Cellular redox state as a critical factor in initiating early embryonic-like program in embryonic stem cells

Dear Editor, Embryonic stem cells (ESCs) are pluripotent stem cells that can efficiently generate all embryonic but not extraembryonic tissues. However, a small percentage (0.1–1%) of totipotent-like cells arise spontaneously in ESC cultures, which have expanded cell fate potential to differentiate into both embryonic and extraembryonic cells. Intriguingly, these cells express high levels of transcripts including MERVL family of retroviruses and Zscan4 that are specifically activated in two-cell stage during embryo development. For these reasons, these rare cells are also called totipotent-like cells or 2C-like cells. Furthermore, these cells can be labeled with a fluorescence protein reporter driven by the LTR of MERVL retroviruses, for example, MERVL-LTR::tdTomato (2C::tdTomato) reporter. Currently, the molecular factors contributing to the emergence of 2C-like state are still not clear. Zscan4 expression marks an intermediate state that precedes the 2C-like state. To identify pathways that initiate the emergence of 2C-like state, we performed RNA-Seq with purified Zscan4::GFP-positive ESCs. Totally there were 721 and 882 genes upregulated and downregulated for more than two fold in Zscan4::GFP positive versus negative ESCs (Supplementary Table S1). Interestingly, KEGG pathway analysis identified glutathione metabolism significantly enriched in downregulated genes with a fold of enrichment as 4.2 and P value as 0.001 (Supplementary Fig. S1a). Gene set enrichment analysis (GSEA) confirmed the overall reduction of glutathione metabolism in Zscan4::GFPpositive ESCs (Supplementary Fig. S1b). Glutathione is one of the most important antioxidants in cells, and its metabolism is known to affect cellular redox state. Based on this finding, we then checked whether the level of reactive oxygen species (ROS) is different between 2C-like cells and normal ESCs. Very strikingly, ROS level was significantly higher in 2C::tdTomato positive than negative cells (Fig. 1a, b), indicated by 2′, 7′-dichlorodihydrofluorescein diacetate (H2DCFDA), a chemically reduced form of fluorescein used as a general indicator for ROS in cells. In addition, a genetically coded fluorescent sensor HyPer showed that hydrogen peroxide level was significantly increased in 2C::tdTomato-positive cells (Supplementary Fig. S2). These results reveal an abnormal redox state characterized by increased ROS level in 2Clike cells that arise spontaneously in ESC culture. To test whether high ROS level can cause the transition of ESCs into 2C-like state, we treated ESCs with hydrogen peroxide and found that the fraction of 2C::tdTomatopositive cells was indeed significantly increased by the treatment (Fig. 1c, d). Consistent with the causative role of ROS, addition of ROS scavenger N-acetyl-cysteine (NAC) significantly repressed the effect of hydrogen peroxide (Fig. 1c, d). To further support that hydrogen peroxide promotes the emergence of 2C-like state, we performed RNA-Seq analysis of hydrogen peroxide-treated ESCs (Supplementary Table S2). The results showed that hydrogen peroxide-treated ESCs significantly enriched 2C-specific ZGA transcripts (Fig. 1e). In addition, a significant fraction of MERVL-LTR-driven genes were also upregulated in hydrogen peroxide-treated ESCs (Fig. 1f). Previously, knocking out miR-34a or G9a and knocking down LINE1 or CAF-1 (p150 and p60) have been shown to activate 2C-like program. Consistently, genes upregulated in these conditions were also significantly induced

Shown are mean ± SD, n = 3. The p-value was calculated by two-tailed Student's t test.   Table S1-3   Supplementary Table S1

RNA extraction and RT-qPCR
Total RNA was extracted from cells with Trizol reagent (Invitrogen, Cat. # 15596026).
For quantitative PCR with reverse transcription (RT-qPCR) analysis, cDNA were obtained from 500 ng total RNA using HiScript II Q RT SuperMix for qPCR (Vazyme, Cat. # R223). qPCR assay were carried out with AceQ qPCR SYBR Green Mater Mix reagent (Vazyme, Cat. # Q141) in 96-well dishes in three biological replicates on StepOne Plus Real-Time PCR System (Applied Biosystems) with standard protocols.
The expression levels were standardized by -actin housekeeping gene expression.
Primers for qPCR are listed in Supplementary Table S3.

ROS level quantification
The probe H2DCFDA (Molecular Probes, Cat. # C6827) was used for ROS measurement according to the manufacturer's instructions with slight modification. In brief, cells were collected and resuspended in pre-warmed HBSS buffer (Gibco™, Cat.

RNA-seq and bioinformatics analysis
Total RNA was purified by poly-T oligo-attached magnetic beads twice and then were used to generate double-stranded (ds) cDNA. The ds-cDNA was ligated to adaptors and sequenced by Illumina Genome Analyzer (Novogene). Reads were aligned to the mouse genome (mm10) with STAR (version 2.5.0) using the GENCODE transcript annotation as transcriptome guide. All programs were perform under the default settings except for special statements. Expression levels were quantified as normalized FPKM using Cufflinks (version 2.2.1). GSEA was used to calculate the enrichment of selected gene sets by java GSEA Desktop Application. R 3.5.1 were used for the generation of MA plot, scatter plot, box plot and venn diagram. The 2C-specific ZGA genes are genes activated during ZGA (the 2C stage) that are also enriched in 2C::tdTomato + cells from ref. 4 . The list of MERVL-LTR driven transcripts was from ref. 5 . The list of genes induced by mir-34a knockout was from ref. 6 . The list of genes induced by G9a knockout was from ref. 4 . The list of genes induced by LINE1 knocking down was from ref. 7 . The lists of genes induced by p150 or p60 knocking down were from ref. 1 . The list of genes up-or down-regulated in siPias4 treated ESCs were from ref. 3 .

Quantification and statistical analysis
The number of independent experimental replications, the definition of center and precisions measures are reported in the figure legends (n, mean ± SD). p<0.05 is generally considered as statistically significant. Statistical analyses were performed using the GraphPad Prism v6 software and R programming. Statistical significance was assessed by two-tailed t test except when specified in the figure legends. For boxplots, upper and lower whiskers are defined as respectively Q3 + 1.5 x IQR and Q1 -1.5 x IQR with Q1 and Q3 being the first and third quartile of the plotted distribution and IQR the inter-quartile range, and the p-value was determined by Wilcoxon signed rank test. For multiple comparison, the p-value was calculated by one-way or two-way ANOVA followed with Dunnett's test. And hypergeometric test was used for venn diagram significant test.