Abstract:

Using the docking module FlexX in the SYBYL molecular modeling suite，a total of 66 enantiomers were docked to four lipases to predict the enantioselectivity of these lipases. In this process，the substrate analogs of tetrahedral intermediates were covalently docked to the active sites of lipases. Docking conformations without all the three essential H-bonds were discarded. The docking result indicated that the binding energy difference between the two enantiomers was not significant enough to accurately predict the enantiopreference of the enzymes when enantioselectivity was lower than 100. And the prediction ratio increased to 81.5% obviously，when enantioselectivity was higher than 100 and the substrate contained fewer backbone carbons. During the docking process，productive conformations were formed between most substrates and lipases，which were in accordance with the theoretical catalytic mode. Besides，owning to its high computing rate，this method can be applied to modeling and prediction of potential substrates of lipases with high throughput.

摘要：

以SYBYL软件包中的FlexX对接模块为研究工具，模拟预测了4种脂肪酶对66对手性底物的水解及转酯反应的对映体选择性。模拟过程中，将四面体中间态底物类似物共价对接到脂肪酶的活性位点，以3个必需氢键的形成为筛选条件，去除不符合标准的酶-底物对接构象。对接结果表明：当催化反应的E值较小时（E＜100），酶与R/S两种底物的结合自由能差不足以准确预测酶的优先反应构型；而当底物的E≥100，并且主链碳原子数目较少时，该方法对脂肪酶对映体选择性的预测率明显提高，达到81.5%。由于模拟中，酶与绝大多数底物形成了含有3个氢键的反应型构象，符合理论催化模型，同时由于该方法计算速度快，因而该方法可用于高通量模拟预测脂肪酶的潜在底物。