H3K9me3 Histone Mark Drives DNA Methylation Recovery After Germline Epigenome Editing

Epigenetic inheritance — the idea that acquired marks in parental cells can influence offspring biology — remains hotly debated because direct mechanistic evidence has been elusive. This study addresses that gap by inventing a germline-specific epigenome editing platform in mice, applying it to the H19 differentially methylated region (H19-DMR), a well-characterized imprinting control locus. The H19-DMR is paternally methylated in healthy individuals; loss of this methylation in sperm causes Silver-Russell syndrome (SRS), characterized by fetal growth restriction due to Igf2 downregulation. By precisely recapitulating this epimutation without any DNA sequence change, the authors could cleanly test whether epigenetic information in sperm is transmitted to offspring.

The editing system uses a dCas9-SunTag construct fused to the catalytic domain of TET1 hydroxylase, guided by nine gRNAs targeting H19-DMR, and driven by the Stra8 premeiotic-specific promoter. This confines editing to spermatogonia through spermatocytes — after the endogenous paternal imprint is established at the pro-spermatogonia stage — ensuring targeted demethylation occurs post-imprinting. Three independent transgenic strains were generated by pronuclear injection of linearized vector. Bisulfite sequencing confirmed near-complete loss of CpG methylation across H19-DMR in sperm from all strains, while six tested off-target loci showed minimal change. Maternal oocytes, as expected, remained unmethylated regardless of genotype.

Offspring derived from transgenic fathers — even those not inheriting the transgene themselves — showed intrauterine growth retardation at birth, consistent with SRS-like phenotypes. CpG methylation levels at CTCF-binding sites m1–m4 within H19-DMR were significantly reduced in these offspring compared to wild-type controls. However, methylation was only partially lost (not completely absent), indicating that de novo remethylation occurred during pre-implantation development. This partial recovery was consistent across multiple offspring and represented genuine intergenerational inheritance at a reduced penetrance, not full transmission. Importantly, no epigenetic alterations were detected in F2 offspring, establishing that transgenerational (beyond F1) inheritance did not occur at this locus.

To mechanistically explain the partial remethylation, the authors interrogated histone modifications at H19-DMR after fertilization. Using targeted histone editing in zygotes — employing a dCas9-KRAB fusion to remove H3K9me3 specifically at H19-DMR shortly after fertilization — they demonstrated that depletion of this mark abolished the subsequent de novo DNA methylation recovery. H3K9me3 is known to be deposited on paternal chromatin after fertilization and is a known recruiter of DNMT3A/3L de novo methyltransferase complexes. These results position H3K9me3 as an instructive intermediate: even when sperm DNA methylation is erased, residual or newly deposited H3K9me3 at the locus guides remethylation in the early embryo. The same mechanism was found to operate at other imprinted loci, suggesting broader relevance.

The study's significance is threefold. First, it provides a clean, sequence-neutral germline epigenome editing tool that generates heritable epimutations without confounding genetic changes — a methodological advance over prior CRISPR-insertion approaches. Second, it demonstrates partial intergenerational (F0→F1) but no transgenerational (F1→F2+) inheritance at H19-DMR, clarifying the boundaries of epigenetic memory transmission. Third, it identifies H3K9me3 as a previously uncharacterized molecular bridge between germline-erased DNA methylation and post-fertilization remethylation, suggesting that histone marks — not just DNA methylation — serve as the true carriers of epigenetic memory through the reprogramming window.