Abstract
I verified the computer simulations of the structure and the
dynamics of beta-cyclodextrins and water in
crystalline structures with a comparison of the root mean square
fluctuations of the molecular dynamics (md) simulations and the
crystal structure experimentally solved at different humidities by Dr
Steiner 1994 in the group of Prof Dr Saenger, which results in good
agreement for the atoms in the glucose rings. The outer atoms have a
larger flexibility in the simulations than in the experiment. I
investigated the different types of disorder in the molecular and in
the crystalline part. This shows that the crystalline part looks like
the molecular part of disorder for the atoms of the glucose rings.
The molecular part of the disorder is larger than the crystalline
part of the disorder for the outer atoms. This additional
characterisation of disorder is possible on simulations, but
difficult in experiments. The effects of the flexibility on the
distribution of some torsion angles and other structural parameters
of the beta-cyclodextrins were investigated
too. Some deviations from the experiment with respect to sugars
became evident. These differences are due to a weakness in the
quality of the energy function in CHARMM. All of my comparisons
indicate, that there is a good principal agreement between the
measurement and the simulation data describing the structure and
dynamics of the crystals of the beta-cyclodextrins.
The pattern and the lifetimes of hydrogen bonds in the crystal
were calculated. The hydrogen bonds stabilize the structure of the
crystal and the beta-cyclodextrin molecules.
I get results for the structure and dynamics of the hydrogen bonds
from the md simulations. These results correlate with the
experimental results qualitatively in many aspects and quantitatively
in some aspects. There is an obvious dependency on the humidity of
the crystal. There are not enough water molecules in the crystal to
stabilize the structure at low humidity. At high humidity the water
molecules compete with one another, so they do not form larger
systems of hydrogen bonds.
The most important point of the evaluation of the md simulations
is the dynamics of the water molecules in the crystal. I determine
the diffusion constant and path of water in the crystal. To do this,
I have developed a fast algorithm to derive the most probable drift
direction. It was possible to find a diffusion path through the
crystal, which goes through the beta-cyclodextrin
rings. The inner and outer water molecules are connected by this
diffusion path.
The newly developed simulation program has advantages in CPU time
compared to CHARMM. This is due to the usage of a list of cells
instead of the Verlet list.
| 
