Single-cell 5-formylcytosine Landscapes of Mammalian Early Embryos at Single-base Resolution

Active DNA demethylation in mammals involves Ten-eleven translocation (TET) family proteins-mediated oxidation of 5-methylcytosine (5mC). However, base-resolution landscapes of 5-formylcytosine (5fC, an oxidized derivative of 5mC) at the single-cell level remains unexplored. Here we present “CLEVER-seq”, which is a single-cell, single-base resolution 5fC sequencing technology, based on biocompatible, selective chemical labeling of 5fC and subsequent C-to-T conversion during amplification and sequencing. CLEVER-seq of mouse early embryos reveals the highly patterned genomic distribution and parental specific dynamics of 5fC during mouse early pre-implantation development and synergistic 5fC production in paternal and maternal genomes. Integrated analysis demonstrates that promoter 5fC production precedes the expression upregulation of a clear set of developmentally and metabolically critical genes. Collectively, our work reveals the dynamics of active DNA demethylation during early mouse pre-implantation development and provides an important resource for further functional studies of epigenetic reprogramming in single cells. For the untreated control samples, the accession number of raw data is under GSE263008 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE263008). We developed a novel single-cell 5fC sequencing technique, termed “CLEVER-seq” (chemical labeling enabled C-to-T conversion sequencing), to map the genome-wide base resolution 5fC landscapes in mouse early embryos at single-cell level. We applied CLEVER-seq to mESCs (2i), mouse sperm, metaphase II oocytes, zygotes, 2-cell embryos, 4-cell embryos, inner cell mass (ICM) and trophectoderm (TE). The male and female pronuclei within each individual PN3-PN4 stage zygote were physically isolated by micromanipulation; the individual blasotomeres of 2-cell and 4-cell embryos were also separated. In addition to mESCs (2i), we also performed CLEVER-seq to mESCs (serum), mEpiSCs, and hESCs. In total, we sequenced more than 1,500 GB data for 139 single cells, and identified 3,216-41,347 5fCpG sites from these samples.