Three-Dimensional Genome Architecture and the Evolution of Vertebrate Regulatory Innovations

Phenotypic diversity in animals arises largely from changes to cis-regulatory landscapes – enhancers, promoters, and their organization into topologically associating domains (TADs). This dissertation asks how such landscapes originated, how they evolved, and how they are deployed to specify cell types and tissues and argues that the architectural principles that generate regulatory innovation are the same ones that constrain genome evolution. I first establish the ancient origin of this architecture, showing that cohesin/CTCF-mediated TADs arose at the base of the deuterostomes ∼590 million years ago. Their emergence enabled a pronounced expansion of regulatory landscapes, plausibly contributing to the increased complexity of deuterostome body plans while simultaneously locking gene order in place, so that three-dimensional architecture acted at once as an evolutionary driver and a constraint. I then examine how the two rounds of whole-genome duplication in early vertebrates relieved this constraint: duplication was followed by a burst of genomic rearrangement preceding gene loss, with jawless and jawed vertebrates resolving the duplicated genome along distinct paths. Using the sea lamprey Petromyzon marinus as a reference jawless vertebrate, I show that conserved syntenic blocks correspond to groups of TADs repositioned as cohesive units, that ancestral three-dimensional organization is typically retained by only one duplicate while its paralogs diverge, and that a rearrangement that brings a gene to a new different cis-regulatory context can place this gene under new cell type-specific control through “enhancer adoption”, a mechanism that underlies widespread neo- and sub-functionalization in vertebrate genomes. I then turn to two vertebrate innovations. First, I develop spatially resolved single-cell transcriptomics of the developing mouse limb and resolve two parallel, deeply interconnected regulatory systems: one specifying cell type and the other spatial position. I show how both reorganize when a limb signaling center is perturbed. Finally, in human medullary thymic epithelial cells, I show that the promiscuous self-antigen ex-pression underlying central immune tolerance is driven by NF-κB co-opting tissue-specific enhancers that are largely derived from transposable elements and that genetic variation in these elements shapes heritable risk of autoimmune disease. Together, these studies trace cis-regulatory architecture from its origin and evolution to its deployment in development and immunity and converge on a single theme: the three-dimensional organization of the genome and the regulatory landscapes it contains are repeatedly remodeled and redeployed, linking genome evolution to the phenotypic and physiological diversity of animals.