| 摘要: | 隨著分子生物學的快速進展與基因和蛋白質快速突變，取得基因的定序與胺基酸的序列對於現今的科技而言已是相當方便；然而，目前蛋白質的立體結構分析，卻仍是一大難關。透過計算生物學(Computational Biology)模式(如: SWISS-MODEL、ΑlphaFold2)，由胺基酸序列推測該蛋白質的結構，可大幅簡化結構的解析流程。本研究以New Delhi metallo-β-Lactamase (NDM)為標的，希望透過更精細的方式，探討以胺基酸序列建立之模擬結構，可否成為精準的立體結構模型，建立整合結構與功能之資料庫，探討序列變異點位對蛋白質結構的影響。本研究透過標準化結構分析流程，建立一套完整的結構分析模式。研究中有三大突破，包含為分析流程設計、臨床變異型/人工突變型的綜合模擬預測與現有結晶結構的價值最大化。第一，進行蛋白質結構標準定位視覺化，並輔以突變位外各位點之胺基酸鍵角差異進行量化分析。在動態反應或環境變因的狀態下，評估細部結構變化。第二，由NDM之臨床與人工突變型序列，建立整合的結構模型資料庫，並將對應結晶結構的結果，分析不同來源的點突變之間的交互影響、結構對功能影響以及NDM催化中心變化。第三，收集以及統整目前結構資料庫中與標的相關的結晶結構，針對常見不同反應條件進行討論，包含臨床突變位結構、蛋白質與受質作用機制和抑制劑作用位置進行分析。希望透過這套系統流程優化，對於快速且新興的蛋白質突變，進行快速精準之模擬預測。
With the rapid advancement of molecular biology and mutation of genes and proteins, obtaining gene sequencing and amino acid sequences has become quite convenient with today's technology. However, the analysis of protein structures remains a major challenge. Many studies, such as SWISS-MODEL and ΑlphaFold2, predict the structure of a protein from its amino acid sequence with computational biology model, simplifying the previously time-consuming and labor-intensive process of structure determination. This study aims to explore a more refined approach to obtaining accurate three-dimensional structure models solely from amino acid sequences. It focuses on the New Delhi Metallo-β-Lactamase (NDM) as the target protein and establishes an integrated database to investigate the impact of specific sequence changes on the protein structure. In this study, a comprehensive structural analysis framework is established through standardized procedures. There are three major breakthroughs in the research, including the design of the analysis workflow, comprehensive simulation prediction of clinical variants/artificial mutants, and maximization of the value of existing crystal structures. Firstly, the analysis workflow involves the standard positioning and visualization of protein surface structures. Quantitative analysis is performed by considering the differences in amino acid bond angles at mutation sites, evaluating the structural changes in specific regions or under specific environmental conditions. Secondly, a consistent structural prediction model database is established using sequences from target proteins' clinical and artificial mutant strains. Corresponding to the results of crystal structures, the interaction effects between different point mutations from various sources, overall structural variations, and changes in binding centers are inferred. Thirdly, relevant crystal structures related to the target protein are collected and integrated from existing structural databases. Analysis is conducted on common crystalline environments, including clinical mutation site structures, protein-substrate interaction mechanisms, and inhibitor binding sites. Our study goal is to simulate and predict rapidly evolving and emerging protein mutations in a time-saving and efficient manner using this system. |
