COMPUTER STUDIES OF PROTEIN SURFACE RECOGNITION

蛋白质和其他生物大分子在 具体的方式。这种特殊性是由几何排列决定的 以及相互作用表面的原子的化学性质 分子。可以产生对分子表面的详细描述 通过结合X射线晶体结构测定， 溶剂可及性计算和计算机制图。这是详细的 分子表面描述可以应用于理解以下系统 大分子相互作用，例如：抗体分子的结合 对于一种蛋白质，基因调控蛋白与DNA的结合， 聚集血红蛋白亚基形成四聚体蛋白，即 黄曲霉毒素对胰蛋白酶和羧肽酶活性部位的封闭 抑制蛋白，以及病毒外壳蛋白单体组装成 冰盖。 分子识别既依赖于化学，也依赖于几何 形成接触的表面区之间的互补性。化学制品 互补性可以通过分析氢键、盐键 以及分子界面上的疏水接触。几何学 互补性可以通过表征外部的形状来研究 大分子的表面。蛋白质的外表面由下列公式计算 在三维空间上滚动一个球，代表一个水分子 蛋白质的结构。对于蛋白质表面的每个区域，都有一组 然后应计算表征区域形状的参数。 对于分子已经共晶的情况和原子 已确定了络合物的结构，几何互补性 将通过比较形状参数来测量界面处 接触中的表面区域。归根结底，应该可以预测 蛋白质对之间或蛋白质与核之间的相互作用 酸，通过计算将表面区域与 形状互补。这一能力将非常有用，因为它将 增加已知的结构信息量，而不需要任何 额外的结晶学工作。

Proteins and other biological macromolecules interact with each other in specific ways. This specificity is determined by the geometric arrangement and chemical properties of atoms at the surfaces of the interacting molecules. A detailed description of the molecular surface can be produced by a combination of X-ray crystallographic structure determination, solvent-accessibility calculations and computer graphics. This detailed molecular surface description can be applied to understanding systems where macromolecules interact, for example: the binding of an antibody molecule to a protein, the binding of gene regulatory proteins to DNA, the aggregation of hemoglobin subunits to form the tetrameric protein, the blocking of the active sites of trypsin and carboxypeptidase enzymes by inhibitor proteins, and the assembly of viral coat protein monomers into capsids. Molecular recognition is dependent on both chemical and geometric complementarity between the surface regions forming the contact. Chemical complementarity can be investigated by analyzing hydrogen bonds, salt links and hydrophobic contacts at molecular interfaces. Geometric complementarity can be studied by characterizing the shapes of the outer surfaces of macromolecules. The outer surface of a protein is computed by rolling a sphere, representing a water molecule, over the three-dimensional structure of the protein. For each region of protein surface, a set of parameters characterizing the shape of the region should then be computed. For the situation where molecules have been co-crystallized and the atomic structure of the complex has been determined, the geometric complementarity at the interface would be measured by comparing the shape parameters of the surface regions in contact. Ultimately, it should be possible to predict associations between pairs of proteins, or between a protein and a nucleic acid, by computationally docking together surface regions with complementary shapes. This ability would be very useful, since it would augment the amount of structural information known, without requiring any additional crystallographic work.