Simulations Of α-helix And β-sheet Proteins By Moleculer Dynamics With United-Residue Model

The protein folding problem has been regarded as a grand challenge for understanding the origin of life in the past decades. From the basic question came various research problems including (1) the protein folding code; (2) the folding mechanism; and (3) the prediction of a protein structure from its amino acid sequence. In addition to other approaches to tackle this problem, molecular dynamics (MD) simulations have been made huge progress to recover the folding processes at atomic details. One of the most significant methods was the united-residue (UNRES) model, which simplified the representation of amino acids by two beads. Meanwhile, the force field of UNRES is purely physics-based, determined by a cluster-cumulant expansion of the effective free energy of a protein plus the surrounding solvent. So this coarse grain method, UNRES, is capable of carrying out large-scale simulations in real time. Using UNRES molecular dynamics simulations, we obtained results as follows:1) We simulated the folding of a large eight-helix-bundle protein with a length of 145 amino acids from extended states. We observed multiple nucleation folding pathways: the formation of secondary structures was followed by the nucleation of helices at the two terminal parts and also at the middle of the chain, and then the nuclei grew and combined with each other to form the tertiary structure. Surprisingly, we also found a vectorial folding pathway that was shown recently for co-translational folding in the ribosome exit tunnel. Furthermore, we found that all three-helix subunits in the chain can fold into native-like conformations independently, especially those at the two terminal parts and the middle of the chain, which may be responsible for the nucleation’s. And more thermodynamic results showed that the nucleus at C-terminal of the chain has a higher propensity. These results may help understand the folding mechanism of large repeat helical proteins. 2) FBP WW domain provides us a typical and excellent prototype to study the folding mechanism ofβ-sheet formation. So we carried out 512 Langevin canonical dynamics trajectories by UNRES simulations for the wild type of FBP WW domain and its mutants at proper temperatures to obtain meaningful and considerable statistics that might provide insights into kinetics and thermodynamics. The results indicated the biphasic kinetics which are two-state and three-state routes. The folding kinetics was affected greatly by the mutations and truncation at C terminus but slightly at N terminus according to the experimental results by Nguyen etc. The distribution of structures at equilibrium states supported their suggestion of tuning the free energy landscape. This is still working in progress. More results will be showed further.3) Previous studies indicated that activation of the peripheral cannabinoid receptor 2 (CB2) had a strong inhibitory effect on acute nociceptive, inflammatory, and neuropathic chronic pain. However its analgesic mechanisms were not clear until now. We attempted to elucidate signal pathways by modeling structure of CB2 and docking programs. The results showed the existence of interaction between CB2 and Gs protein, which were also supported by experimental results, indicating a new signal pathway for inhibitory pain. However, this was just the start of this work. We would further study the soft structure of CB2 and analgesic mechanisms by UNRES approach.As a whole, we studied three specific cases of protein folding problem:the folding mechanisms of an eight-helix-bundle protein and a typical P-sheet protein, and protein structure prediction of CB2. These results provided us some interesting information and insights into the protein folding problem.