Portal de Periódicos da CAPES

Effect of cholesterol on DMPC phospholipid membranes and QSAR model construction in membrane-interaction QSAR study through molecular dynamics simulation

2005; Elsevier BV; Volume: 14; Issue: 7 Linguagem: Inglês

10.1016/j.bmc.2005.11.009

1464-3391

Jianzhong Liu, Yang Liu,

Molecular Sensors and Ion Detection

In this study, both pure DMPC and DMPC/cholesterol mixed membrane monolayer were built to compare the physical–chemical properties and dynamics properties through molecular dynamics simulation and normal-mode analysis. The results show that the addition of cholesterol decreases the area of per molecule of membrane, increases the lipid amplitude motion, and changes the solute diffusion coefficient. It is also found that the addition of cholesterol greatly changes the solute-membrane 1,4-nonbonded interaction energy (ΔE14). MI-QSAR models were constructed based on solute–membrane interaction energy descriptors and other intramolecular descriptors. The results show that ΔE14 substitutes ΔEHB as the second important descriptor compared with the previous study. Final results suggest that short range solute–membrane interaction energy changes due to the uptake of the solute may play an important decision on permeability in DMPC/cholesterol membrane. A test set was applied to evaluate the predictivity of MI-QSAR models. The result suggests that the combination of FH2O and ΔE14 not only improves r2 and q2, but also greatly improves the model predictivity. Based on the combination of q2 and rpre2 values, a two-term model is better used to predict the solute permeability in this study (r2 = 0.859, q2 = 0.803, and rpre2=0.540). Due to the small sample both in training set and test set, more datasets are necessary to make a final decision about the model construction and prediction.