Parental nucleosome segregation and the inheritance of cellular identity

Abstract

Gene expression programmes conferring cellular identity are achieved through the organization of chromatin structures that either facilitate or impede transcription. Among the key determinants of chromatin organization are the histone modifications that correlate with a given transcriptional status and chromatin state. Until recently, the details for the segregation of nucleosomes on DNA replication and their implications in re-establishing heritable chromatin domains remained unclear. Here, we review recent findings detailing the local segregation of parental nucleosomes and highlight important advances as to how histone methyltransferases associated with the establishment of repressive chromatin domains facilitate epigenetic inheritance.

Fig. 1 |. Histone dynamics at the eukaryotic replisome.

During DNA replication in mammals, the nucleosome is disassembled ahead of the replication fork into one histone (H3–H4)₂ tetramer and two histone H2A–H2B dimers by the minichromosome maintenance (MCM) helicase shown at the fork. The leading strand shows the replication fork, simplified by depicting DNA polymerase-ε (Polε) and its interaction with PCNA, while the lagging strand is depicted first by Polα, then by replication protein A (RPA), Okazaki fragments in blue and DNA Polδ with its interaction with PCNA. MCM complex component 2 (MCM2) and anti-silencing function protein 1 (ASF1) chaperoning activities handle parental (H3–H4)₂ tetramer (blue histones and black tails) to the Polε subunits POLE3–POLE4 at the leading strand or chromosome transmission fidelity factor 4 (CTF4) and Polα at the lagging strand. Furthermore, newly synthesized H3–H4 dimers (blue histones and white tails) are chaperoned in the nucleus by ASF1, which further transfers H3–H4 dimers to chromatin assembly factor 1 (CAF1) at the fork. H2A–H2B dimers are chaperoned by NAP1 to form a nucleosome on newly synthesized DNA. Lastly, the ‘read–write’ mechanism for methylated histone H3 lysine 27 (H3K27me) and H3K9me is shown by a dashed arrow behind the replication fork, resulting in the methylation of nearby, newly synthesized H3 tails (white) and restoring a repressive chromatin state.

Fig. 2 |. Applications used for determining the fate of parental nucleosomes across DNA replication.

Four groups conceived of in vitro and in vivo techniques to ascertain whether parental nucleosomes are locally inherited. a | T antigen (for the simian virus 40 (SV40) replication system) or extract from Xenopus laevis was added to a plasmid reconstituted with nucleosomes comprising biotinylated histone H3 or histone H4 positioned at a single, 601 nucleosome positioning site. By following biotin-labelled H3 or H4 onto newly synthesized DNA, the results show the random and dispersed segregation of nucleosomes in the SV40 system and the local segregation (within the 601 sequence) for nucleosomes in the X. laevis extracts. b | The chromatin occupancy after replication (ChOR-seq) technique was used to ascertain the densities of trimethylated histone H3 lysine 4 (H3K4me3) and H3K27me3 in prereplicative and postreplicative DNA in HeLa cells. A time course-dependent analysis of newly replicated chromatin shows the accurate recycling of H3K4me3 and H3K27me3 densities. c | Biotinylation of H3 in Saccharomyces cerevisiae was achieved by engineering yeast cells to contain a tetO locus within the gal10 gene and then inducing the transient expression of TetR–BirA before DNA replication. Following the fate of the biotinylated H3 shows local recycling of parental nucleosomes at the repressed tetO locus after DNA replication. d | Biotinylation of endogenous replication-dependent H3.1- and H3.2-containing nucleosomes was achieved in mouse embryonic stem cells through the transient expression of a nuclease-deactivated Cas9 (dCas9) fusion with BirA and use of distinct guide RNAs (gRNAs) for BirA targeting to endogenous transcriptionally active or repressed loci. The results show that biotinylated H3 nucleosomes from active chromatin domains are dispersed, whereas local segregation is observed for repressive chromatin domains. BAP, biotin acceptor peptide; EdU, 5-ethynyl-2′-deoxyuridine; MCM, minichromosome maintenance.

Fig. 3 |. Proposed mechanisms for the local segregation of parental nucleosomes in facultative late-replicating repressive chromatin.

Here, we illustrate possible mechanisms for how parental histone (H3–H4)₂ tetramers are transferred ahead of the replication fork in the context of chromatin structure and S phase progression (replication timing). For simplicity, the replisome is shown as all-encompassing for the active components of the mammalian replication fork. a | The anti-silencing function protein 1 (ASF1)– minichromosome maintenance complex component 2 (MCM2)–replisome-driven default model shows the MCM helicase unwinding double-stranded DNA and destabilizing nucleosomes ahead of the replisome. Parental (H3–H4)₂ tetramers are then handled by MCM2 and handed over to ASF1 for transfer to the replisome, resulting in the random segregation of (H3–H4)₂ to the leading and lagging strands. b | The model envisioning unknown histone chaperones (indicated by the question mark) shows the specific recruitment of an as yet to be identified factor during the mid to late S phase transition when facultative heterochromatin is beginning to duplicate. In this instance, parental (H3–H4)₂ tetramers are handed over to this putative factor by MCM2, facilitating the transfer of trimethylated histone H3 lysine 27 (H3K27me3)-containing (H3–H4)₂ tetramers behind the replication fork. The ‘read–write’ mechanism of Polycomb repressive complex 2 (PRC2) would then methylate newly synthesized nucleosomes (white histone tails) and re-establish a fully repressed chromatin domain in duplicated chromatin.

Fig. 4 |. Mechanisms for potential to acquire a new cellular identity.

Three scenarios are depicted in which parental nucleosome segregation may influence a distinct chromatin state in daughter cells. a | Asymmetrical nucleosome segregation. This scenario would result in the segregation of parental trimethylated histone H3 lysine 27 (H3K27me3)-containing (H3–H4)₂ tetramers towards the leading strand, while the lagging strand would contain only newly synthesized (H3–H4)₂ with relatively few modifications. b | Tetramer splitting model. This event would occur primarily on bivalent promoters with parental nucleosomes comprising the histone variant H3.3 having both H3K27me3 and H3K4me3 modifications. c | Genetic mutations perturbing heterochromatin. Mutations in genes encoding histone chaperones or H3 itself, such as those found in oncohistones, would result in the loss of local parental nucleosome segregation and/or that of newly synthesized H3–H4 deposition. This might then lead to the loss of heterochromatin in daughter cells and a more accessible chromatin, as observed in cancer cells.