Single-Molecule Mapping and Heterogeneous Dynamics of Epigenetic Modifications in Live Microbes

Single-Particle Tracking (SPT) in living cells informs the dynamics of target molecules and enables the investigation of their functions and interactions with other components in the cell. The quantitative analysis of SPT trajectories from living cells traditionally relies on the Brownian diffusion model. For molecules with homogenous dynamics or in well-studied biological systems whose biophysical mobility states are predictable, SPT analysis is robust and the mobility states of the molecules can be related to their biological functions in cells. However, for complex or poorly understood biological systems such as epigenetic modification systems, an objective SPT analysis method that quantify the heterogeneous dynamics of target molecules is keenly needed to investigate their functions and interactions in vivo. In this Dissertation, I develop a single-molecule tracking analysis framework with nonparametric Bayesian statistics and anomalous diffusion models to investigate epigenetic modifications in live bacterial and yeast cells. Chapter II presents a new SPT analysis method combining nonparametric Bayesian statistics and a supervised recurrent neural network. The method is named NOnparametric Bayesian Inference for Anomalous diffusion in Single-molecule tracking (NOBIAS). The performance of NOBIAS is validated with simulated datasets of heterogeneous dynamics, asymmetric diffusion, and a mixture of anomalous diffusion models. NOBIAS is also applied to experimental datasets from live cells and identifies anomalous diffusion and asymmetric diffusion in the systems. DNA methylation in bacterial cells is a marker for specific protein-DNA interactions. DnmA is a recently characterized DNA methyltransferases (MTase) in Bacillus subtilis, responsible for all detectable N6-methyladenosine DNA methylation. In Chapter III, I use single-molecule tracking and spatial mapping to study of DnmA in live Bacillus subtilis. The results show that DnmA is regulated by the DNA substrate and correlates with DNA replication and DNA-RNA hybrid cleavage. This work combines single-molecule imaging of DnmA and phage predation assays to identify that DnmA is functionally an orphan MTase regulating gene expression. Epigenetic modifications in eukaryotes regulate the chromatin structure and the gene expression level. Histone H3 lysine 9 methylation (H3K9me) is a conserved epigenetic marker for heterochromatin and gene silencing. Epigenetic modifications rely on writer, reader, and eraser proteins to establish, maintain and remove modifications. In Chapters IV and V, I use single-molecule tracking and nonparametric Bayesian statistical analysis to understand the behaviors of these modifiers in vivo. In Chapter IV, I focus on the H3K9me reader protein, Swi6, in the fission yeast cell. I present the dynamics of Swi6 following different perturbations including knockouts of related proteins and the engineering of the Swi6 protein itself. I map Swi6’s distinct mobility states onto their biological roles in living cells and show a high-specificity binding mechanism through Swi6 oligomerization. Chapter V presents the single-molecule dynamics of multiple H3K9me modifiers in fission yeast. By comparing the dynamics of these modifiers and centering on the two H3K9me reader proteins Swi6 and Chp2, I propose that chromatin plays an important role to reinforce interaction and complex assembly on its site rather than just an inserted platform for interaction. Through this dissertation, I show a powerful and informative methodology combining live-cell single-molecule tracking and advanced statistical inference. The dissertation provides quantitative analysis, detailed statistical models, and their application to epigenetic modifications in bacterial and yeast cells. This methodology is also applicable to any system where in vivo single-molecule tracking is feasible.PhDBiophysicsUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/176588/1/ziyuanc_1.pd