Exploring protein dynamics using graph theory and single-molecule spectroscopy
File(s)
Author(s)
Peach, Robert
Type
Thesis
Abstract
Proteins are the workhorses of any living system. They are essential for day-to-day life and yet
we still do not fully understand how they work. Proteins are intricate structures composed of
thousands of atoms that interact and fluctuate over multiple temporal and spatial scales. These
motions are encoded in the peptide sequence; fast and local vibrations develop collectively into
the motion of secondary structures and catalytic domains. Despite the importance of dynamical
fluctuations in mediating function and other biomolecular processes such as allostery, there is
inadequate evidence supporting the various theories and calculational approaches which attempt
to determine the relationships between structure, dynamics and function.
Graph theory provides a way of capturing and representing the 3-dimensional atomistic physico-
chemical details of a protein in a reduced form. In this thesis, we use Markov Stability, a novel
graph-theoretic method for analysing the hierarchy of protein motions. We begin by producing
an energy-weighted atomistic graph representation of a protein. We then use the transient
behaviour of a random-walk to identify regions of atoms/residues that form communities. Using
mutagenesis we are able to identify mutations that result in a large change to the community
structure. We then use experimental methods (primarily single-molecule FRET) to explore the
effect of these predicted mutations and validate the predictive qualities of Markov Stability.
In this thesis, we demonstrate a strong correlation between theoretical and experimental mea-
sures. We focus our study on Adenylate Kinase; a protein that balances molecular stability and
fast large domain motions in the pursuit of cellular energy homeostasis. Given the structural
data taken from the protein data bank we use the Markov stability graph theoretical approach
to identify a number of biologically relevant community structures at different timescales. We
are able to illustrate agreement between Markov Stability and the rate of motions of ADK
subdomains. We then use the graph theoretical analysis to predict the effect of mutations to
global dynamics and validate these predictions using single-molecule FRET; we show excellent
agreement between the calculated properties of the graph theoretical analysis and the variation
in FRET dynamics. Finally, we explore the relationship between the predictive score attributed
to mutants by Markov Stability and the associated change in molecular stability.
we still do not fully understand how they work. Proteins are intricate structures composed of
thousands of atoms that interact and fluctuate over multiple temporal and spatial scales. These
motions are encoded in the peptide sequence; fast and local vibrations develop collectively into
the motion of secondary structures and catalytic domains. Despite the importance of dynamical
fluctuations in mediating function and other biomolecular processes such as allostery, there is
inadequate evidence supporting the various theories and calculational approaches which attempt
to determine the relationships between structure, dynamics and function.
Graph theory provides a way of capturing and representing the 3-dimensional atomistic physico-
chemical details of a protein in a reduced form. In this thesis, we use Markov Stability, a novel
graph-theoretic method for analysing the hierarchy of protein motions. We begin by producing
an energy-weighted atomistic graph representation of a protein. We then use the transient
behaviour of a random-walk to identify regions of atoms/residues that form communities. Using
mutagenesis we are able to identify mutations that result in a large change to the community
structure. We then use experimental methods (primarily single-molecule FRET) to explore the
effect of these predicted mutations and validate the predictive qualities of Markov Stability.
In this thesis, we demonstrate a strong correlation between theoretical and experimental mea-
sures. We focus our study on Adenylate Kinase; a protein that balances molecular stability and
fast large domain motions in the pursuit of cellular energy homeostasis. Given the structural
data taken from the protein data bank we use the Markov stability graph theoretical approach
to identify a number of biologically relevant community structures at different timescales. We
are able to illustrate agreement between Markov Stability and the rate of motions of ADK
subdomains. We then use the graph theoretical analysis to predict the effect of mutations to
global dynamics and validate these predictions using single-molecule FRET; we show excellent
agreement between the calculated properties of the graph theoretical analysis and the variation
in FRET dynamics. Finally, we explore the relationship between the predictive score attributed
to mutants by Markov Stability and the associated change in molecular stability.
Version
Open Access
Date Issued
2017-11
Date Awarded
2018-04
Copyright Statement
Attribution NoDerivatives 4.0 International Licence (CC BY-ND)
Advisor
Willison, Keith
Klug, David
Yaliraki, Sophia
Barahona, Mauricio
Sponsor
Engineering and Physical Sciences Research Council
Publisher Department
Chemistry
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)