In order to make the molecular dynamics calculations faster the C-terminal arm of the protein, beyond the sixtieth residue, from the original structure, as well as one base pair at each end of the DNA were cut. In this way one could reduce the size of the system and thus the number of water molecules needed for the simulation. The terminal residues had to be modified (e.g. an oxygen had to be attached at the C-terminal residue of the polypeptide), in order to have a chemically correct structure.
The first step before the actual molecular dynamics simulations was to energy-minimize the four complexes--the wild type, the one with only the protein mutated, the one with only the DNA mutated, and the one with both mutations. This procedure as well as subsequent molecular dynamics simulations were performed using the OPALp program [16] and the AMBER force field [17]. Energy minimization takes care to a certain degree of most steric and van der Waals conflicts, which otherwise could result in unstable molecular dynamics simulations.
The resulting complex was fixed by applying position constraints to the heavy atoms for the initial equilibration procedure. It was then immersed into a water bath, which was preequilibrated. Periodic boundary conditions were used. The size of the periodic box was determined by the water thickness parameter. In these simulations, a thickness of 12 Å was used. The water molecules, which were in steric conflict with the complex, were removed. This resulted in approximately 8500 water molecules, which were included in the simulations. The next step was to let the water bath adapt to the fixed macromolecules. The water description used in these simulations included energy terms for two real bonds and one virtual bond between the hydrogens. This virtual bond replaced the energy term from the angle between the two real bonds, in order that the SHAKE procedure, which keeps the bond lengths cosnstant and thus speeds the simulations, could be applied. The reference temperature for the first two picoseconds of the run was 50 K, and afterwards 300 K. The temperature relaxation time was 0.1 ps, and the volume was kept constant. The position constraints were removed after 5 ps of the run.
The structures, which served as raw results, were saved after each picosecond. Because of the periodic boundary conditions, the water molecules, which crossed the boundaries, had to be moved to the initial primitive cell by applying a translational symmetry operator. This step was was necessary in order to analyse the data.