In compliance with the highest standards of animal care and welfare, the study was conducted under the strict supervision and authorization of the Committee for the Care and Use of Experimental Animals at Jiangxi Agricultural University, as per protocol number JXAULL-2021036.

A total of twenty-two healthy male Hu sheep, with similar body weight (19.97±0.41 kg) and age (66.75±2.11 days), were divided evenly into two groups. One group was fed a basal diet (n = 11, CON), while the other group received an additional 1.0 g per day of HDMF on top of the basal diet (n = 11, HDMF). The ratio of concentrate to forage in the basal diet was 70:30, with the specific ingredient composition and chemical components as shown in Table 1. HDMF was purchased from Hubei Yongkuo Technology Co., Ltd. (Tianmen, China), with a purity of 99.5%. The entire trial lasted for 75 days, including a 15-day adaptation period and a 60-day experimental period. The sheep were housed individually, with free access to feed and a clean supply of drinking water throughout the trial. Given the requirement to monitor growth performance in this trial, and taking into account that each group comprised 11 biological replicates, we opted not to employ a crossover design.

During the trial, the dry matter intake was recorded daily. The body weights were measured on two consecutive days at the beginning and end of the trial. Two days before the trail’s end, blood was collected from the jugular vein into blood collection tubes and centrifuged at 2,012×g for 15 minutes to obtain serum samples. Before morning feeding, rumen contents were collected using the oral intubation method as described by Paz et al [17], and the the first 50 mL was discarded to reduce saliva contamination. The collected rumen contents included both liquid and solid phases. After passing through four layers of gauze and being divided into centrifuge tubes, they were immediately frozen in liquid nitrogen and then transferred to a laboratory freezer at −80°C for storage.

The rumen pH value is measured immediately after the rumen fluid is obtained using a portable pH meter (Testo 206, Testo AG, Schwarzwald, Germany). The ammonia nitrogen (NH₃-N) concentration was determined following the protocol established by Broderick and Kang [18], utilizing the phenol-hypochlorite colorimetric method. The concentration of microbial crude protein (MCP) was measured using the Folin phenol method, which is an adaptation of Lowry’s assay, as delineated in the work of Makkar et al [19]. In this study, the VFA assessed included acetate, propionate, isobutyrate, butyrate, isovalerate, and valerate, with the collective sum of isobutyrate, isovalerate, and valerate categorized as branched-chain VFA (BCVFA). Each VFA component was identified by its relative retention time and quantified against a standard curve, both established under identical operational conditions. The analysis was conducted on a gas chromatograph (GC-2014, Shimadzu Corporation, Kyoto, Japan), following the scheduled procedures and parameter settings outlined in Qiu et al [20].

Serum parameters included cortisol and antioxidant indicators, which are used to assess stress status and antioxidant capacity, respectively. In this study, the antioxidant indicators measured included total antioxidant capacity (T-AOC), glutathione peroxidase (GSH-Px), superoxide dismutase (SOD), malondialdehyde (MDA), and reactive oxygen species (ROS). All assays were conducted strictly according to the instructions provided with the corresponding assay kits from the Beijing Sinouk Institute of Biological Technology (Beijing, China). The oxidative stress index (OSI) was calculated by dividing ROS by T-AOC, serving as a comprehensive measure of antioxidant capability.

Genomic DNA from rumen fluid samples was extracted utilizing a Bacterial DNA Kit (OMEGA, Omega Bio-Tek, Norcross, GA, USA), employing a two-step protocol that included bead-beating for initial lysis [17]. The integrity and concentration of the DNA were assessed using 1% agarose gel electrophoresis and a spectrophotometer (NanoDrop 2000, Thermo Fisher Scientific, Wilmington, DE, USA), respectively. A total of 22 samples, each exhibiting high purity and quality, were forwarded to Allwegene Gene Technology Co., LTD (Nanjing, China) for further amplification and sequencing. The V3-V4 hypervariable regions of the 16S rRNA gene were targeted for amplification to characterize bacterial diversity and community structure, utilizing barcoded primers as reported by Wei et al [21]: 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACNNGGGTATCTAAT-3′). The amplification protocol and reaction conditions were in accordance with those detailed by Qiu et al [20]. Polymerase chain reaction amplicons were inspected on a 1% agarose gel, purified using Agencourt AMPure XP Kits (Beckman Coulter, Brea, CA, USA), and subsequently used to construct libraries for microbial diversity sequencing. High-throughput sequencing was conducted on an Illumina NextSeq 2000 PE300 platform, generating paired-end reads. The raw sequencing data have been archived in the Sequence Read Archive (SRA) under accession number PRJNA1158230.

The sequencing data were subjected to analysis using Quantitative Insights Into Microbial Ecology (QIIME 2) [22]. The paired-end reads were merged using Fast Length Adjustment of Short reads (FLASH, version 1.2.11) [23], with a maximum mismatch rate of 0.10 and a minimum overlap of 10 base pairs. Sequences were filtered out if they did not meet the following criteria: length between 200 and 500 bp, quality score of at least 20, absence of ambiguous bases or chimeric sequences, and match to primer sequences and barcode tags. The sequences that passed the quality control were denoised into amplicon sequence variants (ASVs) utilizing the Deblur algorithm within QIIME 2. ASVs from all samples were rarefied to a common depth of 31,113 reads, yielding an average of 265 ASVs per sample. Taxonomic classification of each ASV was conducted using the Ribosomal Database Project (RDP) Classifier at a 70% confidence threshold, aligning with the bacterial SILVA 138 database.

The normality of the data distribution was assessed using the Shapiro-Wilk test. As all data, except for the microbiome data, met the assumptions of normality, an independent samples t-test was performed using SPSS software (version 20; IBM Corporation, Armonk, NY, USA) to compare the mean differences between the CON and HDMF groups. Statistical significance was set at p<0.05, with a trend was noted p-values between 0.05 to 0.10 (0.05≤p<0.10). For the microbiome data, Wilcoxon rank-sum test was used to compare the differences between the CON and HDMF groups. QIIME 2 was employed to calculate five alpha diversity indices—Chao 1, observed species, phylogenetic diversity (PD) whole tree, Shannon index, and Simpson index—to characterize the richness and evenness of the rumen bacterial community. To elucidate the differences and similarities between the CON and HDMF groups, principal coordinates analysis (PCoA) and nonmetric multidimensional scaling (NMDS) were performed using R software (version 4.0.2), based on Bray-Curtis distances. An analysis of similarities (ANOSIM) was also conducted to explore the community structure similarities between the CON and HDMF groups, utilizing the vegan package within R software. Furthermore, an Linear Discriminant Analysis (LDA) Effect Size (LEfSe) analysis was introduced to identify biomarkers within each group. This analysis was carried out using LEfSe software, with an LDA score threshold set at 3.0 [24]. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) was utilized to predict the metagenomic contributions of the detected communities, thereby elucidating the intrinsic ruminal functions of the bacterial microbiota.

Table 2 presents the effect of dietary HDMF supplementation on growth performance of Hu sheep. Compared to the CON group, dietary supplementation with HDMF did not result in significant differences in final body weight (FBW), average daily weight gain (ADG), average daily dry matter intake (DMI), or feed to gain ratio (F/G) (p>0.05). However, there was a numerical increase of 4.72% in FBW, 11.12% in ADG, and 4.55% in DMI. Additionally, there was a 7.56% decrease in the F/G.

The effect of dietary HDMF supplementation on serum antioxidant capacity of Hu sheep is listed in Table 3. Dietary supplementation with HDMF significantly increased the levels of GSH-Px and SOD, reduced the levels of MDA and ROS, and ultimately lowered the OSI (p<0.05). Additionally, there was a trend suggesting that HDMF supplementation might decrease cortisol levels (p = 0.055).

The effect of dietary HDMF supplementation on rumen fermentation characteristics of Hu sheep is shown in Table 4. Dietary HDMF supplementation had no significant impact on ruminal pH value, concentrations of MCP, NH₃-N, total VFA or individual VFA (p>0.05). In contrast, valerate concentration tended to decrease in the HDMF group (p = 0.096), while the proportion of valerate was significantly lower in the HDMF group compared to the CON group (p<0.05).

As shown in Figure 1, the AI-2 concentration and the level of biofilm formation were higher in the HDMF group compared to the CON group (p<0.05). No significant differences were observed in microbial density (p>0.05).

HDMF supplementation did not alter rumen bacterial richness or evenness, as indicated by the Chao1, observed species, PD whole tree, Shannon index, and Simpson index (p>0.05, Table 5). A total of eleven phyla were observed with relative abundances greater than 0.1%. Bacteroidota, Firmicutes, and Proteobacteria occupied the top three positions, accounting for 40.18%, 37.86%, and 19.23%, respectively. However, no significant differences were observed between the CON and HDMF groups for any of these phyla (p>0.05, Table 6). Twenty-four genera were observed with relative abundances greater than 0.5%, with Prevotella and Succinivibrio occupying the top two positions, accounting for 32.75% and 13.45%, respectively. The relative abundance of Eubacterium eligens group was significantly lower in the HDMF group compared to the CON group (p<0.05, Table 7).

As shown in Figure 2, both PCoA (Figure 2A) and NMDS (Figure 2B) plots clearly reveal overlaps between the HDMF and CON groups. Additionally, ANOSIM found no significant differences in the microbial composition between the HDMF and CON groups (R = 0.0116, p = 0.303).

The LDA distribution bar graph (Figure 3A) indicates that a total of twenty-one species were observed with an LDA score greater than 3.0 and a significance level of p<0.05. In the CON group, the biomarkers with significant discriminative power were Bacilli, Erysipeloclostridiaceae, Shuttleworthia, Eubacterium eligens group, and Erysipelotrichaceae UCG 002. In the HDMF group, they were Euryarchaeota, Desulfobulbia, Methanobacteria, Desulfobulbales, Methanobacteriales, Desulfobulbaceae, Rikenellaceae, Methanobacteriaceae, Riknellaceae RC9 gut group, Prevotellaceae UCG 003, Desulfobulbus, Methanobrevibacter, and Selenomonas ruminantium. Cladogram (Figure 3B) derived from the LEfSe analysis revealed the similar biomarkers with LDA distribution bar graph.

This study identified 29 gene families in the rumen fluid samples, 23 of which had relative abundances greater than 0.1% (Table 8). Gene families related to the metabolism of cofactors and vitamins, carbohydrate metabolism, and amino acid metabolism were predominant in the rumen microbiome of both HDMF and CON groups. However, no significant differences were observed among these three gene families or any other gene families examined in this study (p>0.05).

HDMF exhibits antimicrobial, anti-inflammatory, and antioxidant properties in the biopharmaceutical field [25,26], making it a promising candidate for animal production. HDMF has a sweet caramel-like odor and is one of the major compounds generated by the Maillard reaction [27]. However, no increase in DMI was observed in this study, suggesting that this caramel-like flavor may not a taste preferred by the Hu sheep. While the numerical enhancements in ADG and feed conversion ratio were observed with dietary HDMF supplementation, they did not reach statistical significance. The brevity of the feeding period may have obscured these differences. Echoing this, Li et al [28] reported that there was only a mere numerical increase in growth performance within the initial 60 days of glutamate supplementation, with growth significantly accelerating after 90 days of feeding Hu sheep. A longer feeding period is required to confirm the growth-promoting effects of dietary HDMF supplementation, as the current short-term feeding has already shown numerical improvements.

Serum hormone levels and antioxidant capacity provided a good explanation for the numerical improvements in growth performance. Cortisol is a crucial glucocorticoid in animals, playing a pivotal role in stress response, metabolism, immune modulation, and cognitive function [29]. It helps animals adapt to various challenges by regulating energy availability and maintaining homeostasis [30]. However, high cortisol levels signify a state of physiological arousal, preparing the animal to cope with perceived threats [30]. Dietary supplementation with HDMF tended to decrease the level of cortisol, implying that HDMF can alleviate chronic stress in animals. GSH-Px and SOD are integral components of the antioxidant defense system in animals [31,32]. GSH-Px uses glutathione to reduce hydrogen peroxide and lipid peroxides, thereby preventing oxidative damage to cellular components [31]. On the other hand, SOD catalyzes the conversion of superoxide radicals into molecular oxygen and hydrogen peroxide, which GSH-Px then further metabolizes [31]. The coordinated action of these enzymes is essential for maintaining cellular redox balance and mitigating the harmful effects of ROS. Higher levels of GSH-Px and SOD were observed in the HDMF group, likely due to the inherent antioxidant properties of HDMF itself. MDA, a byproduct of lipid peroxidation, serves as a biomarker for oxidative stress in animals [33]. The lower levels of MDA in the HDMF group indicate that the addition of HDMF alleviated stress, which corresponds to the trend of decreased cortisol levels. ROS, naturally produced in cells, play a dual role in physiological processes and pathogenesis [34]. Essential for cell signaling and homeostasis, ROS are also implicated in various diseases when present in excessive levels, leading to oxidative damage to lipids, proteins, and DNA [34]. The HDMF group exhibited lower levels of ROS, indicating that dietary supplementation with HDMF mitigated oxidative damage in the body. This reduction in overall oxidative stress is further supported by a decreased OSI. The data suggest that HDMF supplementation in the diet can enhance the antioxidant capacity and reduce oxidative stress in Hu sheep, potentially improving their growth performance.

Rumen fermentation is a critical biological process in ruminants, pivotal for the efficient degradation of fibrous plant material [35]. The complex microbial activity within the rumen transforms cellulose into VFA, which are essential for providing the primary energy substrates for the ruminants [36]. Furthermore, these rumen microorganisms are crucial in nitrogen metabolism, as they convert dietary protein into NH₃-N and amino acids. These nitrogenous compounds are then utilized for the synthesis of MCP, which is vital for the protein nutrition of ruminants [37]. In this study, dietary supplementation with HDMF did not alter most of the rumen fermentation characteristics, except for the concentration and proportion of valerate. Valerate is principally produced through the microbial fermentation of dietary carbohydrates and branched-chain amino acids [36]. However, propionibacteria also convert lactic acid into propionic and valeric acids via the propionic acid metabolic pathway [38]. Both the concentration and proportion of valerate decreased after dietary supplementation with HDMF, possibly due to the antimicrobial effect of HDMF leading to a reduction in the activity of propionibacteria. This reduction can also be indirectly corroborated by the decrease in the numerical values of propionate concentration and proportion. These results suggest that the dietary addition of HDMF may inhibit the activity of some VFA-producing microorganisms, thereby reducing the production of valerate. However, the specific mechanisms remain to be elucidated through further investigation.

In the rumen, the LuxS/AI-2 QS system, identified as the dominant QS mechanism through metagenomic and metatranscriptomic studies, is crucial for interspecies communication among rumen bacteria [6,8]. AI-2, the central and distinctive signaling molecule of the LuxS/AI-2 QS system, allows bacteria to sense its concentration and synchronize their collective actions, such as adjusting cell density and initiating biofilm formation [9,39]. The study revealed that dietary addition of HDMF led to a marked enhancement in AI-2 levels and biofilm formation. These findings suggest that HDMF stimulates rather than suppresses QS among rumen microbial communities, contradicting earlier literature [12]. The discrepancy in outcomes might stem from the fact that previous studies identifying HDMF as an inhibitor of LuxS/AI-2 QS focused on specific bacterial strains in vitro, whereas the rumen houses a diverse microbial community, which could respond differently to HDMF. Furthermore, the complex structure of biofilms not only bolsters bacteria’s resilience against external environmental challenges but also facilitates the adhesion of rumen microorganisms to feed particles [39]. This promotes the efficient breakdown of feed, thereby enhancing feed conversion efficiency [9,39]. The notable increase in biofilm formation observed in the HDMF group sheds light on the numerical upgrades in feed conversion rates and the concurrent increase in serum antioxidant activity following the dietary addition of HDMF.

Alpha diversity metrics quantify the richness and evenness of species within a specific sample or a single ecological community [40]. Alpha diversity plays a pivotal role in preserving the stability and functionality of the rumen ecosystem, significantly impacting the efficiency of feed digestion and nutrient absorption in ruminant animals [41]. There were no significant differences in either species richness or evenness between the CON and HDMF groups, suggesting that dietary supplementation with HDMF did not impact the alpha diversity of the rumen microbial community. Taxonomic annotation offers a more holistic perspective on microbial community structure across various levels. At the phylum level, no significant differences were detected between the CON and HDMF groups. At the genus level, the relative abundance of the Eubacterium eligens group in the HDMF group was significantly lower than that in the CON group. The Eubacterium eligens group is recognized for its pivotal role in producing short-chain fatty acids and its ability to utilize pectin, a complex carbohydrate found in plant cell walls [42]. This pectin-degrading ability is especially crucial in the rumen of ruminant animals, facilitating the fermentation process and bolstering the host’s energy yield. The relative abundance of the Eubacterium eligens group corresponded with the VFA production within the rumen fermentation characteristics, implying that the supplementation with HDMF might not be favorable for the efficiency of nutrient utilization. Nevertheless, this inference necessitates substantiation through further digestive experimentation.

The biomarker analysis revealed a higher discriminatory power at various taxonomic levels, spanning from phylum to species. Bacilli participate in the breakdown of hemicellulose and the utilization of polysaccharides [43], a function analogous to that of the Eubacterium eligens group previously discussed. This similarity further suggests that the dietary inclusion of HDMF could potentially exert adverse effects on nutrient utilization. The presence of Erysipeloclostridiaceae, Shuttleworthia, and Erysipelotrichaceae UCG 002 has been correlated with inflammation or pathological conditions within the body [44,45]. Following the supplementation of HDMF, a notable reduction in their relative abundances was observed, which is indicative of an amelioration in the animal’s health. This observation aligns with the reduced cortisol levels and enhanced antioxidant capabilities reported earlier, pointing towards a positive shift in the physiological balance. The class of Desulfobulbia, along with its affiliated Desulfobulbales, Desulfobulbaceae, and Desulfobulbus, are sulphate-reducing bacteria involved in biofilm formation [46]. Within the HDMF group, the elevated relative abundance of these bacteria coincides with an increased level of biofilm formation, substantiating the notion that HDMF supplementation can bolster the animal’s resilience against external stressors. Members of the Methanobacteria, Methanobacteriales, Methanobacteriaceae, and Methanobrevibacter, which are part of the Euryarchaeota, contribute to greenhouse gas emissions by producing methane through the metabolism of organic compounds in anaerobic environments [47]. The activity of Selenomonas ruminantium plays a vital role in the rumen’s hydrogen gas recycling process, as they contribute to sustaining the low partial pressure of hydrogen, a condition that is essential for the functioning of methanogenic archaea [48]. Methane production signifies an extent of energy loss, as this energy could have otherwise been harnessed for growth and productivity [49]. The elevated relative abundance of these bacteria in the HDMF group implies a higher degree of energy loss, which is reflected in the comparatively reduced levels of VFA production within this group. The Rikenellaceae RC9 gut group stands out as a significant component of the Rikenellaceae family. It exhibits a positive correlation with the uptake and metabolic processing of VFA, thereby indirectly influencing the immune capabilities of the host [50]. Research also indicates a link between the Rikenellaceae RC9 gut group and the generation of metabolic byproducts by rumen microorganisms [51]. Predominantly, these byproducts are concentrated in metabolic pathways that include galactose metabolism, the biosynthesis of unsaturated fatty acids, and the breakdown of fatty acids, all of which are intricately connected to the host’s energy metabolism [51]. The studies suggest that the Rikenellaceae RC9 gut group is capable of bolstering the host’s resistance to diseases and the efficacy of VFA absorption. This could partially account for the observed improvement in the growth performance of sheep with the dietary addition of HDMF, despite no observed increase in VFA output. A wealth of research indicates that an increase in the ruminant’s antioxidant defenses and immune function is correlated with a higher relative abundance of Prevotellaceae UCG-003 [52–54]. In addition, studies have revealed a significant positive correlation between Prevotellaceae UCG-003 and feed efficiency [54,55]. The elevated relative abundance of Prevotellaceae UCG-003 in the HDMF group provides a microbial rationale for the observed enhancement in immune function and improved feed efficiency resulting from dietary HDMF supplementation.