Badger John
Zenobia Therapeutics, San Diego, CA, USA.
Methods Mol Biol. 2012;841:161-77. doi: 10.1007/978-1-61779-520-6_7.
Crystallographic fragment screening is a technique for initiating drug discovery in which protein crystals are soaked or grown with high concentrations of small molecule compounds (typically MW 110-250 Da) chosen to represent fragments of potential drugs. Specific binding of these compounds to the protein is subsequently visualized in electron density maps obtained from analysis of X-ray diffraction data collected from these crystals. Theoretical and practical experience indicate that a suitably diverse library of fragment compounds containing only a few hundred compounds may be sufficient to provide a comprehensive screen of the protein target. By soaking crystals in mixtures of 3-10 compounds a fragment screen may be completed within ∼100 diffraction data sets. This data collection requirement may be met given reproducible well-diffracting protein crystals and robotic sample handling equipment at a high flux X-ray source. The leading practical issue for most crystallography laboratories that wish to launch a fragment screening project is the design and/or procurement of an appropriate fragment library. Although several off-the-shelf fragment libraries are available from chemical suppliers, the numbers, sizes, and solubility of the compounds in relatively few of these libraries are well-match to the specific needs of the crystallographic screening experiment. Informed consideration of the properties of compounds in the screening library, possibly augmented by additional filtering of available compounds with appropriate search tools, is required to design a successful experiment. The analysis of results from crystallographic fragment screening involves highly repetitive application of routine image data processing and structure refinement calculations from many very similar crystals. Efficient handling of the data applies a high-throughput structure determination methodology that conveniently packages the structure solution calculations into a single process that provides the crystallographer-analyst with ready-to-view maps for evaluating crystals for bound compounds.
晶体学片段筛选是一种启动药物发现的技术,其中蛋白质晶体用高浓度的小分子化合物(通常分子量为110 - 250道尔顿)进行浸泡或生长,这些小分子化合物被选来代表潜在药物的片段。这些化合物与蛋白质的特异性结合随后在从这些晶体收集的X射线衍射数据分析得到的电子密度图中显现出来。理论和实践经验表明,一个仅包含几百种化合物的适当多样化的片段化合物库可能足以对蛋白质靶点进行全面筛选。通过将晶体浸泡在3 - 10种化合物的混合物中,大约100个衍射数据集内即可完成片段筛选。在高通量X射线源具备可重复的良好衍射蛋白质晶体和机器人样品处理设备的情况下,可以满足这一数据收集要求。对于大多数希望开展片段筛选项目的晶体学实验室来说,首要的实际问题是设计和/或采购合适的片段库。尽管化学供应商提供了几种现成的片段库,但其中相对较少的库中化合物的数量、大小和溶解度与晶体学筛选实验的特定需求匹配良好。为了设计一个成功的实验,需要明智地考虑筛选库中化合物的性质,可能还需要用适当的搜索工具对可用化合物进行额外筛选。晶体学片段筛选结果的分析涉及对许多非常相似晶体的常规图像数据处理和结构精修计算的高度重复应用。数据的有效处理应用了高通量结构测定方法,该方法将结构解析计算方便地打包成一个单一过程,为晶体学分析人员提供随时可查看的图谱,用于评估结合化合物的晶体。