The RNAseq data on IVV embryos we used are the data we have uploaded to NCBI previously (https://submit.ncbi.nlm.nih.gov/subs/sra/SUB10711406/PRJNA 783716), under accession IDs: MII_1 (SRR17041081), MII_2 (SRR17041080), MII_3 (SRR17041074), IVV_2C_1 (SRR17041073), IVV_2C_2 (SRR17041072), IVV_2C_3 (SRR17041071), IVV_4C_1 (SRR17041070), IVV_4C _2 (SRR17041069), IVV_4C_3 (SRR17041068), IVV_8C_1 (SRR17041067), IVV _8C_2 (SRR17041079), IVV_8C_3 (SRR17041078), IVV_Bla_1 (SRR17041077), IVV_Bla_2 (SRR17041076), and IVV_Bla_3 (SRR17041075). LncRNAs were isolated according to non-coding annotations of pigs in the ALDB database (http://202.200.112.245/aldb/); the specific data are shown in Supplementary Table S1. Differentially expressed genes (DEGs) were analyzed by IDEP, and heat maps were constructed. The pathway enrichment of DEGs was obtained through WebGestalt (http://www.webgestalt.org/option.php). The bubble plot of pathway enrichment was constructed by BioLadder (https://www.bioladder.cn/web/#/pro/cloud); the specific data are shown in Supplementary Table S2. The target genes were predicted by the ALDB database, the information on lncRNAs and coexpressed genes is shown in Supplementary Table S3, and the Venn diagram was constructed by FunRich version 3.1.3.

Antibodies used are as follows: H3K4me2 (Abcam; ab7766, diluted 1: 200), H3K4me3 (Abcam; ab8580, diluted 1: 200), H3K9me3 (Abcam; ab8898, diluted 1: 200), 5mC (Eurogentec; BI-MECY-0100, diluted 1:100), 5hmC (Active Motif; 39,769, diluted 1:100), DNMT1 (Invitrogen, MA5-16169, diluted 1:10), Alexa Fluor 488 goat anti-rabbit (Invitrogen, A-11008, diluted 1: 200), and/or Alexa Fluor 488/594 goat anti-mouse (Invitrogen, A32723/A-11020, diluted 1: 200) antibodies. All the chemicals used to culture were purchased from Sigma-Aldrich (St. Louis, MO, United States), unless otherwise stated.

The pig ovaries we used were taken from the same local slaughterhouse (Changchun Huazheng, Jilin, China); the pigs are the cross of Landrace boar and Large White sow. The ovaries were transported to the laboratory within 2 h. They were kept in 0.9% NaCl supplemented with 200 IU/ml penicillin and streptomycin at 35–36.5°C.

The follicular fluid containing cumulus–oocyte complexes (COCs) from 3–6 mm ovarian follicles was aspirated using an 18-gauge needle. COCs with at least three layers of cumulus cells were selected, washed three times in a manipulation fluid (TCM-199 supplemented with 0.1% polyvinyl alcohol), and then cultured in the media of IVM. Approximately, every 200 COCs were cultured in a 1 ml drop of maturation medium (TCM-199 supplemented with 10 μg/ml epidermal growth factor, 0.5 μg/ml porcine luteinizing hormone, 0.5 μg/ml porcine follicle-stimulating hormone, 26 mM sodium bicarbonate, 3.05 mM glucose, 0.91 mM sodium pyruvate, 0.57 mM cysteine, 0.1% PVA, 10% fetal calf serum, 75 mg/ml penicillin G, and 50 mg/ml streptomycin) for 22–24 h at 38.5°C, 5% CO₂, and 95% air. Then, they were transferred to a hormone-free maturation medium (the formula is consistent with the previous maturation medium without 10 μg/ml epidermal growth factor, 0.5 μg/ml porcine luteinizing hormone, and 0.5 μg/ml porcine follicle-stimulating hormone) for 20 h at 38.5°C, 5% CO₂, and 95% air. Then, cumulus cells were removed from oocytes with a manipulation fluid supplemented with 0.2% hyaluronidase. The oocytes with polar body 1 (PB1) were considered matured and used for the following experiments.

siRNA was injected using the microinjection meter (Eppendorf, FemtoJet 4i, United States). A measure of 5–10 pL of 25 nM lncT siRNA, SIN3A siRNA, or NC siRNA was injected into MⅡ oocytes, and then, the oocytes were incubated with siemens in PGM (porcine gamete medium: 100 ml water with 0.6313 g NaCl, 0.07456 g KCl, 0.00477 g KH₂PO₄, 0.00987 g MgSO₄·7H₂O, 0.2106 g NaHCO₃, 0.07707 g CaC₆H₁₀O₆·5H₂O, 0.0187 g D-Glucose, 0.3 g PVA, 0.00242 g cysteine, 0.04504 g C₇H₈N₄O₂, 0.0022 g C₃H₃NaO₃, and 100 μl/ml penicillin/streptomycin). The lncT siRNA, SIN3A siRNA, and NC siRNA were designed and synthesized by Sangon Biotech (Shanghai, China). The sequences were listed as follows:

lncT: sense: CCA​GAU​GAG​AUG​GUG​AUA​ATT,

Fresh semen was collected from the Jilin University pig farm. Density gradient centrifugation was used to wash them. In brief, percoll was configured in the concentration of 90% and 45%, 2 ml of semen was added and centrifuged at 300 g for 20 min, and the supernatant was removed after centrifuge. Then, 4 ml of DPBS was added and centrifuged at 300 g for 10 min. The sperm were resuspended with PGM. Sixty denuded matured oocytes were each cultured in 400 μl PGM with a final sperm concentration of 1.6 × 10⁵—5.0×10⁵ sperm/ml, at 38.5°C and 5% CO₂ for 5–6 h. After washing off the adherent sperm, the fertilized oocytes were transferred to PZM-3 (Yoshioka et al., 2002).

The smart-seq2 method was used to amplify each sample (6–8 embryos for each group), according to the manufacturer’s instructions. The RNA concentration of library was measured by using a Qubit 2.0 Fluorometer (Life Technologies, CA, United States). The Agilent Bioanalyzer 2100 system was used to assess the insert size, and the quality of the amplified products was evaluated according to the detection results. The amplified product cDNA was used as the input for the library construction of transcriptomes. After the library construction, the Agilent Bioanalyzer 2100 system was used to assess the insertion size, and the TaqMan fluorescence probe of an AB Step One Plus Real-Time PCR system (library valid concentration >10 nM) was used to quantify the accurate insertion size. Clustering of the index-coded samples was performed using a cBot cluster generation system and the HiSeq PE Cluster Kit v4-cBot-HS (Illumina). Then, the libraries were sequenced by Zhejiang Annoroad Biotechnology (Beijing, China) on an Illumina platform, and 150-bp paired-end reads were generated. STAR was used to compare transcriptome data and Cufflinks for quantitative splicing. DEmRNAs were identified by the IDEP website (http://bioinformatics.sdstate.edu/idep/). Genes with false discovery rates (FDRs) ≤ 0.05, |log2FC| ≥ 1.5, and p-value ≤ 0.01 were selected as candidate genes, and the specific data are shown in Supplementary Table S4.

Total RNA was extracted, and complementary DNA (cDNA) was synthesized with the SuperScript™ IV CellsDirect™ cDNA Synthesis kit (11750350, Invitrogen, United States), following the manufacturer’s instructions. The qPCR was performed with FastStart Essential DNA Green Master (06924204001, Roche, United States) via a StepOnePlus Real-Time PCR system. Primers sequences were listed as follows:

lncT: forward-5′-GGTCACTTGGCAGAGATGCT-3′

The results were analyzed by the 2^−ΔΔCT method. The qPCRs were all repeated three times.

A total of 5–8 embryos of IVF were washed with PBS containing 0.1% polyvinylpyrrolidone (PVP). The zona pellucida of embryos was dissolved in an acidic Tyrode solution (pH 2.5). After washing in PBS-PVP, embryos were fixed with 4% paraformaldehyde for 30 min in the dark. After being washed in PBS-PVP, embryos were permeabilized with 0.2% Triton X-100/PBS (v/v) for 20 min and then blocked with 2% BSA/PBS for 1 h. For 5mC/5hmC staining, the embryos were treated with 4N-HCl and Tris-HCl for 30 min each before BSA blocking. Embryos were incubated with primary antibodies at 4°C overnight. After being washed in PBS-PVP, embryos were stained with a secondary antibody at 37°C for 2 h in dark. Then, DNA was stained with 10 μg/ml DAPI for 15 min. All samples (DNMT1 excepted) were observed under a Nikon Eclipse Ti-U microscope equipped with appropriate filters (Nikon, Tokyo, Japan) after mounting. Color images were captured using a DS-Ri2 CCD camera (Nikon, Tokyo, Japan) and an analysis software application (NIS-Elements BR; Nikon, Tokyo, Japan). The same exposure times and microscope settings were used for all captured images. Evaluation of the fluorescence intensity of individual images was performed by ImageJ software (National Institutes of Health, Bethesda, MD). The cytoplasmic background fluorescence intensity was measured as an average intensity level within the cytoplasmic area. Thereafter, the correction for the cytoplasmic background was carried out, and the background-subtracted images were used for further analysis (Zaitseva et al., 2007). The DNMT1 immunofluorescence staining images were obtained by confocal microscopy (ZEISS, Examiner, Z1/LSM880). At least five embryos at each development stage were analyzed.

After the removal of the zona pellucida, fixation by 4% paraformaldehyde, permeabilization by 1% Triton-X 100, and blocked by 1% BSA, the blastocysts were incubated with a term deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) solution from the In Situ Cell Death Detection Kit (Roche, Mannheim, Germany) at 37°C for 1 h in darkness. The blastocysts were washed with PBS-PVP three times and stained with DAPI (10 μg/ml) for 15 min. The stained blastocysts were mounted between a cover slip and a glass slide and observed under a fluorescence microscope (Nikon, Tokyo, Japan), with 5–8 blastocysts per group.

Data are presented as the means ± SEMs. The experiments were repeated at least two times in triplicate. Statistical analysis was performed by GraphPad Prism 6.01 (GraphPad Software, United States). A chi-squared test was performed to analyze the ratio of the developmental capacity and efficiency of the control and SIN3A KD embryo data between the two groups. *p < 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significant.

To explore the differentially expressed lncRNAs in IVV embryos, a heatmap was structured. As shown in Figure 1A, there were 2,000 different expressed lncRNAs, and 699 lncRNAs presented decreased expression when embryos developed to the 4-cell stage, 361 lncRNAs were upregulated in the blastocyst stage, 275 lncRNAs were upregulated in 4-cell and 8-cell stages, and 665 lncRNAs were increased in 8-cell and blastocyst stages (FDR<0.05 fold change>2). The expression pattern of lncRNA was consistent with the time of gene activation, and we speculated that this part of lncRNA played a role in regulating gene expression. The average number of exons of the lncRNAs was 2.5 (Figure 1B), and the average length of the lncRNAs was 310bp (Figure 1C), which was consistent with the characteristics of lncRNAs. To study the functions of the DElncRNAs, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was carried out. As shown in Figure 1D, genes in cluster A were mainly enriched in fatty acid biosynthesis, base excision repair, and cell adhesion molecule pathways. Cluster B was enriched in the pentose phosphate pathway and oxidative phosphorylation carbon metabolism. Genes in cluster C were enriched in nitrogen metabolism, glycosphingolipid biosynthesis, and RNA degradation pathways. Also, cluster D was enriched in ribosome, oxidative phosphorylation, fatty acid biosynthesis, etc. To select the candidate lncRNA for regulating embryonic development, we predicted the targeting genes of these lncRNAs, which are upregulated from the 4-cell stage, and coexpression was analyzed with the mRNAs with the same characteristics. As shown in Figures 1E,F, a total of 37 lncRNAs were selected. These results suggested that the highly expressed lncRNAs have the potential role in embryonic development, and the 37 lncRNAs were the candidate genes.

Since lncT (ALDBSSCT0000005403 in Figure 1F) has a single product when we use PCR to verify whether lncT exists in the pig genome, we chose lncT to conduct experiments to verify its role in embryos (Supplementary Figure S1). The total length of lncT is 2038bp, and the whole sequence is shown in Supplementary Table S5. It contains a polyA tail (Supplementary Figure S1). We detected its expression in IVF embryos, as shown in Figure 2A, and lncT was upregulated in 4-cell and 8-cell stages. Meanwhile, to explore the expression characteristics of lncT, we detected its expression in other organs, as shown in Figure 2B, and it was highly expressed in two germ organs, the testis and ovary. To investigate the functions of lncT, we knocked down the expression of lncT in embryos of IVF via the injection of small interfering RNAs (siRNAs) against it. As shown in Figure 2C, we injected siRNA or NC into MⅡ oocytes, and then investigated its effect on embryonic development. As shown in Figure 2D and Table 1, no difference was observed in the percentage of the blastocyst in the injected NC group (20.4 ± 0.95%) compared to the control group (22.92 ± 1.97%), suggesting that there was no significant effect on the NC group; thus, we will not inject NC in future experiments. However, the percentage of blastocysts in the lncT-KD group (13.65 ± 0.82%) was significantly decreased when compared to the control group (**p < 0.01). As expected, lncT was decreased in lncT-KD embryos compared to NC or the control group (***p < 0.001) (Figure 2E). Meanwhile, we also observed that the embryonic development difference was significant from the 8-cell stage (**p < 0.01) (Figure 2F). These data indicated that the knockdown of lncT inhibited porcine embryonic development.

In order to study the mechanism of lncT in embryonic development, we collected embryos at the 4-8-cell stage, according to the arrest period of cleavage for transcriptome sequencing. As shown in Figure 3A, there were 636 downregulated genes and 1,564 upregulated genes after siRNA injection (FDR<0.05, logFC>2). The heatmap showed the transcriptome profile of the control group and lncT-KD group embryos (Figure 3B). Cluster A was the downregulated gene corresponding with the functions of regulation of cellular components, actin filament organization, and anatomical structure size (Figure 3C), while cluster B and -C were the upregulated genes corresponding with the functions of translation, the amide biosynthetic process, cellular nitrogen compound metabolic process, gene expression, and cellular macromolecule biosynthetic process (Figure 3C). These data revealed that the adverse effects in embryo development after siRNA injection may be due to abnormal gene expression.

Next, we focused on the epigenetic modification genes in the transcriptome, as shown in Figure 4A, and the heatmap showed the differentially expressed genes between the two groups. As shown in Figure 4B, DNMT3B, PRMT1, HDAC3, and CRAM1 were significantly downregulated (*p < 0.05), while KDM5C, ESC O 1, SETD2, and KDM5B were significantly upregulated (*p < 0.05) in lncT-KD embryos. Thus, we detected the expression of H3K4me3, H3K4me2, H3K9me3, and 5mC/5hmC during embryonic development. As shown in Figure 4C, the expression level of H3K4me3 was downregulated when lncT siRNA injection was injected in 2-cell and 4-cell stages (*p < 0.05). Also, H3K4me2 had a similar expression trend after lncT siRNA injection (***p < 0.001) (Figure 4D). The expression of H3K9me3 had no significant difference between the control and lncT-KD group in 2-cell, 8-cell, and blastocyst stages, but it was downregulated in the 4-cell stage (**p < 0.01) (Figure 5A). When we detected the expression of 5mC/5hmC, we observed that the 5mC modification was slowly lost during embryonic development, but it was always present in the lncT-KD embryos (*p < 0.05) (Figure 5B). Taken together, these results revealed that the knockdown of lncT affects epigenetic modification, and these effects were at the 4-cell stage.

To further confirm the key genes that lncT regulated in embryonic development, we evaluated the top changes in genes between the two groups. As shown in Figure 6A, the heatmap showed the top 10 downregulated genes and 10 upregulated genes. Also, SIN3A was observed as the significantly downregulated gene after lncT siRNA injection (*p < 0.05) (Figure 6B). Thus, we speculated that lncT regulated early embryonic development partly via modulation of SIN3A expression in pigs. We selected a siRNA to interfere with the expression of SIN3A, as shown in Figure 6C, and SIN3A was significantly decreased after S1 injection (***p < 0.001). Next, we tested if the embryos had a phenotype as lncT depletion by knocking down SIN3A. As shown in Figures 6D,E, blastocyst formation was severely affected in embryos due to deficiency of SIN3A (**p < 0.01). Meanwhile, we detected the expression of 5mC/5hmC, as shown in Figure 7A, and the modification of 5mC in the 4-cell stage was increased after knockdown of SIN3A (*p < 0.05); although there was no significant difference in 5mC at the blastocyst stage, the modification of 5hmC was decreased in the SIN3A-KD blastocyst (*p < 0.05). The previous study reported the KD of SIN3A-induced DNMT1 in the nucleus (Zhao et al., 2019). Thus, we detected the expression of DNMT1; as expected, DNMT1 was mainly distributed in the nucleus in the SIN3A-KD group compared to the control group. Meanwhile, it was also mainly distributed in the nucleus in the lncT-KD group (Figure 7B). In addition, the pluripotent genes were reduced after the injection of siRNA of lncT (**p < 0.01) and SIN3A (***p < 0.001) (Figure 7C). Also, increased apoptosis was found in all interference groups (Figure 7D). These results indicated lncT and SIN3A regulated global 5mC modification and DNMT1 protein abundance and localization, and lncT regulated porcine early embryonic development partly through the regulation of SIN3A expression.