Fragment-Based Drug Design: How Big Is Too Big?

Much has been discussed about the proper physicochemical properties (e.g., molecular weight, hydrophobicity,
etc.) that should be considered when utilizing fragment leads in drug design. However, little has been reported
as to what emphasis, if any, should be placed on the potency of the resulting fragment leads. In this report,
a retrospective analysis of 18 highly optimized inhibitors is described in which the compounds were
systematically deconstructed until the minimal binding elements could be identified. An analysis of the
potency changes that were observed as the leads were reduced in size indicate that a nearly linear relationship
exists between molecular weight and binding affinity over the entire range of sizes and potencies represented
in the dataset. On the basis of these observations, prediction maps can be constructed that enable critical
and quantitative assessments of the process of lead identification and optimization. These data place well-defined limits on the ideal size and potency of fragment leads that are being considered for use in fragment-based drug design.