Design, synthesis, and calorimetric studies on protein-ligand interactions : apolar surface area, conformational constraints, and application of the Topliss decision tree

A preorganised amino acid derivative containing a cyclopropyl constraint was designed to orient an amino acid into its bound conformation. This constrained mimic was determined by ITC to be equally potent to the native Phe derivative. It was found that a more favorable enthalpy of binding was compensated by an equally unfavorable entropy compared to the native ligand. In order to properly ascertain the effects of the cyclopropane constraint, a flexible control containing the same number of heavy atoms was synthesized and tested, and it was found to be at least 200 fold less potent than the constrained analog. However, without structural data of the flexible control, it is difficult to infer if the differences in ligand binding affinity arose from the ligand constraint or some other unknown complexity to binding. We studied the thermodynamic and structural effects of modifying alkyl chains of n-alka(e)nol and phenylalka(e)nol binders to MUP-I by both the removal of a rotor via deletion of a methylene unit and restriction of a rotor via the installation of an internal olefin. In general, we observed that a similar thermodynamic signature accompanies modifications for both the n-alka(e)nol and phenylalka(e)nol ligands: A favorable T[delta][delta]Sºo̳b̳s̳ is compensated by an unfavorable T[delta][delta]Hºo̳b̳s̳ such that T[delta][delta]Gºo̳b̳s̳ for both removal of a methylene and insertion of an internal olefin are unfavorable and equipotent, respectively. The insertion of an internal olefin into an alkyl chain led to significantly more favorable entropies than does methylene removal, yet enthalpy-entropy compensation leads to nearly equipotent binding energetics. However, we did find a strong correlation between [delta]Ho̳b̳s̳° and buried apolar Connolly Surface Area (CSA). The intrinsic free energies of introducing an internal olefin into the n-alkanols and phenylalkanols differ markedly from the observed data. It was observed that intrinsic affinities are more favorable than the observable because a favorable T[delta][delta]S⁰i̳n̳t̳ dominates an unfavorable [delta][delta]Hºi̳n̳t̳. Also, we discovered that the intrinsic entropies of inserting an internal olefin are nearly double that of removing a methylene group, suggesting that the insertion of an internal olefin results in the restriction of more C-C rotors. We have shown through ITC analysis that the added substituents probed in this study provided binding increases to our Grb2 SH2 ligands as expected, but that the thermodynamic driving force of binding affinities depended greatly upon the specific nature and flexible mobility of the ligands in the binding pocket. Through a combination of X-ray and ITC studies it was shown that ligands containing rigid and aromatic functional groups bound with a higher [delta]H° than the more flexible alkyl ligands, and that this effect correlates well with more direct vdW contacts made in the pocket. Finally, we described a case study where a strict adherence to the Topliss operational schemes led to an expedient development of novel MUP-I binding analogs. The validity of the schemes was also depicted through the synthesis and testing of ligands that were correctly predicted to be weaker/equipotent to the starting ligand. Of important note is that the degree to which the schemes led to affinity boost depended greatly on the starting potency of the initial compound.