Chromatin is the Same in a Relative Way (But You’re Older)

To understand how different cells arise from the same genome, it is essential to determine the structure-function relationships in chromatin, the gargantuan DNA, protein and RNA complex that is the storage matrix evolution selected for the genetic material. Once terminal differentiation is achieved, cells need to remember what cell type they are and also to remember certain life experiences—the substrate for this epigenetic memory is chromatin. Chromosomes in the interphase nucleus occupy discrete territories, forming preferential interactions within and between DNA molecules. Different regions of the nucleus exhibit distinct organizational patterns and accordant functional behavior,¹ such as gene silencing at the nuclear periphery and the formation of distinct sub-organelles, such as transcription factories, which are enriched in transcriptional machinery. Within chromosomes, large swaths of DNA are rendered accessible for protein binding and subsequent regulation by a complex, and mostly opaque, interplay between DNA sequence cues, histone variants (and their modification), ATP-dependent chromatin remodeling complexes, transcription factors, transcriptional machinery (e.g. RNA polymerases), chromatin structural proteins and perhaps non-coding RNAs (Figure 1). Histone post-translational modifications (a.k.a., ‘marks’), notably those on the soluble, unstructured termini which are exposed from the DNA-bound nucleosome core, have been correlated with various gene regulatory behaviors including active transcription, silenced chromatin (temporarily or semi-permanently), transcriptional enhancers (DNA regulatory regions that facilitate transcription by binding proteins), to name a few. Histone marks are coordinated by suites of proteins neologistically referred to as writers, erasers and readers, that add, remove and interpret these modifications.

Figure 1.

Factors controlling chromatin accessibility and transcription. Nucleosomes (two copies each of H2A, H2B, H3 and H4) package genomic DNA into chromatin which is further organized into higher order structures that restrict (heterochromatin) or promote (euchromatin) transcription. The various factors that have been associated with transition of chromatin between these states are indicated, including histone turnover, from Li et al. Non-nucleosome protein/RNA binding not shown.

Many of these scales of chromatin regulation have been examined in the cardiovascular system, including specifically in cardiomyocytes. A general consensus is that development from embryo to young adulthood is associated with greater restriction of transcriptome potential and a more rigid commitment to cardiovascular lineages, whereas diseases of the heart and vasculature can be induced or ameliorated by manipulation of chromatin-targeting proteins, yet disease is not associated with comprehensive reprogramming of chromatin architecture.^2,3

Li et al. in this issue⁴ have taken a novel approach to studying cardiomyocyte chromatin using a genetically encoded GFP-tagged histone H2B protein to examine the upgrading of old chromatin with new histones. Their studies rely on utilization by chromatin machinery of the tagged H2B as a building block of nucleosomes, thereby enabling determination of the rate at which nucleosomes are replaced in cardiac chromatin and to what extent this rate varies across the genome. First, the half-life of a histone (as measured by GFP-tagged histone H2B) and thus, they hypothesize, of the nucleosome (although it is possible different rates of exchange exist for different histone isoforms) is ~2 weeks, with most of this time presumably spent bound to chromatin. The median half life of proteins from the non-mitochondrial proteins in adult mouse heart was determined by mass spectrometry to be 6.48 days, with a 5th-95th percentile range of 1.25 days to 24.75 days (mitochondrial proteins are slightly longer lived; half life 15.4 days), rates that incidentally were increased ~25% in the setting of disease.⁵ Adult cardiomyocytes, which very rarely divide or proliferate⁶ and very rarely replicate their genomes,⁷ are an appealing model system to examine how chromatin is replenished in the absence of the wholesale genome repackaging events that accompany mitosis and cell division. Nucleosome recycling was not uniform across the myocyte epigenome. Instead, gene promoters (defined by proximity to a gene’s transcription start site), enhancers (defined by the enhancer-associated histone mark H3K27ac) and other regulatory regions (defined by H3K4me1 [possible enhancer] or H3K4me3 [active transcription]) exhibited a higher histone turnover rate than was observed in regions of the epigenome not occupied by these marks. Moreover, the rates of exchange at promoters and active histone marks were curiously similar to each other, and greater than at the silencing mark H3K27me3, suggesting that the machinery responsible for exchanging histones is agnostic to the histone mark responsible for opening chromatin. Likewise when transcription factor binding sites were examined: greater histone turnover was observed at cardiac specific transcription factors loci (GATA4, TBX3, TBX20 and NKX2–5) compared with pluripotency transcription factors, implying preferential access at these regions, although whether the cardiac transcription factors are actually bound at these sites during turnover is unclear (for the pluripotency factors, which are not expressed in mature cardiomyocytes, the answer is presumably no).

The investigators performed Gene Ontology analyses on the top 10,000 genes with greatest histone turnover rate, revealing these genes to be involved in a variety of interesting biological processes, some of which have obvious relevance to cardiomyocyte specific biology, and others that do not, raising the intriguing possibility that dynamics of histone exchange are either independent of cell type specific chromatin machinery to target them to genes unique to the cell type, and/or that such machinery does not exist.

Histone marks to not act in isolation. Coincident marks can confer distinct behavior (e.g. presence of both active and inhibitory marks, so-called bivalent regions) and chromatin readers, transcriptional machinery and/or chromatin remodelers bind to specific histone marks to regulate transcription or other genomic process. Extending their analyses into this realm, the investigators examined histone turnover rates at genes occupied by the EED subunit of the polycomb repressive complex (PRC), a multiprotein complex responsible for deposition of the silencing mark H3K27me3 and, as shown previously by these same investigators, for the balance of trimethylation (H3K27me3) or acetylation at this site (H3K27ac) in cardiomyocytes. Gain and loss of function studies targeting EED underscore the complex, nonlinear behavior of chromatin modifying complexes: depletion of EED did not change H3K27me3 but did increase H3K27ac, while at the same time decreasing histone turnover rate. Overexpression of EED on the other hand induced neither a change in total H3K27me3 or H3K27ac (as measured by western blot), nor a change in histone turnover rate. In either context (inactive H3K27me3 or active H3K27ac), the depletion of EED was associated with lower histone replacement. Lastly, the investigators tested another protein commonly localized to accessible chromatin, Brg1, which is an ATPase subunit of the BAF chromatin remodeling complex. As with EED, depletion of Brg1 decreased histone turnover rate, suggesting that actions to deplete proteins that bind open chromatin result in lower turnover rate of nucleosomes, potentially by shifting the equilibrium towards a less accessible conformation.

It will be fascinating to discover whether subcomplexes of BAF and or PRC exist, containing subunits that confer histone turnover activity, differentiating these complexes from the chromatin remodeling and gene silencing activities of these complexes, respectively, both of which, along with other chromatin modifying enzymes, have been implicating in histone turnover in previous studies in replicating and non-replicating cells.⁸ Also intriguing is the possibility that lineage specific transcription factors which, unlike the BAF and PRC complexes, have DNA sequence targeting motifs are responsible for targeting a subpopulation of these complexes to the right loci for timely histone turnover. That chromatin stays the same in a terminally differentiated cell, thereby maintaining the correct transcriptome, is a posteriori true. How this occurs is yet unknown.

Studies from murine neurons—also terminally differentiated—suggest that the histone turnover tools described in Li et al.⁴ may be useful in revealing novel aspects of cardiomyocyte physiology. Distinct histone variants cycled on/off neuronal chromatin at different rates (histone H3.3 was most frequently evicted compared with other variants), these rates are affected by the age of the organism and exposure of mice to conditions that promote synaptic plasticity and enhance learning and memory in turn stimulated increased histone H3.3 turnover.⁹ Do cardiomyocytes exchange different histone variants at different rates and do these rates change in the setting of physiological or pathological stimulation and with the age of the organism (or developmental stage)? Because all the studies in Li et al. were conducted in post-mitotic, non-replicating cells, it remains unknown to what extent the behaviors of histone turnover described in this study are in fact unique to non-replicating, terminally differentiated cells. And after the 2 weeks have got behind you and it’s time to replace your histones, how did the cell monitor this passing of time and how does it know which variants to deposit at which loci? Increasing evidence suggests that many factors correlate with transition of chromatin from accessible for transcription to inaccessible (Figure 1) and this investigation highlights the correlation of histone turnover with this chromatin feature. Yet one of the most important unanswered questions in the field of chromatin biology—regardless of cell type—is the temporal rules by which these factors change chromatin conformation and modulate transcription in response to stress stimulation, normal physiological function, development, and disease.