Dynamics of higher order chromatin structures

Abstract

During the last few decades, the intensive focus on microscopy observations and genome sequencing analyses has proved that the genomic DNA is packaged in the non-random manner in the nucleus of interphase cells. Accumulated evidence have thus documented that the chromatin organization in 3D plays key roles in central biological processes, such as transcription, replication and DNA repair. In the interphase nucleus, each chromosome is expanded and organized in a manner depending on structural hallmarks of the nucleus. Thus, repressed domains localize to the nuclear periphery to form lamina associated domains (LADs) or large organized chromatin K9 modifications (LOCKs). In addition, prevalent chromatin interactions can be formed from same chromosome (in cis) or different chromosome (in trans). It is still not clear how such dynamic interactions between chromatin fibers control the expressivity of the genome and to what extent these depend on epigenetic chromatin states.

The study in this thesis had focused on the dynamics of higher order chromatin structures, particularly on the relationship between the dynamics of chromatin structure and chromatin states. Since the resolution of current single cell techniques in the chromatin organization research, such as DNA FISH and immuno-staining, are limited by the resolution of the microscopy, we invented a new in situ single cell technique termed ChrISP (paper I). Using this technique we could detect chromatin proximities with a resolution less than 17nm even though the analysis was implemented using the low resolution confocal microscope.

In paper II, the scope of the ChrISP technique was extended to include an analysis of chromatin states within a single chromosome in a single cell to document that compacted chromatin at the nuclear periphery depends on the H3K9me2 mark that impinges on the nuclear periphery in finger-like structures. Moreover, upon the removal of these marks the rest of the chromosome showed signs of compaction, potentially related to chromosome condensation. These results are consistent with the interpretation that the H3K9me2 mark regulates pleiotropic features of higher order chromatin structure.

In paper III, we had used the view point of a single locus to explore the dynamics of chromatin interactions in developmental window using the circular chromatin conformation capture (4C) technique. The resulting inter-chromosomal network connected, surprisingly, both active and repressive chromatin domains involving LADs. Moreover, this network depended on the circadian recruitment of active chromatin hubs to the repressed chromatin structures at the nuclear periphery mediated by the physical proximities between CTCF and PARP1. This circadian pattern was required to attenuate transcription of the active chromatin hubs in a rhythmic manner.

In summary, a new high-resolution technique termed ChrISP was invented in this thesis to enable quantitative analyses of dynamic of higher order chromatin structures. This technique could, moreover, be used to visualize specific chromatin marks, notably H3K9me2, within a specific chromosome in relation to structural hallmarks of the nucleus within a single cell. The compact chromatin structure thus identified was discovered to transiently harbor active chromatin hubs, which was recruited to the nuclear periphery in oscillating manner. We show that this feature likely underlies the attenuation of genes under circadian control. These findings open new perspectives to understand the function of dynamics of higher order chromatin structure.