Special case of membrane proteins

In this case it is absolutely necessary to embed the protein in a coherent lipid bilayer model, along with all the necessary water and ions or counter ions. This results in a very large system and the atoms belonging to the environment of the protein account for six to seven times the atoms of the protein (remember that computational times are proportional to the square of the number of atoms). The PBC convention is used throughout, allowing in fact the creation of a quasi-infinite multilayered system (Figme 9.14). 

If you want to know the number of atoms contained in any box, it is only necessary to calculate the volume of this box in and divide this number by 10. This tip works well with water as a solvent. Note that in organic solvents the number of atoms will largely be less than this calculated value. 

First of all, the use of molecular mechanics and dynamics pre-suppose a three-dimensional structure or a model built by sequence homology. Even if molecular mechanics involves large simplifications, it is an extremely powerful and reliable technique to study organic systems, except for reaction pathways. Molecular dynamics is a simple technique which is almost insensitive to temperature but extremely sensitive to the chosen time-step for the integration of the Newton s law. If the main interest of the user is to explore the conformational space of a system, then heating is the best choice (up to 1000 K or more). 

With the most up-to-date supercomputers the space scale of the system will be, at most, 10 X 10 X 30 nm (or equivalent) and the timescale at most 100 ns for systems in the 10 x 10 x 10 nm space scale. In this case, calculations are huge and may imply several weeks or months of nonstop simulations. Apart from these drawbacks, there is almost no limitation to the size or complexity of the simulated systems. However, it has always to be remembered that the behaviour of biological systems is extremely slow at an atomic scale and will always necessitate huge amounts of calculation. 

At present, starting from a protein one-dimensional sequence in order to build a three-dimensional structure is hopeless even for very short sequences. The best way to obtain access to a three-dimensional model is to perform sequence homology building, starting from an X-ray or NMR structure of a similar protein. 

---