Becker et al. (2016). H3K9me3-dependent chromatin: barrier to cell fate changes. Trends in Genetics. 32(1): 29-41.

Models of Developmental Gene Silencing

• Genetic material in the nucleus divided into 2 categories:
  • Euchromatin: DNA with relatively low density, high gene transcription rates
  • Heterochromatin: regions of the chromosome that are compact, transcriptionally repressed
    • Can be constitutive (present in all cell types, phases of cell cycle) or facultative (repression temporally specific or cell-specific)
  •  Large proportion of genome has repeat-rich sequences
    • Risk to genome integrity due to possible recombination, duplication
    • Utility in keeping regions silent (constitutive heterochromatin)
    • Repeat-rich heterochromatin marked by H3K9 methylation (di- and tri-methylation)
    • Mammals: methylation catalyzed by 5 members of the SET-domain containing methyltransferase family

• Heterochromatin protein 1: HP1, three isoforms in mammals
  • Can self-oligomerize, recruit repressive proteins to modify histones
  • Contributes to compaction and spread of heterochromatin
  • Binds H3K9me2/3 via its chromodomain
• Methyltransferases that deposit H3K9me2/3 required to establish hypermethylation at CpGs, low-level histone actylation
  • Two characteristics of heterochromatin
•  H3K27me3: methylation of lysine 27 at histone 3, catalyzed by PRC2 (Polycomb repressor complex)
  • Facilitates facultative silencing in cell-type specific repression
  • Especially present at lineage-specifying TF genes eg. Hox genes
  • H3K27me3 marked promoters are still able to be bound by general TFs, paused RNAP
• H3K9me3 involved in cell type-specific regulation of facultative heterochromatin
  • in differentiated cells, form large contiguous domains called patches
  • Expand in both number, size during differentiation
  • Span numerous genes repressed in cell type-specific manner
  • These domains largely exclusive of H3K27me3
• H3K9me3: repressive modification, also forms megabase-scale domains that include genes
  • called LOCKS (large organized chromatin K9 modifications)
  • Binding sites for the repressor protein CTCF detected at boundaries
  • Unsure if domains expand during differentiation
  • Important to silence lineage-inappropriate genes in differentiation

Heterochromatin: A Barrier to Cell Reprogramming

• Hallmarks of cell identity erased during reprogramming to iPSCs
  • Requires the reprogramming TFs to bind their targets in DNA
  • Reactivation of pluripotency genes, suggest that accessing heterochromatin important to the process
  • Only <0.1% of cells are successfully reprogrammed
• OSKM are the key reprogramming factors
  • All 4 open chromatin sites, but only OSK target sites containing nucleosomes w/o histone marks
  • This makes them pioneer factors
• DBRs: differentially bound regions, megabase-scale chromatin regions in which none of the 4 factors can target DNA in fibroblasts
  • Same domains bound by OSKM in pluripotent cells
  • Overlap with domains enriched for H3K9me3 in fibroblasts but not ESCs
  • Knockdown of SUV39H112 increases Sox2, Oct4 binding
  • Encode diverse genes and elements, including TFs essential to pluripotency
  • Pluripotency genes seem to be more refractory to activation
  • Majority of genome regions found to have altered non-CpG methylation in iPSCs vs. ESCs are DBRs
  • Some H3K9me3 domains stay in iPSCs – indicate incomplete reversion to ESC state
  • H3K9me3 removal may help increase reprogramming efficiency
  • Knockdown of SUV39H1.H2 led to increased iPSC colony formation
  • Also seen with other H3K9 methyltransferases, unclear which is more responsible for stabilizing the differentiated state
• Other factors/components of repressive chromatin acts outside DBRs
  • Demethylation of H3K9me3 needed for reprogramming via Utx
  • Repressive histone variant macroH2A inhibits reprogramming
  • H3K27me3 methyltransferase EzH2 needed for iPSC reprogramming
  • Need deposition of H3K27me3 and removal of H3K9me3 simultaneously
  • MBD3: component of the NuRD histone remodelling and deacetylase complex, mediator of gene silencing
    • Knockdown leads to improved iPSC programming
    • Stops reprogramming factor activity at the sites they already bind
    • May play a role in regulating H3K9me3 – hasn’t been explored

Paucity of Heterochromatin Defines Pluripotent State

• Reduction of inaccessible H3K9me3-marked heterochromatin fundamental hallmark of the pluripotent state
  • Chromatin of pluripotent cells shows increased rate of exchange at chromosomal proteins e.g. linker histones, HP1
    • This indicates a dynamic and accessible state
• Repetitive sequences: DNA sequences with high copy numbers, organized in adjacent near-identical units or dispersed throughout the genome
  • Includes retrotransposons, tandem repeats, satellite repeats, endogenous retroviruses
  • More common expression of these in ESCs, repressed in differentiated cells
  • Deletion of proteins that maintain chromosomal accessibility leads to impaired self-renewal of ESCs
  • Developmental plasticity of ESCs linked to chromatin accessibility
• Partially reprogrammed cells have highly compartmentalized heterochromatin structures
  • Contain dense chromatin fibres similar to diffrentiated cells
  • DNA methylation, H3K9me3 at specific pluripotency loci
  • Erasure of H3K9me3 can allow them to become full iPSCs
• SCNT: somatic cell nuclear transfer, uses factors of egg cytoplasm to restore pluripotency
  • H3K9me3 heterochromatin is a barrier to SCNT as well
  • RRRs – reprogramming resistant regions, silenced only in SCNT condition
  • Reducing H3K9me3 led to improved SCNT success
• Heterochromatin, especially H3K9me3, presents a barrier to reprogramming, regardless of the cell conversion methodology

H3K9me3 as a Regulator of Cell Fate In Vivo

• Patterns of H3K9me3 must be reorganized in cell fate transitions in development
  • Early embryo and terminal lineage maturation
  • TF networks ensure that H3K9me2/3 is regulator
  • Setdb1 occupies and represses genes that encode developmental regulators
    • Also acts as a corepressor of Oct4, suppressing trophoblast genes
• Implantation is followed by a progressive silencing of Oct3/4 and other pluripotency genes (Nanog, Stella, Rex1)
  • Deposition of H3K9me2, DNA methylation dependent on GLP and G9a occurs
  • G9a prevents Oct3/4 reactivation when differentiated ESCs returned to pluripotent state
  • Mutations in GLP that disrupt its ability to recognize H3K9me1 led to decreased dimethylation and a delay in pluripotency silencing, abnormal embryonic development
  • Cross-talk exists between H3K9me3 and H3K27me3
  • A direct role for H3K9me2/3 has been proposed in developmental control of gene expression
    • Reduced H3K9me2 occurs at LADs (lamina-associated domains), coupled to relative depletion of H3K27me3
  • G9a and GLP-null embryos have early lethality
  • SETB1 homozygous inactivation also embryonic lethal
  • Distinct lethal phenotypes in each case, shows they have different developmental contributions
• H3K9me3 contributes to lineage restriction in mature cell types
  • Shown by examining methylation status in Th1 vs. Th2 cells
  • Showed that H3K9me3/H3K27me3 have different roles in the two different lineages

Molecular Control of H3K9me3 Deposition

• Additional factors needed to explain site selectivity of H3K9me3
  • Sequence-specific TFs have been shown to recruit the heterochromatin machinery
  • KRAB-ZNFs: Kruppel-associated box zinc finger proteins, important for establishment of heterochromatin and have mostly lineage-specific expression
  • Noncoding RNA can function as a binding platform for heterochromatin establishment at specific positions
    • In yeast, heterochromatin dependent on RNAi pathway components, requires transcription of a locus to be silenced
    • Not well understood in mammals yet but believe RNA is involved in the H3K9me3 establishment

Concluding Remarks

• Large domains of H3K9me2/3 form in a cell-type specific fashion
• Machinery responsible for this is still mainly mysterious
• Need to understand initiation, delimitation of the H3K9me2/3 domains to develop more targeted ways to reduce H3K9me3-dependent heterochromatin
• RNAi knockdown of all 5 methyltransferases for H3K9 was shown to increase reprogramming efficiency
• Establishment of conditional knockouts of genes alone/in combination could provide insight
• Future studies of lineage-specific H3K9me2/3 domains should look at whether they impair transdifferentiation/conversion