GFP is widely utilized in C. elegans to visualize specific cells and labeled proteins. For co-localization studies, CFP (cyan) and YFP (yellow) variants can be used simultaneously to generate discrete two-color images. Conventional fluorescence filter sets, however, do not allow the use of GFP in combination with either CFP or YFP due to extensive overlap of these emission spectra. Here we describe new optical tools now available in the Zeiss LSM 510 META confocal microscope for simultaneous multicolor imaging with CFP, GFP and YFP. In the META, a diffraction grating disperses the sample fluorescence over a bank of detectors. This configuration subdivides the visible spectrum into 10.7 nm bandwidths and thereby effectively generates an emission spectrum for each pixel in the image plane. A linear unmixing algorithm utilizes reference spectra for each fluorophore (CFP, GFP, YFP) to calculate a best fit to the composite emission spectrum for each pixel. Pixels that exactly match a specific reference spectrum are assigned a single corresponding pseudocolor whereas pixels that register signal from more than one reporter are assigned the appropriate combination of pseudocolors and relative intensities for each contributing fluorophore. We used this strategy to generate images of transgenic animals expressing CFP, GFP, and YFP in specific motor neuron classes:
unc-25::CFP (DD, VD, blue);
unc-53::GFP (DA, AS, green);
acr-2::YFP (DA, VA, DB, VB, red). DA motor neurons were assigned a fourth color, yellow due to overlapping expression of green GFP and red YFP in these cells. An added benefit of the linear unmixing algorithm is that gut autofluorescence which does not match reporter reference spectra can be digitally removed without degrading the fidelity of the image.