https://www.nature.com/articles/s41467-026-72381-0
Lots of potential interest in this paper. Not least Cnidarians some of which (I have not checked this one) have the ability to reverse development.
chatGPT:
Summary
This paper studies gene body methylation (gbM) in the sea anemone Nematostella vectensis, a cnidarian with an ancestral-style animal methylome. In most invertebrates, DNA methylation is not spread across the whole genome as in vertebrates, but is concentrated within the bodies of highly and broadly expressed genes. The core question is: what is gbM actually doing?
The authors show that Nematostella has stable CpG methylation across development and adult tissues, averaging about 12% global mCG, and that highly methylated genes tend to be transcriptionally stable rather than dynamically regulated. This supports the idea that gbM is not mainly a switch controlling ordinary gene expression.
They then depleted DNA methylation using the DNMT1 inhibitor GSK-3484862, achieving roughly a sixfold reduction in methylation at gastrula stage, yet embryos remained viable and developed to primary polyps with survival comparable to controls. The main molecular effect was not massive gene-expression disruption, but chromatin opening inside methylated gene bodies, especially at intragenic transposable elements. The page 3 figure shows new ATAC-seq and CAGE-seq signal inside a representative gene after demethylation, indicating ectopic accessibility and cryptic transcription start sites.
The key mechanistic conclusion is that gbM acts as a permissive repression system: it suppresses spurious intragenic promoters and transposable-element-derived transcription while still allowing the host gene to be actively transcribed. This is distinct from heterochromatin, which would be more likely to interfere with transcription through the gene body.
The inheritance part is also important. After early demethylation, methylation was not efficiently restored in somatic tissues, even through repeated regeneration. It was partially restored in the germline, especially in sperm, and recovery was biased toward transcriptionally active regions marked by gene-body-associated histone modifications such as H3K79me2/3 and H3K36me3. The page 4 figure shows fast- and slow-recovering DMR classes and links recovery to chromatin context.
Finally, crosses between demethylated and untreated animals showed that aberrant methylation states could pass into the next generation. Offspring methylation approximated the average of parental gamete methylation, consistent with limited or absent global post-fertilization methylation reprogramming. Inherited hypomethylated regions retained abnormal chromatin accessibility and partly abnormal transcription in F1 embryos.
Novelty
The main novelty is that the paper gives a direct mechanistic function for invertebrate gene body methylation. Rather than treating gbM as a vague marker of active genes, it shows experimentally that gbM suppresses cryptic intragenic transcription, especially from embedded transposable elements.
A second novel point is the demonstration that experimentally induced methylation defects can be partly inherited across generations in an early-branching animal. This is stronger than a simple correlation between methylation and genotype, because the authors perturb methylation directly and then follow recovery and inheritance.
A third interesting novelty is the proposed role of transcription-associated chromatin in remethylation. The finding that H3K79me2/3 predicts methylation recovery more strongly than expected is potentially important, because DNMT3 targeting is often discussed mainly in relation to H3K36me3.
Critique
The paper is strong because it combines methylation profiling, ATAC-seq, RNA-seq, CAGE-seq, regeneration, germline sampling, crossing experiments and haplotype-resolved long-read methylation. That makes the central story unusually well triangulated.
The main limitation is that the key perturbation relies heavily on GSK-3484862, a DNMT1 inhibitor. The authors do use morpholinos against DNMT1/DNMT3/UHRF1 as supporting evidence, which helps, but chemical inhibition can still have off-target or DNMT1-protein-level effects. The F1 inheritance data partly addresses this because DNMT1 is no longer directly inhibited, but the initial F0 phenotype still depends substantially on pharmacological demethylation.
A second limitation is ecological and evolutionary extrapolation. The authors argue that inherited epimutations may provide regulatory variation with evolutionary consequences. That is plausible, but the paper does not show that these inherited methylation differences produce adaptive phenotypes under natural environmental conditions. It demonstrates inheritance and molecular consequences, not natural selection on those states.
A third issue is species generality. Nematostella is valuable because it is early-branching and retains ancestral-like methylation, but it is also unusual: it has strong regenerative capacity and apparently stable methylation across development. The findings may not generalize cleanly to insects, molluscs, annelids or other invertebrates with different methylome dynamics.
A fourth point is that the biological magnitude of the expression changes is modest. That is actually consistent with the authors’ argument that gbM is more about suppressing cryptic transcription than controlling normal gene expression, but it also means the physiological consequences remain somewhat underdeveloped.
Overall assessment
This is a convincing and important paper. Its best contribution is to reframe invertebrate gene body methylation as a genome-defence mechanism compatible with active transcription, rather than as a conventional gene-expression regulator. The inheritance findings are also significant, because they show how incomplete methylation resetting can generate persistent epigenetic variation. The biggest open question is whether such inherited epimutations matter for organismal fitness or adaptation in natural environments.