Sipilä Julius, Nurmi Harri, Kaukonen Ann Marie, Hirvonen Jouni, Taskinen Jyrki, Yli-Kauhaluoma Jari
Viikki Drug Discovery Technology Center, Faculty of Pharmacy, University of Helsinki, P.O. Box 56, FI-00014 Helsinki, Finland.
Eur J Pharm Sci. 2005 Jul-Aug;25(4-5):417-25. doi: 10.1016/j.ejps.2005.04.002.
Currently there are several compounds used as drugs or studied as new chemical entities, which have an electron withdrawing group connected to a vinylic double bond in a phenolic or catecholic core structure. These compounds share a common feature--current computational methods utilizing the Hammett type equation for the prediction of ionisation constants fail to give accurate prediction of pK(a)'s for compounds containing the vinylic moiety. The hypothesis was that the effect of electron-withdrawing substituents on the pK(a) of p-vinyl phenols is due to the delocalized electronic structure of these compounds. Thus, this effect should be additive for multiple substituents attached to the vinylic double bond and quantifiable by LFER-based methods. The aim of this study was to produce an improved equation with a reduced tendency to underestimate the effect of the double bond on the ionisation of the phenolic hydroxyl. To this end a set of 19 para-substituted vinyl phenols was used. The ionisation constants were measured potentiometrically, and a training set of 10 compounds was selected to build a regression model (r2 = 0.987 and S.E. = 0.09). The average error with an external test set of six compounds was 0.19 for our model and 1.27 for the ACD-labs 7.0. Thus, we have been able to significantly improve the existing model for prediction of the ionisation constants of substituted p-vinyl phenols.
目前有几种化合物被用作药物或作为新的化学实体进行研究，它们在酚类或儿茶酚类核心结构中具有与乙烯基双键相连的吸电子基团。这些化合物有一个共同特点——目前利用哈米特型方程预测电离常数的计算方法无法准确预测含有乙烯基部分的化合物的pK(a)值。假说是吸电子取代基对对乙烯基苯酚pK(a)的影响是由于这些化合物的离域电子结构。因此，对于连接在乙烯基双键上的多个取代基，这种影响应该是可加和的，并且可以通过基于线性自由能关系（LFER）的方法进行量化。本研究的目的是生成一个改进的方程，以降低低估双键对酚羟基电离影响的倾向。为此，使用了一组19种对位取代的乙烯基苯酚。用电位法测量电离常数，并选择10种化合物的训练集来建立回归模型（r2 = 0.987，标准误差 = 0.09）。对于我们的模型，6种化合物的外部测试集的平均误差为0.19，而ACD实验室7.0的平均误差为1.27。因此，我们能够显著改进现有的用于预测取代对乙烯基苯酚电离常数的模型。
