GIST Scholar
Not a member yet
30271 research outputs found
Sort by
Engineering of a Synthetic α-Secretase
No full textThe excessive accumulation of amyloid- (A is implicated in the pathogenesis of Alzheimer’s disease (AD). Recent clinical studies have demonstrated that elimination of A is a viable therapeutic strategy. In the current study, we conceptualized a fusion membrane protein, referred to as synthetic -secretase (SAS), designed to cleave specifically amyloid precursor protein (APP) and A at the -site. In mammalian cells, SAS indeed cleaved APP and A at the targeted -site. Overexpression of SAS in the hippocampus was achieved by stereotaxic injection of recombinant adeno-associated virus serotype 9 (AAV9) encoding SAS (AAV9-SAS) into the bilateral ventricles of mouse brains. Consequently, SAS enhanced the non-amyloidogenic processing of APP, and reduced the levels of soluble A and plaques in the 5xFAD mice. Furthermore, SAS significantly attenuated the cognitive deficits in 5xFAD and App knock-in (NL-G-F/NL-G-F) mice, as demonstrated through novel object recognition and Morris water maze tests. Unlike other A -cleaving proteases, SAS has highly strict substrate specificity. I propose that SAS can be an efficient modality to eliminate excessive A from diseased brains. I further obtained and characterized SAS variants with increased proteolytic activity (SASN70Q, SASScMV, and SASWpMV), altered subcellular localizations (SAS 0, SASGPI, SASSEC, SASFUR, SASDUAL, SASFC5), and reduced immunogenicity (SASK114P, SAST117D). I propose that these findings would be valuable for optimizing SAS as a therapeutic candidate for various A -related diseases.DoctorContents
Abstract . ⅰ
Contents . ii
List of figures . ⅴi
List of tables viii
PART Ⅰ. Therapeutic Effects of the Synthetic α-Secretase
Ⅰ. Introduction 2
Ⅱ. Materials and method . 4
2-1. Plasmid construction . 4
2-2. modRNA synthesis . 4
2-3. AAV production 5
2-4. Oligomeric FITC-Aβ42 preparation . 5
2-5. Transfection 5
2-6. Cell culture and treatment . 6
2-7. Immunocytochemistry (ICC) 6
2-8. Animals . 7
2-9. Treatments 8
2-10 Tissue preparation 8
2-11. Protein electrophoretic analysis and Western blotting 9
2-12. Immunohistochemistry (IHC) . 9
2-13. Enzyme linked immunosorbent assay (ELISA) 10
2-14. Behavior test . 10
2-15. Statistical analysis . 13
Ⅲ. Results 14
3-1. Conceptualization and characterization of SAS 14
3-2. SAS cleaves an artificial substrate and APP at the α-site 15
3-3. SAS cleaves FITC-Aβ42 at the α-site . 16
3-4. Expression of SAS in the 5xFAD mice injected with AAV9-SAS 17
3-5. SAS enhances the non-amyloidogenic processing of APP in the 5xFAD mice 17
3-6. SAS reduces Aβ load in the 5xFAD mice 17
3-7. SAS partially restores abnormal reduction of anxiety-like behavior in the 5xFAD mice . 18
3-8. SAS attenuates cognitive deficits in the 5xFAD mice . 19
3-9. SAS attenuates motor function defects in the 5xFAD mice 20
3-10. SAS attenuates cognitive deficits in the App knock-in (NL-G-F/NL-G-F) mice 20
Ⅳ. Legends, Figures and Tables . 22
Ⅴ. Discussion 43
PART Ⅱ. Characterization of the Synthetic α-Secretase and Its Variants
Ⅰ. Introduction 46
Ⅱ. Materials and method . 48
2-1. Plasmid construction . 48
2-2. Transfection 48
2-3. Cell culture and treatment . 48
2-4. Glycosidase treatment . 48
2-5. Protein electrophoretic analysis and Western blotting . 49
2-6. Glycosylation prediction . 49
2-7. Structure prediction by AlphaFold3 49
2-8. In silico immunogenicity assay . 49
2-9. Statistical analysis . 49
Ⅲ. Results 51
3-1. SAS activity is affected by N-glycosylation 51
3-2. Generation of SAS variants . 52
3-3. Generation of SASSEC variants 52
3-4. Generation of SASSEC variants utilizing BBB penetrating moieties 53
3-5. Generation of SAS utilizing NIas from other viruses 54
3-6. Computational prediction of immunogenicity and de-immunization of SAS . 54
3-7. Generation of de-immunizing variants SAS 55
Ⅳ. Legends, Figures and Tables . 56
Ⅴ. Discussion 73
Reference 75
Abstract in Korean 83
Acknowledgement . 85
Curriculum Vitae . 88
List of figures
PART Ⅰ. Therapeutic Effects of the Synthetic α-Secretase
Figure 1. The design and characterization of SAS 22
Figure 2. SAS cleaves an artificial type-I membrane protein and APP at the α-site . 24
Figure 3. SAS cleaves FITC-Aβ42 at the α-site 26
Figure 4. Expression of SAS in the 5xFAD mice injected with AAV9-SAS 28
Figure 5. SAS enhances the non-amyloidogenic processing of APP in the 5xFAD mice . 30
Figure 6. SAS reduces Aβ load in the 5xFAD mice 32
Figure 7. SAS partially restores abnormal reduction of anxiety-like behavior in the 5xFAD mice 34
Figure 8. SAS attenuates cognitive deficits in the 5xFAD mice . 36
Figure 9. SAS attenuates motor function defects in the 5xFAD mice . 38
Figure 10. SAS attenuates cognitive deficits in the App knock-in (NL-G-F/NL-G-F) mice 40
PART Ⅱ. Characterization of the Synthetic α-Secretase and Its Variants
Figure 1. Generation and characterization of glycoengineered variants of SAS 56
Figure 2. Generation and characterization of SAS variants 58
Figure 3. Generation and characterization of variants of SASSEC . 60
Figure 4. Generation and characterization of variants of SASSEC utilizing BBB penetrating moieties 62
Figure 5. Generation and characterization of SAS utilizing NIas from other viruses . 64
Figure 6. Computational prediction of immunogenicity of SAS for MHC class I and II . 66
Figure 7. Computational prediction of de-immunization of SAS for MHC class I and II 68
Figure 8. Generation and characterization of de-immunizing variants of SAS 70
List of tables
PART Ⅰ. Therapeutic Effects of the Synthetic α-Secretase
Table 1. Antibodies used in this study 42
PART Ⅱ. Characterization of the Synthetic α-Secretase and Its Variants
Table 1. Antibodies used in this study 7
Machine-learning-aided accelerated discovery of energy materials with enhanced properties
No full textThe discovery of novel materials plays a crucial role in advancing cutting-edge technologies, enabling innovative performance improvements in fields such as energy storage and electronics. As existing materials reach their limits, there is a growing demand for novel materials with superior performance to surpass current ones. However, the discovery of novel materials remains a challenging task, with traditional trial-and-error methods being time-consuming and resource-intensive. This dissertation explores the application of mahine learning (ML) to accelerate the discovery of high-performance materials, with a focus on two key areas: metal-organic frameworks (MOFs) for ammonia (NH3) adsorption and cathode active materials for magnesium (Mg) ion batteries. The first study leverages an ML model to efficiently screen over 12,000 MOFs, identifying eight MOFs with high NH3 working capacity through the integration of ML predictions and theoretical validation through GCMC and molecular dynamics (MD) simulations. The second study investigates Mg-based garnet structures for Mg-ion batteries, employing ML to accelerate the discovery process and first-principle calculations for theoretical validation. Together, these studies demonstrate the transformative potential of ML in materials science, offering a pathway to significantly reduce the time and resources required for materials discovery, while ensuring the reliability of predictions through computational validation.Docto
Ice Recrystallization Inhibition via 2D Janus Peptide Nanosheets—A Molecular to Macroscales Approach
No full textAntifreeze proteins (AFPs) are crucial for the survival of polar organisms under subzero temperatures, as they selectively bind to specific planes of ice crystals and regulate formation and growth of the ice. To develop cryoprotective agents that effectively replicate the function of AFPs, it is essential to understand how their chemical functionalities and structural features influence ice crystal formation and growth. Previous studies have focused on the atomic-level interactions between ice-binding moieties (IBMs) and ice crystal surfaces to explain ice recrystallization inhibition (IRI) activity. Self-assembled nanostructures composed of amphiphilic peptides were designed to mimic the functional characteristics of AFPs both locally and structurally on a macro-scale and atomic-scale. The amphiphilic peptides are composed of hexa-phenylalanine, glutamic acid, and an alkyl tail, with the IBMs incorporated at the hydrophilic terminus. The hexa-phenylalanine segment facilitates β-sheet formation via hydrogen bonding and enhances structural stability through π-π stacking of aromatic rings. Simultaneously, the electrostatic repulsion between glutamic acid residues and the hydrophobic effect from the alkyl tail promotes the self-assembly of peptides into two-dimensional (2D) nanostructure with a high specific surface area. The presence of IBMs at the hydrophilic segment enhances interaction with ice, whereas the hydrophobic face acts as a barrier that limits water penetration. These features give rise to well-defined 2D Janus nanosheets. The formation of 2D nanostructures markedly enhanced ice recrystallization inhibition (IRI) activity. Assessment of IRI activity across different ice-binding residues showed that those enabling both hydrogen bonding and hydrophobic interactions led to the highest inhibition efficiency via synergistic molecular effect. Anisotropic ice crystal growth was observed in the presence of amphiphilic peptide nanosheets, indicating interaction between the nanosheet and the ice. Ice nucleation occurred at a lower temperature in the presence of amphiphilic peptide nanosheets than in pure water, implying their modulatory effect on ice formation. Cryo in-situ imaging techniques enabled direct visualization of the adsorption behavior and ice-binding interfaces between the peptides and ice, offering clear structural evidence of their antifreeze activity. Amphiphilic peptide nanoagents can induce chemical interactions at the atomic and molecular scales, mimicking the structural function of natural AFPs by 2D structural formation. These results suggest that self-assembled 2D amphiphilic peptide nanostructures serve as potent ice growth inhibitors and represent a cryoprotective agents for the preservation of sperm, oocytes, stem cells, tissues, and organs
Visualizing speech styles in captions for deaf and hard-of-hearing viewers
No full textSpeech styles such as extension, emphasis, and pause play an important role in capturing the audience's attention and conveying a message accurately. Unfortunately, it is challenging for Deaf and Hard-of-Hearing (DHH) people to enjoy these benefits when watching lectures with common captions. In this paper, we propose a new caption system that automatically analyzes speech styles from audio and visualizes them using visualization elements such as punctuation, paint-on, color, and boldness. We conducted a comparative study with 26 DHH viewers and found that the proposed caption system enabled them to recognize the speaker's speech style in lectures. As a result, the DHH viewers were able to watch lecture videos more vividly and were more engaged with the lectures. In particular, punctuation can be a practical solution to visualize speech styles and ensure legibility. Participants expressed a desire to use our caption system in their daily lives, providing valuable insights for future sound-visualized caption research. © 2024 Elsevier LtdFALSEsciescopu
A Multilayer-Based Visibly Transparent Radiative Cooler with Enhanced Solar Selectivity
No full textTransparent radiative cooling enables passive thermal management by selectively reflecting solar radiation and emitting thermal energy, thereby offering a promising solution for energy-efficient cooling in transparent applications. Despite recent progress in the development of transparent radiative coolers, achieving a balance between high visible transparency and effective solar reflection remains a critical challenge in enhancing the cooling performance. In this study, a transparent radiative cooler comprising a 50 mu m-thick polydimethylsiloxane (PDMS) emitter integrated with a sophisticated metal-dielectric multilayer is evaluated both experimentally and theoretically. The theoretical simulations demonstrate a visible transmittance >82% and a near-infrared reflectance >92%, with a sharp transition between spectral regimes. Experimental characterizations confirm a visible transmittance of 48% alongside an enhanced near-infrared reflectance of 98%, thereby validating robust cooling performance under real-world conditions. Additionally, daytime outdoor measurements indicate that the cooler achieves a maximum temperature reduction of 10.8 degrees C compared to conventional PDMS-coated glass. Furthermore, global simulations illustrate the cooling effect of this cooler under diverse climatic conditions, highlighting its potential for use in various terrestrial areas.TRUEsciescopu
Controlling CD4+ T Cell Responses by T Cell Receptor Signaling
No full textCD28-PI3K 신호전달 경로는 활성화된 CD4+ T 세포의 대사 과정을 조절하여 세포 증식을 위한 요구를 충족시켜 T 세포의 병원체에 대한 면역 반응을 가능하게 한다. 본 연구를 통하여 TCR/CD28 자극에 의한 CD4+ T 세포 증식에서 PANK4의 활성 조절을 통하여 신생 지질 합성을 조절하는 역할을 한다는 것을 발견했다. CD28 자극에 의한 신호는 PDK1과 PANK4의 직접 결합을 유도하고, 결과적으로 PANK4를 억제하여 조효소 A 합성을 촉진함으로써 신생 지질 합성을 조절한다. CD4+ T 세포 증식에서 PANK4의 이러한 조절 역할은 항상성 증식 및 대장염 모델에서 검증되었으며, 생체 내에서 Pank4 결핍 CD4+ T 세포는 야생형 CD4+ T 세포보다 더 활발한 세포분열을 보이는 것이 확인되었다. 본 연구 결과는 TCR/CD28 자극 시 PANK4 조절이 CD4+ T 세포 증식에 필수적이며 CD4+ T 세포 반응을 조절하는 표적으로 연구될 수 있음을 나타낸다.|The CD28-PI3K pathway regulates the metabolic reprogramming of activated CD4+ T cells to meet the demands for optimal proliferation, enabling T cells to respond to pathogens. I found that PANK4 regulation plays a role in TCR/CD28-mediated CD4+ T cell proliferation by regulating de novo lipid synthesis. CD28 signaling negatively regulates PANK4 through direct binding with PDK1, controlling de novo lipid synthesis by modulating coenzyme A synthesis. This regulatory role of PANK4 in CD4+ T cell proliferation was confirmed in an experimental colitis model, where Pank4-deficient CD4+ T cells exhibited greater expansion than their wild-type counterparts when co-transferred. My findings suggest that PANK4 regulation is crucial for TCR/CD28-induced CD4+ T cell proliferation and represents a potential target for modulating general CD4+ T cell responses.Docto
Timing‐Dependent Spiking Neural Network: Board‐Level Hardware Implementation with Photoelectroactive Van der Waals Synapses
No full textThe rapid growth of unstructured data in applications such as autonomous systems and edge AI underscores the urgent need for energy-efficient, real-time computing exemplified by biological brains, where synaptic weights are adjusted according to the timing of neural spikes, known as spike-timing-dependent plasticity (STDP). This work presents the first experimental realization of a multi-channel timing-dependent spiking neural network (TD-SNN) at the board-level by integrating photoelectroactive synaptic devices with an analog leaky integrate-and-fire (LIF) neuron circuit. The synaptic devices exploit the precise timing dependency between electrical presynaptic and optical postsynaptic spikes to emulate STDP, enabling reversible and bidirectional modulation of synaptic weights through photoelectroactive doping. By engineering the shape of presynaptic pulses, the devices demonstrate diverse biological STDP learning rules, including Hebbian, anti-Hebbian, all-LTP, and all-LTD. Integrated single- and multi-channel networks exhibit self-learning, system-level adaptive, and competitive behaviors. Experimentally extracted STDP parameters are implemented in SNN simulations, where network performance is determined by the long-term potentiation/depression area ratio (LTP/D area ratio, PDR) of the STDP curve. When PDR ≥ 1.25, robust pattern classification is achieved, reaching up to 90.9% accuracy on MNIST tasks. These results mark a milestone in timing-dependent neuromorphic hardware, demonstrating device-level feasibility toward adaptive and real-time learning hardware.FALSEsciescopu
Deep metric loss for multimodal learning
No full textMultimodal learning often outperforms its unimodal counterparts by exploiting unimodal contributions and cross-modal interactions. However, focusing only on integrating multimodal features into a unified comprehensive representation overlooks the unimodal characteristics. In real data, the contributions of modalities can vary from instance to instance, and they often reinforce or conflict with each other. In this study, we introduce a novel MultiModal loss paradigm for multimodal learning, which subgroups instances according to their unimodal contributions. MultiModal loss can prevent inefficient learning caused by overfitting and efficiently optimize multimodal models. On synthetic data, MultiModal loss demonstrates improved classification performance by subgrouping difficult instances within certain modalities. On four real multimodal datasets, our loss is empirically shown to improve the performance of recent models. Ablation studies verify the effectiveness of our loss. Additionally, we show that our loss generates a reliable prediction score for each modality, which is essential for subgrouping. Our MultiModal loss is a novel loss function to subgroup instances according to the contribution of modalities in multimodal learning and is applicable to a variety of multimodal models with unimodal decisions. Our code is available at https://github.com/DMCB-GIST/MultiModalLossFALSEsciescopu
TOPOLOGY OPTIMIZATION OF MULTIPLE DEGREES-OF-FREEDOM BREAKAWAY STRUCTURES
No full textBreakaway devices are mechanical “fuses” to protect critical components from damage caused by excessive loads. Current breakaway devices are designed by analytical or empirical approaches, resulting in complicated assemblies with limited degrees-of-freedom (DOF). While the stress-constrained topology optimization (TO) has been extensively studied for designing structures against failure, no prior work has utilized TO for designing breakaway structures that fail under prescribed loads. This work proposes a TO-based formulation for designing monolithic multi-DOF breakaway structures with embedded yield zones, regions intended for failure. Compared to the approach based on the conventional stress-constrained TO, the proposed method keeps the average stress in the yield zones below the material’s yield strength, ensuring consistent breakaway behaviors in the presence of modeling errors and manufacturing variations. Meta-design strategies are developed to facilitate the convergence: sliding loading surfaces to realize asymmetric responses in a single DOF, yield zone and domain separation to decouple responses among DOFs, and incremental addition of load cases for better response interpolation between load cases. Preliminary results demonstrate that the proposed formulation produces the structures that fail at multiple finite element nodes, indicating greater tolerance against modeling errors and manufacturing variations, and the meta-design strategies successfully enable the desired breakaway behaviors in multi-DOF space. © 2025 by ASME
Development of a sequence-based RNA purification method with a single-nucleotide resolution by synthesis and selection
No full textThe accuracy of synthetic RNA is critical for the success of RNA therapeutics. However, current RNA purification methods rely on physicochemical properties such as length or charge, which make it hard to effectively distinguish sequences with single-nucleotide errors. To overcome this limitation, we developed a novel sequence-based RNA purification platform capable of purifying error-free RNAs at single-nucleotide resolution. Our method operates through a synthesis and selection mechanism utilizing modified deoxynucleotide triphosphates (dNTPs). It involves stepwise elongation with sequence-specific dNTPs, combined with blocking steps using a triple combination of dideoxynucleotide triphosphates (ddNTPs) to prevent the extension of erroneous templates. Final labeling with biotin-modified dNTP enables the selective capture of error-free RNA molecules. We validated this strategy using prime editing guide RNAs (pegRNAs), which are over 160 nucleotides in length and have a complex secondary structure. Next-generation sequencing (NGS) analysis of the purified pegRNAs demonstrated a global reduction in errors across most regions of target sequence, with notable decreases in deletion and substitution errors. Compared to unpurified controls, the purified pegRNAs showed an average increase of 2.59 percentage points in overall purity, underscoring the effectiveness of our method. This approach offers an indispensable solution for the precise purification of synthetic RNAs and holds strong potential for enhancing the quality and performance of RNA therapeutics.|합성 RNA의 서열 정확성은 RNA 치료제의 성공적인 적용에 있어 매우 중요한 요소이다. 그러나 현재 널리 사용되는 정제 기술은 길이나 전하 등 물리화학적 특성에 의존하고 있어, 단일 염기 오류와 같은 정밀한 구별에는 한계가 있다. 이러한 문제를 해결하기 위해, 본 연구에서는 단일 염기 분리능을 갖는 새로운 서열 기반 RNA 정제 플랫폼을 개발하였다. 본 기술은 변형된 디옥시뉴클레오타이드(dNTP)를 이용한 합성과 선택 기작을 기반으로 하며, 정확한 서열에 대해 단계적으로 dNTP를 합성하고, 비정상 서열의 경우 목표 서열에 상보적이지 않은 조합의 ddNTP로 신장을 종결했다. 마지막으로 Biotin이 표지된 dNTP를 통해 오류 없는 RNA만을 선택적으로 회수한다. 모든 반응은 자성 비드 위에서 수행되어 용액 교환과 세척이 용이하였다.
이러한 전략은 162 nt 길이와 복잡한 2차 구조를 가진 prime editing guide RNA(pegRNA)에 적용되어 유효성을 검증하였다. 차세대 염기서열 분석(NGS)을 통해 정제된 pegRNA는 전체 서열 영역에서 오류가 감소했으며, 특히 결실(deletion) 및 치환(substitution) 오류가 크게 줄어드는 양상을 보였다. 정제 전후를 비교한 결과, 정제된 RNA의 평균 순도는 2.59 %p 향상되어 합성 RNA에 대한 본 방법의 유의미한 개선을 확인하였다.
본 정제 플랫폼은 기존 RNA 정제 방식의 한계를 극복하고, 정밀한 합성 RNA 정제를 가능하게 하는 기술이다. 또한 본 플랫폼은 높은 유연성을 바탕으로 siRNA, CRISPR guide RNA 등 다양한 합성 RNA의 정제에 활용할 수 있으며, 치료제의 품질과 성능을 향상시킬 수 있는 높은 확장성과 실용성을 지닌다. 본 연구는 서열 특이적 정제를 통한 고정밀 RNA 품질 관리의 새로운 가능성을 제시하며, RNA 생명공학 분야에 중요한 발전을 이끌 수 있는 기술적 기반을 마련하였다.Master1. Introduction 1
1.1. Advances in RNA biology and therapeutics 1
1.2. Limitations of conventional oligonucleotide synthesis and purification 1
1.3. Sequence-based RNA purification mechanism for single-nucleotide resolution 2
2. Materials and Methods 4
2.1. Oligonucleotide design and synthesis 4
2.2. Preparation of DNA-linked magnetic bead 4
2.3. Synthesis and selection 5
2.4. Polyacrylamide Gel Electrophoresis (PAGE) analysis 6
2.5. Next Generation Sequencing (NGS) analysis 7
3. Results and Discussion 8
3.1. Sequence-based RNA purification via stepwise elongation and selective blocking 8
3.2. Evaluation of sequence-based RNA purification method by NGS 9
4. Conclusions 12
Summary 13
Reference 15
Acknowledgement 1