Studies On The Enantioselectivity And Conformation Changes Of Lipases Aided By Molecular Modeling

Using the docking modules in the Sybyl software, we docked the substrates analog of tetrahedral intermediates to the active sites of four lipases, to extensively predict the enantioselectivity of these lipases towards 66 enantiomers. Docking conformations without all the three essential H-bonds were discarded. To improve the docking accuracy and prediction ratio gradually, we used the Surflex, Surflex-GeomX and FlexX covalent-docking mode in turn, and we got the best predictive results when using FlexX covalent-docking mode. The covalent-docking results indicated that, when the enantio selectivity (E value) was lower than 100, the binding energy difference between the two enantiomers was not significant enough to be used to accurately predict the enantiopreference of the enzymes. And the prediction ratio obviously increased to 81.5%, when the E value was higher than 100 and the substrate containd fewer backbone carbons. At the same time, in all the three different docking modes, the lipase CALB obtained the best predictive results, with a correlation coefficient of 0.52 between the predicted and experimental binding free energy differences in the covalent-docking mode, which was consistent with the result got from MM/PBSA method and indicated that it should be a credible predicting result.Then, to predict the enantio selectivity of lipases without crystal structures, three structural models of ROL, Lipase-YS and Lipase-AK were constructed using homology modeling method, and we got high-quality structural models tested by model evaluation softwares. Using these modeled structures as receptors, we docked their substrates to the active sites by covalent-docking mode, to predict their enantioselectivity. A low prediction ratio of 52.8% was achieved due to the lower E values gathered from the papers and limitations of docking, and this proved again that FlexX covalent-docking could only get a better prediction ratio when the E was higher than 100.Finally, to assess the influence of different solvents to lipases with a lid, Amber 11 was used to model the structure changes of lipase HLL in water, methanol and chloroform. Molecular dynamics results indicated that, the core structures of open and closed HLL were stable in all three solvents, with a small rmsd difference; but the lid structures were much flexible, with a big rmsd difference. The a-helix areas of HLL in water increased, and decreased in methanol and chloroform. Comparatively, theβ-sheet areas did not show big difference in all three solvents. Besides, different solvents could affect the SASA of HLL, due to their different characteristics and polaritys. The lid structure could open, fold or keep stable in different solvents, and it controlled the channel of substrates into the enzyme active site and had an effect on the local micro-environment, so the solvents had a serious effect on the catalytic performance of lipases with a lid. To some extent, this might give explanations to the better extensive adaptability of substrates for CALB, compared with other lipases containing a lid structure.