In Qunipa, Dazi District, Lhasa City, Tibet Autonomous Region (Latitude N: 29°48′22.54″, Longitude E: 91°37′14.89″, Altitude: 3,700 m), 16 Jersey cattle with an average body weight (BW) of 385.36 ± 46.25 kg (mean ± SD) were used as experimental animals. The experiment was carried out in a completely random design. All Jersey cattle were randomly assigned into two groups, with 8 animals in each group, the control group (CON group) were feeding basis total mixed rations (TMR) and the experimental group were feeding basis TMR plus 20(g/d/head) N-carbamylglutamate (NCG), the ingredients and chemical composition of the diets are given in Table 1. Jersey cattle were fed two equal meals at 08.00 and 18.00 h daily. All animals had free access to fresh water. The total experimental period lasted for 60 days, the acclimatization period lasted for 12 days until all Jersey cattle reached the stable dry matter (DM) intake according to the standard of metabolic BW, and the formal experimental period lasted for 48 days. The experimental feeds were offered ad libitum and the refusals were collected every day during the formal experimental period to measure the voluntary feed intake, and feed intake was measured.

During the experiment, 2 cows in the CON group developed physiological diseases, and the final samples were 6 cows in the CON group and 8 cows in the experimental group. The diets offered and left behind were recorded daily for calculating DMI and collected residual feed samples were used for chemical analysis. Feed samples were oven-dried at 65°C, ground to pass a 1-mm sieve, and stored for pending laboratory analysis. On day 41 of the formal experiment, before morning feeding, 5 ml blood sample was collected from each cattle in a vacuum tube pre-filled with Ethylene diamine tetraacetic acid (EDTA) for metabolomic assays, and an additional 5 ml blood sample was collected from each cattle in a vacuum tube pre-filled with heparin for blood parameters determinations. After 30 min still standing, samples were centrifuged at 3,000 rpm, 1 ml plasma was taken in the tube and immediately frozen by liquid nitrogen, and 3 replicates tubes of each cattle were stored at −80°C for blood parameters, antioxidant capacity, immunity, parameters associated with altitude sickness determining, and Liquid chromatography tandem-mass spectrometry (LC–MS/MS) analysis. On day 41 before morning feeding, fecal samples were collected in sterile plastic tubes, immediately frozen in liquid nitrogen, and stored at −80°C until mixed up for LC–MS/MS analysis. On day 42, 43, 44, 45, and 46 of the formal experiment, at 0, 3, and 6 h after morning feeding, 50-g fresh feces were taken from the rectum for nutrients determination except for crude protein (CP), another 50-g fresh feces were taken from rectum and added 10 ml 10% H₂SO₄ for CP determination, the feces samples in each day were mixed up and oven dried at 65°C, then ground to pass a 1-mm sieve for nutrients determination. On day 47 and 48 of the formal experiment, rumen contents were collected orally using an oral stomach tube at 0, 3 and 6 h after morning feeding, the first 100 ml of rumen contents were discarded, and the following 100 ml were rapidly collected for sampling (23) and stored at −80°C. After mixing the rumen contents from the 3 time points, approximately 10 ml of the contents was centrifuged at 10,000g for 15 min at 4°C, 1 ml supernatant was transferred into tubes containing 0.1 ml of 25% (w/v) metaphosphoric acid, and this mixture was stored at −20°C for subsequent determination of volatile fatty acids (VFAs) and ammonia. On day 47 and 48 of the formal experiment, 50 ml milk samples were collected at 6.00 and 19.00 h each day, The potassium dichromate preservative was added to the milk and mixed with the four milk samples in equal proportion. Composited milk samples were stored at 4°C and analyzed for milk quality determination.

Feeds and feces were used for nutrients determination, the gross energy (GE) was determined by an isothermal automatic calorimeter (5E-AC8018, Changsha Kaiyuan Instruments Co., Ltd, China). The DM, CP, GE, organic matter (OM), calcium (Ca), and phosphorus (P) were determined according to Association of Official Analytical Chemists methodologies (AOAC, 2002). Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were determined using a Fibretherm Fiber Analyzer (Gerhardt, Bonn, Germany) according to Van Soest et al. (24). Apparent total tract digestibility (ATTD) was calculated using acid insoluble ash (AIA) endogenous indicator method with the following formula (25):

The NH₃-N and VFA concentrations were determined according to Chen et al. (26). Milk samples were measured using Basic Unit MilkoScan FT + Type 76150 (Foss Electric, Hillerd, Denmark), blood biochemical parameters and immunity parameters (IgA, IgG, and IgM) were determined by Shenzhen Mindray BS-190 fully automatic biochemical analyzer (Shenzhen Mindray Bio-Medical Electronics Co., Ltd, Shenzhen, China). Concentration of antioxidant parameters (superoxidase dismutase, catalase, malondialdehyde (MDA), total antioxidant capacity (T-AOC) in plasma were analyzed by kits (Nanjing Jiancheng Bioengineering Institute Co., Ltd, Nanjing, China). Parameters associated with altitude sickness in plasma were determined by ELISA kits (Nanjing Jiancheng Bioengineering Institute Co., Ltd, Nanjing, China). Ammonia nitrogen (NH₃-N) and VFA were measured by the methods of 22. Blood oxygen saturation and heart rate were detected by pulse oximeter (CMS60D-VET, Qinhuangdao CONTEC Instruments Co., Ltd, China).

The LC–MS/MS analyses of plasma and feces were performed using an Ultra High Performance Liquid Chromatography (UHPLC) system (1290, Agilent Technologies) with a UPLC HSS T3 column (2.1 mm × 100 mm, 1.8 μm) coupled to Q Exactive (Orbitrap MS, Thermo). The mobile phase A was 0.1% formic acid in water for positive, and 5 mmol/L ammonium acetate in water for negative, and the mobile phase B was acetonitrile. The elution gradient was set as follows: 0 min, 1% B; 1 min, 1% B; 8 min, 99% B; 10 min, 99% B; 10.1 min, 1% B; 12 min, 1% B. The flow rate was 0.5 ml/min. The injection volume was 2 μl. The Q Exactive (QE) mass spectrometer was used for its ability to acquire Tandem mass spectrometry (MS/MS) spectra on an information-dependent basis (IDA) during an Liquid chromatography coupled with mass spectrometry (LC/MS) experiment. In this mode, the acquisition software (Xcalibur 4.0.27, Thermo) continuously evaluates the full scan survey MS data as it collects and triggers the acquisition of MS/MS spectra depending on preselected criteria. Electrospray ionisation (ESI) source conditions were set as following: sheath gas flow rate as 45 Arb, Aux gas flow rate as 15 Arb, capillary temperature as 320°C, full MS resolution as 70,000, MS/MS resolution as 17,500, collision energy as 20/40/60 eV in Normalized collision energy (NCE) model, spray voltage as 3.8 kV (positive) or −3.1 kV (negative), respectively.

All the MS raw data (.raw) files were converted to the mzML format using ProteoWizard, and processed by R package XCMS (version 3.2). The preprocessing results generated a data matrix that consisted of the retention time (RT), mass-to-charge ratio (m/z) values, and peak intensity. OSI-SMMS (version 1.0, Dalian Chem Data Solution Information Technology Co. Ltd) was used for peak annotation after XCMS data processing with in-house MS/MS database. After obtaining the sorted data, Principal Component Analysis (PCA) was performed and transformed into linear uncorrelated variables through orthogonal transformation, and a low-dimensional projection in PCA was formed. The plasma and feces samples of the NCG group and the CON group were analyzed based on LC–MS technology, using SIMCA software (V14.1, Sartorius Stedim Data Analytics AB, Umea, Sweden) to perform logarithmic conversion and centralization of the data dPCA was performed after formatting. To avoid the difference, variable was scattered to more principal components, the data is further used orthogonal partial least squares–discriminant analysis (OPLS–DA). The corresponding OPLS–DA model to obtain the R² value and Q² value of the random model was established to avoid the over-fitting of the test model and to evaluate the statistical significance of the model. The standard for screening differential metabolites was Student's t-test p < 0.05, and variable importance in the projection (VIP) was calculated in OPLS–DA model greater than 1. By searching the chemical structure of these differential metabolites in the Human Metabolome Database (HMDB), Kyoto Encyclopedia of Genes and Genomes pathway database (KEGG) and other published articles, and classifying them on the Super Class and Class levels based on the HMDB classification criteria. Differential metabolites between 2 groups mapped into their biochemical pathways through metabolic enrichment and pathway analysis based on MetaboAnalyst 5.0 (https://dev.metaboanalyst.ca).

The data of the nutrients digestibility, rumen fermentation, milk quality, and antioxidant parameters and index associated with altitude stress of the two groups were analyzed using the Shapiro–Wilk test for normal distribution with subsequent Student's t-test to compare means in the SPSS Statistics 22 (IBM, Chicago, USA) software. A p ≤ 0.05 was considered to indicate a statistically significant difference, and 0.05 < p < 0.10 represented a tendency.

The nutrients digestibility parameters were presented in Table 2, we found that feed NCG to Jersey dairy cows was tended to increase DMI (p = 0.09) and the digestibility of DM, CP, NDF, ADF, GE, and OM were not altered (p > 0.1) by supplementation of NCG.

As shown in Table 3, the milk fat content was tended to increase by 84.96% (p = 0.09), and the milk protein content in NCG group (5.09%) was also tend to decrease (p = 0.07) when compared with the CON group (6.47%). However, lactose content (p = 0.74), DM content (p = 0.31), DM without fat (p = 0.24), and somatic cells (p = 0.30) had no significant difference between NCG group and CON group.

Rumen fermentation parameters including ammonia nitrogen (NH₃-N) and VFAs were summarized in Table 4. The NCG supplementation had no effect (p = 0.44) on the NH₃-N concentration in the rumen. The molar content of total VFA and acetic/propionic were similar (p > 0.10) between the treatments. For individual VFA, except the molar proportion of butyric acid was tended to decrease (p = 0.08) in NCG group, other VFA like the molar proportion of acetic acid, propionic acid, butyric acid, valeric acid were also similar (p > 0.10) between NCG and CON group.

The oxygen saturation and heart rate were shown in Figure 1. The oxygen saturation (Figure 1A) was tended to increase (p = 0.09) in NCG group when compared with CON group. However, the heart rate (Figure 1B) was not different (p = 0.78) between treatments.

The plasma immunity and antioxidant parameters are shown in Table 5. Immunoglobulin A (IgA), immunoglobulin G (IgG), and immunoglobulin M (IgM) concentrations in plasma did not differ (p > 0.10) between treatments. In terms of antioxidant parameters, such as nitric oxide synthase (NOS), superoxide dismutase (SOD), MDA, catalase (CAT) and T-AOC were similar (p > 0.10) between treatments.

The results of index associated with altitude stress are shown in Table 6. Compared with the CON group, the plasma contents of transferrin, thyroxine, and endothelin in the NCG group were significantly decreased (p < 0.05), while, heat shock protein (p = 0.07) and aldosterone (p = 0.06) also tended to decrease. Plasma levels of C-reactive protein, cortisol, asymmetric dimethylarginine, and hypoxic-inducing factor were not significantly different (p > 0.10).

For plasma and feces samples, the PCA score plot showed that the model interpretation rates for the NCG and CON groups under the positive and negative ion mode conditions were R²X = 0.548 and 0.506, R²X = 0.506 and 0.565, respectively. Except for stool samples in negative ion mode (Figure 3B), other samples were well separated, and samples in the same group were well aggregated together (Figures 2A,B, 3A). An OPLS-DA supervised model was used to assess inter-group sample differences. During the analysis and fitting process of plasma and feces samples, in the positive ion mode of the OPLS–DA score plot, R²Y = 0.99 and Q² = 0.567, R²Y = 0.992 and Q² = 0.603, respectively, whereas in the negative ion mode, R²Y = 0.992 and Q² = 0.603, R²Y = 0.983 and Q² = 0.380, respectively. From the results of OPLS–DA score, whether for serum samples or stool samples, that can be seen that the two sets of samples were distinguished significantly, and the samples were all within the 95% confidence interval (Hotelling's T-squared ellipse), which indicating the model was stable and reliable (Figures 4A,B, 5A,B). The OPLS–DA model R²Y was ~1, indicated that the established model conforms to the real situation of the sample data, and Q² is greater than 0.4, which indicated that the OPLS–DA model could better explain the difference between the two sets of samples and there was no over-fitting (Figures 4C,D, 5C,D).

In the two samples of plasma and feces, with p < 0.05 and the VIP of the first principal component of OPLS–DA model greater than 1 as the critical value, and we identified a total of 114 differential metabolites. Among them, 42 metabolites were identified in plasma samples and 72 metabolites were identified in feces. Compound classification of differential metabolites in the plasma and feces samples were processed based on comparisons with the HMDB. Differential metabolites from plasma can be divided into 5 categories at the Super Class level, mainly including lipids and lipid-like molecules, benzenoids, organic acids, derivatives, etc. (Table 7), among which lipids and lipid-like molecules have the largest number of differential metabolites, which was 28. Compared to the CON group, the levels of differential metabolites classified as benzenoids in the NCG group, all increased significantly in plasma. The significantly increased plasma differential metabolites were 2,5-dihydroxybenzoate and salicyluric acid (fold change > 2). The detailed division of each Super Class can be divided into Class, and the number and proportion of plasma differential metabolites distributed in each Class are shown in Figure 6, lipids and lipid-like molecules mainly include fatty acyls, glycerophospholipids, steroids, and steroid derivatives. In addition, organic acids and derivatives and benzenoids are mainly based on carboxylic acids and derivatives and benzene and substituted derivatives, respectively. The differential metabolites identified from feces samples are more abundant, which can be divided into 9 types in the Super Class level (Table 8), and the significantly increased fecal differential metabolites were Butenafine (fold change > 2). Consistent with the results identified in plasma samples, lipids and lipid-like molecules also accounted for the largest percentage of differential metabolites identified in feces, with a total of 30 identified. Furthermore, there are two types of organoheterocyclic compounds and phenylpropanoids and polyketides also account for a large proportion, which have been identified 15 and 9 different metabolites, respectively. Similarly, the differential metabolites identified in the feces were divided into Class on the basis of Super Class, and the corresponding number and proportion of differential metabolites are shown in Figure 7. Different from the plasma differential metabolites, the differential metabolites from feces accounted for the largest proportion of lipids and lipid-like molecules in the classification of lipids and lipid-like molecules is prenol lipids. Different from plasma differential metabolites, in feces differential metabolites that prenol lipids occupy the largest proportion in lipids and lipoid-like molecules.

To further determine the biological significance of the differential metabolites, we searched differential metabolites in the KEGG and performed a metabolic pathway analysis using MetaboAnalyst 5.0 (https://dev.metaboanalyst.ca). After drawing into a bubble chart, we found that plasma differential metabolites can be enriched in a total of 11 metabolic pathways (Figure 8), mainly lipid metabolism and amino acid metabolism. These include Arg and proline metabolism and Arg biosynthesis, two metabolic pathways directly related to Arg synthesis. Similarly, the differential metabolites in feces were enriched in 7 metabolic pathways, mainly aromatic amino acid metabolism and lipid metabolism pathways (Figure 9). In addition, differential metabolites from plasma and feces can be enriched to two metabolic pathways in common, which are phenylalanine metabolism and glycerophospholipid metabolism.