The Mediterranean Institute for Life Sciences
Split, Croatia

High resolution ray-traced model of a nucleosome, isolated on black.

A nucleosome is the basic unit of DNA packaging in eukaryotes, consisting of a segment of DNA wound in sequence around four histone protein cores.  This structure is often compared to thread wrapped around a spool.

Nucleosomes form the fundamental repeating units of eukaryotic chromatin, which is used to pack the large eukaryotic genomes into the nucleus while still ensuring appropriate access to it.  In mammalian cells approximately 2 m of linear DNA have to be packed into a nucleus of roughly 10 µm diameter.  

Nucleosomes are folded through a series of successively higher order structures to eventually form a chromosome; this both compacts DNA and creates an added layer of regulatory control, which ensures correct gene expression.

(Nucleosome - Wikipedia)


It has already been determined that silencing near tRNA genes requires nucleolar localisation through work with yeast, and now a follow up paper from David Engelke’s lab at the University of Michigan tracks the process ‒ tRNA gene mediated (tgm) silencing ‒ to the nucleosomes.

Elaborating on what makes this form of silencing distinct from other forms of transcriptional silencing or boundary element (cis-acting elements which also serve a structural role) function in yeast, the group discuss models for communication between the tRNA gene transcription complexes and local chromatin.

These RNA polymerase III transcription units are dispersed in high copy throughout nuclear genomes and act counter to RNA polymerase II transcription in their immediate vicinity on a chromosome. 

In some respects tgm silencing may be incompatible with other forms of silencing, since tRNA genes are able to block propagation of silenced chromatin states caused by other silencing mechanisms. It is not clear what the relationship is between tgm silencing and this “insulator” or “boundary element” function.

To test the possible involvement of nucleosomes in an unbiased manner, we screened a comprehensive library of alanine substitution mutations in the four core histones for the ability to release tgm silencing. Over 10% of the substitutions strongly release the silencing, with many residing in the histone H3 and H4 N-terminal tails. In addition, a large group of alleviating mutations defines a surface on the nucleosome disk that has not been implicated in other silencing forms. In light of these observations, we also tested additional mutations in a broad group of chromatin modification and remodeling enzymes, and find that the Rpd3 and Hos1 histone deacetylases, the Glc7 phosphatase, and the RSC nucleosome remodeling activity are also required for silencing.

As can be seen from the images above,

The identified residues map to three general areas of the nucleosome structure: the N-terminal tails of H3 and H4, other residues likely in near contact with the DNA near the nucleosome dyad, and a broad surface on the nucleosome disk face.

The H3 and H4 tails are particularly intensive regions of post-translational modifications that affect transcription and have the densest clusters of mutations that alleviate silencing. They emerge from the nucleosome core nearly opposite each other across the DNA helix and can interact with both the DNA and external proteins.

The left view is from the edge of the disc to optimise visualisation of residues in contact with DNA, while the view alongside it optimises visualisation of residues on the disc surface.

Light blue = DNA, Grey = histones (except residues affecting tgm silencing) : Blue = H2A, Cyan = H2B, Yellow = H3, Red = H4.

An “intriguing possibility" arises from the finding that Maf1, a direct target of TOR signaling which represses RNA pol. III, is required for tgm silencing.

When MAF1 is deleted, tRNA transcription increases, which we originally expected would intensify silencing near tRNA genes. When tgm silencing was instead abolished by maf1Δ , it indicated Maf1 was exerting an influence beyond its physical interaction with pol III and the tRNA gene-bound transcription factors. A link between Maf1 and RSC [chromatin remodelling complex] is thought to be mediated through the Rsc9 subunit, consistent with the involvement of the RSC complex in transcriptional regulation in response to TOR signaling.
Unlike boundary element function, which requires only bound TFIIIC, tgm silencing seems to require the presence of the full RNA polymerase III complex (pol III, TFIIIC, TFIIIB, 1), containing at least 25 polypeptides. Thus, it is not surprising that silencing near the tRNA genes involves interactions with activities not observed in boundary function.

Since both tRNA genes and some tRNA-derived short interspersed elements (SINEs) have been shown to have boundary element properties in larger eukaryotic genomes, it will be interesting to examine whether the other interactions with local chromatin structure are retained as well in these highly reiterated elements.

Good et al., “Silencing near tRNA genes is nucleosome-mediated and distinct from boundary element function”. Gene, vol 526, pp. 7‒15 (2013)

Nucleosome hopping and sliding kinetics determined from dynamics of single chromatin fibers in Xenopus egg extracts (2007)

Chromatin function in vivo is intimately connected with changes in its structure: a prime example is occlusion or exposure of regulatory sequences via repositioning of nucleosomes. Cell extracts used in concert with single-DNA micromanipulation can control and monitor these dynamics under in vivo-like conditions. We analyze a theory of the assembly–disassembly dynamics of chromatin fiber in such experiments, including effects of lateral nucleosome diffusion (“sliding”) and sequence positioning. Experimental data determine the force-dependent on- and off-rates as well as the nucleosome sliding diffusion rate. The resulting theory simply explains the very different nucleosome displacement…

  • Nucleosome displacement in transcription (2007)
    • "It is becoming increasingly clear that the eukaryotic transcriptional machinery is adapted to exploit the presence of nucleosomes in very sophisticated ways."

…kinetics observed in constant-force and constant-pulling velocity experiments. We also show that few-piconewton tensions comparable to those generated by polymerases and helicases drastically affect nucleosome positions in a sequence-dependent manner and that there is a long-lived structural “memory” of force-driven nucleosome rearrangement events.

Q: Briefly outline how methylation marks on histones are regulated (5% of marks). Describe, with emphasis on the experimental evidence, the molecular mechanisms that lead to altered histone methylation in leukaemia (95% of marks).

Methylation is a form of post-translational modification on histone proteins, established dynamically through their laying down by methyltransferases and removal by demethyltransferases (also known as demethylases) — as a general term I’ll refer to them here as modifier enzymes (modifiers).

Histone protein assemblies are strongly basic, such that the entire octamer hydrogen bonds and electrostatically interacts with the (negatively-charged) DNA phosphate backbone (as well as through non-polar contacts with the sugar groups). The ‘tails’ of histone proteins contact neighbouring nucleosomes, and thus fine modifications throughout the structure modulate both protein-protein and protein-DNA interactions. As such these epigenetic ‘marks’ modulate chromatin structure and function.

Histone modifiers can act locally, globally, or in restricted domains (as is the case for histone 3 lysine 27 trimethylation, H3K27me3). As well as modulating transcriptional activity (such as RNA polymerase II recruitment) epigenetic marks can modulate the activities of modifier enzymes: marks begetting marks.

Various such enzymes exist, with specificity for different residues on the core histone proteins’ N-terminal tails. Both acetylation and methylation are possible on the tails’ lysines, such that dynamic modifier activity-dependent competition takes place over the control of histone modification levels.

The complement of histone modifiers exhibits a striking redundancy (many modifiers hold specificity over the same residues), which can produce cell-type specific modification from e.g. increased levels of MLL1 over MLL2 in one tissue type than another.

Histone modifiers don’t act alone, but form multi-subunit complexes, such as the well-known DotCom (a complex housing H3K79 methylase DOT1).

Gratuitous histone modification map (because it's great) from my
previous post on histones, originally Huang et al. (2014) Snapshot:
Histone Modifications. Cell, 159(2), 458-458.

Traditional biochemical approaches may have attempted to monitor enzymatic acitivities of modifiers, but given the genome-wide scale and the resolution to which such activities must be measured at, an immunohistochemistry approach is used, chromatin immunoprecipitation (ChIP).

Leukaemia, a cancer of one bone marrow or blood, is the most common childhood cancer, in which uncontrolled proliferation of immature white blood cells leads to a drop in the number of normal white blood cells (leukocytes). More accurately, leukaemia comprises a family of related conditions, characterised primarily into acute/chronic, lymhoblastoid/myeloid/lymphoid, or mixed lineage.

Disruption of epigenetic marks is one hallmark of leukaemia, and has been taken as the causal factor in Hox (homeobox) gene misregulation (when these developmental genes are not downregulated, the “ectopic” HOX expression prevents maturation of the white blood cells).

One regulator of HOX expression is MLL1 (named for its implication in mixed lineage leukaemia), at chromosomal locus 11q23. MLL1 is frequently involved in chromosomal translocation events, where fusion of genes rearranges the domains in the resultant protein.

MLL’s common fusion partners are AF4, AF9, ENL, AF10 and AF6, all but the last of which belong to a family of serine/proline-rich, nuclear-localised, transcriptional activators.

Other protein families are trithorax group (TrxG) and polycomb group (PcG), which play antagonistic roles in epigenetic regulation (at H3K4 and H3K27, positively and negatively affecting gene expression, respectively) as reviewed by Krivtsov and Armstrong (2007).

The human MLL protein is a homologue of Drosophila melanogaster (fruit fly) trithorax, and frequently targeted for recurrent chromosomal translocations in human acute leukaemias characterised as acute myeloid (AML), acute lymphoblastic (ALL), or biphenotypic mixed lineage leukaemia (MLL).

MLL1’s C-terminal SET domain possesses a lysine methyltransferase activity (for H3K4). It’s separated from various protein- and DNA-interacting domains on MLL1’s C-terminal end by two common cleavage sites, such that translocations often result in separation of these two functions.

Through these inter-chromosomal rearrangements, the MLL N-terminus can be ‘fused’ to the C-terminus of over 50 different partners, losing the original H3K4 domain. A major group of MLL fusion partners interact with the DOT1 (recruit it), inducing its increased deposition of H3K79me2 marks.

MLL fusion proteins may activate a leukaemogenic gene expression program through various mechanisms, as Krivtsov and Armstrong detail in their 2007 review.

Focussing solely on histone methylation rather than the more complex issue of leukaemia’s aetiology, it can be seen that although the C-terminal SET domain is lost in MLL translocations (with it the H3K4 methylation activity), and most MLL translocations appear to produce increased Hox expression, on the whole the mechanism does not seem to be simply to abolish methylation.

The SET domain may be maintained through the new fusion partner, or as has been shown through β-galactosidase fusions, oligomericity of the fusion protein may be important (away from its usual usage as an inducible enzyme).

AF10 associates with DOT1L, as was shown by Okada and colleagues through coimmunoprecipitation with a polyclonal antibody from Raji cells, the first continuous human cell line of haematopoietic origin, which were observed to express each of the proteins highly (protein levels measured through both Northern blots and microarray analysis), with an IgG control coming up empty (confirming the Ab’s specificity).

Having confirmed human association, the experimentalists used a mouse murine model to investigate leukaemic transformation — myeloid progenitor bone marrow cells.

Okada et al. used a FLAG (epitope) tag to specifically pull down the fusion protein [with specific antibodies], after production of each gene construct through a retroviral expression system for fusion proteins (with AF10/a different methyltransferase/methyltransferase activity-defective mutants, or negative controls with just FLAG-tag, just FLAG-tag and MLL, or just FLAG-tag and hDOT1L), each along with a selectable marker (conferring resistance to an aminoglycoside antibiotic, G418.

Transformational capability of each construct was assessed through myeloid colony forming assays: normal untransformed [but virally transfected] cells would die over three rounds of plating, while those cells on the path to oncogenesis would display a hyperreplicative? phenotype and survive (forming visible colonies).

Including a (SAM domain) mutant of the DOT1 fusion partner confirmed it was this enzymatic activity responsible for the cancer phenotype. Similarly, histone methylation (rather than an intrinsic property of AF10) was incriminated through the induction of the self-same phenotype (albeit to a lesser degree) with an alternative lysine methyltransferase-competent fusion partner.

The HoxA9 locus was required for the above assay, and end-point PCR (placing primers along the breadth of the homeotic gene) subsequently demonstrated comparable enrichment (i.e. binding) of the MLL-DOT1 and -AF10 fusion proteins, further supporting AF10’s spatial association with DOT1. Immunofluorescence staining visualised precisely this, with matching foci of high expression.

Fusion of the MLL N-terminus to the DOT1L C-terminus (MLL-DOT1L) was shown to immortalise haematopoetic progenitors, a phenomenon not seen under expression of either MLL or DOT1L alone.

Comparing ChIP-quantified histone modifications in cells harbouring either of the MLL fusion protein against a blank vector-expression control along various HoxA9 amplicons (sampling the upstream, exonic, and downstream portions).

While little H3K79me2 was observed over the locus in the control, there was a degree of H3K4me2. Transgene introduction sparked H3K79 dimethylation over the MLL-DOT1 fusion protein HoxA9 amplicons, and through recruitment of DOT1, the MLL-AF10 fusion protein was effecting the same outcome.

Conversely, H3K4 dimethylation in untransformed cells fell in the presence of the fusion protein(s) — the MLL fusion protein exerted a dominance over the endogenous MLL, displacing it and all but abolishing its residual dimethylation.

Further experiments have shown similar in MLL-AF9 fusion

More recently, leukaemia stem cells (LSKs) have been used to examine the mechanistic underpinnings of histone methylation in this family of blood cancers.

Bernt et al. (2011) expressed an MLL-AF9 fusion protein in granulocyte progenitor cells (GMPs), subsequently termed leukaemic GMPs (L-GMPs).

ChIP-seq over the Hox gene cluster indicated restricted binding of the fusion protein in Hox clusters 9-10. The majority of the (120 or so) MLL-AF9 binding sites overlap with the far larger number of H3K79me2 sites, indicating most of the H3K79 dimethylation came from constitutive DOT1 activity. By contrast, genes not known to be targets of the fusion protein exhibit low levels of this epigenetic mark.

Comparison of control vector expression vs. fusion protein-expressing cell histone methylation by their ChIP-seq profiles shows with excellent resolution that while the broad pattern is indistinguishable, the highly expressing genes have strong peaks in H3K79 dimethylation around the transcription start sites (TSSs): target genes are also more broadly distributed on both sides of the TSS (i.e. would be so in leukaemic cells).

Although the above outlines the molecular mechanisms leading to altered histone methylation leukaemia, it is worth noting that the H3K79 mark is a prevalent, positive regulator of gene expression.

Presumably, loss of this mark would lead to a change of expression: removing the target as well as other genes in an inducible Cre recombination system* has shown that even though H3K79me2 is a widespread histone modification, only the MLL-AF9 target genes are sensitive to disruption of H3K79 methylation in leukaemia cells, and as such it acts a selective switch on their transcriptional capacity.

* I've gone into the molecular details of Cre previously, here

Krivtsov and Armstrong (2007) MLL translocations, histone modifications and leukaemia stem-cell development. Nature Reviews Cancer, 7(11), 823-833

More questions answered

TGIF: Biomedical weekend reading includes some cool papers

TGIF: Biomedical weekend reading includes some cool papers

I’m working on an R01, but I still try to find time to read a wide variety of papers. Below are the ones I’m hoping to get to this weekend.

Less Myc, longer “health span” Cell paper from Sedivy Lab.

ESC Histone H3.3 nucleosomal functions Epigenetics & Chromatin paper from Keji Zhao Lab.

Human PGC specification Cell paper from Jacob Hanna Lab.

SETDB1 & hnRNP K tango to silence ERVs in ESCs PLOS…

View On WordPress