Fragment filtering goes virtual

Scientists have discovered a powerful approach to fragment-based drug discovery, a technique that could provide new drug leads.

Fragment-based drug discovery―identifying small chemical pieces that can then be developed into drug leads―has been very successful. However, the current approaches used to screen fragments are not 'high throughput', so only a small number of fragments can be tested for each drug target.

Online in Nature Chemical Biology this week, Yu Chen and Brian Shoichet wanted to know if computational docking?a method to predict how a chemical will interact with a protein―could be used to screen a large collection of chemical fragments at once. First they used computational docking to search for fragments that might bind to beta-lactamase, a challenging drug target. They then separately determined the enzyme's structure with the computationally identified fragments bound, which revealed that the computer predictions had been accurate. After chemical optimization, the fragments were converted into good beta-lactamase inhibitors.