Kinnings Sarah L, Jackson Richard M
Institute of Molecular and Cellular Biology and Astbury Centre for Structural Molecular Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK.
J Chem Inf Model. 2009 Feb;49(2):318-29. doi: 10.1021/ci800289y.
Methods for analyzing complete gene families are becoming of increasing importance to the drug discovery process, because similarities and differences within a family are often the key to understanding functional differences that can be exploited in drug design. We undertake a large-scale structural comparison of protein kinase ATP-binding sites using a geometric hashing method. Subsequently, we propose a relevant classification of the protein kinase family based on the structural similarity of its binding sites. Our classification is not only able to reveal the great diversity of different protein kinases and therefore their different potential for inhibitor selectivity but it is also able to distinguish subtle differences within binding site conformation reflecting the protein activation state. Furthermore, using experimental inhibition profiling, we demonstrate that our classification can be used to identify protein kinase binding sites that are known experimentally to bind the same drug, demonstrating that it has potential as an inverse (protein) virtual screening tool, by identifying which other sites have the potential to bind a given drug. In this way the cross-reactivities of the anticancer drugs Tarceva and Gleevec are rationalized.
分析完整基因家族的方法在药物发现过程中变得越来越重要，因为一个家族内的相似性和差异往往是理解功能差异的关键，而这些功能差异可用于药物设计。我们使用几何哈希方法对蛋白激酶ATP结合位点进行大规模结构比较。随后，我们基于其结合位点的结构相似性对蛋白激酶家族提出了相关分类。我们的分类不仅能够揭示不同蛋白激酶的巨大多样性，从而显示它们在抑制剂选择性方面的不同潜力，而且还能够区分反映蛋白激活状态的结合位点构象内的细微差异。此外，通过实验抑制谱分析，我们证明我们的分类可用于识别已知在实验中与同一药物结合的蛋白激酶结合位点，表明它有潜力作为一种反向（蛋白）虚拟筛选工具，通过识别哪些其他位点有可能结合给定药物。通过这种方式，抗癌药物特罗凯和格列卫的交叉反应性得到了合理的解释。
