Functional expression of olfactory receptors in yeast and development of a bioassay for odorant screening

Authors


E. Pajot-Augy, INRA, Neurobiologie de l'Olfaction et de la Prise Alimentaire, Domaine de Vilvert, 78352 Jouy-en-Josas Cedex, France
Fax: +33 1 34 65 22 41
Tel: +33 1 34 65 25 63
E-mail: Edith.Pajot@jouy.inra.fr

Abstract

The functional expression of olfactory receptors (ORs) is a primary requirement to examine the molecular mechanisms of odorant perception and coding. Functional expression of the rat I7 OR and its trafficking to the plasma membrane was achieved under optimized experimental conditions in the budding yeast Saccharomyces cerevisiae. The membrane expression of the receptor was shown by Western blotting and immunolocalization methods. Moreover, we took advantage of the functional similarities between signal transduction cascades of G protein-coupled receptor in mammalian cells and the pheromone response pathway in yeast to develop a novel biosensor for odorant screening using luciferase as a functional reporter. Yeasts were engineered to coexpress I7 OR and mammalian Gα subunit, to compensate for the lack of endogenous Gpa1 subunit, so that stimulation of the receptor by its ligands activates a MAP kinase signaling pathway and induces luciferase synthesis. The sensitivity of the bioassay was significantly enhanced using mammalian Golf compared to the Gα15 subunit, resulting in dose-dependent responses of the system. The biosensor was probed with an array of odorants to demonstrate that the yeast-borne I7 OR retains its specificity and selectivity towards ligands. The results are confirmed by functional expression and bioluminescence response of human OR17-40 to its specific ligand, helional. Based on these findings, the bioassay using the luciferase reporter should be amenable to simple, rapid and inexpensive odorant screening of hundreds of ORs to provide insight into olfactory coding mechanisms.

Abbreviations
Endo H

endoglycosidase H

GPCR

G protein-coupled receptor

KLH

keyhole limpet hemocyanin

OR

olfactory receptor

PMSF

phenylmethylsulfonyl fluoride

PNGase F

peptide N-glycosidase F

S14

somatostatin-14

SSTR2

somatostatin receptor subtype 2

PBST

0.05% Tween, NaCl/Pi

The olfactory receptors (ORs) are a large group of proteins belonging to subfamily I of G protein coupled receptors (GPCRs) that bind odorant ligands. These receptors are predicted to contain seven transmembrane helices that change their relative orientation upon odorant stimulation, resulting in the conformational change of the receptor and productive interaction of its intracellular loops with Golf, the α subunit of the heterotrimeric G protein [1–3]. Several lines of evidence suggest that the mechanism of OR activation by an odorant is central to understanding odorant perception and coding. Each OR recognizes multiple odorants and most odorants are recognized by several ORs [4–7]. One OR can discriminate between odorants with different functional groups, molecular size or shape and can even be sensitive to odorant concentration [8–10]. In addition, receptor perception of an odorant can be enhanced or antagonized by the presence of another odorant [8, 11,12]. Despite the importance of OR pharmacology to olfactory detection and discrimination, detailed characterization of ligand–receptor interaction has been achieved for relatively few ORs due to the absence of natural sources providing one receptor in sufficient amounts and to the inherent difficulties associated with the expression of ORs in heterologous cell systems [10].

A major hindrance to functional expression of ORs has been the tendency of ORs to be retained in endoplasmic reticulum of heterologous cells due to inefficient folding. This process results in receptor sequestration through the formation of aggregates and degradation before they can be transported to the plasma membrane [13,14]. The failure of ORs to translocate efficiently to the plasma membrane was also associated with the absence of adequate accessory proteins and chaperones in non-native cells, or with the absence of glycosylation at the N-terminus of the OR [15]. However, ORs do not even traffic well to the plasma membrane when expressed in a cell line derived from olfactory epithelium (ODORA cells) that exhibits some olfactory sensory neuron characteristics [10,16,17].

Plasma membrane trafficking of ORs in commonly cultured cell lines was slightly improved by appending N-terminal protein sequences from other seven transmembrane domain family members. Fusion proteins have been constructed between ORs and either the β2-adrenergic receptor [18], the N-terminal extension of rhodopsin [19] or the membrane import sequence of the serotonin receptor [20]. Functional expression of mouse 71 OR was dramatically increased upon coexpression with the β2-adrenergic receptor, but not that of rat I7 or human OR17-40 receptors [21]. This suggested that different ORs may require distinct GPCR partners to drive surface expression, maybe through their persistent physical association. An alternative approach for the functional expression of ORs utilized an adenovirus vector to deliver OR cDNAs to the sensory neurons of olfactory epithelia [4,8]. However, this approach has practical limitations due to the difficulty in maintaining olfactory neurons in primary culture, the inconsistency of viral-mediated gene transfer, and the cost if it was to be applied to a large number of ORs.

The aim of this study was to optimize the baker's yeast Saccharomyces cerevisiae as a host system for properly expressing an OR at the plasma membrane, and for its efficient coupling to a signaling pathway that produces a measurable response to odorant stimulation. The yeast system was chosen for several reasons. Firstly, S. cerevisiae has been successfully used for functional expression of many GPCRs [22–28]. Secondly, yeast constitutes an attractive system to study membrane receptors providing a null background for mammalian GPCRs and G proteins. Finally, yeast cells may provide a means for detailed investigation of receptor pharmacology in vivo through the use of sensitive reporter systems that take advantage of the functional homologies between yeast pheromone and mammalian GPCR signaling pathways.

In genetically modified yeast strains, the reporter system is activated after receptor–ligand interaction, Gα protein dissociation and activation of the MAP kinase pathway [29]. Several GPCRs have been shown to efficiently couple to the endogenous yeast Gα protein subunit, Gpa1. Yeast Gpa1, Ste4 and Ste18 are structurally and functionally similar to mammalian Gα, β and γ subunits, respectively [30]. In many cases, this functional coupling was improved by replacing Gpa1 by mammalian or chimeric Gpa1/mammalian Gα subunits that have been shown to interact with both the Ste4/Ste18 complex and heterologous GPCR [31,32]. Additionally, other elements of the mating signal transduction pheromone pathway were either deleted or functionally replaced by their mammalian or mutant counterparts to optimize S. cerevisiae for GPCR structure–function investigations [22,26,29,33].

In the present study we have used the rat I7 OR as a model to investigate the OR expression in yeast since its preferred ligands (octanal, heptanal, nonanal) and their effective concentration ranges have already been determined [4,8,10,19]. We recently described how S. cerevisiae can successfully be engineered as a reporter system for odorant detection [34]. Two different yeast strains expressing an odorant receptor were only able to grow on selective media following specific odorant ligand stimulation. However, this growth reporter had very limited sensitivity and was poorly adapted to the transitory nature of the response. Thus, methodological improvements were drastically needed for this system to be ultimately used for pharmacological screening purposes. Here, we use luciferase as a rapid reporter to study the I7 OR pharmacology. We have optimized the experimental conditions for the production of the I7 OR in yeast and used biochemical and immunological methods to estimate the levels of receptor expression and its cellular localization.

Results

Yeast transformations

Functional expression of the I7 or OR17-40 receptor was achieved in the yeast strain MC18 modified to allow sensitive bioassay based on synthesis of luciferase upon odorant stimulation. Strain MC18 has been reported to have an unknown mutation that prevents cell cycle arrest upon activation of the pheromone-mating signaling pathway and to lack GPA1 gene [35]. The yeast strains with the I7 expression vector, pJH2-I7, or the OR17-40 expression vector, pJH2-OR17-40, were then transformed, respectively, with either pRGP-Golf or pRGP-Gα15 vectors to replace the lacking GPA1 gene, or with PRGP-Golf. RT–PCR analysis conducted on RNA extracted from the strains shows mRNA bands for the ORs (Fig. 1) and the two Gα proteins at expected sizes (data not shown). In addition, RT–PCR analysis demonstrated that mRNA of the olfactory receptors is present in both uninduced and induced yeast cells (Fig. 1). This indicates a leakage of GAL1/10 promoter in glucose-containing minimal medium. The reporter plasmid pRHF-luc was cotransformed in the yeast strains at the same time as the pRGP vector. It places expression of luciferase under control of the FUS1 promoter, which is activated downstream of the MAP kinase cascade (illustrated in Fig. 2).

Figure 1.

mRNA of the rat I7 OR or human OR17-40 receptor in transformed yeast strains. Total RNAs from either uninduced or induced strains transformed with pJH2-I7 or pJH2-OR17-40 expression vectors were subjected to RT-PCR using specific primers to demonstrate that mRNA synthesis of the ORs occurs even in uninduced cells. I7 cDNA in pJH2-I7 plasmid and OR17-40 cDNA in pJH2-OR17-40 plasmid were amplified in parallel as positive controls.

Figure 2.

Modifications of yeast pheromone and signal transduction pathway to yield agonist-induced luciferase activity. The heterologous OR receptor activated by its odorant ligand couples to an heterotrimetric G-protein consisting of a mammalian Gα subunit (Golf or Gα15) and the yeast βγ subunit, Ste4/Ste18. Gα dissociates from the complex and allows Ste4/Ste18 to activate the mating MAP kinase cascade. This in turn induces luciferase synthesis under the control of the FUS1 promoter, allowing quantitative readouts of the receptor–ligand interaction.

Biochemical characterization of the yeast I7 OR

To examine the presence of the I7 protein in uninduced and induced yeast cells, membrane preparations were analyzed by immunoblotting using a polyclonal antibody raised against I7. No immunoreactivity was detected in control membrane preparations from either nontransformed MC18 yeast cells or cells transformed with the initial pJH2-somatostatin receptor subtype 2 (SSTR2) plasmid (Fig. 3A). In the case of yeast cells transformed with the pJH2-I7 expression vector two immunoreactive bands were observed at approximately 40 and 51 kDa (Fig. 3A). The calculated molecular weight for the I7 OR is 39 kDa, it is therefore likely that the 40 kDa band corresponds to the receptor monomer. The 51 kDa band may correspond to a glycosylated form of the receptor since I7 OR contains two N-glycosylation sites, one at the N terminus and another in the second extracellular loop.

Figure 3.

Immunoblot analysis of the I7 OR in yeast membrane extracts. (A) Membranes were prepared from nontransformed yeast (MC18), and transformed with either control expression vector, pJH2- SSTR2 or pJH2-I7. Each lane was loaded with 30 µg of membrane proteins. Samples were analyzed by SDS/PAGE followed by Western blotting using anti-I7 IgG. Yeast were grown at 15 °C. (B) The glycosylation status of the I7 OR was investigated on membranes from yeast transformed with pHJ2-I7 expression vector induced with 2% galactose at 15 °C. Membrane proteins were digested with either Endo H or PGNase F. The corresponding controls were performed by using water instead of deglycosylation enzymes in respective incubation buffer. Each lane was loaded with 10 µg of protein. Immobilized samples were probed with anti-I7 IgG.

To determine if the I7 receptor is glycosylated in S. cerevisiae, the membrane fraction from the induced yeast cells was digested with either endoglycosidase H (Endo H) or peptide N-glycosidase F (PNGase F). Figure 3B shows that the 51-kDa band is sensitive to both Endo H and PNGase F digestion, resulting in almost complete deglycosylation to yield the 40-kDa band. The 51-kDa band thus represents the receptor monomer with the exclusive addition of mannose residues.

The presence of the immunoreactive bands in the lanes relative to uninduced yeast cells (Fig. 3A) is a further sign of GAL1/10 promoter leakage, as already seen at the mRNA level.

Luciferase bioassay and functional expression of the ORs in yeast

To address the functional integrity of the ORs expressed in yeast, we developed a luciferase reporter bioassay. Initially, the bioassay was configured using a control yeast strain transformed to coexpress luciferase under control of FUS1 with the SSTR2 receptor and Gpa1. When this strain was incubated with either α-factor to stimulate endogenous α-factor receptor (Ste2), or with somatostatin 14 (S14) to stimulate SSTR2 receptor, luciferase activity was observed to increase in a dose-dependent manner (data not shown). Thus, the bioassay provides an effective readout of GPCR–ligand interaction and we therefore applied the assay to monitor the I7 or OR17-40 activity.

Figure 4A summarizes differential luciferase-mediated luminescence detected in the yeast strains expressing I7 OR, grown at various conditions, following their stimulation with 5 µm heptanal, octanal or nonanal. In this set of experiments the effects of yeast growth temperature and GAL1/10 promoter induction were tested in order to optimize I7 OR functional expression. As some GPCRs were reported to fold and traffic better to the plasma membrane when their expression level is restricted by reduced temperatures [36,37], we examined I7 OR activity in yeasts grown at 30 or 15 °C. Indeed, luciferase-mediated responses to odorants were dependent on the yeast growth temperature. The temperature shift to 15 °C markedly improved the functional response of the receptor (Fig. 4A).

Figure 4.

Effects of galactose induction and temperature on functional expression of the I7 OR in yeast. (A) Luciferase bioluminescence was measured following yeast stimulation by odorants in two strains expressing either I7 OR and Golf, or I7 OR and Gα15. Strains were grown in minimal media with 2% glucose (uninduced) or subsequently in minimal media with 2% galactose (induced) at 30 or 15 °C. Differential bioluminescence for each sample was calculated with respect to controls that were prepared by replacing the odorant by water. The data were recorded as the mean ± SEM of three separate experiments. (B) Western blot analysis of the I7 receptor, Golf and Gα15 expression levels in the corresponding yeast membrane lysates as in (A) (20 µg of protein per lane). Note that the highest level of I7 receptor expression is obtained in induced strains at 15 °C while the levels of Golf and Gα15 do not change significantly.

The luciferase reporter activity was compared in uninduced and galactose-induced conditions since a GAL1/10 promoter leakage had been detected in uninduced yeasts (Figs 1 and 3A). The luciferase-mediated luminescence responses to odorant stimulations were increased in induced cells compared to uninduced cells (Fig. 4A), indicating that galactose induction increases the yield of functional I7 OR relative to the leakage level.

Figure 4A also shows that functional responses to odorants were several-fold higher in the strain coexpressing the Golf subunit in comparison to the strain coexpressing the Gα15 subunit when grown in the same conditions.

In addition, the protein levels of I7 OR, Golf and Gα15 in strains from Fig. 4A were compared. The immunoblot analysis indicated that the levels of Golf and Gα15 in membrane fractions are constant regardless of the temperature and galactose induction, while the level of the I7 OR is higher in membrane fractions from yeasts induced by galactose at 15 °C (Fig. 4B). Thus, the highest level of both membrane-associated and active I7 OR were obtained in yeast induced by galactose at 15 °C. Therefore, we chose to perform pharmacological analysis on the yeast coexpressing I7 and Golf under these conditions.

Pharmacological characterization of the yeast I7 OR

Previous pharmacological investigations of the I7 OR expressed in mammalian cells have shown that receptor responses to odorants are dose dependent [10,38]. In this study, the yeast-borne I7 OR was stimulated by heptanal, octanal or nonanal over the concentration range from 5 × 10−14 to 5 × 10−4 m. All three ligands evoked luciferase reporter activity in a dose-dependent manner as presented in Fig. 5A. Response thresholds for heptanal, octanal and nonanal were 3 × 10−9, 7 × 10−9, and 5 × 10−8 m, respectively. The maximal amplitude was detected for 5 × 10−7 m octanal.

Figure 5.

Differential bioluminescence dose–response upon odorant stimulation of yeast-expressed olfactory receptors. Measurements were performed on yeast transformed to coexpress the I7 OR, Golf and the luciferase reporter (A), and on yeast coexpressing the human OR17-40, Golf, and the luciferase reporter (B). These strains were induced with 2% galactose at 15 °C. Dose–response curves are plotted as a difference of bioluminescence response to odorants relative to controls obtained by replacing odorants with water.

Interestingly, as observed with the I7 OR expression in mammalian cells, dose–response curves of the yeast OR I7 were bell-shaped instead of exhibiting a plateau at high ligand concentrations. As shown in Fig. 5A no significant response was detected for odorant concentrations 5 × 10−4 m or higher. This could be due to odorant and/or its solvent toxicity to yeast cells or to receptor desensitization at the highest ligand concentrations. It was checked that a final odorant concentration of 5 × 10−5 m corresponds to a concomitant dilution of the organic solvent that is not deleterious to the yeast cells. This was tested by cell shape examination and count after incubation in the odorant dilutions, and also by monitoring the influence of odorant dilutions on the luciferase bioassay performed with S14 stimulation of a control yeast strain coexpressing SSTR2, Gpa1 and the luciferase reporter. Only final odorant concentrations of the three aldehydes above 10−4 m were toxic for yeast cells, and this was ascribed to the presence of the organic solvent (dimethyl sulfoxide) itself, which was deleterious for the yeast cells. So the bioluminescence response as a function of odorant concentration is significant up to 5 × 10−5 m. The narrow bell-shaped dose–response curves in the range from 5 × 10−8 m to 5 × 10−5 m indeed define the operational range of the I7 receptor.

We also examined receptor specificity by testing whether a panel of nine various odorants might be recognized by the yeast I7 OR. Among these, octanol, octanon and octanoic acid were selected as they possess the same carbon chain length but have different functional groups. None of them induced any luciferase activity when tested over a wide range of concentrations (5 × 10−12 to 5 × 10−4 m).

Similarly, the commonly used odorants isoamyl-acetate, lyral, lilial, pyridine, diacetyl and cyclohexyl-acetate, tested over the same concentration range, failed to induce any luciferase activity. These findings are consistent with those obtained with the I7 OR expressed in mammalian cells [8,10] and strongly suggest that yeast expressed I7 OR retains the ligand selectivity and specificity equivalent to its mammalian expressed counterpart.

The yeast-expressed OR17-40 was stimulated with helional in the concentration range 5 × 10−14 to 5 × 10−4 m, yielding a bioluminescence dose–response curve shown in Fig. 5B, with a threshold concentration of 6 × 10−8 m, and maximal amplitude for 5 × 10−6 m. As in the case of I7 OR, this curve is bell shaped, and finely tuned for helional concentrations between 5 × 10−5 m and 5 × 10−7 m. In order to test OR17-40 specificity, heptanal was used as a negative control [20] over the whole range of concentrations studied (data not shown).

Cellular localization of the I7 OR in yeast

Immunoblot analysis showed that the I7 receptor is associated with the yeast membrane fraction. In order to check that the I7 receptor is, at least partly, associated with yeast plasma membrane, the I7-specific antibody was used to immunostain nonpermeabilized spheroplasts. Spheroplasts of nontransformed MC18 yeast cells showed no staining with the anti-I7 IgG(Fig. 6). In contrast, all the spheroplasts of I7 OR yeast strain cells that had been induced with galactose at 15 °C showed an intense cortical labeling (Fig. 6) indicating a clear presence of the receptor at the plasma membrane.

Figure 6.

Immunofluorescence confocal microscopy of the I7 OR in yeast. Optical (left panels) and confocal (right panels) visualization of spheroplasts from nontransformed MC18 yeast and yeast transformed with pJH2-I7 expression vector (induced at 15 °C). Immunolabeling was performed with the anti-I7 IgG and an Alexa488-coupled secondary antibody on nonpermeabilized spheroplasts. Scale bar, 5 µm.

In order to examine the ultracellular localization of the receptor, immunogold electron microscopy was performed on induced I7-transformed yeast cells grown at 15 °C (Fig. 7). The presence of the I7 OR was obvious at the plasma membrane (two to four gold particles per section) demonstrating that the I7 molecules are targeted to their functional location (long arrows). A few gold particles were associated with vesicular structures located at the plasma membrane (double arrows), consistent with the membrane trafficking of the I7 OR molecules. In the cytoplasm, gold particles were associated with endoplasmic reticulum cisternae (short arrows), thus localizing the I7 OR molecules to their secretion pathway. The receptor was also present in vacuoles (arrowheads) and sometimes associated with vacuole membranes (open arrowheads). No gold grains were observed on sections where the primary or the secondary antibody was omitted. The presence of the gold particles associated with the plasma membrane indicates that at least some of the I7 molecules produced are inserted at the site commensurate with their ability to sense the external environment.

Figure 7.

Ultrastructural localization of the I7 OR in yeast. Yeast strain transformed with pJH2-I7 expression vector and induced at 15 °C was immuno-labeled with the primary anti-I7 IgG and 10 nm gold-conjugated secondary antibody. Gold grains are present on the plasma membrane (long arrows) and vesicles near the plasma membrane (double arrows). They are also associated with endoplasmic reticulum cisternae (short arrows), the vacuole (arrowheads) and sometimes with vacuolar membrane (open arrows). Bar, 0.25 µm.

The I7 OR quantification by ELISA-type test

To quantify the level of I7 OR associated with membranes, an ELISA-type test was carried out using the specific anti-I7 IgG. As purified I7 receptor is not available the calibration curve was generated by serial dilution of keyhole limpet hemocyanin (KLH)-coupled I7 antigenic N-terminal 15-amino acid peptides. Fig. 8A shows this calibration curve as well as the negative control obtained by probing KLH alone. Using the KLH-coupled I7 peptide as a standard, the I7 antigen concentration in the range 1–100 µm could be measured accurately (SD < 10%). Fig. 8B shows the ELISA measurements collected from the serial dilution of membrane preparations from yeast cells expressing the I7 OR induced at 15 °C. Membrane proteins from control SSTR2-strain were included as a negative control. Increasing amounts of membrane proteins from the control yeast failed to elicit an ELISA signal indicating the specificity of the reaction (Fig. 8B). In contrast, the I7 expressing yeast produced a dose-dependent signal that could be saturated with higher amounts of membrane fraction bearing I7 OR. The concentration of I7 receptor produced in induced yeast was deduced to be 327 pmol·mg−1 of membrane protein, i.e. 1.44 × 105 receptor per cell. This compares to 352 pmol·mg−1 of membrane protein for recombinant expression of the α-factor receptor Ste2p itself in S. cerevisiae[25], which is the best level ever achieved for any GPCR membrane expression in yeast.

Figure 8.

ELISA-type test for quantification of I7 expressed in yeast cells. The ELISA test was performed by using an anti-I7 primary antibody and a biotin-conjugated antibody plus horseradish streptavidine peroxidase to quantify the presence of I7 OR. (A) Calibration curve obtained by probing KLH-coupled I7 antigenic N-terminal 15-amino acid peptides. Uncoupled KLH was probed as a control. (B) ELISA signal plotted as a function of increasing amount of total proteins in membrane fractions from I7 yeast strain induced at 15 °C, and from yeast expressing SSTR2 as a control.

Discussion

In this study we developed a novel, robust and sensitive yeast-based bioassay for odorant screening. Yeast was engineered to functionally express an olfactory receptor in conjunction with a mammalian Gα subunit and to exhibit agonist-dependent luciferase reporter activity. By taking advantage of structural and functional similarities between yeast and mammalian GPCR signaling pathways, this assay enables the quantitative measurement of receptor activity, or alternately the detection of its ligands. Using known ligands of I7 OR (heptanal, octanal and nonanal), we successfully demonstrated that they act as agonists as already experienced in mammalian cells. Odorants of the same carbon chain length but with different functional groups failed to induce any luciferase activity demonstrating that the yeast borne receptor retains its ligand specificity and selectivity. In addition, validation of the system was completed by demonstrating that six commonly used odorants do not stimulate luciferase activity. OR17-40 also exhibited an intense and specific bioluminescence in response to helional stimulation. This corroborates the adequacy of the bioassay for a totally different olfactory receptor.

By fusing the pheromone inducible FUS1 promoter sequence to the coding sequence of luc, the expression of luciferase was regulated through activation of the MAP kinase signaling pathway upon OR–odorant interaction, to allow quantification of the dose–response to odorants. The luciferase reporter was chosen for its sensitivity, rapidity and easy to perform enzymatic reaction. In a previous study we used the FUS1–His1 or FUS1–Hph reporters to provide odorant-dependent yeast growth on histidine-deficient or hygromycin-containing medium, respectively [34]. Such assays are commonly used for GPCRs but have low sensitivity and are time consuming as they include a delay of 24–48 h in response [24,34]. Therefore, they are not best adapted to study ORs regarding the transitory nature of their response to odorants, their desensitization and recycling observed in mammalian cells and possible degradation of odorant molecules at yeast growth temperatures. β-galactosidase assay for GPCR agonist screening is more rapid but requires relatively expensive fluorogenic substrates for sensitive readout [31]. We believe that our luciferase sensor can be at the basis of efficient, rapid and low cost screening of a large range of odorants.

The signal can be significantly enhanced through Gα subunit engineering. The intense reporter activity registered demonstrates that the receptor naturally coupling Gα protein, Golf, is able to interact efficiently with both the heterologous OR and the endogenous Gβγ complex, Ste4/Ste18. The efficient coupling of Golf to the pheromone response pathway was previously demonstrated when it complemented a Gpa1 null mutation in S. cerevisiae[35]. This is in contrast to a chimeric Gpa1–Golf, which showed poor coupling efficiency with either the OR and/or Ste4/Ste18 [34]. We also observed higher sensor sensitivity with Golf than with the promiscuous Gα15 commonly used for pharmacological studies of recombinant ORs. This probably arises from the poor affinity of Gα15 for yeast Gβγ, as such a lack of affinity has already been reported for Gpa1/Gα15,16 chimeras in S. cerevisiae[31,32].

Another aspect of this study was the optimization of OR functional expression in S. cerevisiae. During the last decade only a handful of mammalian ORs have been functionally expressed in heterologous systems due to inefficient receptor insertion into the plasma membrane [10,13,14,17]. Here, we found that I7 OR functional responses to odorants were notably enhanced when yeasts were induced in galactose-containing medium at 15 °C. The achievement of high I7 response to ligand stimulation was correlated to its improved expression, since on Western blots a significant increase in receptor level was observed in the membrane fraction. Under these conditions, neither aggregation of possibly misfolded receptors within the yeast, nor yeast vacuole overloading with species intended for degradation were observed by immunogold labeling. Thus, it appears that galactose induction at 15 °C provides adequate conditions for functional receptor expression. It remains unclear how S. cerevisiae responds to mild low temperatures and at which stage of the folding/trafficking process the reduced temperatures have an effect. Recently it was reported that in S. cerevisiae a temperature downshift to 10–18 °C leads to an induction of specific ‘cold shock proteins’, some of which are able to serve as molecular chaperones [39]. Such proteins could be involved in the upregulation of I7 OR functional expression observed at 15 °C. However, other mechanisms that arise upon lowering the temperature must also be considered. For instance, lower temperature may positively affect the yield of properly folded proteins [40–42]. Also, it is interesting to note that reduced temperatures increase the content in higher sterols within yeast cell membranes [43]. This may not only improve receptor insertion into the plasma membrane [44], but also allow correct receptor activity [45].

The achievement of receptor plasma membrane insertion was demonstrated by confocal immunofluorescence microscopy of nonpermeabilized spheroplasts and by ultrastructural immunogold analysis. In addition, immunological analysis of raw and deglycosylated samples showed that the predominant receptor form in the membrane fraction is the mannose-glycosylated monomer. Indeed, only high mannose elongation of core sugars can occur in S. cerevisiae, contrary to mammalian cells [46]. However, considering the yeast I7 receptor discrimination between closely related ligands strongly suggests the authenticity of its ligand binding and the maintenance of the coding ability at the receptor level. Consequently this suggests that glycosylation of I7 OR is not a major determinant of receptor pharmacology.

Odorant concentrations giving rise to responses in yeast cells are several orders of magnitude higher than those observed in COS or ODORA cells [10]. This discrepancy in the behavior of I7 and OR17-40 receptors expressed in yeast vs. mammalian cells could be due to differences in the lipid membrane composition and organization between the two heterologous systems [45]. Nevertheless, by comparing the threshold concentrations for I7 OR response to odorant stimulation, we find that heptanal ranks first as in COS cells, where as octanal and nonanal are less potent ligands. Thus, although higher odorant concentrations are necessary to activate the receptor in this nonmammalian cellular background, the receptor affinity ranking and selectivity are close to those in mammalian cells.

The yeast system was optimized for functional expression and sensitive characterization of the olfactory receptors. It could be amenable to a rapid, inexpensive screening assay with an extended dynamic range, in which the many orphan ORs could be investigated against the extraordinary large number of naturally occurring odorants. Although optimization is certainly required for transfer to a high throughput format, this method demonstrates a potential for conveniently screening a large number of organic molecules as novel GPCR ligands which could serve as leads for drug discovery.

Experimental procedures

Odorants and other reagents

Odorant solutions were prepared just before use as described previously [10,34]. Octanal, nonanal, heptanal, diacetyl, cyclohexyl-acetate, octanol, octanon, octanoic acid, isoamyl-acetate, pyridine were from Sigma-Aldrich (Saint Quentin, Fallavier, France). Helional was a generous gift from Givaudan-Roure (Dübendorf, Switzerland), courtesy of B Schilling. Lyral and lilial were kindly provided by Roche (Meylan, France).

Complete protease inhibitor cocktail, Endo H and PNGase F were from Roche Diagnostics GmbH (Mannheim, Germany). NaCl/Pi pH 7.4 was from Oxoid (Basingstoke, Hampshire, England). Phenylmethylsulfonyl fluoride (PMSF), KLH, poly(l-lysine) (Mr > 300 000), meta-periodate and Tween 20 were from Sigma-Aldrich. RQ1 RNAse-free DNAse was from Promega (Charbonnieres-les-Bains, France). Enzymes used for molecular cloning were from Promega and New England Biolabs (Beverly, MA, USA). The DNA size marker was DRIgest III from Amersham Pharmacia Biotech Europe (Orsay, France). Protein size markers were Broad Range Prestained SDS/PAGE Standards from Bio-Rad (Marnes la Coquette, France).

Expression vectors

All plasmid manipulations were performed in E. coli strain DH5α from Gibco BRL (Invitrogen, Cergy Pontoise, France). After selection, a single colony was used to isolate the plasmid for yeast transformation. The multicopy plasmid construct, pJH2-I7, for I7 receptor expression, was obtained by homologous recombination in the pJH2-SSTR2 expression vector (kindly provided by MH Pausch, Cyanamid Agricultural Research Center, Princeton, NJ, USA) as described previously [34]. OR17-40 full-length sequence was cloned into a pGEM-T vector, then inserted in the pCMV-Tag3 expression vector for N-terminal c-myc tagging using sites BamHI and XhoI of the multiple cloning site as described previously [10]. pJH2-OR17-40 expression vector was obtained from pJH2-SSTR2 by homologous recombination introducing the c-myc-OR17-40 coding sequence, using primers (5′-CGTCAAGGAGAAAAAACCCCGGATCTAAAAAATGGAGCAGAAACTCATCTCTGAAGAGGATCTG-3′) and (5′-GCATGCCTGCAGGTCGACTCTAGAGGATCTCAAGCCAGTGACCGCCTCCC-3′), and checked for the presence and sequence of the new insert, as in the case of pJH2-I7.

Plasmids pJH2-I7 and pJH2-OR17-40 carry a galactose inducible GAL1/10 promoter. The expression vector also contains a GAL4 gene under the control of the GAL10 promoter. The induction of the yeast, by galactose containing media, results in overexpression of GAL4, in turn inducing an increase of the expression of the OR gene under control of GAL1. The pJH2 vector contains the URA3-selectable marker.

Two Gα protein expression vectors, pRGP-Golf, or pRGP-Gα15 were used. The pRGP-Golf vector with the cDNA of Golf under control of Gpa1 promoter was reported by Crowe et al. [35]. The pRGP-Gα15 expression vector was obtained by replacing Golf coding sequence by the Gα15 coding sequence. The pRGP vector contains HIS3-selectable marker. To endow the yeast strain with a reporter capacity, a pRHF-luc plasmid was constructed by replacing the hph coding sequence from the pRHF-hph plasmid [35] by the luciferase coding sequence. In this vector the Photinus pyralis cDNA sequence is placed downstream the FUS1 promoter. The pRHF-luc vector contains TRP1-selectable marker.

Yeast transformation, growth and galactose induction

The S. cerevisiae strain MC18 (MATa gpa1::lacZ[LEU2]ade2-1 his3-11, 15 leu2-3112 trp1-1 ura3-1 can1–100) [35] was transformed with either pJH2-I7 or pJH2-OR17-40, pRGP-Golf, pRHF-luc or with pJH2-I7, pRGP-Gα15, pRHF-luc expression vectors using the lithium acetate method [47]. Transformed cells were plated on 2% agar in media A: yeast nitrogen base (Difco, Detroit, MI, USA), synthetic drop-out CSM media without HIS, LEU, TRP, URA (Bio101, Inc., Vista, CA, USA), 40 mg·mL−1 adenine, complemented with 2% glucose. The colonies were grown in liquid media A complemented with 2% glucose at either 30 or 15 °C, until they reached exponential growth phase (attenuance at 600 nm, D600, in the range 1–2). The presence of plasmids in transformed cells was verified by PCR on nucleic acid extracts. Induction of I7 expression was performed as reported for the SSTR2 induction [24] with the exception of the temperature. In brief, the cells were washed to remove glucose and cultured for 4–6 h in the selection media containing 3% lactate, then pelleted and diluted to a D600 0.5 and finally cultured in the selection media A containing 2% galactose at either 30 or 15 °C for about 18 or 60 h, respectively. All subsequent experiments with either uninduced or induced yeasts were carried out with cells in exponential growth phase (D600 in the range 1–3).

RNA extraction and RT-PCR

RNA was extracted from yeast cells following the hot acidic phenol procedure. RT-PCR was performed on DNAse-treated RNA extracts. Primers used for RT-PCR were: for the I7 OR (5′-CGTCAAGGAGAAAAAACCCCGGATCTAAAAAATGGAGCGAAGGAACCACAG-3′) and (5′-AGCTGCCTGCAGGTCGACTCTAGAGGATCCTAACCAATTTTGCTGCC-3′); for OR17-40 (5′-CGTCAAGGAGAAAAAACCCCGGATCTAAAAAATGGAGCAGAAACTCATCTCTGAAGAGGATCTG-3′) and (5′-GCATGCCTGCAGGTCGACTCTAGAGGATCTCAAGCCAGTGACCGCCTCCC-3′); for Golf (5′-GGTACCGCTGCAATGGGGTGTTTGGGCAAC-3′) and (5′-GCGGCCGCCTCAGATCACAAGAGTTCGTACTGC-3′); for Gα15 (5′-ATGGCCCGGTCCCTGACTTGG-3′) and (5′-TCACAGCAGGTTGATCTCGTCC-3′). Negative controls for the presence of remaining DNA were provided by RT-PCR with the same primers performed on nonreverse transcribed mRNA.

Isolation of yeast membranes

Membranes were prepared from yeast cells washed twice with ice-cold water, harvested by centrifugation and resuspended in an equal volume of ice-cold lysis buffer (50 mm Tris/HCl pH 7.5, 1 mm EDTA, 0.1 mm PMSF, 250 mm sorbitol) and the Complete protease inhibitor cocktail. Glass beads (425–600 µm, Sigma) were added and cells were disrupted by seven cycles of 1 min of vigorous vortexing/1 min of cooling on ice. Samples were pooled and centrifuged at 5000 g for 10 min at 4 °C to remove unbroken cells and cell walls. The supernatant was further centrifuged at 40 000 g for 40 min at 4 °C. This second pellet, enriched in membranes, was resuspended in the lysis buffer with a Dounce homogenizer, and stored in aliquots at −80 °C. The protein concentration of the membrane preparation was determined using the BCA reagent (Pierce, Brebieres, France) with BSA as a standard.

Immunoblot analysis

Proteins of the membrane fraction were separated by electrophoresis on 12% SDS polyacrylamide gels and electrotransferred onto Hybond-C Extra membrane (Amersham Pharmacia Biotech Europe). The membrane was blocked with 5 μg·mL−1 polyvinyl alcohol for 1 min. This reaction was stopped by soaking the membrane in 4.5% nonfat dried milk in NaCl/Pi. Membranes were incubated overnight at 4 °C with rabbit anti-I7 polyclonal antibody raised against its N-terminal 15 amino acids (custom made by Neosystem, Strasbourg, France), rabbit anti-Golf (1 : 500, Santa Cruz Biotechnology, Santa Cruz, CA, USA) or goat anti-Gα16 (1 : 500, Santa Cruz Biotechnology) at 1 µg·mL−1 in 4.5% nonfat dried milk in NaCl/Pi. After washing, membranes were incubated for 1 h at room temperature with either biotin conjugated anti-rabbit IgG (Sigma) (1 : 1000) and streptavidin-horseradish peroxidase conjugate (Amersham Pharmacia Biotech Europe) (1 : 1000) or anti-goat IgG conjugated to horseradish peroxidase (1 : 2000) diluted in the same buffer. Blots were revealed using the enhanced chemiluminescence (ECL) detection kit from Amersham Pharmacia Biotech Europe.

Glycosylation studies

Deglycosylation with Endo H was performed by incubation of 2 mg·mL−1 membrane proteins with 0.14 U·mL−1 Endo H in 40 mm sodium citrate buffer, pH 5.5, 0.5% SDS, 2 mm PMSF for 3 h at 37 °C. Deglycosylation with PNGase F was performed by incubation of 3 mg·mL−1 membrane proteins with 7 U·mL−1 of PNGase F in NaCl/Pi pH 7.5, 0.5% SDS, and 2 mm PMSF overnight at 37 °C. Samples were subsequently immunoblotted as described above.

ELISA quantification of I7 expressed at yeast membrane

Quantification of the I7 receptor expression was performed using the I7 antibody in an ELISA calibrated against the initial antigen comprising the N-terminal 15-amino acid peptide, coupled with glutaraldehyde to KLH as a carrier protein. Serial dilutions of membrane fraction of yeast expressing the I7 OR (3.5 mg·mL−1 total protein) or the KLH-coupled-antigenic peptide (0–1 × 10−13 mol) were deposited in the poly(l-lysine) (0.01%) coated wells of a 96-well plastic plate for 1 h at 37 °C. Corresponding dilutions of KLH alone, membrane fraction of nontransformed yeast, or membrane fraction of yeast transformed with the pJH2-SSTR2 plasmid and thus expressing the SSTR2 receptor instead of the I7 OR [24] were also deposited as negative controls. The plates were saturated for 2 h in the blocking buffer [3% (v/v) goat serum, 3% (w/v) BSA in NaCl/Pi], then incubated overnight at 4 °C with the anti-I7 IgG in blocking buffer (1 : 200). After rinsing three times with 0.05% (v/v) Tween, NaCl/Pi (PBST) and three times with NaCl/Pi the plates were incubated for 1 h at 37 °C with the secondary anti-rabbit biotinylated antibody (1 : 500) and horseradish streptavidine peroxidase (1 : 500) in blocking buffer. After extensive washing with PBST and NaCl/Pi, 3,3′,5,5′-tetramethylbenzidine kit from Kirkegaard & Perry Laboratories (Gaithersburg, MD, USA) was used to yield a colorimetric reading.

Immunodetection and confocal microscopy

Yeast cells were fixed in 1/10 (v/v) formaldehyde for 30 min and washed twice with 1.2 m sorbitol, 1% (v/v) 2-mercaptoethanol, 0.1 m potassium phosphate buffer, pH 6.5. Cells were resuspended in this buffer and transformed into spheroplasts by incubating the cells with 50 U·mL−1 lyticase (Sigma) for 15 min at 30 °C while shaking. The spheroplasts were then washed twice with 1.2 m sorbitol, 0.1 m potassium phosphate buffer, pH 6.5, and deposited on 0.01% (w/v) poly(l-lysine)-coated glass slides. Slides were treated with blocking buffer [3% (v/v) goat serum, 3% (w/v) BSA in NaCl/Pi] for 1 h at room temperature. Each slide was then incubated with the primary anti-I7 IgG (0.001 mg·mL−1) diluted in the blocking buffer overnight at 4 °C. After washing three times with 1% (w/v) BSA, PBST, and once with 1% (w/v) BSA, NaCl/Pi, the slides were incubated with secondary Alexa488-labelled anti-rabbit IgG (Molecular Probes, Eugene, OR, USA) diluted 1 : 3000 in the blocking buffer for 1 h at room temperature in the dark. After incubation, the slides were washed three times with PBST, twice with 0.1 m NaHCO3, 0.15 m NaCl, pH 8.2, and twice with NaCl/Pi. Slides were mounted with Vectashield antifading mounting medium (Vector Laboratories, Inc., Burlingame, CA, USA) and stored at 4 °C in the dark until viewed. Immunolabeled spheroplasts were observed with a Carl Zeiss LSM 310 confocal laser scanning microscope. Images were treated using imagej and Adobe photoshop (Adobe Systems, San Jose, CA, USA) softwares.

Immunodetection and electron microscopy

Yeast cell fixation and embedding were carried out according to the protocol described by Sander et al. [23]. Briefly, cells were fixed with 4% (v/v) paraformaldehyde, 2.5% (v/v) glutaraldehyde and 1% (w/v) meta-periodate in 0.1 m cacodylate buffer pH 7.3, for 3.5 h at room temperature. Afterwards, the cells were washed twice with this buffer and incubated overnight in buffered glycine (2%). The following day, cells were postfixed in 1% OsO4 in cacodylate buffer for 1 h, washed with water, subsequently treated with aqueous 2% (w/v) uranyl acetate for 1 h and enclosed in 2% (w/v) agar-agar. After consolidation at 4 °C and fixation in 2.5% (v/v) glutaraldehyde for 15 min, the pellet was cut into 1 mm3 pieces. These were dehydrated in a graded ethanol series and embedded in epoxy resin (LX112, Ladd Research Industries, Inland Europe, Conflans/Lanterne, France). Ultra-thin sections (50–100 nm) were cut with an ultra-microtome (Ultracut, Reichert, Vienna, Austria) and collected on nickel grids for immunogold labeling. For ultrastructural localization of the I7 OR, sections collected on nickel grids were permeabilized at room temperature with saturated meta-periodate water solution for 1 h, washed with water then with 0.1 m HCl (10 min) and again with water. Free aldehydic sites were quenched by incubation with 2% (w/v) glycine in NaCl/Pi. Afterwards, nonspecific sites were blocked with incubating buffer consisting of 10% (v/v) normal goat serum and 5% (v/v) BSA, 0.5% (v/v) Triton X-100 and 0.5% (v/v) Tween 20 in NaCl/Pi for 1 h. After several washes in the incubation buffer, sections were incubated overnight with the anti-I7 IgG (1 : 50) in the incubation buffer in a wet chamber at 4 °C. The sections were washed in NaCl/Pi containing 0.1% (w/v) acetylated BSA (NaCl/Pi/BSAc, Aurion, Wageningen, the Netherlands), and then incubated with 10 nm gold-conjugated goat anti-rabbit F(ab′)2 fragments (Aurion) diluted 1 : 40 in NaCl/Pi/BSAc applied 1.5 h at room temperature. After extensive washing in NaCl/Pi/BSAc and NaCl/Pi, the antigen–antibody complex was stabilized with 2.5% glutaraldehyde in NaCl/Pi. The sections were then contrasted using Reynolds' lead citrate before observation. Controls for the immunocytochemical reaction were carried out by replacing either the primary or the secondary antibody by the incubation buffer in the reaction sequence. The sections were finally viewed under a CM12 Philips electron microscope.

Functional assay in vivo

Two million cells in 200 µL culture media were incubated with an odorant for 60 min at room temperature to induce the reaction scheme summarized in Fig. 1. The yeast cells were then pelleted and resuspended in 200 µL 25 mm glycylglycin buffer (pH 7.8), 1 mm EDTA, 8 mm MgSO4, 1% (v/v) Triton X-100, 15% (v/v) glycerol, 1 mm dithiothreitol. Samples were homogenized for 20 s with a Potter in an Eppendorf tube and luciferase activity was recorded from 100 µL placed in a Lumat LB 9501 luminometer (Berthold Technologies, Bad Wildbad, Germany). The reaction was initiated by injection of 2.2 mm luciferin in 25 mm glycylglycin buffer (pH 7.5), 15 mm MgSO4, 5 mm ATP. In control experiments, stimulation was performed using solutions in which the odorant had been replaced by water. Relative bioluminescence values provided by the luminometer were averaged and expressed as a differential between the sample and its corresponding control. All experiments were performed in triplicate and the results shown in the figures are representative of at least two independent experiments.

Acknowledgements

We thank Dr M.H. Pausch for providing the pJH2-SSTR2 plasmid, and Dr B. Schilling (Givaudan-Roure) for the gift of helional. We also wish to warmly thank D. Grebert for her dedicated and skilful support and Dr L. McCartney for critical comments on the manuscript. This work was financially supported by Institut National de la Recherche Agronomique, the SPOT-NOSED Project of the European Community (IST-2001-38739), the Action Concertée Incitative ‘Olfactory biosensors’ of the French Ministry of Research, and the Ile-de-France region, in the framework of a SESAME contract. J.M. is a postdoctoral fellow supported by a grant within the SPOT-NOSED Project.