SRA

The TET dioxygenases erase mediate DNA dedemethylation in pre-implantation embryos and in primordial germ cells, yet limited studies address their contribution to the global gain of DNA methylation following implantation. Here, we show that Tet1 is expressed and non-redundantly contributes to 5-hydroxymethylctyosine (5hmC) non-redundantly in the pre-gastrulation mouse epiblast. Ablation of Tet1 in primed epiblast cells results in widespread loss of 5hmC associated with gain of 5-methylcytosine at CpG islands and promoters. Moreover, Tet1 is expressed, albeit at lower levels, in the extra-embryonic ectoderm. Tet1-deficiency in the pre-streak mouse embryos causes dysregulation of early lineage regulators in the epiblast and increased expression of metabolic genes in the extra-embryonic ectoderm. Our studies reveal a distinct role of Tet1 in regulating the methylome landscape of the post-implantation mammalian epiblast and a hitherto unknown gene repressive effect in the extra-embryonic lineage, providing insights into the early developmental origins of epigenetic-based basis of imprinting and developmental disorders. Overall design: Lysates of E6.25 epiblast (E) and extra-embryonic (ExE) tissues were pooled, based on retrospective genotyping, from at least 6 embryos (collected from 2-3 litters). Four pooled biological replicates were collected for each tissue compartment and each genotype, i.e. wild type (wt), Tet1gt/wt (Zwt), Tet1gt/gt(ZZ); in 3 of 4 pooled replicates, the E and ExE samples were obtained from the same litters and given the same numbering. One ExE sample of ZZ was lost. RNA was purified using the RNeasy micro Kit (Qiagen) to obtain 1-10 ng of RNA per pool. RNA quality was assessed using a Bioanalyzer 6000 Pico chip and only samples with RNA integrity number (RIN) > 8.0 were processed further. First-strand cDNA synthesis and amplification was performed by using the SMARTer Ultra Low Input RNA amplification kit v3 (Clontech). Libraries were prepared using NEBNext Ultra DNA library prep for Illumina and sequenced on NextSeq 500 in high-output to generate 75-bp single-end reads.  Low quality ends and adapter sequences were trimmed off from the Illumina reads with FastX 0.0.13 and cutadapt 1.7.1. Using FastX and ShortRead 1.16.3, we filtered subsequently small reads (length < 35 bp), polyA-reads (>90% of the bases equal A), ambiguous reads (containing N) and low quality reads (>50% of the bases < Q25).