The most common application of fluorescent proteins in biological research is to visualize the localization of the FP, fused to the protein of interest (thus learning about the localization of your protein). Though looking at the localization of one protein can be informative, it is even more informative if one can detect co-localization with another protein or molecular marker (e,g, a dye that stains a specific organelle).
However, in order to detect co-localization it is required that both FPs will have different emission spectra (e.g. GFP & RFP). The problem of finding non-overlapping colors becomes more difficult when we want to use three colors (or more).
Using the spectrum viewer (see previous post), you can see that in some cases the excitation and/or the emission spectra overlap. Such is the case with CFP and GFP, for instance. Although the excitation maximum for GFP is 488nm and for CFP is 436nm, using a 436 laser will excite some GFP (~35% at peak emission). More importantly, though CFP emission peak is 477nm, the spectra of CFP goes all the way to 600nm (~5%), and definitely overlaps that of GFP. Therefore, the signal (photons) that we detect at the 509nm (GFP peak) can come from CFP emission.
In order to reduce the overlap excitation or emission, we use special filters and dichromatic beamsplitters (“mirrors”) that allow passage of light at specific band of wavelengths. These filters are housed in a filter cube. There are different kinds of filters with different ranges of bandpasses for each color.
However, even with filters, overlap can occur, particularly if the filters in your microscope are wide-range. A possible solution to a three color problem (assuming you expect all three to co-localize) is to use a large Stokes shift (LSS) protein. Stokes shift (as mentioned in an earlier post) is the shift in wavelength from excitation to emission. For most FPs, Stokes shift is less than 50nm (usually much less). For LSS proteins, the shift is over 100nm. For example, the fluorescent protein mKeima has an excitation maximum of 440nm (blue), and emission maximum at 620nm (red). Thus, using the correct emission filter for> 600nm, excitation at 440nm will result in emission of CFP, GFP and mKeima, but only mKeima emission will be detected. However mKeima is not a good choice if you have RFP in your system as well, since mKeima also has a small excitation peak at 590nm that give a similar emission peak at 620nm. This is true for acidic pH, and therefore crucial for the experiment I design in yeast (yeast cytosol is mildly acidic, pH between 5.5 to 7.5). Furthermore, mKeima has low photostability and low brightness.
Therefore, I am going to use a protein called LSSmKate1, with ex/em peaks at 463/624nm or LSSmKate2 (460/605) with no other excitation peak at longer wavelength (except at very basic pH=11).
Two papers on LSSmKates:
Monomeric red fluorescent proteins with a large Stokes shift (Kiryl D. Piatkevich, James Hulit, Oksana M. Subach,Bin Wu, Arian Abdulla, Jeffrey E. Segall, and Vladislav V. Verkhusha. PNAS (2010) 107(12):5369)
Engineering ESPT Pathways Based on Structural Analysis of LSSmKate Red Fluorescent Proteins with Large Stokes Shift (Kiryl D. Piatkevich, Vladimir N. Malashkevich, Steven C. Almo and Vladislav V. Verkhusha. J. Am. Chem. Soc., (2010), 132 (31):10762)