Living organisms are made of cells that are capable of responding to external signals by modifying their internal state and subsequently their extracellular context. Understanding the spatio-temporal dynamics of these complex interaction networks in biological systems is the subject of a field known as systems biology. To investigate these interactions (a necessary step before understanding or modeling them), one needs to develop means to control or interfere spatially and temporally with these processes and to monitor their responses on a fast timescale and with single-cell resolution. Among all the great tools to precisely perturb the dynamics of biological networks, light activation has been witnessed as an extremely powerful, non-invasive methodology, providing an accurate control over biomolecules’ activity at an unprecedented resolution. This dissertation exemplifies the advantages of using optical tools to investigate complex bioprocesses with studies in two different fields which are developmental biology and cancer biology. In the first project, we show how photo-isomerization of the 13-cis retinoic acid to the trans-isomer (and vice versa) could be used in studying retinoic acid’s role in hindbrain development during zebrafish embryogenesis. The second project employs photon uncaging to induce oncogenic expression, and models tumor evolution from single cells in the zebrafish. Developing and validating these tools in these two projects greatly facilities our understanding of several biological networks. Photo-isomerization of retinoic acid between all-trans and cis form helps illuminate retinoic acid’s unique spatial distribution and function on hindbrain development in a developing zebrafish embryo. And this again clarifies the long-standing controversy of retinoic acid acting as a morphogen in patterning vertebrate embryo. Photo-uncaging of caged cyclofen proves to be a novel and powerful tool to non-invasively manipulate tumor-associated gene expression. The work shown here could shed new light on cancer initiation and growth, and provide new tools for target validation and testing of novel anti-cancer drugs.

Plasmonic nano-objects have shown great potential in enhancing sensing, energy transfer and computing, and there has been much e↵ort to optimize plasmonic systems and exploit their field enhancement properties. Super-resolution imaging with quantum dots (QDs) is a promising method to probe plasmonic near-fields. However, due to the strong coupling between QDs and plasmons, this technique is hindered by the formation of distorted point spread functions (PSFs) and QD mislocalizations. Chapter 4 of this dissertation investigates the coupling between QDs and ‘L-shaped’ gold nanostructures, and demonstrates both theoretically and experimentally that this strong coupling can induce polarization- / wavelength-dependent changes to the apparent QD emission intensity, polarization and position. From the magnitude and direction of the PSF shift under emission polarization modulation, the coupling strength can be extracted, and the true PSF location can be back-calculated from tabulated theoretical and experimental values. This discovery helps to better apply super-resolution imaging techniques to detect the plasmonic near-fields.Besides using fluorescence intensity as the local-field intensity indicator, photophysical properties of the emitter (e.g. on-time ratio) have shown to be a great candidate as well. Super-resolution fluctuation imaging (SOFI) has great potential in extracting the photophysical properties of emitters with super-resolution. In chapter 5, I discuss an open-source, modular SOFI analysis package we built for both reconstructing super-resolved plasmonic near-fields and engaging the SOFI community with a wide range of applications. Chapter 6 demonstrates how we characterize the photophysical properties of a specific fluorescent protein suitable for SOFI analysis. Our work provides a practical method with higher precision for plasmonic near-field mapping, which benefits many applications like biosensing and optical quantum computing.

Visualizing neural activities is an important step towards studying and understanding the human brain. Since neural signals are transmitted through modulations of their membrane potentials, a membrane potential sensor capable of translating such signals into experimental observables is essential for recording neural signals. Moreover, the capability of observing membrane potential in the functional structures of a neuron, which can be as small as femto-liters in volume, is crucial for studying important brain functions such as memory. For this reason, we developed inorganic nanosensors based on semiconductor nanorods as membrane potential sensors. These nanorods, which operate via the quantum confined Stark effect, display large voltage sensitivities by changes in fluorescence intensity, spectra and lifetime, allowing non-invasive observation of minute fluctuations in the membrane of live cells. The extreme brightness of these nanorods also allow single particle recordings, which will enable studies of membrane potentials in tiny neuronal structure such as synapses. In this thesis, we will first introduce the background, review the existing membrane potential sensors and introduce the relevant literatures regarding solid-state membrane potential sensors. For our development, we will first describe the development of a spectrally-resolved microscope for measuring the spectra of single nanoparticles in Chapter 2. Then, the characterizations of the quantum confined Stark effect in a variety of nanoparticles will be described in Chapter 3. The temporal response, long-term stability and the capability of recording electric field modulation at 1 kHz using a single nanorod will be reported and discussed. In Chapter 4, we will demonstrate a surface functionalization approach utilizing designed alpha-helical peptides and zwitterionic ligands for facilitating insertion of the nanorods into lipid membranes. These functionalized nanorods were shown to spontaneously insert into cell membranes and report membrane potential in live cells. The detailed characterization of the functionalized nanorods in the membrane and the capability of recording the membrane potential using a single functionalized nanorod will be demonstrated in Chapter 5.

Super resolution Optical Fluctuation Imaging (SOFI) has been widely acknowledged and

advanced over the past years. Comparing to other extensively adopted super resolution

techniques such as PALM, STORM, STED and SIM, advantages of SOFI include compatibility

with different imaging platforms, suitability for a wide variety of probes, flexibility in imaging

conditions, and a user-controlled trade-off between spatial- and temporal- resolutions. SOFI

therefore holds great promise for ‘democratizing’ super resolution imaging for broad

applications by non-expert practitioners. The theoretical resolution enhancement of SOFI scales as the square root of the cumulant order n, and once combined with a post-processing deconvolution algorithm, the resolution enhancement factor increases up to n. In this dissertation I will discuss the fundamental challenges faced by high order SOFI applications including pixel intensity dynamic range expansion, associated artifacts, point-spread function (PSF) estimation, and deconvolution. Several approaches for solving these challenges will be presented, that together constitute what we dub as ‘SOFI-2.0’. The power of SOFI-2.0 will be demonstrated for focal-adhesion dynamics (at super resolution) in live cells.

RNA Polymerase (RNAP) is a crucial protein molecule that decodes the information stored in DNA as base pair sequences, through a process called transcription. The study of the mechanics and composition of RNAP in a cell is crucial to understanding the complex and arduous process of life. Most biological systems respond to a change by adding heterogeneity to its population states. One of the most effective ways of studying a biological system is to compare the population states of proteins as a function of time or chemical environment. Unfortunately, some of the subpopulations induced by external stimulation are too small and are hidden under ensemble averaging when using traditional methods like gel-electrophoresis. I have developed single-molecule methods capable of resolving subpopulations and monitoring the transition of individual molecules to its various function states. Using these methods we can (1) rapidly screen RNAP activity in a large reagent space and (2) map the transcription factor binding sites on DNA with ultrahigh accuracy (ó < 300base pairs) in the genomic scale with the ability to differentiate single cell variations.

The actin cytoskeleton is an essential component of the cell aiding in a range of biological functions from movement and maintaining the shape of the cell to transporting sub-cellular cargo. In our work, we have investigated lung cancer and pre-cancer cells to understand the importance of actin cytoskeleton in identifying and tracking sub-cellular processes. In lung cancer cells, we have identified a novel intermediate phenotype during Epithelial Mesenchymal Transition (EMT) with a unique cytoskeletal architecture based on the relative orientation of stress fibers. From localization of actin binding proteins (APBs) we have demonstrated that the intermediate and final state has predominantly different types of stress fibers. To quantify the cytoskeletal difference between these phenotypes, we have developed a method Statistical Parametrization of the Cell Cytoskeleton (SPOCC) that uses Orientational Order Parameter (OOP) as a figure of merit to track the progression of EMT. Our image quantification technique has improved throughput and is non-destructive compared to other existing techniques to track biological changes, such as mass-cytometry or RNA-sequencing. Due to the extensive inter-dependence of the cell cytoskeleton with its mechanical properties and in turn their invasiveness, we have correlated the OOP data with Young’s modulus derived from AFM experiments to further validate the intermediate nature of the novel phenotype. In pre-cancer cells, it has been reported that cells derived from non-cancerous epithelial tissue samples also demonstrate enhanced invasiveness similar to mesenchymal cells. Here we have delved into the cytoskeletal architecture of this high migratory cells that demonstrated almost mesenchymal-like stress fiber structure. Using the cytoskeleton as a marker, we have identified that the MAPK/ERK pathway is active in these high migratory cells and is likely responsible for their modified morphological as well as biological characteristics.

The numerous complex molecular processes occurring inside living cells are primarily carried out by proteins and other biopolymers, such as ribonucleic acids (RNA). The identity and quantity of the different proteins and RNA determine the cell's phenotype and changes in response to the environment. Therefore, the internal composition of the cell in terms of the type and concentration of proteins and RNA is tightly regulated. Gene expression is the process of using the DNA sequence information to produce these biopolymers. Transcription, the initial step in gene expression, where one strand of DNA is used as template by the enzyme RNA polymerase (RNAP) for synthesizing a complementary RNA or transcript. Since cell phenotype is mostly determined by transcription, a complex regulatory mechanism exists involving a large number of factors to control the level of transcription of a gene. Although most studies are focused on multiple cycles of either transcription or association of DNA and RNA Polymerase (RNAP) to make RNAP-Promoter open complex (RPO), single round transcription studies are crucial in elucidating the mechanism of sophisticated RNAP-DNA interactions and its kinetics in transcription. In this context, we have developed a novel in vitro quenching based single round transcription assay using single molecule detection. Using this, we could successfully dissect initiation kinetics starting from different initial transcribing stages and found that transcription initiation doesn’t follow a sequential model (as commonly believed). Instead, we identified a previously uncharacterized state that is unique to initial transcribing complexes and associated with the backtracked RNAP-DNA complex. Also, we have investigated the size/concentration effects of various osmolytes and macromolecular crowding agents, which mimic the crowded cellular environment, on actively-transcribing RNAP and found enhancement in transcription kinetics by larger crowding agents at the same viscosity.

Transcription of genomic DNA of all organisms is carried out by members of the multi-subunit RNA polymerase family. Regulation of RNA polymerase localization and activity underlies cellular homeostasis, division, and response to environmental cues. The catalytic mechanism, overall architecture, and many sequence and structural features of bacterial RNA polymerase are conserved in its Archaeal and Eukaryotic counterparts. The human RNA polymerase II (Pol II) is responsible for transcription of all protein-coding and many non-coding genes. The majority of current knowledge on RNA polymerases and their mechanism at different steps in transcription derives from extensive work done using classical biochemical, genetic and structural biology methods. However, the use of single-molecule approaches addressed crucial questions on the function and mechanism of RNA polymerases during transcription, which were not possible to answer with ensemble-based approaches due to averaging effects. A useful fluorescence-based single-molecule technique to measure distances on the molecular scale and monitor dynamics is F�rster resonance energy transfer (FRET). Here, I report on the development of diffusion-based single-molecule FRET (smFRET) methods to investigate different steps in transcription by the in vitro reconstituted human Pol II system. Using an assay that monitors the FRET changes between fluorescent dyes in the unwound region of promoter DNA (transcription bubble), I demonstrated the effect of certain components of the reconstituted system on the relative size of the transcription bubble. I also detail the optimizations done to enhance the affinity of single-stranded DNA (ssDNA) FRET probes to complementary target sequences. These ssDNA FRET probes were used to investigate the effect of certain components of the reconstituted system on Pol II activity by measuring the relative levels of RNA product. In addition to studies on the Pol II system, I report on the effect of the 5’-group of nascent RNA on the stability of the Escherichia coli RNA polymerase (RNAP) transcription bubble. I show how the presence of a 5’-monophosphate appears to destabilize the open bubble while a 5’-hydroxyl has no effect. Finally, I describe the work done on a project I took part in that identified a previously uncharacterized RNAP paused complex in initiation. We demonstrate that RNAP complexes undergoing initial transcription can enter the inactive paused state by backtracking. I also demonstrate how the presence of a 5’-triphosphate rapidly enhances entrance of RNAP complexes undergoing initial transcription into an inactive paused complex.