Hit identification by screening
A drug most often is an organic molecule with a m. w. of a few hundreds of Dalton (less than 500 according to Lipinski rule of 5). But first we need to identify a compound, also called a hit, which by binding to the protein will modify its function. SInce most initial hits usually have low affinity towards their target, they need to be modified in order to increase their affinity towards the protein in question and to optimize other properties, like solubility, toxicity, etc. Apparently we also need an activity assay, which will allow us to study the effect of the compounds on protein activity as well as its activity in cell cultures. In a traditional drug discovery project people often use the method of high-throughput screening (HTS), when libraries of drug-like compounds, which may consist of hundreds of thousands, or even millions of compounds, are screened against the protein target. Such screening may identify one or few compounds with the ability to inhibit the activity of the drug target. If the three-dimensional structure of the protein in question and crystallization conditions are available, the compound or compounds, if more than one has been identified, may be co-crystallized with the protein to obtain the structure of the protein-ligand complex. This will help in mapping the structure of the binding site and will provide the information required for optimization of the compound identified. This will be the beginning of a structure-based drug design project.
Figure showing an inhibitor bound to the Plasmodium falciparum spermidine synthase (Dufe et al. 2007)
Another, continuously growing in popularity approach in drug discovery is screening of the drug target against libraries of small molecule fragments (molecular weight not larger than 300 Da), a so-called fragment-based drug design (FBDD). Since the molecules used in this case are much less complex in structure, than drug-like molecules, the probability of finding good binders is much higher. Fragment-based drug design also allows keeping the size of the compound library much smaller (few hundreds compounds, rather than hundreds of thousands) and may provide a better sampling of the chemical space. The task in this case is to find weak, but efficient binders (high binding energy for each atom), which may subsequently be grown into more potent compounds with drug-like properties. This process allows a much better control of the properties of the molecule in terms of solubility, toxicity, etc. After a decade of positive results, starting from the identification of the first leads and extending through to the stage when drug candidates can enter clinical trials, it has become evident that the combination of fragment-based drug discovery and structure-based drug design is more superior to “traditional” drug discovery methods (P.J. Hajduk & J.Greer, 2007).
Structural information in drug design
Although homology modeling may be used to guide the design of new inhibitors, the success of a structure-based design project relies heavily on the availability of crystals of the drug target. The structure of the complex with the ligand will provide detailed knowledge on the binding interactions, like the number of hydrogen bonds, hydrophobic interactions, distances between the interacting groups, the presence of water in the active site, etc. A protein-ligand complex may also contribute to a better understanding of the mechanism of enzymatic activity if, for example, the drug target under investigation is an enzyme. Such knowledge may in turn contribute to improve the potency and selectivity of the inhibitor. In drug discovery and structure-based drug design the information on the ligand and its interactions may also be used in the design of smaller, focused small-molecule libraries of compounds with higher potential to bind to the protein target.
Compound libraries are screened in activity assays to identify a series of best binders with closely related molecular structure, which will help in the formulation of the structure-activity relationships (SAR) hypothesis, and eventually result in the identification of best ligand structure. This may be further optimized with regard of its toxicity, solubility, etc., to produce a so-called lead candidate, and finally a lead molecule or lead series. This process normally involves a repeated determination of the structure of the drug target in complex with many compounds. Sometimes hundreds such structures may be determined before a desired molecule is created.
This is just a short outline of the process. A separate chapter on drug design will be included in the up-coming iBook on structural bioinformatics.