Significance of Knotted Structures for Function of Proteins and Nucleic Acids
Thursday Abstracts
Geometrical and Topological Aspects of Protein Structures
Alexey G. Murzin
.
MRC Laboratory of Molecular Biology, Cambridge, United Kingdom.
Proteins are biological objects subjected to natural selection. Protein structures are also governed
by the laws of physics and chemistry and, in general, combine both singular and regular features.
The former usually contribute to protein function and therefore tend to be evolutionarily
conserved, the latter result from the general principles of protein folding and contribute to overall
structural stability. The capability of protein structures to accommodate potentially unfavourable
sigularities greatly increases the number of possible folds suitable for functional selection. It also
allows the formation of peculiar architectures and topologies occasionally found in the protein
folds.
I will illustrate my overview of protein structure principles and evolution with the examples of
rare and unusual structural and topological features discovered since the creation of our
Structural Classification of Proteins (SCOP) database twenty years ago.
Experiment and Modeling: Competitive or Complementary Approaches to Structural
Biology?
Wladek Minor
.
University of Virginia, Charlottesville, USA.
The three-dimensional structures determined by X-ray crystallography play a central role in
understanding protein-small molecule and protein-protein interactions at the molecular level.
Each unique structure deposited to the Protein Data Bank (PDB) increases the number of models
that can be calculated (predicted) for experimentally unknown structures. The experimental
verification of models produced by the CASP competition shows that top experts can accurately
predict the overall structure of proteins where there is a similar protein of known structure, and
in some cases, even when a protein is not similar to any protein with a known structure.
However, the experimental verification of applicability of automatic methods developed for
meta-servers shows that the accuracy of a predicted model significantly drops when the sequence
identity between the model and an experimentally derived structure drops below 30%.
A combined approach of multidisciplinary experimental and computational methods will lead to
a dramatic increase in accuracy of structural predictions and computational screening. The
process will be discussed in the context of knotted proteins.
- 17 -
