Insulated Neighborhoods: Structural and Functional Units of Mammalian Gene Control

Abstract

Understanding how transcriptional enhancers control over 20,000 protein-coding genes to maintain cell-type-specific gene expression programs in all human cells is a fundamental challenge in regulatory biology. Recent studies suggest that gene regulatory elements and their target genes generally occur within insulated neighborhoods, which are chromosomal loop structures formed by the interaction of two DNA sites bound by the CTCF protein and occupied by the cohesin complex. Here, we review evidence that insulated neighborhoods provide for specific enhancer-gene interactions, are essential for both normal gene activation and repression, form a chromosome scaffold that is largely preserved throughout development, and are perturbed by genetic and epigenetic factors in disease. Insulated neighborhoods are a powerful paradigm for gene control that provides new insights into development and disease.

Figure 1. The enhancer-gene specificity conundrum

A. Model of a genomic region encompassing an enhancer and two genes. The features that cause an enhancer to regulate only specific genes are still not fully understood, which we refer to as the enhancer-gene specificity conundrum B. Model of a genomic region encompassing an enhancer and two genes with the transcription factor CTCF bound in-between. CTCF is a component of enhancer-blocking insulators, but which CTCF-bound sites function as an insulator in vivo is still unclear.

Figure 2. Insulated neighborhoods

A. Hierarchy of chromosome structures: Chromosome territories, Topologically Associating Domains and insulated neighborhoods. Anchor refers to the CTCF-bound site interacting with another CTCF-bound sites, both co—bound by a cohesin ring. B. Features of insulated neighborhoods in human embryonic stem cells (ESCs). The values displayed for the size range and number of genes represent the middle 95% of the data range. C. Evidence for insulation of insulated neighborhoods: 97% Enhancer-gene interactions occur within insulated neighborhoods in human ESCs. D. Evidence for insulation of insulated neighborhoods: Deletion of insulated neighborhood anchors leads to gene misregulation. E. Evidence for insulation of insulated neighborhoods: Mutations of insulated neighborhood anchors in tumor cells leads to oncogene activation.

Figure 3. Insulated neighborhoods in development

Cell-specific enhancer-gene interactions occur within insulated neighborhoods that are generally maintained in different cell types. Left side displays a linear model of a genomic region encompassing a gene associated with cell type-specific enhancers, the right side displays the insulated neighborhood model of the locus.

Figure 4. Insulated neighborhoods are the mechanistic basis of TADs

A. Hi-C and cohesin ChIA-PET identify similar Topologically Associating Domains (TAD). Bars indicate the TADs identified using Hi-C and ChIA-PET in human ESCs at the genomic region whose co-ordinates are indicated in the bottom. B. Model of a TAD that consists of an insulated neighborhood C. Model of a TAD that consists of nested insulated neighborhoods D. Model of a TAD that consists of two insulated neighborhoods, nested within a TAD-spanning CTCF-CTCF loop

Figure 5. Relationships between insulated neighborhoods and other DNA loop models

DNA loops at the EYA1 genomic locus generated using three different types of chromatin contact data in lymphoblastoid cells. Displayed are the cohesin ChIA-PET interactions (Heidari et al., 2014) used to identify insulated neighborhoods, Hi-C data (Rao et al., 2014) used to identify “Loop Domains”, and CTCF ChIA-PET data (Tang et al., 2015) used to identify “CTCF contact domains”. Increased stringency filtering of the CTCF ChIA-PET data reveals a chromosome structure similar to the insulated neighborhoods and Loop Domain. The coordinates of the genomic region are displayed at the bottom.

Figure 6. Insulated neighborhoods at the IGF2/H19 and ß-globin locus

A. Insulated neighborhood model at the maternal and paternal alleles of the imprinted IGF2/H19 locus. On the maternal allele, CTCF binding at the imprint control region upstream of the H19 gene creates an insulated neighborhood around H19 and an enhancer, which prevents the enhancer from activating the IGF2 gene. On the paternal allele the imprint control region is methylated which leads to repression of the H19 gene, and prevention of CTCF binding. On this allele, a large insulated neighborhood is formed allowing the downstream enhancer to activate the IGF2 gene. Black lollipops indicate DNA methylation. The insulated neighborhood models are displayed on the right. B. Lack of methylation at the imprint control region upstream of H19 and the presence of the large insulated neighborhood on the maternal IGF2/H19 allele occur in patients with Beckwith-Wiedemann syndrome. C. Methylation at the imprint control region upstream of H19 and the presence of the small insulated neighborhood on the paternal IGF2/H19 allele occurs in patients with Silver-Russell syndrome. D. Insulated neighborhood model at the β-globin locus containing a cluster of globin genes and an upstream locus control region (LCR).

Figure 7. Insulated neighborhoods as a method to identify target genes of disease-associated enhancer variation

A. (Top) Assignment of an enhancer-associated Single Nucleotide Polymorphism (SNP) to a gene based on linear proximity. (Bottom) Assignment of a SNP to a gene based on the insulated neighborhood model. B. Model of the insulated neighborhood organization at the FTO-IRX3-IRX5 locus. C. Model of the insulated neighborhood organization at the CDC123-CAMK1D locus.

Figure 8. Neighborhood perturbation and repair through site-specific DNA methylation

A. Targeting a dCas9-DNA-methyltransferase 3a/3l (Dnmt3a/3l) fusion to an insulated neighborhood anchor leads to DNA methylation, abrogation of CTCF binding and loss of neighborhood integrity. Black lollipops indicate DNA methylation. B. Targeting a dCas9-TET (Ten-eleven translocation) fusion to an aberrantly methylated insulated neighborhood anchor leads to DNA de-methylation, and restoration of neighborhood integrity.