Abstract
We have developed the system to measure many fluorescent components in their mixtures, based on the microspectroscopic recording of excitation-emission matrix (EEM) fluorescence data and on parallel factor analysis (PARAFAC) as the blind spectral decomposition method. This EEM/PARAFAC system can easily separate >10 heavily overlapping spectral components, without any preknowledge about the individual component spectra. It has proven to be very effective in the simultaneous measurement of multiple fluorescence components (either intrinsic or extrinsic) in single cells (Shirakawa and Miyazaki, Biophys J, 2004). In the present study, we tried to apply the same algorithm to multispectral imaging data for the analysis of the fluorescent components in living cells. The EEM fluorescence images were acquired with the conventional fluorescence microscope, changing and emission wavelengths by switching bandpass filters, or with the laser-scanning confocal microscope that equipped a multichannel spectrometric sensor as the fluorescence detector. PARAFAC modeling applied to time-series data of EEM fluorescence images for single mouse eggs could blindly separate signals of, e.g., functional proteins labeled with GFP and conventional calcium indicators, as well as autofluorescence components. The results showed that this approach will be a powerful methodology for the simultaneous measurement of many intracellular molecules in physiological experiments. [Jpn J Physiol 55 Suppl:S70 (2005)]
