An investigation of the single-molecule biophysics of membrane systems with atomic force microscopy

Abstract

Cell membranes define the boundary of the fundamental unit of all organisms. They maintain the biochemical gradients of the cell and are carefully regulated by various membrane proteins and factors. Understanding of these components and their mechanisms is a broad and multiphasic field of biophysical investigation, both for fundamental research and for potential therapeutic benefit. While bulk assays provide insight into overall behavior, single-molecule methods probe the underlying complexities of these systems. The atomic force microscope (AFM) is a powerful tool which can directly visualize active biomolecules on a surface in physiological buffer and temperature. It is capable of high spatial (approx 1 Å vertical; approx 10 Å lateral) and temporal resolution (approx 100 ms), and is thus poised for critical single-molecule research. This dissertation utilizes the AFM for two membrane-related studies, namely, understanding the mechanisms underlying i) P-glycoprotein (Pgp) mediated efflux which causes multi-drug resistance in cancer cells and ii) membrane remodeling induced by fungal peptide toxin candidalysin (CL). Pgp is an ABC transporter overexpressed in cancer cells that non-selectively binds and effluxes drugs from the cytosol, leading to multi-drug resistance. Crucial to inhibiting this effect and developing cancer therapies is an in-depth examination of the structure-function relationship of this protein. Pgp reconstituted into lipid bilayers is imaged using AFM in the absence and presence of multiple ligands. To identify specific conformations along the hydrolysis cycle, ATP analogs are used; for substrate-bound conditions, Pgp is imaged in the presence of chemotherapeutic drugs. We show that Pgp has a wide range of conformational flexibility, and characterize shifts in states populated by the protein. Kymographs reveal the dynamics of individual cytosolic side features with high temporal resolution. Precision particle detection and statistical analysis techniques differentiate conformational states in both images and kymographs. Novel ATP activity measurements tracking the release of inorganic phosphate product confirm that the protein remains active on a surface. Taken together, the data provide insight into a pharmaceutically relevant membrane protein in near-native conditions. Human fungal pathogen Candida albicans secretes CL as a virulence factor which forms pores in epithelial cells, triggering uncontrollable Ca2+ influx and a cascade of immune responses. AFM studies reveal a pore-assembly process by which the peptide oligomerizes into a fundamental subunit, which then polymerizes into linear arrangements. These linear features close, forming loops which are proposed to insert into the membrane and are a critical structure component of the pore. Pore development progresses in stages, from initial depressions in the membrane to pores ringed in positive punctate features, called rims. Sequential imaging gives evidence that this transition may stabilize the pore. Site-specific mutagenesis identifies a loss-of-function mutant which fails to permeabilize membranes. AFM imaging reveals that the mutant does not polymerize or form pores. A gain-of-function mutant with enhanced loop formation showed increased membrane damage over the wild-type, supporting our theory that the loops are involved in perforation. These investigations integrated several biophysical techniques in addition to single-molecule AFM and represent a collaborative effort which unveiled a novel pore-forming mechanism and a possible therapeutic target.