How to rule the nucleus: divide et impera

Introduction

There is growing evidence that spatial organization is the key to genome function. In addition to other epigenetic factors [1], the arrangement of chromatin within the nucleus is important for transcription regulation and for establishing and maintaining cellular identity during differentiation. Recent genome-wide molecular studies have revealed a variety of distinct chromatin units, or domains, that build up the interphase nucleus. Chromatin is subdivided into two compartments, A and B, of juxtaposed transcriptional activity and apparent spatial segregation with loci from the same compartment contacting each other frequently but avoiding contacts with loci from the other compartment [^{2 and 3••}]. Within compartments, chromatin is self-organized in topologically associated domains (TADs) [4 and 5], which are structurally conserved and serve as functional platforms for physical interactions between regulatory elements [6, 7 and 8]. Various other domains, such as lamina-associated domains (LADs) [^{9, 10•• and 11}], nucleolus-associated domains (NADs) [12] or pericentromere-associated domains (PADs) [13^•], have been implicated in organizing and anchoring the genome within the nucleus. In cycling cells, genomic regions with time-coordinated replication form replication domains [14, 15, 16 and 17].

Although the fine details of genome organization and its functional connections to transcription regulation have been thoroughly discussed in a plethora of recent reviews (e.g., this COCB issue, see also [17, 18, 19, 20, 21, 22, 23 and 24], it remains a challenge to integrate the genomics-based and microscopy-based views into a cohesive concept. Like stepping back from a pointillist painting, we zoom out to see the nucleus in the pattern of its many domains. First, we define euchromatin (EC) and heterochromatin (HC) domains, discuss how chromatin is spatially organized in the nucleus and propose a model in which the repeat repertoire, together with scaffolding structures, determine chromatin folding. Next, we expand on how EC and HC segregation impacts on chromosome topology and is compatible with the invariant subdivision of chromatin into TADs. Finally, we ask how chromatin folding is regulated during differentiation and development, emphasizing the differences between cycling and postmitotic cells.

Spatial arrangement of EC and HC in the nucleus

To achieve a functional nuclear architecture, the genome is segregated into active EC and inactive HC. EC is gene-rich, includes mostly housekeeping genes, and replicates early in S-phase. By contrast, HC is gene-poor, includes tissue-specific genes, and replicates late in S-phase (Figure 1a; see also [17]). Additionally, EC and HC are differentially marked by interspersed repetitive sequences, which account for up to 45% of the mammalian genome [25 and 26]. Short interspersed repetitive sequences (SINEs) reside mostly in gene-rich EC, whereas retrotransposon-related long elements (LINEs) and LTRs locate preferably in HC (Figure 1a). Analysis of mammalian genomes by Hi-C revealed that every chromosome can be subdivided into two sets of loci forming A and B compartments, within which loci preferentially interact but avoid interactions with the other compartment [^{2 and 3••}]. A-compartment and B-compartment closely match EC and HC regarding compaction, gene-richness, expression, replication timing and, consequently, also repeat repertoire (Figure 1a).