RIKEN-Seiko Collaboration Center
JPSingle-Molecule Biophysics of Biomolecules
Cells function as systems, maintaining metabolic and cellular homeostasis while diverse biomolecules continually change their expression, localization, and interactions. Understanding these systems requires both tracking molecular dynamics at the cellular scale and examining the structure and function of each molecular component. We isolate these components and refine single-molecule measurement technologies to observe and analyze them. By following individual molecules, we can track structural and functional changes during activation and turnover and reveal differences between molecules. We also study the effects of amino acid mutations and environmental factors at the single-molecule level, deepening our understanding of molecular properties and informing the development of improved molecules and technologies for practical use.
Single-molecule analysis of proteins
Proteins deftly regulate functions such as chemical reactions and molecular transport while dynamically changing their structures. We develop original technologies to directly observe and manipulate biomolecules one molecule at a time. By visualizing the motions, reactions, and structural fluctuations of individual molecules—which are averaged out and cannot be resolved by conventional measurements of large molecular populations—we seek to uncover how they work at the molecular level.
We have directly observed and manipulated the rotation of ATP-hydrolyzing enzymes, which rotate as they break down ATP, at the single-molecule level. This allowed us to distinguish, at the level of elementary steps, how structural changes relate to chemical reactions such as ATP binding, hydrolysis, and product release. We found that structural fluctuations and the order of these reactions are important factors controlling protein catalytic activity. We have also developed original devices, including microchambers that form lipid bilayers, and used them to directly measure membrane-protein functions that were difficult to capture one molecule at a time, including proton transport by ATP synthase and lipid transport by scramblases.
A defining feature of our research is that we do not simply apply existing measurement methods to biomolecules; we create new technologies tailored to the biological phenomena we want to observe. By developing new measurement principles and devices for phenomena that current technologies cannot reveal, we aim to make the invisible visible. We will continue to combine original technology development centered on single-molecule measurement with life science, uncover unknown functions and operating principles of diverse biomolecules, and open new avenues in life science.
Understanding CRISPR-Cas, a bacterial immune system
CRISPR-Cas is a sensor molecule in the bacterial immune system that protects bacteria from foreign nucleic acids, such as those from viruses. It recognizes and cleaves target nucleic acids that match its guide RNA in a sequence-specific manner. We focus on this remarkable molecular recognition mechanism and directly measure, at the single-molecule level, how CRISPR-Cas searches for and recognizes its target among vast numbers of nucleic acid sequences and proceeds to cleave it. Our research aims to clarify the reaction mechanism from its elementary steps.
For Cas13a, which cleaves RNA, we directly tracked an individual Cas13a molecule activated by binding its target RNA as it successively cleaved nearby reporter RNAs in a microchamber, and quantified its enzyme activity molecule by molecule (Commun Biol, 2021). This enabled us to analyze the reaction properties of individual molecules, which are difficult to discern from average measurements of large molecular populations alone. For Cas12a, which cleaves DNA, we are also systematically examining at the single-molecule level how many mismatches in the target sequence it tolerates and how their positions and types affect the reaction rate (Anal Chem, 2026c).
This fundamental understanding helps address important questions: Why does CRISPR-Cas sometimes react with non-target sequences, and how can it distinguish sequences that differ by just one base with high accuracy? We also use insights from single-molecule analyses of molecular recognition and reaction rates as design principles to maximize CRISPR-Cas performance and further improve the sensitivity and accuracy of nucleic acid tests for digital liquid biopsy.
Analyzing and engineering molecular function with AI and laboratory automation
Small changes in sequence can dramatically alter the function of proteins and nucleic acids. Which amino acids, base sequences, and domains determine function? We address these questions by combining precise experimental data from single-molecule measurements, in silico simulation data, and machine learning. Our laboratory has multiple GPU systems for AI and machine learning, as well as computing resources for first-principles calculations and coarse-grained simulations. This infrastructure lets us carry out experiments, measurements, simulations, data analysis, and machine learning within the laboratory.
We are also developing a rapid Design–Build–Test–Learn (DBTL) cycle, an increasingly important approach in synthetic biology. By combining machine-learning-based sequence design with rapid synthesis and evaluation, we aim to repeat this cycle in a short time. To achieve this, we work closely with the RIKEN-Seiko Collaboration Center (RSCC), established in April 2026, to develop laboratory automation that seamlessly connects robotic sample preparation, measurement, data analysis, AI-based interpretation, and the next round of design without manual intervention. This will deepen our understanding of molecular function and enable the rational, rapid creation of molecules with desired properties.
