Ensiled diet improved the growth performance of Tibetan sheep by regulating the rumen microbial community and rumen epithelial morphology

Abstract

The aim of this study was to investigate the effects of ensiled agricultural byproducts from Qinghai-Tibet plateau on growth performance, rumen microbiota, ruminal epithelium morphology, and nutrient transport-related gene expression in Tibetan sheep. Fourteen male Tibetan sheep were randomly assigned to one of two diets: an untreated diet (without silage inoculum, CON, n = 7) or an ensiled diet (with silage inoculum, ESD, n = 7). The total experimental period lasted for 84 d, including early 14 d as adaption period and remaining 70 d for data collection. The ESD increased average daily gain (P = 0.046), dry matter intake (P < 0.001), ammonia nitrogen (P = 0.045), microbial crude protein (P = 0.034), and total volatile fatty acids concentration (P < 0.001), and decreased ruminal pH value (P = 0.014). The proportion of propionate (P = 0.006) and the copy numbers of bacteria (P = 0.01) and protozoa (P = 0.002) were higher, while the proportion of acetate (P = 0.028) was lower in the sheep fed ESD compared to CON. Pyrosequencing of the 16S ribosomal RNA gene revealed that ESD increased the relative abundance of Firmicutes, Ruminococcus, Lachnospiraceae_AC2044_group, Lachnospiraceae_XPB1014_group, and Christensenellaceae_R-7_group in the rumen (P < 0.05), while decreased the relative abundance of Bacteroidota, Prevotellaceae_UCG-003, and Veillonellaceae_UCG-001 (P < 0.05). Analyses with PICRUSt2 and STAMP indicated that the propionate metabolism pathway was enriched in the sheep fed ESD (P = 0.026). The ESD increased the rumen papillae height (P = 0.012), density (P = 0.036), and surface area (P = 0.001), and improved the thickness of the total epithelia (P = 0.018), stratum corneum (P = 0.040), stratum granulosum (P = 0.042), and stratum spinosum and basale (P = 0.004). The relative mRNA expression of cyclin-dependent Kinase 2, CyclinA2, CyclinD2, zonula occludens-1, Occludin, monocarboxylate transporter isoform 1 (MCT1), MCT4, sodium/potassium pump, and sodium/hydrogen antiporter 3 were higher in the rumen epithelial of sheep fed ESD than CON (P < 0.05). Conversely, the relative mRNA expressions of Caspase 3 and B-cell lymphoma-2 were lower in the sheep fed ESD than CON (P < 0.05). In conclusion, compared with an untreated diet, feeding an ensiled diet altered the rumen microbial community, enhanced nutrient transport through rumen epithelium, and improved the growth performance of Tibetan sheep.

This research demonstrated that ensiled diet improved the growth performance of Tibetan sheep by changing the rumen microbial structure and by promoting nutrient transport through rumen epithelium.

Introduction

Tibetan sheep (Ovis aries), residing on the Qinghai-Tibet Plateau at elevations exceeding 3,000 m, are integral to the local communities, providing meat, fur, fuel, and other living materials. Presently, more than 50 million animals graze on the Qinghai Plateau year-round, relying on natural grass as their main feed (Wang et al., 2019). The harsh climatic environment (low temperature, low oxygen, and strong ultraviolet) on the Qinghai-Tibet Plateau results in a seasonal imbalance in the forage supply (Sha et al., 2023). During the cold season (from October to April of the following year), the region experiences a 7-month period of dry grass period, leading to a severe shortage of forage and an acute deficiency of nutrient intake in Tibetan sheep (Sha et al., 2023). Tibetan sheep fall into a vicious cycle of “alive in summer, strong in autumn, thin in winter, and died in spring” (Jing et al., 2018). This inefficient mode of production has severely limited the performance of Tibetan sheep and reduced the economic return for local people.

Local agricultural byproducts such as barley lees, corn straw, and potato pulp are rich in proteins, crude fibers, and beta-glucans (Xu et al., 2020; Chen et al., 2024). However, despite their nutritional value, these byproducts are not used in ruminant diets due to their complex cell wall structure (Jin et al., 2014). The cell wall structure composed of cellulose, hemicellulose, and lignin is difficult to degrade by rumen microorganisms (Yu et al., 2016). Ensiling is a common method for preserving fodder and other feed materials for livestock that actively promotes rumen fiber degradation and is widely used in animal feeding (Xu et al., 2022; Gao et al., 2023). The use of probiotic inoculants such as yeast, lactobacillus, bacillus, and other lactic acid-producing bacteria is a common practice in the production of ensiled ration (Yan et al., 2019; Cui et al., 2022).

Diet is the main factor that alters the microbial community in the rumen (Newbold and Ramos-Morales, 2020). He et al. (2023) reported that feeding fermented lees enhanced the abundance of cellulolytic bacteria and improved the microbial community structure in the gastrointestinal tract of Guanling cattle. Xu et al. (2023) reported that silage treatment disrupted the structure of rice straw and increased the abundance of cellulolytic bacteria in the rumen of Hu sheep. Results of aforementioned studies suggest that an ensiled diet may promote fiber degradation and utilization by altering rumen microbiota. The rumen microorganisms ferment fiber to produce volatile fatty acids (VFA), which provide 70% to 80% of the total energy required by ruminants (Li et al., 2023). Approximately 50%-80% of VFA are absorbed and transported through rumen epithelium, where they also contribute to the development and structural changes of rumen epithelium (Jing et al., 2018). Based on the available information, it appears possible to improve the nutritive value of local agricultural byproducts by ensiling, on Qinghai-Tibet Plateau. This study examines the effects of an ensiled diet on production performance, rumen fermentation parameters, and microbiota, as well as rumen epithelial morphology and expression of genes related to nutrient transport in Tibetan sheep.

Materials and Methods

Ethics approval

The use of animals in this experiment was approved by the Animal Care and Use Committee of Nanjing Agricultural University, Nanjing, Jiangsu, China [SYXK (SU) 2021–0086].

Preparation of experimental diets

Two total mixed rations were prepared according to the nutrient requirements of meat-type sheep and goats (NY/T 816-2021, Table 1). The two experimental rations consisted of an untreated diet for the CON group and an ensiled diet for the ESD group. The silage inoculum was obtained from Qinghai Regenerative Nutrition Biotechnology Co., LTD (Hu Zhu, China). The commercial silage inoculum, comprising Saccharomyces cerevisiae, Lactobacillus plantarum, and Lactobacillus casei, at concentrations of 2 × 10⁹, 4 × 10⁹, and 1 × 10⁹ cfu/kg, respectively, was applied to untreated diet. After inoculation, the untreated diet was sealed in plastic bags and ensilaged at a temperature range of 20 to 30 °C for 10 d to produce the ensiled diet.

Table 1.

Ingredients and chemical compositions of experimental diets fed to Tibetan sheep

Feed components	Ingredients, % of DM	Items	Groups1
			CON	ESD
Highland barley lees	40.00	Metabolic energy2, MJ/kg DM	9.67	9.51
Corn straw	15.00	Crude protein, % DM	15.45	15.52
Potato pulp	10.00	Ether extract, % DM	2.69	2.71
Corn	15.00	Neutral detergent fiber, % DM	29.71	28.44
Rice bran	5.00	Acid detergent fiber, % DM	16.06	17.01
Corn germ meal	7.00	Calcium, % DM	0.89	0.91
Seabuckthorn seed meal	3.00	Phosphorus, % DM	0.45	0.46
Corn steep liquor	2.00	Water-soluble carbohydrates, % DM	5.93	6.10
Stone powder	0.50	pH	5.82	4.68
NaHCO3	0.60
Urea	0.20
NaCl	0.20
Premix3	1.50

¹Groups: CON, untreated diet; ESD, ensiled diet.

²Metabolic energy values were calculated value using the data for each ingredient obtained from the China Feed Database (http://www.chinafeeddata.org.cn).

³Premix contains vitamin A: 100 KIU; vitamin D3:30 KIU; vitamin E: ≥300 IU; CuSO₄: 275 mg; ZnSO₄: 700 mg; FeSO₄: 400 mg; and MnSO₄: 500 mg/kg.

Experiment design and sample collection

The experiment was conducted at Mohe camel farm (36°49ʹN, 98°54ʹE, 3,059 m above sea level), Chaka Town, Wulan County, Haixi Mongolian and Tibetan Autonomous Prefecture, Qinghai Province from March to June 2023. Fourteen healthy male Tibetan sheep (5 months old, 29.37 ± 1.16 kg live weight) were selected as experimental animals. Sheep were randomly assigned to two groups: one fed an untreated diet (CON, n = 7) and the other an ensiled diet (ESD, n = 7), following a completely randomized design. All sheep were housed in separate pens (0.7 m × 1.2 m), each equipped with a water tank and feed trough. The pens were located within a semi-enclosed house maintained at a temperature similar to that of the external environment. The total experimental period lasted for 84 d, including an initial 14 d adaptation period for the experimental diets, with the remaining 70 d allocated for data collection. During the experimental period, diets were provided at 8:30 a.m. and 7:30 p.m. All animals had free access to feed and water. Sheep were weighed on days 14 and 84 to estimate weight gain, while dry matter intake (DMI) was measured twice weekly. Feed samples were collected weekly and stored at −20 °C for subsequent chemical analysis. On day 84, all experimental animals were slaughtered by halal carotid exsanguination before the morning feeding. Rumen fluid samples were collected immediately after slaughter and filtered through four layers of gauze. The pH value of the rumen fluid was measured immediately using a pH electrode meter (PHS-3C, Hangzhou Qiwei Instruments Co., Ltd., Hangzhou, China). Three pieces of rumen epithelial tissue (5 × 5 cm) were obtained from the rumen dorsal sac after removing the muscular and serosal layers. One tissue piece was placed in normal saline solution for subsequent determination of rumen epithelial papillae index, while the second piece was fixed with 4% paraformaldehyde solution and stored at room temperature for subsequent histological analyses. The third tissue piece was cut into smaller pieces (1 × 1 cm) for subsequent RNA extraction and analysis. All epithelial tissues and rumen fluid samples were briefly preserved in liquid nitrogen before being transferred to −80 °C for later analysis.

Chemical analysis of feeds

Feed samples were dried at 65 °C for 48 h, crushed, and passed through a 1 mm sieve. Dry matter and ether extract were determined according to the method of Association of Official Analytical Chemists (AOAC, 2012). Neutral detergent fiber and acid detergent fiber were determined by the method of Van Soest et al. (1991). Crude protein was measured by the Kjeldahl methods (AOAC, 2012). Calcium and phosphorus were determined by an element analyzer (iCAP 7400, Thermo Fisher Scientific Inc, USA). The water-soluble carbohydrates concentration was determined by the method of Arthur Thomas (1977). The feed samples (10 g) were blended with 90 mL of distilled water in a juice extractor for 30 s, and filtered through four layers of medical gauze. The pH of the filtrates was measured immediately using a pH electrode meter.

Determination of rumen fermentation parameters and microbial populations

Ruminal lactate concentration was measured by a commercial kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Ruminal ammonia nitrogen (NH₃-N) and microbial crude protein (MCP) were determined according to the methods of Broderick and Kang. (1980) and Makkar et al. (1982), respectively. Ruminal VFAs were analyzed using a 7890B gas chromatography system (Agilent Technologies, Santa Clara) equipped with a fused-silica capillary column (length: 30 m; internal diameter: 0.25 mm; film thickness: 0.25 mm; Supelco), according to the method described by Jin et al. (2018). The copy number of rumen microorganism was determined by quantitative real-time polymerase chain reaction (qRT-PCR) using Quant StudioTM 5 Flex System (QuantStudio 5 Flex, Thermo) and calculated with a standard curve using Standard plasmid (Shenggong Bioengineering Co., Ltd., Shanghai, China). The primers for bacteria, fungi, archaea, and protozoa refer to previous studies (Sylvester et al., 2004; Denman and McSweeney, 2006; Jeyanathan et al., 2011; Metzler-Zebeli et al., 2013) and are listed in Supplementary Table S1.

Microbial DNA extraction and 16S ribosomal RNA amplicon sequencing

Total microbial DNA in rumen fluid was extracted using the TIANamp stool DNA kit (DP328, TIANGEN Biotechnology, Beijing, China). The quality and concentration of DNA were measured by a Nanodrop (2000C, Thermo, USA) and stored at −20°C. The Primers 515 F (5ʹ-GTGCCAGCMGCCGCGTAA-3ʹ) and 806 R (5ʹ-GGACTACHVGGTWTCTAAT-3ʹ) were used to amplify the V4 region of bacterial 16S Ribosomal RNA (16S rRNA) gene (Caporaso et al., 2011). About 30 ng DNA per sample was used for PCR amplification. The PCR amplification products were purified by Agcourt AMPure XP microbeads, checked for library quality, and sequenced using Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA). 16S rRNA sequencing of microbial DNA was performed at BGI Genomics (Shenzhen, China).

Bioinformatics analysis

Raw sequencing reads were filtered using Cutadapt (version 2.6) and clean data were obtained after removing the primer sequence and barcode. Bacterial amplicon sequence variants (ASVs) were generated using the QIIME2 (Bolyen et al., 2019) at a 100% similarity level. Species annotation was performed with SILVA database (version 138; Quast et al., 2012) to obtain the relative abundance of bacteria at the phylum and genus levels (Supplementary Table S2). The shared and unique ASVs between the two groups were analyzed using the Venn diagram. Alpha diversity and the principal coordinate analysis (PCoA) of bacterial communities between groups based on Bray–Curtis distance were calculated with QIIME2. Linear discriminant analysis (LDA) effect size (LEfSe) was used to detect the difference in bacterial community between the two groups based on relative abundance > 1%, LDA > 2, and P < 0.05 (Segata et al., 2011). The correlations between phenotypic indices and differential bacteria genera were analyzed using Spearman’s correlation heatmap. The Kyoto Encyclopedia of Genes and Genomes (Kanehisa and Goto, 2000) pathway hierarchy level 3 of rumen bacteria was applied and predicted using PICRUSt2 software (Douglas et al., 2020). The differential metabolic pathways between rumen bacterial communities using the statistical analysis of taxonomic and functional profiles (STAMP) software (P-value were adjusted by Bonfferoni method, Parks et al., 2014).

Morphology analysis of rumen epithelial tissue

The epithelial tissue samples in normal saline solution were randomly cut into three pieces of epithelial tissue (1 × 1cm). The average density of each tissue was measured using a magnifying glass and recorded as (N). Subsequently, five rumen papillae were randomly selected from each tissue and their height and width were measured using vernier calipers. The surface area of rumen epithelial papillae/cm² was calculated as 2 × height × width × density of the papillae according to the method of Wang et al (2022). The tissue samples, immersed in 4% paraformaldehyde solution, were subjected to alcohol gradient dehydration before being embedded in paraffin. Subsequently they were sliced into 6 μm thickness and stained with hematoxylin/eosin. The morphology of epithelial tissue was observed using an Olympus BX45 microscope (Olympus Optical, Tokyo, Japan). Five rumen papillae per animal were randomly selected for analysis, and three images (representing three different areas within one papilla) were captured per papilla, totaling 15 replicates per animal. ImageJ (Version 1.53c) software was used to measure predefined criteria previously described by Malhi et al (2013). The thickness of each stratum and total epithelia were measured to calculate their average values.

RNA extraction and qRT-PCR analysis of rumen epithelial tissue

The total RNA of rumen epithelial tissue was extracted using commercial kits (RC-112-01, Vazyme, Nanjing, China). Extracted RNA was qualified using a Nanodrop (2000C, Thermo, USA) and agarose gel electrophoresis. RNA was reverse-transcribed into cDNA by using HiScript II First Strand cDNA Synthesis Kit (+ gDNA wiper; R212-01, Vazyme). Next, qRT-PCR was performed using Quant StudioTM 5 Flex System (QuantStudio 5 Flex, Thermo, USA) and ChamQTM SYBR qPCR Master Mix kit (Q711–02, Vazyme). The reaction system was 20 μL, including SYBR qPCR Master Mix 10 μL, forward and reverse primers 0.4μL each, DNA template 2 μL, and ddH₂o 7.2 μL. The fluorescence reaction procedure was 95°C 30 s; 95°C 10s, 60°C 30 s, 40 cycles; 95°C 15 s, 60°C 1 min, 95°C 15 s. Each sample contained three technical replicates. The primers for CyclinA2, CyclinD1, CyclinE1, cyclin-dependent Kinase 2 (CDK2), CDK4, CDK6, Caspase 3, Caspase 8, B-cell lymphoma-2 (Bcl-2), and Bcl-2 associated agonist of cell death were quoted as described by Xu et al (2018). The primers for zonula occludens-1 (ZO-1), tight junction protein claudin (Claudin-1 and Claudin-4), and Occludin were quoted as described by Chen et al. (2023). The primers for monocarboxylate transporter isoform 1 (MCT1), MCT4, down-regulated in adenoma, anion exchanger 2, sodium/hydrogen antiporter isoform 1 (NHE1), NHE3, vacuolar-type ATPase, and beta actin (β-actin) were quoted as described by Jing et al. (2020). The primer for sodium/potassium pump (Na⁺/K⁺-ATPase) was quoted by Jing et al. (2018) and NHE2 was quoted as described by Wang et al. (2022). All primers were synthesized by Shenggong Bioengineering Co., Ltd. (Shanghai, China). β-Actin was used as a housekeeping gene to normalize the mRNA expression levels of each target gene, and the 2^−ΔΔCT method was used to analyze the data. The primers mentioned above are listed in Supplementary Table S3.

Statistical analysis

The data, including growth performance, rumen fermentation parameters, microbial populations, rumen epithelial morphology, and the relative mRNA expression of genes in the rumen epithelium, were statistically analyzed using Student’s t-test in the SPSS software version 27.0. The PCoA based on the Bray–Curtis distance was analyzed by QIIME2, and the differences between two groups were measured by the ANOSIM. LEfSe analysis was employed to determine the significant differences in the bacterial community between the two groups. Specific visualization steps were carried out through R software (version 4.2.3.), Graph Prism v8, and the online data visualization website (https://www.omicstudio.cn). Results were presented as means and SD and considered significantly different at P < 0.05.

Results

Growth performance

Final body weight (P = 0.042), DMI (P < 0.001), and average daily gain (ADG; P = 0.046) were higher in sheep fed ESD than on CON (Table 2). However, no significant difference was observed in the feed conversion ratio (P = 0.504) between the two groups during the experiment period.

Table 2.

Effects of ensiled diet on growth performance of Tibetan sheep¹

Items	Groups2		P-value
	CON	ESD
Initial body weight, kg	30.91 ± 0.72	31.19 ± 0.86	0.535
Final body weight, kg	42.03 ± 2.70	45.34 ± 2.74	0.042
Average daily gain, g/d	158.78 ± 35.57	202.25 ± 37.55	0.046
Dry matter intake, kg/d	1.21 ± 0.11	1.46 ± 0.10	<0.001
Feed conversion ratio3	7.98 ± 2.14	7.37 ± 1.00	0.504

¹Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences.

²Groups: CON, untreated diet; ESD, ensiled diet.

³The feed conversion ratio represents the ratio between dry matter intake and average daily gain.

Analysis of rumen fermentation parameters and microbial populations

The ESD increased the NH₃-N (P = 0.045), MCP (P = 0.034) and decreased ruminal pH (P = 0.014), while the concentration of lactate (P = 0.607) between the two groups remained unaffected (Table 3). Additionally, ESD significantly increased total VFA concentration (P < 0.001). The proportion of propionate increased (P = 0.006), while that of acetate (P = 0.028), iso-butyrate (P = 0.003), iso-valerate (P = 0.003), and the acetate/propionate (A/P) ratio decreased (P = 0.015) in the sheep fed ESD. Ensiled diet significantly increased the copy number of bacteria (P = 0.010) and protozoa (P = 0.002) in the ruminal fluid but did not affect archaea (P = 0.501) and fungi (P = 0.099).

Table 3.

Effects of ensiled diet on rumen fermentation parameters and microbial populations of Tibetan sheep¹

Items	Groups2		P-value
	CON	ESD
Rumen fermentation parameter
pH	7.07 ± 0.16	6.88 ± 0.09	0.014
Ammonia nitrogen, mg/dL	10.76 ± 1.99	13.35 ± 2.08	0.045
Microbial crude protein, mg/dL	63.41 ± 2.99	69.53 ± 6.08	0.034
Lactate, mmol/L	1.79 ± 0.44	1.61 ± 00.80	0.607
Total VFA3	75.43 ± 16.38	109.62 ± 10.76	<0.001
Proportion of individual VFA, %
Acetate	67.71 ± 2.84	64.75 ± 1.32	0.028
Propionate	18.26 ± 2.45	21.80 ± 1.69	0.006
Butyrate	9.05 ± 0.92	9.18 ± 2.44	0.900
Iso-butyrate	1.74 ± 0.22	1.38 ± 0.15	0.003
Valerate	1.15 ± 0.19	1.27 ± 0.14	0.221
Iso-valerate	2.07 ± 0.23	1.63 ± 0.22	0.003
Acetate/propionate ratio	3.78 ± 0.72	2.98 ± 0.20	0.015
Microbial copy number, log10 (copy numbers)/mL
Bacteria	11.27 ± 0.11	11.47 ± 0.13	0.010
Protozoa	6.22 ± 0.88	7.87 ± 0.68	0.002
Funji	6.56 ± 0.11	6.68 ± 0.39	0.501
Archaea	5.82 ± 1.09	6.57 ± 0.27	0.099

¹Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences.

²Groups: CON, untreated diet; ESD, ensiled diet.

³VFA = volatile fatty acid.

Bacterial communities in the rumen

There were 737 common taxa between the two groups (Figure 1A). Additionally, the CON and ESD treatments have 192 and 158 unique taxa, respectively. There was no significant difference in alpha diversity indices (ACE, Chao1, Shannon, and Simpson) between the two groups (Figure 1B). There was significant structural separation of rumen bacterial communities between the treatments at the phylum (P = 0.009) and genus level (P = 0.045; Figure 1C and D). At the phylum level, ruminal bacteria were dominated by the phyla Firmicutes and Bacteroidota (Figure 2A), while the major genera were Prevotella and Rikenellaceae_RC9_gut_group (Figure 2B). LEfSe analysis showed that ESD increased the relative abundance of Firmicutes, Ruminococcus, Lachnospiraceae_AC2044_group, Lachnospiraceae_XPB1014_group, and Christensenellaceae_R-7_group in the rumen (P < 0.05), while decreased the relative abundance of Bacteroidota, Prevotellaceae_UCG-003, and Veillonellaceae_UCG-001 (P < 0.05; Figure 2C and D). Prevotellaceae_UCG-003 was positively correlated with iso-valerate, iso-butyrate, and A/P ratio, but negatively correlated with MCP, DMI, and the proportion of propionate (|r| > 0.6, P < 0.05, Figure 3A). Veillonellaceae_UCG-001 was positively correlated with A/P ratio but negatively correlated with DMI and the proportion of propionate (|r| > 0.6, P < 0.05, Figure 3A). Ruminococcus was negatively correlated with iso-valerate, iso-butyrate, and A/P ratio, but positively correlated with MCP, DMI, TVFA concentration, and the proportion of propionate (|r| > 0.6, P < 0.05, Figure 3A). Christensenellaceae_R-7_group was positively correlated with the proportion of propionate and ADG but negatively correlated with A/P ratio (|r| > 0.6, P < 0.05, Figure 3A). Lachnospiraceae_AC2044_group was positively correlated with the proportion of propionate but negatively correlated with A/P ratio (|r| > 0.6, P < 0.05, Figure 3A). Furthermore, the propionate metabolism pathway in the ESD was significantly higher than in the CON (P = 0.026, Figure 3B).

Figure 1.

(A) Venn diagram. (B) Alpha diversity analysis between two groups. PCoA analysis at phylum level (C) and genus level (D). Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences. Groups: CON, untreated diet; ESD, ensiled diet.

Figure 2.

(A) Relative abundance of the bacteria at the phylum level in the rumen fluid. (B) Relative abundance of the bacteria at the genus level in the rumen fluid. (C) Differential phyla identified by the LEfSe analysis (relative abundance > 1%, LDA > 2, and P < 0.05). (D) Differential genera identified by the LEfSe analysis (relative abundance > 1%, LDA > 2, and P < 0.05). Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences. Groups: CON, untreated diet; ESD, ensiled diet.

Figure 3.

(A) The correlation heatmap of rumen fermentation, growth performance, and six differential genera of rumen bacteria (|r| > 0.6, P < 0.05). (B) The STAMP analysis of KEGG pathway hierarchy level 3 of rumen bacteria. Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences. Groups: CON, untreated diet; ESD, ensiled diet.

Morphology analysis of rumen epithelial tissue

Ensiled diet increased the height (P = 0.012), density (P = 0.036), and surface of rumen papillae (P = 0.001) in sheep, but had no effect on their width (P = 0.459; Table 4; Figure 4A). The thickness of the total epithelia (P = 0.018), stratum corneum (P = 0.040), stratum granulosum (P = 0.042), and stratum spinosum and basale (P = 0.004) were higher in the sheep fed ESD than CON (Table 4; Figure 4B).

Table 4.

Effects of ensiled diet on rumen epithelial papillae related parameters of Tibetan sheep¹

Items	Groups2		P-value
	CON	ESD
Papillae morphology
Height, mm	3.13 ± 0.21	3.47 ± 0.22	0.012
Width, mm	2.01 ± 0.14	2.07 ± 0.11	0.459
Density, N/cm2	52.57 ± 4.79	58.57 ± 4.72	0.036
Surface, mm2/cm2	663.18 ± 75.37	838.92 ± 94.81	0.001
Thickness of different stratum, μm
Total epithelia	113.18 ± 14.64	143.05 ± 16.51	0.018
Stratum corneum	21.71 ± 1.40	24.79 ± 2.60	0.040
Stratum granulosum	22.73 ± 4.02	28.45 ± 5.17	0.042
Stratum spinosum and basale	68.73 ± 14.72	89.82 ± 19.54	0.004

¹Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences.

²Groups: CON, untreated diet; ESD, ensiled diet.

Figure 4.

Representative exterior photos (A) and micrographs (B) of the rumen epithelium morphology in Tibetan sheep. Effects of ensiled diet on mRNA expression of Cell cycle regulation and Apoptosis-related genes in rumen epithelium of Tibetan sheep (C). Effects of ensiled diet on mRNA expression of Tight junction and VFA transport protein-related genes in rumen epithelium of Tibetan sheep (D). CDK2, cyclin-dependent Kinase 2; Bcl-2, B-cell lymphoma-2; BAD, Bcl-2 associated agonist of cell death; ZO-1, zonula occludens-1; DRA, down-regulated in adenoma; AE2: Anion exchanger 2; MCT1, monocarboxylate transporter isoform 1; Na⁺/K⁺-ATPase, sodium/potassium pump;.νH⁺-ATPase, vacuolar-type ATPase; NHE1, sodium/hydrogen antiporter isoform 1. Data results are presented as means ± SD, n = 7. P < 0.05 represents statistically significant differences. β-Actin was used as a housekeeping gene to normalize the mRNA expression levels of each gene, and the 2^−ΔΔCT method was used to analyze the data. Groups: CON, untreated diet; ESD, ensiled diet.

Gene expression related to nutrient transport in rumen epithelial tissues

The genes related to cell cycle regulation (CDK4, CyclinA2, and CyclinD1), tight junction (ZO-1 and Occludin), Bcl-2, and VFA transport protein (MCT1, MCT4, Na⁺/K⁺-ATPase, and NHE3) were upregulated (P < 0.05, Figure 4C and D), while Caspase 3 was downregulated (P = 0.025, Figure 4C) in rumen epithelial tissues of sheep fed ESD.

Discussion

Lignocellulose consists of cellulose, hemicellulose, and lignin. Rumen microorganisms degrade lignocellulose, providing primary energy source for the ruminants (Gharechahi et al., 2023). The degradation process requires rumen microorganisms to physically attach to their substrate. However, complex structure of plant cell wall hinders this interaction, leading to low utilization of lignocellulose (Gharechahi et al., 2023). Ensiled pretreatment is an effective method for improving the utilization of lignocellulosic feeds. Zhou et al.(2024) also found that the addition of tail vegetable silage to lamb diets increased ADG and average daily feed intake. Xu et al. (2023) reported that ensiled rice straw resulted in a loose porous structure and a larger surface area, which enhanced the sites for the attachment of rumen microbiota and improved the digestibility of rice straws. In the present study, ensiled diet increased the ADG of Tibetan sheep. The improved ADG could be attributed to two factors: 1) primarily due to the higher organic acid content, the diet’s palatability increased, resulting in increased DMI (Zhang et al., 2022), and 2) ensiled diet may have facilitated the physical attachment of the rumen microorganisms to the diet’s surface, leading to more effective degradation of lignocellulose (Vahidi et al., 2021).

Rumen is the main organ of the gastrointestinal tract and previous research has found that the silage can change rumen fermentation parameters (Song et al., 2023). In the present experiment, the total VFA concentration in Tibetan sheep was higher on ESD, which may be due to the higher DMI. Ruminal pH was lower in the ESD group as a result of the higher VFA levels. ESD resulted in an increased proportion of propionate but decreased the proportion of acetate and the A/P ratio. Since propionate serves as a precursor for gluconeogenesis in the liver of ruminants, a higher proportion of propionate suggests improved energy efficiency in sheep (Fan et al., 2021). Rumen microorganisms degrade ammonia to form NH₃-N, and then use NH₃-N as nitrogen source to synthesize MCP (Zhou et al., 2019). Additionally, the increase in the copy numbers of bacteria and protozoa in the ESD group may also lead to higher levels of NH₃-N and MCP.

In this study, PCoA analysis showed a significant separation of microbiota at both the rumen phylum and genus levels. Specifically, at the phylum level, the relative abundance of Firmicutes was higher and that of Bacteroidota was lower in the ESD group. Although both Firmicutes and Bacteroidota are the dominant phyla in the rumen, energy absorption is mainly caused by the increase in the relative abundance of Firmicutes rather than Bacteroidota (Ma et al., 2023). Bacteroidota mainly degrade carbohydrates and proteins, while Firmicutes are important cellulolytic bacteria (Wang et al., 2023b). Firmicutes degrade fiber and produce large amounts of VFA, which could explain the greater ADG in sheep on ESD (Turnbaugh et al., 2006). At the genus level, the relative abundance of Ruminococcus, Lachnospiraceae_AC2044_group, and Lachnospiraceae_XPB1014_group was higher in sheep fed ESD. Ruminococcus, belongs to Ruminococcaceae, which are one of the major cellulolytic bacteria in rumen, rich in glycoside hydrolases and carbohydrate-binding modules (Stewart et al., 2018). Most members of the Lachnospiraceae, including the Lachnospiraceae_AC2044_group and Lachnospiraceae_XPB1014_group, as well as members of the Ruminococcus, have been shown to degrade cellulose and hemicellulose into VFA (Biddle et al., 2013). This is consistent with the result of Spearman’s correlation analysis in this study. In addition, the Christensenellaceae_R-7_group, which taxonomically belongs to Clostridia, has also been reported to produce propionate (Wang et al., 2023a). The changes in genera between the two groups are also evident in Prevotellaceae_UCG-003. As a member of the Prevotella genus, Prevotellaceae_UCG-003 can efficiently utilize starch and protein to produce acetate (Xue et al., 2020). In this study, Prevotellaceae_UCG-003 was found to be positively correlated with A/P ratio and negatively correlated with the proportion of propionate, which is consistent with the study of Wang et al. (2023a). The decrease in Prevotellaceae_UCG-003 and the increase in Christensenellaceae_R-7_group may be the main reasons for the higher proportion of propionate in the rumen of sheep fed ESD. Jiang et al. (2023) speculated that Veillonellaceae_UCG-001 may also be a cellulolytic bacterium in the rumen. Nevertheless, the specific reason for lower relative abundance of Veillonellaceae_UCG-001 on ESD than CON cannot be explained and needs further investigation. The results of STAMP analysis demonstrated that the microbiota in the rumen of sheep fed ESD was enriched in the propionate metabolism pathway. These findings suggest that ESD has the potential to enhance rumen propionate production, energy efficiency and subsequently improve the growth performance of Tibetan sheep. However, it is important to note that although PICRUSt2 offers tools for predicting microbial function, additional research is necessary to validate these predictions.

The VFA not only provide a lot of energy for ruminants, but also promotes the growth and development of rumen epithelium (Yang et al., 2020). Malhi et al. (2013) demonstrated that ruminal epithlium is composed of four stratums: stratum corneum, stratum granulosum, stratum spinosum, and basale. By examining the rumen papillae, we found that the height, density, and surface area of rumen papillae were improved in the rumen of sheep fed ESD. The morphological analysis also showed that the ESD increased the thickness of different stratums and total ruminal epithlium. The larger the surface area of the papillae, the greater is the contact area with digesta and thus more VFA absorption through rumen wall. This could explain the greater ADG of sheep on ESD from the perspective of rumen tissue morphology. Furthermore, we also examined the relative expression of genes related to cell cycle regulation, apoptosis, tight junctions, and VFA transport proteins in the rumen epithelium.

Ruminal epithelial development is associated with cell proliferation, which is mainly regulated by cyclins and CDKs (Sherr, 1993; Yang et al., 2020). In the present study, the gene expression levels of CDK4, CyclinA2, and CyclinD1 were significantly increased in sheep fed ESD. The five successive phases, G₀, G₁, S, G₂, and M constitute the animal cell cycle, while CyclinD1 binds to CDK4 to form a complex that accelerates the passage of cells through G₁ to S phase (Mathew et al., 2010). Malhi et al. (2013) demonstrated that the increased expression levels of CyclinD1 and CDK4 in rumen epithelial cells of goats contribute to the proliferation induced by butyrate. The expression level of CyclinA2 gradually increases after entering S phase, forming the CyclinA2–CDK complex to drive chromosome replication and subsequently participates in the initiation of mitosis (Loukil et al., 2015; Kanakkanthara et al., 2016). Apoptosis is an important factor affecting ruminal epithelial development, and we found that the expression level of Caspase 3 was downregulated while Bcl-2 was upregulated in sheep fed ESD. Caspase 3 is a key enzyme in the mitochondrial-dependent apoptosis pathway, and Bcl-2 expression is reduced to activate Caspase3, which ultimately leads to apoptosis (Wu et al., 2016). In the present study, ESD accelerated cell cycle progression leading to cell proliferation and inhibit cell apoptosis. This eventually leads to an increased surface area and thickness of the ruminal papillae. However, an intact ruminal epithelial barrier is a prerequisite for nutrient absorption and body health. Chen et al.(2023) reported that when the intestinal epithelial barrier is damaged, the expression levels of ZO-1, Occludin, Claudin-1, and Claudin-4 are significantly reduced. In the present study, the expression levels of ZO-1 and Occludin in the sheep fed ESD were significantly higher than those on CON, indicating that ESD was more conducive to rumen homeostasis and animal health.

In the rumen epithelium, the main roles of MCT1 and MCT4 are to mediate VFA transport and maintain intracellular pH. The difference is that MCT1 is present on the blood-oriented side, while MCT4 is located on the luminal-oriented side (Benesch et al., 2014). These two monocarbonic acid transporters cooperate to transport H⁺/VFA⁻ from the lumen to the blood (Kirat et al., 2007). In previous studies (Aschenbach et al., 2011; Yang et al., 2012), it has been demonstrated that Na⁺/K⁺-ATPase can expel Na⁺ from the cell and provide the driving force for NHE3 to expel intracellular H⁺, which plays an important role in maintaining intracellular pH. Our results showed that MCT1, MCT4, Na⁺/K⁺-ATPase, and NHE3 were significantly increased on the ESD, which may be due to the synergy of several proteins to promote VFA transport from the rumen epithelium. Jing et al. (2018) demonstrated that concentrate supplementation increased the mRNA expression levels of MCT1, MCT4, and NHE3 in ruminal epithelial cells. In previous studies (Yang et al., 2012; Yan et al., 2014), it has been reported that the mRNA expression level of NHE3, MCT1, MCT4, and Na⁺/K⁺-ATPase was proportional to the concentration of VFA. In the present study, the ESD promoted cell proliferation and VFA transport, and reduced cell apoptosis. On the other hand, the intracellular H⁺ load was increased, so the H⁺ discharge rate was accelerated to stabilize the intracellular pH. The ESD improved the relative expression of genes related to nutrient transport, but the amount of intracellular functional proteins is not necessarily consistent with molecular gene adaptation (Steele et al., 2012; Wang et al., 2022). Therefore, further investigation into the number of nutrient transport functional proteins in the rumen epithelium is needed.

Conclusion

Compared to the CON group, sheep fed ESD exhibited higher DMI and ADG, primarily due to the improved palatability of ESD. Additionally, ESD increased the relative abundance of cellulolytic bacteria and the proportion of propionate in the rumen. Furthermore, it altered the morphology of the rumen epithelium and upregulated the relative mRNA expression of genes related to nutrient transport, thereby enhancing the energy utilization efficiency of Tibetan sheep. These findings offer theoretical guidance for improving the nutritive value of agricultural byproducts fed to Tibetan sheep on the Qinghai-Tibet Plateau.

Glossary

Abbreviations

A/P

acetate/propionate ratio

ADG

average daily gain

ASV

amplicon sequence variant

Bcl-2

B-cell lymphoma-2

CDK2

cyclin-dependent kinase 2

CON

untreated diet

DMI

dry matter intake

ESD

ensiled diet

MCP

microbial crude protein

MCT1

monocarboxylate transporter isoform 1

Na⁺/K⁺-ATPase

sodium/potassium pump

NH₃-N

ammonia nitrogen

NHE1

sodium/hydrogen antiporter isoform 1

PCoA

principal coordinate analysis

qRT-PCR

quantitative real-time polymerase chain reaction

STAMP

statistical analysis of taxonomic and functional profiles

VFA

volatile fatty acids

ZO-1

zonula occludens-1

16S rRNA

16S ribosomal RNA

β-actin

beta actin

Conflict of interest statement

The authors declare no conflicts of interest.

Data Availability

Raw sequencing reads of bacterial 16S rRNA gene of rumen fluid samples are deposited in the NCBI Sequence Read Archive database (BioProject accession number PRJNA1063203).