Study Of The Structure-Function Relationship Of LipK107 By Molecular Modeling And Its Rational Design

More and more intersections are found between information technology and biotechnology. As the computer technology gets mature, computer simulation (also called molecular modeling) is now playing more and more important role in the study of structure-function relationship of proteins. In the present thesis, we adopted versatile molecular modeling tools (including homology modeling, molecular docking, molecular dynamics simulation and MM/PBSA) to study extensively the structure-function relationship of LipK107, a microbial lipase screened from the soil. Rational design of the LipK107 was performed on the basis of the above structure-function analysis, and two mutants were found to improve the activity of the enzyme.Homology modeling is the more precise method of predicting the 3-D structure of a target protein. In this thesis, this modeling tool was applied to build both the active and the inactive conformation of LipK107. By comparing the two theoretic structures, key structural features of LipK107 were identified, including a catalytic triad (S79-D232-H254), an oxyanion hole (L13, Q80) and a lid substructure (I118-D152). The existence of the lid was further verified by observing the "interfacial activation" of LipK107. When superimposing the two structures, we could observe the conformational change from active conformation to inactive conformation of the lipase. Furthermore, molecular dynamics simulation was used to simulate the lipase in water surrounding. The result displayed directly the conformational change of the protein from active conformation to the inactive one.Molecular docking was widely employed to explore the binding mode of ligand and its corresponding receptor as well as their action mechanism. In the present thesis, we developed a computer-aided substrate screening system (CASS) for lipase using the flexible docking program—Affinity on the basis of the 3-D structure and reaction mechanism of Candida antartica lipase B (CALB). When screening the substrates of CALB with CASS, we found that the success rate was as high as 95.7%. Besides, the CASS was used to screen the substrates of LipK107 as well. The predicted result showed that four out of the ten candidate compounds were identified as the potential substrates while the rest six compounds were found not to be catalyzed by the enzyme. The results of the following biocatalysis of these ten compounds showed that only one was predicted wrongly, suggesting that the success rate of CASS was as high as 90% for LipK107. As far as we know, CASS was the only method of substrate virtual screening of enzyme.Biocatalysis involves the process of bond breaking and forming. Molecular docking, however, could not fulfill the goal. So we used molecular dynamics simulation of the enzyme-substrate transition state to explore the enantioselectivity of CALB catalyzed reaction. A new productive binding mode——"M/H permutation" between the enzyme and the slow-reacting enantiomer was discovered. Using the same molecular dynamics simulation method, we predicted the enantioselectivity of LipK107. R-enantiomer was identified as the fast-reacting enantiomer. The in silico result was in consistent with the result of the following biocatalysis experiment.On the basis of the hydrophobic and charge distributions of the lid of LipK107, a series of molecular modeling method (including virtual mutation, molecular docking, molecular dynamics and MM/PBSA) were employed to study their effects on the activity of the lipase. Besides, corresponding mutated proteins (I128E/V129D, E130L/K131I, Y138V, R120P/K121P, H146S/R147T) were built virtually. By comparing the binding energy of the mutants, we showed that only two mutants (E130L/K131I, Y138V) would probably improve the activity of LipK107. The results of the following point mutation suggested that only mutants E130L/K131I and Y138V of all five mutants could showed improved reaction conversion and eep when the enzyme was used to catalyzed the resolution of 1-phenyl ethanol. The experimental result was in consistent with the result of theoretical prediction. This not only guaranteed the accuracy of the in-silico prediction, but also implied that the combination of the versatile molecular modeling tools ("virtual mutation+molecular docking+molecular dynamics+MM/PBSA") would provide a new method for rational design of proteins.In addition to the rational design of the lid of LipK107, virtual mutation was used to design a mutant (D232A/N103D) with a new catalytic triad (S79-D103-H254) based on positional conservation of the catalytic triad of LipK107. Contrary to the predicted result, the mutant lost its activity in the actual biocatalysis experiment. Using molecular dynamics, we found that the hydrogen bond between D103 and H254 of the newly design catalytic triad (S79-D103-H254) was unstable in the mutant D223A/N103D. This was the main reason that resulting in the loss of activity of the mutant...