In theory the 3-dimensional structure holds the key to the complete understanding of the function of the enzyme, e.g. stability, activity and substrate preferences. We do, however, still lack the understanding and sufficient theories to completely utilize these structures in the process of engineering new functionalities. We utilize a number of computational tools in order to assist the extraction of this information. These tools include electrostatic calculations, substrate docking, active site architecture and molecular dynamics of the molecule.
Many structural features cannot be unveiled without the aid of advanced stereographic visualisation, which allows total freedom in inspection of 3-dimensional structures. Graphic visualisation tools confirms, refute and inspires new ideas and is therefore an important point of human intervention in the design process and as such completely non-replaceable by any computational method in designing new enzyme variants.
For the purpose of designing new enzyme variants much information can be extracted from the natural diversity in sequence space. Comparing enzymes of homologous sequence, but originating from species adapted to different environments, can disclose features related to different functionalities, i.e. conserved features may relate to enzymatic activity or structural integrity. Whereas sequences from thermo-stable, alkaline or psychrophilic organisms may provide clues for engineering such desired features into the structure of interest.
We use all this information extensively in combination with sequence-derived data in the design of new enzyme variants.
In our analysis of dynamic properties we use high-performance computational facilities, enabling simulating dynamics o
f complex solvated systems and detailed electrostatic characterisation. Our computational clusters allow us to develop in silico screening assays, improving both speed and quality of subsequent engineering work.
Occasionally the desired structural information is not available, either because the protein structure has not yet been solved or because the studied object is a protein-protein / protein-ligand complex. In the latter case computational docking methods can be employed to obtain a structural model. Our requirement to docking is somewhat different as compared to pharmaceutical small molecule docking, since a typical enzyme substrate is significant larger than a drug molecule and thus computationally a much more difficult task.
Another important distinction between docking as used in drug design and enzyme substrate docking is that in the latter case the substrate is invariable whereas the protein is variable (mutable) and vice versa in drug design. Thus our purpose of substrate docking is to change substrate specificity/affinity by tweaking active site of the enzyme.
Building homology models, referred to as homology modeling or molecular modeling can compensate lack of a 3-dimensional structure. The prerequisite for homology modelling is the availability of one or more closely related protein sequences with known 3-dimensional structures. In such cases an alignment between a target sequence of unknown structure and sequences of known structure are created and subsequently used as guide for calculating a structural model of the target sequence. Reliability of homology-modelled structures depends on the degree of homology between sequence of known structure and the target sequence and as such determines what the model can be used for. For example structural models based on low homology should
never be used for substrate docking studies. On the other hand, models based on high homology can be nearly as good as structures determined by X-ray crystallography.