Development of a method to predict protein function based on module classification

本研究的目的是(1)阐明蛋白质功能的关键成分定位于紧凑的短片段、模块，以及(2)阐明以模块为构建块的蛋白质设计原理。结果，我们发现不同类型的蛋白质功能被分配给模块。例如，(1)木聚糖酶家族10的催化结构域的功能部位定位于某些模块；(2)在许多情况下，某些金属离子的配体存在于一个或两个模块中。此外，在不同的蛋白质中经常观察到结构和功能相似的模块，这表明这些模块可能被洗牌了。此外，(3)共同的模块用于蛋白质-蛋白质相互作用；(4)只有两个模块用于血红蛋白的亚基相互作用，这可以解释为什么血红蛋白的亚基相互作用在不同的谱系中是多样化的。我们还通过实验证明了模块可以被删除或洗牌。例如，(1)删除了Barnase中的模块M2，而不损失与天然结构相似的稳定三维结构和协同折叠行为。(2)10号家族的两种木聚糖酶之间通过模数M10交换而形成的嵌合体木聚糖酶，虽然酶活性降低了十分之一，但仍保持了结构和功能。本研究从计算机科学和实验科学两个角度阐明了蛋白质模块化设计的基本原理。

The aim of this study is (1) to clarify that key components of protein functions are localized to compact short segments, modules, and (2) to elucidate principle of protein design using modules as building blocks. In consequence, we found various kinds of protein functions were assigned to modules. For example, (1) the functional sites of the catalytic domain of family 10 xylanase are localized to some of the modules; (2) in many case, ligands for certain metal ions exist within one or two modules. Furthermore, modules with similar structure and function are frequently observed in various proteins, which indicates the possibility that the modules were shuffled. In addition, (3) common modules are found to be used for protein-protein interaction; (4) only two modules are used in subunit interaction of hemoglobin, which could explain why subunit interactions of hemoglobin are diversified in various lineage. We also showed that modules can be deleted or shuffled by experiments. For example, (1) module M2 in barnase was deleted without loss of the stable three-dimensional structure similar to the native structure and the cooperative folding behavior. (2) a chimera xylanase, which was made by exchange of module M10 between two kinds of xylanases in family 10, maintained the structure and function, though the enzyme activity decrease by a tenth. This study has clarified a basic principle for module-based protein design from both views of computer science and experimental science.