A Minimal Approach to the Dynamic Regulation of Biomolecular Conformation
von Maximilian Vossel
Datum der mündl. Prüfung:2022-07-15
Erschienen:2022-12-14
Betreuer:Dr. Aljaz Godec
Gutachter:Dr. Aljaz Godec
Gutachter:Prof. Dr. Peter Sollich
Gutachter:Prof. Dr. Matthias Krüger
Gutachter:Prof. Dr. Stefan Klumpp
Gutachter:Prof. Dr. Jörg Enderlein
Gutachter:Dr. Andreas Neef
Dateien
Name:diss.pdf
Size:26.7Mb
Format:PDF
Description:Dissertation
Zusammenfassung
Englisch
Allostery is the key mechanism enabling a “remote” regulation of the activity of proteins and other biological macromolecules. It forms the basis for controlling various inter- and intracellular processes. Allosteric proteins accommodate changes at their active site as a result of ligand binding at a second, often spatially distant binding site—the allosteric site. Despite the ubiquity and fundamental role of allostery in proteins, the general physical principle underlying the efficient and precise long-range transmission of mechanical sig- nals from the allosteric to the active site remains elusive. In addition to contributing to the unraveling of fundamental processes essential for sustaining life, the elucidation of a gen- eral mechanism holds a significant practical potential: It allows the prediction of allosteric sites in proteins based on single structures and hence addresses the current bottleneck for allosteric drug design. Here, building on the previous success of spectral methods in explaining the dynamics of proteins, we hypothesize a simple physics-based principle that unifies allosteric behavior in trained artificial and protein-derived elastic networks. The mechanism rests on the concept of a collective lever that couples the interaction of stiff and soft modes in a sophisticated way. Input displacements at allosteric source pockets efficiently load collective stiff springs while the response occurs via soft modes, which convert the stored energy into large, non- local but specific displacements at the target pocket. To test this hypothesis, we develop a fast and accurate algorithm for determining the full (nonlinear) response of elastic networks to perturbations caused by displacements at the source pockets and employ it to evolutionarily train networks to display allosteric re- sponses. We observe nonlinear and non-reciprocal behavior during the responses and dis- cuss the origin and the implications this may have for allostery. Using spectral and perturbative approaches, we find convincing evidence confirming the predicted properties that characterize the uniqueness of the real source compared to other possible binding pockets in protein-derived and artificial, trained networks. Finally, we demonstrate the applicability of the concept in drug design by using it to predict allosteric source pockets in proteins—with remarkable success, considering that only the initial structure and a single-parameter model is employed. These findings shift the paradigm of allosteric signal propagation in networks and proteins from a purely soft-mode based interpretation towards a two-step thinking, which differentiates strongly between the allosteric in- and output. We conclude that allostery is only fully comprehensible if viewed as what it actually is—a non-equilibrium effect that requires the perturbation to be accounted for explicitly in the spectral interpretation.
Keywords: allostery; proteins; conformational motion; elastic network; collective modes; mode-coupling; nonlinearity; non-reciprocity; evolutionary optimization; network design; drug design