IntroductionThe drug discovery industry has never been under such pressure to deliver as it is today. While R&D costs spiral higher, attrition statistics in clinical development show no signs of improvement and may actually be getting worse. For example, the FDA highlighted in March 2004 that a drug entering phase 1 clinical development in the year 2000 was less likely to reach the market than one that entered clinical development in 1985 [1]. It is a truism that much of the fate of a drug candidate in clinical development is embedded in the chemical structure and, hence, is in the control of the chemist at the point of design. Clearly, there are serious limitations on what drug designers have so far learnt about the relationship between chemical structure and attrition through clinical development.Drug designers are ever-optimistic, and the hope is that an increasing understanding of the interactions of drug candidates with their protein targets, at the molecular level, may allow quality to be built into drug candidates at the design stage. Indeed, there are a growing number of targets that we are trying to inhibit or activate for which we do understand the relationship between inhibition or activation of the target at the molecular level and their pathophysiological effect. Advances in structural and molecular biology, as well as in biophysics, have led to the determination of high-resolution atomic structures of many of the protein targets of drug discovery projects. For instance, p38 kinase is a target that is well validated in the clinic with respect to its role in inflammation, and many high-resolution crystal structures are available. The same is true for thrombin for anti-thrombosis and renin for hypertension, iNOS for inflammation, EGF receptor tyrosine kinase for cancer, as well as many antibacterial and antiviral targets. Crystal structures of some of the most important drug-metabolising enzymes are also known [2-7], which potentially enables the structure-based rational optimisation of potency, selectivity and metabolism. However, routine control of potency, selectivity and metabolism based on the use of structural information has yet to become a reality.While high-resolution protein structure information can be derived by both NMR spectroscopy and X-ray crystallography, most structures by far have been determined by crystallographic methods. The application of NMR spectroscopy to routine protein structure determination is limited, as it requires large quantities of soluble, multiply labelled protein, considerable time and is limited to comparatively small proteins. Solid-state NMR spectroscopy is being applied to
Reviews
KEYNOTE REVIEWThe structure model of the designed compound shown with the 2Fo-Fc map contoured at 1s, showing only weak density around the putative position of the ethylamine sidechain.