All experiments were approved by the Animal Care and Use Committee of Nanjing Agriculture University and were performed in accordance with Animal Research Institute Committee guidelines.

Porcine ovaries were collected from prepubertal gilts at a local slaughterhouse and transported to the laboratory within 2h in prewarmed physiologic saline supplemented with 800 IU/mL gentamicin. The contents of antral follicles (3–5 mm in diameter) were aspirated with 20-gauge needles. Only fully grown oocytes with an evenly granulated cytoplasm, surrounded by at least three uniform layers of compact cumulus cells, were selected for experiments. The pooled column oocyte complexes (COCs) were washed three times with maturation medium, which was a TCM199 based medium supplemented with 10% porcine follicular fluid (PFF), 10 ng/ml of epidermal growth factor (EGF), 10 IU/ml of PMSG, 10 IU/ml of HCG, 0.57 mM cysteine, 0.91 mM sodium pyruvate, 3.05 mM D-glucose, 75 mg/ml of penicillin and 50 mg/ml of streptomycin. A group of COCs were transferred to 4-well plates containing 500 μL of maturation medium covered with 200 μL mineral oil, and incubated at 38.5°C under 5% CO₂ in 95% humidified air for in vitro maturation (IVM).

After a period of incubation, COCs were removed in medium containing 0.02% (w/v) hyaluronidase at 38.5°C for 5 min. The denuded oocytes were isolated from COCs by repeatedly pipetting and then collected for subsequent experiment after 3~4 rinses.

Respectively in vitro mature for 0h, 22h, 46h, denuded oocytes isolated from COCs after 3~4 rinses by Dulbecco’s Phosphate Buffered Saline (DPBS). Samples (400 oocytes per sample, 5 replicates for each stage) were transferred to the bottom of Eppendorf tubes, rapidly frozen by liquid nitrogen and stored at −80°C. In order to extract metabolite, samples were thoroughly mixed with 200 μL of methanol/water (80/20, vol/vol) and vortexmixed, followed by homogenized using a mechanical homogenizer. After incubation on ice for 10 min, samples were centrifugated at 16,000g for 15 min at 4°C. Supernatant was transferred to concentrated and desiccated, prior to storage at -80°C until instrumental analysis.

Mass spectrometry analysis was performed using the UltiMate3000 high -performance liquid chromatography system coupled with the Q-Exactive mass spectrometer. Subsequent chromatographic separation was performed on a Hypersil GOLD C18 column (1.9 μm, 100 mm ×2.1 mm), with temperature at 40°C. The binary solvent system used was solvent A of acetonitrile containing 0.1% formic acid, and solvent B of water containing 0.1% formic acid, which was at a flow rate of 0.4 mL/min. A multistep gradient was used for the metabolites analysis. Briefly, the column mobile phase was held 1% mobile phase A for 3 min (t=3 min), followed by an increase to 99% in 7 min lasting for 5 min (t=15 min), and finally immediately reduced to 1% and held for 2 min (t=17 min).

MS data were collected by the high-resolution mass spectrometer equipped with an electrospray ionization source. The instrument was operated in both positive and negative electrospray ionization (ESI) modes in a mass range 70-1050 m/z with a resolution of 70,000 for MS detection. The metabolites were identified by the comparison of accurate mass and retention time with commercial standard compounds using the author-constructed library. All samples were analyzed in a randomized fashion to avert complications of the injection order. ‘‘R’’ (V2.15) was used to perform all statistical analysis. The Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) was conducted by SIMCA-P software (V14.0; Umetrics AB, Umea, Sweden) to observe Intra-group repeatability and between-group difference. T test (p < 0.05) coupled with a variable importance in projection (VIP) analysis (VIP >1.00) were considered statistically significant. KEGG Mapper (V4.1) (https://www.genome.jp/kegg/) was used to identified metabolic pathway information.

COCs that were collected at 0h, 22h, 46h of IVM were stripped of cumulus cells in the same manner. The single-cell samples (20 oocytes per sample, 4 replicates for each stage) were removed in tubes with lysis buffer containing protease inhibitor and ribonuclease inhibitor. Then we carried out the amplification by the SMART-Seq2 method.

For cDNA synthesis, an Oligo-dT primer was used to the reverse transcription reaction, followed by PCR amplification to enrich the cDNA and magbeads purification step to clean up the production. Primer sequences used include: oligo-dTV: 5’-AAGCAGTGGTATCAACGCAGAGTACTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTVN-3’; Template Switching Oligo (TSO): 5’-AAGCAGTGGTATCAACGCAGAGTACATrGrG+G-3’; ISPCR: 5’-AAGCAGTGGTATCAACGCAGAGT-3’. Then the cDNA concentration was preliminarily measured using Qubit^® 3.0 Flurometer and Agilent 2100 Bioanalyzer to ensure the expected production with length around 1~2kbp. For library preparation, the pooled and purified cDNA was fragmented by sonication and then converted to sequencing libraries according to the standard Illumina library preparation protocol, including DNA fragmentation, end repair, 3’ ends A-tailing, adapter ligation, PCR amplification and library validation. PCR primers included P5 PCR Primer (5’TGATACGGCGAOCACCGAG) and P7 PCR Primer (5’AAGCAGAAGACGGCATACGAG). Library preparation integrity was verified with PerkinElmer LabChip^® GX Touch and Step OnePlus™ Real-Time PCR System. After libraries were qualified, sequencing libraries were performed on the Illumina Hiseq platform for PE150 sequencing.

Raw reads were first subjected to quality control with FastQC (v0.11.9). Based on the evaluation of the raw data, Trim galore (v0.6.4) was then used to trim sequencing adapters and filter out low-quality reads. The clean data were next aligned to the swine reference genome (Sscrofa11) by STAR (v2.7.6a) (12), followed by gene expression quantification by HTSeq (v1.99.2) (13) for counts and RSEM (v1.3.3) (14) for FPKM values. Differential expression analysis was performed with count data using the DESeq2 (v1.30.1) (15) R package, and the genes with absolute fold change > 1.5 and adjusted P value < 0.05 were treated as differentially expressed genes (DEGs). Enrichment analysis of DEGs for Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways was conducted by DAVID (v2022q1) (16), and significantly enriched GO terms and KEGG pathways were satisfied with p-value < 0.05. FPKM values were used to generate heatmaps in R (v4.0.4). For network reconstruction between genes and metabolites, we only considered DEGs and differentially expressed metabolites (DEMs) for the network reconstruction. In brief, we took the intersection of top 300 DEGs between any paired time points, and, similarly, the intersection of DEMs for correlation analysis. We kept the connection between genes and metabolites if the Pearson’s correlation coefficients were greater than 0.95. Finally, the network between genes and metabolites based on correlation analysis was visualized with Cytoscape (v3.7.1).

The profiling of statistics was performed by the software GraphPad Prism (V7.0) for Windows. Student’s t test was used to identify the differential metabolites and genes, and data are presented as means ± SD (standard deviation), unless otherwise stated. A p value less than 0.05 was considered statistically significant.

Although existing studies have identified several cellular pathways, we believe that the integrated profiling of metabolomics and transcriptomics contributes to a comprehensive insight of the metabolic dynamics as the oocytes progress through meiosis. Our previous study (17) has analyzed statistically the time-dependent nuclear maturation proportion of porcine oocyte during in vitro culture. About 96.3%, 89.2% and 73.2% of oocytes were at GV (0h), Pre-MI (24h), and MII (44h) stages, respectively. In the present study, we collected a great number of oocytes (totally 8,000 oocytes) isolated from cumulus-oocyte complexes (COCs) at three key time points (0h, 22h, 46h, Figure 1A ). They were subjected to identification of intracellular metabolome using ultra-high-performance liquid chromatography-tandem high-resolution mass spectrometry (UHPLC-HRMS) ( Figure 1B ). A total of 441 metabolite components were detected, among which 197 differential metabolites were determined by the combination of a t test (p <0.05) and a variable importance in projection (VIP) analysis (VIP >1.00) ( Supplementary Table 1 ). Scatted plots of orthogonal partial least-squares-discriminant analysis (OPLS-DA) clearly distinguished three groups, suggesting stage-dependent separation ( Figures 1C–E ). A heatmap was utilized to analyze and visualize the changes in metabolite level during meiotic maturation ( Figure 1F ). In order to extract significantly altered metabolic pathways, we mapped differentially expressed metabolites to Kyoto Encyclopedia of Genes and Genomes (KEGG), displaying the distinct metabolic patterns upon oocyte maturation. Generally, the levels of most amino acids and carbohydrate increased during meiotic maturation, whereas a notable reduction was observed in both lipid metabolites and nucleotides in the transition from GV to MII stage ( Figures 1F–H ).

Gene transcription plays a crucial role in predicting and evaluating the dynamics of cellular processes during oocyte development. To explore the expression of the enzymes related to metabolism, we simultaneously conducted transcriptional profiling of oocytes at three stages mentioned above ( Figure 2A ). Principal Component Analysis (PCA) clearly showed stage-dependent separation of the three oocyte groups ( Figure 2B ). Changes in gene transcript abundance were assessed by differential expression analysis, and of the 31,908 genes identified, 2,686 exhibited significant alterations, based on p<0.05 couple with |log2(foldchange)| >0.59 ( Figures 2C, D and Supplementary Table 2 ). Gene Ontology (GO) and KEGG pathway enrichment analyses were conducted for differential expression genes, exhibiting a significant enrichment for numerous gene categories critical for metabolic pathways ( Supplementary Figures 1A, B ). Differentially expressed genes were mapped to KEGG metabolic pathways, and 9 significantly altered pathways were uncovered, involving metabolism of bile acid, tryptophan, purine and pyrimidine, and etc. ( Supplementary Figures 1C–I ). Collectively, by integrating metabolomics and transcriptomics, we have provided a broad picture of global metabolic characteristics and dynamic changes in different pathways during porcine oocyte maturation.

The dynamic characteristics of lipid metabolism during porcine oocyte maturation were poorly defined. As shown in Figure 1F , temporal metabolome profiles clearly displayed the alternations in level of lipid metabolites as the oocytes progress through meiosis, with enrichment primarily in fatty acid beta-oxidation ( Figure 3 ) and metabolism of bile acid ( Supplementary Figure 2 ) and steroid hormones ( Supplementary Figure 3 ). The content of lipid metabolites involved in fatty acid beta-oxidation (i.e., traumatic acid, tetradecanedioic acid, carnitine, Figures 3A–C ) declined dramatically during meiotic resumption, whereas the abundance of bile acid and steroid hormones metabolism-related products (i.e., cholic acid, glycocholic acid, ursodeoxychoic acid, testosterone, and progesterone, Supplementary Figures 2 and 3 ) displayed a notable reduction in MII oocytes compared to GV oocytes. Such the dynamic change of lipid-related products might play a vital role in nutrient adjustment, hormone regulation, and cellular homeostasis (18).

Amino acids play important roles in the metabolism of all organisms, including protein synthesis, energy production (26), organic osmolytes (27), and intracellular buffer (28). Nevertheless, very little is known about metabolic dynamics of amino acids during porcine oocyte maturation. A combined metabolome and transcriptome analysis of porcine oocyte reveals the diverse characteristics of amino acid metabolism during maturation ( Figures 4 – 6 ; Supplementary Figure 4 ), reflecting the sophisticated and fine-tuned regulation in oocytes. Four metabolic pathways significantly changed in meiotic oocytes were presented below.

In mammals, a close relationship between carbohydrate metabolism and maturation process in oocyte has been proposed (40). However, the metabolite dynamics of carbohydrate during oocyte maturation remain poorly studied. Glucose, utilized through both pentose phosphate pathway (PPP) and tricarboxylic acid (TCA) cycle, is essential for oogenesis in pigs (41, 42). Our metabolomic analysis demonstrated that several key components of glycolysis and PPP experienced dramatic alterations upon meiotic maturation. For instance, an ~15-fold-increase was observed for glucose 6-phosphate and fructose 1,6-bisphosphate, an ~40-fold-increase in sorbitol 6-phosphate, and particularly ~800-fold increase in gluconolactone were identified in MII oocytes compared to GV oocytes ( Figures 7A–E ). Meanwhile, transcriptome analysis identified the increased mRNA levels of 5 (i.e., IDNK, PGD, RPIA, PRPS and ADPGK) out of 7 enzymes related to the metabolic pathways mentioned above ( Figures 7F–M ). In addition, as a metabolism hub, TCA cycle serves to connect the processes of glycolysis, amino acid synthesis and other biosynthetic pathways (43). The level of malate was decreased in maturing oocytes ( Supplementary Figure 5A ). Nonetheless, the abundance of citrate underwent ~3-fold-increase during meiotic resumption ( Supplementary Figure 5B ). Consistent with this observation, the abundance of differential expression genes (i.e., CS, MDH2, SDHB, SUCLG2 and ACO1) involved in TCA cycle showed the significant upregulation during meiotic maturation, implying the enhancement of TCA activity ( Figures 7N–R and Supplementary Figures 5C–E ). Carbohydrate metabolism occupies the core of bio-molecular metabolism, and the substrates involved in carbohydrate metabolism provide the essential saccharides and energy for oocyte maturation. Cumulatively, our results indicate the active carbohydrate metabolism, specifically, the PPP and TCA cycle, during porcine oocyte maturation.

Purine and pyrimidine nucleotides play critical roles as precursors for the synthesis of DNA and RNA, major energy carriers, and core elements of cofactors in metabolic pathways (44). In the de novo synthesis of purine nucleotides, activated ribose-5-phosphate (R5P) is then converted to Inosinic acid (IMP), which is further transformed into AMP and GMP. Similarly, the de novo synthesis of pyrimidine ends with the production of UMP, from which other pyrimidine nucleotides are converted. Temporal metabolome profiles clearly revealed that out of 17 differential metabolites involved in nucleotide metabolism, the abundance of 15 experienced the significantly decrease during meiotic maturation ( Figures 8A–L and Supplementary Figures 6A–E ). Consistent with the metabolite changes, the majority of enzymes responsible for purine and pyrimidine metabolism presented the reduced mRNA levels in matured oocytes ( Figures 8M –S and Supplementary Figures 6F–M ). For instance, transcriptomic data showed a 95% decrease in GMP synthase (GMPS), an enzyme in guanine ribonucleotide biosynthesis, and cytosolic 5’-nucleotidase 3A (NT5C3A), an enzyme in purine salvage pathway, in matured oocytes compared to GV oocytes. Together, the results clearly reveal the progressive reduction in nucleotide metabolism during porcine oocyte maturation.