Molecular recognition by the EWS transcriptional activation domain

Interactions between Intrinsically Disordered Protein Regions (IDRs) and their targets commonly exhibit localised contacts via target-induced disorder to order transitions. Other more complex IDR target interactions have been termed "fuzzy" because the IDR does not form a well-defined induced structure. In some remarkable cases of fuzziness IDR function is apparently sequence independent and conferred by amino acid composition. Such cases have been referred to as "random fuzziness" but the molecular features involved are poorly characterised. The transcriptional activation domain (EAD) of oncogenic Ewing's Sarcoma Fusion Proteins (EFPs) is an ≈280 residue IDR with a biased composition restricted to Ala, Gly, Gln, Pro, Ser, Thr and Tyr. Multiple aromatic side chains (exclusively from Try residues) and the particular EAD composition are crucial for molecular recognition but there appears to be no other major geometrically constrained requirement. Computational analysis of the EAD using PONDR (Molecular Kinetics, Inc. http://www.pondr. com) complements the functional data and shows, accordingly, that propensity for structural order within the EAD is conferred by Tyr residues. To conclude, molecular recognition by the EAD is extraordinarily malleable and involves multiple aromatic contacts facilitated by a flexible peptide backbone and, most likely, a limited number of weaker contributions from amenable side chains. I propose to refer to this mode of fuzzy recognition as "polyaromatic", noting that it shares some fundamental features with the "polyelectrostatic" (phosphorylation-dependent) interaction of the Sic1 Cdk inhibitor and Cdc4._I will also speculate on more detailed models for molecular recognition by the EAD and their relationship to native (non-oncogenic) EAD function.