Importance of hierarchical interactions in concentrated protein systems

Abstract

Protein-protein interactions play important roles in signal transduction, chemical catalysis, cell metabolism, membrane transport, muscle contraction and more. While many of these interactions can be understood as two-body interactions, other important properties can emerge when large numbers of molecules interact together. In this dissertation, we investigate three such systems. We find that a common characteristic in the emergent behavior is interactions occurring on multiple length and energy scales. Biomolecular condensates appear throughout the cell serving a wide variety of functions. Many condensates form by interactions between multivalent proteins, which produce phase-separated networks with liquidlike properties. We use a model system of poly-SUMO and poly-SIM proteins in which the basic unit of assembly is a zipperlike filament due to the interaction of alternating poly-SUMO and poly-SIM molecules. These filaments have defects of unsatisfied bonds that allow crosslinking for both the formation of a 3D network and the recruitment of additional molecules, called clients. We observe a nonmonotonic client binding response that can be tuned independently by the client valence and binding energy. These results show how the interactions within liquid states can be disordered yet still contain structural features that provide functionality to the condensate. The nucleolus is a large membrane-less organelle in the nucleus which is responsible for the assembly of ribosomes. We have developed models to understand how interactions among major proteins assist pre-ribosomal assembly and proper rRNA folding in the nucleolus. The first of these models describes the electrostatic interaction between two major nucleolar proteins (NPM1 and SURF6-N) as a driving force for the phase separation in the nucleolus. The second model describes the role of NPM1 as a chaperone in the rRNA folding process. We have developed analytic, numerical, and computational methods which confirm that NPM1 lowers the zipper barrier to unzip misfolded rRNA. Protein aggregation is a major problem in drug formulation for the pharmaceutical industry. Antibody molecules form elongated complexes due the interaction between their domains. These complexes entangle with each other causing a sharp rise in viscosity of solution. To reduce the production cost, we need to predict the viscous behavior of molecules in the early stages of drug development. We develop a method to use dilute solution measurements to predict the antibody viscosity. We translate the strength of self-association measured by dilute solution experiments (ACSINS, DLS) into the binding affinity of molecules. A theoretical model based on entanglement physics predicts viscosity from these binding affinities. Predictions are in good agreement with the test set of molecules with a few outliers. Using the theoretical model, we predict the physical mechanism of these outliers, and we have proposed solutions to account for these mechanisms in a refined method.