All the required permissions regarding Animal care and management for the current research were received from the Institutional Animal Care and Sichuan Agricultural University committee (Ethics No. SYXK (chuan) 2019–187).

Fifty 4-month-old Chuannan black goats (Capra hircus) with similar body weight (13.75 ± 0.27 kg) were selected from Jintang County, Chengdu, China. The goats were numbered with ear tags and randomly divided into 5 groups with 10 replicates per group, and animals in each group were housed in individual cages. The study included 5 diets with CP levels of 8.12% (T8), 10.15% (T10), 12.17% (T12), 14.13% (T14), and 16.18% (T16), respectively, corresponding to 66.7, 83.3, 100, 117.7, and 133.3% of the recommended protein level based on the weight of 15 kg and average daily gain (ADG) 150 g specified in the Chinese Feeding Standard of Meat-type Sheep and Goat (NY/T 816–2021). The composition and nutritional value of diet was presented in Table 1. A 14-day adaptation period was provided to goats before the research trial, and the research trial lasted 56 days. During the entire study, a total mixed ration was provided to all goats twice daily on 8:00 and 16:00. The leftover in the trough was collected and weighed before feeding the next morning, and recorded as the leftover feed weight. The feed provided in the morning was maintained about 5% feed refusal.

A randomized complete design was used to determine the effect of protein level on various parameters using the Mixed procedure with SAS version 9.4 (SAS Institute Inc., Cary, NC). The model used to analyze these data was:

where μ = the general mean, Prot_i = the effect of the protein level i, and E_ij = the random error. Means were separated using orthogonal polynomial contrasts to examine linear or quadratic effects of protein level after protein level effect was detected significant. Data normality test was performed using the Shapiro–Wilk procedure and significance was declared at p < 0.05, while a trend was regarded at 0.05 ≤ p < 0.10. The Spearman correlation analysis was conducted on the 35 selected metabolites and the top four microbial genera with significant differences using Origin software (version 9.8.0.200).

Table 2 presented the results of the growth performance. The protein content in the diet had a quadratic impact on FBW (p < 0.05), ADG (p < 0.001) and FCR (p < 0.001). An increasing trend in ADFI was observed with the rise of dietary CP content (Not Significant, NS). The ADG (p < 0.001) and FCR (p < 0.05) of group T14 was the best compared to other groups.

Table 3 presented the results of the apparent digestibility of nutrients and nitrogen metabolism. A linear effect was noticed in the apparent digestibility of ADF (p < 0.05), while quadratic effects were observed in the apparent digestibility of DM (p < 0.05) and NDF (p < 0.05). A quadratic trend in the digestibility of NDF was observed as the dietary CP level increase (NS).

The dietary crude protein content had quadratic effect on nitrogen intake (NI) (p < 0.001), fecal nitrogen (FN) (p < 0.001), urinary nitrogen (UN) (p < 0.001), nitrogen retention (NR) (p = 0.001) and NR/NI (p < 0.05) (Table 3).

The results of the rumen fermentation parameters are presented in Table 4. A quadratic effect was observed in NH₃-N (p < 0.05), MCP (p < 0.05), TVFA (p < 0.01), acetate (p < 0.05), valerate (p < 0.01) and isovalerate (p < 0.001) with the increase in dietary protein content. Nevertheless, no substantial variation was noted in the ruminal pH value, propionate and butyrate, and the A: P.

Since the production performance of the T14 and T8 groups was the best and the worst respectively, six heads from each of these 2 groups were selected for rumen microorganisms metagenomic analysis and serum metabolomic analysis. Utilizing the Illumina sequencing platform and following quality control procedures, a total of 195,150,361 and 192,696,573 high-throughput sequences were acquired from rumen fluid samples of groups T8 and T14. The percentage of sequences with an average quality score exceeding Q30 was 93%. A Venn diagram was constructed to illustrate the intersections between the two groups (Figure 1a), it revealed that 127 OTUs (0.04% of the total sequence) were exclusive to group T14, and 115 OTUs (0.58% of the total sequence) were unique to group T8, and 3,615 OTUs (99.38% of the total sequence) were shared between the two groups.

The alpha diversity indices of microbial taxa are presented in Figure 1b. There was no significant difference in ACE and Chao indexes between the two groups (NS). However, the Simpson and Shannon indexes (p < 0.05) of group T14 were higher than those of group T8. The beta diversity result showed distinguishable microbial communities between groups T8 and T14 (Figure 1c), with PCo1 and PCo2 explaining 51.4 and 15.3% of the total variance, respectively.

As depicted in Figure 1d and Table 5, the predominant bacteria in groups T8 and T14 at the phylum level were Firmicutes, Bacteroides, Actinobacteria, and Proteobacteria, collectively constituting over 95% of the total bacterial sequence. The relative abundance of Proteobacteria in group T14 was higher than that in group T8 (p < 0.05), whereas the relative abundance of Firmicutes in group T14 was lower than that in group T8 (p < 0.05). At the genus level, the top 10 taxa were depicted in Figure 1e and Table 5. The relative abundance of Pseudomonas and Salmonella was higher (p < 0.05), whereas the relative abundance of Burkholderia and Bacillus was lower (p < 0.05) in group T8 compared to group T14. The LefSe analysis yielded comparable results, as shown in Figure 1f. To gain a deeper insight into the functional alterations of rumen microbiota between group T14 and group T8, Kyoto Encyclopedia of Genes and Genomes enrichment analysis was conducted. At the level 3 (Table 6), it was observed that pathways related to amino acid metabolism, including alanine, aspartate, and glutamate metabolism, glycine, serine, and threonine metabolism, cysteine and methionine metabolism, arginine and proline metabolism, and histidine metabolism, were down regulated in group T14 compared to group T8.

The PCA score plots comparing group T8 and T14 was presented in Figure 2a. The T8 and T14 group were entirely distinguishable, with the first and second principal components explaining 19.8 and 15.8% of the overall variance, respectively. The OPLS-DA also showed a clear differentiation between the two groups (Figure 2b). The parameters of the model indicated high reliability and predictivity (R2X = 0.605, Q2 = 0.806). The permutation test indicated the absence of an over-fitting (Figure 2c).

The identification of differently expressed metabolites between group T8 and T14 was conducted by utilizing the OPLS-DA model to obtain a VIP score >1, as well as with p values < 0.05. As depicted in Figure 2e and Supplementary Table S1, a total of 35 metabolites were found to be different. Among them, 14 metabolites such as genistein, 5-hydroxyindoleacetic acid, m-cresol, ornithine, and citric acid were upregulated, whereas 21 metabolites, including L-isoleucine, L-cysteine and L-histidine were downregulated. As depicted in Figure 2d and Supplementary Table S2, the differentially expressed metabolites were implicated in the pathway of histidine metabolism, arginine and proline metabolism, glutathione metabolism, and the citrate cycle. The schematic depiction of the metabolic network was illustrated in Figure 3.

In the present study, the ADFI was not affected by the dietary CP levels, and similar results were reported before by Zhu et al. (2020). However, the ADG and FCR of the goats were significantly improved as the dietary protein level increased from 8 to 14%, which was consistent with previous studies. Titi et al. (2000) observed the highest ADFI and ADG in the group fed a 16% protein diet when Jordan black goat were provided with diets containing four different protein levels (12, 14, 16 and 18%). Wang et al. (2014) also reported that a 17% CP diet led to a higher ADG and an improved FCR compared to the 15% CP group of Hainan black goats. Our results showed that 14% CP was the optimal, while previous studies reported that 16 and 17% were the best. The difference is speculated to be due to the different breeds and ages of goats.

Our results showed that the apparent digestibility of DM, CP, NDF, and ADF was significantly affected by the dietary protein level. This finding was consistent with Zhu et al. (2020), who reported improved apparent digestibility of nutrients in Anhui white goats as dietary protein level increased. In this study, the apparent digestibility of NDF and ADF of group T14 was higher than that of group T8, this may be due to the higher amount of degradable protein entering the rumen, resulting in more nitrogen source available for the growth of microorganisms. Proper dietary protein content can meet the requirements of microbial growth better (Liu et al., 2021). With the increase of dietary protein content, the activity of fiber-degrading bacteria increased, leading to an increase in the apparent digestibility of NDF and ADF.

In the present study, the NI and FN increased with the increase of dietary protein level. Zhang et al. (2020) reported that the FN of goats increased by 20.9 and 36.45% when fed with 11.5 and 13.0% CP compared to the control group (CP, 10.0%). Additionally, as the dietary protein content increased from 10 to 16%, NR of goat kids increased from 3.6 to 5.6 g/day, respectively. Furthermore, the UN of goats increased with the increase of dietary protein content, which was in line with the results reported by Katongole et al. (2022), who observed in dairy heifers. These findings suggested that an appropriate dietary protein level is beneficial for enhancement of nitrogen retention in goats.

Maintaining the rumen pH within the normal range is crucial for the health of rumiants, as any decrease or increase can disrupt the growth of rumen microbes. In this study, the dietary protein level did not have a significant effect on rumen pH, which remained within the normal range of 5.5–7.5 (Zhang et al., 2022). Putri et al. (2021) also reported that an elevation in dietary protein intake did not have significant impact on rumen pH levels. In the present study, the NH₃-N concentration exhibited a linear increase in response to dietary protein levels, which was consistent with the findings of Aguerre et al. (2016), and elevated NH₃-N concentrations were attributed to accelerated protein fermentation associated with higher rumen degradable protein content (Souza and White, 2021).

The current study demonstrated that the higher dietary protein led to elevated concentrations of TVFA in the rumen, which was consistent with the findings of Yang et al. (2016) observed in Chahaer lambs. In the present study, the concentration of propionic acid in group T14 was higher than that in group T8. As we all known that propionic acid serves as the primary substrate for gluconeogenesis in ruminants (Gleason et al., 2022). Consequently, it is suggested that the goat in group T14 exhibits greater energy utilization efficiency, potentially contributing to the superior growth performance. Zhou et al. (2019) and Zhang et al. (2022) documented that an increase in dietary protein level from 111 to 136 g/kg DM led to an elevated ruminal concentration of isobutyrate and isovalerate in beef cattle. The branched chain fatty acids in the rumen are primarily produced through breakdown of amino acids (Liu et al., 2023; Lu et al., 2024). The increase in dietary protein undoubtedly enhances the production of branched-chain fatty acids.

The observed shifts in rumen fermentation parameters are closely synchronized with alterations in the microbial community structure. In this study, the T14 group showed a higher abundance of beneficial bacteria such as Bacillus and Burkholderia, which are known to possess robust enzymatic capabilities for degrading complex polysaccharides. This microbial shift likely drives the observed quadratic increase in TVFA and acetate concentrations, as these taxa facilitate more efficient carbohydrate fermentation (Wang et al., 2022; Aguerre et al., 2016). The significant elevation in MCP levels in the T14 group suggests that the optimized nitrogen-to-energy ratio favored the proliferation of nitrogen-utilizing bacteria, enhancing the conversion of ammonia nitrogen into bacterial biomass (Zhu et al., 2020). These findings indicate that dietary protein levels modulate the rumen’s metabolic output by restructuring the microbial ecosystem, specifically by promoting taxa that optimize energy harvest and nitrogen assimilation.

The composition and structure of the microorganisms in the rumen are directly linked with ruminants’ ability to utilize different kinds of feed. The increase in protein content was associated with a concurrent increase of Shannon and Simpson indices, indicating a promotion of rumen microbial diversity. This finding aligned with the results reported in Holstein bulls by Costa-Roura et al. (2020). The current study demonstrated that Firmicutes the main phylum, rather than Bacteroidetes and Proteobacteria, which is consistent with a previous study conducted in Chahaer lambs (Yang et al., 2016). Research has shown that with age, the rumen microbiota of ruminants undergoes significant changes. Numerous studies have confirmed that Firmicutes contain some pathogenic bacteria, which to some extent reflect the dysbiosis or instability of the gastrointestinal microbiota community structure (Byndloss and Baumler, 2018). Bradley and Pollard (2016) demonstrated an association between Firmicutes and inflammation and metabolic syndrome. In this study, we observed that the T14 group exhibits a strong anti-inflammatory effect, suggesting that a high-protein diet can enhance digestion and absorption in goats. We speculate that a low-protein diet can cause inflammation in the body, which in turn affects changes in the abundance of gut microbiota. Another reason is that Firmicutes is associated with the breakdown of proteins (Wang et al., 2022), which is due to the presence of urease in Proteus, which can quickly break down urea and effectively utilize proteins in goats. Liu et al. (2022) reported that Proteobacteria are significantly involved in the degradation of cellulose, hemicellulose, starch, and oligosaccharides, suggesting that goats in group T14 exhibit enhanced degradability of fibrous materials.

At the genus level, the prevalence of Salmonella and Pseudomonas was lower in group T14 compared to group T8, while Burkholderia and Bacillus were more abundant. Salmonella plays a significant role as a pathogenic factor in gastrointestinal diseases within the US goat industry (Hempstead et al., 2022). Pseudomonas is frequently linked to sporadic mastitis in goats (Leitner and Krifucks, 2007). A study by Yusuf et al. (2018) found that the population of Pseudomonas in goat meat decreased when the diet was supplemented with 1.5% Andrographis paniculata leaves. On the other hand, Bacilli, functioning as a probiotic, can directly enhance host health through the production of diverse bioactive substances, regulation of the immune system, and influence on nutrient absorption (Todorov et al., 2022). da Silva et al. (2016) documented that continuous ruminal administration of Burkholderia provided protection against poisoning by Amorimia. Septentrionalis in goats. In summary, our findings suggest that increasing dietary protein from 8 to 14% leads to improvements in gastrointestinal health by decreasing the relative abundance of pathogenic bacteria and enhancing the relative abundance of beneficial bacteria. This ultimately enhances the digestion and absorption of nutrients, including fiber and protein.

Surprisingly, at KEGG level 3, there was a notable increase in the relative abundance of rumen microbiota genes associated with amino acid metabolism in group T8 compared to group T14. The observed phenomenon is hypothesized to be a result of the compensatory increase that occurs when the protein content of group T8 is in sufficient to support normal microbial growth. This suggests that the rumen microbiome exhibits adaptability under varying diet conditions, despite the differences in rumen microbiome composition between sheep with high and low residual feed intake (Zhang et al., 2021).

The results of the serum metabolomics analysis indicated a significantly higher concentration of substances with anti-inflammatory and anti-oxidative properties, such as genistein and 5-hydroxyindoleacetic acid, in group T14 compared to group T8. Cong et al. (2022) noted that the inclusion of genistein led to a decrease in miR-21 expression by inhibiting NF-kB, consequently leading to a reduction in the inflammatory response. Moreover, Liu et al. (2023) documented that genistein has the potential to reinstate the functions of anti-oxidases and reduce the expressions of NOD-like receptor pyrin domain-containing 3 inflammasome components, thereby alleviating inflammation. Furthermore, Yang et al. (2022) noted that 5-hydroxyindoleacetic acid has been identified as a new ligand for the G protein-coupled receptor (GPR35). This ligand has the ability to enhance neutrophil recruitment to inflammatory sites, facilitating the clearance of pathogenic bacteria through the activation of GPR35. Valente et al. (2021) asserted that the swift breakdown of serotonin in tissues to 5-hydroxyindoleacetic acid could impact the cycling of glucose and insulin, potentially leading to changes in the metabolism of developing cattle. In the current study, the concentration of serum γ-glutamylcysteine and L-Cysteine was found to be lower in the T14 group compared to the T8 group. Gamma-glutamylcysteine serves as the precursor for the synthesis of glutathione, which aids in the body’s elimination of detrimental oxygen-free radicals and alleviation of various inflammatory injuries (Braidy et al., 2019). The reduction in levels of γ-glutamylcysteine and L-Cysteine may be attributed to their utilization in the biosynthesis of glutathione. The present study observed enrichment of the glutathione metabolism pathway (Figure 2f). Furthermore, our findings indicated a negative correlation between the relative abundance of Pseudomonas and the anti-inflammatory metabolite genistein, while a positive correlation was observed between Burkholderia and Bacillus with the same metabolite (Supplementary Figure S2). Therefore, it is recommended that increasing the dietary protein level from 8 to 14% enhances the body’s anti-inflammatory and antioxidant capacity, thereby improving production performance. The serum concentration of sulfamethoxazole in group T8 was found to be lower compared to that in group T14. This observation suggests that the goat in group T14 may have possessed a more robust metabolic system, leading to a faster elimination of sulfamethoxazole.

Arginine is recognized as one of the most important amino acids in cells, as it acts as a precursor for the synthesis of various compounds such as nitric oxide, urea, polyamines, proline, glutamate, and creatine (Sawano et al., 2019). Additionally, the interconversion of arginine and ornithine in the body may account for the observed changes in ornithine and imino argininein the context of this study. Furthermore, the enrichment of the arginine and proline metabolism pathway was observed in group T14 compared to group T8. In the body, most of the energy is provided by the crucial metabolic pathways called tricarboxylic acid (TCA) cycle (Liu et al., 2023). Citric acid serves as a primary intermediate in the TCA, and plays a crucial role in mediating glycolysis and gluconeogenesis (Zhao et al., 2020). The rise in gluconic acid levels, as a derivative of glucose, indicates an increase in the availability of glucose. According to Michiels et al. (2023), the inclusion of 9 g/kg of gluconic acid in the diet resulted in improved performance of weaned piglets. The current investigation revealed a significant increase in the levels of citric acid and gluconic acid in group T14 compared to group T8. This finding suggests a potential enhancement in energy metabolism, which could result in increased energy availability for the growth of the T14 group goat (Figure 3). This further elucidates the reasons behind the superior production performance of group T14 goats compared to group T8.