Dynamical-nonequilibrium simulations: an emerging approach to study time-dependent structural changes in proteins

Proteins are neither static nor work in isolation in physiological conditions. In fact, it is the opposite; proteins are continuously moving and switching between different conformations. Moreover, changes in the environment can shift the balance between their multitude of conformations.

For example, changes in pH and the binding of ions or small molecules to a protein can promote specific structural changes and ultimately determine the protein's macroscopic behaviour. This ability to respond to external changes by fluctuating between conformations is a fascinating feature and is crucial for protein's function and regulation. A detailed description of a protein's conformational rearrangements is essential to understand its working mechanism and function.

Even though it is possible to experimentally determine the positions of the atoms in a protein (e.g. using cryogenic electron microscopy or X-ray crystallography), in some cases, the effects of making changes to the protein (e.g. mutations or the binding of ligands and ions) are not obvious. An alternative to experimental approaches is the application of computer simulations.

I have been at the forefront of developing and employing computational methods to map structural changes in proteins. For this project, in particular, the approach of combining computer simulations in different conditions (e.g. in the presence and absence of a ligand) is the keystone. This approach permits for a detailed mapping of the time evolution of the structural changes in a protein in response to an external perturbation (e.g. ligand unbinding).

The proposed research, undertaken at the University of Bristol, will develop and apply new computational approaches to transform the study of conformational changes in proteins. Three fundamentally different biomolecular systems will be studied during the timeframe of this Fellowship, ranging from soluble enzymes (beta-lactamases) to membrane channels (cystic fibrosis transmembrane conductance regulator (CFTR) channel) and receptors (nicotinic acetylcholine receptors).

The diversity of the systems under investigation perfectly highlights the flexibility and general applicability of the computational approaches to be used. Beta-lactamases are bacterial enzymes capable of hydrolysing antibiotics (e.g. penicillin) and are an important cause of resistance to these drugs. The CFTR channel, which sits on the surface of cells, transports chloride and bicarbonate, and its malfunction causes cystic fibrosis.

Nicotinic acetylcholine receptors are ion channels widely distributed in the nervous system and are associated with many diseases and conditions, including nicotine and alcohol addiction. In this work, I will map the communication networks connecting functionally important regions within each protein and understand how small molecules and mutations impact those networks.

My computational findings will unlock a diversity of interactions that will be explored experimentally by my collaborators, namely Profs Spencer (University of Bristol), Sheppard (University of Bristol), Bermudez (Oxford Brookes University), Sine (Mayo Clinic) and Gallagher (University of Bristol). Their experimental results will feed into my computational models, helping to refine and enhance them.

This is a highly collaborative, multidisciplinary project that combines computational (in partnership with Oracle) and experimental expertise, which provides a unique opportunity to expand fundamental knowledge of all three systems' working mechanisms. In the longer term, this knowledge will foretell the properties of newly emerged beta-lactamases mutants, which cause antimicrobial resistance, and inform the design of new therapeutics (e.g. drugs to treat cystic fibrosis, non-opioid drugs for chronic pain, anti-addiction agents and beta-lactamases inhibitors to fight antibiotic resistance).