Files
Abstract
Allostery is a fundamental property of proteins, describing the functional coupling between distantly separated sites within the molecule. In different manifestations, this property enables signal transduction, gene expression, and regulation. Although decades of research have advanced our understanding of allosteric proteins, fundamental questions persist. In this thesis, I present my work on two of these questions: What physical principles enable allostery? And how can allosteric function emerge in proteins through evolution? I begin by presenting a reduced theoretical framework that physically represents generic proteins. By making minimal assumptions, I derive the essential physical conditions necessary for allostery. What emerges is a basic classification of allostery into two forms: mechanisms that couple distant sites through soft normal modes within a single state, and mechanisms that rely on multiple states, with effective interactions emerging only upon switching between them. This framework unifies all previously proposed models of allostery. Building on this work, I investigate what determines the form of allostery that emerges through evolution. I analyze an elastic network model of a protein across a range of physical and evolutionary constraints and find that a continuum of mechanisms exists, operating on a combination of the soft mode and multi-state principles. I find that when the system has the capacity for frustration, the multi-state mechanism produces stronger allosteric effects and is selectively favored. Finally, I consider how allosteric function can emerge in proteins even without direct selection. This question is motivated by the finding that many proteins have latent allosteric function—properties that exist in the protein but are not part of its native role. Previous computational and experimental data suggest that allosteric architectures enhance a protein's adaptability by wiring distant sites to functional regions. I hypothesize that fluctuating environments select for evolvable designs and, as a byproduct, allosteric architectures. Using a minimal binding model, I show that periodically switching selective pressures promote the emergence of evolvable, allosteric sequences. These results show how a protein's design reflects not only its present function but also the statistical structure of its evolutionary past.