Molecular Dynamics Simulations On Two Classes Of Proteins And Protein Docking Studies

With the development of the computer technology and computational methodology, molecular modeling has been widely used in scientific research fields. Computational methods can be thought of as the third category of scientific research methods along with empirical and theoretical ones. In biology field, computational approaches provide a series of manner, homology modeling, molecular dynamics (MD) simulations and automatic docking, to construct protein three-dimensional structures and illustrate the function of proteins etc. In this thesis, three kinds of researches have been conducted with these methods and outlined as following.1. Homology modeling and docking studies on P450 enzyme.Mouse CYP2C38 and CYP2C39 are two closely related enzymes with 91.8% sequence identity. But they exhibit different substrate binding features. The 3-dimentional models of CYP2C38 and CYP2C39 were constructed based on the crystal structure of 2C8 (PDB code: 1PQ2) and refined by energy minimization and MD simulations. With the flexible docking approaches, four substrate–enzyme complexes were obtained, including tolbutamide–CYP2C38 tolbutamide–CYP2C39, atRA–CYP2C38 and atRA–CYP2C39.Based on these models, we have shown that: the main difference between two related enzymes is the cavity volume above the heme group of CYP2C38 being smaller than that of CYP2C39. This may be one factor contributing to CYP2C38's catalytical inactivity towards atRA. Another possible factor based on our calculations is the spatial effect of Phe237 in CYP2C38. Arg241, Glu300, Leu366 (Ile366) and Leu476 could be the most important residues based on this study.We used molecular docking approach combined with molecular dynamics (MD) simulations to model three-dimensional (3D) complex structures of SPD-304, zafirlukast and L-745,870 into CYP3A4, respectively. The results showed that these three drugs can stably bind into the active site and the 3-methylene carbons of the drugs keep a reasonable reactive distance from the heme iron. The complex structure of SPD-304-CYP3A4 is in agreement with experimental data. For zafirlukast, the calculation results indicate that 3-methylene carbon might be the dehydrogenation reaction site. Docking model of L-745,870-CYP3A4 shows a potential possibility of L-745,870 dehydrogenated by CYP3A4 at 3-methylene carbon which is in agreement with experiment in vivo. In addition, residues in the phenylalanine cluster as well as S119 and R212 play a critical role in the ligands binding based on our calculations.2. Protein-Protein Docking using a Brownian Dynamics Simulations ApproachWe present a newly adapted Brownian Dynamics (BD) based protein docking method for predicting native protein complexes. The approach includes three steps: 1) Global BD conformational sampling, 2) Compact complex selection, 3) Local energy minimization. A shell-based grid force field was developed to represent receptor protein and solvation effect. This shell-based grid map has the advantage of largely reducing the number of grids compared to the 3D box grid map and thus can cover whole surface area of the receptor for conducting a global search. Soft potential maps were introduced to account partial flexibility of the proteins. The performance of this newly developed BD protein docking approach has been evaluated on a test set of 24 crystal protein complexes and on seven CAPRI targets. The results from the re-docking experiment show that the Root Mean Square Deviation (RMSD) between the predicted lowest energy and the crystal structures are within 2 ?. All the tested Capri targets can be sampled to the near native configurations. Six of seven targets can be predicted correctly within top 300 ranking. These results indicate that this approach can be useful for prediction of protein-protein interactions.3. Molecular dynamics simulations of PIP₂-driven Kir channel activationInwardly rectifying K+ (Kir) channels are gated by the PIP₂ molecule. The molecular mechanism of how PIP₂ induce Kir channels structural transition from the closed to the open state remains unclear. A high resolution structure of a Kir3.1 chimera revealed the presence of a cytosolic (G-loop) gate captured in the closed or open conformations. Here we conduct 100ns Molecular Dynamics simulations of these two channel states in the presence and absence of PIP₂:1) Constricted form of BacKir3.1; 2) Constricted form of BacKir3.1 + PIP₂; 3) Dilated form of BacKir3.1 + PIP₂. The simulation results revealed that PIP₂ is able to open and stabilize the G-loop gate. The opening of G-loop gate is accompanied by the structural transition of cytosolic interface between the subunits from latched to the unlatched conformation triggered by PIP₂. The constricted holo system exhibited several transitional characteristics between the constricted apo conformation and the dilated holo one. A molecular mechanism of PIP₂ controlling G-loop gate was proposed: The N-terminus initially interacted with the CD-loop in the absence of PIP₂ (latched conformation). The closed G-loop gate was stabilized by the conserved salt bridge E304-R313 and had loosely interaction with CD-loop. The presence of PIP₂ drove anchoring of the N-terminus through R52, which began to weaken the interactions between the CD loop and the N-terminus, engaging the G-loop into the network of new interactions. The weakened CD-loop interactions with the Slide Helix and N-terminus allowed the CD-loop to rearrange itself relative to the G-loop. Furthermore intrasubunitr interactions between the CD loop and the G loop further stabilized the two loops in the new conformation. This enhanced interaction by the CD and G loops was further stabilized by the formation of a salt bridge between R219 and PIP₂ thus driving the G-loop conformation to the open state. The N-terminus, no longer interacting with the CD-loop could form stable interactions withβM (unlatched interface). This is the first mechanistic view of PIP₂-induced channel gating that is consistent with existing experimental evidence.