Jun Kozuka

A Computational Framework for Bioimaging Simulation.

Using bioimaging technology, biologists have attempted to identify and document analytical interpretations that underlie biological phenomena in biological cells. Theoretical biology aims at distilling those interpretations into knowledge in the mathematical form of biochemical reaction networks and understanding how higher level functions emerge from the combined action of biomolecules. However, there still remain formidable challenges in bridging the gap between bioimaging and mathematical modeling. Generally, measurements using fluorescence microscopy systems are influenced by systematic effects that arise from stochastic nature of biological cells, the imaging apparatus, and optical physics. Such systematic effects are always present in all bioimaging systems and hinder quantitative comparison between the cell model and bioimages. Computational tools for such a comparison are still unavailable. Thus, in this work, we present a computational framework for handling the parameters of the cell models and the optical physics governing bioimaging systems. Simulation using this framework can generate digital images of cell simulation results after accounting for the systematic effects. We then demonstrate that such a framework enables comparison at the level of photon-counting units.

Integrin LFA-1 regulates cell adhesion via transient clutch formation

Integrin LFA-1 regulates immune cell adhesion and trafficking by binding to ICAM-1 upon chemokine stimulation. Integrin-mediated clutch formation between extracellular ICAM-1 and the intracellular actin cytoskeleton is important for cell adhesion. We applied single-molecule tracking analysis to LFA-1 and ICAM-1 in living cells to examine the ligand-binding kinetics and mobility of the molecular clutch under chemokine-induced physiological adhesion and Mn(2+)-induced tight adhesion. Our results show a transient LFA-1-mediated clutch formation that lasts a few seconds and leads to a transient lower-mobility is sufficient to promote cell adhesion. Stable clutch formation was observed for Mn(2+)-induced high affinity LFA-1, but was not required for physiological adhesion. We propose that fast cycling of the clutch formation by intermediate-affinity integrin enables dynamic cell adhesion and migration.

Automated single-molecule imaging in living cells

An automated single-molecule imaging system developed for live-cell analyses based on artificial intelligence-assisted microscopy is presented. All significant procedures, i.e., searching for cells suitable for observation, detecting in-focus positions, and performing image acquisition and single-molecule tracking, are fully automated, and numerous highly accurate, efficient, and reproducible single-molecule imaging experiments in living cells can be performed. Here, the apparatus is applied for single-molecule imaging and analysis of epidermal growth factor receptors (EGFRs) in 1600 cells in a 96-well plate within 1 day. Changes in the lateral mobility of EGFRs on the plasma membrane in response to various ligands and drug concentrations are clearly detected in individual cells, and several dynamic and pharmacological parameters are determined, including the diffusion coefficient, oligomer size, and half-maximal effective concentration (EC50). Automated single-molecule imaging for systematic cell signaling analyses is feasible and can be applied to single-molecule screening, thus extensively contributing to biological and pharmacological research.Large scale live cell screens often lack single-molecule resolution. Here the authors present an artificial intelligence-assisted TIRF microscope with automated cell searching and focusing, and use it for high-throughput single-molecule imaging of EGFR dynamics in response to various stimuli.

Intracellular Protein-Labeling Probes for Multicolor Single-Molecule Imaging of Immune Receptor–Adaptor Molecular Dynamics

Single-molecule imaging (SMI) has been widely utilized to investigate biomolecular dynamics and protein-protein interactions in living cells. However, multicolor SMI of intracellular proteins is challenging because of high background signals and other limitations of current fluorescence labeling approaches. To achieve reproducible intracellular SMI, a labeling probe ensuring both efficient membrane permeability and minimal non-specific binding to cell components is essential. We developed near-infrared fluorescent probes for protein labeling that specifically bind to a mutant β-lactamase tag. By structural fine-tuning of cell permeability and minimized non-specific binding, SiRcB4 enabled multicolor SMI in combination with a HaloTag-based red-fluorescent probe. Upon addition of both chemical probes at sub-nanomolar concentrations, single-molecule imaging revealed the dynamics of TLR4 and its adaptor protein, TIRAP, which are involved in the innate immune system. Statistical analysis of the quantitative properties and time-lapse changes in dynamics revealed a protein-protein interaction in response to ligand stimulation.