Pharmaceutical industry is facing a key challenge in terms of declining pipelines of NCEs despite exponential increase in R&D spending and breakthroughs achieved in several technological fronts. One of the key factors for the future success depends on the ability to generate quality lead molecules that addresses late stage attrition during the development phase.
Novel approaches and newer technologies are being tried by pharmaceutical companies to address this issue. A structure based strategy for new drug discovery has assumed a wider importance in this regard with progress in high throughput structure determination and screening methods. A relatively novel technique for lead generation using structural biology methods is fragment based drug discovery. Fragment based screening is an exciting development which is increasingly used for generation of high quality lead molecules.
Fragments are small organic molecules (typically 150-250 Da), which bind to the target protein with a low affinity (mM-mM). Compared to the conventional mature compound libraries, screening of fragments allows coverage of greater chemical diversity space with fewer molecules. Because of the low binding affinities, often it is not possible to detect these fragments in an HTS campaign. However, advancements in high throughput NMR and crystallography have led to the possibility of using these techniques to detect very low affinity ligands.
Advancements in structural biology techniques such as automation in protein expression, purification and miniaturization of crystallization and the availability of automatic structure solving algorithms have made high throughput crystallography realistic. Availability of high field NMR with auto-sampler, cryo and nano probes combined with better methods of protein labeling has made faster and high quality data collection by NMR a viable technology for the lead discovery strategies.
Fragment screening and Optimization
The success of this approach depends on careful selection of fragment libraries for screening. Structural information and computational techniques are employed for selection and prioritization of compound libraries. Typically screening of fragments can be done using both NMR spectroscopy and crystallography. Protein-ligand interactions can be monitored by NMR using either ligand based or protein based techniques.
In the ligand based method, the proton spectrum of the ligand is monitored using Saturation Transfer Difference (STD) and water LOGSY techniques. In the protein based approach, the perturbation in the chemical shift of amide (NH) peaks in the two dimensional heteronuclear single quantum coherence spectrum (HSQC) is monitored.
Although precise structural information of the fragment binding is not revealed by STD and water LOGSY methods, the throughput is fairly high to allow screening of fragment libraries. HSQC experiment provides the structural information but requires labeled protein and assignment of the amide peaks. Screening by crystallography is carried out either by soaking a cocktail of fragments with the pre-formed crystals or by co-crystallization.
To improve the throughput of fragment screening a combination of NMR and crystallographic screening can be a used to quickly identify and prioritize the fragments. An additional advantage of this approach is the potential for identification of secondary or allosteric binding sites. These non-competitive binding events may not be specifically detected by enzymatic assays.
The secondary/allosteric binding sites can be exploited while developing selective molecules. Further optimization of fragments can be achieved using structure guided medicinal chemistry approaches by linking or evolution of fragments.
Recent developments in fragment based approaches can accelerate the early stages of drug discovery. The key success of this method is the utilization of biophysical methods, most commonly NMR for screening of fragments and crystallography for the determination of bound conformation of fragment. This highly integrated approach can be adapted to various target classes to meet the needs of any drug discovery programme. However, a few key challenges still remain in this area. Structures of some of the major protein families such as GPCRs need to be determined. Predictive power of computational techniques needs to improve in tandem with advancements in structural biology in order to fully exploit the benefits of fragment based discovery.
(The authors are part of Structure Based Drug Design Group, Aurigene Discovery Technologies Limited, Bangalore.)