An ancestral feature of eukaryotic genomes is the presence of nucleosomes that wrap DNA around an octamer of two copies of each of the four core his- tones H2A, H2B, H3 and H4. Nucleosomes constrain negative DNA supercoils and limit DNA accessibility, necessitating nucleosome mobilization to carry out gene regulation 1 . The core histones are each character- ized by a histone fold domain (HFD) that enables them to dimerize in specific antiparallel pairs, H3 with H4 and H2A with H2B, that can further assemble by forming four-helix bundles between dimers, leading to a central H3–H4 tetramer flanked by two H2A–H2B dimers 2 . In addition to HFDs, core histones have unstructured tails that are subject to numerous post-translational modifi- cations (PTMs) with important roles in gene regulation. Histone variants, especially of H2A and H3, may replace the corresponding core histone to form nucleosomes with distinct properties 3 . Fossil stromatolites, interpreted as microbial mats and fossil fibres from hydrothermal vents, suggest that bacterial cells with early forms of gene expres- sion were extant ~3.7 billion years ago 4–6 . By contrast, steranes — probably of eukaryotic (or proto-eukaryotic) origin — date from 2.5–2.7 billion years ago 7 , and large ornamented fossil cells confidently related to an extant eukaryote group (red algae) date from 1.2–1.6 billion years ago 8, 9 . Although putatively early- diverging eukaryotes such as metamonads and kinetoplastids may have separated from other eukaryotes before the appearance of algae 10 , the fossil record suggests a 1–2 billion year gap between the early origin of prokary- otic transcription factor (TF)-based gene expression sys- tems and the nucleosome-based regulation of modern eukaryotic genes. This gap raises the question of how the acquisi- tion of eukaryotic nucleosomes fundamentally altered gene- regulatory processes. Except for octameric nucleosomes, nearly all enzymatic components of the eukaryotic chromatin landscape, including TFs, poly- merases, topoisomerases, acetyltransferases, deacetylases, SET domain methyltransferases and even homologues of ATP-dependent chromatin remodellers, are present in prokaryotes. The traditional view is that nucleosomes act as repressors of gene expression 1 and that certain acti- vating PTMs of histones, histone variants, and chroma- tin remodellers promote gene expression. To the extent that they do this, what are their mechanisms of action? How did they acquire these roles in an emerging nucleo- some landscape? The last eukaryotic common ancestor (LECA) was a complex nucleated cell with an endomem- brane system, cytoskeleton, mitochondrion, and linear chromosomes that underwent mitosis and meiosis 11 . The origin of eukaryotic cells is controversial, and the first eukaryotic common ancestor (FECA) has been variously proposed to be a hypothetical cell equally ancient as, but independent of, bacteria and archaea (urkaryote), a bacte- rium with an archaeal endosymbiont, or an archaeon with a bacterial endosymbiont (BOX 1); however, in any scenario, Histone fold domain (HFD). A protein dimerization domain of three helices separated by two loops that is characteristic of archaeal and core eukaryotic histones, TATA- binding protein-associated factors and some other proteins. Four-helix bundles Structures formed by two helices of each of two histones that enable dimers to assemble into more complex structures. Old cogs, new tricks: the evolution of gene expression in a chromatin context Paul B. Talbert, Michael P. Meers and Steven Henikoff * Abstract | Sophisticated gene-regulatory mechanisms probably evolved in prokaryotes billions of years before the emergence of modern eukaryotes, which inherited the same basic enzymatic machineries. However, the epigenomic landscapes of eukaryotes are dominated by nucleosomes, which have acquired roles in genome packaging, mitotic condensation and silencing parasitic genomic elements. Although the molecular mechanisms by which nucleosomes are displaced and modified have been described, just how transcription factors, histone variants and modifications and chromatin regulators act on nucleosomes to regulate transcription is the subject of considerable ongoing study. We explore the extent to which these transcriptional regulatory components function in the context of the evolutionarily ancient role of chromatin as a barrier to processes acting on DNA and how chromatin proteins have diversified to carry out evolutionarily recent functions that accompanied the emergence of differentiation and development in multicellular eukaryotes. Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA. *e-mail: steveh@fhcrc.org https://doi.org/10.1038/ s41576-019-0105-7 REVIEWS NATURE REVIEWS | GENETICS