Wu Zhufeng, Zhang Xingwang, Ma Zhiguo, Wu Baojian
Division of Pharmaceutics, College of Pharmacy, Jinan University , Guangzhou , China.
Xenobiotica. 2015;45(8):653-62. doi: 10.3109/00498254.2015.1016136. Epub 2015 Apr 2.
1.UDP-glucuronosyltransferase1A3 (UGT1A3) catalyzes glucuronidation of numerous xenobiotics/drugs. Here, we aimed to establish substrate selectivity models for UGT1A3 using the pharmacophore and VolSurf approaches. 2.Fifty structurally diverse substrates of UGT1A3 were collated from the literature. These substrates were divided into training (n=34) and test sets (n=16). The pharmacophore model was developed using the Discovery Studio 2.5 software. A user-defined feature (i.e. the glucuronidation site) was included in the program for model generation. The VolSurf model was derived using the VolSurf program implemented in SYBYL 8.0 software. 3.The pharmacophore model consisted of three features (i.e. one glucuronidation site and two hydrogen-bond acceptors). The activities of 81% of test set substrates were adequately predicted (deviated by less than one-log unit) by the model, suggestive of a satisfactory predictive power. The refined VolSurf model based on 22 molecular descriptors was statistically significant (r(2)=0.793, q(2)=0.606). It also processed a good predictability as the activities of 14 test set compounds were well predicted. The VolSurf model highlighted the chemical features (including large molecule size, hydrophilic regions and hydrogen-bonding groups) contributing to favored glucuronidation by UGT1A3. 4.In conclusion, two predictive 3D-QSAR models (i.e. the pharmacophore and VolSurf models) for UGT1A3 were successfully established. These models contributed to an improved understanding of the substrate preference of UGT1A3 and a more comprehensive prediction of UGT-mediated metabolism.
1.尿苷二磷酸葡萄糖醛酸基转移酶1A3（UGT1A3）催化多种异生物素/药物的葡萄糖醛酸化反应。在此，我们旨在运用药效团和VolSurf方法建立UGT1A3的底物选择性模型。2.从文献中整理出50种结构各异的UGT1A3底物。这些底物被分为训练集（n = 34）和测试集（n = 16）。使用Discovery Studio 2.5软件开发药效团模型。在模型生成程序中纳入了一个用户定义特征（即葡萄糖醛酸化位点）。VolSurf模型是使用SYBYL 8.0软件中实现的VolSurf程序推导得出的。3.药效团模型由三个特征组成（即一个葡萄糖醛酸化位点和两个氢键受体）。该模型能够充分预测81%的测试集底物的活性（偏差小于一个对数单位），表明其具有令人满意的预测能力。基于22个分子描述符的优化VolSurf模型具有统计学意义（r(2)=0.793，q(2)=0.606）。由于14种测试集化合物的活性得到了良好预测，所以它也具有良好的预测性。VolSurf模型突出了有助于UGT1A3进行有利葡萄糖醛酸化反应的化学特征（包括大分子尺寸、亲水区和氢键基团）。4.总之，成功建立了两种用于UGT1A3的预测性三维定量构效关系模型（即药效团模型和VolSurf模型）。这些模型有助于更好地理解UGT1A3的底物偏好，并更全面地预测UGT介导的代谢过程。
