Structure-based selection of building blocks for array synthesis via the World-Wide Web

In this article we are concerned with the selection of chemical entities for array synthesis in a structure-based design project. We have extended our conformational searching algorithm to permit the enumeration of a set of substituents at a particular position for a fixed template. The conformational space of each of the resulting structures is then explored within the confines of the binding site to identify conformations that do not interact unfavorably with the surrounding protein. The template remains fixed in its original orientation within the binding site. The interaction between each conformation and the binding site can also be quantified using various calculated properties. Each substituent for which one or more acceptable conformations can be found is retained for further analysis. Use of the program is facilitated by a Web-based interface that enables nonexpert molecular modelers to perform searches, view the results in a platform-independent manner (via VRML), and perform simple cluster analysis on the resulting sets of molecules. The approach is illustrated using a series of penicillin-based HIV-1 protease inhibitors.