Project Summary BET proteins are directly involved in pathologies such as viral infection and different types of cancer. Although the different BET proteins satisfy different roles in the cell and are preferentially expressed in different tissues, current BET inhibition strategies are non-specific – resulting in toxicity. The ET domain of BET proteins has recently emerged as a protein interaction hub with promising selectivity of binders towards specific BET proteins – making the ET domain an interesting target towards the design of novel drug therapeutics. Yet, little is known about the ET interactome or its binding mechanism. This proposal aims to increase our understanding of the ET interactome (Aim 1) by using computational tools to: a) identify possible binders; b) select the strongest binders using machine learning and physics-based approaches; and c) characterize the most promising leads through NMR experiments. The challenge lies in addressing the polymorphic nature of ET: it can undergo conformational changes to bind different peptide sequences – which in turn bind along different binding modes. Such binding plasticity concomitantly leads to a wide range of binding affinities. Furthermore, a particular peptide sequence binds the ET domain of different BET proteins with different binding affinities, setting the foundation for the design of inhibitors specific to each BET protein. However, it is not clear where the origin of specificity lies, as the ET domains have high sequence and structural similarity across BET members. Thus, Aim 2 will unveil the binding mechanism for peptides to ET and elucidate the origin of specificity through the use of adaptive sampling molecular dynamics strategies and Markov State Models. The relationship between binding affinity and binding mode is currently unknown. It could be that virus are exploiting a binding mode that leads to higher binding affinity than those used by regulatory host proteins. Aim 3 of the proposal addresses this issue by designing novel peptide binders to identify the limits of binding affinity amongst the different binding modes. These computational designs will lead to the identification of hot-spot regions in the binding domain to exploit towards the long-term goal of therapeutic design strategies.