目的 設(shè)計(jì)DPP-4和SGLT-2雙靶點(diǎn)抑制劑。方法 根據(jù)現(xiàn)有的DPP-4和SGLT-2活性化合物訓(xùn)練BiRNN模型，使其生成具有潛在活性的雙靶點(diǎn)新分子，并使用已經(jīng)報(bào)道的活性化合物構(gòu)建HipHop藥效團(tuán)模型，通過藥效團(tuán)模型對(duì)新分子進(jìn)行初篩，然后使用分子對(duì)接、分子動(dòng)力學(xué)模擬以及結(jié)合自由能計(jì)算等方法對(duì)初篩小分子進(jìn)行進(jìn)一步篩選最終得到候選化合物。結(jié)果BiRNN模型生成7 494個(gè)新分子，經(jīng)過虛擬篩選發(fā)現(xiàn)化合物NM186、NM21、NM249、NM107是相對(duì)理想的DPP-4和SGLT-2雙靶點(diǎn)抑制劑。結(jié)論NM186、NM21、NM249、NM107可能是一種具有新型結(jié)構(gòu)的DPP-4/SGLT-2雙靶點(diǎn)抑制劑，這一研究豐富了DPP-4和SGLT-2雙靶點(diǎn)抑制劑的多樣性，為其開發(fā)提供新的設(shè)計(jì)思路。

Objective To design a dual-target inhibitor of DPP-4 and SGLT-2. Methods The BiRNN model was trained according to the existing DPP-4 and SGLT-2 active compounds to generate new dual-target molecules with potential activity, and the HipHop pharmacophore model was constructed using the reported active compounds, and the new molecules were pre-screened through the pharmacophore model, and then the small molecules were further screened by molecular docking, molecular dynamics simulation and combined free energy calculation, and the candidate compounds were finally obtained. Results A total of 7 494 new molecules were generated by the BiRNN model, and the compounds NM186, NM21, NM249, and NM107 were found to be relatively ideal dual-target inhibitors of DPP-4 and SGLT-2. Conclusion NM186, NM21, NM249 and NM107 may be dual-target inhibitors of DPP-4/SGLT-2 with novel structures, which enriches the diversity of DPP-4 and SGLT-2 dual-target inhibitors and provides new design ideas for their development.

R914.2

陜西省秦創(chuàng)原“科學(xué)家+工程師”隊(duì)伍建設(shè)（2023KXJ-080號(hào)）