摘要
应用分子力学方法MM,和半经验量子化学AM1法,获得22种4-X-N-Y-6-氮杂雄-17-羰基-Z-4-烯-3-酮衍生物的优势构象,再利用量子化学算法和分子图形学技术,获得电子结构参数和几何结构参数,采用多元线性回归分析和人工神经网络误差反传算法(BP),将这些参数和衍生物对3β-羟类固醇脱氢酶(3BHSD)的抑制活性相关联。结果表明,衍生物对3BHSD的抑制活性大小和分子最高占用轨道能(E_(HOMO))、16号碳原子的净电荷(Q)及5号碳原子和6号氮原子间键长的相关性较好,成功地建立了22种衍生物的构效关系式。HOMO轨道组成分析表明,A环上的4号、5号碳原子和羰基氧原子(O_(1(?)))及B环上的6号氮原子和7号碳原子可能是与受体作用时的活性位点。
To analyze the QSAR of 22 4-X-N-Y-6-azaandrost-17-CO-Z-4-en-3-ones, MM ＋ molecular mechanics method and semi-empirical AM1 method had been performed. Using multivariate regression analysis （MLR） and standard back-propagation algorithm of artificial neural network（BP） , we obtained a significant QSAR. It could be concluded as bellows： the inhibition activity of 4-X-N-Y-6- Azaandrost-17-CO-Z-4-en-3-ones toward 3BHSD were highly correlated with the highest occupied orbital energy, net charge of C16 and the bond length between C5 and N6. Successful QSAR with correlation for 22 4-X-N-Y-6-azaandrost-17-CO-Z-4-en-3-ones was developed. The analysis of HOMO of 4-X-N-Y-6-azaandrost-17-CO-Z-4-en-3-ones indicated that the C4, the C5 and the O18 of ring A together with the N6 and the C7 of ring B might be active centers that influenced acceptors.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2007年第8期1099-1102,共4页
Computers and Applied Chemistry
基金
江苏省滩涂生物资源与环境保护重点建设实验室(JLCBE05033)
