Xpressed the Ste2p in relatively low expression manner [13], our result did not show evidence of dimerization for the Ste2p expressed under the control of the CYC1 promoter in the split-ubiquitin system. This is assumed that attached transcription factor might be not present at levels adequate to activate the response. Then, we replaced the full-length Ste2p receptors with a truncated version that lacks the C-terminal tails (Ste2DC; amino acids 1?04) to adjust the distance between the C-termini of the receptors [13], providing an increased signal-to-noise (S/N) ratio (under the control of the TPI1 promoter; Fig. S2A and B). Furthermore, we replaced the TPI1 promoter with the CYC1 promoter of the bait vector again, which resulted in much lower background cell growth and a drastically improved S/N ratio (Fig. 3A and B). Since the truncation of the C-termini of Ste2p receptors had been reported to increase in the number of receptor sites [14], the greatly enriched receptors at the plasma membrane might have provided the drastic improvement of S/N ratio. These results indicate that the bait receptor is predominant for successful detection of the dimerized receptors. Thus, the consideration of the receptor’s expression manner is important to screen the GPCR dimer partners, because the leaky background cell growth often brings 16985061 unanticipated candidates. In the optimized system, Hxt1p never presented background cell growth in the dimerization 18204824 assay with the Ste2DC receptor (Fig. 3A). Previously, Overton et al. had used the Hxt1p as the negative control for FRET analysis of Ste2p dimerization and confirmed the subcellular localization of it at the plasma membrane [15]. The b-galactosidase assay that reflects lacZ reporter enzyme activity also displayed similar trends (Fig. 3B). The addition of ligand had no effect on the dimerization events of Ste2DC, since the MAPK-defective NMY62 yeast strain displayed unchanged growth in the presence of a-factor (Fig. 3A). We additionally constructed two types of deletion mutants in which the TM6? domains and the TM1? domains were removed, respectively (TM1? (amino acids 1?36) and TM6? (amino acids 237?04)) (Fig. 4A). Overton et al. also indicated that selfassociation of TM6? had not been detected, and plasma membrane localization of the YFP-tagged TM1? and TM6? was observed [16]. As previously indicated [16], only TM1? formed the homodimer (Fig. 4B and C), showing that our system is applicable to examine critical domains involved in the dimerization of 7TM receptors. Subsequently, we validated the capability of our system to detect human GPCR heterodimer pairs. To avoid competitive dimerization with the endogenous yeast Ste2p receptor, we constructed an NMY63 mutant strain in which the STE2 gene was additionallyFigure 3. Detection for dimerization of yeast truncated Ste2p lacking the carboxy-terminal tail (Ste2DC) receptor in NMY62 strain. Growth and quantitative b-galactosidase activity of yeast cells expressing various combinations of Cub and NubG fusions. The control bait plasmid was pBT3-C mock vector (empty vector). The control prey plasmids were Title Loaded From File pPR3-C mock vector (empty vector) and pPR3-HXT1. (A) Growth assay without a-factor (left panels) and with 5 mM a-factor (right panels). Each cell was spotted in Title Loaded From File serial 10-fold dilutions on SD eu, Trp, Ade and His dropout plate. (B) Quantitative b-galactosidase assay. Error bars represent the standard deviations (n = 3). doi:10.1371/journal.pone.0066793.gdelete.Xpressed the Ste2p in relatively low expression manner [13], our result did not show evidence of dimerization for the Ste2p expressed under the control of the CYC1 promoter in the split-ubiquitin system. This is assumed that attached transcription factor might be not present at levels adequate to activate the response. Then, we replaced the full-length Ste2p receptors with a truncated version that lacks the C-terminal tails (Ste2DC; amino acids 1?04) to adjust the distance between the C-termini of the receptors [13], providing an increased signal-to-noise (S/N) ratio (under the control of the TPI1 promoter; Fig. S2A and B). Furthermore, we replaced the TPI1 promoter with the CYC1 promoter of the bait vector again, which resulted in much lower background cell growth and a drastically improved S/N ratio (Fig. 3A and B). Since the truncation of the C-termini of Ste2p receptors had been reported to increase in the number of receptor sites [14], the greatly enriched receptors at the plasma membrane might have provided the drastic improvement of S/N ratio. These results indicate that the bait receptor is predominant for successful detection of the dimerized receptors. Thus, the consideration of the receptor’s expression manner is important to screen the GPCR dimer partners, because the leaky background cell growth often brings 16985061 unanticipated candidates. In the optimized system, Hxt1p never presented background cell growth in the dimerization 18204824 assay with the Ste2DC receptor (Fig. 3A). Previously, Overton et al. had used the Hxt1p as the negative control for FRET analysis of Ste2p dimerization and confirmed the subcellular localization of it at the plasma membrane [15]. The b-galactosidase assay that reflects lacZ reporter enzyme activity also displayed similar trends (Fig. 3B). The addition of ligand had no effect on the dimerization events of Ste2DC, since the MAPK-defective NMY62 yeast strain displayed unchanged growth in the presence of a-factor (Fig. 3A). We additionally constructed two types of deletion mutants in which the TM6? domains and the TM1? domains were removed, respectively (TM1? (amino acids 1?36) and TM6? (amino acids 237?04)) (Fig. 4A). Overton et al. also indicated that selfassociation of TM6? had not been detected, and plasma membrane localization of the YFP-tagged TM1? and TM6? was observed [16]. As previously indicated [16], only TM1? formed the homodimer (Fig. 4B and C), showing that our system is applicable to examine critical domains involved in the dimerization of 7TM receptors. Subsequently, we validated the capability of our system to detect human GPCR heterodimer pairs. To avoid competitive dimerization with the endogenous yeast Ste2p receptor, we constructed an NMY63 mutant strain in which the STE2 gene was additionallyFigure 3. Detection for dimerization of yeast truncated Ste2p lacking the carboxy-terminal tail (Ste2DC) receptor in NMY62 strain. Growth and quantitative b-galactosidase activity of yeast cells expressing various combinations of Cub and NubG fusions. The control bait plasmid was pBT3-C mock vector (empty vector). The control prey plasmids were pPR3-C mock vector (empty vector) and pPR3-HXT1. (A) Growth assay without a-factor (left panels) and with 5 mM a-factor (right panels). Each cell was spotted in serial 10-fold dilutions on SD eu, Trp, Ade and His dropout plate. (B) Quantitative b-galactosidase assay. Error bars represent the standard deviations (n = 3). doi:10.1371/journal.pone.0066793.gdelete.
