Cells of the immune system operate in dynamic microenvironments where the timing, concentration and order of signaling molecules constantly change. Despite this complexity, immune cells manage to communicate accurately and control inflammation and infection. It is unclear how these dynamic signals are encoded and decoded, and if individual cells retain memory of past exposure to inflammatory molecules. Here, we use live cell analysis, ATAC sequencing and an in vivo model of sepsis to show that sequential inflammatory signals induce memory in individual macrophages through reprogramming the NF-κB network and the chromatin accessibility landscape. We use transcriptomic profiling and deep learning to show that transcription factor and chromatin dynamics coordinate fine-tuned macrophage responses to new inflammatory signals. This work demonstrates how macrophages retain memory of previous signals despite single-cell variability and elucidates the mechanisms of signal-induced memory in dynamic inflammatory conditions like sepsis.

In this talk, I will provide an overview of microfluidics and discuss its applications in advancing biophysical studies. Specifically, I will highlight our efforts in cryo-electron microscopy (cryo-EM) sample preparation techniques by leveraging micro-electromechanical systems (MEMS) and microfluidic technologies. Cryo-EM has revolutionized structural biology by enabling high-resolution visualization of biological macromolecules in their native, hydrated states. Technological advancements in detectors and software algorithms have propelled cryo-EM to near-atomic resolution, making it a compelling alternative to X-ray crystallography. Integrating unconventional technologies like microfluidics further enhances cryo-EM’s potential, particularly for studying rapid molecular dynamics. For instance, time-resolved cryo-EM using microfluidic mixing-and spray devices enables the investigation of fast processes on the millisecond timescale. I will introduce innovative sample preparation methods, including nanofluidic TEM chips for microsecond-scale time resolved cryo-EM, parylene thin-film microfluidics for low-sample-consumption spray techniques, and microfluidic systems for cryo-focused ion beam (FIB) lamella preparation. These advancements demonstrate the exciting possibilities of combining microfluidics and cryo-EM for biophysical research

In each human cell, 2 meter DNA is physically compacted into chromosomes that are packed into a micrometer-sized nucleus, but the mechanism by which the extremely long and negatively charged polymer is compacted into this tiny structure remains elusive. Structural Maintenance of Chromosome (SMC) protein complexes such as cohesin and condensin are the key organizers of the spatiotemporal structure of chromosomes by extruding DNA loops1. In our work using liquid-phase High-Speed Atomic Force Microscopy (HS AFM) and magnetic tweezers (MT)1,2, we were the first to obtain experimental data for yeast condensin acting on individual DNA molecules. The findings suggest a scrunching model in which the SMC complex extrudes a DNA loop by a cyclic switching of its conformation between open and collapsed shapes. In addition, we show that yeast cohesin complexes unexpectedly can exhibit a new type of phase separation, called a Bridging-Induced Phase Separation (BIPS) that is induced by bridging of distant regions on a long DNA by proteins that possess multiple DNA-binding sites3. Our findings suggest that both DNA loop extrusion and phase separation by SMC complexes are fundamental building blocks of chromosome organization. Lastly, for future works, I propose various biophysical studies to shed light on the universal and fundamental principles of phase separation that constitute membraneless organelles in chromosome.

Memories are thought to be encoded in populations of neurons called memory trace or engram cells. However, little is known about the dynamics of these cells because of the difficulty in real-time monitoring of them over long periods of time in vivo. To overcome this limitation, we present a genetically-encoded RNA indicator (GERI) mouse for intravital chronic imaging of endogenous Arc mRNA—a popular marker for memory trace cells. We used our GERI to identify Arc-positive neurons in real time without the delay associated with reporter protein expression in conventional approaches. We found that the Arc-positive neuronal populations rapidly turned over within two days in the hippocampal CA1 region, whereas ~4% of neurons in the retrosplenial cortex (RSC) consistently expressed Arc following contextual fear conditioning and repeated memory retrievals. Dual imaging of GERI and a calcium indicator in CA1 of mice navigating a virtual reality environment revealed that only the population of neurons expressing Arc during both encoding and retrieval exhibited relatively high calcium activity in a context-specific manner. This in vivo RNA imaging approach opens the possibility of unraveling the dynamics of the neuronal population underlying various learning and memory processes.

Satisfactory characterization of quantum transitions involving electronic processes in complex molecular environments remains challenging despite great advances made over many decades. Depending on the nature of systems and environments, such quantum transitions can be characterized by rates or more complete quantum dynamical description. As a recent example of such rate theories, this talk will introduce our new quantum rate theory for nonradiative decay, which can account for new quantum effects that were missing in previously established theories. Applications of this theory to organic dye molecules demonstrate the success of improved rate theories for quantitative description of exciton transfer and nonradiative decay rates. For transitions that go beyond simple rate description, more advanced quantum dynamics approaches become necessary. This talk will also provide a brief overview of our ongoing efforts to develop advanced open system quantum dynamics methods addressing such issues and their utility for future study of driven open system quantum dynamics processes. Of particular interest in this respect is quantum sensing (QS) broadly defined. QS utilizes the sensitivity of unique quantum properties such as quantum coherence and entanglement on external perturbations and seeks for detection of environmental properties through manipulation and measurement of clearly understood quantum systems. So far, control mechanisms utilized for QS have been mostly based on intuition or those adopted from traditional spectroscopic techniques. The objects for sensing have also been fairly limited. However, from the theoretical point of view, there is substantial prospect for the capability of QS to evolve further in the near future, especially at molecular scale, through more satisfactory integration of advanced theories and experimental protocols. The latter part of this talk will address this possibility, provide brief overview of the quantum estimation theory as a key guiding principle for QS, and describe preliminary ideas for future theory development and computational studies. The talk will conclude by summing up recent advances we have made in simulating driven open system quantum dynamics processes and future challenges.

첨단 소재의 성능은 원자 수준의 구조와 전자적 특성에 의해 결정되지만, 실제 물질은 복잡 한 구조와 다양한 반응 경로를 지니므로 이를 정밀하게 규명하고 예측하는 일은 여전히 도전 적이다. 본 발표에서는 인공지능(AI)과 전산모사 기법을 결합하여 복잡한 재료의 원자 단위 메 커니즘을 규명하고 신소재를 발굴하는 본 연구실의 연구를 소개한다. 비정형적 표면 구조를 가진 Pt jagged nanowire의 수소 발생 반응(HER) 메커니즘을 기계학습 포텐셜로 분석하여 기 존 DFT 기반 접근의 한계를 극복하는 방법을 제시한다. 또한, 4D-STEM(4차원 주사 투과 전자 현미경)에 딥러닝 분석을 적용하여 원자 배치와 국소 전기적 특성을 정밀 재구성하고, 전자빔 단일 회절 패턴만으로 고분해능 토모그래피를 구현하는 기법을 설명한다. 마지막으로, 범용 기 계학습 원자간 퍼텐셜을 개발·활용하여 촉매 및 소재의 반응 동역학과 구조–성능 상관성을 대 규모로 예측하고, 합성 방법론 이해를 확장하여 합성 가능성을 수치화하는 차세대 연구 방향 을 제안한다.