Green fluorescent protein (GFP) has many advantages as a reporter molecule, but its stability makes it unsuitable for monitoring dynamic changes in gene expression, among other applications. Destabilized GFPs have been developed for bacterial and mammalian systems to counter this problem. Here, we extend such advances to the yeast model. We fused the PEST-rich 178 carboxyl-terminal residues of the G(1) cyclin Cln2 to the C terminus of yEGFP3 (a yeast- and FACS-optimized GFP variant), creating yEGFP3-Cln2(PEST). We tested the hybrid protein after integrating modules harbouring the yEGFP3 or yEGFP3-CLN2(PEST) ORFs into the Saccharomyces cerevisiae genome. yEGFP3- Cln2(PEST) had a markedly shorter half-life (t(1/2)) than yEGFP3; inhibition of protein synthesis with cycloheximide lead to a rapid decline in GFP content and fluorescence (t(1/2) approximately 30 min) in cells expressing yEGFP3-Cln2(PEST), whereas these parameters were quite stable in yEGFP3-expressing cells (t(1/2) approximately 7 h). We placed yEGFP3-CLN2(PEST) under the control of the CUP1 promoter, which is induced only transiently by copper. This transience was readily discernible with yEGFP3-Cln2(PEST), whereas yEGFP3 reported only on CUP1 switch-on, albeit more slowly than yEGFP3-Cln2(PEST). Cell cycle-regulated transcriptional activation/inactivation of the CLN2 promoter was also discernible with yEGFP3- Cln2(PEST), using cultures that were previously synchronized with nocodazole. In comparison to CLN2, expression from the ACT1 promoter was stable after release from nocodazole. We also applied a novel flow-cytometric technique for cell cycle analysis with asynchronous cultures. The marked periodicities of CLN2 and CLB2 (mitotic cyclin) transcription were readily evident from cellular yEGFP3-Cln2(PEST) levels with this non-perturbing approach. The results represent the first reported successful destabilization of a yeast-GFP. This new construct expands the range of GFP applications open to yeast workers.
Copyright 2000 John Wiley & Sons, Ltd.