The impact of protein fluctuations on molecular recognition

The effect of protein fluctuations on molecular recognition is poorly understood. Prediction of useful properties such as binding affinity using rigid structures has produced sporadic success. Although attempts have been made to model the effect of\r\nconformational fluctuations, capturing the impact of backbone relaxation has remained\r\nparticularly elusive. In order to investigate these effects, a series of surface exposed\r\nAla/Gly mutants were designed in the flexible RT loop of the C-terminal SH3 domain of\r\nSEM5. One set of mutations was designed to perturb the ensemble of accessible\r\nconformations in the unbound ensemble while leaving the interaction surface with the\r\nligand unchanged. The other set was designed to perturb both the interaction surface as\r\nwell as the ensembles of bound and free conformations. The effects of these mutations\r\nwere investigated by generating random conformations of the RT loop and performing\r\nprincipal component analysis to organize the randomly generated conformational states\r\ninto a coherent landscape. To predict the effect of these mutations, we developed a\r\nstatistical mechanical technique using a simplified energy function that only applied the\r\neffects of excluded volume and implicit solvation. This energy function was utilized to\r\nweight an ensemble of conformational states from which aggregate thermodynamic\r\nproperties could be derived. The computed effects of the mutations on the binding\r\naffinity agreed with experimentally determined values (R= 0.97) from isothermal titration\r\ncalorimetry. The results indicate that the bound state of SEM5 SH3 domain contains a\r\nconsiderable repertoire of conformational variants of the high-resolution structure and\r\nthat the determinants of binding cannot be elucidated from the static structure of the\r\nbound complex.\r\n