Condensates Biophysics

St. Jude Children'S Research Hospital, Memphis TN

Investigators

Abstract

PROJECT SUMMARY - PROJECT 1 Fusion oncoproteins (FOs) are observed in ~17% of all cancers and are commonly oncogenic drivers in pediatric cancers, including those with poor prognosis. Kriwacki, Mullighan, Klco and Pounds (Projects 1 & 2, Core 1), and others, discovered that some FOs drive oncogenesis by forming aberrant biomolecular condensates. Biomolecular condensates—membraneless, meso-scale assemblies—mediate diverse biological processes in normal cells, including regulation of gene expression. FOs acquire oncogenic function, often aberrant gene regulation, through fusion of portions of two unrelated proteins, which we hypothesized commonly involves gaining the ability to form condensates. Kriwacki and Mullighan (Projects 1 & 2) tested this hypothesis by expressing 166 FOs in mammalian cells, with 58% (96) shown to form condensates. We leveraged this rich dataset to develop in silico tools based on analysis of sequence-derived physicochemicalfeatures (e.g., disorder content, interaction probability, hydropathy, charge characteristics, etc.) to predict condensate formation by other, untested FO. We did this by developing an accurate Machine Learning (ML) model, termed FO-PunctaML model. Applying this model revealed that 67% (1,999) of 2,999 additional FOs harbor physicochemical features positively correlated with condensate formation, supporting our hypothesis. However, although we know that 96 FOs form condensates in cells, and 1,999 others are predicted to do so, we do not yet broadly understand how distinct patterns of physicochemical features within FO sequences cause condensate formation, representing a knowledge gap. Further, we and others have shown that FOs interact with functional partners within condensates to promote their oncogenic functions; however, the mechanisms of these interactions remain elusive, representing a second knowledge gap. We hypothesize that different patterns of physicochemical features in different types of FOs (termed Groups) encode their ability to undergo multivalent interactions that promote not only condensate formation (by FOs) but also their interactions with functional partners. To test these hypotheses, we propose two Specific Aims: Aim 1 seeks to elucidate how folded domains and intrinsically disordered regions (IDRs) contribute to condensate formation by FOs, and Aim 2 seeks to discover the mechanisms underlying interactions between FOs and their protein partners. To accomplish these aims, we are studying FOs from the set of 96 that tested positive for condensate formation but are otherwise poorly characterized. Importantly, in addition, we have obtained novel FO sequences identified in pediatric cancers by other investigators under this FusOnc Program Project for our studies, including those associated with acute leukemias (Mullighan & Klco, Project 2), ependymoma (Mack, Project 3), and mesenchymal chondrosarcoma (Wang, Project 1). Our studies will elucidate the fundamental biophysical mechanisms underlying condensate formation and partner interactions by FOs. This knowledge will be applied broadly across the FusOnc Program Project to understand how condensate formation by FOs underlies oncogenesis in diverse pediatric cancers.