Welcome back to Peking MOOC courses. This video is supplementary material , in this video we'll introduce you one of the This year’s Nobel Laureates in chemistry--- Michael Levitt, and his typical paper "Computer simulation of protein folding". Here is our team members. Our presentation is divided into two parts: Part 1：Introduction of Michael Levitt and his research, Part 2：Introduction of his representative papers published in 1969 and 1975 in detail. This is Michael Levitt's resume，he got his Ph.D at Cambrige University in 1971. Now,he is working at Stanford University. His research can be divided into three areas: physical simulation, heuristic structure prediction and structural bio-informatics which was most relevant with us. In his early research, he had committed to refine protein onformations using energy minimization procedure. His paper,"Refinement of Protein Conformations Using a Macromolecular Energy Minimization Procedure", which was published on J.Mol.biol in 1969 became one of the most representative paper in this area. Meanwhile,he was the first person who used energy minimization rules to refine an entire protein molecules. The Royal Swedish Academy of Sciences has decided to award the Nobel Prize in Chemistry for 2013 to Martin Karplus,Michael Levitt and Arieh Warshel. “for the development of multiscale models for complex chemical systems”. Here we briefly introduce the development of structure identification of proteins In the early studies, the method of analysis protein structure were X-ray crystallography of crystals or NMR-spectroscopy. And now the research focus gradually shifted from the study of protein structure to function studies. However, the chemical processes state is a microscopic process that occurs in a few microseconds in many chemical reactions. Everyone knows that hydrogen plus oxygen combustion water, but most reactions are larger and more complex. Thus, the details of this process are virtually impossible to describe using traditional methods of chemistry. That is why we need the help of the theoretical modeling to complement experiment. There is a simple addition formula to explain multiscale model which is referred in previous slide. Classical physics + quantum mechanical theory = multiscale model. The model is used for analyzing large and complex chemical systems and reactions. This two cartoons were Newton's apple and Schrödinger's cat. Previously, classical physics and quantum chemistry belonged to rivalling worlds. The Nobel Laureates in Chemistry 2013 have opened a gate between those worlds and have brought about a flourishing collaboration. Specific combination of classical physics and quantum chemistry showed on this figure. It's very simple, we can treat both as two labors. The central region in the system, the spacefilling atoms (red and gray), is described using a Quantum Chemical method. While the surrounding atoms are estimated by classical physics, and the outermost was seen as homogeneous dielectric medium. Next,we're going to introduce Historical development of multi-scale model: The first step in the development of multiscale modelling was taken in the beginning of the 70’s, when Arieh Warshel came to visit Martin Karplus at Harvard. Warshel had a background in inter- and intramolecular potentials and Karplus had the necessary quantum chemical experience. They constructed a computer program that could calculate the π-electron spectra and the vibration spectra of a number of planar molecules with excellent results. This was the first work to show that it is possible to construct hybrid methods that combine the advantages of classical and quantum methods to describe complex chemical systems. In 1975,Levitt and Warshel's study showed the folding of the protein from an open conformation to a folded conformation was studied, and it was shown that it is possible to group atoms in a classical system into rigid units and to treat these as classical pseudoatoms. Obviously, this approach further speeds up the modelling of a system. In 1976,Warshel and Levitt showed that it is possible to construct a general scheme for a partitioning between electrons that are included in the classical modelling and electrons that are explicitly described by a quantum chemical model. We will introduce this work in detail, which published on JMB in 1975 in the subsequent video by my paterner. Multiscale modelling in today，the methodology has been used to study not only complex processes in organic chemistry and biochemistry, but also for heterogeneous catalysis and theoretical calculation of the spectrum of molecules. Levitt said in Nobel banquet speech,"I thank the 2013 Chemistry Committee selected me and my fellow laureates, Martin Karplus and Arieh Warshel not for the field we pioneered, it is because we were brave and daring enough to recognize a new field." Here is some representative papers of Levitt,that relates to protein sequence alignment, structural alignment and 3D structure prediction. In the following vedio,my team members will introduce two classic papers in Levitt's early work. Hello， Now I will introduce the early work of Michael Levitt on manifold of protein. In his work of 1969, he proposed the idea of simulating the process of protein manifold by introducing energy minimization. For a protein molecular having thousands of atoms, it is important for computing purposes that the structure be conveniently and efficiently specified. As it is shown in the figure, the present procedure uses the common chemical notation, with additional symbols to describe the atomic connectivity uniquely. The total energy of the protein could be expressed as follows: the first term represents the bond energy for all bonds. The second term represents the contribution of angular of bonds to the total energy. The third term could be expressed as the contribution of Dihedral angle. The forth term represents the interaction between the non-bonding atoms, which could be expressed as Lennard Jones potential. The fifth term represents the constraint energy for all atoms. The figure below shows the contribution of different energy to the total energy of the system, from which we can find that during the simulation process, the total energy of the system continue to decrease the contribution of the different energy is different. This work adopted the MD simulation by classical mechanics. Therefore, it is of low efficiency and of finite size. In the following work, they adopted a quite different way to increase the efficiency. In their 1975 paper, Michael Levitt and his coworker tried to understand the relationship between protein sequence and conformation. Firstly, they simplified the representation of a protein by averaging over the fine details. Firstly, they simplified the representation of a protein by averaging over the fine details. Secondly, they simulated the folding of this simple structure by the combinant of convergent energy minimization and normal mode thermalization. Here they based on the following assumptions: Much of the protein’s fine structure can be eliminated by averaging The overall chain folding can be obtained by considering only the effective variables. Averaging over groups of atoms in the full structure gives a simplified structure with each residue represented by only two centres, the Calpha atom, and the centroid of the side chain. So simple a model can represent the stable conformation of a folded protein, Now let’s simulate the actual process of folding. The folding of this idealised protein can be simulated by solving the equations of molecular dynamics at sufficiently small time intervals. In a viscous medium like water, these equations of motion can be approximated by Langevin equations, where the change in the variables is directed down the energy gradient with a random deflection due to Brownian motions. For greater computational efficiency they neglected thermal fluctuations during simulation, and the end point of the trajectory is the potential energy minimum accessible from the starting conformation. The figure below showed the results of simulation: Simulation of PTI folding from an extended starting conformation with the terminal helix The 8 ribbon diagrams, which show the Ca chain om left to right to the 8 conformations at the circled points on the r.m.s. deviation curve, respectively. The last five conformations have progressively lower energies and are each a little closer to the native structure The solid dots at the end of a minimisation indicate that a perfect minimum was reached Now let’s give these work a summary: A model based on time averaged forces can account for the stability and folding of a molecule as complicated as a protein.Folding must depend on a very rare random fluctuation that happened tobring the right residues close together with sufficient precision for the short-range forces to take effect. The group members are grateful for the help from Professor Ge Gao and Liping Wei. We also appreciate the hard work and the patience of the teaching assistant group. Thanks for your attention!