Computational chemistry & structural biology
Finding and optimizing target proteins on the computer
An active substance interferes with a pathological process by inhibiting or stimulating a target protein (target) that is actively involved in this process. For this it needs to bind to this target – very specifically, like a key in a lock. Methods of structural biology and computational chemistry (also known as molecular modeling) are indispensable today in the search and development of such active-substance molecules. By providing modern measurement techniques and making digitally calculated predictions, they enable synthesis chemists to concentrate in the lab on the most promising drug candidates.
X-ray structure analysis reveals the 3D structure
The scientists must first determine the exact molecular structure of the target protein before they can start searching by computer – i.e. in silico. This is done with the help of X-ray structure analysis: each protein crystal's lattice structure diffracts the X-ray beam in a characteristic way. By analyzing the resulting diffraction pattern, it is possible to read off the density of electrons in different parts of the target protein and thus determine the position of the atoms. This process is repeated several times, further refining the image until the researchers can see the precise 3D structure of the target molecule.
The molecular structure is examined, both alone and together with a bound active substance. This enables the researchers to find out where in the protein the so-called binding pockets offer the possibility of a reaction with the active substance and what form the interaction between protein pocket and the active substance will take.
Computational chemistry: in-silico drug discovery
The crystallographic data and the digital 3D model of the target protein form the basis of chemists' work in the field of computational chemistry. They use the computer to search – for example in virtual compound libraries – for molecules whose structure fits the target molecule's binding pockets.
Drug candidates that are found digitally must subsequently prove their suitability in the real world. They are either produced by synthesis chemists in Bayer HealthCare's laboratories or purchased from outside suppliers. They are then tested in a test tube (in vitro) to check the computer-predicted effect.
Digital forecasts improve success rates
Computational chemistry is also a valuable tool in the subsequent steps of the drug-discovery process. Once a molecule with the desired binding properties has been found as a starting point, this lead compound's suitability can be further improved on the computer; for example, it is possible to calculate digitally what changes could improve the drug candidate's binding affinity, i.e. its ability to bind to the target. The scientists can also use this method to calculate in advance any biophysical or toxic properties that would result from certain structural changes.
Forecasts and in-silico molecular optimization save time and enable researchers to concentrate on suitable drug candidates at an early stage. This can lead to a decisive increase in the success rate of the subsequent – more complex – laboratory tests.