Molecular recognition by the EWS transcriptional activation domain

Abstract

Ewing’s sarcoma fusion oncoproteins (EFPs) are produced from aberrant chromosomal translocations involving the EWS gene. Transcriptional activation domain of these proteins, EAD, is provided by EWS and is central to oncogenesis. It was previously established that EAD is an Intrinsically Disordered Protein Region (IDPR) and its transcriptional activity requires multiple aromatic groups. However, biophysical basis for the molecular recognition by EAD remained unclear and is the focus of this study. Generally, IDPRs are important for protein-protein interactions and diseases. Some IDPRs, including EAD, remain disordered even when bound to their partners and such interactions are termed “fuzzy”. Fuzzy interactions are characterized by IDPR polyvalency, sequence-insensitivity and a dynamic ensemble of disordered bound-state conformations. The latter characteristic limits classical structural analyses. Cation-π interactions between aromatic π electrons (in Y/F/W) and cationic amino acids (R/K) are prevalent at protein-protein interfaces and we hypothesized that cation-π interactions are central to molecular recognition by EAD (polycation-π model). This hypothesis was evaluated by functional interrogation of a range of EAD sequence variations, including charge perturbations, aromatic number, density and moiety. The results are strongly supportive and EAD-target binding is further seen to be influenced by EAD conformational entropy. The model was also tested by molecular dynamic simulations (in collaboration with others) and this approach broadly captured functional effects, thus strengthening the model. EWS protein, and therefore EAD, most likely acts as scaffold that interacts with a large number of protein partners, but functionally relevant EAD partners have yet to be identified. To address this question, a two-hybrid screen was performed to search for specific EAD-binding proteins (EABPs) that interact via EAD aromatics. Novel EABPs were identified and the binding determinants of one (DYNLT1) were determined and found to be generally consistent with the polycation-π model. Functional relevance of DYNLT1 is not clear but it provides a potential system for direct analysis of EAD-target interactions. In conclusion, the mode of molecular recognition by EAD is unprecedented and may reflect a general solution for fuzzy interactions and promiscuous target binding by polyvalent IDPRs. The findings are also of general interest in that the mode of molecular recognition described is specific for abnormal, oncogenic EAD function and is not strongly shaped by evolution. Perhaps polycation-π mode of EAD in EFPs reflects a primordial mode of protein-protein interaction that was subsequently tailored by evolution for alternative roles in native EWS.

Date of Award	2013
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology