The physical properties of intrinsically disordered proteins (IDPs) affect their positive (functional) and negative (disease-causing) roles in cell function. Yet despite intense effort, we still lack a predictive understanding of the physical properties that govern whether a given polypeptide chain sequence will adopt an expanded or collapsed conformational ensemble under physiological conditions ? and by extension, how collapse affects protein function. The Sosnick and Clark labs have formed a collaboration to develop precisely this understanding. This project consists of experimental studies of IDPs tightly integrated with new data analysis procedures and computational modeling tools. This project builds on our recent findings with ?PNt?, a 334 residue IDP that under physiological conditions adopts an expanded ensemble of conformations well approximated by a self-avoiding random walk, despite having an amino acid content that, according to the current paradigm, should lead to a collapsed, self-associated state. We hypothesize that current models fail to predict the behavior of PNt due to current knowledge gaps regarding which sequence patterns, beyond global sequence composition, lead to collapse. We propose that deviations from an expanded state, e.g., adopting a collapsed globule, are due to specific sequence patterns including local stretches of hydrophobic residues. We will test this hypothesis by reordering (?shuffling?) the amino acid sequence of PNt to induce collapse, and likewise shuffle the amino acid sequence of maltose binding protein (MBP) to promote expansion of its denatured state, which we recently demonstrated is highly collapsed. We will use a battery of biophysical methods (SAXS, hydrogen-deuterium exchange, NMR) to measure the extent of local versus global collapse, assessing the sensitivity of conformational ensembles to subtle changes in sequence order, and the relationship between collapse and hydrogen bonding. We will determine which hydrophobicity scales yield the best predictions of experimental results for polypeptide chain collapse, and use the effects of shuffling to parameterize simulations. Finally, we will test the impact of altering IDP collapse on protein function in vivo, specifically the efficient secretion of autotransporter virulence proteins to the surface of Gram-negative bacteria. Our overall goal is to accurately predict the conformational ensemble for a user-inputted amino acid sequence and solvent condition.
Many proteins contain disordered segments, which play an important role in molecular recognition, function and/or folding. Much remains unknown regarding how a protein sequence affects its propensity for disorder, versus collapse and self-association. Our goal is to build upon our recent findings that challenge the existing paradigm regarding the physical forces that govern disorder in order to develop a predictive understanding of which protein sequences will collapse, and under what conditions.
