diversity, microbial domains and centrality measures between the two groups, with P < 0.05 indicating significance. Fit the neutral community model (NCM) using the minpack.lm package in R and visualize it using ggplot2 (version 3.5.2). To assess β-diversity, PCoA was performed using the Bray–Curtis distance algorithm, and the ANOSIM (analysis of similarities) was used to evaluate the differences between groups, after verifying the assumption of equal within-group dispersion using betadisper. Use Linear Discriminant Analysis effect size (LEfSe) to compare microbial phyla, genera, and species, where an LDA > 2 and a P value < 0.05 are considered to indicate a significant difference. Also use LEfSe to compare the relative abundance of KEGG pathways, KEGG modules, KEGG enzymes, and CAZymes between two groups, where an LDA > 2 and a P value < 0.05 are considered to indicate a significant difference. Spearman correlation was used to evaluate the relationship between dairy cow production phenotypes and serum and fermentation parameters. Using GraphPad Prism 10.1 to analyze the linear correlation (Pearson’s correlation, P < 0.05) between serum factors, fermentation parameters and MY. Identification of MY-related microbiota and metabolites is performed using Spearman’s rank correlation (|r|> 0.50, P < 0.05). All correlation heatmaps and multiple correlation network analyses are performed using Spearman’s rank correlation (|r|> 0.50, P < 0.05), and visualized using the ComplexHeatmap package (version 2.22.0) and Cytoscape 3.9. Construct the symbiotic network using the SparCC method in the ggClusterNet R package [48], with |r|> 0.7, followed by network randomization testing and robustness evaluation through the removal of key nodes. Perform module analysis using “cluster fast greedy”, and simultaneously assess within-module degree (Zi) and participation coefficient (Pi) (ZiPi) along with network properties, and visualize the network using Cytoscape. Use the microeco R package (version 1.14.0) to assess network robustness, cohesion, and vulnerability [49]. Perform Procrustes analysis between microbial communities and metabolites using vegan R package (version 2.6.10). All Partial Least Squares Path Modeling (PLS-PM) models are constructed using the plspm R package (version 0.5.1).
Results
Long-lived dairy cows with different levels of persistent production ability exhibit phenotypic and metabolic differences
LH and LL cows were examined for a diverse range of parameters to assess phenotypic and metabolic differences in cows with dissimilar levels of persistent production. No significant differences were observed between LH and LL cows regarding milk components, dry matter intake (DMI), and BCS (P > 0.05; Table S3). LH cows exhibited significantly higher MY, historical average 305-day MY, milk fat yield (MFY), MPY, energy-corrected milk (ECM), fat-corrected milk (FCM), and feed efficiency (FE) compared to LL cows (P < 0.05; Table S3). Among long-lived dairy cows, serum ALT, ALP, TC, HDL-C, and LDL-C were positively correlated with MY, whereas INS and GCG were negatively correlated with MY (|r|> 0.5, P < 0.05; Fig. 1A, B). However, only TC, HDL-C, LDL-C, INS, and GCG levels were different between LH and LL cows, with the former having higher concentrations of TC, HDL-C, and LDL-C, and lower concentrations of INS and GCG (P < 0.05; Fig. 1C and Fig. S1). Regarding rumen fermentation parameters, propionate proportion was positively correlated with MY, while the acetate:propionate ratio was negatively correlated with MY (|r|> 0.5, P < 0.05; Fig. 1D, E). Additionally, LH cows exhibited a higher propionate proportion and a lower acetate:propionate ratio compared to LL cows (P < 0.05; Fig. 1F).
Relationship between serum biomarkers, rumen SCFAs and performance traits in long-lived dairy cows. A Associations between serum biomarkers and apparent factors (Spearman’s correlation). *P < 0.05; **P < 0.01. B Serum biomarkers that are linearly correlated with MY (Pearson’s correlation, P < 0.05). C Differences in serum biomarkers that associated with MY between the LH and LL groups. Error bars represent mean ± SEM. *P < 0.05. D Associations between rumen fermentation parameters and apparent factors (Spearman’s correlation). *P < 0.05. E Rumen fermentation parameters that are linearly correlated with MY (Pearson’s correlation, P < 0.05). F Differences in rumen fermentation parameters between the LH and LL groups. Error bars represent mean ± SEM. *P < 0.05
Comparison of rumen microbial community diversity and composition between LH and LL dairy cows
A total of 930,617,104 reads were generated during metagenomic sequencing, with an average of 58,163,569 ± 603,343 reads per sample (mean ± SEM). After quality control and host DNA removal, 830,790,440 reads were retained, with an average of 51,924,403 ± 961,022 reads per sample. De novo assembly subsequently yielded 14,848,586 contigs, with an average of 928,037 ± 30,421 contigs per sample (Table S4). No significant differences were observed in the Chao1, Shannon, and Simpson diversity indices for the microbiome, bacteria, viruses, archaea, and eukaryotes of LH and LL cows (P > 0.05; Fig. 2A and Fig. S2A). Moreover, PCoA showed no distinct separation between the two groups for the microbiome, bacteria, viruses, archaea, and eukaryotes at the species level (P > 0.05; Fig. 2B and Fig. S2B). The R2 value of the NCM for the rumen microbiome was higher in the LH group (0.801) than in the LL group (0.750). Furthermore, LH cows had a higher migration rate (Nm) than the LL group (2,199,503 > 1,609,371). The data from the NCM suggest that the rumen microbiome of the LH cows is less constrained by diffusion compared to LL cows (Fig. 2C). No significant differences were observed in the relative abundances of bacteria, viruses, archaea, and eukaryotes domains between the two groups (P > 0.05; Table S5).
Differences in rumen microbiota diversity and structure between LH and LL dairy cows. A The comparison of alpha diversity indices of rumen microbiota at the species level between LH and LL dairy cows. Error bars represent mean ± SEM. B PCoA of rumen microbiota at the species level based on Bray–Curtis dissimilarity (ANOSIM, 999 permutations). C Fitting of the NCM to the LH and LL groups of rumen microbiota. R2 = goodness of fit, Nm = metacommunity size (N) × migration rate (m). D, E Composition of the rumen microbiota at the phylum and genus levels. F Identification of signature species between the LH and LL groups by LDA analysis (LDA > 2, P < 0.05). *P < 0.05; **P < 0.01. G Associations between rumen microbiota and apparent factors (Spearman’s correlation). *P < 0.05; **P < 0.01; ***P < 0.001
The dominant phyla of the rumen microbiome in both groups of cows were Bacteroidota, Bacillota, and Euryarchaeota (relative abundances > 5% in both groups; Fig. 2D). The predominant genera in the rumen microbiome included Prevotella, unclassified f Lachnospiraceae, unclassified c Clostridia, unclassified o Bacteroidales, unclassified f Oscillospiraceae, and unclassified f Bacteroidaceae (relative abundances > 5% in both groups; Fig. 2E). LEfSe was performed to assess differences in the rumen microbiome at the phylum, genus, and species levels which revealed that Thermodesulfobacterota was increased significantly at the phylum level, whereas Fibrobacterota was decreased significantly in the rumen of the LH group (LDA > 2, P < 0.05; Fig. S3A). At the genus level, 50 differential bacterial genera were detected in the rumen microbiome, with 41 genera enriched in the LH group and nine genera enriched in the LL group (LDA > 2, P < 0.05; Fig. S3B). At the species level, 48 bacterial species were enriched in the LH group, while 12 bacterial species and one archaeal species were enriched in the LL group (LDA > 2, P < 0.05; Fig. 2F, Fig. S3C, and Fig. S4). Examination of the relationships between all differential microbial species and cow phenotypes revealed 53 MY-associated bacterial species (|r|> 0.5, P < 0.05; Fig. 2G and Table S6).
Co-occurrence networks and properties of rumen microbiota interactions in LH and LL dairy cows
Rumen microbial co-occurrence networks were constructed for LH and LL dairy cows by selecting species with a relative abundance > 0.01% which enabled capture of intermicrobial associations and comparison of network structures between the two groups (Fig. 3A, E). The power-law distribution observed in both networks indicates that the constructed network structures are non-random (Fig. S5). The microbial interaction networks of LH cows exhibited fewer vertices and edges, lower connectance, average degree, network diameter, mean clustering coefficient, centralization degree, and negative correlation ratio compared with LL cows, while displaying higher average path length, centralization betweenness, and centralization closeness (Table S7). Furthermore, the robustness, cohesion, and vulnerability of the networks did not differ significantly between the LH and LL groups (Fig. S6A and S6B). Based on node centrality metrics (betweenness centrality, closeness centrality, and degree), the centrality distributions were displayed (Fig. S7A). After removing the key nodes (Fig. S7B and S7C), robustness evaluation showed that the LL network performed better. These results suggest that the stability of the LH rumen microbial network is more dependent on key nodes than of the LL group.
Co-occurrence networks of rumen microbiota in LH and LL dairy cows and their associations with phenotypic traits. A Co-occurrence network of rumen microbiota in LH dairy cows. The color of connection lines between two nodes represents a positive (green) or a negative (gray) correlation (|r|> 0.70). B Zi and Pi are used to identify module and network hubs in the LH group. C Spearman’s correlations between network modules in the LH group and apparent factors (|r|> 0.50, P < 0.05). D Relative abundance differences of LH group network modules between the two groups. Error bars represent mean ± SEM. *P < 0.05; **P < 0.01. E Co-occurrence network of rumen microbiota in LL dairy cows. The color of connection lines between two nodes represents a positive (green) or a negative (gray) correlation (|r|> 0.70). F Zi and Pi are used to identify module and network hubs in the LL group. G Spearman’s correlations between network modules in the LL group and apparent factors (|r|> 0.50, P < 0.05). H Relative abundance differences of LL group network modules between the two groups. Error bars represent mean ± SEM. **P < 0.01
Modular analysis was performed using the cluster fast greedy algorithm, and both Zi and Pi were calculated to identify module and network hubs. Neither network exhibited module or network hubs (Table S8), although the LL network contained more connectors (Fig. 3B and F). Eleven microbial modules were identified in the LH cows, but only module 5 exhibited positive correlations with production phenotypes (MY, ECM, FCM, and MFY; |r|> 0.5, P < 0.05). In addition, module 5 was positively correlated with serum INS and GCG concentrations (|r|> 0.5, P < 0.05). Modules 3, 4, and 9 were positively correlated with the acetate:propionate ratio (|r|> 0.5, P < 0.05), whereas modules 6 and 11 correlated with lipid metabolism-related indicators (TC, HDL-C, and LDL-C; |r|> 0.5, P < 0.05; Fig. 3C). Across all modules, only modules 2, 5, and 9 demonstrated significant differences in relative abundance between the LH and LL groups, with all three modules were more abundant in the LH group (P < 0.05; Fig. 3D). Eight microbial modules were delineated in the LL group. Modules 3 and 5 were associated negatively with acetate:propionate ratio and propionate proportion (|r|> 0.5, P < 0.05), whereas Module 8 was negatively correlated with lipid metabolism-related indicators (TC, HDL-C and LDL-C; |r|> 0.5, P < 0.05; Fig. 3G). Among all modules identified in the LL group, only Module 6 exhibited differential abundance between the LH and LL groups and was more abundant in the LH group (P < 0.05; Fig. 3H).
Functional differences in the rumen microbiome between LH and LL cows
Notable differences in the functional characteristics of the rumen microbiome were observed between LH and LL cows. PCoA analysis based on KEGG level 3 pathways did not show distinct separation between LH and LL cows (P > 0.05; Fig. 4A). However, “environmental information processing” was enriched significantly in the LH group at KEGG level 1 pathways, whereas “human diseases” and “organismal systems” were enriched in LL cows (LDA > 2, P < 0.05; Table S9). Moreover, “cell motility”, “amino acid metabolism”, and “global and overview maps” were enriched in LH cows at KEGG level 2 pathways, whereas “excretory system”, “digestive system”, “information processing in viruses”, “transport and catabolism”, and “environmental adaptation” were enriched in the LL group (LDA > 2, P < 0.05; Fig. 4B). 12 pathways differed significantly between the LH and LL cows at KEGG level 3, of which 10 and two were enriched in the LH and LL cows, respectively (LDA > 2, P < 0.05; Table S9). It is worth noting that, the one “global and overview maps” pathway (biosynthesis of amino acids), one “carbohydrate metabolism” pathway (C5-branched dibasic acid metabolism), three “amino acid metabolism” pathways (arginine biosynthesis, histidine metabolism, and valine, leucine, and isoleucine biosynthesis), three “metabolism of cofactors and vitamins” pathways (one carbon pool by folate, pantothenate and CoA biosynthesis, and riboflavin metabolism), two “cell motility” pathways (flagellar assembly and bacterial chemotaxis) were LH-enriched pathways (LDA > 2, P < 0.05; Fig. 4C). Moreover, two “carbohydrate metabolism” pathways (starch and sucrose metabolism and glycolysis/gluconeogenesis) were LL-enriched pathways (LDA > 2, P < 0.05; Fig. 4C).
Differential metabolic pathways identified in the rumen metagenomes of LH and LL dairy cows. A PCoA of rumen microbiota functions at KEGG Level 3, based on Bray–Curtis dissimilarity, was performed to compare LH and LL groups. Dissimilarity between the LH and LL groups was assessed using ANOSIM with 999 permutations. B Fold changes in LH versus LL groups for significantly enriched KEGG Level 2 pathways. *Means LDA > 2 and P < 0.05. C Fold changes in LH versus LL groups of KEGG level 3 pathways within the categories global and overview maps, carbohydrate metabolism, amino acid metabolism, metabolism of cofactors and vitamins, cell motility, and energy metabolism. *Means LDA > 2 and P < 0.05
Comparison of KEGG modules revealed that 24 KEGG modules were enriched in LH cows, whereas four modules were augmented in LL cows (LDA > 2, P < 0.05; Fig. S8A). Interestingly, only one module, M00008 (Entner-Doudoroff pathway: glucose-6P = > glyceraldehyde-3P + pyruvate), associated with pyruvate and propionate production, was found to be enriched in the LH group (LDA > 2, P < 0.05). Comparative analysis at the CAZymes class level revealed the enrichment of carbohydrate-binding modules (CBMs) in LL cows. 37 differential CAZymes were identified at the family level, including 18 glycoside hydrolases (GHs), seven glycosyltransferases (GTs), three carbohydrate esterases (CEs), two polysaccharide lyases (PLs), and one auxiliary activity (AA), all of which were enriched in the rumen of LH group cows (LDA > 2, P < 0.05). Additionally, two GHs, three CBMs, and one AA were enriched in the rumen of LL group cows (LDA > 2, P < 0.05; Fig. S8B and S8C). Spearman correlation analysis indicated that 53 MY-associated microbiota may exert their functions through these specific KEGG modules and CAZymes (|r|> 0.5, P < 0.05; Fig. S9A and S9B).
The multiple Spearman rank correlation network further elucidated the relationships among 53 MY-related microbiota, 10 differential microbial functions, and 6 production phenotypes (Fig. S10A). Specifically, 42 MY-related bacteria positively correlated (|r|> 0.5, P < 0.05) with eight microbial functions related to amino acid metabolism, carbohydrate metabolism, and cofactors and vitamins metabolism (biosynthesis of amino acids, arginine biosynthesis, histidine metabolism, valine, leucine, and isoleucine biosynthesis, C5-branched dibasic acid metabolism, one carbon pool by folate, pantothenate and CoA biosynthesis, and riboflavin metabolism). These eight rumen microbial functions were positively correlated (|r|> 0.5, P < 0.05) with six production phenotypes (MY, ECM, FCM MFY, MPY, and FE). Additionally, the 42 MY-related bacteria also showed significant positive associations with KEGG module M00008 (|r|> 0.5, P < 0.05), which is involved in pyruvate and propionate biosynthesis. This module correlated positively with rumen propionate proportion (|r|> 0.5, P < 0.05), which exhibited positive correlations with six production phenotypes (MY, ECM, FCM MFY, MPY, and FE; |r|> 0.5, P < 0.05; Fig. S10B).
An integrated analysis of key KEGG pathways, CAZymes, KEGG modules, and MY-related bacteria was performed to comprehensively characterize carbohydrate metabolism in the LH cows (LDA > 2, P < 0.05). We observed metabolic functional reprogramming in the rumen microbiome of LH cows which was characterized by reductions in starch and sucrose metabolism and glycolysis/gluconeogenesis (ko00500 and ko00010). Decreased abundance of GH119 which hydrolyzes amylose was accompanied by increased abundance of GH13_9 which degrades amylopectin. Concurrently, the abundance of GH5_2 that targets soluble oligosaccharides increased, whereas the prevalence of the cellulose-binding module CBM63 declined. The degradative capacity of the rumen microbiome in LH cows toward specific substrates, notably starch, was diminished which led to a concomitant decrease in glycolytic flux, whereas the glycolysis-independent propionate-producing module M00008 became enriched. These findings indicate that the rumen microbial community in LH cows has become better adapted to a high-concentrate diet by preferentially fermenting rapidly soluble substrates (Fig. 5A). Moreover, the composition of lactate-producing and lactate-utilizing bacterial populations in the rumen of LH cows was reorganized markedly. Among taxa associated with MY, Anaerolactibacter massiliensis, Galactobacillus timonensis, Olsenella sp., Prevotella lacticifex, and Lactimicrobium massiliensis, all of which are implicated in lactate production, were significantly enriched, whereas Aristaeella lactis declined. Concurrently, the lactate-utilizing taxa Veillonellaceae bacterium and Selenomonadales bacterium increased in abundance. Furthermore, the composition of bacterial taxa involved in succinate production and utilization was reorganized: Succinate-producing bacteria Prevotella sp. AGR2160 was significantly enriched whereas Bacteroidales bacterium, Paludibacteraceae bacterium, and Rikenellaceae bacterium declined. In parallel, the succinate-utilizing taxon Succiniclasticum sp. increased in abundance. In addition, a methane-producing bacterium, Methanobrevibacter sp. UBA188, was enriched in the rumen of LL cows (Fig. 5A). KEGG enzymes mediating the biosynthesis of amino acids (isoleucine, leucine, valine, histidine, and arginine), of C5-branched dibasic acid (itaconate), and key cofactors and vitamins (coenzyme A, vitamin B12 coenzyme, and s-adenosylmethionine) were also significantly enriched in LH cows (LDA > 2, P < 0.05; Fig. 5B and Table S10).
Difference in functional capacities of the rumen microbiome between LH and LL dairy cows. A Reconstruction of the metabolic pathways associated with SCFAs biosynthesis and methanogenesis (LDA > 2, P < 0.05). B Reconstruction of differentially enriched metabolic pathways. The red text represents significantly enriched KEGG enzymes (LDA > 2, P < 0.05). EC = Enzyme Commission number
Rumen metabolic characteristics differed significantly between LH and LL cows
OPLS‐DA was employed to discriminate metabolic profiles between the LH and LL groups (Fig. 6A). A total of 100 metabolites were identified as significantly different, with 50 upregulated and 50 downregulated in LH cows compared to LL cows (VIP > 1, P < 0.05; Fig. 6B). Subsequent correlation analysis between these differential metabolites and MY revealed 61 MY‐associated metabolites, of which 35 exhibited positive associations and 26 negative associations (|r|> 0.5, P < 0.05; Table S11). MSEA revealed that only the purine metabolism pathway was significantly enriched among the top 10 enriched metabolic pathways (P < 0.05; Fig. 6C and Table S12). Procrustes analysis revealed significant concordance between the rumen microbiome and the rumen metabolome (P < 0.001; Fig. 6D). Tracing the origins of significantly different metabolites using MetOrigin 2.0, we identified 16 host-derived metabolites, 23 microbiota-derived metabolites, 17 drug-related metabolites, 54 food-related metabolites, and two environment-related metabolites. Of these, 15 metabolites were identified as microbiota-host co-metabolites (Fig. S11A–B and Table S13). Notably, purine metabolism emerged as the most significantly enriched pathway in microbiota-host co-metabolism pathways (P < 0.001; Fig. S11C). Two co-metabolized purine metabolites, guanosine and adenosine, were identified as MY-associated metabolites within the microbiota-host co-metabolized purine metabolism pathway (Fig. S11D and Table S14). Furthermore, 35 MY-associated bacterial taxa exhibited significant correlations with guanosine, whereas 46 MY-associated bacterial taxa correlated with adenosine (|r|> 0.5, P < 0.05; Fig. 6E). Bacterial species that contributed > 0.1% to purine metabolism and which were associated with MY were selected for mediation effect analysis (Table S15). The results indicate that guanosine mediates 65.56% of the significant positive effect of Acidaminococcaceae bacterium on MY (P < 0.05; Fig. S12). In contrast, no significant mediating effects of adenosine were observed in the microbial influence on MY (P > 0.05; Fig. S12).
Rumen metabolites and their relationships with rumen microbiota. A OPLS-DA of rumen metabolites between LH and LL groups of dairy cows. B Volcano plot of differential rumen metabolites (VIP > 1, P < 0.05). C The top 10 enriched pathways from MSEA of rumen metabolites. D Procrustes analysis to assess the concordance between rumen microbiome structure and rumen metabolite structure. E Spearman’s rank correlation between MY-associated rumen microbiota and MY-associated rumen metabolites. *P < 0.05; **P < 0.01; ***P < 0.001
Comparison of rectum microbial community diversity and composition between LH and LL dairy cows
A total of 900,615,304 reads were generated during the metagenomic sequencing, with an average of 58,788,456 ± 608,648 reads per sample (mean ± SEM). After quality control and host DNA removal, 821,935,230 reads were retained, with an average of 51,370,952 ± 955,931 reads per sample. Subsequently, de novo assembly generated 14,034,542 contigs, with an average of 877,159 ± 24,216 contigs per sample (Table S16). No significant differences (P > 0.05) were observed between LH and LL cows for the microbiome, bacteria, viruses, archaea, and eukaryotes in either α-diversity (Chao1, Shannon, Simpson; Fig. 7A and Fig. S13A) or β-diversity (Bray–Curtis PCoA; Fig. 7B and Fig. S13B). The NCM revealed that the overall model fit for the rectum microbiota was higher in the LH group (R2 = 0.465) than in the LL group (R2 = 0.427). Moreover, Nm was higher in the LH group (1,493,930 vs. 1,424,961), indicating reduced dispersal limitation and greater adaptability of the LH rectum microbial community (Fig. 7C). The relative abundance of archaea in the rectum microbiota of LH cows was significantly reduced compared to LL cows (P < 0.05; Table S17). The predominant phyla in the rectum microbiota were Bacillota, Bacteroidota, Spirochaetota, Uroviricota, Actinomycetota, and Euryarchaeota (relative abundances > 1% in both groups; Fig. 7D). The predominant genera in the rectum microbiota were unclassified f Oscillospiraceae, unclassified c Clostridia, unclassified o Bacteroidales, unclassified f Lachnospiraceae, and unclassified f Bacteroidaceae (relative abundances > 5% in both groups; Fig. 7E). LEfSe analysis was performed to evaluate the differences at the phylum, genus, and species levels in the rectum microbiota. At the phylum level, Actinomycetota was enriched in the LL group (LDA > 2, P < 0.05; Fig. S14A). At the genus level, three differential genera were detected in the rectum microbiota, with Parabacteroides and Parafannyhessea enriched in LH cows, and Bifidobacterium enriched in LL cows (LDA > 2, P < 0.05; Fig. S14B). At the species level, 11 microbial species differed significantly between the two groups (LDA > 2, P < 0.05; Fig. 7F). Three bacterial species (Parabacteroides sp., Parabacteroides distasonis, and Parafannyhessea umbonata) and two archaeal species (archaeon and Methanobrevibacter olleyae) were enriched in LH cows, whereas three bacterial species (Bifidobacterium pseudolongum, unclassified g Bifidobacterium, and uncultured Bifidobacterium sp.) and three archaeal species (Candidatus Methanomethylophilaceae archaeon, Methanobrevibacter millerae, and Methanobrevibacter sp.UBA188) were enriched in LL cows (LDA > 2, P < 0.05; Fig. S14C and S14D). The relationships between all differential microbial species and the cow phenotypes indicated that seven MY-associated microbial species were identified (|r|> 0.5, P < 0.05; Fig. 7G and Table S18). No correlations were found between rectum fermentation parameters and production phenotypes (|r|< 0.5, P > 0.05; Fig. 7H). Furthermore, no significant differences were observed in any rectum fermentation parameters between the LH and LL cows (P > 0.05; Fig. 7I).
Differences in rectum microbiota diversity, structure, and SCFAs production between LH and LL dairy cows. A The comparison of alpha diversity indices of rectum microbiota between LH and LL dairy cows. Error bars represent mean ± SEM. B PCoA of rectum microbiota at the species level based on Bray–Curtis dissimilarity (ANOSIM, 999 permutations). C Fitting of the NCM to the LH and LL groups of rectum microbiota. R2 = goodness of fit, Nm = metacommunity size (N) × migration rate (m). D, E Composition of the rectum microbiota at the phylum and genus levels. F Identification of signature species between the LH and LL groups by LDA analysis. LDA > 2 and P < 0.05. *P < 0.05; **P < 0.01. G Associations between rectum microbiota and apparent factors (Spearman’s correlation). *P < 0.05; **P < 0.01; ***P < 0.001. H Associations between the rectum fermentation parameters and production phenotypes (Spearman’s correlation). I Differences in rectum fermentation parameters between LH and LL groups. Error bars represent mean ± SEM
Co-occurrence networks and network properties of rectum microbiota interactions in LH and LL cows
Rectum microbial symbiosis networks were constructed to capture the interrelationships within the rectum microbiota and to analyze the differences between the LH and LL networks (relative abundances > 0.01% in both groups; Fig. 8A and E). The power-law distribution also observed in both networks indicates that the constructed network structures are non-random (Fig. S15). LH cow network exhibited lower values of vertices, edges, connectance, average degree, centralization degree, and negative correlation ratio compared to the LL cow network. However, the LH network exhibited higher values for average path length, diameter, mean clustering coefficient, centralization betweenness, and centralization closeness (Table S19). Furthermore, the robustness of the LH network was consistent with the LL network (Fig. S16A), with no significant differences observed in cohesion and vulnerability (P > 0.05; Fig. S16B). The node centrality distributions were displayed (Fig. S17A). After removing the key nodes (Fig. S17B and S17C), robustness evaluation showed that the LL network performed better. These results suggest that the stability of the LH rectum microbial network is more dependent on key nodes.
Co-occurrence networks of rectum microbiota in LH and LL dairy cows and their associations with phenotypic traits. A Co-occurrence network of rectum microbiota in LH dairy cows. The color of connection lines between two nodes represents a positive (green) or a negative (gray) correlation (|r|> 0.70). B Zi and Pi are used to identify module and network hubs in the LH group. C Spearman’s correlations between network modules in the LH group and apparent factors (|r|> 0.50, P < 0.05). D Relative abundance differences of LH group network modules between the two groups. *P < 0.05. E Co-occurrence network of rectum microbiota in LL dairy cows. The color of connection lines between two nodes represents a positive (green) or a negative (gray) correlation (|r|> 0.70). F Zi and Pi are used to identify module and network hubs in the LL group. G Spearman’s correlations between network modules in the LL group and apparent factors (|r|> 0.50, P < 0.05). H Relative abundance differences of LL group network modules between the two groups
The "cluster fast greedy" method was employed for modular analysis which demonstrated that neither the LH nor LL network exhibited module or network hubs (Table S20), although the LL network contained more connectors (Fig. 8B and F). In the LH group, 16 microbial modules were identified, with only module 8 showing a negative correlation with LH production phenotypes (MY, ECM, FCM and MFY; |r|> 0.5, P < 0.05). This module was also negatively correlated with INS and GCG serum concentrations (|r|> 0.5, P < 0.05). Additionally, Module 15 and 16 were positively correlated with the rectum butyrate proportion (|r|> 0.5, P < 0.05), and Module 5 was positively correlated with lipid metabolism-related indicators (TC, HDL-C, and LDL-C; |r|> 0.5, P < 0.05) in serum (Fig. 8C). Among all modules, only Module 2 and 6 exhibited differences in relative abundance between the two groups, both of these modules were enriched in LL cows (P < 0.05; Fig. 8D). In the LL group, 13 microbial modules were constructed, with Module 4 negatively correlated with serum ALP concentration (|r|> 0.5, P < 0.05) and Module 8 and 12 positively correlated with lipid metabolism-related indicators (LDL-C/HDL-C and LDL-C; |r|> 0.5, P < 0.05) in the serum (Fig. 8G). No modules in the LL group differed in abundance between LH and LL cows (P > 0.05; Fig. 8H).
Functional differences in the rectum microbiomes between LH and LL cows
PCoA analysis of KEGG level 3 pathways revealed no significant separation between LH and LL cows (P > 0.05; Fig. S18A). Additionally, no differential pathways were identified between the two groups in KEGG pathway at levels 1, 2, and 3 (LDA < 2, P > 0.05; Fig. S18B-C and Table S21). Two KEGG modules (M00049: Adenine ribonucleotide biosynthesis, IMP = > ADP, ATP and M00052: Pyrimidine ribonucleotide biosynthesis, UMP = > UDP/UTP, CDP/CTP) and one KEGG name (Trimeric autotransporter adhesin) were enriched in the LL group, while one KEGG module (M00844: Arginine biosynthesis, ornithine = > arginine) was enriched in the LH group (LDA > 2, P < 0.05; Fig. S19A–19B). A comparison of CAZymes at the class level showed no significant differences between the LH and LL cows (LDA < 2, P > 0.05; Fig. S19C). However, 15 differential CAZymes were identified at the family level, with 10 enriched in the LH group (8 GHs, 1 CBM, 1 GT) and 5 enriched in the LL group (3 GHs, 1 CBM, 1 GT; LDA > 2, P < 0.05; Fig. S19D). Correlation analysis revealed that seven rectum microbiota species associated with MY might exert functions through these specific KEGG modules and CAZymes (|r|> 0.5, P < 0.05; Fig. S20A and S20B). Multiple Spearman rank correlation network further revealed that the MY-associated species M. millerae, M. sp. UBA188, B. pseudolongum, unclassified g Bifidobacterium, and uncultured Bifidobacterium sp. were positively correlated with KEGG module M00049, whereas Parabacteroides sp. and P. distasonis exhibited a negative correlation (|r|> 0.5, P < 0.05). Among the seven MY-associated microbiota, three species (M. millerae, M. sp. UBA188, and uncultured Bifidobacterium sp.) were positively correlated with KEGG module M00052, whereas Parabacteroides sp. and P. distasonis showed a negative correlation (|r|> 0.5, P < 0.05). The M00049 and M00052 differential microbial KEGG modules were negatively correlated with six production phenotypes (MY, ECM, FCM, MFY, MPY, and FE; |r|> 0.5, P < 0.05; Fig. S20C).
Rectum metabolic characteristics differ between LH and LL cows
OPLS-DA was applied to differentiate the rectum metabolites between LH and LL cows (Fig. 9A). A total of 69 metabolites differed significantly between the groups, with 22 metabolites upregulated and 47 downregulated (VIP > 1, P < 0.05; Fig. 9B). Correlation analysis between the differential metabolites and MY was conducted to identify MY-associated metabolites, identifying 36 MY-related metabolites, 17 of which were upregulated and 19 downregulated (|r|> 0.5, P < 0.05; Table S22). The differential metabolites were significantly enriched in the purine metabolism, aspartate metabolism, and nicotinate and nicotinamide metabolism pathways (P < 0.05; Fig. 9C and Table S23). Procrustes analysis found no significant concordance between the structure of the rectum microbiome and the rectum metabolome (P > 0.05; Fig. 9D). Differential metabolites analyzed using MetOrigin 2.0 revealed 16 host-derived metabolites, 19 microbiota-derived metabolites, 16 drug-related metabolites, 48 food-related metabolites, and three environment-related metabolites. Among these, 14 metabolites were co-metabolized by both microbiota and host (Fig. S21A–B and Table S24). Notably, purine metabolism was the most significantly enriched co-metabolism pathway shared by the microbiota and host (P < 0.01; Fig. S21C and Table S25). Two MY-associated metabolites, D-ribose-1-phosphate and Xanthosine, were enriched in this pathway (Fig. S21D). Furthermore, seven MY-associated rectum species were correlated with D-ribose-1-phosphate, while four MY-associated rectum species showed an association with Xanthosine (|r|> 0.5, P < 0.05; Fig. 9E). Microbial species contributing > 0.1% to the metagenomic purine metabolism and associated with MY were selected for mediation effect analysis (Table S26). The mediation effect analysis revealed that D-ribose-1-phosphate mediates 67.55% of the significant positive effect of Parabacteroides sp. on MY. This metabolite additionally mediates 104.90% and 99.50% negative significant effects of unclassified g Bifidobacterium and B. pseudolongum, respectively, on MY (P < 0.05). The mediation effects of xanthosine in the impact of microbiota on MY were all not significant statistically (P > 0.05; Fig. S22.
Rectum metabolites and their relationship with rectum microbiota. A OPLS-DA of rectum metabolites between LH and LL groups of dairy cows. B Volcano plot of differential rectum metabolites (VIP > 1, P < 0.05). C The top 10 enriched pathways from MSEA of rumen metabolites. D Procrustes analysis to assess the concordance between rectum microbiome structure and rectum metabolite structure. E Spearman’s rank correlation between MY-associated rectum microbiota and MY-associated rectum metabolites. *P < 0.05; **P < 0.01; ***P < 0.001
Metabolomic profiling of serum and milk in LH and LL cows and associations between differential metabolites and microbiota
OPLS-DA was applied to distinguish the serum and milk metabolite profiles of the LH and LL cows (Fig. S23A and S23B). A total of 81 serum metabolites exhibited significant differences, of which 42 were up-regulated and 39 were down-regulated. In milk samples, 56 metabolites exhibited significant differences, with 44 up-regulated and 12 down-regulated (VIP > 1, P < 0.05; Fig. S23C and S23D). Differential metabolites were subjected to correlation analysis with MY to identify MY-associated metabolites. In serum, 62 MY-associated metabolites were identified, of which 25 were up-regulated and 37 down-regulated, whereas 28 MY-associated were identified in milk, including 20 up-regulated and eight down-regulated metabolites (|r|> 0.5, P < 0.05; Table S27 and S28). Differential serum metabolites were significantly enriched in the valine, leucine and isoleucine degradation and beta oxidation of very long chain fatty acids pathways. Differential milk metabolites were significantly enriched in the betaine metabolism, phospholipid biosynthesis, and methionine metabolism pathways (P < 0.05; Fig. 9A, B and Table S29 and S30). Procrustes analysis indicated that the rumen microbiome structure did not exhibit significant concordance with serum metabolome structure (P > 0.05; Fig. 10C), but did show significant concordance with milk metabolome structure (P < 0.05; Fig. 10D). In contrast, rectum microbiome structure demonstrated no significant concordance with either serum or milk metabolome structure (P > 0.05; Fig. 10E and F). Most MY-associated rumen and rectum microbiota exhibit significant correlations with MY-associated serum and milk metabolites. (|r|> 0.5, P < 0.05; Fig. S24A, B and Fig. S25A, B).
Enriched metabolic pathways in the serum and milk metabolomes, and the consistency comparison between rumen and rectum microbial microbiomes and the rumen and rectum metabolomes. A and B The top 10 enriched pathways from MSEA of serum and milk metabolites. C, D Procrustes analysis to assess the concordance between rumen microbiome structure with serum and milk metabolite structures. E, F Procrustes analysis to assess the concordance between rectum microbiome structure with serum and milk metabolite structures
Comprehensive analysis of multi-omics and phenotypic associations
Venn diagram analysis was used to evaluate potential metabolite transfer among the rumen, rectum, serum, and milk metabolomes in LH dairy cows (Fig. S26A). Subsequent correlation analysis of these shared metabolites revealed a significant positive association between cnidilide levels in the rectum and serum (|r|> 0.5, P < 0.05; Fig. S26A). Furthermore, LH cows exhibited upregulated cnidilide levels in both the rectum and serum (VIP > 1, P < 0.05; Fig. S26A). PLS-PM was applied to explore relationships among rectum MY-associated microbiota, rectum and serum cnidilide levels, and MY (Fig. S26B). The model indicated that rectum MY-associated microbiota exerted a non-significant negative effect on rectum cnidilide (P > 0.05), whereas rectum cnidilide exerted a significant positive effect on serum cnidilide (P < 0.01). However, serum cnidilide did not significantly influence MY (P > 0.05). These results suggested that the rectum MY-associated microbiota contributes minimally to the transfer of cnidilide from the rectum to the serum, and that the cnidilide transferred into the serum did not significantly affect MY. Multilevel correlation networks stratified by metabolite class were constructed to examine interactions among rumen, rectum, serum, and milk metabolites. This analysis demonstrated that, whether originating from rumen or rectum metabolites, only those metabolites classified as lipids and lipid-like molecules or organic acids and derivatives formed multilevel correlation networks with serum and milk metabolites (Fig. 11A, B). PLS-PM was constructed to evaluate the relationships among rumen microbial modules related to lipid metabolism in LH cows, MY-associated rumen metabolites, MY-associated serum metabolites, MY-associated milk metabolites, and MY (Fig. 11C). Module 6, which was negatively correlated with serum lipid metabolism indices (TC, HDL-C, and LDL-C) and includes MY-associated species (Hallella. mizrahii, H. absiana, H. multisaccharivorax, and Prevotella sp. AGR2160), had a significant positive effect on rumen GPEtn(12:0/18:3) (P < 0.01). Moreover, rumen GPEtn(12:0/18:3) positively influenced serum eicosapentaenoic acid (EPA; P < 0.05), and serum EPA positively influenced milk Pe(20:5/0:0) (P < 0.05). Finally, milk Pe(20:5/0:0) had a significant positive impact on MY (P < 0.05). All other modeled paths were non-significant (P > 0.05). These findings suggest that the rumen microbiota in LH cows may enhance the synthesis of GPEtn(12:0/18:3), thereby increasing the availability of α-linolenic acid in the small intestine for absorption. Once absorbed, α-linolenic acid is transported to the liver and converted to EPA which enters the circulation and serves as a precursor for milk Pe(20:5/0:0) biosynthesis. Elevated milk Pe(20:5/0:0) levels may play a role in supporting the MY of long-lived dairy cows.
Integrated omics analysis based on metagenomics and metabolomics. A Multiple correlation analysis of rumen, serum, and milk metabolites. Edges between nodes represent Spearman correlations: green lines indicate positive correlations and gray lines indicate negative correlations (|r|> 0.50, P < 0.05). B Multiple correlation analysis of rectum, serum, and milk metabolites. Edges between nodes represent Spearman correlations: green lines indicate positive correlations and gray lines indicate negative correlations (|r|> 0.50, P < 0.05). C The PLS-PM module provided an integrated analysis of the relationships among rumen microbial module 6, rumen GPEtn(12:0/18:3), serum EPA, milk Pe(20:5/0:0), and MY. Numbers above arrows represent standardized effect coefficients. R.2 = coefficient of determination; GoF = goodness-of-fit index. EPA = eicosapentaenoic acid; MY = milk yield. *P < 0.05; **P < 0.01
PLS-PM models were built to explore relationships among microbial composition and function, microbial and host metabolomes, and production phenotypes. Rumen microbes had a significant positive effect on rumen microbial functions (P < 0.001), and rumen metabolites had a significant positive impact on rumen SCFAs (P < 0.01). Serum metabolites exhibited a significant negative effect on serum factors (P < 0.01). All other modeled paths were non-significant. The strongest positive influences on milk production performance were exerted by rumen microbes and rumen metabolites, whereas serum metabolites had the strongest negative effect (Fig. S27A and S27B). Rectum microbes and rectum microbial functions had significant positive effect on rectum metabolites (P < 0.01). Both rectum metabolites and rectum microbial functions exerted strong negative effects on milk production performance (Fig. S27C and S27D).
Discussion
The production lifespan of dairy cows is influenced by numerous factors, including milk production, health status, and reproductive performance [50]. These factors are associated closely with the structure, composition, and function of the cow gastrointestinal microbiota [22, 51, 52]. This study integrated rumen and rectum metagenomes with rumen, rectum, serum, and milk metabolomes to investigate the microbial composition and structural characteristics of dairy cows with sustained production capacity and the potential roles of these features in maintaining productive performance. Multi-omics approaches were employed to assess microbial composition, functionality, metabolism, and host metabolism contributions to variations in dairy cow production traits.
Blood metabolic biomarkers and gastrointestinal fermentation parameters in dairy cows serve as direct indicators of physiological status. Increased mobilization of adipose reserves in LH cows heightens risk of hepatic lipid accumulation and infiltration which may lead to hepatocellular damage. Serum concentrations of hepatic injury biomarkers, including ALT and ALP, correlate strongly with lipid mobilization and the degree of ketosis during lactation [17, 53]. However, no significant differences in ALT and ALP were observed between the two groups in this study, possibly reflecting the robust health status of cows with sustained productivity [54, 55]. Furthermore, none of the serum markers of oxidative stress, inflammation, immune response, or energy metabolism correlated with MY, and the absence of intergroup differences in these parameters further supported the health status of both LH and LL cows in the study. Moreover, serum concentrations of TC, HDL-C, and LDL-C were positively correlated with MY and were significantly elevated in the LH group. Cholesterol, as a structural component of mammary epithelial cell membranes and a precursor for milk fat synthesis [56], may undergo increased synthesis and transport during peak lactation. LDL-C serves as the primary carrier of cholesterol into the mammary gland [57]. HDL mediates reverse cholesterol transport from peripheral tissues to the liver, thereby maintaining circulating cholesterol homeostasis and supporting vascular health in aging cows. The elevated HDL-C levels in LH cows partially reflect enhanced HDL-mediated cholesterol transport [58, 59]. Additionally, INS and GCG concentrations were negatively correlated with MY and were significantly decreased in the LH group. The postpartum decline in serum INS forms part of a homeorhetic metabolic regulation mechanism that redirects nutrients, particularly glucose, toward mammary tissue, which is relatively INS-independent [60]. Increased fat mobilization also impairs INS sensitivity and responsiveness [61, 62]. However, no significant differences in circulating NEFA or BHB were observed between the two groups, and the reduced serum INS levels in LH cows may reflect elevated lactational demands. High-yielding dairy cows in mid-lactation may preferentially utilize fatty acids and ketone bodies as energy substrates [63, 64], thereby reducing glucose requirements. This metabolic adaptation may explain the sustained low serum GCG levels observed in LH cows. Previous studies have similarly documented decreased GCG concentrations in high-producing dairy cows [65]. SCFAs generated by ruminal fermentation constitute the primary energy source for ruminants [66]. Total ruminal SCFA and propionate concentrations are elevated in high MPY cows [52]. In LH cows, rumen propionate concentration was positively correlated with MY and were significantly higher than in LL cows. Propionate undergoes gluconeogenesis to glucose in the liver and supplies 60–74% of the carbon for glucose synthesis in dairy cows [67]. Glucose absorbed across the basolateral membrane of mammary epithelial cells is utilized for lactose synthesis, and lactose, as the principal osmolyte in milk, maintains mammary osmotic pressure to support milk secretion [68, 69]. Alterations in serum metabolic markers and enhanced rumen propionate availability satisfy the requirements for lactation in LH cows.
Differential occupation and dominance of available ecological niches by distinct species drive changes in the alpha diversity of the rumen microbiota in cows with varying production performance, with microbial richness significantly lower in cows exhibiting high MPY, enhanced SCFA production, and superior FE [52, 70, 71]. However, no significant differences in the overall rumen microbiome structure were observed between LH and LL cows in this study. Sustained high productivity may not be attributable to restructuring of the overall rumen microbiome in cows demonstrating sustained production capacity. However, LH cows harbored a greater number of core keystone taxa under positive selection in the rumen microbiome, which suggests that adaptive selection of specific microbial populations may underlie sustained high productivity. The rumen microbial co-occurrence network in LH cows had fewer connections and relied on a limited set of key nodes. This core-node-dominated microbial interaction network may be more efficient while maintaining approximately similar network robustness, facilitating the metabolic functions of specific adaptive microorganisms [72]. An MY-associated rumen microbial module was detected in LH cows, centered on Acidaminococcaceae bacterium. This module also included four additional MY-associated taxa from the Acidaminococcaceae family, i.e., Acidaminococcus provencensis, Acidaminococcus sp., A. fermentans, and uncultured Acidaminococcus sp. Members of the Acidaminococcaceae family ferment amino acids to produce acetate and butyrate [73]. Acetate produced by these bacteria not only generates ATP for bovine metabolic demands but also is harnessed by mammary epithelial cells for de novo fatty acid synthesis [14, 74]. The butyrate that is produced contributes to the maintenance of rumen epithelial integrity and immune homeostasis [75, 76]. Furthermore, A. fermentans ferments trans-aconitate, thereby reducing toxic tricarballylate accumulation in the rumen of LH cows that are fed long-term high-concentrate diets and maintaining rumen environmental homeostasis [77]. In addition, the abundance of the methanogenic archaeon Methanobrevibacter sp. UBA188 was found to be decreased in the rumen of LH cows. A similar decline in rumen methanogen abundance has also been reported in high MPY cows and high average daily gain (ADG) goats [18, 78]. Although the present study did not identify a significant association between Methanobrevibacter sp. UBA188 and the production phenotypes of long-lived dairy cows, it is important to note that methanogenic archaea are the principal contributors to methane production within the ruminant digestive system [79]. Therefore, the increased prevalence of LH cows within the population could potentially enhance the environmental sustainability of livestock production by reducing methane emissions [80]
Functional annotation of the rumen microbiota suggests that the rumen microbiota in LH dairy cows exhibits enhanced flagellar assembly and bacterial chemotaxis, thereby experiencing reduced dispersal limitation, which may facilitate more efficient movement and localization of rumen bacteria within the rumen environment and enables bacteria to sense and respond to nutrients present in the rumen [81, 82]. Additionally, functions related to amino acid metabolism, including the biosynthesis of amino acids, valine, leucine, and isoleucine, histidine metabolism, and arginine biosynthesis, are also enriched in the rumen microbiota of LH dairy cows. Branched-chain amino acids account for 50% of the essential amino acids in dairy cows. These amino acids are involved in cell and milk protein synthesis, and also help alleviate postpartum ketosis and oxidative stress-induced damage in mammary epithelial cells [83,84,85]. Histidine is a limiting amino acid in high-producing dairy cows fed diets based on corn silage and alfalfa hay [86]. Histidine deficiency can impair DMI, MY, and MPY in dairy cows [87]. Arginine is a conditionally essential amino acid in cows and dietary supplementation with rumen-protected arginine promotes mammary growth and enhances the expression of proteins associated with fatty acid and milk protein synthesis. The yields of ECM, FCM, MFY, and MPY increase linearly with the supplementation of arginine [88]. The enrichment of these pathways suggests that microbial proteins synthesized in the rumen may more effectively meet the amino acid requirements of LH dairy cows. Similarly, an enhanced rumen microbial capacity for branched-chain amino acids, histidine, and arginine biosynthesis has also been observed in calves with higher ADG [89]. Considering the reported association between early-life ADG and subsequent lactation performance in dairy cows [90], cows with sustained productivity may likewise exhibit superior ADG during early growth stages. Furthermore, carbohydrate metabolism-related pathways, including starch and sucrose metabolism and glycolysis/gluconeogenesis, were enriched in LL dairy cows. This observation is consistent with previous studies which suggested that the microbiome of low MPY dairy cows may have a greater capacity to break down complex carbohydrates [18]. The abundance of GH119, which is a CAZyme involved in the degradation of linear starch [91], and CBM63, which is a cellulose-binding module [92], decrease in LH cows, while the abundance of GH13_9 and GH5_2, which are enzymes involved in the degradation of branched starch and soluble oligosaccharides [93, 94], increase. In contrast to the high-MPY cows, the Entner–Doudoroff pathway, which bypasses classical glycolysis and facilitates the rapid generation of NADPH and pyruvate [95], is upregulated in the rumen of LH cows. This carbohydrate-degrading capacity, more adapted to high-grain diets and more direct in pyruvate production, may better accommodate the high demand for propionate in LH cows [96]. Interestingly, this study also demonstrated that pathways related to the metabolism of cofactors and vitamins, including pantothenate and CoA biosynthesis, riboflavin metabolism, and the one-carbon pool by folate, were significantly enriched in LH dairy cows. The KEGG enzymes enriched in the LH group further support the potential enhancement of coenzyme A, vitamin B12 coenzyme, and S-adenosylmethionine synthesis. Coenzyme A is involved in the citric acid cycle, the synthesis of lipids and glycogen, and also acts as a protective thiol in bacterial antioxidant defense [97, 98]. In addition, coenzyme A is also involved in the synthesis of vitamin B12 coenzyme [99]. Vitamin B12 coenzyme participates in the synthesis of methionine through methyl transfer reactions, thereby promoting the formation of S-adenosylmethionine [100, 101]. Insufficient methyl supply may result in impaired amino acid methylation in dairy cows which subsequently affects the synthesis of hepatic phosphatidylcholine and the output of triglycerides and which may lead to systemic oxidative stress and decreased milk production [102]. Metatranscriptomic analyses have also revealed the enrichment of pathways related to the metabolism of cofactors and vitamins in the rumen microbiome of high-FE cows [103].The rumen microbiota in LH cows may enhance the synthesis of methionine and S-adenosylmethionine to meet the demand for methyl donors, thereby supporting protein and lipid synthesis during the high-lactation period and improving FE.
Metabolites derived from the gut microbiota modulate dairy cow production performance [89, 104]. In this study, we demonstrated that the purine metabolite guanosine which is co-metabolized by the rumen microbiota and the host mediates the positive effect of Acidaminococcaceae bacterium on MY. Exogenous guanosine taken up by mouse intestinal epithelial cells increased the intracellular total guanine pool, guanosine triphosphate levels, and energy charge, thereby promoting the differentiation and maturation of intestinal cells [105]. Guanosine also helps to maintain intestinal fluid homeostasis and intestinal epithelial barrier function by activating guanylate cyclase C signaling which limits the systemic dissemination of intestinal pathogens [106,107,108]. Previous studies have shown that dimethyl itaconate alleviates uterine inflammation via gut microbiota-mediated guanosine [109]. Interestingly, aconitate decarboxylase (EC 4.1.1.6) that is involved in itaconate synthesis within rumen bacteria was also enriched in LH cows in this study [110]. This observation suggests that rumen bacteria in LH cows may promote guanosine production through endogenous itaconate synthesis, thereby supporting the development, differentiation, and barrier function of the ruminal and subsequent intestinal epithelium. Once absorbed by the rumen and gut, guanosine may also exert anti-inflammatory and tissue-repair effects on distal organs such as the uterus, thereby alleviating the major factors limiting the productive lifespan of dairy cows—reproductive disorders [4], and ultimately supporting the productive lifespan of LH cows.
P. distasonis and Parabacteroides sp. were enriched in the rectum of the LH group. P. distasonis can increase the proportion of propionate and butyrate in the rumen, improving the ADG in pre-weaning calves [111]. In addition, P. distasonis has been reported to alleviate insulin resistance [112]. Excessive insulin resistance and ketosis mutually reinforce each other in dairy cows [113], which can reduce oocyte quality and consequently impair reproductive performance [114], thereby limiting the productive lifespan of cows [115]. The enrichment of P. distasonis in the hindgut of LH cows may contribute to the maintenance of their productive lifespan by modulating the progression of insulin resistance and ketosis. Moreover, D-ribose-1-phosphate serves as a significant mediator of the positive effect exerted by Parabacteroides sp. on MY, as well as the negative effects mediated by B. pseudolongum and unclassified g Bifidobacterium on MY. D-ribose-1-phosphate commonly is derived from the nucleotide salvage pathways of gut microbes which catalyze the conversion of purine nucleosides originating from diet, microbes, and the host into D-ribose-1-phosphate and free purine bases, thereby recycling the building blocks of purine nucleosides [116]. D-ribose-1-phosphate within the intestinal lumen is hydrolyzed to D-ribose by host alkaline phosphatases [117]. D-ribose is known to be absorbed by the gut where it exerts anti-inflammatory and antioxidant effects and improves intestinal barrier integrity [118]. Parabacteroides sp. may generate D-ribose-1-phosphate via the nucleotide salvage pathway and export it into the intestinal lumen, thereby increasing its concentration and subsequent hydrolysis into D-ribose which confers protective effects on the gut. In contrast, Bifidobacterium, which is known for the ability to degrade oligosaccharides, efficiently utilizes luminal D-ribose-1-phosphate and D-ribose as substrates for nucleotide salvage synthesis or as carbon sources which reduces the accumulation of these compounds in the intestinal lumen [119, 120].
Integrated multi-omics analyses have revealed structural congruence between the rumen microbiome and the milk metabolome. Furthermore, lipids and lipid-like molecules form multilevel correlation networks across the rumen, blood, and milk metabolomes. In the rumen of LH cows, the microbial module associated with lipid metabolism (including MY-associated bacteria H. mizrahii, H. absiana, H. multisaccharivorax, and P. sp. AGR2160) positively influences the synthesis of ruminal GPEtn(12:0/18:3). Certain bacteria incorporate exogenous polyunsaturated fatty acids (PUFAs) into membrane components which increases the sensitivity of these species to antibiotics [121]. This incorporation delays the availability of PUFAs as substrates for biohydrogenation and consequently reduces the instantaneous hydrogenation rate of PUFAs in the rumen to some extent [122]. The rumen microbial module facilitates the conversion of dietary α-linolenic acid to GPEtn(12:0/18:3), which may represent a mechanism to delay biohydrogenation, thereby permitting increased delivery of protected α-linolenic acid to the small intestine for absorption and conversion into EPA [123], which subsequently enters the bloodstream and is utilized by the mammary gland to synthesize Pe(20:5/0:0). EPA in the bloodstream and mammary gland can be converted into metabolites (3-series prostaglandins and 5-series leukotrienes) and anti-inflammatory mediators (resolvins, protectins, and maresins) that inhibit inflammatory cell activation and decrease the production of proinflammatory mediators [124, 125]. Supplementation with EPA in the in vitro maturation medium reduces embryonic reactive oxygen species levels and improves embryo quality [126]. Moreover, dietary EPA supplementation improves folliculogenesis and in vitro fertilization performances in dairy cows [127]. Increased EPA levels in plasma and milk may reduce systemic and mammary gland inflammation. The elevated circulating EPA that reaches ovarian tissue may further enhance reproductive performance, which is consistent with the observation that dairy cows with sustained production capacity typically exhibit superior fertility [4, 128].
PLS-PM has been widely employed to elucidate the contributions of omics data to phenotypic traits [129, 130]. To simplify the model and improve its reliability, we separately evaluated the effects of the rumen and rectum microbiota, their metabolic functions, and host metabolism on phenotypic traits. Previous studies have demonstrated that rumen metabolism is the primary contributor to MPY in dairy cows [18], whereas the functional activities of the rumen microbiome predominantly determine MFY [34]. However, in the present study, the composition of the rumen microbiota was identified as the most critical target for modulating the productive performance of long-lived dairy cows, due to its predominant contribution to phenotypic traits. Host-gut microbial interactions stabilize during the later stages of the production cycle in dairy cows [131, 132]. Nevertheless, targeted supplementation with rumen direct-fed microbials reduces methane emissions [133, 134], increase MY [134, 135], and improve average ADG [111, 136]. Additionally, for studies aimed at improving production performance via the hindgut, rectum metabolism may play a primary role in shaping production phenotypes in dairy cows. This differs from previous studies in which the microbial function and composition of the hindgut in ruminants were considered to play a primary role in shaping the phenotype [129]. Feeding certain dietary additives has also been shown to be feasible for improving dairy cow gut health, systemic inflammatory response, immune homeostasis, and production performance by modulating hindgut metabolic processes [137,138,139]. The positive effect of rumen microbes on milk production performance indicates a stronger contribution to milk production performance compared to the negative effect of rectum microbes. Rumen microbes ferment dietary carbohydrates into SCFAs, providing the primary source of metabolizable energy for the host and serving as key precursors (acetate and propionate) for milk fat and lactose synthesis [140, 141]. Additionally, microbial protein synthesized in the rumen provides a major source of metabolizable protein in dairy cows and supports milk protein synthesis [142]. Given that rumen microbial fermentation and metabolic activity directly determine lactation capacity and milk composition [18, 143], modulating the rumen microbiota may represent a more effective approach to improving the production performance of long-lived dairy cows.
Lifetime productivity in dairy cows is a trait that combines production performance and longevity, and serves as a key objective in dairy cattle breeding programs [144]. Current studies on dairy cow sustained productivity focus on genomic characteristics by integrating phenotypic, pedigree, and genomic datasets to identify genomic regions associated with lifetime MY with the aim of elucidating the genetic architecture of lifetime production performance and to identify candidate genes for selective breeding [6]. However, this study, through integrated multi-omics analysis, revealed the physiological phenotypic characteristics of dairy cows with high potential lifetime productivity and identified several specific gastrointestinal microbes and their host co-metabolites that may act as potential targets involved in regulating the productive performance of long-lived dairy cows via host–microbiome interactions. Although potential confounding factors such as feed, management, and lactation stage were carefully controlled in this study, the sample size may still limit the generalizability of the findings. In addition, heritable microbial taxa have been reported to be associated with methane emissions [145], and lactation traits [20, 146]. However, the heritability of these related microbial communities and metabolites, as well as their contributions to milk production traits, remains unclear. Future research will involve larger sample sizes and evaluate the heritability of these factors to further validate the results and provide more solid evidence for breeding and management strategies aimed at improving the sustained productivity of dairy cows.
Conclusions
These findings provide new insights into the influence of the rumen and hindgut microbiota on the productive performance of long-lived dairy cows. In summary, rumen Acidaminococcaceae bacterium and rectum Parabacteroides sp. may positively influence productive performance through host–microbiota co-metabolites guanosine and D-ribose-1-phosphate. Moreover, the defined rumen microbial module may enhance EPA enrichment in serum and milk, thereby supporting productive performance in LH dairy cows. Collectively, these results identify potential microbial and metabolic targets for developing precision nutritional management strategies aimed at breeding and feeding dairy cows with long-term productive resilience from the perspective of host–microbiota co-metabolism.
Data availability
The metagenomic raw sequencing data have been deposited in the National Center for Biotechnology Information (NCBI) under the BioProject accession number PRJNA1288150. The metabolomics raw data generated in this study have been deposited in MetaboLights (The European Bioinformatics Institute Metabolomics Database, EMBL-EBI MetaboLights) under the accession number MTBLS12879. This dataset is publicly available at the MetaboLights repository (http://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12879).
References
Aubé L, Mialon MM, Mollaret E, Mounier L, Veissier I, de Boyer s Roches A. Review: assessment of dairy cow welfare at pasture: measures available, gaps to address, and pathways to development of ad-hoc protocols. Animal. 2022;16:100597.
Willett W, Rockström J, Loken B, Springmann M, Lang T, Vermeulen S, et al. Food in the Anthropocene: the EAT-Lancet Commission on healthy diets from sustainable food systems. Lancet. 2019;393:447–92.
Heerwagen LR, Christensen T, Sandøe P. The prospect of market-driven improvements in animal welfare: lessons from the case of grass milk in Denmark. Animals (Basel). 2013;3:499–512.
De Vries A, Marcondes MI. Review: overview of factors affecting productive lifespan of dairy cows. Anim. 2020;14:s155-64.
De Vries A. Economic trade-offs between genetic improvement and longevity in dairy cattle. J Dairy Sci. 2017;100:4184–92.
Zhang H, Wang A, Xiao W, Mi S, Hu L, Brito LF, et al. Genetic parameters and genome-wide association analyses for lifetime productivity in Chinese Holstein cattle. J Dairy Sci. 2024;107:9638–55.
Pinedo PJ, De Vries A, Webb DW. Dynamics of culling risk with disposal codes reported by Dairy Herd Improvement dairy herds. J Dairy Sci. 2010;93:2250–61.
Schuster JC, Barkema HW, De Vries A, Kelton DF, Orsel K. Invited review: academic and applied approach to evaluating longevity in dairy cows. J Dairy Sci. 2020;103:11008–24.
Grandl F, Furger M, Kreuzer M, Zehetmeier M. Impact of longevity on greenhouse gas emissions and profitability of individual dairy cows analysed with different system boundaries. Anim. 2019;13:198–208.
Adamie BA, Owusu-Sekyere E, Lindberg M, Agenäs S, Nyman AK, Hansson H. Dairy cow longevity and farm economic performance: evidence from Swedish dairy farms. J Dairy Sci. 2023;106:8926–41.
Heikkilä AM, Nousiainen JI, Jauhiainen L. Optimal replacement policy and economic value of dairy cows with diverse health status and production capacity. J Dairy Sci. 2008;91:2342–52.
McConnel CS, Lombard JE, Wagner BA, Garry FB. Evaluation of factors associated with increased dairy cow mortality on United States dairy operations. J Dairy Sci. 2008;91:1423–32.
Bergman EN. Energy contributions of volatile fatty acids from the gastrointestinal tract in various species. Physiol Rev. 1990;70:567–90.
Urrutia NL, Harvatine KJ. Acetate dose-dependently stimulates milk fat synthesis in lactating dairy cows. J Nutr. 2017;147:763–9.
Wiltrout DW, Satter LD. Contribution of propionate to glucose synthesis in the lactating and nonlactating cow. J Dairy Sci. 1972;55:307–17.
Kay RN. Digestion of protein in the intestines of adult ruminants. Proc Nutr Soc. 1969;28:140–51.
Amin AB, Zhang L, Zhang J, Mao S. Metagenomic and metabolomic insights into the mechanism underlying the disparity in milk yield of Holstein cows. Front Microbiol. 2022;13:844968.
Xue MY, Sun HZ, Wu XH, Liu JX, Guan LL. Multi-omics reveals that the rumen microbiome and its metabolome together with the host metabolome contribute to individualized dairy cow performance. Microbiome. 2020;8:64.
Bai H, Lai Z, Zhang J, Zheng X, Zhang J, Jin W, Lin L, Mao S. Host genetic regulation of specific functional groups in the rumen microbiome of dairy cows: Implications for lactation trait. J Adv Res. 2024;11:012.
Zhang C, Liu H, Jiang X, Zhang Z, Hou X, Wang Y, et al. An integrated microbiome- and metabolome-genome-wide association study reveals the role of heritable ruminal microbial carbohydrate metabolism in lactation performance in Holstein dairy cows. Microbiome. 2024;12:232.
Hu X, Li S, Mu R, Guo J, Zhao C, Cao Y, et al. The rumen microbiota contributes to the development of mastitis in dairy cows. Microbiol Spectr. 2022;10:e0251221.
Zhao C, Hu X, Bao L, Wu K, Zhao Y, Xiang K, et al. Gut dysbiosis induces the development of mastitis through a reduction in host anti-inflammatory enzyme activity by endotoxemia. Microbiome. 2022;10:205.
Zhao C, Hu X, Qiu M, Bao L, Wu K, Meng X, et al. Sialic acid exacerbates gut dysbiosis-associated mastitis through the microbiota-gut-mammary axis by fueling gut microbiota disruption. Microbiome. 2023;11:78.
Sutton JD. Digestion and absorption of energy substrates in the lactating cow. J Dairy Sci. 1985;68:3376–93.
Elias PM. Stratum corneum defensive functions: an integrated view. J Invest Dermatol. 2005;125:183–200.
Erdman RA. Dietary buffering requirements of the lactating dairy cow: a review1. J Dairy Sci. 1988;71:3246–66.
Gu F, Zhu S, Hou J, Tang Y, Liu JX, Xu Q, et al. The hindgut microbiome contributes to host oxidative stress in postpartum dairy cows by affecting glutathione synthesis process. Microbiome. 2023;11:87.
Gu F, Zhu S, Tang Y, Liu X, Jia M, Malmuthuge N, et al. Gut microbiome is linked to functions of peripheral immune cells in transition cows during excessive lipolysis. Microbiome. 2023;11:40.
Huzzey JM, Mann S, Nydam DV, Grant RJ, Overton TR. Associations of peripartum markers of stress and inflammation with milk yield and reproductive performance in Holstein dairy cows. Prev Vet Med. 2015;120:291–7.
Jeon SJ, Cunha F, Vieira-Neto A, Bicalho RC, Lima S, Bicalho ML, et al. Blood as a route of transmission of uterine pathogens from the gut to the uterus in cows. Microbiome. 2017;5:109.
Ma C, Sun Z, Zeng B, Huang S, Zhao J, Zhang Y, et al. Cow-to-mouse fecal transplantations suggest intestinal microbiome as one cause of mastitis. Microbiome. 2018;6:200.
Zhao C, Bao L, Qiu M, Wu K, Zhao Y, Feng L, et al. Commensal cow Roseburia reduces gut-dysbiosis-induced mastitis through inhibiting bacterial translocation by producing butyrate in mice. Cell Rep. 2022;41:111681.
Han R, Mourits M, Steeneveld W, Hogeveen H. The association of herd performance indicators with dairy cow longevity: an empirical study. PLoS One. 2022;17:e0278204.
Jiang B, Qin C, Xu Y, Song X, Fu Y, Li R, et al. Multi-omics reveals the mechanism of rumen microbiome and its metabolome together with host metabolome participating in the regulation of milk production traits in dairy buffaloes. Front Microbiol. 2024;15:1301292.
Sofyan A, Uyeno Y, Shinkai T, Hirako M, Kushibiki S, Kanamori H, et al. Metagenomic profiles of the rumen microbiota during the transition period in low-yield and high-yield dairy cows. Anim Sci J. 2019;90:1362–76.
Edmonson AJ, Lean IJ, Weaver LD, Farver T, Webster G. A body condition scoring chart for Holstein dairy cows. J Dairy Sci. 1989;72:68–78.
Yang J, Li Y, Sun M, Zhang Y, Guo S, Zhou D, et al. Associations of rumen and rectum bacteria with the sustained productive performance of dairy cows. Front Microbiol. 2025;16:1565034.
Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34:i884–90.
Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60.
Li D, Liu CM, Luo R, Sadakane K, Lam TW. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics. 2015;31:1674–6.
Hyatt D, Chen GL, Locascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11:119.
Fu L, Niu B, Zhu Z, Wu S, Li W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics. 2012;28:3150–2.
Li R, Li Y, Kristiansen K, Wang J. SOAP: short oligonucleotide alignment program. Bioinformatics. 2008;24:713–4.
Buchfink B, Xie C, Huson DH. Fast and sensitive protein alignment using DIAMOND. Nat Methods. 2015;12:59–60.
Han C, Shi C, Liu L, Han J, Yang Q, Wang Y, et al. Majorbio cloud 2024: update single-cell and multiomics workflows. Imeta. 2024;3:e217.
Xia J, Wishart DS. Msea: a web-based tool to identify biologically meaningful patterns in quantitative metabolomic data. Nucleic Acids Res. 2010;38:W71-77.
Yu G, Xu C, Wang X, Ju F, Fu J, Ni Y. MetOrigin 2.0: advancing the discovery of microbial metabolites and their origins. Imeta. 2024;3:e246.
Wen T, Xie P, Yang S, Niu G, Liu X, Ding Z, et al. GgclusterNet: an R package for microbiome network analysis and modularity-based multiple network layouts. Imeta. 2022;1:e32.
Liu C, Cui Y, Li X, Yao M: microeco: an R package for data mining in microbial community ecology. FEMS Microbiol Ecol. 2021;97(2):fiaa255.
Shrestha B, Paudyal S, Kaniyamattam K, Grohn YT. Graduate student literature review: organic dairy cattle longevity and economic implications-contemporary perspectives. J Dairy Sci. 2025;108:3734–45.
Fu Y, He Y, Xiang K, Zhao C, He Z, Qiu M, Hu X, Zhang N: The Role of Rumen Microbiota and Its Metabolites in Subacute Ruminal Acidosis (SARA)-Induced Inflammatory Diseases of Ruminants. Microorganisms. 2022;10:10081495.
Xue MY, Sun HZ, Wu XH, Guan LL, Liu JX. Assessment of rumen bacteria in dairy cows with varied milk protein yield. J Dairy Sci. 2019;102:5031–41.
Andjelić B, Djoković R, Cincović M, Bogosavljević-Bošković S, Petrović M, Mladenović J, Čukić A: Relationships between Milk and Blood Biochemical Parameters and Metabolic Status in Dairy Cows during Lactation. Metabolites. 2022;12:12080733.
Hu HH, Mu T, Zhang ZB, Zhang JX, Feng X, Han LY, et al. Genetic analysis of health traits and their associations with longevity, fertility, production, and conformation traits in Holstein cattle. Animal. 2024;18:101177.
Kulkarni PS, Mourits MCM, Slob J, Veldhuis AMB, Nielen M, Hogeveen H, et al. Dutch dairy farmers’ perspectives on culling reasons and strategies. Prev Vet Med. 2023;218:105997.
Maheshwari A, Mantry H, Bagga N, Frydrysiak-Brzozowska A, Badarch J, Rahman MM. Milk fat globules: 2024 updates. Newborn Clarksville. 2024;3:19–37.
Bionaz M, Loor JJ. Gene networks driving bovine milk fat synthesis during the lactation cycle. BMC Genomics. 2008;9:366.
Bardagjy AS, Steinberg FM: Relationship Between HDL Functional Characteristics and Cardiovascular Health and Potential Impact of Dietary Patterns: A Narrative Review. Nutrients. 2019;11:11061231.
Nazir S, Jankowski V, Bender G, Zewinger S, Rye KA, van der Vorst EPC. Interaction between high-density lipoproteins and inflammation: function matters more than concentration! Adv Drug Deliv Rev. 2020;159:94–119.
Weber C, Schäff CT, Kautzsch U, Börner S, Erdmann S, Görs S, et al. Insulin-dependent glucose metabolism in dairy cows with variable fat mobilization around calving. J Dairy Sci. 2016;99:6665–79.
De Koster J, Hostens M, Van Eetvelde M, Hermans K, Moerman S, Bogaert H, et al. Insulin response of the glucose and fatty acid metabolism in dry dairy cows across a range of body condition scores. J Dairy Sci. 2015;98:4580–92.
Pires JA, Souza AH, Grummer RR. Induction of hyperlipidemia by intravenous infusion of tallow emulsion causes insulin resistance in Holstein cows. J Dairy Sci. 2007;90:2735–44.
Hammon HM, Stürmer G, Schneider F, Tuchscherer A, Blum H, Engelhard T, et al. Performance and metabolic and endocrine changes with emphasis on glucose metabolism in high-yielding dairy cows with high and low fat content in liver after calving. J Dairy Sci. 2009;92:1554–66.
Loor JJ, Dann HM, Everts RE, Oliveira R, Green CA, Guretzky NA, et al. Temporal gene expression profiling of liver from periparturient dairy cows reveals complex adaptive mechanisms in hepatic function. Physiol Genomics. 2005;23:217–26.
Dong J, Liu Y, Li S, Sun Z, Chen X, Wang D, et al. The physiological dissimilarities of Holstein dairy cows with different milk yields. Vet Med Sci. 2023;9:429–42.
Hackmann TJ. New biochemical pathways for forming short-chain fatty acids during fermentation in rumen bacteria. JDS Commun. 2024;5:230–5.
Aschenbach JR, Kristensen NB, Donkin SS, Hammon HM, Penner GB. Gluconeogenesis in dairy cows: the secret of making sweet milk from sour dough. IUBMB Life. 2010;62:869–77.
Costa A, Lopez-Villalobos N, Sneddon NW, Shalloo L, Franzoi M, De Marchi M, et al. Invited review: milk lactose-current status and future challenges in dairy cattle. J Dairy Sci. 2019;102:5883–98.
Osorio JS, Lohakare J, Bionaz M. Biosynthesis of milk fat, protein, and lactose: roles of transcriptional and posttranscriptional regulation. Physiol Genomics. 2016;48:231–56.
Li F, Guan LL: Metatranscriptomic Profiling Reveals Linkages between the Active Rumen Microbiome and Feed Efficiency in Beef Cattle. Appl Environ Microbiol. 2017;83:e00061-17.
Shabat SK, Sasson G, Doron-Faigenboim A, Durman T, Yaacoby S, Berg Miller ME, et al. Specific microbiome-dependent mechanisms underlie the energy harvest efficiency of ruminants. ISME J. 2016;10:2958–72.
Honjo M, Suzuki K, Katai J, Tashiro Y, Aoyagi T, Hori T, Okada T, Saito Y, Futamata H: Stable States of a Microbial Community Are Formed by Dynamic Metabolic Networks with Members Functioning to Achieve Both Robustness and Plasticity. Microbes Environ. 2024;39:ME23091.
Marchandin H, Teyssier C, Campos J, Jean-Pierre H, Roger F, Gay B, et al. Negativicoccus succinicivorans gen. nov., sp. nov., isolated from human clinical samples, emended description of the family Veillonellaceae and description of Negativicutes classis nov., Selenomonadales ord. nov. and Acidaminococcaceae fam. nov. in the bacterial phylum Firmicutes. Int J Syst Evol Microbiol. 2010;60:1271–9.
Cheng Z, Meng Z, Tan D, Datsomor O, Zhan K, Lin M, et al. Effects of supplementation of sodium acetate on rumen fermentation and microbiota in postpartum dairy cows. Front Microbiol. 2022;13:1053503.
Dai H, Liu X, Yan J, Aabdin ZU, Bilal MS, Shen X. Sodium butyrate ameliorates high-concentrate diet-induced inflammation in the rumen epithelium of dairy goats. J Agric Food Chem. 2017;65:596–604.
Górka P, Kowalski ZM, Pietrzak P, Kotunia A, Jagusiak W, Holst JJ, et al. Effect of method of delivery of sodium butyrate on rumen development in newborn calves. J Dairy Sci. 2011;94:5578–88.
Cook GM, Wells JE, Russell JB. Ability of Acidaminococcus fermentans to oxidize trans-aconitate and decrease the accumulation of tricarballylate, a toxic end product of ruminal fermentation. Appl Environ Microbiol. 1994;60:2533–7.
Wang D, Chen L, Tang G, Yu J, Chen J, Li Z, et al. Multi-omics revealed the long-term effect of ruminal keystone bacteria and the microbial metabolome on lactation performance in adult dairy goats. Microbiome. 2023;11:215.
Haque MN. Dietary manipulation: a sustainable way to mitigate methane emissions from ruminants. J Anim Sci Technol. 2018;60:15.
Matthews C, Crispie F, Lewis E, Reid M, O’Toole PW, Cotter PD. The rumen microbiome: a crucial consideration when optimising milk and meat production and nitrogen utilisation efficiency. Gut Microbes. 2019;10:115–32.
Rodenes-Gavidia A, Lamelas A, Bloor S, Hobson A, Treadway S, Haworth J, et al. An insight into the functional alterations in the gut microbiome of healthy adults in response to a multi-strain probiotic intake: a single arm open label trial. Front Cell Infect Microbiol. 2023;13:1240267.
Zheng Y, Cao X, Zhou Y, Ma S, Wang Y, Li Z, et al. Purines enrich root-associated Pseudomonas and improve wild soybean growth under salt stress. Nat Commun. 2024;15:3520.
Leal Yepes FA, Mann S, Overton TR, Ryan CM, Bristol LS, Granados GE, et al. Effect of rumen-protected branched-chain amino acid supplementation on production- and energy-related metabolites during the first 35 days in milk in Holstein dairy cows. J Dairy Sci. 2019;102:5657–72.
Wohlt JE, Clark JH, Derrig RG, Davis CL. Valine, leucine, and isoleucine metabolism by lactating bovine mammary tissue. J Dairy Sci. 1977;60:1875–82.
Wu S, Liu X, Cheng L, Wang D, Qin G, Zhang X, et al. Protective mechanism of leucine and isoleucine against H(2)O(2)-induced oxidative damage in bovine mammary epithelial cells. Oxid Med Cell Longev. 2022;2022:4013575.
Lee C, Hristov AN, Cassidy TW, Heyler KS, Lapierre H, Varga GA, et al. Rumen-protected lysine, methionine, and histidine increase milk protein yield in dairy cows fed a metabolizable protein-deficient diet. J Dairy Sci. 2012;95:6042–56.
Giallongo F, Harper MT, Oh J, Parys C, Shinzato I, Hristov AN. Histidine deficiency has a negative effect on lactational performance of dairy cows. J Dairy Sci. 2017;100:2784–800.
Zhang J, Lang J, Bu L, Liu Y, Huo W, Pei C, et al. Impacts of dietary arginine supplementation on performance, nutrient digestion and expression of proteins related to milk fatty acid and casein synthesis in early lactating dairy cows. Animal Nutr. 2025;21:267–78.
Xu SY, Feng XR, Zhao W, Bi YL, Diao QY, Tu Y. Rumen and hindgut microbiome regulate average daily gain of preweaning Holstein heifer calves in different ways. Microbiome. 2024;12:131.
Hayes CJ, McAloon CG, Kelly ET, Carty CI, Ryan EG, Mee JF, et al. The effect of dairy heifer pre-breeding growth rate on first lactation milk yield in spring-calving, pasture-based herds. Animal. 2021;15:100169.
Watanabe H, Nishimoto T, Kubota M, Chaen H, Fukuda S. Cloning, sequencing, and expression of the genes encoding an isocyclomaltooligosaccharide glucanotransferase and an alpha-amylase from a Bacillus circulans strain. Biosci Biotechnol Biochem. 2006;70:2690–702.
Georgelis N, Yennawar NH, Cosgrove DJ. Structural basis for entropy-driven cellulose binding by a type-a cellulose-binding module (CBM) and bacterial expansin. Proc Natl Acad Sci U S A. 2012;109:14830–5.
Matsumoto A, Nakajima T, Matsuda K. Mode of action of glycogen branching enzyme from Neurospora crassa. J Biochem. 1990;107:118–22.
Shin NR, Doucet D, Pauchet Y: Duplication of horizontally acquired GH5_2 enzymes played a central role in the evolution of longhorned beetles. Mol Biol Evol. 2022;39:msac128.
Conway T. The Entner-Doudoroff pathway: history, physiology and molecular biology. FEMS Microbiol Rev. 1992;9:1–27.
Mu YY, Qi WP, Zhang T, Zhang JY, Mao SY. Gene function adjustment for carbohydrate metabolism and enrichment of rumen microbiota with antibiotic resistance genes during subacute rumen acidosis induced by a high-grain diet in lactating dairy cows. J Dairy Sci. 2021;104:2087–105.
Barritt SA, DuBois-Coyne SE, Dibble CC. Coenzyme A biosynthesis: mechanisms of regulation, function and disease. Nat Metab. 2024;6:1008–23.
Gout I. Coenzyme A: a protective thiol in bacterial antioxidant defence. Biochem Soc Trans. 2019;47:469–76.
Martens JH, Barg H, Warren MJ, Jahn D. Microbial production of vitamin B12. Appl Microbiol Biotechnol. 2002;58:275–85.
Li C, Gui G, Zhang L, Qin A, Zhou C, Zha X. Overview of Methionine Adenosyltransferase 2A (MAT2A) as an anticancer target: structure, function, and inhibitors. J Med Chem. 2022;65:9531–47.
Weissbach H. The isolation of the vitamin B12 coenzyme and the role of the vitamin in methionine synthesis. J Biol Chem. 2008;283:23497–504.
McFadden JW, Girard CL, Tao S, Zhou Z, Bernard JK, Duplessis M, et al. Symposium review: one-carbon metabolism and methyl donor nutrition in the dairy cow. J Dairy Sci. 2020;103:5668–83.
Xue MY, Xie YY, Zhong Y, Ma XJ, Sun HZ, Liu JX. Integrated meta-omics reveals new ruminal microbial features associated with feed efficiency in dairy cattle. Microbiome. 2022;10:32.
Zhuang Y, Gao D, Jiang W, Xu Y, Liu G, Hou G, et al. Core microbe Bifidobacterium in the hindgut of calves improves the growth phenotype of young hosts by regulating microbial functions and host metabolism. Microbiome. 2025;13:13.
María Vieites J, Torre RL, Ramírez MC, Torres MI, Sánchez-Pozo A, Gil A, et al. Exogenous nucleosides accelerate differentiation of rat intestinal epithelial cells. Br J Nutr. 2008;99:732–8.
Han X, Mann E, Gilbert S, Guan Y, Steinbrecher KA, Montrose MH, et al. Loss of guanylyl cyclase C (GCC) signaling leads to dysfunctional intestinal barrier. PLoS One. 2011;6:e16139.
Ikpa PT, Sleddens HF, Steinbrecher KA, Peppelenbosch MP, de Jonge HR, Smits R, et al. Guanylin and uroguanylin are produced by mouse intestinal epithelial cells of columnar and secretory lineage. Histochem Cell Biol. 2016;146:445–55.
Mann EA, Harmel-Laws E, Cohen MB, Steinbrecher KA. Guanylate cyclase C limits systemic dissemination of a murine enteric pathogen. BMC Gastroenterol. 2013;13:135.
He Y, Cai J, Xie X, Zhang X, Qu L, Liu J, et al. Dimethyl itaconate alleviates escherichia coli-induced endometritis through the guanosine-CXCL14 axis via increasing the abundance of norank_f_Muribaculaceae. Adv Sci. 2025;12:e2414792.
Dwiarti L, Yamane K, Yamatani H, Kahar P, Okabe M. Purification and characterization of cis-aconitic acid decarboxylase from Aspergillus terreus TN484-M1. J Biosci Bioeng. 2002;94:29–33.
Feng X, Liu Y, Xu S, Ma J, Yuan H, Wang H, et al. Functional analysis of parabacteroides distasonis F4: a novel probiotic strain linked to calf growth and rumen fermentation. J Anim Sci Biotechnol. 2025;16:50.
Sun Y, Nie Q, Zhang S, He H, Zuo S, Chen C, et al. Parabacteroides distasonis ameliorates insulin resistance via activation of intestinal GPR109a. Nat Commun. 2023;14:7740.
Youssef M, El-Ashker M. Significance of insulin resistance and oxidative stress in dairy cattle with subclinical ketosis during the transition period. Trop Anim Health Prod. 2017;49:239–44.
Baruselli PS, Vieira LM, Mf SF, Mingoti RD, Ferreira RM, Chiaratti MR, et al. Associations of insulin resistance later in lactation on fertility of dairy cows. Theriogenology. 2016;86:263–9.
Seifi HA, Leblanc SJ, Leslie KE, Duffield TF. Metabolic predictors of post-partum disease and culling risk in dairy cattle. Vet J. 2011;188:216–20.
Timofeev VI, Zhukhlistova NE, Abramchik YA, Fateev II, Kostromina MA, Muravieva TI, et al. Crystal structure of Escherichia coli purine nucleoside phosphorylase in complex with 7-deazahypoxanthine. Acta Crystallogr F Struct Biol Commun. 2018;74:355–62.
Trembacz H, Jezewska MM. The route of non-enzymic and enzymic breakdown of 5-phosphoribosyl 1-pyrophosphate to ribose 1-phosphate. Biochem J. 1990;271:621–5.
Liu H, Yan C, Teng Y, Guo J, Liang C, Xia X. Gut microbiota and D-ribose mediate the anti-colitic effect of punicalagin in DSS-treated mice. Food Funct. 2024;15:7108–23.
Liu D, Wang S, Xu B, Guo Y, Zhao J, Liu W, et al. Proteomics analysis of Bifidobacterium longum NCC2705 growing on glucose, fructose, mannose, xylose, ribose, and galactose. Proteomics. 2011;11:2628–38.
Pokusaeva K, Neves AR, Zomer A, O’Connell-Motherway M, MacSharry J, Curley P, et al. Ribose utilization by the human commensal Bifidobacterium breve UCC2003. Microb Biotechnol. 2010;3:311–23.
Herndon JL, Peters RE, Hofer RN, Simmons TB, Symes SJ, Giles DK. Exogenous polyunsaturated fatty acids (PUFAs) promote changes in growth, phospholipid composition, membrane permeability and virulence phenotypes in Escherichia coli. BMC Microbiol. 2020;20:305.
Bettero VP, Del Valle TA, Barletta RV, de Araújo CE, de Ferreira Jesus E, de Almeida GF, et al. Use of protected fat sources to reduce fatty acid biohydrogenation and improve abomasal flow in dry dairy cows. Anim Feed Sci Technol. 2017;224:30–8.
Goyens PL, Spilker ME, Zock PL, Katan MB, Mensink RP. Compartmental modeling to quantify alpha-linolenic acid conversion after longer term intake of multiple tracer boluses. J Lipid Res. 2005;46:1474–83.
Kousparou C, Fyrilla M, Stephanou A, Patrikios I: DHA/EPA (Omega-3) and LA/GLA (Omega-6) as Bioactive Molecules in Neurodegenerative Diseases. Int J Mol Sci. 2023;24:10717.
Jeong HY, Moon YS, Cho KK. ω-6 and ω-3 polyunsaturated fatty acids: inflammation, obesity and foods of animal resources. Food Sci Anim Resour. 2024;44:988–1010.
Nikoloff N, Carranza AC, Fabra MC, Campagna A, Anchordoquy JP, Anchordoquy JM, et al. Eicosapentaenoic acid supplemented to in vitro maturation medium results in lesser lipid content and intracellular reactive oxygen species in blastocysts of cattle. Anim Reprod Sci. 2021;229:106765.
Moallem U, Shafran A, Zachut M, Dekel I, Portnick Y, Arieli A. Dietary α-linolenic acid from flaxseed oil improved folliculogenesis and IVF performance in dairy cows, similar to eicosapentaenoic and docosahexaenoic acids from fish oil. Reproduction. 2013;146:603–14.
Heise J, Liu Z, Stock KF, Rensing S, Reinhardt F, Simianer H. The genetic structure of longevity in dairy cows. J Dairy Sci. 2016;99:1253–65.
Cheng X, Liang Y, Ji K, Feng M, Du X, Jiao D, et al. Enhanced propionate and butyrate metabolism in cecal microbiota contributes to cold-stress adaptation in sheep. Microbiome. 2025;13:103.
Zhuang Y, Liu S, Xiao J, Chen T, Gao D, Xu Y, et al. Metagenomics reveals the characteristics and potential spread of microbiomes and virulence factor genes in the dairy cattle production system. J Hazard Mater. 2024;480:136005.
Xue M, Sun H, Wu X, Guan LL, Liu J: Assessment of Rumen Microbiota from a Large Dairy Cattle Cohort Reveals the Pan and Core Bacteriomes Contributing to Varied Phenotypes. Appl Environ Microbiol. 2018;84(19):e00970-18.
Zhu Z, Difford GF, Noel SJ, Lassen J, Løvendahl P, Højberg O. Stability assessment of the rumen bacterial and archaeal communities in dairy cows within a single lactation and its association with host phenotype. Front Microbiol. 2021;12:636223.
Philippeau C, Lettat A, Martin C, Silberberg M, Morgavi DP, Ferlay A, et al. Effects of bacterial direct-fed microbials on ruminal characteristics, methane emission, and milk fatty acid composition in cows fed high- or low-starch diets. J Dairy Sci. 2017;100:2637–50.
Vyas D, McGeough EJ, Mohammed R, McGinn SM, McAllister TA, Beauchemin KA. Effects of Propionibacterium strains on ruminal fermentation, nutrient digestibility and methane emissions in beef cattle fed a corn grain finishing diet. Animal. 2014;8:1807–15.
Bulnes M, Bonilla J, Suazo M, Michelotti TC, Paz A, Lefler J, et al. Effects of peripartal rumen-derived direct-fed microbial supplementation on lactation performance, metabolism, ruminal fermentation, and microbial abundance in dairy cows. J Dairy Sci. 2025;108:2354–70.
Cappellozza BI, Copani G, Boll EJ, Queiroz O. Supplementation of direct-fed microbial Enterococcus faecium 669 affects performance of preweaning dairy calves. JDS Commun. 2023;4:284–7.
He Z, Ma Y, Yang S, Zhang S, Liu S, Xiao J, et al. Gut microbiota-derived ursodeoxycholic acid from neonatal dairy calves improves intestinal homeostasis and colitis to attenuate extended-spectrum β-lactamase-producing enteroaggregative Escherichia coli infection. Microbiome. 2022;10:79.
Zhao Y, Yu S, Li L, Zhao H, Li Y, Jiang L, et al. Feeding citrus flavonoid extracts decreases bacterial endotoxin and systemic inflammation and improves immunometabolic status by modulating hindgut microbiome and metabolome in lactating dairy cows. Animal Nutr. 2023;13:386–400.
Zhao Y, Yu S, Zhao H, Li L, Li Y, Liu M, et al. Integrated multi-omics analysis reveals the positive leverage of citrus flavonoids on hindgut microbiota and host homeostasis by modulating sphingolipid metabolism in mid-lactation dairy cows consuming a high-starch diet. Microbiome. 2023;11:236.
Reyes GC, Innes DJ, Ellis JL, Fox MK, Cant P. Relationship between rate of glucose or propionate infusion and milk protein yield and concentration in dairy cows: a meta-regression. J Dairy Sci. 2024;107:2785–96.
Urrutia N, Harvatine KJ. Effect of conjugated linoleic acid and acetate on milk fat synthesis and adipose lipogenesis in lactating dairy cows. J Dairy Sci. 2017;100:5792–804.
Schwab CG, Broderick GA. A 100-year review: protein and amino acid nutrition in dairy cows. J Dairy Sci. 2017;100:10094–112.
Golder HM, Thomson J, Rehberger J, Smith AH, Block E, Lean IJ. Associations among the genome, rumen metabolome, ruminal bacteria, and milk production in early-lactation Holsteins. J Dairy Sci. 2023;106:3176–91.
Fuerst C, Sölkner J. Additive and nonadditive genetic variances for milk yield, fertility, and lifetime performance traits of dairy cattle. J Dairy Sci. 1994;77:1114–25.
Difford GF, Plichta DR, Løvendahl P, Lassen J, Noel SJ, Højberg O, et al. Host genetics and the rumen microbiome jointly associate with methane emissions in dairy cows. PLoS Genet. 2018;14:e1007580.
Bai H, Lai Z, Zhang J, Zheng X, Zhang J, Jin W, et al. Host genetic regulation of specific functional groups in the rumen microbiome of dairy cows: implications for lactation trait. J Adv Res. 2025;75:111–22.
Acknowledgements
The authors would like to thank the reviewers for their valuable suggestions and insightful comments on this work.
Funding
This study was supported by grants from the National Key R&D Program of China (No. 2023YFD1801100). Shaanxi Livestock and Poultry Breeding Double-chain Fusion Key Project (Grant No. 2022GD-TSLD-46). The Key R&D Program of Ningxia Hui Autonomous Region (Grant No. 2018BBF33001).
Author information
Authors and Affiliations
Contributions
JHY: Conceptualization; Data curation; Visualization; Writing-Original draft preparation. YFF, TFG, SSG, MMY, DZ, PFL and AHW: Writing-review and editing. YPJ: Supervision; Writing-reviewing and editing; Funding acquisition. All the authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Ethics approval and consent to participate
This experiment was conducted at the Animal Research and Technology Center of Northwest A&F University, and it was performed in accordance with the recommended guidelines from the Administration of Affair Concerning Experimental Animals (Ministry of Science and Technology, China, revised 2004). The protocol was approved by the Institutional Animal Care and Use Committee at Northwest A&F University.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
40168_2025_2309_MOESM1_ESM.xlsx (download XLSX )
Additional file 1: Table S1. Ingredients of the basal diet (%). Table S2. Detailed methods used in metagenomic and metabolomic analyses. Table S3. Comparison of physiological parameters between LH and LL cows. Table S4. Summary of sequence data generated from rumen samples of LH and LL cows. Table S5. Wilcoxon rank-sum test was used to compare the rumen microbial domains. Table S6. Spearman correlation analysis to identify MY-associated rumen microbiota. Table S7. Comparison of rumen network metrics between LH and LL networks. Table S8. The topological roles of rumen microbiota in the LH and LL networks. Table S9. The relative abundances of rumen KEGG pathways between LH and LH group. Table S10. KEGG enzymes significantly different between the LH and LL groups. Table S11. Identification of rumen MY‐associated metabolites with significant differences between LH and LL cows. Table S12. Top 10 enriched metabolic pathways in rumen metabolome. Table S13. Identification of microbial and host-derived metabolites in rumen. Table S14. Metabolite origin-based pathway enrichment in rumen metabolome. Table S15. Rumen microbial species contributing more than 0.1% to purine metabolism. Table S16. Summary of sequence data generated from rectum samples of 8 LH and 8 LL cows. Table S17. Wilcoxon rank-sumtest was used to compare the rectum microbial domains. Table S18. Spearman correlation analysis to identify MY-associated rectum microbiota. Table S19. Comparison of rectum network metrics between LH and LL networks. Table S20. The topological roles of rectum microbiota in the LH and LL networks. Table S21. The relative abundances of rectum KEGG level 1 pathways between LH and LL cow. Table S22. Identification of rectum MY‐associated metabolites with significant differences between LH and LL cows. Table S23. Top 10 enriched metabolic pathways in rectum metabolome. Table S24. Identification of microbial and host-derived metabolites in rectum. Table S25. Metabolite origin-based pathway enrichment in rectum metabolome. Table S26. Rectum microbial species contributing more than 0.1% to purine metabolism. Table S27. Identification of serum MY‐associated metabolites with significant differences between LH and LL cows. Table S28. Identification of milk MY‐associated metabolites with significant differences between LH and LL cows. Table S29. Top 10 enriched metabolic pathways in serum metabolome. Table S30. Top 10 enriched metabolic pathways in milk metabolome.
40168_2025_2309_MOESM2_ESM.docx (download DOCX )
Additional file 2: Figure S1 Comparison of serum biomarkers between LH and LL dairy cows. Differences in serum biomarkers between the LH and LL groups. Error bars represent mean ± SEM. Figure S2 The differences in rumen microbiota diversity and structure between LH and LL groups. Figure S3 Comparison of rumen bacterial taxa between LH and LL groups. Figure S4 Comparison of rumen archaeal species between LH and LL group. Figure S5 Comparison between the rumen microbial networks (LH and LL cows) and the randomized Erdős–Rényi (E–R) networks. Figure S6 Comparison of rumen microbial network robustness between LH and LL groups. Figure S7 Distributions of centrality measures and robustness tests after key-node removal in the rumen microbial networks of LH and LL dairy cows. Figure S8 Comparison of rumen microbiome KEGG modules and CAZymes between LH and LL dairy cows. Figure S9 Spearman correlation analysis of rumen microbial KEGG modules and CAZymes with MY-associated microbiota. Figure S10 Relationships between rumen MY-associated microbiota, differential microbial functions, and production phenotypes. Figure S11 Screening of significantly enriched differential rumen metabolites. Figure S12 Mediation model of rumen metabolites mediating the effects of specific microbial species on MY. Figure S13 The differences in rectum microbiota diversity and structure between LH and LL groups. Figure S14 Comparison of rectum microbial taxa between LH and LL groups. Figure S15 Comparison between the rectum microbial networks (LH and LL cows) and the randomized Erdős–Rényi (E–R) networks. Fig. S16 Comparison of rectum microbial network robustness between LH and LL groups. Figure S17 Distributions of centrality measures and robustness tests after key-node removal in the rectum microbial networks of LH and LL dairy cows. Figure S18 Metabolic pathways identified in the rectum metagenomes between LH and LL dairy cows. Figure S19 Comparison of rectum microbiome KEGG modules, name, and CAZymes between LH and LL dairy cows. Figure S20 Spearman correlation analysis of rectum microbial functions with MY-associated microbiota. Figure S21 Screening of significantly enriched differential rectum metabolites. Figure S22 Mediation model of rectum metabolites mediating the effects of specific microbial species on MY. Figure S23 Comparison of serum and milk metabolites in LH and LL cows. Figure S24 MY-associated serum and milk metabolites and their relationship with MY-associated rumen microbiota. Figure S25 MY-associated serum and milk metabolites and their relationship with MY-associated rectum microbiota. Figure S26 Metabolite transfer and its impact on milk production phenotypes. Figure S27 PLS-PM-based integrative multi-omics analysis.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Yang, J., Feng, Y., Guo, T. et al. The impact of rumen and hindgut microbiomes on the persistent productivity of long-lived dairy cows. Microbiome 14, 60 (2026). https://doi.org/10.1186/s40168-025-02309-1
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1186/s40168-025-02309-1













