Comparison of fluorescent markers as second markers for GFP-based reporter systems.
C. A. Axtell and G. A. Beattie.

The protein GFP (green fluorescent protein) and its derivatives are useful reporters of gene expression due to their substrate-independence, stability, low toxicity and ease of detection. For many applications, a GFP-reporter system would benefit from the presence of a second fluorescence marker on the GFP-reporter containing cell. For example, we have constructed bacterial biosensors that produce GFP in response to water stress, and the usefulness of these biosensors would be increased by a second fluorescence marker that simply labels the biosensor cells. We compared various fluorescent proteins and dyes to identify one that meets the following criteria: 1) it is non-toxic, 2) it is detectable for at least 4 generations, and 3) it exhibits a distinct excitation and/or emission spectrum from a red-shifted variant of GFP (hereafter called GFP). We evaluated two derivatives of GFP, sapphire and YFP (yellow fluorescent protein). No toxicity effects were observed in cells constitutively expressing individual GFP derivatives. Expression of the GFP derivatives was stable for multiple generations and cells producing sapphire or YFP could be distinguished spectrally from cells producing GFP. Dual expression of sapphire and gfp in a cell reduced cell growth under some conditions. Dual expression of yfp and gfp in a cell altered the expected fluorescence output of GFP and YFP; specifically, the expected GFP fluorescence was greatly reduced and the expected YFP fluorescence was increased. This observation suggests that a fluorescence resonance energy transfer (FRET) from GFP to YFP can occur between two unlinked GFP derivatives within a cell. We also evaluated two fluorescent dyes, a nucleic acid dye, SYTO Red Stain #64 (Molecular Probes), and a cell membrane dye, PKH26 (Sigma). Both dyes exhibited toxicity effects, species-specificity in the uniformity with which they labeled cells grown in broth, and loss from labeled cells within 24 to 48 hours. Thus, we have not yet identified an ideal second marker for use in GFP-based systems.