Fluorescence fluctuation analysis of mixed chromophores from a line-scanning hyperspectral imaging system

Fluorescence fluctuation analysis of dilute biomolecules can provide a powerful method for fast and accurate determination of diffusion dynamics, local concentrations, and aggregation states in complex environments. However, spectral overlap among multiple exogenous and endogenous fluorescent species, photobleaching, and background inhomogeneities can compromise quantitative accuracy and constrain useful biological implementation of this analytical strategy in real systems. In order to better understand these limitations and expand the utility of fluctuation correlation methods, spatiotemporal fluorescence correlation analysis was performed on spectrally resolved line scanned images of modeled and real data from mixed fluorescent nanospheres in a synthetic gel matrix. It was found that collecting images at a pixel sampling regime optimal for spectral imaging provides a method for calibration and subsequent temporal correlation analysis which is insensitive to spectral mixing, spatial inhomogeneity, and photobleaching. In these analyses, preprocessing with multivariate curve resolution (MCR) provided the local concentrations of each spectral component in the images, thus facilitating correlation analysis of each component individually. This approach allowed quantitative removal of background signals and showed dramatically improved quantitative results compared to a hypothetical system employing idealized filters and multi-parameter fitting routines. © 2010 Copyright SPIE - The International Society for Optical Engineering.