RMSD test

One such criterion is the time dependence of the root mean square deviation between the complex at a given time and at the initial time. This criterion gives information on the stability of the complex or of a part of it.

  

Figure 3.1: Time dependence of root mean square deviation in two runs (wild type and protein mutant). The parts selected for the analysis are protein backbone from residue 8 to 59, DNA bases from base pair 4 to 11 and both of them.

If the complex (or its selected part) is stable, the RMSD, after an initial increase, should fluctuate around a constant value. The results for the wild type complex and for the complex with the protein mutant provide an example of such behaviour (Figure 3.1).

  

Figure 3.2: Time dependence of root mean square deviation in two runs (DNA mutant and double mutant). Selections are the same as in the previous figure.

  

Figure 3.3: Time dependence of root mean square deviation for all DNA bases in the four runs. The DNA of the double mutant complex starts dissociating at the termini toward the end of the run (cyan curve).

Figures 3.1 and 3.2 also indicate that the protein backbone, excluding the flexible N-terminal arm and the C-terminus, is stable in all four runs. On the other hand, the DNA is less stable. Especially the doubly mutated complex shows a constant increase in the RMSD for the DNA bases (base pairs 4-11). The latter result suggests that caution should be taken in analysing the behaviour of the doubly mutated complex.

Figure 3.3, where all bases of the DNA are considered (base pairs 2-13), also points to the doubly mutated complex as ``ill-behaved'' toward the end of the run (after 160 ps). The results of the double mutant complex simulation are therefore dealt with more cautiously in the following sections. A more general conclusion, which may be drawn from the above figures, is that the mutation in the DNA appears to distort the system more than the mutation in the protein. One reason for this is that the mutation in the protein is achieved by changing few atoms, which belong to the end of a flexible side chain of the protein and are far from its backbone, whereas the mutation of the DNA involves significantly more atoms, which constitute the core of the molecule. Moreover, the resulting protein mutant may be energy-minimized by rotating the flexible side chains, a possibility almost lacking in the DNA, where the only adjustment was the rotation around the glycosidic bonds of the mutated bases. An intrinsic property of these simulations, that could have contributed to the instability of the DNA, was the absence of counter ions, which essential in vivo.