Generating Chromosome Geometries at the Single-Cell Level from Cryo-Electron Tomograms

Abstract

JCVI-syn3A is a genetically minimal bacterial cell consisting of 493 genes and 543 kbp derived from a gram-positive pathogen, Mycoplasma mycoides. Previous work established genome scale gene essentiality and proteomics data along with its essential metabolic network (Hutchison et al. Science 2016, Breuer et al. eLife 2019,) and a kinetic model of genetic information processing (Thornburg et al. Frontiers in Molecular Biosciences 2019), solved stochastically under the well-stirred approximation. Spatially-resolved kinetic models require cellular architecture including spatial distributions of ribosomes and the circular DNA. Based on cryo-electron tomography, Syn3A does not have a single, condensed nucleoid region. Instead the ribosomes and DNA appear to be nearly randomly distributed throughout the cells. To create DNA geometries for our spatial models, we introduce a method of generating self-avoiding circular chromosome configurations in a lattice model with a resolution of 11.8 bp per monomer on a 4 nm cubic lattice. Realizations of the chromosome geometries are constrained by the ribosome locations and cell membrane measured in the tomograms. Using ensembles of these constrained chromosome configurations, we predict Hi-C contact maps for individual cells at resolutions greater than 500 bp. The 543 kbp chromosome of JCVI-syn3A corresponds to 46188 effective monomers in the 4nm lattice representation. The initial configurations of the chromosomes are grown as self-avoiding polygons constrained by the cell boundaries and ribosome locations, before being equilibrated. During the equilibration, the chromosomes are modeled as a stiff heteropolymers. Polymer configurations are sampled from the canonical ensemble using a Metropolis-Hastings algorithm through a series of circularity-preserving moves. Initial comparisons of the computationally-generated contact maps to experimental Hi-C contact maps show good agreement and the chromosome model provides a framework to incorporate chromosomal interactions in the spatially-resolved kinetic modeling of whole JCVI-syn3A cells.