Bayesian Analysis of Imaging FCS Investigates the Interaction of Monomeric HIAPP with Live Cell Membrane

Abstract

Human islet amyloid polypeptide (hIAPP), co-secreted with insulin from pancreatic β cells, is linked to the pathogenesis of type II diabetes mellitus. While the mechanism of hIAPP-membrane interaction at aggregating concentrations of the peptide is well established, the same is not true for monomeric hIAPP. We used imaging total internal reflection-Fluorescence Correlation Spectroscopy (ITIR-FCS) to monitor the effect of monomeric hIAPP on live cell membrane dynamics. The complex interaction of hIAPP with the plasma membrane operates via formation of domains. After hIAPP addition, overall membrane diffusion first increases and then decreases along with concomitant formation and enlargement of the domains which encompass the entire membrane in a time dependent manner. However, the current limitation of ITIR-FCS data analysis does not allow the complete characterization of the temporal evolution of the domains. We implemented Bayesian inference tests which were previously applied to confocal FCS. This allowed us to observe three phases in the evolution of hIAPP-membrane interaction - i) Pre-nucleation phase (1-10 min) where diffusion is single component at most of the pixels and the diffusion coefficient increases monotonically. ii) Propagation phase (15-40 min) where the two-component model fits better for the pixels within the domains. The components exhibit distinct but constant diffusion coefficients but their molecular fraction change with the slow diffusing fraction increasing in time. iii) Saturation phase (40-60 min) where pixels within the domain fit better with single-component model. Therefore, we conclude that hIAPP instantaneous increases membrane diffusion. This is followed by the formation of diffusion-restricted domains that eventually slows down overall membrane dynamics. The novel Bayesian inference analysis of imaging FCS successfully resolved hIAPP-induced domain formation and concomitant modulation of plasma membrane dynamics, which was not explained by the conventional fitting routine.