Computer Assisted Drug Design

Department of Pharmaceutical Chemistry,

University of Kuopio

Drug design is a three-dimensional puzzle where small

drug molecules, ligands, are adjusted to the binding
site of a protein. The factors which affect the protein-
ligand interaction can be characterized by using
molecular docking and different quantitative
structure-activity relationships (QSAR) methods.
These results can be represented by property maps as
those shown in Fig. 1.

The most commonly used tool to model biological system

is molecular dynamics. The model of a receptor
refined with molecular dynamics simulations is presented
in Fig. 2.

Figure 1. In CoMFA map the colored fields describe how

molecular structure can be modified to increase biological
activity.

Virtual Screening is a computational technique to find

novel drug candidates. Data from virtual screening can
be used to develop predictive models in order to
optimize ADMET (adsorption, distribution, metabolism,
elimination, toxicology) properties of the candidate
molecules.

The ultimate goal of this procedure is to find

interesting lead molecules that are worth for further
drug research and synthesis by either our laboratory or
industrial partners (Fig. 3).

Figure 2. 3D models of membrane receptors can be refined and validated

in a realistic lipid-water-salt environment using molecular dynamics
simulations.

Applications:

•Find interesting lead molecules quickly

•Predicting properties and activities of untested
molecules
•Propose compounds for synthesis
•Validate models of receptor binding sites
•Optimize pharmacokinetic properties of Figure 3. New potent inhibitor for the Human Sirtuin Type 2 enzyme was
found using a virtual screening technique.
compounds

Further information
Prof. Antti Poso, email: Antti.Poso@uku.fi
http://www.uku.fi/farmasia/fake/modelling/index.shtml