Marenduzzo, Davide

Evans, Martin

Brackley, Chris

Morozov, Alexander

Chiang, Chi Hang Michael

2022-03-23T11:15:01Z

2022-03-23T11:15:01Z

2022-03-23

In eukaryotes, chromosomes reside in the crowded environment of the cell nucleus. Understanding the principles or mechanisms governing the three-dimensional (3D) folding of chromosomes and the role that this plays in genome function and disease has been a long-standing challenge in molecular biology and biophysics. Despite recent advances in experimental technologies that can probe chromatin structure at unprecedented resolutions, our knowledge of these principles remains incomplete. In this thesis, I model chromatin as a coarse-grained polymer and perform molecular dynamics simulations to investigate the mechanisms shaping genome organisation in several contexts. I first study the interplay between chromatin folding and the development of epigenetic patterns using a “recolourable” polymer model. Here, each segment possesses a histone mark or colour that can be updated by “writer” proteins, while “reader” proteins mediate 3D folding by bridging segments with similar marks. Coupling the action of readers and writers leads to the spreading of epigenetic information, facilitated by a change in chromatin conformation. By introducing “genomic bookmarks”, factors which associate with chromatin and recruit specific readers and writers, the model produces stable yet plastic epigenetic domains that can be maintained faithfully across replication events, and removal of bookmarks destabilises these domains. Remarkably, with bookmarking the model can reproduce the profile of Polycomb associated modifications along a whole chromosome arm in Drosophila. I then study the mechanisms driving chromatin reorganisation in cellular senescence, a pathological condition in which cells permanently exit from the cell cycle. I consider a minimal model to dissect the role of heterochromatin- and lamina-mediated interactions in oncogene-induced senescence. By varying these two ingredients alone, the model recapitulates typical organisations observed in growing and senescent cells. It demonstrates that the difference in the locality of chromatin interactions in these different conditions can be explained by polymeric phase transitions. It also shows that lamina-associated domains are highly stochastic, as observed in experiments. Crucially, the model offers a biophysical mechanism for the metastability of senescent phenotypes and may explain why it is challenging for senescent cells to return to the growing condition. Finally, I employ the highly predictive heteromorphic polymer (HiP-HoP) model to examine the elusive link between the spatial interactions of gene regulatory elements and transcription by building a compendium of simulated 3D structures of individual genes within the human genome. The model predicts the frequent associations, interaction topologies, and transcription probability for each regulatory element. It shows that the number of associations correlates significantly with transcription, consistent with the picture of transcription hubs or factories. Interestingly, it indicates that loop extrusion activity is related to the transcriptional variability of a gene across a population of cells. Overall, this pan-genomic analysis offers new insight into the connection between chromatin structure and function.

https://hdl.handle.net/1842/38795

http://dx.doi.org/10.7488/era/2049

en

The University of Edinburgh

Molecular dynamics simulations on the principles governing chromosome organisation

Thesis or Dissertation

Doctoral

PhD Doctor of Philosophy