Binding Properties Of Proteins Studied By Coarse-grained Model

Proteins are important molecular machines with functions of catalysis, transport, immunity and so on. These special functions are determined by the unique spatial structures of proteins. Therefore, it is important to maintain the stability of protein structures. In vivo, proteins are surrounded by different organelle, biomolecules (such as proteins, DNAs, and RNAs) and foreign matters (such as nanoparticles). Some might stabilize the structures of proteins, while some might destruct them. Thus, it is meaningful to study the interaction between proteins and other molecules. In this thesis, we studied the interactions of protein-protein system and protein-nanoparticle system by molecular dynamic simulation method. Since proteins and nanoparticles are representatives of organic and inorganic materials, it is helpful for understanding the folding and binding properties of proteins in complex surrounding by studying these two interactions. Proteins are mimicked by coarse-grained Go model based on natural structures. This model can enhance sample efficiently, which helps to reveal further physical properties.As for the protein-protein system, we studied the binding process of Arc repressor. The free energy curve of binding process shows the properties of "fly-casting" mechanism. That is, the curve has a flat and wide region. This suggests that interaction between two subunits occurs at the early stage of binding process when the two monomers are relatively far away from each other. As "fly-casting" is a new developed theory, there are few comprehensive methods to describe the process. Thus, we focused on finding useful analysis methods in our work. In this thesis, we mainly discussed whether Φ-value analysis is suitable for "fly-casting" mechanism. O-value analysis has been widely used to describe protein folding process, and has achieved many successes in finding key residues during transition states. However, our results suggest that Φ-value analysis can only roughly depict the properties of "fly-casting" process, while it fails to give detailed information. It is due to the fact that, the transition state that can be well described by Φ-value analysis is usually located at the relatively late stage of "fly-casting" process, while the key residues that play important roles in the binding process should interact with each other at the early stage. Therefore, Φ-value analysis is not fit for "fly-casting" mechanism. According to the properties of free energy curve, we defined effective capture radius Reff. Residues with large Reff interact with other residues at the early stage of binding, while those with small Reff interact with others late. This indicates that Reff is an effective value to depict "fly-casting" mechanism. To further verify the independence of our results on Go model, we simulated the system with a modified Go model, which include attraction between nonnative contacts. As a result, we obtained similar conclusions. This suggests that our results are robust regardless of models.As for the protein-nanoparticle interaction, we did two studies. The first one studied the interaction between three proteins of different topology, a+P, all a and all β, and two nanoparticles (NPs) of different hydrophobicity. For each protein-NP system, we drew the conformational phase diagram, and found four regions, including DF(desorbed-folded state)、AF(adsorbed-folded state)、AU(adsorbed-unfolded state) and WU(wrapped-unfolded state).The shapes and positions of each region are determined by NP size and interaction strength. We further found that, during the transition between DF and AU, the folding and binding process of proteins coordinate with each other. Moreover, secondary structures of proteins are also influenced by NP. Generally, in the folded region, βstructure increases while a structure decreases slightly. This results from geometrical properties of secondary structure instead of crowding effect or hydrophobicity of proteins and NPs. β structure is consisted of inter-chain H-bonds. Thus, it is flexible to fit for the curvature of NP without unfolding. However, a structure is consisted of intra-chain H-bonds, and it is rigid. When it adsorbs onto NP, it is easy to unfold.The second one studied the interaction between three proteins and Janus NP with half hydrophobic surface and half hydrophilic surface. Our simulation resulted that, for α+β and all a proteins, the conformational phase diagrams contain two regions, which are DF and AU. Due to the inconsistency of the hydrophobicity distribution between proteins and NPs, proteins will twist upon binding and lose their structures. While for all β protein, the diagram has three regions, which are DF, AF and AU. The appearance of AF may due to the flexibility of β structure. When adsorbing onto smaller NP with weaker interaction strength, the protein can tolerate the curvature and keep most of its structure. Other results are consistent to the previous work.Although there have been large amount of excellent studies on protein-NP interaction, most of them only focus on the interaction between specific protein and NP. Our work systematically studied how the factors including protein’s topology, NP size, surface hydrophobicity and interaction strength influence protein-NP system. This may provide some reference for future design of bio-nano-materials.The contents of this thesis are arranged below. In the first chapter, we will briefly introduce the theory of protein structure, protein folding, binding mechanism and protein-NP interaction. In the second chapter, we will introduce the work about binding process of Arc repressor and the effectiveness of Φ-value analysis. In the third and fourth chapters, we will introduce the work about protein-NP interaction. Finally, in the fifth chapter, we will make conclusions about this thesis and suggest for future work.