| 520 | |
▼a The structural aspects of biological systems are tightly paired with their functions. This understanding has been demonstrated over a broad range of length scales, spanning the ultrastructure of a cell to the macroscopic architecture of organs. Connecting structure and function relies on the integration of physical and biological sciences to analyze the fundamental arrangement and cooperation of specific sets of biomolecules, frequently at the nanoscale. One crucial area within this nano regime is the study of the folding of chromatin and its relation to critical biological processes such as transcription, replication, differentiation, DNA repair, and apoptosis. In molecular biology, angstrom resolution imaging through cryo-electron microscopy has been routinely performed to solve the structure of virus, protein, and macromolecules. On the other end, in cellular biology, the conventional electron microscopy has been utilized to provide ultrastructure for organelles for many decades. For epigenomics, the unique challenges in quantifying the chromatin organization are twofold: 1. The demand for ultra-high resolution and large imaging volume. 2. The demand for identifying structures based on their molecular functions. While the fundamental components of chromatin, the nucleotides, are only ~ 1 nm (the DNA double-helix is 2 nm across), they self-organize into a massive hierarchical polymer complex, the chromosomes, which are distributed over a distance of tens of microns within the cell nucleus. For most ultra-thin TEM samples, the contrast in electron micrographs originates from the phase shift of electrons passing through the specimen and inerfering on the detector. On the other hand, for thick specimens like chromatin, large-angle scattering of electrons passing through the biological sample dominates the image contrast, and the image intensity reflects primarily the mass-thickness distribution. Although it is possible to identify specific organelles through morphological information such as mitochondria, it is difficult, if not impossible, to differentiate different types of nucleic acids from electron micrographs, let alone genes with different transcription states. Within this work, we introduce multiple methods specifically designed to overcome the issues in adapting electron microscopy in chromatin imaging. In Chapter 1, we review needs in chromatin characterization and the major advances in electron microscopy over the last few years, as well as outstanding challenges in comprehensive quantification of genome organization. We highlight the demand for novel imaging methods for ultra-thick, low-contrast, and beam sensitive samples, such as chromatin. Expanding on this, Chapter 2 discusses the practical implementation of a label-free three-dimensional tomography reconstruction of a whole mammalian cell. Chapter 3 further advances this topic by introducing molecular specific labeling into electron tomography reconstruction. Combing previously reported DNA labeling ChromEM with quantitative high angle annular dark field (HAADF) imaging mode in the scanning transmission electron microscopy (STEM) for thick samples, the hybrid method, ChromSTEM, effectively kills two birds with one stone. The work shifts gears slightly in Chapter 4 for introducing a shortcut to obtain the label-free chromatin distribution for the whole cell statistically instead of deterministically, in comparison with Chapter 2. Chapter 5 will discuss the potential of utilizing novel sparse-sampling and inpainting to reduce beam damage in 3D tomography reconstruction even further. Finally, in Chapter 6, we will discuss the potentials of a consolidated nanoimaging platform featuring Spectroscopic Intrinsic Contrast photon-localization Nanocsopy (SICLON), 3D Spectroscopic Photon-Localization Microscopy (SPLM), and ChromSTEM. Future directions and potential expansions on the preliminary work is discussed in this thesis. The consistent theme of this work is the development and adaptation of advanced microscopy in quantifying the genome organization over a broad range of length scales, particularly with an eye towards reducing the time or integrated electron dose for critical information. In many cases, we find that vast new information can be obtained with simple innovations in conventional electron microscopy, and the future of epigenomics will likely be at the interface of computation, microscopy, and sequencing studies. |