Zygotes lacking maternal Dppa3 activity fail to reach the blastocyst stage (Payer et al., 2003; Bortvin et al., 2004), demonstrating that Dppa3 plays a role during the cleavage stages of pre-implantation development. However, the contribution of Dppa3 to 2CLCs reprogramming in ESCs has not been studied in detail. We hypothesized that Dppa3 may regulate a 2C-like state. To address this hypothesis, we first generated ESCs with exogenous Dppa3 overexpression (Dppa3 OE ESCs). Morphologically, Dppa3 OE ESCs showed compacted cell colonies (Figure 1A). Increased expression levels of Dppa3 in Dppa3 OE ESCs were confirmed by qPCR and western blotting (Figures 1B, C).

Next, we performed the mRNA-seq analysis. We identified 476 upregulated genes and 174 downregulated genes, respectively, more than a 1.5-fold change in Dppa3 OE ESCs (Supplementary Table S4). RNA-seq results showed that transcripts of 2C-specific genes were significantly elevated in Dppa3 OE ESCs, which were normally repressed by DNA methylation such as Zscan4, Tcstv1, and Tcstv3 (Figure 1D). A gene set enrichment analysis for the stage specific genes also revealed 2C-specific genes were enriched in Dppa3 OE ESCs (Figure 1E). The population of upregulated genes were also enriched with genes associated with germ cell development, such as spermatid development, spermatogenesis, and meiotic cell cycle in gene ontology (GO) terms (Supplementary Figure S1). We performed a detailed examination of the expression of 2C-like genes and endogenous transposable element (TE) in Dppa3 OE ESCs, and found that MERVL, LINE1, and IAP were significantly activated in Dppa3 OE ESCs (Figure 1F). MERVL-LTR transcriptional network activation occurs not only in two-cell embryos in vivo but also in vitro in a subset of mouse ESCs and during iPSC reprogramming (Eckersley-Maslin et al., 2016). Zscan4, expressed transiently in sporadic ESCs (1–5%) at any given time, marks transient 2C-state cells (Zalzman et al., 2010; Rodriguez-Terrones et al., 2018). The most significant difference in the transcription factor expression between the totipotent 2CLCs and the pluripotent ESCs was the Zscan4 expression (Falco et al., 2007). So, we next investigated the expression pattern of Zscan4 in Dppa3 OE ESCs. Western blot results indicated that Dppa3 OE promoted the expression of Zscan4 (Figure 1G). The elevation protein level of Zscan4 in Dppa3 OE ESCs was also verified in another ESCs’ line (J1ES cells) (Figure 1H). We next examined the percentage of Zscan4-positive (Zscan4⁺) cells in Dppa3 OE ESCs. An increased percentage of Zscan4⁺ ESCs in Dppa3 OE ESCs was detected by immunofluorescence (IF) and cytometry analysis (Figures 1I,J; Supplementary Figure S2). Moreover, it can be seen from the immunofluorescence results that Zscan4⁺ ESCs also have a higher intensity of Dppa3 signal (Figure 1I). These findings suggested that the ectopic expression of Dppa3 activated the 2C-specific genes and enhanced the percentage of 2CLCs in ESCs.

To further evaluate the effect of Dppa3 in promoting the 2CLCs conversion, we generated Dppa3 knockdown (Dppa3 KD) ESCs by RNA interference (RNAi) and assessed the capacity for 2C-like state conversion. The mRNA and protein levels of Dppa3 in Dppa3 KD ESCs were effectively reduced to about 30% of that in the control ESCs (Figures 2A, B). Compared with Dppa3 KD1, the knockdown efficiency of Dppa3 KD2 was more significant, so Dppa3 KD2 ESCs were used for further analysis. Zscan4 protein levels were significantly decreased in Dppa3 KD ESCs, compared to control ESCs, by using the western blot analysis (Figure 2B). It was also confirmed in another embryonic stem cell line, J1 ESCs, that the Dppa3 knockdown resulted in a decrease in Zscan4 protein levels (Figure 2C). Using immunofluorescence, we observed a pronounced reduction in the population of Zscan4⁺ cells in Dppa3 KD ESCs (Figure 2D). Flow cytometry also showed a significant reduction of Zscan4⁺ cells in Dppa3 KD ESCs (0.209%) compared with that in control ESCs (1.12%) (Figure 2E). We also examined the expression of 2C-related genes and endogenous transposable element (TE) in Dppa3 KD ESCs, and found that MERVL, LINE1, and IAP were significantly reduced in Dppa3 KD ESCs (Figure 2F). Taken together, combined overexpression of the Dppa3 increased Zscan4⁺ cells, and we conclude that Dppa3 is a driver of the 2C-like state.

DNA methylation globally loss in 2C-like cells (Dan et al., 2017). The molecular functions of Dppa3 on DNA methylation are demonstrated in zygotes, oocytes, and ESCs (Nakamura et al., 2012; Li et al., 2018; Mulholland et al., 2020), but little is known in 2CLCs conversion. To explore the underlying mechanisms of how 2C-like genes are activated in Dppa3 OE ESCs, we next analyzed the DNA methylation pattern in Dppa3 OE and control ESCs using immunofluorescence (IF) and reduced representation bisulfite sequencing (RRBS), respectively. While analyzing DNA methylation in ESCs with IF using a 5-methylcytosine (5mC) antibody, we noticed that DNA methylation levels were markedly lower in Dppa3 OE ESCs compared to that in control ESCs (Figure 3A). The number of 5mC foci in each cell was also quantified, which was significantly reduced in Dppa3 OE ESCs less than half of that in the control ESCs (Con), suggesting that the ectopic expression of Dppa3 reduced the level of DNA methylation in ESCs (Figure 3A).

To unravel DNA methylation changes, we next performed RRBS of control and Dppa3 OE ESCs to identify regions that are regulated by Dppa3. The RRBS analysis also revealed that the mCG levels were significantly reduced in Dppa3 OE ESCs, mainly distributing at repeats, intron, and 3′UTR (Figure 3B). CpG island (CGI) showed the lowest levels of methylation difference between the Dppa3 OE ESCs and the control ESCs, consistent with the fact that they are constitutively unmethylated in embryonic stem cells (Deaton and Bird, 2011). We specifically compared DMR (differentially methylated regions) methylation in Dppa3 OE ESCs with that in the control ESCs. Approximate 60% of DMRs were methylated in control ESCs, and a lower proportion (30%) of methylated DMRs was found in the Dppa3 OE ESCs (Figure 3C). DMRs were broadly distributed on all 19 autosomes (data not shown). The DMRs analysis identified 6,920 (98.2%) hypomethylated regions, whilst only 129 (1.8%) hypermethylated regions in Dppa3 OE ESCs (Supplementary Tables S5, 6). We therefore focused on the hypomethylated DMRs (hypo–DMRs). The majority hypo–DMRs overlapped with repeat regions (∼2,900 regions, 41.91%). Only 6.36% of the regions mapped to the promoter (Figure 3D). The decrease in DNA methylation was particularly pronounced at LINE-1 (L1) elements. This widespread DNA hypomethylation was reminiscent for the global decrease in DNA methylation accompanying the 2CLCs reprogramming (Eckersley-Maslin et al., 2016). Next, we performed the gene ontology (GO) analysis using genes with differentially methylated promoters (DMPs). Among the hypomethylated DMPs (440), 97 genes (22%) were 2C-like genes (Figure 3E). These data strongly suggest that Dppa3 regulates DNA hypomethylation in ESCs, especially for 2C-like genes and retroviral elements.

Remarkably, by the GSEA analysis, DNA damage and recombination pathways were also enriched in Dppa3 overexpression ESCs (Figure 4A). Recently, it has been reported that DNA damage is required for 2C-like cells’ induction in ESCs (Grow et al., 2021). p53 has been shown to be necessary for DNA damage-mediated Dux induction and the emergence of 2CLCs (Atashpaz et al., 2020; Grow et al., 2021). There are many sources of endogenous DNA damage in early embryos (Ziegler-Birling et al., 2009; Srinivasan et al., 2020), among which p53 has been found to be activated shortly after fertilization. p53 plays an important role in Dux activation and Dux targeted expression during ZGA, and the loss of p53 can lead to the reduction of endogenous fluctuating 2CLCs (Storm et al., 2014; Hendrickson et al., 2017; Grow et al., 2021). Phosphorylated H2AX (γH2AX) is also commonly used as a DNA damage response marker. To assess whether Dppa3 is responsible to DNA damage response, we performed immunofluorescence and western blot analysis of γH2AX and p53 in Dppa3 OE and Dppa3 KD ESCs. The immunofluorescence intensity and protein levels of γH2AX and p53 were both significantly increased following by Dppa3 overexpression and decreased by Dppa3 knockdown in ESCs (Figures 4B–D). Immunofluorescence results also indicated that Dppa3 closely co-localized with γH2AX (Figure 4C), demonstrating that Dppa3 promotes DNA damage response in ESCs.

DNA demethylation always couples with histone modification and chromatin configuration. Next, we profiled the chromatin accessibility of Dppa3 OE ESCs by assay for transposase-accessible chromatin sequencing (ATAC-Seq) (Li et al., 2019; Yan et al., 2020). After peaks in each condition were called, regions of different ATAC sensitivities were identified. We identified 3,117 regions that gained an ATAC signal (ATAC gained) in Dppa3 OE ESCs compared to that in Con ESCs. Furthermore, we found that Dppa3 OE cells exhibited more extensive and specific opening of chromatin specifically at the 2CLCs-related gene set (Figure 5A). The Zscan4 cluster has nine paralogous genes (including pseudogenes) located at heterochromatin (sub-telomere region). Three of them, Zscan4c, Zscan4d, and Zscan4f, encode full-length ORFs with 506 amino acids. Through the analysis of RNA-seq, we found that Dppa3 overexpression elevated the transcription level of Zscan4 (Figure 5B; Figure 1D). Zscan4 was activated after Dppa3 overexpression in ESCs and has been identified by prior studies as the marker of the 2C-like states (Figure 1). Our ATAC-Seq results also indicated that Zscan4 elements are ATAC signal-enriched loci in Dppa3 OE ESCs (Figure 5C). Like Zscan4, MERVL elements, which are located near the heterochromatin regions, are also ATAC signal-enriched loci in Dppa3 OE ESCs, further suggesting that Dppa3 facilitate the epigenomic remodeling especially for the heterochromatin (Figure 5D). Dppa3-induced 2C-like cells exhibited an extensive and specific opening of chromatin at MERVL and Zscan4 elements, which are specifically expressed in 2C stage embryos to regulate 2CLCs conversion (Hendrickson et al., 2017).

In previous observations, non-methylated CpG chromatin was usually enriched with H3K4me3. In the two-cell embryonic stage, the number of H3K4me3-marked promoters increased dramatically, and many promoters retained this modification in the following stages till the blastocyst stage. So, we next investigated the histone modification in Dppa3 OE ESCs. Initially, we examined the H3K4me3 by immunofluorescence, which revealed that the global levels of H3K4me3 were increased in Dppa3 OE ESCs (Figures 5E,F). Consistent with this, Dppa3 knockdown reduced the H3K4me3 modification (Figures 5E,F). Furthermore, Dppa3 had a negative regulation on heterochromatin markers, such as H3K9me3 and H3K27me3 (Figure 5G), further indicating that Dppa3 promotes epigenetic remodeling in ESCs. Together, these data demonstrate that Dppa3 facilitates DNA hypomethylation in ESCs and induces ESCs acquiring chromatin accessibility at Zscan4 and MERVL elements promoting 2C-like genes expression and 2CLCs conversion (Figure 6) (Eckersley-Maslin et al., 2016; Zhang et al., 2021).

Mouse ESCs (J1 or D3) were all cultured in high glycose DMEM (HyClone) supplemented with 15% fetal bovine serum (HyClone), 2 mM l-glutamine (Gibco), 5,000U/mL penicillin and streptomycin (Gibco), 0.1 mM NEAA (Gibco), 0.1 mM 2-mercaptoethanol (CAS No. 60-24-2, Sigma), and 1,000U/ml LIF (ESG1107, Millipore Crop) at 37 °C in 5% CO2 incubator.

Murine Dppa3 CDS was cloned into expression vector pLch3.7 (CAG promoter). pLch3.7–Dppa3 and empty vectors (served as control) were transfected into ESCs by Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After 2 weeks selection (1.5 μg/ml puromycin), Dppa3 overexpression (Dppa3 OE) ESCs colonies were picked and cultured in the ESCs culture medium for stable Dppa3 OE ESCs lines. ESCs were transfected with pLch3.7-Dppa3 or empty vectors and then selected by puromycin (2 μg/ml) for 24 h, which were used as transient Dppa3 overexpression ESCs.

Control (Con) and shRNA sequences (Supplementary Table S3) against Dppa3 mRNA were used for the Dppa3 knockdown (Dppa3 KD) experiment. The sequences were cloned into pSIREN-RetroQ vector (Clontech). The reconstructed pSIREN-RetroQ vectors were then transfected into ESCs using Lipofectamine 2000 (Invitrogen). The cells were then cultured by the non-penicillin or streptomycin culture medium for 2 days. ESCs were then selected by puromycin (1.5 μg/ml) for 24 h, and positive clones were next used in the experiment.

To analyze the endogenous Zscan4 expression profile (percentage of Zscan4⁺ cells and fluorescence intensity), ESCs were collected and washed with cold PBS, then fixed in cold 4% PFA, permeabilized in 0.1% Triton X-100 in blocking solution (4% BSA in PBS) for 30 min, washed three times, and left in the blocking solution for 1 h. ESCs were incubated 1 h at room temperature with the primary antibody against Zscan4 (1:1000, Millipore), washed three times, and incubated for 30 min with secondary antibodies, FITC goat anti-rabbit IgG (1:1000, Beyotime). Samples were washed three times with PBS and the FACS analysis was performed using a FACSAria Flow Cytometer (BD Biosciences).

ESCs were fixed in 4% paraformaldehyde for 30 min at room temperature and then permeabilized with 0.1% Triton X-100 for 30 min. Then cells were blocked in 10% goat serum for 1 h at room temperature. Primary antibodies against Dppa3 (1:200; R&D systems, MAB2566), Zscan4 (1:1000, Millipore, AB4340), H3K4me3 (1:1000, Abcam, ab8580), γH2AX (1:1000, Cell Signaling Technology, 80312), and 5mC (1:200, Abcam, ab10805, the cells were treated with 2 M HCl for 30 min, then neutralized with Tris-HCl of pH 8.0) were diluted in 1% normal serum (v/v) and incubated overnight at 4 °C and detected with the respective secondary antibodies including Alexa Fluor 594 secondary antibody (1:500; Proteintech, SA00006-4), Alexa Fluor 647 secondary antibody (1:500; Abcam, ab150159), Alexa Fluor 555 secondary antibody (1:500; Beyotime, A0460) at room temperature for 1 h and labeled with DAPI (Sigma) to visualize the nuclei. Sample images were captured using laser scanning confocal microscopy.

Total RNA was prepared using TRIzol (Invitrogen) and samples were treated with DNase (DNA-free, Ambion) following the instructions. 2 μg RNA was reverse-transcribed using the first-strand cDNA Synthesis System (Roche) with oligonucleotides. Real-time PCR was performed using the TransStart Green qPCR SuperMix Kit (TransGen Biotech). All quantitative PCR (qPCR) reactions were performed in the Opticon® System (Bio-Rad). Primer sequences are listed in Supplementary Table S2. The expression level of genes was determined from the threshold cycle (CT), and the relative expression levels were normalized to the expression of 18S ribosomal RNA, respectively, and calculated by the 2^−ΔΔCT method as described in the literature.

ESCs were harvested and lysed in the Western lysis buffer (Beyotime). For Western blots, Samples were boiled at 95 °C for 10 min with 1×SDS sample buffer before loading onto 4–20% Tris-glycine gels (Bio-Rad). Resolved proteins were transferred to the polyvinylidene difluoride membrane (0.25 mm pore, Millipore). Membranes were then blocked for 1 h in 1% casein prepared in Tris-buffered saline and 0.1% Tween-20 (TBST) before blotting with respective primary antibodies diluted in TBST, overnight at 4 °C. Blots were washed three times with TBST and incubated with the secondary antibody in the same buffer for 1 h at room temperature. Post three TBST washes, the membranes were imaged on an automatic chemiluminescence image analysis system. Quantitation of signal and analysis was performed using the ImageJ software. The following primary antibodies were used: anti-Dppa3 (1:1000, Abcam, ab19878), anti-H3 (1:5000, Abcam, ab1791), anti-γH2AX (1:1000, Cell Signaling Technology, 80312), anti-Zscan4 (1:1000, Millipore, AB4340), and anti-Tubulin (1:10,000, Proteintech, 66031-1-Ig). The following secondary antibodies conjugated to horseradish peroxidase were used: goat polyclonal anti-rabbit IgG (Bio-Rad), and rabbit polyclonal anti-mouse IgG (Sigma, 1:5000).

ESCs were harvested and total RNA was extracted using TRIzol (Invitrogen). Quality control of the extracted RNA, construction of an RNA-sequencing library and sequencing on BGISEQ-500 was performed at Shenzhen Genomics Institute (BGI). For the bioinformatics analysis, the clean reads were mapped to the Mus musculus mm10 reference genome using Bowtie2. Reads were assigned and counted to genes using the RSEM. Functional enrichment (GO annotation, KEGG, and GESA) of gene sets with different expression patterns was performed using cluster Profiler and GSEA (R version 4.1.0), respectively. Scatter plots were generated using Excel to graphically reveal genes that differ significantly between two samples. The RNA-seq data have been uploaded in GEO (GSE number: 197967).

After the ESCs were counted, they washed once with 50 ml of cold PBS. The nuclei from 100,000 ES cells were lysed in 50 μL of lysis buffer (10 mM Tris–HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2 and 0.1% NP-40). Immediately after lysis, the nuclei were spun at 500×g for 5 min to remove the supernatant. Nuclei were then incubated with Tn5 transposase and segmentation buffer (Illumina) at 37 °C for 30 min. After segmentation, the transposed DNA was purified with a Min Elute kit (Qiagen). Samples were then amplified by PCR. Post-PCR clean-up was performed by adding 1.2 volume of AMPure XP bead (Beekman). Library concentration was measured with Qubit. Library integrity was checked by the Agilent bioanalyzer, and 75 bp single-read sequencing was performed using an Illumina HiSeq 2,500 platform per standard operating procedures.

DNA concentration was measured using Qubit® DNA Assay Kit in Qubit® 2.0 Flurometer (Life Technologies). A total amount of 5.2 μg genomic DNA from each sample was digested using MspI (New England Biolabs, United States ), followed by end repair and adenylation. Cytosine-methylated barcodes were ligated to DNA following the manufacturer’s instructions. Fragments with the 40–220 bp size were selected by gel electrophoresis, purified, and bisulfite-treated using EZ DNA Methylation-Gold Kit (Zymo Research, United States ). The bisulfite converted DNA was then PCR amplified by KAPA HiFi HotStart Uracil + ReadyMix (2X). Library concentration was quantified by Qubit® 2.0 Flurometer (Life Technologies) and the insert size was assayed on an Agilent Bioanalyzer 2,100 system (Agilent Technologies, United States ). The library preparations were sequenced on an Illumina Hiseq (High Output Mode).

All presented results were obtained from at least three independent experiments for each condition. Data are expressed as mean ± SEM. Statistical analyses were performed by one- or two-way analysis of variance using GraphPad Prism software. Differences were considered statistically significant at p < 0.05.