Plank, J.L., Dean, A. (2014). Enhancer function: mechanistic and genome-wide insights. Mol Cell. 55(1): 5-14

• Enhancers: cis-acting regulatory elements that increase transcriptional output of target genes to affect cells in development
  • May reside far away from targets
• Models proposed for interaction with promoters:
  1. Enhancer looping
  2. Linking via protein complexes
  3. Combination of the two models
• Enhancer sequences contain binding sites for TFs, confers tissue specificity
  • Factors binding to enhancers and genes could stabilize chromatin loops
  • Tissue-specific proteins are critical to looping, identified in some model systems
    • e.g. beta globin locus, Th2 cytokine loci, human IFNy locus
• Looping interactions of chromatin-binding proteins (e.g. CTCF) facilitate gene-enhancer contacts
• Genome-wide studies help to place interactions in 3D context
  • Chromosome conformation capture technology: measures physical interaction frequencies between enhancers and targets
  • Refinements in 3C approach detects chromatin loops at multiple levels
  • Enhancer transcription into eRNA may function as part of gene activation, looping
  • Specific features of enhancers, targets, chromatin structures found

Long-Range Interactions with Target Genes

• TFs bind enhancers in clusters
• Exclude nucleosomes, contribute to the DNase I hypersensitivity
• Certain loci show specific enhancers binding to proteins required for looping
• Complex including GATA1 and cofactor FOC1 required for beta clobin LCR coding
• Erythroid cell transcriptional activation requires enhancer action
  • Complex includes LBD1, TAL1, LMO2
  • Dimerization domain of LBD1 underlies enhancer-gene proximity
  • When linked to the beta globin promoter, was capable of driving loop formation and partial activation of transcription
  • Large cohort of erythroid genes are activated by the LBD1 complex
• CTCF/cohesin binding elements and enhancers bind T-bet, a lineage-specific Th1 factor in the IFNy locus
  • shows proteins can participate directly in enhancer-gene looping
  • CTCF promotes a Th1 specific IFNy locus looped conformation
  • Need to join both CTCF and enhancer sites to activate IFNy transcription
  • CTCF: transcriptional repressor, creates boundaries between topologically associated domains in chromosomes, facilitates interactions b/w/ transcription regulation sequences
• During cell division, enhancer-gene interactions are disassembled
  • FOXA1 (lineage factor for hepatoma cells), GATA1 (key erythroid cell lineage factor) remain associated with chromosomal sites containing enhancers

Enhancer Loops and Transcriptional Activation

• Enhancers function primarily to promote increased transcriptional output
  • Interaction with promoters may involve transcriptional machinery components
  • Mediator occupies the enhancers of many ESC genes, pluripotency factors
  • Links promoters to Pol II at target promoters thru direct interaction with cohesin
• Enhancers may be involved in the initiation of transcription
  • Some function to release Pol II in order to allow elongation
  • Loop to target promoters, permit activation of the P-TEFb complex
  • PTEF-b: complex required for the release of Pol II into the elongation stage
• Study showed enhancer-promoter gene interactions enriched for cell-specific genes
  • Gene families regulated by common TF were overexpressed
  • Overrepresentation of multigene complexes, linked to Pol II foci
  • Pluripotency genes in ESCs were also connected within one hub
• Shows enhancer looping key, maybe required for transcription activation
  • Formation of new enhancer loops precedes transcription at the beta-globin locus
  • Long-range enhancer-gene interactions may drive nuclear relocalization and clustering

Enhancer Loops within the Nucleus

• Genomic context: long range interactions between loci predominantly within TADs (topologically associating domains)
  • TADs largely conserved across range of cell types in development
  • Borders enriched for CTCF sites (also found within TADs)
  • Long range interactions within the same TAD more common than between TADs
  • Intra-TAD contacts may be involved in cell-specific transcriptional activation
  • Intra-TAD more variable among different cell types
• EPUs: enhancer-promoter units, clusters of coregulated enhancers and promoters
  • Super enhancers: large domains of up to 50 kb that contain clusters of individual enhancer elements, highly occupied by Mediator, pluripotency factors
  • Enhancers within these regions are associated with genes that encode cell identity regulators
  • Sub-TADs have also been identified, vary in tissue-specific manner
• CTCF/cohesin and Mediator play roles at different levels of long-range interactions contributing to TAD or sub-TAD organization
  • HoxA genes and enhancers occupy same TAD but are grouped into specific topological domains in the limb buds (where the genes are active)
  • Enhancer-gene long range interactions are strengthened by but do not depend on enhancer activity
• Are TADs functionally relevant for long-range enhancer activation of genes?
  • Deletion of a TAD border in Xist locus led to new ectopic contacts, long-range transcriptional misregulation
  • CTCF (not cohesin) knockdown in vitro led to increased inter-TAD interactions
  • Cohesin knockdown in post-mitotic thymocytes had no change in TAD organization
  • CTCF may help maintain TAD borders
• Is an enhancer able to fucntion outside of its normal TAD context?
  • Ectopic B-globin LCR re-established contact with the gene and increased transcription
  • Inter-TAD interactions may help enhancer function
  • HoxD cluster: some genes switch enhancer contacts from one TAD to another as gene expression changes across locus, close to the TAD border
  • There may therefore be a flexibility to the border of TADs
• Within TADs enhancers cluster together into topological domains, relevant gene targets, EPUs, sub-TADs
  • May all be different examples of same enhancer-dependent clustering phenomenon
• Enhancers, promoters must scan a limited nuclear area in order to make contact
  • Interruption of contact sites within a cluster of coregulated genes affects transcription of other interacting genes
  • Consistent w/ model that enhancer-gene clustering within TADs, association of TADs of similar character serve to nucleate transcription factors
  • Emphasizes role of enhancers and looping in the spatial organization of transcription in the nucleus

Temporal and Spatial Regulation of Enhancers

• Enhancers are progressively modified to activate transcriptional programs
  • Occurs specifically through acquisition of H3K27ac mark
  • “poised” enhancers in ESCs have p300, BRG1 occupancy, H3K4me1, low nucleosome density
  • Enhancers also have H3K27me3 mark, PRC2
• Later in development loss of PRC2 and H3K27me3, acquisition of H3K27ac and ability to activate gene expression
• “Inactive” state of ESC enhancers with high nucleosome density identified, have occupancy by ELL3
• ELL3: polymerase II elongation factor, may mark enhancers for subsequent activation
  • As ESCs differentiate, enhancers of pluripotent genes inactivated
  • Some mechanisms proposed:
    1. L-SID1 (histone H3K4/9 demethylase) removes H3K4me1
    2. PRC2 deposits H3K27me2 mark at ESC enhancers before the H3K27ac mark
    3. OTX2 transcription factor pioneers new enhancers, allows Oct4 to move to new site and activate new targets
       • Permits cells to exit their naive state
       • Occurs during transition between naive and “primed” cells
• Changes occurring genome wide within regulatory landscape in regulation
  • ~90,000 enhancers show specific tissue/stage-specific activity windows
  • DNase I hypersensitivity used as a proxy to measure enhancers, gave similar results
  • Enhancer usage varies among cell lineages, linked to lineage-determining TFs
• Enhancers in differentiated cells capable of responding to external stimuli
  • TFs can mark “inactive” enhancers in unstimulated differentiated cells, required for later activation of targets
  • Unclear how latent enhancers are recognized and activated
  • Also unclear how enhancer modifications poise/activate enhancers
  • HoxD had some enhancer-gene contacts prior to gene activation, so cannot say that activation depends on the establishment of loops to target promoters
    • May be necessary but not sufficient?

DNA Methylation as a Modulator of Enhancer Activity