The main research contents included the stable establishment of a zinc-deficient rat model; metabolomic detection for serum, urine, and feces samples; and metabolite annotation in metabolic pathways. The overall research flowchart is presented in Supplementary Figure 1.
2.1 Animal model and samples preparation
Five-week-old growing Sprague–Dawley (110–130 g) rats (n=30) were obtained commercially (Beijing Vital River Laboratory Animal Technology Co., Ltd.). The average weight of the rats on initiation of the experiment was 160.78±10.29 g. All rats were housed individually in stainless steel metabolic cages. To minimize trace metal contamination, silicon plugs and stainless-steel drinking water pipes were used in water bottles. Before the experiment, a seven-day adaptive feeding period with a standard rat diet (AIN-93G, containing 30 mg Zn/kg) was conducted, with ad libitum access to distilled water. The animal room was controlled for temperature (21–23 °C), humidity (53 %), and light (12 h light/dark cycle). The 30 rats, which were numbered from 1 to 30 according to weight, were randomly divided into two groups (n=15/group) at day 8, and then 30 random numbers were read from any row or column in the random number table. All selected random numbers were numbered from small to large. Rats with odd and even numbers were assigned to the normal zinc diet group (NZG) and low-zinc diet group (LZG), respectively. The NZG and LZG were fed a diet containing 30 mg and 10 mg Zn/kg, respectively. The animal diets were obtained commercially from the Beijing KeAoLiXie Animal Food Co., Ltd., China. Simultaneously, the Zn2+ content of the diets was confirmed by mineral analysis using atomic absorption spectroscopy. The normal diet had 30.40±2.86 mg/kg of zinc and the low zinc diet had 9.70±1.50 mg/kg. The contents of minerals, trace elements, and compositions in the diets are detailed in Supplementary Table 5. Both groups were fed distilled water without metal ions. Feed intake was recorded daily, and body weight was measured weekly throughout the course of the study. After being fed the respective diets for four weeks, the animals were transferred to metabolic cages, using a sieve and funnel, to collect urine and feces samples. The 24 h urine and stool samples were collected twice. Stool samples of all the rats were collected and placed in non-ionic Eppendorf tubes. Urine samples were collected in 50 mL ion-free plastic centrifuge tubes (Corning Incorporated) and centrifuged at 3000 rpm for 10 min, and the supernatant was collected. Subsequently, all the rats were fasted for 12 h and euthanized using intraperitoneal injection of 10 % chloral hydrate (0.3 mL/kg body weight). Blood samples were collected from the abdominal aorta and placed at room temperature for 2 h. After centrifugation for 15 min at 3000 rpm, serum was collected. Serum, urine, and fecal samples were stored at -80 ℃ for further biochemical and metabolic profile analysis. The animal study was reviewed and approved by the Animal Protection and Use Committee of the Harbin Medical University.
2.2 Serum biochemical analysis
Serum, urine, and feces Zn2+ and serum metallothionein concentrations were used to ascertain zinc levels in the two groups. Samples were prepared by wet digestion, and the levels of serum, urine, and feces Zn2+ concentrations were measured using an atomic absorption spectrophotometer. Briefly, HNO3:HClO4 (4:1) solution (Guaranteed reagent, Xilong Scitenfic, China) was added to 200 μL serum, and the mixtures were digested at 100 ℃ until the solutions became colorless. To reduce the acid content, 3 mL deionized water was added to the digestive solution and dried twice with heat. The digested samples were dissolved in 5 mL 5 % HNO3. The pretreatment methods for feces and urine samples were similar to those used for serum samples. The concentration of metallothionein was determined using enzyme-linked immunosorbent assay kits (Summus Biotechnology Development Co. Ltd.) and a Hitachi 7100 automatic biochemistry analyzer (Hitachi High-Technologies, Shanghai, China). We analyzed the differences between the two groups using independent sample t-tests. Data were considered statistically significant at P<0.05.
2.3 Metabolomics analysis of serum, urine, and feces
2.3.1 Sample preparation
To precipitate out protein from serum, 3× volume of methanol was added to the serum (350 mL). After vortexing for 2 min, the mixture was centrifuged at 10,000 rpm for 10 min at 4 °C, and the supernatant was collected. The supernatant was dried by evaporation using N2 at 37 °C, and the residues were re-dissolved in 350 mL of acetonitrile/water (1:2), vortex-mixed for 2 min, and centrifuged at 14,000 rpm for 10 min at 4 °C. The pretreatment methods for urine samples were similar to those used for serum samples. The fecal metabolite was extracted at a ratio of 1:3 (weight of fresh feces-to-methanol) in methanol (chromatography grade). The samples were homogenized by whirl mixing for 2 min and then centrifuged at 10000 rpm for 10 min at 4 °C. Serum, urine, and fecal samples were filtered through a 0.22 µm ultrafiltration membrane (Millipore) for further removal of contaminants. The supernatants were transferred to Eppendorf tubes and stored at -80 °C.
2.3.2 UPLC−QTOF-MS/MS Analysis
UPLC-Q-TOF-MS/MS analysis was performed using a Waters Acquity UPLC system coupled to a Waters Micromass Q-TOF micro™ mass spectrometer, with electrospray ionization (ESI) in the positive and negative ion modes. Chromatographic (2 µL) separation was performed on an Acquity UPLC™ BEH C18 column (100 mm × 2.17 mm, 1.7 μm, Waters, Milford, MA, USA), maintained at a temperature of 35 °C and a flow rate of 0.35 mL/min. Analytes were eluted using a gradient of mobile phase A—water-formic acid (100: 0.1, v/v) and mobile phase B—acetonitrile. The elution gradient was as follows: 2 % B for 0.5 min, 2−20 % B over 0.5−6.0 min, 20−35 % B over 6.0−7.0 min; 35−70 % B over 7.0−9.0 min; 70−98 % B over 9.0−10.5 min; 98 % B for 2.0 min, and back to 2 % B for 6.0 min.
For mass spectrometry analysis, a Waters Micromass Q-TOF 13 spectrometer was used, with a full scan over 50−1000 m/z in positive and negative ion modes (Waters, Manchester, U.K.), source temperature of 110 °C, and a cone gas flow of 15 L/h. A desolvation gas temperature of 320 °C and gas flow of 650 L/h were used. The capillary voltage was set at 3.0 kV in the ESI+ mode and 2.8 kV in the ESI− mode, and the cone voltage was 35 V. All analyses were performed using a locking spray to ensure accuracy and reproducibility. A lock mass of leucine enkephalin for the ESI+ ([M+H] + = 556.2771) and ESI− modes ([M+H]− = 554.2615) were used via a lock spray interface. The lock spray frequency was set at 10 s, and the lock mass data were averaged over ten scans for correction.
UPLC-Q-TOF-MS/MS analysis of urine and feces samples was performed in a manner similar to the serum analysis, with minor modifications.
2.3.3 Data processing
The raw UPLC-TOF-MS/MS data were analyzed using MarkerLynx Application Manager 4.1 (Waters Corporation, Milford, MA, USA). MarkerLynx ApexTrack peak integration was used for peak detection and alignment. The peak parameters (peak width at 5 % height and peak-to-peak baseline noise) were automatically calculated by the system. The collection parameters were set as described in our previous study .
The data matrix of samples was imported to SIMCA-P 14.0 software (Umetrics, Umeå, Sweden) for multivariate analysis. Unsupervised principal component analysis was used for all the samples to determine the general separation. Supervised multivariate analyses of orthogonal partial least square discriminant analysis (OPLS-DA) were applied to highlight the maximal different metabolites associated with zinc deficiency.
2.3.4 Identification of differential metabolites
To ensure data availability, only metabolites with less than 20 % missing values were used for the analysis. With respect to the identification of differential metabolites, variables were selected based on a threshold of VIP value (VIP > 1.0) and S-plot from the OPLS-DA model. Simultaneously, these differential metabolites from the OPLS-DA model were validated at the univariate level using Student’s t-test, with the critical P-value set to 0.05. Metabolites were identified based on the accurate mass, retention time, and matching MS spectra of the unknowns to the standard model compounds. The following databases were used to support metabolite identification: HMDB (http://www.hmdb.ca/) and METLIN (http://metlin-scripps.edu/). Finally, variables with fold change less than 0.8 and greater than 1.2 were included in the metabolic profile analysis to be considered as biologically significant.
2.3.5 Metabolite profiling
MetaboAnalyst (http://www.metaboanalyst.ca/) was used for the pathway analysis. Kyoto Encyclopedia of Genes and Genomes (https://www.kegg.jp/) and Small Molecule Pathway Database (http://www.smpdb.ca/) were used for functional analysis and mapping significantly altered metabolites in the metabolic pathway.
2.3.6 Data analysis
The analysis of food intake and weight development was expressed as β (95 % confidence interval). Differences between groups were analyzed using linear mixed models. Data were analyzed using R software (version 4.1.2).