有机小分子化合物解离常数pK_a的预测研究进展  

A review on predicting pK_a values of small organic compounds

作　　者：于海瀛[1] 陈伟[1] 袁泉[1] 陈建荣[1] 林红军[1] 洪华嫦[1] 

机构地区：[1]浙江师范大学地理与环境科学学院,金华321004

出　　处：《科学通报》2015年第14期1261-1271,1-3,共11页Chinese Science Bulletin

基　　金：国家自然科学基金(21207119);浙江省教育厅科研项目(Y201226156);浙江师范大学博士启动基金(ZC304012043)资助

摘　　要：对于弱酸和弱碱化合物,解离常数(p Ka)是最重要的理化性质参数之一,其决定化合物的溶解度、亲脂性、生物富集性、毒性以及药物分子的吸收、分布、代谢和排泄(ADME)性质.通过实验方法测定化合物水溶液中的p Ka受到物质稳定性、仪器测定范围以及人力物力消耗等多方面的限制,因此过去几十年间发展了大量的p Ka预测方法.本文以有机小分子化合物为研究对象,回顾了20年来p Ka预测的研究成果,包括pK a实验数据的来源、质量、测定方法,重点介绍3类预测方法(线性自由能关系模型、定量结构-性质关系模型和第一性原理方法),并简单总结了常用的商业软件,最后提出未来p Ka预测研究需要关注的问题.The dissociation degree of weak acids and bases, pKa, is an important physicochemical property that determines their solubility in water, lipophicity, bioaccumulation, toxicity and ADME properties of pharmaceuticals. However experimental measurement of pKa is time-consuming and limited by the purity and stability of chemicals as well as instrument application scope. In this paper we reviewed the development of pKa prediction for small organic compounds. First, we introduced the thermodynamic background of pKa and the dissociation processes, micro-pKas and macro-pKas for chemicals with multi-ionization centers, taking cetirizine as an example. Then we summarized the resources of experimental pKa values, tabulating both electronic databases and books, analyzed the data quality, and listed seven error sources for inaccurate data. Potentiometry, spectrophotometry, conductometry, capillary electrophoresis （CE）, nuclear magnetic resonance （NMR） and high performance liquid chromatography （HPLC） are commonly used and favored analytical methods for pKa measurement. This study mainly focused on the review of three classes of pKa prediction approaches, including linear free energy relationships （LFERs）, quantitative structure-property relationships （QSPRs） and first principle methods. LFERs were predominantly used in the early study of pKa prediction, and remain a powerful approach which is widely and successfully employed in popular commercial and freely available software packages. QSPR are another one of the most common techniques used in pKa prediction. Benefit from the development of quantum chemistry, a great number of descriptors, such as superdelocalizability, polarizability, group philicity, molecular electrostatic potential （MEP） have been developed and calculated at different theoretical levels. These descriptors were employed to establish numerous linear QSPR models through partial least-squares （PLS） or multiple linear regressions, or non-linear QSPR models by machine learning method

关 键 词：酸碱解离常数 pKa预测 线性自由能关系 定量结构-性质关系 第一性原理 

分 类 号：O621.2[理学—有机化学]