Dietary protein sources in concentrate supplementation influence growth performance by manipulating gut microbiota and serum metabolites in suckling Donkey foals

Abstract

Background

Protein is a primary nutrient in concentrate supplementation for donkey foals, and the source of this protein significantly influences their growth and development. Milk-derived protein sources, such as milk powder, casein, and whey protein, are widely used in milk replacers for donkey foals due to their balanced nutritional profiles, high digestibility, and high bioavailability. However, the increasing costs of milk powder and whey protein have prompted researchers to explore alternative protein sources, with soy protein being a particularly promising option. This study compared the effects of soybean meal and milk-derived ingredients as protein sources in concentrate supplementation on the growth performance, rectal microbiota, and serum metabolites of suckling donkey foals. A total of 42 Dezhou donkey foals, aged 10 days, were randomly assigned to three groups: SP (soybean meal as the main protein source in the diet), MP (milk-derived ingredients as the main protein source in the diet), and SMP (a combination of the SP and MP diets at a ratio of 6:4 used as the dietary component). Each group consisted of 14 replicates, with one donkey in each replicate. The foals were raised from 10 days of age to 130 days of age, and the entire experimental period lasted 120 days.

Results

The final body weight (at 130 days of age) and average daily gain (ADG) were significantly higher (P < 0.001) in the SP group compared to the MP and SMP groups. The feed conversion ratio (FCR) in the SP and SMP groups was significantly lower than that in the MP group (P < 0.05). Among the three groups, the serum levels of thyroxine, growth hormone (GH), insulin-like growth factor-І (IGF-І), and vitamin B6 were significantly enhanced (P < 0.05) in the SP group, whereas the cortisol levels were significantly decreased (P < 0.05). Rectal microbiota analysis further demonstrated that the SP intervention reshaped the gut microbial composition and enriched several genera, including Oscillospiraceae_UCG-005, Oscillospiraceae_NK4A214_group, Akkermansia, Porphyromonas, Streptococcus, Bacteroides, Fusobacterium, and Christensenellaceae_R-7_group. Metabolomic profiling identified 15 differential metabolites, which were considered the key differential metabolites in this study and were related to phenylalanine, tyrosine and tryptophan biosynthesis, phenylalanine biosynthesis, vitamin B6 metabolism, biotin metabolism, tryptophan metabolism, and some amino acid metabolic processes. Notably, the rectal microbial genera Akkermansia, Porphyromonas, Oscillospiraceae_NK4A214_group, Oscillospiraceae_UCG-005, and Streptococcus, which were most abundant in the SP group, showed significant positive correlations with ADG, serum concentrations of thyroxine, IGF-I, and vitamin B6, as well as with the levels of serum metabolites serotonin and pyridoxine.

Conclusions

Compared to the milk-derived protein in concentrate supplementation for donkey foals, soybean meal protein improved the abundance of beneficial gut microbiota, further affected serum hormones and metabolites, and probably resulted in improved body weight and feed efficiency. This study provides a new approach to modulating gut bacteria to enhance growth performance in donkey foals.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s42523-025-00457-1.

Introduction

In China, rising living standards have fueled increased demand for donkey meat and skin. However, the donkey population has been declining annually due to high feeding costs, which are primarily associated with low fertility rates, low average daily gain (ADG), and an elevated risk of disease outbreaks [1]. Notably, the growth and health of suckling herbivorous animals are critically affected by their gut microbiome, potentially leading to lifelong consequences [2, 3]. Consequently, effective management, particularly nutritional strategies during the suckling period, is essential for the profitability and long-term sustainability of the herbivorous animal farming industry. Early supplementary feeding, particularly with protein-rich feed, is recognized as an effective strategy to meet the increasing nutritional demands of growing donkey foals, compensating for declining milk intake and supporting their transition to solid feed after weaning [4, 5].

Protein is a key nutrient in concentrated feed supplements for donkey foals, and proteins from animal and plant sources result in different physiological effects and metabolic responses due to their unique amino acid profiles [6]. Milk-derived proteins, such as milk powder and whey protein, are widely used as the main protein source in suckling animal feed due to their balanced nutrition, high digestibility, and efficient utilization [7]. Studies indicate that incorporating milk-derived animal protein into the diet more closely approximates the amino acid composition of natural mammalian milk, thereby reducing the digestive burden for young animals [5, 8]. However, due to the high cost of milk powder and animal protein feeds, plant proteins, notably soy protein, are increasingly being investigated for use in milk replacers [9]. Soybean meal serves as a cost-effective, plant-based protein source in animal diets and is abundant in essential amino acids, including lysine and methionine, which play critical roles in supporting muscle development in foals [10]. Additionally, the isoflavones and saponins present in soybean meal exhibit antioxidant and anti-inflammatory properties, thereby enhancing immune function in animals [11]. Although soybean meal may contain anti-nutritional factors and heat-sensitive components, these can be effectively inactivated during the pelleting process when mixed with other feed ingredients [12, 13]. Moreover, the small peptide content increases during pelleting, which enhances protein digestibility and absorption [14].

The gut microbiota plays a central role in the health and development of equines [15, 16]. As equine animals, donkeys possess an efficient microbial fermentation system located in the hindgut (including the cecum and colon). Specifically, microbial fermentation in the cecum plays a vital role by providing an optimal environment for the digestion of fibrous materials. The gut microbiota is particularly susceptible to environmental and dietary factors in donkey foals and may influence future health [17]. Dietary protein sources probably influence both the composition and functional activity of the gut microbiota. Previous studies have demonstrated that plant-based protein sources tend to enhance gut microbial diversity and stimulate the production of short-chain fatty acids, whereas animal-derived protein sources exhibit the opposite effect [18]. Meanwhile, specific protein sources can markedly influence the composition of fecal microbiota and systemic metabolic profiles [19]. The difference in protein sources directly affects metabolic regulation [20]. Furthermore, alterations in the gut microbiota and its associated nutritional functions could potentially influence host metabolism, which can be reflected by serum metabolites [21]. These studies provide direct scientific evidence that different protein sources can influence host phenotypes through the regulation of intestinal microbiota.

Although the effects of concentrate supplementation on growth performance in suckling donkeys have been recognized, the specific impacts of protein sources from milk-derived protein and soybean meal remain unclear. This study investigated how these two protein sources affect growth performance, rectal microbiota, serum biochemical indicators, growth-related hormones, and metabolites in donkey foals. We hypothesized that different protein sources (milk-derived protein and soybean meal) influence donkey foal growth by altering gut microbiota and serum metabolites. The study aimed to: (1) analyze changes in the composition and function of rectal microbes and serum metabolites among groups fed SP, MP, and SMP groups; and (2) integrate metagenomic and metabolomic data to elucidate how microbial and metabolite changes affect growth. These findings elucidated how different protein sources in concentrates influence foal growth through alterations in gut microbiota and serum metabolites.

Methods

Animal trials, design, and diet

The research protocols were rigorously reviewed and approved by the Animal Nutrition Department of Shandong Agricultural University, strictly adhering to the National Guidelines for Animal Research (Protocol No. SDAUA-2024-224).

The study was conducted from March to June at the donkey breeding base of Dong-e-e-jiao Co., Ltd., located in Liaocheng, Shandong Province, China. A total of 42 Dezhou donkey foals, aged 10 days and with similar body weights (38.30 ± 0.16 kg), were randomly allocated into three treatment groups. Each group consisted of 14 replicates, with one donkey foal in each replicate.

Based on the varying protein sources in the concentrate supplementation for donkey foals, three dietary treatments were designed: soybean meal as the main protein source (SP group), milk-derived ingredients including milk powder, whey powder, and whey protein concentrate as the main protein source (MP group), and a combination of soybean meal and milk-derived ingredients as the protein sources (SMP group).

The experiment lasted 120 days. Each donkey foal was housed with its dam in an individual stall to facilitate natural suckling. The animals were fed at 08:00 and 18:00, with ad libitum access to both feed and water. The concentrate supplementation was provided exclusively to the donkey foals through a designated feed tank. On the way to the feed trough, there was a 35-cm-wide restriction pen that allowed only donkey foals to pass.

Additionally, the donkey foals had free access to the roughage provided for the dam, whereas the dam was unable to consume the concentrate-supplemented diet intended for the foals. The concentrate supplementation (Table 1) was specifically formulated to meet the nutritional requirements as recommended by the NRC for horses (2007) and was prepared in pelleted form as total mixed rations (TMR). The milk compositions (Table S1) of the dam donkeys in the study were analyzed by multi-functional dairy analyzer (FOSS, Hillerød, Denmark).

Table 1.

Compositions and nutrient levels of the diets (DM basis)

Items (%)		Treatment
	SP1	MP	SMP
Basic diet ingredients, %
Corn	53.50	33.00	45.30
Wheat middling	4.00	12.00	7.20
Soybean meal	15.00	0.00	9.00
Fermented soybean meal	8.00	8.00	8.00
Expanded soybean	5.00	5.00	5.00
Milk powder	0.00	12.00	4.80
Whey powder	0.00	10.00	4.00
Whey protein concentrate	0.00	4.50	1.80
Glucose	7.50	8.00	7.70
Soybean oil	2.00	2.50	2.20
CaHPO4·2 H2O	0.80	0.80	0.80
Calcium formate	0.60	0.60	0.60
L-Lysine	0.60	0.60	0.60
DL-Methionine	0.20	0.20	0.20
Trace mineral permix2	1.30	1.30	1.30
Vitamin premix3	1.50	1.50	1.50
Chemical composition, % of DM4
Metabolizable energy (MJ/kg)	16.09	16.17	16.12
Dry matter	90.98	91.98	91.49
Crude protein	18.34	18.18	18.28
Ether extract	5.31	5.31	5.31
Calcium	0.72	0.79	0.75
Total phosphorus	0.56	0.54	0.55

¹SP: the soybean meal as the primary protein source of the diet; MP: the milk-derived ingredients as the primary protein source of the diet; SMP: The feed formula of SP (including each component) and MP groups (including each component) in a ratio of 6:4 was used as the dietary component

²Mn (as manganese sulfate) 40 mg, Fe (as ferrous sulfate) 70 mg, Zn (as zinc sulfate) 80 mg, Cu (as copper sulfate) 20 mg, I (as potassium iodide) 0.25 mg, Se (as sodium selenite) 0.25 mg

³The vitamin premix provided: Vitamin A 8,000 IU; Vitamin D3 3500 IU; Vitamin E 30 IU; Vitamin K₃ 1.0 mg; Vitamin B₁ 1.5 mg; Vitamin B₂ 5.5 mg; Vitamin B₆ 2.0 mg; Vitamin B₁₂ 0.02 mg; Pantothenic acid 7.50 mg; Niacin 26.5 mg; Biotin 0.07 mg; Folic acid 0.6 mg

⁴Metabolizable energy was calculated according to NRC Nutrient Requirements of Horses [22], and other values were measured

Growth performance analysis

The donkey foals were weighed at 10 d (initial body weight), 70 d, and 130 d (final body weight) of age. The initial and final body weights were used to calculate the average daily gain (ADG). The amounts of the concentrate offered and refused for each foal were recorded daily to calculate concentration average daily feed intake (ADFI). Although the foal had access to the roughage (a 2:1 mix of peanut straw and soybean stalk) provided for the female donkeys, intake was minimal and difficult to measure. Additionally, since the foal suckled freely while housed with the mare, milk intake could not be accurately quantified. Therefore, this study focused only on the concentrate supplementation given directly to the foals. The feed conversion ratio (FCR), calculated as F/G, was determined by dividing ADFI by ADG.

Sample collection and determination

Over the six-week period prior to the conclusion of the experiment, milk production was measured at two-week intervals. During each measurement, female donkeys and their foals were separated from 7:00–10:00 a.m. and from 2:00–5:00 p.m., respectively. Throughout the isolation period, visual, olfactory, and vocal contact were maintained between the dam donkeys and their foals. The milk from the female donkey was collected by hand-milking at 10:00 a.m. and 5:00 p.m., respectively. Then, daily milk yield was calculated using the following formula: estimated milk yield (EMY, kg/d) = milking yield (the sum of the milking yield twice a day) (kg/d) × 4 [23]. The milk yield for each group of female donkeys was calculated as the average of three measurements.

During each measurement, the morning and afternoon milk samples were mixed in a 1:1 ratio for subsequent milk composition analysis. The contents of milk protein, fat, solids-not-fat, total solids, lactose, relative density, and titratable acidities were determined using infrared spectrophotometry (FOSS, Hillerød, Denmark).

At the end of the feeding trial, when the donkey foals were 130 days old, seven donkey foals from each group were randomly selected for sample collection (n = 7). Blood for each donkey foal was collected before the morning feeding (fasted for 12 h) via the jugular vein using evacuated tubes without anticoagulant, then centrifuged at 3,500 × g for 10 min at 4 °C to separate serum. The serum was subsequently aliquoted into 1.5 mL microcentrifuge tubes and stored at -80 °C for further analysis. Fresh fecal samples were then collected from the rectum and immediately transferred into 5 mL sterile centrifuge tubes. The tubes were subsequently placed into liquid nitrogen for rapid freezing. Upon arrival at the laboratory, the samples were stored at -80 °C for further DNA extraction and microbial community analysis.

Serum index detection

The serum levels of thyroxine, cortisol, growth hormone (GH), insulin-like growth factor-І (IGF-І), and vitamin B6 were measured using ELISA kits provided by Jingmei Biotechnology Co., Ltd. All procedures were conducted according to the manufacturer’s instructions. Additionally, the serum concentrations of triglycerides (TG), total cholesterol (TCHO), low-density lipoprotein cholesterol (LDL-Cho), and high-density lipoprotein cholesterol (HDL-Cho) were measured by an automated biochemistry analyzer (Hitachi 7020, Tokyo, Japan).

16 S rRNA gene sequencing and data analysis

The genomic DNA of the microbes in fecal samples was extracted using an extraction kit (Omega, Georgia, USA). The quality and concentration of the DNA were then assessed by 1.0% agarose gel electrophoresis and a NanoDrop2000 spectrophotometer (Thermo Scientific, Wilmington, USA), respectively. The hypervariable region V3-V4 of the bacterial 16 S rRNA gene was amplified with primer pairs 338 F (5’-ACTCCTACGGGAGGCAGCAG-3’) and 806R (5’-GGACTACHVGGGTWTCTAAT-3’) by T100 Thermal Cycler PCR thermocycler (BIO-RAD, California, USA) [24].

The PCR product was excised from a 2% agarose gel and purified using the PCR Clean-Up Kit (YuHua, Hangzhou, China). The concentration of the purified amplicon was quantified using the Qubit 4.0 Fluorometer (Thermo Fisher Scientific, Waltham, USA). Purified amplicons were pooled in equimolar amounts and subjected to paired-end sequencing on the Illumina NextSeq2000 platform (Illumina, San Diego, USA) following the standard protocols provided by Majorbio Bio-Pharm Technology Co., Ltd. The raw sequencing reads were deposited into the NCBI Sequence Read Archive (SRA) database. Following demultiplexing, the sequences underwent quality filtering using fastp and were subsequently merged using FLASH [25, 26]. Then, high-quality sequences were de-noised using the DADA2 plugin in Qiime2 [27]. DADA2-denoised sequences are commonly referred to as Amplicon Sequence Variants (ASVs). Subsequently, Qiime2 was employed to calculate alpha diversity indices, including Shannon, Sobs, and ACE [28]. The similarity between microbial communities across different samples was assessed using principal coordinate analysis (PCoA) based on Bray-Curtis dissimilarity. The Heatmap Illustrator program (version 1.0.3.7) was utilized to generate bar charts and heat maps. Significant differences in microbiota composition among treatment groups at each taxonomic level were identified using one-way ANOVA followed by linear discriminant analysis (LDA) effect size (LEfSe) [29].

Metabolomic profiling analysis

The serum metabolome was analyzed by Novogene Corporation Ltd., Beijing, China, using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The serum samples (100 µL) were placed in EP tubes and resuspended with prechilled 80% methanol using a vortex mixer. Then, the samples were incubated on ice for 5 min and centrifuged at 15,000 × g at 4 °C for 20 min. A portion of the supernatant was diluted to a final concentration of 53% methanol using LC grade water. The samples were subsequently transferred to fresh Eppendorf tubes and centrifuged again at 15,000 × g at 4 °C for 20 min. Finally, the supernatant was injected into the LC-MS/MS system for analysis [30, 31]. LC-MS/MS analyses were performed using a Vanquish UHPLC system (Thermo Fisher, Bremen, Germany). Samples were injected onto a Hypersil Gold column (100 × 2.1 mm, 1.9 μm) and separated using a 12-minute linear gradient at a flow rate of 0.2 mL/min. The mobile phases consisted of eluent A (0.1% formic acid in water) and eluent B (methanol) for both positive and negative ionization modes. The solvent gradient program was as follows: 2% B for 1.5 min; 2–85% B over 3 min; 85–100% B over 10 min; 100–2% B over 10.1 min; and 2% B at 12 min. The Q Exactive HF Hybrid Quadrupole-Orbitrap mass spectrometer was operated in alternating positive/negative polarity mode with the following parameters: spray voltage, 3.5 kV; capillary temperature, 320 °C; sheath gas flow rate, 35 psi; auxiliary gas flow rate, 10 L/min; S-lens RF level, 60; and auxiliary gas heater temperature, 350 °C. The data files generated by UHPLC-MS/MS were processed using XCMS for peak alignment, peak picking, and quantitation of each metabolite. The identified metabolites were annotated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The orthogonal partial least squares discriminant analysis (OPLS-DA) was utilized to investigate the overall distribution patterns of metabolites among the three treatment groups. The enrichment and pathway analyses of the metabolites in this study were conducted using the online MetaboAnalyst platform, version 6.0 (https://www.metaboanalyst.ca/MetaboAnalyst/ModuleView.xhtml).

Statistical analysis

The data were analyzed using SPSS 27.0 (SPSS Inc., USA) via one-way ANOVA and Tukey’s HSD test to identify significant differences. Differences with P < 0.05 were considered statistically significant. The results are expressed as mean and the standard errors of the mean (SEM). Spearman’s rank correlation analysis was employed to evaluate the correlations between specific growth performance, growth-related hormones, key differential microbes, and serum metabolites.

Results

Growth performance

As shown in Table 2, at the start of the study, the initial body weight (BW) of the donkey foals in each group showed no significant difference. However, compared to the MP and SMP groups, the SP group exhibited significantly higher final BW (P < 0.001), average daily gain (ADG) (P < 0.001), and average daily feed intake (ADFI) (P < 0.001). The feed conversion ratio (FCR) in the SP and SMP groups significantly decreased compared to the MP group (P < 0.05).

Table 2.

Effects of different protein sources in concentrate supplementation on growth performance of Donkey foals (n = 14)

Items2	Treatments1			SEM3	P-value
	SP	MP	SMP
Initial BW, kg	38.41	38.24	38.25	0.160	0.891
Final BW, kg	125.11a	116.11b	115.32b	0.843	< 0.001
ADG, g	722.46a	665.54b	658.72b	6.123	< 0.001
ADFI, g	661.42a	617.84b	588.37c	5.020	< 0.001
FCR	0.88b	0.93a	0.89b	0.006	< 0.05

¹SP, soybean meal as the primary protein source in the diet; MP, milk-derived ingredients as the primary protein source in the diet; SMP, the feed formula of SP and MP groups at a ratio of 6:4 was used as the dietary component

² BW, body weight; ADG, average daily gain; ADFI, average daily feed intake of the concentrate supplementation; FCR, feed conversion ratio

³ SEM, standard error of the mean

^{a, b, c} Mean values within a row that are labeled with different superscripts (a, b, c) are significantly different (P < 0.05). The absence of a shoulder label or the presence of identical letters on the shoulder labels indicates no significant difference (P > 0.05)

Serum growth-related hormones and biochemical indicators

The serum levels of growth-related hormones are presented in Fig. 1A and B. The concentrations of thyroxine (P < 0.05) and growth hormone (GH) (P < 0.05) in the SP and SMP groups were significantly higher than those in the MP group. Additionally, compared to the MP group, the SP group exhibited an increased concentration of serum IGF-І (P < 0.01) and a decreased concentration of cortisol (P < 0.05).

Fig. 1.

Effects of different protein sources in concentrate supplementation on the concentration of serum growth-related hormones and biochemical indices. (A) Serum thyroxine concentration in the three groups. (B) Serum cortisol, GH, and IGF-I concentrations in the three groups. (C) Serum vitamin B6 concentration in the three groups. (D) Serum TG, TCHO, HDL-Cho, and LDL-Cho concentrations in the three groups. Differences in serum growth-related hormones and biochemical indices were considered statistically significant at the levels of *P < 0.05, **P < 0.01, and ***P < 0.001

As shown in Fig. 1C, the serum vitamin B6 concentration (P < 0.001) in the SP group was significantly higher than that in both the MP and SMP groups. As depicted in Fig. 1D, there were no significant differences in serum triglyceride (TG) concentrations (P > 0.05) among the three groups. However, compared to the SMP group, the serum total cholesterol (TCHO) and high-density lipoprotein cholesterol (HDL-Cho) concentrations (P < 0.05) in the SP and MP groups significantly increased. The serum low-density lipoprotein cholesterol (LDL-Cho) concentration (P < 0.05) in the SP and SMP groups was significantly lower than that in the MP group.

Diversity of the rectum microbiota

Figure 2 presents the diversity of the rectum microbiota for the SP, MP, and SMP groups. The alpha diversity indices (P < 0.05), including Abundance-based coverage estimator (ACE), observed species richness (Sobs), and Shannon index were significantly higher in both the SP and SMP groups compared to the MP group (Fig. 2A and B, and 2C). Notably, the Venn diagram in Fig. 2D indicates that 1,408 amplicon sequence variants (ASVs) were shared across all groups. Furthermore, the SP, MP, and SMP groups contained 2,507, 2,052, and 2,774 unique ASVs, respectively. Figure 2E shows the ASVs distribution at the genus level. A total of 294 ASVs were shared among all groups. The SP, MP, and SMP groups exhibited 70, 100, and 42 unique ASVs, respectively. As the beta diversity index, principal coordinate analysis (PCoA) (P < 0.05) based on Bray-Curtis distances revealed significant differences in rectum microbiota composition among the three groups (Fig. 2F).

Fig. 2.

Effects of different protein sources in concentrate supplementation on rectum microbiota diversity in donkey foals. (A) to (C) Abundance-based coverage estimator (ACE) index, observed species richness (Sobs) index, and Shannon index in SP, MP, and SMP groups. (D) Venn diagram of total amplified sequence variants (ASVs). (E) The Venn diagram of ASVs at the genus level. (F) Principal coordinate analysis (PCoA) of the SP, MP, and SMP groups. Differences in rectum microbiota were considered statistically significant at the levels of *P < 0.05, **P < 0.01, and ***P < 0.001

Compositions and structures of the rectum microbiota

The relative abundances of the top 10 phyla and 15 genera of the three groups were analyzed (Fig. 3 and Table S2). At the phylum level (Fig. 3A), Firmicutes and Bacteroidota were the dominant phyla, accounting for approximately 80% of the total bacterial community across all groups. At the genus level (Fig. 3B), Bacteroides, Christensenellaceae_R-7_group, Oscillospiraceae_UCG-005, Oscillospiraceae_UCG-002, Rikenellaceae_RC9_gut_group, Fusobacterium, Oscillospiraceae_NK4A214_group, and Akkermansia were the predominant microbial populations in the three groups.

Fig. 3.

Effects of different protein sources in concentrate supplementation on rectal differential microbes in donkey foals. (A) The relative abundances of rectal microbes for the top 10 phyla. (B) Relative abundances of rectal microbes for the top 15 genera. (C) and (D) Random forests analysis of rectal microbiota (the relative abundances of the top 10 phyla and 15 genera), red indicates Mean Decrease in Accuracy (MDA) > 2

To further analyze the structure and composition of rectum microbiota, random forest analysis was also employed to predict whether these microorganisms exhibited significant variation across different groups, based on the importance of features for model accuracy. A threshold of mean decrease in accuracy (MDA) > 2 was applied. At the phylum level, the Verrucomicrobiota, Actinobacteriota, Fusobacteriota, Campylobacterota, Proteobacteria, Firmicutes, Bacteroidota, and Desulfobacterota were identified (Fig. 3C). At the genus level, Oscillospiraceae_UCG-005, Oscillospiraceae_NK4A214_group, Bacteroides, Christensenellaceae_R-7_group, Akkermansia, Porphyromonas, Streptococcus, Fusobacterium, and Phascolarctobacterium were detected (Fig. 3D).

Differential microbes in the rectum

The relative abundance analysis for the top 10 phyla and 15 genera was performed by one-way ANOVA to identify the differential microbes in the rectum (Fig. 4). At the phylum level, compared to the MP group, the relative abundance of Firmicutes (P < 0.05) in the SP and SMP groups significantly increased, whereas the relative abundance of Fusobacteriota, Campylobacterota, Proteobacteria, and Actinobacteriota (P < 0.05) decreased. Additionally, Verrucomicrobiota was significantly higher in the SP group than in the MP and SMP groups (P < 0.001) (Fig. 4A). Notably, the ratio of Firmicutes to Bacteroidota (F/B) (P < 0.05) was significantly lower in the MP group compared to the SP and SMP groups (Fig. 4B). At the genus level, SP exhibited significant increases in the relative abundance of Oscillospiraceae_UCG-005, Oscillospiraceae_NK4A214_group, Akkermansia, Porphyromonas, and Streptococcus compared to MP and SMP (P < 0.05). Meanwhile, MP showed significant increases in the relative abundance of Bacteroides and Fusobacterium compared to both SP and SMP (P < 0.05), as well as a notable rise in the relative abundance of the Christensenellaceae_R-7_group compared to SP (P < 0.05) (Fig. 4C).

Fig. 4.

Differential microbes in the rectum. (A) Differential microbes among the three groups at the phylum level. (B) The Firmicutes to Bacteroidota (F/B) ratio. (C) Differential microbes among the three groups at the genus level. (D) The histogram represents the linear discriminant analysis (LDA) values (LDA score > 3). Differences in rectum microbes were considered statistically significant at the levels of *P < 0.05, **P < 0.01, and ***P < 0.001

Additionally, linear discriminant analysis (LDA) effect size (LEfSe) with an LDA score greater than 3 was utilized to identify the differential microbes in the rectum across the three groups (Fig. 4D). Specifically, the SP group was characterized by higher relative abundances of Akkermansia, Porphyromonas, Oscillospiraceae_NK4A214_group, Oscillospiraceae_UCG-005, and Streptococcus. The MP group showed increased levels of Bacteroides, Fusobacterium, Christensenellaceae_R_7_group, Phascolarctobacterium, and Actinomyces. Meanwhile, the SMP group demonstrated a significantly higher relative abundance of norank_f_Bacteroidales_UCG_001.

Therefore, based on the results of random forests analysis, ANOVA analysis, and LEfSe analysis, Oscillospiraceae_UCG-005, Oscillospiraceae_NK4A214_group, Akkermansia, Porphyromonas, Streptococcus, Bacteroides, Fusobacterium, and Christensenellaceae_R-7_group were considered as characteristic differential rectal microorganisms.

Profile of serum metabolites and enrichment of metabolic pathways

A total of 1669 serum metabolites (including both positive and negative ions) were identified across the three groups. The metabolite profiles were distinctly separated using Orthogonal Projections to Latent Structure Discrimination Analysis (OPLS-DA) (Fig. 5A.1 to 5A.3). Subsequently, differential serum metabolites were identified by comparing SP versus MP (SP vs. MP), SP versus SMP (SP vs. SMP), and MP versus SMP (MP vs. SMP), based on fold change thresholds (FC > 1.2 or FC < 0.833), variable importance in projection (VIP > 1), and P-value < 0.05. A total of 149 differential metabolites were detected between SP and MP, comprising 39 upregulated and 110 downregulated metabolites (Fig. 5B.1, Table S3). Between SP and SMP, a total of 250 differential metabolites were identified, including 94 upregulated and 156 downregulated metabolites (Fig. 5B.2, Table S4). Similarly, 226 differential metabolites were found between MP and SMP, with 114 upregulated and 112 downregulated metabolites (Fig. 5B.3, Table S5).

Fig. 5.

Comprehensive analysis of untargeted serum metabolites of the three groups. (A.1) to (A.3) Scores scatter plot of the Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) models comparing SP versus MP, SP versus SMP, and MP versus SMP. (B.1) to (B.3) Volcanic plots illustrating differential metabolites from the pairwise comparisons among the three groups. (C.1) to (C.3) The metabolic pathways of differential metabolites from the pairwise comparisons were analyzed using the KEGG pathway database for Equus asinus (donkey). (D) Venn diagram of the key differential metabolites identified in KEGG metabolic pathway

The differential metabolites were analyzed using the KEGG database for Equus asinus (donkey). The P values and importance values of the corresponding metabolic pathways were obtained, and the metabolic pathways with an impact value exceeding 0.1 were selected. In the SP versus MP comparison, the identified metabolic pathways included phenylalanine, tyrosine, and tryptophan biosynthesis; phenylalanine metabolism; vitamin B6 metabolism; biotin metabolism; tryptophan metabolism; one-carbon pool by folate; glycerophospholipid metabolism; and cysteine and methionine metabolism (Fig. 5C.1, Table S6). In the SP versus SMP comparison, the metabolic pathways encompassed histidine metabolism; phenylalanine metabolism; vitamin B6 metabolism; biotin metabolism; and arginine and proline metabolism (Fig. 5C 5C.2, Table S7). In the MP versus SMP comparison, the metabolic pathways involved alpha-linolenic acid metabolism; histidine metabolism; arginine and proline metabolism; and cysteine and methionine metabolism (Fig. 5C.3, Table S8). Then, 8, 8, and 6 differential metabolites were identified from the selected metabolic pathways (impact value > 0.1) in the comparisons of SP vs. MP, SP vs. SMP, and MP vs. SMP, respectively (Fig. 5D), which were denoted as the key differential metabolites in this study.

Furthermore, considering the relatively favorable growth performance of the SP group, a detailed analysis was performed to elucidate the differential metabolite pathways between this group and the other two groups. Specifically, eight differential metabolic pathways were identified between the SP and MP groups, while five differential metabolic pathways were detected between the SP and SMP groups. These pathways were closely linked to vitamin metabolism, as well as amino acid metabolism and synthesis processes (Fig. 6).

Fig. 6.

The metabolites and metabolic pathways of the SP group differ from those of the other two groups. The red color of the metabolites indicates upregulation in the SP group, green indicates downregulation, and the black compounds represent nonsignificance. The rounded rectangles with orange color represent pathways related to amino acid metabolism and synthesis, and the purple color represents pathways related to vitamin metabolism

Correlations between production performance and serum hormones, rectal microbiota, and serum metabolites

Spearman correlation analyses were performed to investigate the relationships between differential production performance and serum hormones, rectal microbes, and specific serum metabolites in donkey foals across the three groups (Fig. 7). As shown in Fig. 7B and C, specific differential microbes, including Oscillospiraceae_NK4A214_group, Oscillospiraceae_UCG-005, Porphyromonas, Streptococcus, and Akkermansia, exhibited significant positive correlations with ADG, serum levels of thyroxine, IGF-І, vitamin B6, as well as the serum metabolites serotonin and pyridoxine. In contrast, these microbes showed negative correlations with other differential serum metabolites such as L-Phenylalanine, biotin, L-Tryptophan, pyridoxal phosphate, L-Methionine, and 2-Lysolecithin (Fig. 7C). Notably, these metabolites were negatively associated with ADG, serum thyroxine and vitamin B6 (Fig. 7D). Additionally, pyridoxine and serotonin were positively associated with ADG, serum thyroxine and vitamin B6 (Fig. 7D).

Fig. 7.

Association of differential production performance, rectum microbes, serum hormones, and metabolites in donkey foals among three groups (n = 7). (A) The microbes were enhanced by concentrate supplementation with soybean meal as a protein source. (B) Correlations between differential rectum microbes and growth performance, as well as serum hormones. (C) Correlations between differential rectum microbes and serum metabolites. (D) Correlations between differential rectum metabolites and growth performance, as well as serum hormones. Differences in association were considered statistically significant at the levels of *P < 0.05, **P < 0.01, and ***P < 0.001

Discussion

The pre-weaning development of the gastrointestinal tract critically influences the growth, production performance, and health of animals in adulthood [32, 33]. Weaning represents a highly stressful transition for young animals, primarily due to the abrupt separation from their mothers and the need to adapt to a new diet. To alleviate weaning stress, the gastrointestinal tract must adapt to solid feed through adaptations that include absorption, transportation, pH regulation, and immune function [34]. Consequently, strategies such as the mixed feeding of breast milk with concentrated starter feed supplements or plant-based solid feed have been introduced to promote the morphological and functional maturation of the gastrointestinal tract, thereby enhancing the efficiency of solid feed digestion and absorption [35–37]. Soybean meal serves as the a critical protein source in starter feed for dairy calves due to its optimal amino acid profile, high digestibility, and excellent palatability. Extruded soybean meal has been shown to enhance dry matter intake, body weight gain, and ADG without compromising the FCR [38]. In this study, concentrate supplementation with soybean meal as the primary protein source, compared to supplementation based on milk-derived ingredients, also enhanced the growth rate of donkey foals by promoting ADG and ADFI. Soybean meal more closely aligns with the natural dietary composition compared to milk-derived ingredients and effectively promotes the secretion of digestive enzymes. Consequently, when donkey foals transition from breastfeeding to solid food, the inclusion of soybean meal in starter feed provides superior support for their growth and development [39, 40]. Additionally, during the heat treatment of soybean meal in the pelleted feed production process, anti-nutritional factors (such as trypsin inhibitors) were broken down, and the content of small peptides increased, which was beneficial for improving the growth performance [41].

Serum indices, including growth-related hormones and biochemical indicators, play a pivotal role in regulating the growth and development of donkey foals. In this study, concentrate supplementation with soybean meal as a protein source increased the levels of thyroxine, GH, and IGF-І while decreasing the cortisol level in the serum. Soybean meal is a rich source of soy protein, which has been shown to stimulate GH synthesis [42]. Moreover, isoflavones in soybean meal may promote compensatory thyroxine synthesis through feedback regulation mechanisms [43]. As a direct precursor for thyroxine synthesis, tyrosine plays a crucial role in enhancing the biosynthesis of thyroid hormones [44]. Concurrent supplementation with tryptophan and phenylalanine has significantly elevated IGF- І levels in infants [45]. Cortisol, an adrenal steroid hormone, serves as a key indicator of hypothalamic-pituitary-adrenal (HPA) activity and is widely used to assess stress responses [46]. The relatively lower serum cortisol level observed in the SP group may indicate a potential reduction in the stress response. In the current study, the serum lipid biochemical markers also exhibited alterations in TCHO, HDL-Cho, and LDL-Cho levels. The MP group showed increases in serum TCHO and LDL-Cho levels. Saturated fatty acids abundant in milk-derived ingredients can upregulate genes related to cholesterol synthesis by activating sterol-regulatory element binding protein 2 (SREBP-2), thereby promoting cholesterol synthesis and increasing the level of LDL-Cho in the serum [47, 48]. Compared with the MP group, TCHO and LDL-Cho in the serum of the SP group were significantly decreased. Soybean meal is a rich source of plant-based protein, and the dietary fiber in soybean meal effectively decreased serum TCHO and LDL-Cho levels by enhancing bile acid metabolism and promoting the resynthesis of bile acids from cholesterol, while phytosterols competitively inhibited the absorption of cholesterol in the intestine [49, 50]. In the SMP group, the combination of soybean meal and milk-derived ingredients may reduce cholesterol synthesis and enhance the transport of HDL-Cho and cholesterol to the liver, potentially via the synergistic interaction between soybean protein and milk protein [51, 52].

The gut microbiota plays a crucial role in optimizing the growth and development, gut health, and feed efficiency of donkey foals [53]. Donkeys can extract more energy from diet via microbially mediated strategies [34]. Previous research has demonstrated that varying protein sources in the diet can significantly influence the fecal microbiota of piglets [54]. Soy protein in soybean meal has the potential to modulate gut microbial composition by enhancing microbial diversity and promoting the proliferation of beneficial microorganisms [55]. In this study, supplementation with concentrate containing soybean meal reshaped the composition of the gut microbial community, enhanced microbial diversity, and increased the Firmicutes-to-Bacteroidota ratio in donkey foals. Equines rely on the hindgut (cecum and colon) for fiber fermentation, where a high proportion of Firmicutes is typically associated with efficient fiber degradation. This process enhances short-chain fatty acid (SCFA) production, thereby strengthening the intestinal barrier, which serves as a critical indicator of healthy digestion [56, 57]. Supplementation with soybean meal protein, compared to milk-derived ingredients, also altered the composition of the rectal microbiota at the genus level in donkey foals. For instance, the relative abundance of Oscillospiraceae_UCG-005, Oscillospiraceae_NK4A214_group, Akkermansia, Porphyromonas, and Streptococcus increased. Conversely, the relative abundance of Bacteroides, Christensenellaceae_R-7_group, and Fusobacterium decreased. Oscillospiraceae_UCG-005 and Oscillospiraceae_NK4A214_group, which belong to the phylum Firmicutes, play protective roles in intestinal cells by producing butyrate and are crucial participants in cellulose decomposition processes [58, 59]. Akkermansia, a mucin-degrading genus within the phylum Verrucomicrobia, plays an essential role in maintaining the integrity of the mucin layer and alleviating intestinal inflammation [60]. Unlike milk-derived ingredients, soybean meal contains insoluble fiber that can promote the growth of beneficial intestinal probiotics such as Akkermansia, thereby enhancing intestinal barrier function, energy metabolism, and nutrient absorption efficiency [61]. Furthermore, Akkermansia modulates tryptophan metabolism to produce aryl hydrocarbon receptor agonists, which suppress HPA axis activity and reduce cortisol levels [62]. Streptococcus supports the maintenance of mucus layer homeostasis through metabolite exchange with Akkermansia, thereby fostering a symbiotic relationship that enhances gut health [63]. Porphyromonas, a genus of proteolytic bacteria within the phylum Bacteroidota, exhibits the ability to degrade complex host-derived proteins and may be involved in interactions within the gut microbiota [15]. Christensenellaceae, including the genus Christensenellaceae_R-7_group, was significantly enriched in individuals with a normal body mass index compared to obese individuals and showed an increase following diet-induced weight loss [64, 65]. Bacteroides may contribute to the development of metabolic disorders, and an increased abundance of certain species may be associated with inflammatory processes [66–68]. Fusobacterium may potentially contribute to visceral hypersensitivity and colonic dysmotility [69].

In this study, serum metabolites were altered due to the variation in protein sources within the concentrated feed supplements (soybean meal or milk-derived ingredients). Notably, the SP group exhibited the highest production performance, characterized by increased levels of pyridoxine and serotonin, as well as decreased levels of L-phenylalanine, biotin, L-tryptophan, pyridoxal phosphate, L-methionine, and 2-lysolecithin. Serum pyridoxine and pyridoxal phosphate are both involved in vitamin B6 metabolism. Pyridoxine is the predominant form of vitamin B6 and plays a critical role in the proper functioning of the nervous system [70]. Pyridoxal phosphate, a versatile coenzyme, plays a pivotal role in various biological processes, including biomolecular synthesis and amino acid metabolism, which are essential for protein synthesis [71]. Consequently, serum vitamin B6 levels were measured in this study to verify if changes in pyridoxine and pyridoxal phosphate levels affect vitamin B6 synthesis. Consistent with the prediction, serum vitamin B6 in the SP group was significantly higher than that in the other two groups. Vitamin B6 protects cells from oxidative damage, promotes cellular growth, participates in neurotransmitter synthesis, and modulates insulin secretion, thereby regulating energy metabolism and supporting overall growth [72]. Intestinal microorganisms may modulate the serum vitamin B6 and are potentially associated with alterations in metabolic pathways [73, 74]. Furthermore, pyridoxal phosphate is a cofactor that catalyzes the decarboxylation reaction, thereby converting tryptophan to serotonin [75]. Supplementation of tryptophan in sheep diets can enhance serotonin levels and improve performance [76]. Serotonin can promote the growth of young ruminants by affecting glucose and insulin metabolism [77]. In addition, the consumption of dietary fiber has been shown to enhance insulin sensitivity, which is positively correlated with plasma serotonin levels [78]. L-phenylalanine may be rapidly utilized for muscle development and protein synthesis during phases of rapid body growth [79]. The downregulation in serum L-methionine levels in the SP group may indicate rapid uptake for protein synthesis, improved one-carbon metabolism efficiency, and enhanced antioxidant capacity, thereby supporting better growth performance [80].

This study identified significant correlations between specific growth performance and serum growth-related hormones, differential rectal microbes, and serum metabolites. These findings indicated that soybean meal, as a protein source in concentrated supplements, modified the rectal microbiota and serum metabolites compared with milk-derived ingredients used as the protein source in supplements for donkey foals. This modification further regulated amino acid and vitamin metabolism, thereby promoting the growth and development of foals.

Conclusion

In summary, concentrate supplementation with soybean meal in donkey foals altered the rectal microbiota, specifically driving a substantial increase in Firmicutes abundance. This is consistent with the fermentative response to plant-derived fibrous carbohydrates, which enhances microbial energy production and serves as a primary driver of microbial diversity. The overall changes increased the abundance of specific genera of microorganisms in the rectum, such as Akkermansia, Porphyromonas, Oscillospiraceae_NK4A214_group, Oscillospiraceae_UCG-005, and Streptococcus. These alterations in the microbiota were correlated with changes in microbiota-derived serum metabolites, which were identified as key differential metabolites involved in amino acid and vitamin metabolic pathways. Moreover, changes in the rectal microbiota and serum metabolome were associated with alterations in serum hormone (thyroxine, GH, IGF-I, and cortisol) concentrations and vitamin B6 levels. These variations were also linked to improvements in ADG and feed efficiency. The findings of this study suggest that soybean meal is an effective protein source in concentrate supplementation for suckling donkey foals and a potential alternative to milk-derived ingredients in these feeds.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

SP

Soybean meal as the primary protein source of diet

MP

Milk-derived ingredients as the primary protein source of diet

SMP

Both soybean meal and milk-derived ingredients the protein sources of diet

ADG

Average daily gain

ADFI

Average daily feed intake

TMR

Total mixed rations

FCR

Feed conversion rate

GH

Growth hormone

IGF-Ⅰ

Insulin-like growth factor-Ⅰ

TG

Triglycerides

TCHO

Total cholesterol

LDL-Cho

Low-density lipoprotein cholesterol

HDL-Cho

High-density lipoprotein cholesterol

ASVs

Amplicon sequence variants

PCoA

Principal coordinate analysis

LDA

Linear discriminant analysis

LEfSe

Linear discriminant analysis effect size

LC-MS/MS

Liquid chromatography-tandem mass spectrometry

KEGG

Kyoto Encyclopedia of Genes and Genomes

OPLS-DA

Orthogonal partial least squares discriminant analysis

SEM

Standard errors of the mean

ACE

Abundance-based coverage estimator

Sobs

Observed species richness

SREBP-2

Sterol-regulatory element binding protein 2

HPA

Hypothalamic-pituitary-adrenal

Author contributions

CHG and CYZ were contributed to the study design and interpretation of the findings; YPL, HFZ, and HL conducted the animal work; YPL and HFZ participated in lab analysis and manuscript preparation; YPL and CHG oversaw the development of the experiment and wrote the final version of the manuscript. GGZ, YJ, and BYD had supervision. All authors read and approved the final version of the manuscript.

Funding

This research was supported by the National Key R&D Program of China-Korea Cooperative Project (Grant Nos. 2019YFE0107700 and NRF-2019K1A3A1A20081146), the Shandong Province Modern Agricultural Technology System Donkey Industry Innovation Team (Grant No. SDAIT-27), and the Key R&D Program of Shandong Province (Grant No. 2022TZXD0018), the Research and Development Project on Full-Priced Mixed Pellet Feed for Donkeys (Grant No. DEEJ-YB-2024-01-033).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.