Thomas Cremer
3D Nuclear Topography of Active and Inactive Regulatory Sequences Studied with Super-Resolution Fluorescence Microscopy
The coordinated expression of genes requires the coordinated action of transcription regulatory elements (TREs) including non-transcribing sequences such as promoters, enhancers, insulators, silencers and locus control regions. Genome targets with active TREs show an increased sensitivity to DNAse I digestion, called DNAse I hypersensitive sites (DHS+) with an average size between 100-1000 bp. Genome wide maps of DHS+ sites have been published for diverse human cell types. Many such DHS+ targets in one cell type are not DNase I sensitive in another cell type (DHS-), indicating that regulatory sequences are not actively used there.
In the present study we investigated for the first time the 3D nuclear topography of active and inactive regulatory sequences. We demonstrate significant differences between the 3D topography of active and inactive targets. We have recently proposed a model for a functionally defined nuclear organization based on two co-aligned three-dimensional networks: an active and an inactive nuclear compartment (ANC and INC) (Cremer et al., 2015. FEBS Letters 589, 2931–2943).
Experimental evidence for this model shows that chromosome territories (CTs) are built up from chromatin domain clusters (CDCs), which form still higher networks pervading the nuclear space. Whereas the compacted chromatin core of CDCs, called the INC, is enriched in repressive histone marks, a peripheral peripheral layer of low density chromatin, called the perichromatin region (PR) is enriched in epigenetic marks for transcriptionally competent chromatin and represents the nuclear domain, where transcription, splicing, chromatin replication and DNA repair occur. The PR lines a contiguous channel system, the interchromatin compartment (IC), which starts at nuclear pores, permeates the nuclear space between the higher order chromatin network and serves a role in nuclear import and export functions. The IC carries nuclear bodies and splicing speckles and interacts with the PR. Accordingly, the PR together with the IC is called the ANC.
In line with the ANC-INC model we demonstrate that active regulatory sequences are exposed at the outer periphery of CDCs with loops penetrating into the IC. By contrast, inactive regulatory sequences are more embedded within the interior of CDCs, although still excluded from the most compact core.
​
​
Irina Solovei
The Major Chromatin Classes Blueprint the Nuclear Architecture
The spatial arrangement of chromatin is crucial for the regulation of nuclear processes. In particular, the existence of hierarchically organized chromatin domains, from TADs to A/B compartments, as well as the scaffolding of chromatin via LADs, provide evidence for the importance of eu- and heterochromatin segregation in the nucleus. However, the mechanisms of chromatin segregation and the relations between chromatin domains and the primary genomic sequence remain poorly understood.
We interrogated these questions by studying the spatial arrangement of small Mb-ranged chromosomal regions, consisting of eu- and heterochromatic subregions, in both con- and xenospecific backgrounds. For this purpose, we used retinal neurons and fibroblasts of mice carrying a human artificial chromosome. We analyzed the spatial intranuclear arrangement of two versions of the human artificial chromosome, circular and linear, and used for comparison an orthologous region from the respective endogenous mouse chromosome. Both the human and mouse regions included three subregions, each corresponding to a major chromatin class: gene-rich and SINE-rich euchromatin, gene-poor and LINE-rich heterochromatin, gene-depleted constitutive heterochromatin consisting of satellite DNA.
We show that segregation of the three major chromatin classes is highly autonomous and chromosome subregions with a size of 0.6 - 3 Mb are able to correctly locate within their own chromatin class irrespectively of the chromosomal context and xenospecific background. In contrast, establishment of nuclear lamina association patterns needs a broader chromosomal context, and in case of the human artificial chromosome is apparently influenced by its geometrical constraints.
Our data allow us to speculate that major chromatin classes, marked by certain repeat repertoires, provide a link between the sequence of the genome and its spatial distribution in mammalian nuclei, and thus blueprint the overall nuclear architecture, which is further specified and tuned by other epigenetic factors.
​
​
Tom Misteli
Deep Imaging to Probe Genome Organization and Function
The genome is one of the major physical entities in the eukaryotic cell and is spatially and temporally highly organized. The molecular mechanisms that determine genome architecture and its relationship to function are poorly understood. High-throughput imaging approaches are emerging as powerful tools to elucidate the cell biological properties of genomes and to link genome architecture to function at a single-cell level.
Deep Imaging methods are based on the development of high-capacity, high-precision automated microscopes which allow acquisition of large imaging datasets and the implementation of computational image analysis and data mining methods to quantitatively capture morphological phenotypes. Deep Imaging enables new experimental strategies for the study of the genome including visualization and analysis of rare events such as chromosome breaks and translocations, use of large-scale imaging-based screens to probe molecular mechanisms of genome organization and function in an unbiased fashion, and they allow mapping of the genome in 3D space.
These approaches are powerful tools to probe the cell biology of genomes and provide novel insights into genome architecture and function.
​
​
Denis Duboule
The Functions of TADs and Boundaries at the HoxD Locus
Sergei Nechaev
Spectra of Random Clusters in Highly Dissolved Solutions: Number-Theoretic Manifestation of Noise
Luca Giorgetti
Structural Organization of the Inactive X-Chromosome in the Mouse
X-chromosome inactivation (XCI) entails major reorganization of the X chromosome as it becomes silent and heterochromatic. During female development, XCI is triggered by up-regulation of the non-coding Xist RNA from one of the two X’s. Xist coats the chromosome in cis and induces silencing of almost all genes via its A-repeat region, although some genes (constitutive escapees) avoid silencing in most cell types, and others (facultative escapees) escape XCI only in specific contexts. A role for Xist in organizing the inactive X (Xi) has been proposed.
Recent chromosome conformation capture approaches revealed global loss of local structure on the Xi and formation of large mega-domains, separated by a region containing the DXZ4 macrosatellite. However the molecular architecture of the Xi, at silent and expressed regions remains far from understood.
We investigated the structure, chromatin accessibility and expression status of the mouse Xi using allele-specific Hi-C, ATAC-Seq and RNA-Seq in highly polymorphic, clonal neural progenitor cells (NPCs) and embryonic stem cells (ESCs). We demonstrate a critical role for the DXZ4-containing boundary and Xist in shaping the Xi structure: deletion of the boundary disrupts mega-domain formation, and induction of Xist RNA, with its A-repeat region, initiates formation of the boundary and induces loss of DNA accessibility. We also find that the Xi lacks active/inactive compartments and topologically associating domains (TADs), except around genes that escape XCI. Escapee gene clusters display TAD-like structures and retain DNA accessibility at promoter-proximal and CTCF binding sites. Furthermore, altered patterns of facultative escape in different NPC clones correlate with the presence of different TAD-like structures following XCI. These findings point to a key role for transcription and CTCF in the formation of TADs in the context of the Xi.
Angelo Rosa
Chromosome Organization and the Physics of Crumpled Polymers
​
Chromosome structure and dynamics make the objects of considerable experimental investigation. In the latest years, this approach has been paralleled by computer simulations of polymer models which provide a remarkable accurate description of chromosome behavior under various conditions and over wide ranges of length- and time-scales.
In this talk, I will discuss the analogy between chromosome conformations and the Physics of crumpled polymers in entangled solutions. In particular, I will show how one can exploit this analogy (1) to explain the essential observed behavior of chromosomes in cell nuclei during interphase and (2) to design a fast computational scheme for building model conformations of large chromosomes with different degrees of resolution.
Peter R. Cook
Transcription Factories: Genome Organization and Gene Regulation
We suggest transcription ‘factories’ are major organizers of the human genome during interphase. The nucleolus is the prototypic factory; it is a place where many rRNA genes are efficiently co-transcribed by local concentrations of RNA polymerase I. Analogous clusters of RNA polymerase II in nucleoplasmic factories make protein-coding transcripts. Then, a promoter is only likely to initiate if tethered near a factory containing appropriate factors. As motifs like enhancers, silencers, insulators, barriers, and boundaries are transcription units, they would work by tethering target promoters close to, or distant from, suitable factories; although we might name the motifs differently, they are all just transcription units influencing promoter-factory distance (and so initiation frequency).
We also describe a new force able to drive genome organization. Here, we use a minimal model in which bivalent red and green spheres (‘transcription factors’) bind to different cognate sites in runs of beads (‘chromatin’) to form molecular bridges that stabilize loops. In the absence of additional explicit forces, molecular dynamic simulations reveal these ‘factors’ bind to ‘chromatin’ and then (extraordinarily) spontaneously cluster into structures like factories’; red factors cluster with other red ones, green with green, but rarely red with green. Binding of just two ‘factors’ to active and inactive regions of human ‘chromosomes’ yields rosettes, topological domains, and contact maps like those seen using ‘Hi-C’. This unforeseen ‘bridging-induced attraction’ provides a robust, simple, and generic force able to organize interphase chromosomes at all scales.
Alexander Y. Grosberg
TBA
Cristina Cardoso
Elementary Units of DNA Replication and Repair: A Mirror of Chromosome Higher Order Structure?
Half a century ago the elucidation of the DNA double helix structure was quickly followed by the visualization of replicons in DNA fibers. To connect 1D DNA replication/repair information with whole cell 3D data in mammalian cells, we combined superresolution 3D structured illumination microscopy and time-lapse analysis of S-phase dynamics with DNA replication fiber and genome wide ChIP-seq analyses.
We found that the subnuclear replication structures can be optically resolved down to single replicons during all S-phase stages. This sets aside the conventional interpretation of nuclear replication foci as replication factories, i.e., the complex entities that process multiple clustered replicons. Further, our data suggest that S-phase dynamics is primarily dictated by chromatin folding and synthetic replisome complexes assemble on template DNA. Accordingly, individual replicons within the chromatin context represent the elementary units of 3D genome duplication and organization.
Combining experimental data and theoretical modeling we developed a minimal comprehensive model for DNA replication in mammalian cells based on stochastic initiation and domino-like DNA replication progression that can reproduce the observed temporal progression of genome replication and its spatial dynamics in cells. We have extended these analyses to the DNA damage response by integrating 3D-SIM data with genome-wide sequencing data and correlation with genomic features. We identified phospho-H2AX chromatin domains in the nanometer range and found that these nano-domains arrange preferentially into higher-order clustered structures resulting in discontinuously phosphorylated chromatin.
Our correlation analysis further indicates that chromatin structural and molecular determinants can be uncoupled. Finally, we define the elementary structural chromatin unit read by the DNA damage response.
Bas van Steensel
Architecture and Regulatory Functions of Lamina-Associated Domains
​
​
Olga Dudko
The Contraction of Space and Time in Genomic Interactions
Christoph Cremer
Localization Microscopy of Nuclear Structure: Imaging the Epigenetic Landscape at the Nanoscale
Models of functional nuclear genome structure have provided a variety of predictions on various length scales, from the micrometer to the nanometer range experimental tests at the individual cell level using far field light microscopy were limited by the conventional resolution limit of about 200 nm in the object plane and 600 nm along the optical axis (“Abbe/Rayleigh-limit”). These limits have now been overcome by various super-resolution fluorescence microscopy (SRM) methods, such as 4Pi-, Stimulated Emission Depletion (STED), Photoactivated Localization Microscopy (PALM), or Stochastic Optical Reconstruction Microscopy (STORM).
Here, we report on the quantitative imaging of the epigenetic nuclear landscape based on Spectral Precision Distance/Position Determination Microscopy (SPDM). SPDM, a variant of localization microscopy, makes use of conventional fluorescent proteins or single standard organic fluorophores in combination with standard (or only slightly modified) specimen preparation conditions, allowing to use the same laser frequency for both photoswitching and fluorescence read out. Presently, this Single Molecule Localization Microscopy (SMLM) approach allowed us to light-optically resolve nuclear structures down to few tens of nanometer in 3D, and to perform quantitative analyses of individual small chromatin domains; of the nanoscale distribution of histones, chromatin remodeling proteins, transcription, splicing and repair related factors, as well as newly replicated DNA. The experimental results support recent models of functional nuclear.
Dual-color SPDM was applied to monitor in mouse cardiomyocyte cell nuclei the environmental effects of ischemia on chromatin nanostructure. Short-term oxygen and nutritient deprivation (OND) induced a previously undescribed nuclear architecture, consisting of large, chromatin sparse voids interspersed between DNA-dense hollow helicoid structures of the order of 40 to 700 nm in dimension. The OND induced chromatin compaction was reversible, and upon restitution of normoxia and nutrients, chromatin transiently adopted a significantly more open nanostructure than in untreated cells, exemplifying the dynamic capacity of nuclear genome architecture to physically respond to environmental conditions. In another recent application, we combined dual-color localization microscopy with statistical evaluation methods to image at the nanoscale the epigenetic landscape of individual pachytene chromosomes of meiotic prophase I nuclei in female mouse oocytes, or to study nanoscale genome structure using oligonucleotide labeling.
These super-resolving light microscopic approaches open an avenue to study the epigenetic landscape directly on the individual cell level at unprecedented optical and structural resolution; to validate predictions by molecular high-throughput sequencing methods in real space coordinates; and to perform quantitative tests of numerical models of nuclear genome nanostructure down to the single molecule level.
Andrzej Stasiak
Transcription-Induced Supercoiling and TADs Formation
Musa Mhlanga
TBA
​
​
​
Marina Lusic
Genome Organization in Function of Viral Integration and Transcription
​
To replicate its genome and to complete its life cycle, retro-transcribed viral genome in the form of viral DNA (vDNA) enters the nucleus. Nuclear space, in which chromatin is organized to support different structural and functional aspects of the cell represent a big challenge for the incomi g viral genome, which needs to be integrated into cellular chromatin in order to continue the productive infection. Integration of the viral genome into host chromatin depends on the enzymatic activity of HIV-1 integrase and involves different cellular factors, which influence the selection of integration sites. The functional consequence of integration site selection relate to the viral transcription, which usually follows the integration event. However, not always is integration followed by transcription and viral genome can be silenced for long periods of time, generating latent reservoir of quiescent integrated HIV-1 DNA.
​
We recently showed that HIV-1 provirus integrates into the host cell genome in a non-random manner with a spatial preference for the active genes, which are localized in nuclear periphery. HIV-1 integration is determined by the three-dimensional architecture of chromatin in such a way that the virus avoids lamin-associated domains (LADs) whereas it largely prefers the transcriptionally active regions positioned proximally to the NPC. H3K9me2 mark, deposited by G9a histone methyltransferase, is particularly prominent in the nuclear periphery. Our previous result showed that silenced HIV-1 genome is marked by this mark and that treatment of cells with BIX01294, which blocks G9a, reactivates HIV-1 transcription.
We reasoned that treatment of cells with BIX01294 prior to HIV-1 infection would perturb the peripheral chromatin organization and could give us further insight about HIV-1 integration preferences in activated primary CD4+ T cells. BIX01294 treatment resulted in an overall increase in the number of integrated viral DNA copies per cell, as visualized by 3D DNA FISH; FISH signals also revealed that viral genome changed its position, from preferentially peripheral to evenly distributed in the nucleus. We next reasoned that changed nuclear distribution could reflect HIV-1 integration into different genomic sequences. We probed this by sequencing HIV-1 integration sites in cells treated with BIX01294 and found that integration patterns did not change significantly. In fact, localization of some of the prominent HIV-1 integration genes such as BACH2, STAT5, MKL2, NPLOC4 and NFATC3 in CD4+ T cells treated with BIX01294 also changed, and by becoming more central, reflected the new localization of HIV-1. Highly targeted HIV-1 genes aside from their peripheral nuclear localization also possess certain sequence specificities, which largely determine viral integration.
Our new data suggest that the underlying genomic sequence plays a pivotal role in proper targeting of replication competent HIV-1. We are further exploring the genomic sequences marked with H3K9me2 chromatin mark with a goal to further understand how these underlying sequences affect the fate of HIV-1 integration.
​
Gerd Blobel
Functions of Chromatin Looping
Mario Nicodemi
Hierarchical Folding of Chromosomes in Neuronal Differentiation and its Link to Epigenetics
Mammalian chromosomes fold into arrays of megabase-sized topologically associating domains (TADs). By combining Hi-C data and modeling techniques, we investigate TAD higher-order interactions through neuronal differentiation and find that they form a hierarchy of domains-within-domains (metaTADs) extending across genomic scales up to the range of entire chromosomes. TAD interactions are well captured by tree-like, hierarchical structures irrespective of cell type.
The structure of metaTAD trees correlates with genetic, epigenomic and expression features, and tree rearrangements during differentiation are linked to transcriptional state changes. Using polymer modelling, we show that hierarchical folding promotes efficient chromatin packaging without the loss of contact specificity, highlighting a role far beyond the simple need for packing efficiency.
Torsten Waldminghaus,
Building Synthetic Secondary Chromosomes to Understand Chromosome Maintenance in Bacteria
François Képès
Modelling and Engineering Genome Architecture
Shall we soon be able to design and achieve desired ensembles of DNA conformations in a live cell?
This talk will describe some steps towards analyzing, and subsequently synthesizing, features of the genome architecture, with a view to optimize genes co-regulation in transcription factories.
Amos Tanay
Single Cell and Single Molecule Chromosome Conformation Capture
Chromosome conformation capture experiments are used to study contact structure on scales ranging from chromosomal territories to intra-loop promoter-enhancer interactions.
Using single cell Hi-C and single molecule mapping of 3C ligation chains, we study the dynamics and of pairwise and higher order relationships affecting contacts across these scales. We show how contact landscapes are modulated by the mitotic and replication cycle, and evaluate scenarios involving stable chromosomal hubs in repressive contexts and transient chromosomal contacts in transcriptionally active domains. The data suggest that in order to understand how complex chromosomal conformation participates in determining and stabilizing complex transcriptional programs, such dynamics must be considered.
We motivate further work toward integration of dynamic chromosome confirmation and complex transcriptional states using single cell RNA-seq data defining precisely transcriptional niches in mouse embryos and analysis of key topological domains driving such niche definition.
Josée Dostie
Chromatin State and Architecture Guides the Function of a Long Non-Coding RNA
Thousands of long noncoding RNA (lncRNA) genes have been identified in mammalian genomes, and several have been shown to encode important regulators of gene expression. However, the mechanisms by which lncRNAs control transcription remain largely uncharacterized.
Here, we investigate and compare how HOTAIRM1 regulates the expression of HOXA genes in two cellular differentiation systems. We show that while HOTAIRM1 promotes the expression of proximal HOXA genes during all-trans retinoic acid (RA) induction in NB4, it activates HOXA1/2 but represses HOXA4/5/6 in NT2-D1 cells. We find that distinct forms of the HOTAIRM1 lncRNA are expressed in these cell types, and show they associate with either the UTX/MLL or PRC2 histone-modifying complexes to regulate the levels of activating H3K4me3 and silencing H3K27me3 marks, respectively.
Our study demonstrates that lncRNAs can have distinct functions in different cell types. While HOTAIRM1 simply enhances proximal HOXA gene expression in NB4, it contributes to their temporal collinear activation with RA in NT2-D1. We suggest that transcriptional regulation by a lncRNA can be guided by multiple factors including the type of variant expressed, the chromatin landscape surrounding both lncRNA and target genes, and by their physical proximity in three-dimensional space.
Daniel Jost
Epigenomics in 4D: A Functional Role for the Dynamic Coupling between Epigenome and Chromatin Organization
Cellular differentiation occurs during the development of multicellular organisms and leads to the formation of many different tissues where gene expression is modulated without modification of the genetic information. These modulations are in part encoded by chromatin-associated proteins or biochemical tags that are set down at the chromatin level directly on DNA or on histone tails. These markers are directly or indirectly involved in the local organization and structure of the chromatin fiber, and therefore may modulate the accessibility of DNA to transcription factors or enzymatic complexes, playing a fundamental role in the transcriptional regulation of gene expression.
Statistical analysis of the repartition of this epigenomic information along the chromosomes have shown that genomes of higher eukaryotes are linearly partitioned into domains of functionally distinct chromatin states. In particular, experimental evidence has shown that the pattern of chromatin markers along chromosomes is strongly correlated with the 3D chromatinorganization inside the nucleus. This suggests a coupling between epigenomic information and large-scale chromatin structure that could statistically quantified.
Recently, using polymer physics and numerical simulations, we showed that attractive interactions between loci of the same chromatin state might be the driving forces of the folding of chromatin inside the nucleus. In this study, we assumed that the epigenomic information pre-exists to the 3D organization. However, increasing number of experimental results suggests that chromatin marks are themselves highly dynamic during cell cycle or developmental stages and that 3D organization of chromatin might play a key role in the stabilization and function of chromatin markers.
We will describe our efforts to better understand the dynamical crosstalk between the epigenome and the 3D organization and we will illustrate the modularity of our framework in several biological contexts. In particular, we show that epigenomic-driven contacts and the formation of interacting compartments coupled to a reader-writer mechanism of epigenetic maintenance lead to a better and more robust control of epigenome, suggesting that 3D organization of chromosome plays a functional role at the epigenetic regulation level.
​
​
Pietro Liò
Chromatin Capture and Epigenetic Data Evidence Integration using Multi Layer and Hierarchical Block Matrices
Benjamin Audit
Structural Organization of Human Replication Timing Domains
Understanding how the DNA double helix is spatially and dynamically organized in the nucleus of eukaryotic cells and how this affects genome functions is one of the main challenges of cell biology.
In this seminar, I will illustrate that when combining concepts and methodologies coming from physics with waveletbased multiscale signal processing, we are able to extract original information about the structure/function relationship in the nucleus. I will review some recent analyses of genomewide epigenetic modification data, mean replication timing (MRT) profiles and chromosome conformation (HiC) data that have established some link between the spatio-temporal replication program, gene expression and chromatin structural domains.
This provides an analysis framework for the understanding of the epigenetically regulated global chromatin reorganization that underlies the loss of pluripotency during cell differentiation in human. The proposed view reconciles the dichotomic picture of early transcriptionally active and late heterochromatin constant timing regions that replicate by multiple rather synchronous origins in separated nuclear compartments of open and closed chromatins, with the model proposed by our group where Ushaped MRT domains are bordered by ‘‘master’’ replication origins from which a replication wave initiates and propagates toward the domain centre via a cascade of origin firing.
I will finally discuss different modes of chromatin folding depending on the cells' differentiation status.
​
Ana Pombo
Genome Architecture Mapping: A Spatial Approach to Map Chromatin Contacts
​
The organization of the genome in the nucleus and the interactions of genes with their regulatory elements are key features of transcriptional control and their disruption causes disease. Technologies based on chromosome conformation capture (3C) have profoundly expanded our understanding of the role of genome architecture in gene regulation. However, 3C-based techniques have important limitations, many of which arise from their reliance on digestion and ligation of the interacting DNA segments.
​
Genome Architecture Mapping (GAM) is a novel genome-wide method for measuring three-dimensional chromatin topology without ligation. GAM extracts spatial information by sequencing DNA from a large collection of thin nuclear sections and quantifying the frequency of locus co-segregation. We apply GAM to mouse embryonic stem cells and show that it independently identifies topologically associating domains (TADs) and compartments A/B.
Strikingly, exploration of the most prominent chromatin contacts using GAM in combination with a statistical model (SLICE) identifies specific chromatin contacts enriched for interactions between active genes and enhancers across very large genomic distances. GAM also reveals abundant three-way contacts genome-wide, especially between the enhancers most highly occupied by pluripotency transcription factors and highly transcribed genomic regions. These complex, multi-region contacts are most prominent for sequences further away from the nuclear lamina.
Our results highlight a previously inaccessible complexity in genome architecture and a major role for gene-expression specific contacts in organizing the genome of mammalian nuclei.
​
​
Jean-Marc Victor
Finite-Size Scaling Analysis of Super-Resolution Imaging and Simulations of Epigenetic Domains
Colin Semple
Elevated Rates of Substitution at Regulatory and Architectural Sites across Cancer Types
Regulatory regions of the genome are important players in cancer initiation and progression. We have examined the patterns of mutations accumulating at active transcription factor binding sites across many cancer types and in normal human populations.
We find strikingly high rates of mutation at active regulatory sites across different cancers, relative to matched control sequences. This excess of mutations disproportionately disrupts the bindingsites of particular factors, such as CTCF, and is likely to be driven by selectively neutral processes, such as the replication timing of the genomic regions concerned.
However, binding sites involved in specific chromatin structures suffer particularly high levels of mutation, suggesting widespread disruption of nuclear organisation in cancers.
​
​
François Spitz
Molecular Mechanisms Organizing and Regulating Long-Distance Relationships
In vertebrates, the regulatory elements that control gene expression can lie several hundreds kilobases away from the promoter they influence. Because of the distances involved, the activity of these elements is defined not only by their intrinsic regulatory potential, but also by their ability to transfer it to their target genes. Hence, their biological function is closely associated with the 3D-organization of the genome.
Using an in vivo enhancer sensor system, we showed that distant enhancers are not intrinsically locked on specific target genes, but instead operate throughout large genomic intervals. Remarkably, these regulatory domains overlap with the topologically-associating domains (TADs) recently highlighted by chromosomal conformation studies, suggesting a tight link between the structural and regulatory architecture of the genome.
To understand how TADs influence enhancer-promoter interactions, we have systematically dissected few large genomic loci with in vivo chromosomal engineering techniques. We found that some TAD boundaries appeared indeed to block enhancer-promoter interactions and therefore contribute to enhancer selectivity by regulating their range of action. Besides this insulating activity, we now also show that TADs promote long-distance functional interactions. Within the context of these highly self-interacting domains, the genomic distance separating enhancer and promoter does not seem to be a defining parameter. In contrast, when these topological compartments are disrupted, linear distance becomes a limiting factor for enhancer-promoter interactions, and enhancers act erratically.
Several complexes have been proposed to contribute to TAD formation, but their precise contribution has so far remained unclear. We found that preventing the loading of cohesin complexes on interphase chromosomes has a dramatic influence on the 3D-organisation of the genome, with a striking disappearance of most topological domains and alteration of gene expression. Our data show that these expression changes are not caused by reprogramming of the regulatory inputs themselves, but result mostly from the inability of enhancers to reach their normal target genes.
Altogether, our data demonstrate that the 3D-organisation of the genome mediated by cohesin complexes is essential to establish specific and stable regulatory interactions between distant elements, conferring precision and robustness to gene regulatory programs.
​
Marie-Odile Baudement
The HRS-Seq: A Novel Method for Genome-Wide Profiling of Nuclear Compartment-Associated Sequences
​
We are interested in how mammalian nuclear organization controls gene expression in physiological and pathological situations.
We have suggested that supranucleosomal genome organization in gene-rich domains is statistically constrained and displays on average an helical conformation. Functional folding would then be achieved by chromatin looping mediated by locus-specific factors and/or by recruitment into specific nuclear compartments that may confine specific chromatin contacts in order to facilitate genomic regulations. However, most nuclear compartments are difficult to isolate and methods that provide general overviews of these genomic regions are quite complex to handle.
We have developed the HRS-seq method, a novel straightforward genome-wide approach whereby HRS sequences (HRS=High-salt Recovered Sequences) associated to nuclear compartments are isolated from the rest of the genomic DNA upon high-salt treatments and subjected to high-throughput sequencing. In order to determine how nuclear compartments influence chromatin organization during cell differentiation, we used the HRS-seq method on three distinct cell types during differentiation of mouse Embryonic Stem Cells (ESC) into Neural Precursor Cells (NPC) and neurons (collab. with T. Bouschet/L. Journot, IGF, Montpellier, France) as well as in mouse liver cells. We thus obtained the first genome-wide profiling of HRS in the mouse.
We found that the HRS are strongly associated to highly expressed genes. Interestingly, gene ontology analyses show that they correlate with the cell fate. Among genes that are found in common to all cell types, throughout differentiation, we found histone genes, suggesting that the Histone Locus Bodies, a specific class of Cajal body, is retained in our HRS assays. We evidenced the presence of sequences known to be associated to other well known nuclear compartments such as the nucleolus (olfactory receptor genes), and also possibly the paraspeckles and nuclear speckles (Neat1 and Malat1 genes).
Furthermore, thanks to a cross-analysis with Hi-C data available in the literature (collab. with A. Cournac/J. Mozziconacci, UPMC, Paris, France), we have shown that the HRS display a high contact probability in the 3D space of the nucleus. They are also highly enriched in some specific repeat sequences like tRNA genes, which are known to spatially cluster within the nucleus.
Globally, these results allow us to validate our experimental approach and to demonstrate that the HRS-seq method is a powerful tool to explore the genomic composition of nuclear compartments and to understand the impact of this level of chromatin organization on gene regulation and cell fate determination.
​
​
Diego Cattoni
Nanoscale Structuration and Regulation of Chromatin Folding Revealed by Super-Resolution Imaging
​
Chromatin accessibility is regulated by DNA sequence and specific modifications (epigenetic marks) that define chromatin types or states. In recent years, a new level of hierarchical organization at the Mb and sub-Mb regions was uncovered (Topologically Associating Domains, TADs). In Drosophila, TADs display well-defined epigenetic marks that strongly correlate with epigenetic chromatin types.
Here, we employed multi-color super-resolution microscopy to study the nanoscale organization epigenetic domains in Drosophila. We found that active (H3K4me3) and inactive (H3K27me3) chromatin form domains of distinctive sizes displaying multi-scale organization. These domains are segregated from each other but show a high degree of interdigitation. The size and number of domains show a high degree of correlation with the one-dimensional epigenetic code defined from genome-wide methods.
Consistently, chromatin insulators preferentially localized at TADs borders located in the periphery of inactive chromatin domains, whereas they highly colocalized with active transcription sites. By genome wide and pair-wise labelling of TADs barriers we did not observe long range interactions between TADs borders, in contrast to previous observations in mammalian genomes.
The distances between consecutive and non-consecutive barriers increased following a power-law scaling, with active and inactive chromatin exhibiting distinct physical behaviours. Our data strongly suggest a model in which active and inactive TADs are dynamically segregated from each other with their spatial positioning intimately associated with the linear epigenetic code of the chromatin fiber and not necessarily regrouped according to their epigenetic state.
​
​
Francesco Ferrari
Comparison of Computational Methods for the Analysis of Hi-C Data
​
More than a decade after the completion of the human genome-sequencing project, it is clear that, while eukaryotic genomes encode genetic information in their linear sequence, higher order chromatin organization inside the nuclear space and sub-nuclear compartments has a relevant role in genome functionality. This includes large structural domains (known as TADs, Topologically Associated Domains, representing densely interacting regions separated by boundaries), as well as short-range chromatin loops, such as those connecting distal regulatory elements (enhancers) to their target genes. These loops bring in close physical proximity the transcriptional regulatory complexes bound at promoters and enhancers, which are crucial for cell-type-specific regulation of gene expression.
In the last decade, different techniques known as Chromosome Conformation Capture technologies have been developed to investigate nuclear structure/function relationships. The most promising of these techniques is Hi-C, which combines DNA proximity ligation with high-throughput sequencing in a genome-wide fashion to interrogate all genomic loci at once. Its advantage over the other assays is that the generated contact probability matrix allows the unbiased investigation of chromatin organization at multiple levels (i.e. from chromosome territories to short-range interactions). However, Hi-C potential is hampered by several factors affecting the data analysis.
Firstly, biases generated by the experimental protocol and intrinsic properties of the data must be accurately taken into account. Secondly, a higher resolution in defining interactions or domains requires higher sequencing depth; this, in turn, has an impact on the computational resources requested for the analysis. Finally, Hi-C is still evolving and changing fast both in terms of experimental protocols and of statistical methodologies to analyze the data. In particular, despite the increasing number of available computational pipelines, there is still no consensus on what is the best approach to analyze Hi-C data. The comparative assessment of Hi-C analysis pipeline is therefore necessary not just to identify the optimal pipeline, but also to propose guidelines and good practices such as quality controls and file formats to conveniently store results.
Here, we describe the comparison of different approaches for the identification of chromatin interactions and TADs from Hi-C data, quantifying the impact of algorithms, data resolution, and protocol variations. In particular, we present the benchmarking results obtained comparing 5 pipelines for the characterization of chromatin loops and 5 methods to identify TADs on 6 publicly available HiC datasets, comprising data from different species (Homo sapiens and Drosophila melanogaster) and cell lines, Hi-C protocols, and resolutions. From this analysis we derive practical information useful for experimentalists and bioinformaticians as well.
​
​
Chiara Lanzuolo
Impairment of LaminA/C-Polycomb Crosstalk Affects Genome Compartmentalization in Laminopathies
​
Laminopathies are rare and incurable human degenerative disorders due to a mutation of Lamin A/C gene with a large variety of clinical symptoms including skeletal muscular dystrophy and premature senescence. Although in the last years scientists have been accumulated a high number of evidence showing nuclear architecture alterations correlated with laminopathy pathogenesis, how different layers of genome organization evolve and influence each other during disease progression is essentially unknown.
Our results demonstrate a functional cross talk between Lamin A/C and key epigenetic repressors involved in the maintenance of cell identity, the Polycomb group of proteins (PcG). We have shown that Lamin A/C is evolutionarily required for PcG proteins nuclear compartmentalization and that Lamin A/C knock-down leads to PcG bodies disassembly and PcG protein dispersion. This causes detachment from chromatin and defects in PcGmediated higher order structures, thereby leading to impaired PcG repressive functions.
Here we show how the impairment of this crosstalk can influence genome and nuclear architectures affecting physiological processes such as muscle regeneration and cellular senescence ultimately leading to laminopathy progression.
​
​
Dario G. Lupiáñez
Understanding the Pathogenic Effects of Structural Variations on Chromatin Organization
​
Mammalian genomes are organized into megabase-scale topologically associated domains (TADs) that have been proposed to partition the genome into large regulatory units. We recently demonstrated that the disruption of TAD structures can cause rewiring of functional interactions between genes and distant-acting enhancers, resulting in pathogenic phenotypes. At the extended WNT6/IHH/EPHA4/PAX3 locus, distinct human and mouse limb malformations are caused by deletions, inversions or duplications with an effect on chromatin architecture. It remains unclear however, to what extend these structural variants remodel the preexisting TAD organization at the EPHA4 locus and originate novel domains of interaction.
To investigate this, we generated high resolution interaction profiles in developing limbs from mutant mice using Capture-HiC technology. Our results show that deletions including a boundary element result in a complete fusion of adjacent TADs originating a novel chromatin domain delimited by the boundaries located immediately up and downstream of the deleted region. Furthermore, a large inversion including a boundary element originates a profound chromatin reorganization between adjacent TADs resulting in the formation of two novel chromatin domains. Strikingly, these domains are physically isolated by the rearranged boundary element, which completely retains its functionality despite its inversion. In all the studied variants multiple interacting loops are de-novo established, mainly involving previously characterized intraTAD elements.
Our results advance in the understanding of how structural variations can remodel the chromatin landscape of the mammalian genome and originate aberrant gene patterns of expression that lead to disease.
​
​
Karen Lipkow
Cellular Systems Biology of Chromosome Dynamics
​
The questions of how genes are regulated remains fundamental even after many decades of intense study. Rather than just studying the linear, one-dimensional sequence of DNA to inform us about regulatory mechanisms, we can now investigate the complex 3-dimensional organisation of whole genomes. It has become clear that this organisation is non-random and highly dynamic.
To address new questions in genome architecture, we are taking a systems biology approach, combining the bioinformatic determination of chromatin states with quantitative experiments such as HiC and fluorescent microscopy, and dynamic, stochastic models of whole genome organisation.
Comparing these results with our experimental data, this has led us to understand how biophysical properties of the chromatin fibre lead to significant and biologically relevant self-organisation of the genome.
​
​
Vittore F. Scolari
Quantitative Assessment of Irreversible Crosslink Effect on Cromatin Structure
​
The structural investigation of large macro-molecular assemblies is an essential step towards understanding the molecular mechanisms at work in cells. Due to the very unstable and labile nature of these assemblies most assays rely on a fixation step, typically achieved using formaldehyde cross-linking, to freeze and capture contacts made by proteins and nucleic acids. Propelled by rapid technological advances such as ChIP-seq and Hi-C chromatin structure is today field of intense activity.
​
Here we use Hi-C which maps contacts between genomic regions both within and between chromosomes in order to quantify the chromatin distortion induced by irreversible cross-link. The analysis of the polymeric structure emerging from the contact maps show the presence of two different organizations at short and long distances. The large distance behaviour reflects the in-vivo structure of the chromosomes and can be interpreted in terms of classical equilibrium polymer model. On the other hand, the short distance behaviour depends on the concentration of the cross-linking agent and on exposure time and cannot be interpreted by an equilibrium dynamics.
By modeling the cross-linking effect as a polymer irreversible collapse we were able to quantitatively describe this short distance polymeric structure. We find that the collapse shows a dynamical exponent which is independent from the concentration and reminiscent of a glassy dynamics.