CARM1 inhibitor (Millipore, 217531) was dissolved in DMSO (Sigma, D2650) and stored at −20°C. Embryo culture medium was used to dilute the stock solution to obtain the desired working solution. The same volume of DMSO was added into the medium as a control when the chemical was used.

Ovaries were collected from a local slaughterhouse. Follicular fluid was aspirated from antral follicles at 3–6 mm in diameter. Cumulus-oocyte complexes (COCs) were selected under a stereomicroscope. Subsequently, COCs were cultured in one well of 4-well plate containing 400 μL in vitro maturation medium for 44 h at 38.5°C, 5% CO₂ and saturated humidity. Cumulus cells surrounding oocytes was removed using 1 mg/mL hyaluronidase following maturation.

Oocytes at metaphase II stage were stimulated using two pulses of direct current (1.56 kV/cm for 80 ms) in activation medium. Subsequently, the activated oocytes were washed three times in the porcine zygote medium (PZM-3) medium and were incubated in the chemically assisted activation medium for 4 h. Then, embryos were cultured in PZM-3 droplets at 38.5°C, 5% CO₂ and 95% air with saturated humidity.

Oocytes at the metaphase II were washed in the modified Tris-buffered medium (mTBM) containing 2 mg/mL BSA and 2 mM caffeine. The oocytes were incubated in mTBM for 4 h at 38.5°C in 5% CO₂ in air. Semen of two boars was mixed and centrifuged at 1,900 g for 4 min in DPBS supplemented with 1 mg/mL BSA. Following the removal of supernatant, the sperm concentration was adjusted with mTBM to 1 × 10⁶ sperms/mL. The semen was then added to the mTBM droplets containing oocytes. After co-incubation of oocyte and sperm for 6 h, sperms surrounding oocytes were washed out and presumptive zygotes were cultured in PZM-3 at 38.5 C in 5% CO₂ in air.

CARM1-mCherry mRNA used for microinjection was synthesized in vitro. pIVT- CARM1-mCherry plasmid containing T7 promoter was linearized by digestion with BspQI. Linearized DNA templates were purified using a DNA clean and concentrator Kit (ZYMO RESEARCH, D4003, Tustin, CA, United States). According to the manufacture’s manual, in vitro transcription of CARM1-mCherry mRNA was performed through using mMESSAGE MACHINE TM T7 kit (Ambion, AM1344, shanghai, China) and Poly (A) Tailing Kit (Ambion, AM1350, Shanghai, China). Then, mRNA was treated with TURBO DNase to remove the DNA templates and was further purified using MEGAclear Kit (Ambion, AM1908, Shanghai, China). After mRNA was dissolved in RNase-free water, mRNA concentration was determined by a NanoDrop instrument (Thermo Fisher Scientific, Shanghai, China) and was then aliquoted and stored at −80°C.

Total RNA was extracted from oocytes and embryos using RNeasy Mini Kit (Qiagen, 74104). RNA was transcriptionally reversed into cDNA using QuantiTect Reverse Transcription Kit (Qiagen, 205311). The primers used in this study were listed in Supplementary Table 1. The assembly of polymerase chain reaction was prepared in FastStart SYBR Green Master (Roche, 04673514001) and was run on StepOne Plus (Applied Biosystems, Foster, United States). The samples were collected three times and three biological replicates were conducted for each gene. EF1A1 was used as the internal reference gene. The Cq values were obtained and analyzed using the 2^–ΔΔCt method.

Embryos were fixed in 4% paraformaldehyde solution for 15 min, permeabilized with 1% Triton X-100 for 30 min at room temperature (RT) and then blocked with 2% BSA at RT for 1 h. Samples were incubated in solution containing primary antibodies overnight at 4°C. Following washing, samples were incubated for 1 h in solution containing secondary antibodies in the dark at 37°C. Following washing, samples were counterstained using 4, 6-diamidino-2-phenylindole dihydrochloride or propidium iodide for 10 min and were then loaded onto glass slides. Finally, samples were imaged using laser scanning confocal microscopy (Olympus, Japan). Information regarding primary and secondary antibodies used was provided in Supplementary Table 2.

siRNA species was designed to target three different sites of the porcine CARM1 coding region (GenePharma, Shanghai, China). Microinjection was performed in a T2 (TCM199 with 2% FBS) medium containing 7.5 μg/mL Cytochalasin B on an inverted microscope (Olympus, Japan). Approximately 10 pl of siRNA solution (50 μM) was microinjected into the cytoplasm of MII oocytes. Embryos were cultured in PZM-3 medium for 7 days.

Single embryo at day 5 (non-injected and CARM1-siRNA injected embryos) was collected for RNA-seq analysis. RNA was extracted using the RNeasy Mini Kit (Qiagen, 74104). Pre-amplified cDNA was fragmented using Fragmentase (NEB, M0348S) via the incubation at 37°C for 20 min. cDNA libraries were constructed by TruSeq Nano DNA LT Library Preparation Kit (FC-121-4001). Then, the libraries were sequenced on the Illumina HiSeq 2500 instrument (LC-Sciences, Hangzhou, China). The reads were mapped to the pig reference genome. Differential gene expression between non-injectedand CARM1 siRNA injected embryos was determined using Cufflinks. The threshold for significance was a false discovery rate ≤0.05 and a fold expression change ≥2. GO analysis was performed using DAVID Bioinformatics Resources 6.8. RNA-seq data are presented in Supplementary Tables 3–5.

Statistical analyses were performed using one-way ANOVA or Student’s t-test (SPSS 17.0). All experiments were carried out at least three times and were presented as mean ± standard error of mean (mean ± S.E.M). P < 0.05 was considered to be statistically significant.

To quantify the relative abundance of CARM1 mRNA in early embryos, qPCR was performed to determine the expression of CARM1 mRNA. The results revealed that CARM1 is expressed in oocytes and embryos, but the expression levels of CARM1 mRNA are significantly higher in embryos at the 4-cell and 8-cell stages compared to oocytes, morulae, and blastocysts (Figure 1A) (P < 0.05). Furthermore, immunofluorescence staining was performed to determine the localization of CARM1’s putative histone substrates including H3R2me2, H3R17me2, and H3R26me2 in porcine embryos. The specificity of the arginine dimethylation antibodies in porcine blastocysts had been verified before undergoing immunostaining experiment (Supplementary Figure 1A). Immunofluorescence analysis showed that the dimethylation modifications of three histone arginine residues simultaneously localize at the cytoplasm and nuclei of both GV and MII oocytes whereas they are present in nuclei throughout early embryonic development (Figures 1B–D). Additionally, the dimethylation modifications of three histone arginine residues are also present in both the ICM and TE cells of blastocysts (Figures 1B–D). Therefore, these results indicate that CARM1 mRNA and its putative dimethylation modifications of histone arginine residues are dynamically present in early embryonic development.

To determine the specific histone arginine substrates of CARM1 in porcine embryos, a chemical inhibitor with different concentrations (9, 18, and 27 μM) was used to treat embryos. The developmental results revealed that treatment of 9 μM CARM1 inhibitor did not affect the developmental rates of early embryos (Figures 2A,B and Supplementary Figures 2A,B). Treatment of 18 μM CARM1 inhibitor partially hindered the development of 2-cell to morula whereas it significantly reduced the blastocyst rates (Figures 2A,B and Supplementary Figures 2A,B) (P < 0.05). Treatment of 27 μM CARM1 inhibitor significantly blocked the development of 2-cell to blastocyst stages (Figures 2A,B and Supplementary Figures 2A,B) (P < 0.05). Considering the developmental phenotypes of embryos exposed to different concentrations of CARM1 inhibitor, the dosage of 27 μM was used in the subsequent experiments. To further determine the levels of H3R2me2, H3R17me2, and H3R26me2, immunofluorescence staining was performed in embryos at the 2-cell, 4-cell, and blastocyst stages. The results showed that CARM1 inhibition did not affect H3R2me2 and H3R17me2 levels, but significantly reduced H3R26me2 levels in embryos at the indicated stages (Figures 2C–E and Supplementary Figures 2C,D) (P < 0.05). Therefore, these data demonstrate that CARM1 activity is required for blastocyst formation and the generation of H3R26me2 modification in porcine embryos.

To further confirm whether CARM1 protein specifically recognizes the H3R26 residue in porcine embryos, RNAi approach was used to delete the CARM1 protein. MII oocytes were microinjected with CARM1 siRNA or negative control (NC) siRNA. Uninjected oocytes served as control groups. A subset of embryos at the 4-cell and 8-cell stage was subject to qPCR to detect the expression levels of CARM1 mRNA. The results revealed that siRNA injection significantly reduced the levels of CARM1 mRNA in embryos at the 4-cell (Figure 3A) and 8-cell (Figure 3B) stage compared to the control groups (P < 0.05). No differences in expression levels of CARM1 mRNA were observed between NC group and uninjected group (Figures 3A,B). There are not available for porcine specific CARM1 antibodies so that we could not directly evaluate the levels of CARM1 protein in embryos. Instead, we constructed the fusion expression vector of CARM1-mCherry and synthesized in vitro the fusion mRNA. Both CARM1-mCherry mRNA and siRNA was injected into oocytes and served as the experimental group, uninjected, mCherry mRNA or CARM1-mCherrymRNA injection served as the control groups. Fluorescence intensity analysis showed that CARM1 siRNA injection significantly reduced the levels of CARM1-mCherry protein at the 4-cell stage compared to the control groups (Figure 3C) (P < 0.05). Next, immunofluorescence staining was performed to evaluate whether CARM1 KD affected the levels of H3R2me2, H3R17me2, and H3R26me2 in embryos. As shown in Supplementary Figures 3A,B, CARM1 KD did not alter the levels of both H3R2me2 and H3R17me2 in embryos at the 2-cell, 4-cell, and blastocyst stage. However, CARM1 KD significantly decreased H3R26me2 levels in embryos at the 2-cell (Figure 3D), 4-cell (Figure 3E), and blastocyst (Figure 3F) stage compared to the control groups (P < 0.05). Collectively, these results document that CARM1 protein specifically catalyzes H3R26me2 in porcine early embryos.

To elucidate whether CARM1 KD affected embryo development, the developmental rates of PA embryos at each stage were subject to statistical analysis. We found that CARM1 KD had no effect on embryo development to 2-cell, 4-cell, 8-cell, and morula stage (Supplementary Figures 3C,D), but reduced the blastocyst rates (Days 5–7) compared to the control groups (Figures 4A,B) (P < 0.05). Additionally, we did not observe differences in the rates of embryo development between NC siRNA injected and uninjected control groups (Figures 4A,B and Supplementary Figures 3C,D). To confirm the specificity of CARM1 KD, rescue experiments were performed by coinjection of CARM1-mCherry mRNA and CARM1 siRNA in oocytes. The results revealed that coinjection of CARM1-mCherry mRNA and CARM1 siRNA could restore blastocyst formation of CARM1 KD embryos (Figures 4C,D), demonstrating a specific role of CARM1 in early embryonic development. To determine whether CARM1 KD impaired lineage allocation in blastocysts, embryos were stained with a CDX2 antibody to determine the TE cell number (Figure 4E). The number of ICM cells was indirectly determined by subtracting the TE number from the total cell number. The results showed that CARM1 KD resulted in a significant reduction in the number of total cells, ICM and TE cells (Figure 4F) (P < 0.05). However, the ratio of ICM cells to TE cells in the CARM1 KD blastocysts was similar to that in the control groups (Figure 4F). Similarly, CARM1 KD did not impair the development of IVF embryos before the morula stage (data not shown), while CARM1 KD led to a significant reduction in both embryos that developed to blastocyst stage (days 5–7) (Supplementary Figures 4A,B) and lineage cells of IVF blastocysts (Supplementary Figures 4C,D) (P < 0.05). Hence, these results indicate that CARM1 is indispensable for blastocyst formation and normal lineage allocation.

To decipher the molecular mechanisms of CARM1 regulating embryo deve0lopment, single-embryo RNA sequencing was utilized to characterize tanscriptomic changes in CARM1 KD embryos. To validate the single-embryo RNA sequencing data, the expression levels of 6 differentially expressed genes (DEGs) (3 downregulated and 3 upregulated genes), namely ID2, CCND2, CARM1, PLK5, PPP2R2C, and CCNA1, were analyzed by qPCR. The expression patterns of these selected genes are highly consistent with the treads obtained from RNA sequencing data (Supplementary Figure 5), suggesting the robustness of the results obtained by RNA sequencing. We identified a total of 154 DEGs between CARM1 KD and uninjected control groups, of which there were 89 down-regulated genes and 65 up-regulated genes, respectively (Figure 5A and Supplementary Table 3). Furthermore, GO analysis was performed to annotate the potential function of the DEGs in embryo development. The DEGs were classified into three main categories (biological process, cellular component, and molecular function) according to the GO database. The top-ranking 10 biological processes, 10 cellular components, and 10 molecular functions involved in each GO term were provided (Figure 5B and Supplementary Table 4) (P < 0.05), such as positive regulation of cell population proliferation (10 genes, e.g., AREG, CCNA1, FGF19, CARM1, CCND2), negative regulation of epithelial cell proliferation (4 genes, e.g., WNT10B, GPC3), cell adhesion (4 genes, e.g., ADAM8, TGFBI). Lastly, KEGG analysis revealed that the DEGs were mainly enriched in 10 significant pathways (Figure 5C and Supplementary Table 5) (P < 0.05), such as Hippo signaling pathway (6 genes, e.g., WNT10B, ID2, PPP2R2C, AREG, ITGB2, CCND2), PI3K-Akt signaling pathway (6 genes, e.g., AREG, FGF19, EFNA4, CCND2, CCNA1). Altogether, these data demonstrate that CARM1 regulates the expression of key genes important for blastocyst formation.