The interaction of amyloid beta 40 peptide with anionic phospholipid Langmuir monolayers

The amyloid beta (Abeta) family of peptides is theorized to be the primary cause of neurotoxicity in Alzheimer's disease (AD). Despite strong circumstantial evidence pointing to the central role of Abeta in AD, to date there are several competing theories as to the mechanism of neurotoxicity. Previous work has shown that lipid-protein interactions may mediate the neurotoxicity of the peptide. We have studied the insertion of the peptide into anionic monolayers in both the liquid and the condensed phase at physiological concentrations of the 40 residue Abeta peptide (Abeta40). We saw a larger insertion into dimyristoylphosphatidylserine (DMPS) films than we did for dipalmitoylphosphatidylglycerol (DPPG) films under the same conditions. Using X-ray reflectivity, fluorescence microscopy, Brewster angle microscopy, and grazing incidence X-ray diffraction, we were able to determine detailed parameters for the DMPS system that illustrates how the peptide interacts with the film.;We have also studied the insertion process into mixed monolayers of DMPS with dihydrocholesterol (dChol) and DPPG with dChol. With an increasing mole percentage of dChol, the insertion drops precipitously. After 35 mole percent dChol for DMPS:dChol mixed monolayers, we observe essentially the same insertion as there is into pure dChol. For DPPG:dChol mixed monolayers, this drastic drop in insertion occurs at approximately 40 mole percent dChol. We explain this effect within the context of the condensed complexes theory of lipid cholesterol interactions.;We also discuss additional studies that were needed to analyze the X-ray reflectivity (XR) data for the pure lipid systems and the mixed DMPS:dChol systems. Fitting XR data is a typically laborious task that is easily skewed by the preconceptions of the experimenter. We implemented a stochastic model-independent method for fitting XR reflectivities that minimizes user bias and requires very little information about the system. We have also developed a unique method that allows us to take the model-independent data and move it to a completely model-dependent methodology for further analysis.