SummaryGlobal DNA demethylation is an integral part of reprogramming processes in vivo and in vitro, but whether it occurs in the derivation of induced pluripotent stem cells (iPSCs) is not known. Here, we show that iPSC reprogramming involves both global and targeted demethylation, which are separable mechanistically and by their biological outcomes. Cells at intermediate-late stages of reprogramming undergo transient genome-wide demethylation, which is more pronounced in female cells. Global demethylation requires activation-induced cytidine deaminase (AID)-mediated downregulation of UHRF1 protein, and abolishing demethylation leaves thousands of hypermethylated regions in the iPSC genome. Independently of AID and global demethylation, regulatory regions, particularly ESC enhancers and super-enhancers, are specifically targeted for hypomethylation in association with transcription of the pluripotency network. Our results show that global and targeted DNA demethylation are conserved and distinct reprogramming processes, presumably because of their respective roles in epigenetic memory erasure and in the establishment of cell identity.

Transcription of protein coding genes is accompanied by recruitment of COMPASS to promoter-proximal chromatin, which methylates histone H3 lysine 4 (H3K4) to form H3K4me1, H3K4me2 and H3K4me3. Here, we determine the importance of COMPASS in maintaining gene expression across lifespan in budding yeast. We find that COMPASS mutations reduce replicative lifespan and cause expression defects in almost 500 genes. Although H3K4 methylation is reported to act primarily in gene repression, particularly in yeast, repressive functions are progressively lost with age while hundreds of genes become dependent on H3K4me3 for full expression. Basal and inducible expression of these genes is also impaired in young cells lacking COMPASS components Swd1 or Spp1. Gene induction during ageing is associated with increasing promoter H3K4me3, but H3K4me3 also accumulates in non-promoter regions and the ribosomal DNA. Our results provide clear evidence that H3K4me3 is required to maintain normal expression of many genes across organismal lifespan.

SummaryCritical roles for DNA methylation in embryonic development are well established, but less is known about its roles during trophoblast development, the extraembryonic lineage that gives rise to the placenta. We dissected the role of DNA methylation in trophoblast development by performing mRNA and DNA methylation profiling of Dnmt3a/3b mutants. We find that oocyte-derived methylation plays a major role in regulating trophoblast development but that imprinting of the key placental regulator Ascl2 is only partially responsible for these effects. We have identified several methylation-regulated genes associated with trophoblast differentiation that are involved in cell adhesion and migration, potentially affecting trophoblast invasion. Specifically, trophoblast-specific DNA methylation is linked to the silencing of Scml2, a Polycomb Repressive Complex 1 protein that drives loss of cell adhesion in methylation-deficient trophoblast. Our results reveal that maternal DNA methylation controls multiple differentiation-related and physiological processes in trophoblast via both imprinting-dependent and -independent mechanisms.

Extrachromosomal circular DNA (eccDNA) facilitates adaptive evolution by allowing rapid and extensive gene copy number variation and is implicated in the pathology of cancer and ageing. Here, we demonstrate that yeast aged under environmental copper accumulate high levels of eccDNA containing the copper-resistance gene CUP1. Transcription of the tandemly repeated CUP1 gene causes CUP1 eccDNA accumulation, which occurs in the absence of phenotypic selection. We have developed a sensitive and quantitative eccDNA sequencing pipeline that reveals CUP1 eccDNA accumulation on copper exposure to be exquisitely site specific, with no other detectable changes across the eccDNA complement. eccDNA forms de novo from the CUP1 locus through processing of DNA double-strand breaks (DSBs) by Sae2, Mre11 and Mus81, and genome-wide analyses show that other protein coding eccDNA species in aged yeast share a similar biogenesis pathway. Although abundant, we find that CUP1 eccDNA does not replicate efficiently, and high-copy numbers in aged cells arise through frequent formation events combined with asymmetric DNA segregation. The transcriptional stimulation of CUP1 eccDNA formation shows that age-linked genetic change varies with transcription pattern, resulting in gene copy number profiles tailored by environment.

Understanding species' responses to environmental change underpins our abilities to make predictions on future biodiversity under any range of scenarios. In spite of the huge biodiversity in most ecosystems, a model species approach is often taken in environmental studies. To date, we still do not know how many species we need to study to input into models and inform on ecosystem-level responses to change. In this study, we tested current paradigms on factors setting thermal limits by investigating the acute warming response of six Antarctic marine invertebrates: a crustacean Paraceradocus miersi, a brachiopod Liothyrella uva, two bivalve molluscs, Laternula elliptica, Aequiyoldia eightsii, a gastropod mollusc Marseniopsis mollis and an echinoderm Cucumaria georgiana. Each species was warmed at the rate of 1°C h À1 and taken to the same physiological end point (just prior to heat coma). Their molecular responses were evaluated using complementary metabolomics and transcriptomics approaches with the aim of discovering the underlying mechanisms of their resilience or sensitivity to warming. The responses were species-specific; only two showed accumulation of anaerobic end products and three exhibited the classical heat shock response with expression of HSP70 transcripts. These diverse cellular measures did not directly correlate with resilience to heat stress and suggested that each species may have a different critical point of failure. Thus, one unifying molecular mechanism underpinning response to warming could not be assigned, and no overarching paradigm was supported. This biodiversity in response makes future ecosystems predictions extremely challenging, as we clearly need to develop a macrophysiology-type approach to cellular evaluations of the environmental stress response, studying a range of well-rationalized members from different community levels and of different phylogenetic origins rather than extrapolating from one or two arbitrary model species.