Concepts in Receptor Optimization: Targeting the RGD Peptide

Synthetic receptors have a wide range of potential applications, but it has been difficult to design
low molecular weight receptors that bind ligands with high, “proteinlike” affinities. This study uses novel
computational methods to understand why it is hard to design a high-affinity receptor and to explore the
limits of affinity, with the bioactive peptide RGD as a model ligand. The M2 modeling method is found to
yield excellent agreement with experiment for a known RGD receptor and then is used to analyze a series
of receptors generated in silico with a de novo design algorithm. Forces driving binding are found to be
systematically opposed by proportionate repulsions due to desolvation and entropy. In particular, strong
correlations are found between Coulombic attractions and the electrostatic desolvation penalty and between
the mean energy change on binding and the cost in configurational entropy. These correlations help explain
why it is hard to achieve high affinity. The change in surface area upon binding is found to correlate poorly
with affinity within this series. Measures of receptor efficiency are formulated that summarize how effectively
a receptor uses surface area, total energy, and Coulombic energy to achieve affinity. Analysis of the
computed efficiencies suggests that a low molecular weight receptor can achieve proteinlike affinity. It is
also found that macrocyclization of a receptor can, unexpectedly, increase the entropy cost of binding
because the macrocyclic structure further restricts ligand motion.