Hyperspectral Imaging Approaches for Measuring Three‐Dimensional FRET

Abstract

Förster Resonance energy transfer (FRET) imaging has become a standard method for assessing protein‐protein interactions as well as changes in the concentration of signaling molecules, such as cAMP. However, many FRET probes or pairs have low signal‐to‐noise ratios and limited dynamic range. Our previous work suggested that hyperspectral imaging and analysis allows quantitative measurement of FRET efficiency in 3D (x, y, and t) within single cell. Here, we present data demonstrating that hyperspectral imaging approaches allow 4D (x, y, z, and t) assessment of FRET efficiency in living cells. In these studies, we have utilized a FRET biosensor consisting of a cAMP binding domain from Epac sandwiched between donor and acceptor fluorophores, Turquoise and Venus. Binding of cAMP induces a conformational change that reduces FRET efficiency. The FRET probe was transfected into HEK293 or pulmonary microvascular endothelial cells. To further demonstrate the potential of 4D hyperspectral approaches to simultaneously assess multiple fluorescence signals in single cells, we co‐transfected the calcium probe RCaMP into cells. 4D hyperspectral images were acquired using a Nikon A1R confocal microscope as follows: Excitation was provided using 405 and 561 nm laser lines. Spectral emission was detected from 424 to 724 nm in increments of 10 nm using a 32‐wavelength spectral detector on a Nikon A1R confocal microscope. A 3D (x, y, z) hyperspectral image was acquired every 15 seconds for 15 minutes. Following one minute measurement of baseline signals, we treated the cells with either 50 μM forskolin (an adenylyl cyclase activator), 0.1 μM or 1 μM isoproterenol (a beta adrenergic agonist), 10 μM rolipram (a PDE4 inhibitor), 1 μM prostaglandin, and/or calcium modulating agents such as 0.1 or 1 μM thapsigargin. Spectral images were processed using custom MATLAB scripts. Briefly, spectral images were unmixed to individual endmembers including Turquoise, Venus, DRAQ5 (nuclear label), WGA‐TRITC (plasma membrane label), and RCaMP. 4D FRET images were calculated from unmixed images. Data were re‐sliced to visualize fluorescence signals in different orientations/projections through each cell. This overall approach allowed assessment of changes in FRET signals in 4D. We observed little or no gradients in the lateral direction with real time 2D (x, y, and t) imaging. However, with 3D imaging (x, y, and z), we found axial gradients (from apical to basolateral) in the cell. With 4D (x, y, z, and t) imaging, we observed that cAMP gradients move from apical to basolateral side over time. Our data suggest that 3D (x, y, and t) imaging approaches may provide only a limited perspective of cAMP gradients and that 4D (x, y, z and t) studies are required to study cAMP gradients and signaling mechanisms in a cell.Support or Funding InformationThis work is supported by P01HL066299, S10RR027535, S10OD020149, R01HL058506, and the Abraham Mitchell Cancer Research Fund.