I. ALL-ATOMISTIC MD CALCULATIONS
We are extending all-atomistic MD calculation study to viruses which are conventionally a target of experimental studies only in the field of microbiology. Picorna viruses such as polioviruses and mouth and foot disease viruses are small with the diameter of about 30 nm consisting of about ten million atoms including electrolyte solutions as an environment. It is possible to handle entire viruses using highly parallelized high performance supercomputers such as the K-computer at RIKEN combined with our high performance highly parallelized MD program, MODYLAS. The figure is a snapshot of MD calculation for poliovirus in electrolyte solution. Molecular structure and its fluctuation in the solution, stability, response to high pressure, and chemical environment inside the capsid have been investigated. Further, interaction between the capsid and its receptor has also been studied since it is very important as an initial process of infection. The interactions are evaluated quantitatively based on the free energy calculations. Now, the calculation is being extended to hepatitis B viruses (HBV) containing RNA and its antiviral reagents.
Fig. All-atomistic MD calculation of poliovirus. (left) The calculated empty capsid in solution formed by 240 capsid proteins. (right) Stroboscopic picture of a water molecule crossing the capsid.
Polymers are one of the typical materials whose molecular weight is very large and their motion is very slow. This is the reason why, conventionally, the polymers were out of scope of the all-atomistic MD calculation studies. So far, universal properties of the polymers have been actively investigated based on coarse-grained models and statistical mechanical models. However, peoples are very eager to describe diversity, that is, chemical details of the polymer properties starting from chemical structure of the monomer. This is very important in the development of new polymers. Now, we are doing all-atomistic MD calculations for polymers focusing our attention on very high speed phenomena such as impact fracture, standing on a different position from the traditional way of polymer simulation. The figure is a snapshot of impact fracture of polyethylene. The yellow circles are radicals formed by the breakage of chemical bonds. Polymer electrolyte membranes such as Nafion and Flemion have been investigated, too, focusing on their morphology or microscopic phase separation structure as a function of water content.
Fig. Impact fractures of polyethylene(center) and polycarbonate(right). Yellow circles are radicals formed by the breakage of the chemical bonds.
3. LIPID BILAYERS
(1) Real cell membranes
Physico-chemical properties of normal cell membranes and cancered ones have been investigated based on all-atomistic MD calculations. Leukemic and hepatoma cell membranes have been compared with their normal membranes. The figures are snapshots of the leukemic and normal thymocyte cell membranes of mouse. The cancered cell membrane shows relatively more disordered structure and is found to be, from the dynamics analysis, more fluid than the normal one. Penetration and absorption of the reagents are also investigated by the calculated free energy profile based on thermodynamic integration method.
Fig. Normal thymocyte (left) and leukemic cell (right) membranes. The cancered cell membrane shows disordered structure and is more fluid.
(2) Permeability, Transportation of small molecules across lipid membranes
Permeability of small molecules through lipid membranes has been investigated by molecular dynamics simulations on the basis of inhomogenous solubility-diffusion model. In this approach, free energy profile and local diffusion coefficient of the small penetrant are needed to be calculated. Several free energy methods have been well established for a small molecule. For example, free energy profile of a single water to go across a lipid membrane is precisely calculated by a combination of overlapping distribution method and cavity-based Widom insertion method. (Figure) However, it is still not straightforward to compute the free energy profile for rather large molecules such as peptides and nanoparticles. In this case, the permeation can be a collective dynamics involving multiple molecules; namely, the choice of the reaction coordinate or collective variables is nontrivial. We are trying to develop a simulation method to understand such a complex molecular mechanism. We are also interested in the kinetic factors to characterize the permeation process.
Fig. Free energy profile of a water molecule to go across the DPPC membrane (J. Comp. Chem. 2008)
Micelles are formed in solutions as assemblies of surfactant molecules having both hydrophilic and hydrophobic groups. They are key materials in a number of practical and industrial applications such as soap, cosmetics, medicines, ink, flavors, and foods. In our group, structure and dynamics of the spherical micelles as well as their formation dynamics have been investigated. Further, solubilization of insoluble solutes has been studied based on free energy calculations. Structural stability has also been studied evaluating surface tension and free energy.
Fig. Snapshots of MD calculation for SDS micelle formation in water and ithe free energy profile as a function of micelle size.
5. MEMBRANE PROTEINS
Interactions between protein and reagent have been investigated. In particular, the interactions have been analyzed from a view point of side effect of the reagent. For example, a reagent should inhibit a particular protein, but it shouldn’t inhibit other proteins. Free energies have been evaluated taking account of structural change of the protein of interest.
Fig. A GPCR protein in the model lipid bilayer.