Effects of Substituting Dietary Corn with Grain Byproducts on Fattening Hu Sheep: Growth Performance, Rumen Fermentation, Energy-Nitrogen Metabolism and Greenhouse Gas Emissions

Simple Summary

Replacing corn with grain byproducts in fattening Hu sheep diets numerically reduced formula cost and nitrogen utilization while increasing methane emissions, but did not significantly alter growth performance. Consequently, it failed to improve overall economic returns compared to the control diet. Based on this, bacterial-enzymatic fermentation treatment of the byproducts would lower the relative abundance of methanogens and greenhouse gas emissions, and improve finishing economic returns. Thus, bacterial-enzymatic fermentation presents a potential strategy for achieving cost-effective corn substitution.

Abstract

Grain byproducts can serve as cost-effective alternatives to corn, but may lead to reduced production performance and increased greenhouse gas emissions. This study aimed to investigate the effects of replacing corn with the grain byproducts (wheat bran, sprayed corn bran) subjected to bacterial-enzymatic fermentation treatment or not in Hu sheep, mainly focusing on production performance, energy-nitrogen metabolism, rumen fermentation and greenhouse gas emissions. A total of fifty-four 6-month-old Hu sheep were divided into three groups, with 6 pens per group and 3 sheep per pen, and then randomly allocated to one of the three dietary groups for 60 days, i.e., a control group (CON), a group (RC) that corn was partially (~42%) replaced with grain byproducts, and a group (BF) that corn was partially replaced by fermented grain byproducts. Compared with the CON group, the RC group showed numerically lower rumen total volatile fatty acid (TVFA) concentration and its propionate proportion, nitrogen retention content (NR; −10.22%) and its retention ratio (NR/NI decreased by 4.27 percentage points, absolute reduction from 24.30% to 20.04%), corresponding to a relative decrease of 17.6%.) as well as a numerically reduced net profit (−2.18%) with a decreased feed price (−¥0.16/kg TMR). Meanwhile, the RC group showed a significant increase in the relative abundance of Methanobrevibacter (p < 0.05), accompanied by numerically higher daily methane emissions (+6.14%) and emission intensity (+4.08%), although these methane-related differences did not reach statistical significance (p > 0.05). Compared to the RC group, the BF group resulted in a numerical increase in feed price (+¥0.03/kg TMR), net profit (+27.93%), TVFA concentration, propionate proportion, NR (+28.17%), NR/NI (an increase of 5.38 percentage points), the relative abundance of Prevotella, Shuttleworthia and Succinivibrio as well as the decrease of fecal nitrogen (FN; −12.29%), daily methane emissions (−8.75%), emission intensity (−5.83%) and the relative abundance of Methanobrevibacter. In summary, replacing dietary corn by 42% with wheat bran and sprayed corn bran numerically reduced formula cost and nitrogen utilization, while increasing methane emissions and methanogens abundance, without significantly affecting growth performance. This combination led to no improvement in economic returns for fattening Hu sheep. Bacterial-enzymatic fermentation treatment of these byproducts could mitigate these drawbacks, being superior energy-nitrogen metabolism and lower greenhouse gas emissions intensity, presenting a potential strategy for cost reduction and efficiency enhancement. Further research with larger sample sizes is warranted to confirm these findings and support broader application.

1. Introduction

In recent years, the trend of feed resource shortages and rising raw material prices have posed significant challenges to the livestock industry. Corn, as a mainstream feed ingredient, faces dual pressures from high production costs and competition for grain resources [1], prompting us to seek innovative approaches to mitigate its impact. Grain byproducts, being inexpensive and abundant, can serve as alternative raw materials. The utilization of grain byproducts such as wheat bran, rice bran and corn bran in animal feed holds significant potential for reducing carbon emissions in the livestock sector and improving economic returns [2,3]. However, replacing corn with such by-products typically reduces the proportion of concentrate in the diet, which may lead to inferior production performance and increased greenhouse gas emissions, especially methane (CH₄) [4]. Methane is a key contributor to the global greenhouse effect. Although its total emissions are lower than those of CO₂, methane has a global warming potential approximately 28 times greater than CO₂ per unit mass over a 100-year period [5]. Moreover, methane emissions also represent a loss of 2–15% of dietary energy [6,7], leading to wastage of feed resources. Based on the above, replacing corn with lower-quality and inexpensive grain byproducts may reduce feed costs but lower feed utilization efficiency and increase greenhouse gas emission, finally failing to achieve cost reduction in whole.

Feed processing is proved to be an effective way to enhance the feeding value of feedstuff and thus improve animal performance, primarily involving physical, chemical and biological treatments. Fermentation and enzymolysis are commonly-used methods to pretreat forage and roughages, especially unconventional one [8]. Bacterial-enzymatic fermentation technology achieves an organic integration of fermentation and enzymatic hydrolysis, not only enhancing microbial conversion efficiency of macromolecular substances in feed but also improving feed nutritional value and utilization [9,10]. Studies have shown that supplementing Bacillus licheniformis in the diet of sheep can reduce CH₄ production and improve energy and protein utilization [11]. Combined treatment with Lactiplantibacillus plantarum and cellulase promotes silage fermentation and enhances nutritional value, and also reduces methane yield and increases degradation rate [12]. Dietary supplementation with L. plantarum enhances ruminal propionate concentration and modulates rumen fermentation patterns in ruminants [13]. L. plantarum modulates the bacterial community during ensiling, improving the fermentation quality of whole-plant corn silage. It further regulates rumen microbial interactions, effectively inhibiting methanogenesis and enhancing feed digestibility [14].

Small ruminants, particularly sheep, play a vital role in meat and fiber production [15], while also contributing to methane emissions [16,17]. The Hu sheep, an indigenous Chinese breed, is considered the breed of choice for indoor sheep farming in China [18]. Investigating the disease in all animals. efficient utilization of by-products holds significant importance for reducing costs and enhancing production efficiency in Hu sheep operations.

Given the potential that bacterial-enzymatic fermentation can enhance the availability of nutrients and modify the rumen function, we hypothesize that pre-treating by-products with bacterial-enzymatic fermentation can improve energy and nitrogen utilization in Hu sheep while promoting beneficial changes in rumen fermentation. Bacterial-enzymatic fermentation may be an effective way to increase the utilization rate of by-products and thereby reduce the environmental impact of animal husbandry. Therefore, this study aimed to investigate the effects of replacing corn with the grain byproducts (wheat bran, sprayed corn bran) subjected to microbial-enzymatic fermentation treatment or not in Hu sheep, mainly focusing on production performance, energy-nitrogen metabolism, rumen fermentation and greenhouse gas emissions, ultimately making clear the impacts of grain byproducts inclusion in diets.

2. Materials and Methods

2.1. Ethics Committee Approval

This study was approved by the Animal Welfare Committee of China Agriculture University, Beijing, China (Aw31015202-1-5). These procedures adhered to the principles and regulations for ethical protection in human and animal biological science and technology in China.

2.2. Animals, Diets and Management

A total of 54 six-month-old Hu male sheep with similar initial body weight (34.30 ± 1.27 kg) were divided into three groups, with 6 pens per group and 3 sheep per pen, and then randomly allocated to one of the three dietary groups (Table 1): a control group (CON) fed a basal diet, a group (RC) in which corn was partially replaced (~42% of dietary corn) with a mixture of unfermented grain byproducts (wheat bran and sprayed corn bran), and a group (BF) in which corn was replaced at the same percentage (~42%) with an identical mixture of byproducts that had undergone bacterial-enzymatic fermentation. The fermentation process was conducted as follows: the grain byproduct mix was thoroughly mixed with a commercial fermentation agent (AP600A/B, Borui Technology Co., Ltd., Changchun, China) at an inclusion rate of 0.05% (wt/wt, on a dry matter basis). This agent contained a defined consortium of microorganisms, including Lactiplantibacillus plantarum (>6.0 × 10⁹ CFU/g), Bacillus subtilis (>1.8 × 10¹⁰ CFU/g), and Saccharomyces cerevisiae (>2.0 × 10⁹ CFU/g), along with fibrolytic enzymes (cellulase, xylanase, mannanase, amylase). The mixture was then adjusted to approximately 50% moisture with water, packed tightly into specialized oxygen-barrier bags to create anaerobic conditions, and fermented at ambient temperature (~20 °C) for 10 days, consistent with the manufacturer’s recommended protocol for producing stable fermented feed. To ensure strict dietary consistency, the entire experiment was conducted using a single, large batch of this fermented material, which was prepared before the trial commenced and stored appropriately after fermentation.

Table 1.

Ingredients and nutrient levels of experimental diets.

Items	Treatments
	CON	RC	BF
Ingredients (% DM basis)
Corn	38.00	22.10	22.10
Soybean meal	6.00	6.00	6.00
canola seed meal	5.00	5.00	5.00
Corn germ meal	3.18	3.18	3.18
Distillers dried grains with solubles	3.60	3.60	3.60
Wheat bran	——	12.40	12.40
Sprayed corn bran	——	3.60	3.60
Calcium carbonate	1.20	1.50	1.50
Dicalcium phosphate	0.80	0.40	0.40
Sodium bicarbonate	0.90	0.90	0.90
Salt	0.60	0.60	0.60
Corn straw silage	22.00	22.00	22.00
Dry distillers’ grains	18.00	18.00	18.00
Mold remover	0.12	0.12	0.12
Premix 1	0.6	0.60	0.60
Nutrient content 2 (% DM basis)
ME, MJ/kg	9.25	9.12	9.13
CP, %	14.14	14.78	14.35
NDF, %	31.26	38.69	37.54
ADF, %	18.19	21.69	21.52
Ca, %	1.09	1.14	1.18
P, %	0.50	0.52	0.50

¹: Each kilogram of product contains 800–2400 mg Cu, 4000–20,000 mg Fe, 4000–10,000 mg Mn, 5000–15,000 mg Zn, 70–200 mg I, 20–40 mg Se, 40–60 mg Co, Vitamin D3 300,000–700,000 IU, Vitamin A acetate 800,000–1,500,000 IU, d1-α-tocopheryl acetate 3000 mg. ²: Data are measured values.

This study was conducted at the sheep farm of Runlin Animal Husbandry Co., Ltd., Linqing, China, and lasted for 67 days including a 7-day adaptation period and 60 days of data collection (Figure 1). The sheep were fed on total mixed ration (TMR) at 08:00 and 16:00 daily, with free access to clean water and unified disinfection and immunization procedures. All sheep remained in good health throughout the experiment. Daily health monitoring confirmed the absence of any clinical signs of disease in all animals.

Figure 1.

A schematic overview of the experimental timeline.

Throughout the metabolism and gas sampling periods, all sheep were maintained under ad libitum feeding conditions without physical restraint. No feed refusal or intake depression was evident during these procedures, and daily DMI remained stable and comparable across groups over the entire trial period (Table 2), confirming that the measurement protocols did not confound treatment effects on intake or derived parameters.

Table 2.

Effects of dietary treatments on growth performance of fattening Hu sheep (n = 6).

Items	Treatments			SEM	p Value
	CON	RC	BF
Initial body weight (IBW), kg	34.62	34.36	33.96	0.253	0.601
Final body weight (FBW), kg	46.61	45.91	46.13	0.430	0.811
ADG, g/d	199.91	192.52	202.69	5.635	0.771
DMI, kg/d	1.69	1.75	1.70	0.022	0.449
F/G	8.62	9.16	8.42	0.216	0.348
TMR price, ¥/kg DM	2.20	2.04	2.07	-	-
Daily feed cost, ¥/head	3.72	3.57	3.52	-	-
Total feed cost, ¥/head	223.47	214.24	211.28	-	-
Revenue from weight gain, ¥/head	287.87	277.23	291.87	-	-
Net profit, ¥/head	64.40	62.99	80.59	-	-
Δ Profit *	0.00	−1.41	16.19	-	-

*: The difference in net profit per head for each treatment group (RC, BF) compared to the control group (CON). Values are presented as means.

2.3. Growth Performance

On day 0 and day 60 of the experimental period, all sheep were respectively weighed before morning feeding to record initial and final body weights, and then to calculate average daily gain (ADG). The amounts of total mixed ration (TMR) offered and the refusals were weighed daily to calculate dry matter intake (DMI) and feed-to-gain ratio (F/G). Diet samples were collected periodically to determine dry matter (DM) content and nutritional components.

The formulas for growth performance and economic parameters were as follows:

DMI (kg/(head·d)) = Total DMI per pen/(days × number of animals per pen) (1)

ADG (g/(head·d)) = (Final BW − Initial BW)/days (2)

F/G = DMI/ADG (3)

Feed cost (¥/head) = TMR price (¥/kg DM) × DMI (kg/head) (4)

Revenue from weight gain (¥/head) = (Final BW − Initial BW) × 24 (5)

Net profit (¥/head) = Revenue from weight gain − Feed cost (6)

Economic analysis was based on a lamb market price of ¥24 per kg live weight. The economic analysis was based on fixed market prices and directly calculated from group-mean performance data; therefore, no statistical inference was applied to net profit or feed cost.

2.4. Digestion and Metabolism

On day 40 of the trial, one sheep per pen, with a body weight closest to the pen’s average weight on day 30, was selected for the subsequent 3-day metabolism trial. Fecal and urine samples were collected using the spot sampling method, with urine collected via urinary bags and fecal samples obtained by rectal sampling at 10:00 and 16:00 daily. Finally, all fecal samples collected from each sheep were thoroughly mixed with 10% sulfuric acid added for nitrogen fixation, dried at 65 °C until constant weight, and ground through a 40-mesh sieve. Urine samples were also mixed with 10% sulfuric acid added, and stored at –20 °C for subsequent analysis. The gross energy of the samples was determined using an oxygen bomb calorimeter [19] (5E-AC8018, Changsha Kaid Measurement and Control Instrument Co., Ltd., Changsha, China), and nitrogen content was measured with an automated Kjeldahl nitrogen analyzer (KN680, Jinan Alva Instrument Co., Ltd., Jinan, China).

At each sampling time, approximately 30–50 g of fresh feces were obtained via rectal grab sampling from each selected sheep. To facilitate gentle and rapid collection while minimizing stress and discomfort, a small amount of glycerin enema (Kaisailu^®, Fuxin Longfei Pharmaceutical Co., Ltd., Fuxin, China) was applied to lubricate the anal region and stimulate defecation when necessary. This volume (approximately 200 g) was fully sufficient for the determination of nitrogen, gross energy, and acid-insoluble ash (AIA), and avoided excessive manipulation of the animals.

Daily fecal DM output was estimated using acid-insoluble ash (AIA) as an indigestible marker, with AIA concentrations determined in diet and fecal samples. Fecal nitrogen output (FN) was calculated as the product of fecal DM output and fecal N concentration. Daily urine volume was estimated from urinary creatinine concentration using a fixed creatinine excretion rate of 29 mg/kg BW/d. Urinary nitrogen output (UN) was calculated as urine volume multiplied by urinary N concentration. Feed intake during the metabolism trial was measured using the same procedure described in Section 2.3; daily DMI per sheep was derived from pen-based measurements.

The following formulae were applied to calculate nutrient digestibility and metabolic rates:

Digestible nitrogen (DN) = Nitrogen intake (NI) − Fecal nitrogen (FN) (7)

Nitrogen retention (NR) = NI − FN − Urinary nitrogen (UN) (8)

Nitrogen digestibility (%) = DN/NI × 100 (9)

Nitrogen retention rate (%) = (NR/NI) × 100 (10)

Digestible energy (DE) = Gross energy intake (GE) − Fecal energy (FE) (11)

Metabolizable energy (ME) = DE − Urinary energy (UE) − Methane energy (CH₄-E) (12)

Gross energy digestibility (%) = (DE/GE) × 100 (13)

Gross energy metabolizability (%) = (ME/GE) × 100 (14)

Digestible energy metabolizability (%) = (ME/DE) × 100 (15)

FN (g/d) = Fecal N concentration (g/kg DM) × Fecal DM output (kg/d) (16)

AIA intake (g/d) = DMI (kg/d) × AIA content in diet (g/kg DM) (17)

Urine volume(L/d) = Creatinine excretion rate (mg/kg BW/d) × BW (kg)/Urinary creatinine concentration (mg/L) (18)

2.5. Greenhouse Gas Emission

On day 43 of the trial, 2 sheep were selected from each pen (close to the average weight of that pen) for a 16-day monitoring period (including a 2-day adaptation period and a 14-day data collection period) to measure greenhouse gas emissions. Gas emission of each sheep was measured for 5 min at 3-h interval following the method described by Wang et al. [20,21], using an Automated Head-Chamber (AHC) system (AHC3.0, Institute of Subtropical Agriculture, Chinese Academy of Sciences, Changsha, China). It comprises a fully enclosed feed bin, fan, air filter, collection pipe, airflow meter, and gas sampler (Supplemental Figures S1 and S2). The fan induces airflow at 300 L/s for complete breath collection. Air is filtered, measured by a vortex flowmeter, and sampled at 2 L/min through a 2.0 μm filter, with real-time analysis every second using a GGA-30p Greenhouse Gas Analyzer (Los Gatos Research, Los Gatos, CA, USA). Calibration is performed by instantaneously releasing 0.5 L of pure CH₄ (>99.9%) at the trough base and bin center. Methane production is derived from detected concentration versus injected volume. Each site is tested three times (six total measurements), ensuring a capture rate between 90% and 110% [20]. To ensure the accuracy of the acquired data, capture rate calibration must be performed daily.

Greenhouse gas emissions (g/d) = Greenhouse gas emission rate (g/h) × 24 h (19)

Greenhouse gas emissions (g/kg DMI) = Greenhouse gas emissions (g/d)/DMI (kg/d) (20)

Greenhouse gas emissions (g/kg ADG) = Greenhouse gas emissions (g/d)/ADG (kg/d) (21)

Emission intensity (kg CO₂e) = (CH₄ emission × 28 + CO₂ emission)/ADG (22)

2.6. Rumen Fermentation

On the final day of the trial, rumen fluid samples (50 mL/animal) were collected via oral-gastric tubes before morning feeding. The rumen tube was flushed with clean water before each new sample collection, and the first 10–15 mL of each sample was discarded. Collected samples were stored at −80 °C for subsequent analyses. Ammonia nitrogen (NH₃-N) concentration was determined using colorimetric methods [22]. Volatile fatty acid (VFA) concentration was measured using a gas chromatography-mass spectrometry system [23] (Agilent 5975C, Agilent Technologies, Santa Clara, CA, USA). Microbial crude protein (MCP) was determined using the Bradford assay [24]. Subsequently, 16S rDNA sequencing analysis was performed.

Microbial crude protein (MCP) in rumen fluid was determined using the differential centrifugation and alkaline extraction method described by Makkar et al. (1982) [24], with protein quantified by the Coomassie Brilliant Blue (Bradford) assay. Briefly, strained rumen fluid was shaken to detach particle-associated bacteria, then centrifuged at 500× g for 10 min to remove protozoa and residual feed particles. The supernatant was centrifuged at 25,000× g for 20 min to pellet the bacterial fraction. The pellet was washed with McDougall’s buffer, re-centrifuged, and resuspended in 0.25 N NaOH. After heating in a boiling water bath for 10 min, the lysate was centrifuged (25,000× g, 30 min), and the supernatant was used for protein determination. This procedure minimizes contamination from soluble dietary nitrogen and provides a reliable estimate of relative microbial protein synthesis across treatments.

2.7. 16S rDNA Sequencing and Analysis

The collected rumen fluid samples were subjected to sequencing analysis conducted by Shanghai Ling’en Biotechnology Co., Ltd. Total genomic DNA was extracted using the E.Z. N.A.^® Soil DNA Kit (Omega Bio-Tek, Norcross, GA, USA) according to the manufacturer’s instructions. The hypervariable V3-V4 region of the bacterial 16S rDNA was amplified by PCR using primers 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GACTACHVGGGTATCTAATCC-3′). PCR products were purified using the AxyPrep DNA Gel Extraction Kit (Corning, Corning, NY, USA) and quantified using the Quantus™ Fluorometer (Promega Corporation, Madison, WI, USA). Sequencing was performed on Illumina’s MiSeq PE300 platform. Raw sequencing reads underwent quality control using fastp software (version 0.20.0), assembly with FLASH software (version 1.2.7), and ASV clustering via UPARSE software (version 7.1). Species composition analysis, α-diversity analysis, β-diversity analysis and differential analysis were completed on the LingEnBio platform (http://delivery.biozeron.com) [25]. Alpha and beta diversity metrics were computed in QIIME2. Bacterial taxonomic profiles were analyzed using relative abundance. Beta-diversity was analyzed using Weighted UniFrac distances and the resulting distance matrix was visualized via Principal Coordinate Analysis (PCoA). Differentially abundant genera and phyla were identified via Kruskal–Wallis tests, with statistical significance defined as p < 0.05. Linear discriminant analysis effect size (LEfSe; LDA score ≥ 3.0, p < 0.05) was executed using the Segata LEfSe tool (version 1.1.2) [26].

2.8. Statistical Analyses

Data (growth performance, digestion and metabolism, greenhouse gas emission, rumen fermentation) were organized using Excel software and subjected to single-factor analysis of variance (ANOVA) using SPSS 26.0 (SPSS, Inc., Chicago, IL, USA), followed by Duncan’s multiple range test for multiple comparisons. Statistical significance was defined as p < 0.05.

3. Results

3.1. Growth Performance and Economic Benefits

As summarized in Table 2, replacing approximately 42% of corn in the diet had no significant effect on the production performance of fattening Hu sheep. No significant differences (p > 0.05) were observed in FBW, ADG, DMI, or F/G among the three groups. Numerically, compared with the CON group, the RC group showed a lower ADG and a higher F/G. The BF group exhibited a numerically higher ADG (202.69 g/day) and the lowest F/G (8.42). In terms of economic benefits, the RC and BF groups reduced the TMR price by ¥0.16/kg and ¥0.13/kg, respectively. The BF group achieved the numerically highest net profit (¥80.59/head), representing a 25.14% and 27.93% numerical increase over the CON group (¥64.40/head) and RC group (¥62.99/head), respectively. Among the three groups, BF showed the best numerical economic benefits.

3.2. Energy and Nitrogen Metabolism

The effects of replacing corn with cereal by-products on energy and nitrogen metabolism in fattening Hu sheep are shown in Table 3. Both the RC and BF groups exhibited significantly higher nitrogen intake (NI) compared to the CON group (p < 0.05). Digestible nitrogen (DN) was significantly greater in the BF group than in the CON group (p < 0.05), while the RC group showed an intermediate value that did not differ significantly from either the CON or BF groups. The BF group exhibited the highest DN/NI (58.59% vs. 61.10% vs. 66.35%) and NR/NI (24.30% vs. 20.04% vs. 25.42%). Compared with the CON group, the RC group reduced NR by 10.22% and NR/NI decreased by 4.27 percentage points, absolute reduction from 24.30% to 20.04%, while increasing CH₄-E by 6.14%. Compared with the RC group, the BF group increased NR by 28.17% and 5.38 percentage points of NR/NI, while decreasing FN by 12.29% and CH₄-E by 8.75%.

Table 3.

Effects of dietary treatments on energy and nitrogen metabolism in fattening Hu sheep (n = 6).

Items	Treatments			SEM	p Value
	CON	RC	BF
NI, g/d	37.31 b	40.35 a	40.67 a	0.638	0.049
FN, g/d	15.47	15.60	13.69	0.467	0.176
UN, g/d	12.83	16.65	16.62	0.823	0.085
DN, g/d	21.84 b	24.74 ab	26.99 a	0.785	0.016
NR, g/d	9.01	8.09	10.37	0.620	0.338
DN/NI, %	58.59 b	61.10 ab	66.35 a	1.310	0.036
NR/NI, %	24.30	20.04	25.42	1.552	0.348
GE intake, MJ/d,	33.88	34.25	33.91	0.468	0.946
FE output, MJ/d	15.79	15.63	15.93	0.378	0.956
UE output, MJ/d	1.06	1.11	0.88	0.096	0.647
CH4-E, MJ/d	1.80	1.91	1.74	0.044	0.312
DE, MJ/d	18.09	18.62	18.07	0.396	0.831
ME, MJ/d	15.24	15.61	15.45	0.326	0.904
DE/GE, %	53.46	54.22	53.36	0.894	0.721
ME/GE, %	45.03	45.45	45.61	0.766	0.957
ME/DE, %	84.25	83.85	85.49	0.463	0.367

NI = Nitrogen intake; FN = Fecal nitrogen output; UN = Urinary nitrogen output; DN = Digestible Nitrogen; NR = Nitrogen retention; GE = gross energy; FE = fecal energy; UE = urinary energy; CH₄-E = methane energy; DE = digestible energy; ME = metabolizable energy. Values are presented as means. ^a,b different superscript letters: p < 0.05.

3.3. Greenhouse Gas (CH₄ and CO₂) Emissions

As shown in Table 4, the substitution of grain by-products for corn had no significant effect (p > 0.05) on greenhouse gas emissions from fattening Hu sheep. However, numerically, the RC group exhibited increased methane emissions and emission intensity, whereas the BF group showed an opposite trend. Compared to the CON group, no significant treatment effects were detected for any methane- or CO₂-related emission variable (p > 0.05). However, numerically, the RC group showed the highest daily methane emission (34.35 g/d) and emission intensity (9.74 kg CO₂e/kg gain), whereas the BF group exhibited the lowest values for both metrics (31.99 g/d and 9.17 kg CO₂e/kg gain, respectively).

Table 4.

Effects of dietary treatments on greenhouse gas emissions in fattening Hu sheep (n = 6).

Items	Treatments			SEM	p Value
	CON	RC	BF
CH4 emission, g/d	32.36	34.35	31.99	0.764	0.424
CH4 emission, g/kg DMI	19.19	19.60	18.91	0.484	0.859
CH4 emission, g/kg ADG	165.74	179.79	159.97	6.445	0.461
CO2 emission, g/d	928.57	898.39	937.49	18.613	0.694
CO2 emission, g/kg DMI	550.29	513.11	555.69	14.891	0.473
CO2 emission, g/kg ADG	4719.49	4708.18	4695.11	159.678	0.998
Emission intensity, kg CO2e/kg	9.36	9.74	9.17	0.321	0.784

DMI = dry matter intake; ADG = average daily gain; CO₂e = carbon dioxide equivalent; Values are presented as means.

3.4. Rumen Fermentation and Microbial Community

The effects of replacing corn with grain byproducts on ruminal fermentation parameters were presented in Table 5. Compared with the CON group, the RC group showed a reduction in TVFA concentration and propionate proportion with an increased acetate-to-propionate (A:P) ratio (p > 0.05). The BF treatment appeared to mitigate these effects, showing a tendency for increased TVFA and a fermentation profile characterized by a higher molar proportion of propionate. Notably, significant differences were detected in the molar proportions of branched-chain volatile fatty acids. The BF treatment resulted in a significantly lower molar proportion of isobutyrate compared to the CON and RC treatments (p < 0.05). Furthermore, the proportion of isovalerate was highest in the RC group, intermediate in the CON group, and lowest in the BF group (p < 0.05).

Table 5.

Effects of dietary treatments on rumen fermentation in fattening Hu sheep (n = 6).

Items	Treatments			SEM	p Value
	CON	RC	BF
NH3-N, mg/dL	13.39	13.76	11.63	0.488	0.163
MCP, mg/dL	87.48	102.54	97.21	3.307	0.154
TVFA, mM	85.45	81.70	90.36	2.767	0.432
VFA, mol/100 mol
Acetate(A)	66.41	66.65	66.69	0.437	0.996
Propionate(P)	18.24	16.85	18.68	0.479	0.256
Isobutyrate	1.05 a	1.06 a	0.88 b	0.031	0.025
butyrate	11.35	12.48	11.12	0.366	0.262
Isovalerate	1.95 ab	2.07 a	1.67 b	0.065	0.026
valerate	1.00	0.90	0.96	0.039	0.611
A/P	3.66	4.02	3.72	0.106	0.156

TVFA = Total volatile fatty acid; A/P = Acetate/propionate; Values are presented as means; ^a,b different superscript letters: p < 0.05.

Replacement of corn with grain byproducts significantly altered the microbial community in the rumen fluid of Hu sheep. Compared to the CON group, the RC group exhibited significant increases in both the Chao1 index (species richness) and the Shannon index (diversity and evenness). No significant difference in these alpha-diversity indices between the BF and RC groups (Figure 2A). β-diversity analysis of rumen microbiota revealed a significant structural shift in microbial composition among the groups (p < 0.01; Figure 2B). At the Phylum level, Bacteroidota and Firmicutes were the predominant phyla across all groups (Figure 2D), compared with the CON group, the RC group showed a significant increase in the abundance of Firmicutes and a significant decrease in Bacteroidetes, whereas the BF group exhibited no significant difference from the control and demonstrated a trend toward reversing the microbial shift observed in the RC group (Figure 2C). The relative abundance of Euryarchaeota was significantly higher in the RC group than in the CON group (p < 0.05; Figure 2C), whereas the BF group demonstrated a significant reduction in Euryarchaeota compared to the RC group (p < 0.05). At the genus level, the RC group showed significantly lower abundances of Prevotella and Shuttleworthia (p < 0.05), but higher abundances of Methanobrevibacter and Ruminococcus relative to CON group (p < 0.05). The BF treatment partially counteracted these shifts, significantly reducing the abundance of Methanobrevibacter and increasing that of Succinivibrio (p < 0.05), while showing higher Prevotella abundance compared to RC. Based on LDA > 3.0 analysis using LEfSe, this study identified 22 microbial taxa that exhibited intergroup differences (Figure 2E). Specifically, RC group enriched the relative abundance of Methanobrevibacter, and BF group enriched the relative abundance of Succinivibrio.

Figure 2.

Effects of partial replacement of corn in the diet on microbial community of rumen fluid. (A) Alpha diversity assessed by the Chao1 and Shannon indices; (B) Beta diversity visualized by principal coordinate analysis (PCoA); (C) Phyla and genera with significant differences among the top 10 phyla and genera by relative abundance; (D) Relative abundance of the top 10 microbial phyla and genera; (E) Taxonomic biomarkers identified by linear discriminant analysis effect size (LEfSe) (LDA score > 3.0). Groups: CON (control), RC (corn replacement), and BF (bacterial-enzymatic fermentation). The different superscript letters on the histogram represent a significant difference (p < 0.05).

4. Discussion

4.1. Effects of Replacing Corn with Grain Byproducts on Production Performance of Fattening Hu Sheep

Given the current challenges of a depressed lamb market and persistently high feed costs, enhancing growth performance and economic efficiency is paramount for improving production profitability [27]. Compared to corn, wheat bran and sprayed corn bran feed possess higher fiber and nitrogen content, which induced a shift in the microbial structure within the rumen of the RC group. Specifically, the relative abundance of Ruminococcus, a primary fiber-degrading genus, was enriched. Concurrently, an increase in the relative abundance of Methanobrevibacter was observed, leading to elevated methane production and a consequent loss of dietary energy. Furthermore, although NI increased, it was accompanied by a greater UN excretion. This resulted in a reduction in NR/NI and a numerical decrease in ADG, although the difference in ADG was not statistically significant. Consequently, although growth performance was not statistically impacted, the reduction in feed cost was offset by these metabolic inefficiencies, resulting in a slight numerical decrease in net economic return (approximately −¥1.41).

The BF treatment numerically increased ADG by 5.28% and reduced F/G compared to the RC group, achieving a numerically higher net profit of ¥80.59 per head—a 25.14% numerical improvement over conventional diets, suggesting potential for cost reduction and efficiency enhancement. Previous studies have demonstrated that fermented agricultural byproducts can enhance sheep growth performance and economic returns [28,29,30]. For instance, replacing 15% of soybean straw with fermented rice husk powder modulated the rumen microbiota and improved growth performance in fattening Hu sheep [28]. Similarly, corn silage inoculated with L. buchneri and L. plantarum reduced the acetate-to-propionate ratio in ruminal fluid and increased lamb ADG [29]. Whole-plant corn silage treated with bacterial inoculant exhibited higher in situ NDF digestibility than untreated silage, resulting in significantly greater ADG, DMI, and feed conversion efficiency in sheep [30]. These findings are consistent with the economic advantages observed in the present study, bacterial-enzymatic fermentation of grain byproducts may enhance nutrient digestibility and absorption by pre-digesting macromolecules in feed ingredients, improve rumen fermentation parameters, and ultimately lead to increased average daily gain and economic returns.

4.2. Effects of Replacing Corn with Grain Byproducts on Energy and Nitrogen Metabolism of Fattening Hu Sheep

Nitrogen and energy metabolism, along with enteric methane emissions, are critical factors determining both the economic and environmental sustainability of ruminant production [31,32]. In the study, BF group achieved the numerically highest DN and NR/NI values, with DN/NI being significantly higher than CON (p < 0.05), while NR/NI showed only a numerical increase. The RC group showed intermediate values, and compared with the RC group, the BF group exhibited a 5.38 percentage-point higher NR/NI, which aligns with the observed increase in ADG. Fermented feed improves nitrogen utilization through multiple mechanisms: it supplies highly digestible protein and carbohydrate fractions, optimizes ruminal fermentation patterns, and enhances the efficiency of microbial protein synthesis [33,34]_. Studies demonstrate that fermenting total mixed ration silage containing agricultural by-products with epiphytic lactic acid bacteria improves nitrogen utilization in ewes, the application of silage additives demonstrated a neutral effect on feed intake, nutrient digestibility, and ruminal fermentation parameters [35]. Related study indicates that substituting soybean meal with fermented citric waste significantly elevated rumen pH, ammonia nitrogen, plasma urea nitrogen, and bacterial populations 4 h post-feeding, while volatile fatty acid profiles and energy allocation remained unaffected across treatments [36]. Multiple feeding trials demonstrate that microbial-enzyme treatment of feed ingredients consistently maintains or enhances animal production performance while mitigating nitrogen excretion in feces and urine, thereby reducing environmental impact [37,38]. During fermentation, microbial and enzymatic activity pre-degrades proteins into smaller peptides and free amino acids, enhancing their bioavailability. Concurrently, this process degrades anti-nutritional factors, thereby improving protein utilization and promoting greater nitrogen absorption by the intestinal epithelium rather than fecal or urinary excretion [39].

The improved nitrogen retention (NR) and digestibility (DN/NI) observed in the BF group, despite similar nitrogen intake across RC and BF groups, suggest a shift in the metabolic handling of dietary protein. This improvement can be functionally interpreted through our rumen fermentation and microbial data. The numerically lower ruminal NH₃-N concentration in the BF group (Table 5), coupled with a maintained MCP yield, indicates a more efficient capture of degraded nitrogen into microbial biomass, reducing ammonia loss. This enhanced efficiency is likely supported by the synchronized availability of fermentable energy from pre-digested fiber and protein macromolecule in the fermented byproduct (Supplemental Table S1), promoting microbial growth. Furthermore, the increased relative abundance of Prevotella in the BF group (Figure 1) provides a microbial basis for this improvement. Prevotella species are pivotal in peptide and amino acid metabolism and are major contributors to microbial protein synthesis in the rumen [40,41]. Their enrichment suggests a microbial community better adapted to utilize the nitrogenous components of the fermented diet, leading to greater incorporation of nitrogen into microbial protein and subsequently improving whole-body nitrogen retention, as reflected in our balance data.

No significant differences in methane emissions were detected among treatment groups (p > 0.05), a result consistent with the limited sample size (n = 6) and the inherently high inter-animal variability characteristic of enteric methane production. Although statistical significance was not attained, the numerically lower daily methane emission (−6.87%) and emission intensity (−5.83%) observed in the BF group relative to RC align directionally with a significant reduction in Methanobrevibacter abundance (p < 0.05) and a significant increase in propionate-producing genera (Succinivibrio; p < 0.05). This concordance between microbial shifts and emission trends supports the biological plausibility of bacterial-enzymatic fermentation as a mitigating factor in methanogenesis; however, confirmatory studies with larger sample sizes are warranted. This reduction in methane intensity, combined with the most favorable emission intensity, suggests that the bacterial-enzymatic fermentation contributes to a more environmentally benign fermentation pattern in the rumen. This observation aligns with established mechanisms wherein enzyme supplementation modifies feed physicochemical properties, diminishes hydrogen production, and redirects metabolic hydrogen toward propionate synthesis [42,43,44]. The specific role of microbial-enzyme fermentation in promoting propionate-oriented rumen metabolism will be further elaborated in subsequent sections.

Therefore, the BF group reduced feed energy costs while enhancing nitrogen utilization efficiency in Hu sheep. This improvement, coupled with reduced dietary energy loss, led to better growth performance and thus greater economic returns.

4.3. Effects of Replacing Corn with Grain Byproducts on Rumen Fermentation of Fattening Hu Sheep

The ruminal fermentation parameters and microbial community structure work in concert to determine the efficiency of nutrient utilization and environmental impact in ruminant production [45,46]. In the present study, while no significant differences were observed in most ruminal fermentation parameters, the BF diet demonstrated a distinct pattern that aligns with improved nutrient utilization. The numerically highest total VFA concentration in the BF group (90.36 mM) suggests a trend toward enhanced fermentative activity, which may potentially contribute to the improved nitrogen utilization observed in this group [47]. Notably, the BF treatment significantly reduced the molar proportions of branched-chain VFAs (isobutyrate and isovalerate) relative to other dietary groups. This reduction may reflect a complex ecological shift within the rumen, given that these compounds function as essential growth factors for numerous fibrolytic bacteria, their decline could signal a restructuring of the fiber-degrading microbial consortium [48]. This effect may have been partially compensated by the enzymatic pre-treatment in the BF diet, which externalized a portion of the fiber degradation process [48,49]. Ruminal pH was not measured in this study, which represents a limitation in fully interpreting the fermentation dynamics. Future studies should include continuous or spot pH monitoring to better understand the relationships among diet composition, microbial structure, and acid-base balance in the rumen. It is important to clarify that, while DDG contributed substantially to the total dietary NDF content, they do not provide physically effective fiber (peNDF) capable of stimulating rumination or enhancing salivary buffer secretion, due to their fine particle size and high digestibility. In the present study, the absence of digestive disorders, stable feed intake, and normal rumen fermentation parameters suggest that the total fermentable fiber supply was adequate to support basal rumen function under the conditions of this large-scale commercial farm, where DDG is routinely used as a cost-effective fiber source.

Further analysis revealed that bacterial-enzyme fermentation of corn-replaced diets effectively modulated ruminal fermentation patterns by enhancing propionate synthesis. An increase in the molar proportion of propionate was observed in the BF group (18.68 mol/100 mol) compared to the RC group (16.85 mol/100 mol), restoring values to levels comparable with the CON group. This shift toward a propionate-type fermentation represents a critical metabolic adjustment with substantial implications for methane mitigation [50]. Consistent with this finding, supplementation with 0.5 g/kg DM of yeast and 4.44 g/kg DM of dried kratom leaves has been shown to reduce acetic acid concentration and CH₄ production, while increasing propionate [51]. Similarly, increasing dietary DL-malic acid resulted in a significant decrease in the molar proportions of acetic and butyric acids and a concomitant increase in propionate, ultimately reducing total daily CH₄ emissions [52]. The microbial basis for this shift is supported by the enrichment of Succinivibrio in the BF group, a genus known to contribute to succinate production, a direct precursor for propionate synthesis [53,54].

The CH₄ emissions and production performance in sheep are strongly influenced by the rumen microbiome [55,56]. In this study, compared with the control group, the RC group exhibited a significant increase in the relative abundance of Firmicutes and a significant decrease in Bacteroidetes. Members of the Firmicutes phylum are often associated with enhanced fibrolytic capacity and energy harvest from structural carbohydrates [57], which may partially explain the maintained dry matter intake in the RC group despite the inclusion of more fibrous byproducts. However, this compositional shift towards a Firmicutes-dominant community has also been frequently linked to metabolic pathways favoring hydrogen production, potentially creating a more favorable environment for hydrogenotrophic methanogens [58]. This provides a plausible ecological mechanism for the concurrent significant increase in Euryarchaeota (the phylum containing methanogens like Methanobrevibacter) and the numerically higher methane emissions observed in the RC group. In contrast, Bacteroidetes are key players in the degradation of soluble carbohydrates and proteins and are crucial for propionate production via the succinate pathway [59]. The significant reduction in Bacteroidetes in the RC group aligns functionally with the observed numerical decrease in ruminal propionate molar proportion, a key metabolic hydrogen sink alternative to methanogenesis.

4.4. Effects of Replacing Corn with Grain Byproducts on Rumen Microbial Community of Fattening Hu Sheep

At the genus level, the BF group exhibited a significant reduction in the abundance of Methanobrevibacter and an increase in Succinivibrio, along with a trend toward higher Prevotella abundance relative to the RC group. The marked decrease in Euryarchaeota abundance, particularly the suppression of Methanobrevibacter, provides a fundamental microbial explanation for the numerically lower methane emissions [60]. This microbial restructuring illustrates a key mechanism through which fermented diets can mitigate methanogenesis. The enrichment of Succinivibrio establishes a direct link between community-level shifts and altered fermentation patterns. As a recognized propionate-producing bacterium, Succinivibrio contributes to succinate accumulation, a key intermediate in propionate synthesis, demonstrating how targeted microbial modulation can redirect metabolic hydrogen toward propionogenesis rather than methanogenesis [53,54]. The partial restoration of Prevotella abundance in the BF group represents another ecologically relevant modification. As one of the most abundant and metabolically versatile genera in the rumen, Prevotella species contribute significantly to dietary protein degradation and microbial protein synthesis [40,41]. This may account for the improved nitrogen utilization efficiency observed in the BF group, particularly the higher digestible nitrogen values. The recovery of Prevotella abundance suggests that bacterial-enzyme fermentation creates a favorable environment for this beneficial genus, potentially through enhanced substrate availability or modified fermentation kinetics.

Critically, the restoration of a microbial community structure at the phylum level that was statistically indistinguishable from the corn-based control diet is a key finding in the BF group. This stability is reflected in the improved functional outcomes: the BF group supported a genus-level enrichment of beneficial taxa like Succinivibrio (a propionate producer within the Bacteroidetes phylum) and showed a trend toward higher Prevotella abundance, while simultaneously suppressing Methanobrevibacter. Therefore, the lack of significant phylum-level disturbance in the BF group, compared to the pronounced dysbiosis in the RC group, underlies the moderated rumen fermentation (higher TVFA and propionate), improved nitrogen metabolism, and reduced methanogenesis. It demonstrates that the beneficial effects of bacterial-enzymatic fermentation are mediated not through a simple reversion to the control state, but through the promotion of a new, functionally superior microbial equilibrium that enhances metabolic efficiency while minimizing environmental trade-offs.

The absence of statistically significant differences in total volatile fatty acids, acetate-to-propionate ratio, or microbial crude protein concentration, despite marked compositional shifts in the ruminal microbiota, notably reduced Methanobrevibacter and enriched Succinivibrio, warrants careful interpretation. This pattern aligns with the principle of functional redundancy, wherein phylogenetically distinct microbial assemblies can yield comparable aggregate metabolic outputs under stable dietary regimens [61,62]. While 16S rRNA amplicon sequencing is inherently sensitive to -taxonomic variation [63], ruminal VFA and ammonia concentrations reflect steady-state equilibria shaped by the integrated processes of microbial production, epithelial absorption, and ruminal outflow [64]. Consequently, moderate but metabolically relevant shifts in hydrogen-partitioning taxa may not suffice to perturb bulk fermentation parameters in spot samples, particularly under ad libitum feeding conditions. Notably, significant treatment effects were observed for branched-chain VFAs, sensitive indicators of protein catabolism [65], which corroborate the enhanced nitrogen utilization efficiency documented in the BF group. Thus, the observed microbial restructuring was not functionally inert; rather, its metabolic consequences were selectively expressed in discrete pathways (i.e., nitrogen metabolism and hydrogen partitioning) rather than in global fermentation indices.

The concurrent elevation in relative abundances of Ruminococcus and Methanobrevibacter observed in the RC group exemplifies a well-characterized syntrophic association within the ruminal microbial ecosystem. Ruminococcus spp. function as primary cellulolytic and hemicellulolytic agents, depolymerizing plant structural carbohydrates into fermentable monosaccharides [66], which are subsequently metabolized to hydrogen (H₂), carbon dioxide (CO₂), and short-chain fatty acids, predominantly acetate [67]. The generation of H₂ during fibrolysis establishes a reduced microenvironment conducive to the proliferation of hydrogenotrophic methanogens, particularly Methanobrevibacter, which reduce CO₂ with H₂ to form methane [68]. Accordingly, the enrichment of Ruminococcus provides a sustained substrate flux that supports methanogen growth and elevates methanogenic capacity, thereby accounting for the parallel increase in both taxa and the numerically higher methane emissions in the RC group. In contrast, the BF group exhibited suppression of Methanobrevibacter without a concomitant reduction in Ruminococcus abundance, suggesting that bacterial-enzymatic fermentation pretreatment may disrupt this syntrophic coupling, likely through metabolic hydrogen redirection toward alternative sinks such as propionate synthesis, consistent with the concurrent enrichment of Succinivibrio and the numerical increase in propionate molar proportion.

The partial restoration of Prevotella abundance in the BF group constitutes a functionally consequential shift with direct relevance to methane mitigation. As one of the most prevalent and metabolically versatile bacterial genera within the ruminal ecosystem, Prevotella plays a pivotal role in the fermentation of non-fibrous carbohydrates (e.g., starch, pectin, and soluble sugars) and the proteolysis of dietary protein [69]. Notably, Prevotella species predominantly generate succinate as a major fermentative end product, which serves as a critical intermediate in both the acrylate and succinate pathways of propionate synthesis [70,71]. The subsequent decarboxylation of succinate to propionate [72], mediated principally by succinate-utilizing taxa such as Succinivibrio and Selenomonas, competes directly with methanogenesis for metabolic hydrogen (H₂) [73]. By diverting H₂ from carbon dioxide reduction toward propionate formation, this alternative hydrogen sink effectively reduces substrate availability for hydrogenotrophic methanogens, including Methanobrevibacter.

Furthermore, elevated Prevotella abundance has been consistently associated with reduced ruminal ammonia concentrations and enhanced nitrogen utilization efficiency [74], both of which were recapitulated in the BF group. This concordance aligns with the genus’s well-documented proteolytic capacity and its role in integrating carbohydrate and protein fermentation, thereby potentiating microbial protein synthesis and curtailing nitrogen excretion. Collectively, the synergistic effects of enhanced propionate production, diminished methanogenic substrate supply, and improved nitrogen economy underpin the numerical reductions in methane emissions and emission intensity observed in the BF group. Accordingly, the concurrent enrichment of Prevotella and Succinivibrio denotes a community-wide restructuring toward a hydrogen-efficient fermentative phenotype that is both energetically advantageous and environmentally sustainable.

4.5. Study Limitations

Several limitations warrant acknowledgment but, rather than diminishing contributions, delineate clear priorities for future work. The six-replicate design provided adequate power for key nitrogen and microbial parameters, though it may have underpowered more variable traits (e.g., ADG, methane), yielding non-significant yet directionally consistent trends. Similarly, the RC group exhibited numerically lower concentrations of TVFA and propionate, accompanied by a higher acetate-to-propionate ratio relative to the CON group, although these differences did not reach statistical significance. These non-significant trends are directionally consistent with a fibrolytic, methanogenic shift. Further research can be conducted through metagenomic sequencing to elucidate the functional roles of microorganisms. The economic assessment, though deterministic, provides empirically grounded profit estimates and the first integrated evaluation of fermented byproduct-based diets in ovine systems. Likewise, while a single formulation, its capacity to enhance nitrogen utilization and remodel the rumen microbiome under commercial conditions substantiates translational relevance, with principles transferable to other ruminants. Collectively, this study provides the evidence that bacterial-enzymatic fermentation enhances nitrogen economy, remodels microbiota, and improves economic returns without compromising growth; the limitations thus offer an evidence-anchored roadmap for confirmatory trials and cross-system validation, accelerating translation from proof-of-concept to deployment.

5. Conclusions

This study suggests that directly replacing 42% of dietary corn with grain byproducts reduced feed cost but numerically impaired nitrogen utilization and increased methane emissions in fattening Hu sheep. Although growth performance was not significantly affected, this substitution ultimately failed to improve economic returns. In contrast, bacterial-enzymatic fermentation of these grain byproducts numerically mitigated these adverse effects by improving ruminal propionate production, enhancing nitrogen retention, and suppressing Methanobrevibacter relative abundance, which collectively led to numerically improved growth performance and a 27.93% increase in net profit. Thus, fermented grain byproducts represent a promising corn-substitution strategy for cost-effective and environmentally sustainable sheep production. Nonetheless, given the statistical limitations associated with sample size, further validation in larger-scale studies is recommended to confirm these findings and support practical adoption.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani16050786/s1, Table S1: Chemical composition of the grain byproducts. Figure S1: The diagram of the automated head-chamber (AHC) system used for measuring the greenhouse gas emissions of Hu sheep. Figure S2: Schematic diagram of the automated head-chamber (AHC) system for measuring CH4 production in dairy cows.

Author Contributions

Conceptualization, L.H.; methodology, Z.S., C.D., X.D., J.Q., W.Z. and L.H.; software, L.H.; validation, Z.S., C.D., X.D., J.Q. and W.Z.; formal analysis, X.W., N.R. and Z.Z.; investigation, L.H.; resources, Z.S., C.D., X.D., J.Q. and W.Z.; data curation, X.W., N.R. and Z.Z.; writing—original draft preparation, X.W., N.R. and Z.Z.; writing—review and editing, X.W., N.R., Z.Z. and L.H.; visualization, Z.S., C.D., X.D., J.Q. and W.Z.; supervision, Z.S., C.D., X.D., J.Q. and W.Z.; project administration, Z.S., C.D., X.D., J.Q., W.Z. and L.H.; funding acquisition, L.H. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

All experimental procedures involving animals were conducted in accordance with institutional guidelines and approved by the Animal Welfare Committee of the Agricultural Research Organization, China Agricultural University (Approval No. Aw31015202-1-5).

Informed Consent Statement

Not applicable.

Data Availability Statement

All raw data and sequencing information can be requested by con-tacting the corresponding author Liwen He (helw@cau.edu.cn).

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research was funded by National Key R&D Program of China (2023YFD1300905-1) and the China Agriculture Research System-39 (CARS-39).