3D or Not 3D: Shaping the Genome during Development

Abstract

One of the most fundamental questions in developmental biology is how one fertilized cell can give rise to a fully mature organism and how gene regulation governs this process. Precise spatiotemporal gene expression is required for development and is believed to be achieved through a complex interplay of sequence-specific information, epigenetic modifications, trans-acting factors, and chromatin folding. Here we review the role of chromatin folding during development, the mechanisms governing 3D genome organization, and how it is established in the embryo. Furthermore, we discuss recent advances and debated questions regarding the contribution of the 3D genome to gene regulation during organogenesis. Finally, we describe the mechanisms that can reshape the 3D genome, including disease-causing structural variations and the emerging view that transposable elements contribute to chromatin organization.

Figure 1.

Principles of 3D genome organization. (A) Schematic of the 3D genome organization inside the nucleus. (Left) At higher-order scales, the transcriptionally active and silenced chromatin segregates into A (red) and B (blue) compartments, respectively. DNA in the A compartment organizes around nuclear speckles. The B compartment localizes at the nuclear envelope and frequently overlaps with lamina-associated domains (LADs) and nucleolar-associated domains (NADs). (Right) At smaller genomic scales, chromatin folds into topologically associating domains (TADs). Chromatin loops inside a TAD are formed by the binding of CTCF and cohesin. Chromatin folding allow the interaction of regulatory elements and their target genes. (B) Loop extrusion model. Cohesin extrudes chromatin through its ring-like shape to form a loop. The loop grows in size until it reaches two convergent CTCFs, which block cohesin activity. This is a dynamic process that can be disassembled. (C) Schematic representation of a Hi-C map where A/B compartment (red/blue), loop, TAD, and sub-TAD can be visualized.

Figure 2.

Establishment of the 3D genome organization during early mammalian development. Following fertilization, embryos are subject to a dramatic epigenetic reprogramming during the preimplantation period. Transcriptional input in the zygote is exclusively maternal until zygotic genome activation (ZGA) occurs at the two-cell stage. The establishment of the 3D architecture in the early embryo is also dynamic: A/B compartments exist in the sperm but are then lost and reestablished in the embryo and TADs become defined only from the eight-cell stage. The two-component model suggests that parental-specific preformed domains associated with Polycomb exist transiently from the zygote to four-cell stage and TADs are then established progressively. (ICM) Inner cell mass, (TE) trophectoderm, (PrE) primitive endoderm. (Figure created with permission from modified images by Marius Walter.)

Figure 3.

3D genome organization and gene regulation. (A) Schematic of a hypothetical regulatory landscape of developmental genes along with a gene expression pattern in the mouse postimplantation embryos. Gene A and B are in the same topologically associating domain (TAD). Gene A is not expressed during development. Gene B is activated by its enhancers and is expressed in the developing limb from E10.5 to E13.5. Gene C is in the neighboring TAD and is activated by two different sets of tissue-specific enhancers in the brain and the developing limb. (B) Dynamic versus preformed chromatin folding. (Upper panel) Linear representation of the hypothetical TAD of gene B from A. (Left panel) Model of dynamic enhancer–promoter interaction showing two different chromatin conformations. Enhancer and promoter do not interact in cell type A where the gene is inactive, whereas their physical proximity mediates gene expression in cell type B. (Right panel) In the model of preformed enhancer–promoter communication, the chromatin conformation is the same in cell type A and B and is independent of gene expression.

Figure 4.

Consequences of structural variations (SVs) on the 3D genome. (A) Deletion including a topologically associating domain (TAD) boundary between the Epha4 and the Pax3 TAD leads to TAD fusion. As a consequence, Epha4 enhancers are in the same TAD as Pax3, which is expressed in an Epha4-like pattern. (B) Large inversion creates TADs reshuffling at the Epha4 locus and induces proximity of Epha4 enhancers and Wnt6, which is ectopically expressed in Epha4 territories. (C) Inter-TAD duplication at the Kcnj2-Sox9 locus creates a neo-TAD and relocalize Sox9 enhancers next to Kcnj2 gene, leading to ectopic expression. (D) Inversion at the Epha4 locus relocate Epha4 enhancers in a gene-dense region in the neighboring TAD leading to architectural stripes and gene misexpression. In A–D, the gray square represents the genomic region that is subjected to SV.

Figure 5.

Two mechanisms of transposable element (TE)-mediated formation of domain boundary. (A) Example of TE-derived CTCF participating in the formation of a species-specific loop. The insertion of a TE containing a CTCF site occurred in species 2 but not in species 1. This TE-derived CTCF form a loop with the existing CTCF, dividing a topologically associating domain (TAD) in two sub-TADs. As a consequence, the gene involved in this TAD is less expressed in species 2 than in species 1. (B) TE transcription drives the formation of a boundary that exists only in undifferentiated cells. In differentiated cells where the TE is inactive, this boundary is not present.