Computational study of membrane driven secondary structural changes in proteins

includes bibliographical references and index

PhD; 2023; Chemical Engineering

Conformational changes in proteins, the most abundant biomolecule found in all living organisms, are ubiquitous and triggered by several factors essential for protein function. Protein conformational changes typically occur on time scales of tens of microseconds to milliseconds, lying well outside the sampling regime of conventional molecular dynamics (MD) simulations. Although MD simulations have been extensively used to study protein folding to obtain free energy landscapes, membrane assisted protein folding, the primary focus of this thesis, has received less attention. In this thesis, we present a finite temperature string method path based approach to obtain the free energy of protein conformational changes utilizing path collective variables. We rigorously test and validate our approach and demonstrate its ability to capture the α-helix to β-sheet transformation in the mini G-protein in a reduced two-dimensional collective variable space. We apply the method to study phospholipid membrane driven protein conformational changes associated with the assembly of bacterial pore forming toxins (PFTs) and antimicrobial peptides (AMPs). The mammalian cell membrane contains cholesterol, and several proteins of the PFT family require cholesterol recognition for lytic activity. Although cholesterol has been shown to enhance lytic activity, the molecular underpinning of the role of cholesterol for cytolysin A (ClyA) activity, an α-PFT expressed by E. coli, remains elusive. Using the string method, we unravel the critical role played by cholesterol by obtaining the free energy of the β-tongue transformation to the helix-turn-helix motif of the pore state. Cholesterol was found to assist pore formation by stabilizing an unfolded on-pathway intermediate of the membrane inserted β-tongue motif. Specifically, a tyrosine residue located at the phospholipid protein interface was found to be critical in catalyzing unfolding. Using extensive thermal unfolding MD studies on point mutations of the protein, we concluded that inherent flexibility in key membrane binding domains is essential for pore formation. Point mutations that reduced flexibility were detrimental to pore formation, concurring with experimental observations where a point mutation of tyrosine implicated in cholesterol binding completely abrogated lytic activity. We next applied the string method approach to study the insertion free energy and mechanism of insertion of the AMP ‘CM15’ in the inner bacterial membrane. Our free energy analysis showed that a single membrane-bound peptide unfolded state is more stable than a membrane-inserted folded state, with the insertion mechanism triggered by the N-terminus interactions with the cardiolipin lipid molecules of the bacterial membrane. Cardiolipin has not been considered in the previous studies, and our study points to the vital role of this four tail lipid in AMP-membrane interactions. We also report strong interactions of water molecules with one side of the membrane-inserted amphiphilic peptide, which can potentially be responsible for bacterial cell lysis. In summary, the string method based approach developed in this thesis can be applied to a wide variety of protein conformational changes and can be used to study complex membrane driven protein unfolding, refolding, and conformational changes.