The ability to visualize cellular events by linking them to color or fluorescence changes has been an invaluable tool for biology. We describe a novel plasmid-borne color marker whose expression in yeast leads to purple-colored cells that are also brightly fluorescent. This dominant marker provides a useful tool for rapidly screening plasmid maintenance using a visual or fluorescence assay in both Saccharomyces cerevisiae and Candida albicans.
MONOMERIC (m) variants of the tetrameric Discosoma sp. (Ds) red fluorescent protein (RFP) are widely used as fluorescent protein tags in mammalian cells, but are not well expressed in Saccharomyces cerevisiae and other fungi with AT-rich genomes. Directed mutagenesis experiments that led to development of the bright monomeric mRFP variants used a DsRFP template (DsRed-N1) that had been codon optimized for expression in human cells (Bevis and Glick 2002; Campbell et al. 2002). We reasoned that poor mRFP expression in yeast might be due to incompatible codon usage since synonymous codons of yeast are biased toward an A or T at the third position. In an attempt to enhance mRFP expression in S. cerevisiae, as well as other AT-rich yeast and fungi, mutations were introduced in the mCherry mRFP variant (Shaner et al. 2004) to maximize the codon bias toward the more frequent AT-rich codons. Ninety-eight percent of mCherry codons contain a C or G at the third position so, in total, 218 of 237 codons were altered to produce a yeast-enhanced mRFP (yEmRFP) (supplemental Figure S1 and supplemental Table S1).
Unexpectedly, we discovered that yeast that overexpress a plasmid-borne copy of yEmRFP, not fused to any other gene, display an unusual vivid purple color phenotype (Figure 1A). These results were surprising because an RFP-dependent color phenotype in the visual range has not been reported. Initial experiments were performed in an ade2 mutant strain, so we considered the possibility that this yEmRFP-induced purple color phenotype was somehow related to the reddish color caused by accumulation of purine precursors in ade2 mutants (Roman 1956). However, this possibility was ruled out because a wild-type ADE2 (white colored) strain that expresses plasmid-borne yEmRFP is also purple, although of a different hue (Figure 1A). Notably, ade2, ade2 yEmRFP, and ADE2 yEmRFP strains could be easily distinguished from one another by their distinct colony color. These differences are exemplified by the spectrum of color phenotypes of the random haploid spores produced from an ADE2/ade2 diploid harboring yEmRFP on a URA3 2μ plasmid (yEpGAP-Cherry) (Figure 1B). When plated on nonselective medium, the genotype of each colony could be easily assigned solely on the basis of its color phenotype (Figure 1B). These predicted genotypes were confirmed by plate assay (data not shown). We also noted that another difference between the purple and red phenotypes in these different strains is color development time; purple cells harboring the yEmRFP plasmid could easily be distinguished from those carrying the vector control plasmid after an overnight incubation and were vibrant after 2 days either on plates or in liquid medium. In contrast, ade2-dependent red color development required at least 3–5 days to develop. Taken together, these results demonstrated that the purple color phenotype is due to the expression of yEmRFP. The bright color conferred by yEmRFP suggested that it might be a useful new marker for monitoring genetic events that influence plasmid uptake, maintenance, and stability. The only other reliable color phenotype used in yeast is the red/brown color that is caused by an ade2-dependent accumulation of a red pigment because of a block in adenine production. Since it was described by Roman (1956) >50 years ago, the utilization of red color to monitor genotypes that result from combinations of ade2 and other genes has been widely employed as a nonselective phenotype to visualize myriad genetic events and has provided an invaluable tool for the study of gene dosage and plasmid stability. The experiments described below were designed to test the feasibility of using yEmRFP as a phenotypic color marker and to further characterize it.
To determine if the purple color phenotype correlates with the fluorescence properties of mRFP, yeast cells were assayed by fluorescence. Individual cells that expressed yEmRFP on a 2μ plasmid were visualized by fluorescence microscopy. These cells were brightly fluorescent, with a diffuse pattern that appeared largely restricted to the cytoplasm and excluded from the vacuole (Figure 2A). Fluorescence was specific to the emission and excitation spectra of RFP; no bleed-through was observed using GFP or other filter sets.
The correlations among color, fluorescence, and plasmid maintenance were further examined by characterization of the sectoring phenotype of cells harboring the yEmRFP plasmid. If purple color is tightly linked to expression of the yEmRFP gene, then loss of the plasmid carrying this gene will result in white colonies or white sectors that arise from clones within a purple colony that have lost this plasmid. Yeast cells transformed with URA3-marked yEmRFP plasmid were isolated and then plated on nonselective medium. After 2 days of growth on YPD plates, white and purple/white sectored colonies were detected (Figure 2B), suggesting that retention of the plasmid carrying yEmRFP is required for purple colony color. Examination of these colonies with an RFP-specific filter also demonstrated that white sectors, as well as white colonies, were devoid of fluorescence, while the purple portions of the colonies were brightly fluorescent. These results provide further evidence that there is a tight correlation among color, fluorescence, and yEmRFP expression.
Fluorescence, colony morphology, and sectoring were also examined in yeast expressing yEmRFP from a low-copy CEN vector (pRS316 GAP-Cherry; see supplemental Table S2). When plated on selective medium, cells expressing yEmRFP at low copy did not display the vivid purple color seen in cells that express yEmRFP from a 2μ plasmid. However, individual cells were still brightly fluorescent when viewed by microscopy (data not shown). When plated on nonselective medium, sectored colonies were also easily detected by fluorescence (Figure 2C). Taken together, these results demonstrated that the vivid purple colony color requires yEmRFP overexpression, but that the more sensitive fluorescence-based assays can be used to monitor plasmid stability of both 2μ and low-copy CEN-based plasmids.
To test the sensitivity of this sectoring assay, we asked if an event that is predicted to bias plasmid retention could be visually screened. To test this idea, yeast were transformed with a plasmid carrying two genes: the yEmRFP gene and, in addition, the PTDH3-driven ALG7 gene whose overexpression leads to a mild, dominant-negative growth phenotype in wild-type cells (C. Noffz, unpublished observations). If the purple color phenotype of cells is strictly dependent on plasmid maintenance, then on nonselective medium there should be a higher frequency of white colonies and/or colonies with more white sectors that were derived from these cells that would benefit from plasmid loss. The results of this experiment demonstrated that this prediction was borne out. While yeast transformed with plasmids carrying yEmRFP or yEmRFP–PTDH3-ALG7 cells were bright purple on selective medium (not shown), on nonselective medium there was a large increase (>500-fold) in the frequency of white colonies from cells transformed with the yEmRFP–PTDH3-ALG7 plasmid compared to those transformed with yEmRFP plasmid (Figure 2D). In the former case, most cells were white, although some had a slight purplish hue, suggesting the residual presence of some yEmRFP–PTDH3-ALG7 plasmids. Indeed, individual sectors of cells retaining this plasmid could be clearly observed by a more sensitive fluorescence assay (Figure 2D). Taken together, these data demonstrate that cells that maintain this plasmid can be reliably distinguished from those that lose it with high sensitivity by either a visual or a fluorescence assay. These results also provide support for the suitability of this system as a screening method for genetic interactions that affect the loss or gain of function of additional genes that are linked to this yEmRFP marker.
To further characterize the potential utility of this reporter as a color marker, we wanted to determine if overexpression of yEmRFP conferred any selective disadvantage on cells. Three aspects of the yeast life cycle were analyzed: cell division, mating, and meiosis (Table 1). Cell division was assayed by comparing the growth rate during the logarithmic phase of cells harboring the 2μ yEmRFP or parental vector plasmid. In addition, when cells containing the 2μ yEmRFP or parental vector plasmid were co-inoculated in selective medium lacking uracil at a 1:1 ratio, the ratio of purple cells to white cells remained constant over a 36-hr period (data not shown). Mating efficiency was determined by comparing the frequency of diploid formation. Sporulation efficiency was measured quantitatively by counting the number of tetrads in a culture of sporulated cells and qualitatively by visual inspection (Table 1). The results of these experiments, shown in Table 1, demonstrated no quantitative or qualitative effects of the 2μ yEmRFP plasmid on these processes. Although we cannot rule out effects on other biological processes, these results suggest that overexpression of yEmRFP does not adversely affect the normal growth and reproductive properties of yeast.
Expression in Candida albicans of RFP has not been possible thus far because it has a high frequency of CTG leucine codons that are decoded as serine instead of leucine by the noncanonical C. albicans genetic code. The yEmRFP synthetic gene lacks this codon completely (supplemental Table S1). To examine expression in C. albicans, a CaADH1 promoter-driven yEmRFP was cloned in a CaURA3-marked plasmid containing an autonomously replicating sequence (ARS). C. albicans ARS-containing plasmids are reported to be maintained as high-copy multimeric plasmid molecules that are extrachromasomal (Kurtz et al. 1987). In contrast to the human codon-optimized mRFP, which is not expressed in C. albicans (data not shown), we found that yEmRFP could be easily detected in both budding and hyphal cells (Figure 3A and data not shown). A large percentage of uracil protrophs of a ura3Δ/ura3Δ C. albicans strain transformed with this ARS URA3 ADH1 promoter-driven yEmRFP plasmid was distinguished by a distinct pale pinkish color and bright fluorescence (Figure 3). These colonies did not display the vivid purple color seen in S. cerevisiae, presumably because of a lower yEmRFP copy number. When viewed by microscopy, budding C. albicans cells displayed bright cytoplasmic fluorescence that did not diminish even during 2 hr of hyphal induction; hyphal cells were brightly fluorescent throughout the cell body and filament (Figure 3A).
The mechanism of plasmid stability in C. albicans is not understood. Although the original literature that reported a characterization of ARS-containing plasmids in C. albicans concluded that these plasmids remain episomal as large multimeric plasmids, whether or not these multimers are integrated in the chromosome has not been adequately resolved. To address this question, we performed a sectoring assay of C. albicans transformed with an ARS plasmid containing the yEmRFP gene. If this plasmid remains extrachromosomal, then sectored colonies should arise when transformed cells are grown in the absence of plasmid selection. The results of this experiment, shown in Figure 3B, demonstrate that no sectored colonies of C. albicans transformed with yEmRFP on an URA3, ARS-containing plasmid could be detected, even after multiple platings on nonselective medium. This complete absence of sectoring provides strong genetic evidence that this DNA is stably integrated in chromosomes. These results also provide a proof of principle for this system as a potential genetic tool for investigating factors that contribute to extrachromosomal plasmid maintenance in C. albicans.
Because of the rapidity of color development and the ease of detection, we considered the possibility that yEmRFP expression could be used as a visual screen for identifying yeast transformants, much like the blue-white color associated with lacZ-encoded β-galactosidase production in bacteria (although without the need for chromogenic additives). To determine the feasibility of this idea, yEmRFP-transformed cells were diluted 1:2000 with untransformed cells and plated on nonselective medium. On the basis of this dilution and the number of cells plated, we calculated one or fewer purple colony per plate. As shown in Figure 4A, a single purple colony per plate could be easily identified both visually and by fluorescence under these conditions. Further dilutions, taken out to 1:10,000, demonstrated that the sensitivity of this plate assay is limited by colony size, and hence the number of colonies that can be plated. In the case of S. cerevisiae, confluence of individual colonies occurs beyond ∼5000 colonies/100-mm plate. These results demonstrate that individual yEmRFP colonies can be detected with high specificity (Figure 4B). Low-efficiency transformation methods typically give rise to ∼3 × 103 transformants/μg of plasmid DNA/108 cells, so high-efficiency methods, in which up to 106 transformants/μg of plasmid DNA/108 cells are obtained (Gietz and Schiestl 2007), must be employed for this color marker to be a useful indicator of transformation.
In conclusion, we describe a new purple colony color marker that has high potential for genetic studies in yeast. The ade2-dependent colony-sectoring assay (Hieter et al. 1985; Koshland et al. 1985), which allows the visual monitoring of plasmid loss, has been a powerful and widely applied strategy for identifying numerous types of genetic interactions, including synthetic lethality and dosage suppression (Bender and Pringle 1991). Our studies suggest that this new purple marker can be similarly applied and has several important advantages over ade2. First, cells expressing this protein can be identified both visually by their purple color and by very sensitive fluorescence-based assays. Second, purple color development occurs more rapidly than ade2-induced red. Purple cells can be identified after an overnight incubation, rather than after 3–5 days, as is required for ade2-induced red. Third, unlike ade2-induced red color, which requires engineered strains that contain a recessive ade2 allele, this purple color is plasmid dependent and dominant and therefore can be introduced into any yeast strain background. This system holds promise as a powerful new genetic tool for studies in S. cerevisiae and other fungi in which it can be expressed.
We thank Nicole Averbeck for helpful comments and reading of the manuscript and Catherine Menzies for her technical assistance in the early part of the project. This work was supported by a grant from the National Institutes of Health (RO1-GM048467) to N.D.
- Received January 14, 2008.
- Accepted February 19, 2008.
- Copyright © 2008 by the Genetics Society of America