A Microdomain Formed by the Extracellular Ends of the Transmembrane Domains Promotes Activation of the G Protein-Coupled α-Factor Receptor

Strains and media.The yeast strains used for the analysis of receptor mutants were yLG123 (MAT a ade2-1^o his4-580^a lys2^o trp1^a tyr1^o leu2 ura3 SUP4-3^ts bar1-1 mfa2::FUS1-lacZ ste2::LEU2), PMY1 (MAT a bar1::hisG far1 ste2Δ mfα1::LEU2 mfα2::his5⁺ ade2 his3 leu2 ura3 mfa2::FUS1-lacZ), and 7416-6-3 (MAT a ade2-1^o his4-580^a lys2 tyr1 cry1 SUP4-3^ts leu2 ura3 ste2::LEU2 sst2-1). Mating tests were performed using lys1α (MATα lys1) as a mating partner. Cells were grown in medium that was prepared as described previously (46). Plasmids were introduced into the yeast using the lithium acetate method, and then the cells were grown in synthetic medium containing adenine and amino acid additives but lacking uracil to select for plasmid maintenance.

Cysteine-scanning mutagenesis.A set of Cys substitution mutants was constructed in pJL147 (YEp, URA3, STE2, and T7-3XHA) (29). This plasmid carries a modified STE2 in which the two endogenous Cys residues at positions 59 and 252 were replaced by other amino acids, and it is also C-terminally tagged with a triple hemagglutinin (HA) epitope (12). This modified STE2 gene displays normal signaling activity and was used as the wild type for this study. Site-directed mutagenesis was carried out using PCR with mutagenic oligonucleotides that were designed to be complementary to the STE2 sequence except for the substitutions required to change the appropriate codons to TGT to code for Cys. All mutations were confirmed by DNA sequence analysis using the Big Dye cycle-sequencing kit (Applied Biosystems Inc.).

MTSEA-biotin reactions.The abilities of the mutant receptors to react with MTSEA-biotin (2-[{biotinoyl} amino] ethyl methanethiosulfonate; Biotium) were analyzed essentially as described previously (29). Logarithmic-phase cells (10⁸) were harvested by centrifugation, lysed by agitation with glass beads in 250 μl of cold phosphate-buffered saline (PBS; 10 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 3 mM KCl, 150 mM NaCl, pH 7.4) containing protease inhibitors (1.5 μM Pepstatin A, 1 mM benzamidine, 0.5 mM phenylmethanesulfonyl fluoride), and then the lysate was cleared by centrifugation at 1,000 × g for 1 min. The membrane fraction was harvested by centrifugation at 15,000 × g for 30 min and then resuspended in 200 μl of PBS and incubated with MTSEA-biotin (0.1 mM final concentration) at room temperature for 2 min. MTSEA-biotin was freshly dissolved in dimethyl sulfoxide at a final concentration of 20 mM. To stop the reactions, freshly prepared Cys was added to a final concentration of 10 mM, and samples were incubated for 5 min. The samples were then extracted in ice-cold RIPA buffer (0.1% sodium dodecyl sulfate [SDS], 1% Triton X-100, 0.5% deoxycholic acid, 1× PBS, pH 7.4, 1 mM EDTA), and then the biotinylated proteins were harvested using UltraLink Immobilized Streptavidin Plus beads (Pierce, Rockford, Ill.). The beads were washed four times in RIPA buffer and one time in buffer containing 2% SDS to reduce nonspecific sticking of Ste2p. During the washes, the beads were recovered by allowing them to settle by gravity for 20 min. Bound proteins were eluted using gel sample buffer (8 M urea, 50 mM Tris, pH 6.8, 2% SDS, 10% 2-mercaptoethanol). The extent of reaction with MTSEA-biotin was determined by comparing the amount of Ste2p present in the total membrane fraction to the amount of Ste2p eluted off the streptavidin beads by Western blot analysis as described below. As a control, total membrane protein biotinylation was examined by adding samples of total extract to gel sample buffer lacking 2-mercaptoethanol and then analyzing them on blots probed with streptavidin-horseradish peroxidase (Amersham Pharmacia Biotech). The bands were detected by chemiluminescence using 5-amino-2,3-dihydro-1,4-phthalazinedione (Sigma) as a substrate. Similar analysis of the streptavidin bead supernatant showed that there was essentially complete recovery of the biotinylated proteins on the beads.

Western blot analysis of α-factor receptors.Western blotting was carried essentially out as described previously (23). Samples were separated by electrophoresis on SDS-10% polyacrylamide gels and then electrophoretically transferred to Hybond-P membrane (Amersham Pharmacia Biotech). The HA-tagged receptor proteins were detected on the blots by probing them with the anti-HA antibody 12CA5 (Roche Molecular Biochemicals). The protein bands were detected using alkaline phosphatase-conjugated goat anti-mouse immunoglobulin G (Zymed) secondary antibody and an AttoPhos AP fluorescent-substrate system (Promega). Quantitative analysis of the signals was carried out using ImageQuant computer software. This analysis included both the monomer and the dimer forms of Ste2p that are typically seen on Western blots (29). To facilitate quantitative analysis, four different dilutions of the total membrane fraction were compared to three dilutions of the sample eluted off the streptavidin beads. For each set of mutant receptors analyzed, wild-type and STE2-T199C cells were analyzed in parallel with the other mutants as the negative and positive controls, respectively. The accessibility of each mutant receptor was then calculated as the degree of biotinylation relative to T199C receptors.

Ligand-induced responses.Mating assays were carried out by mixing yeast strain yLG123 (MAT a ste2Δ) containing the indicated wild-type or Cys-substituted mutant version of STE2 on a plasmid with MATα strain lys1α cells and then selecting for the growth of diploid cells on minimal medium. Halo assays for α-factor-induced cell division arrest were performed by spreading 6 × 10⁵ yeast cells on solid-medium agar plates and then placing sterile filter disks containing the appropriate amount of α-factor on the lawn of cells. The halo assays were performed with strain yLG123 carrying the appropriate wild-type or Cys mutant plasmid. A similar approach was used to perform halo assays for cell division arrest induced by novobiocin (Sigma), except that the plasmids were carried in strain 7416-6-3. The formal name for novobiocin is N-(7-((3-O-(aminocarbonyl)-6-deoxy-5-C-methyl-4-O-methyl-β-l-lyxo-hexopyranosyl)oxy)-4-hydroxy-8-meth-yl-2-oxo-2H-1-benzopyran-3-yl)-4-hydroxy-3-(3-methyl-2-butenyl) benzamide.The plates were incubated at 30°C for 48 h, and then the diameters of the zones of ligand-induced cell division arrest were measured. FUS1-lacZ induction assays were carried out by growing PMY1 cells carrying the appropriate receptor plasmid to log phase, adjusting the culture to 3 × 10⁶ cells/ml, and then adding the appropriate concentrations of α-factor (Bachem) for 2 h. The cells were permeabilized with 0.05% SDS and CHCl₃, and then β-galactosidase activity was quantified using o-nitrophenyl-β-d-galactopyranoside (Sigma) as a substrate (34). The results represent the average of two or three independent experiments, each done in duplicate.

α-Factor binding assays.The abilities of the mutant receptors to bind α-factor were assayed essentially as described previously (43). Cells were grown to logarithmic phase, collected by centrifugation, washed twice with ice-cold inhibitor medium (yeast extract-peptone-dextrose medium containing 10 mM KF and 10 mM NaN₃), and then resuspended at a density of either 2 × 10⁸ or 1 × 10⁹ cells/ml. Binding assays were initiated by mixing 50 μl of cells with 50 μl of ³⁵S-α-factor. Samples were incubated for 30 min, and then the cells were collected on a Whatman GF/C filter and the unbound α-factor was removed by washing. Nonspecific binding was determined by performing reactions in the presence of a 100-fold excess of cold α-factor. The bound radioactivity was determined by scintillation counting. Binding data presented in this study represent the averages of three independent assays, each done in duplicate. ³⁵S-labeled α-factor was purified from the supernatant of MATα cells labeled with [³⁵S]methionine (Translabel; ICN, Irvine, Calif.) by chromatography on a Bio-Rex 70 (Bio-Rad; Hercules, Calif.) column as described previously (43).

Cysteine-scanning mutagenesis across extracellular ends of TMD1, -2, -3, -4, and -7.To analyze the functions of residues near the extracellular ends of the TMDs of Ste2p, scanning mutagenesis was carried out to construct 46 new Cys substitution mutants (Fig. 1). Residues near the predicted extracellular ends of TMD1 (Leu⁴⁴ to Ile⁵³), TMD2 (Lys¹⁰⁰ to Ser¹⁰⁷), TMD3 (His¹²⁶ to Glu¹³⁵), TMD4 (Thr¹⁷⁹ to Ile¹⁹⁰), and TMD7 (Val²⁷⁶ to Ala²⁸¹) were mutated. These mutations were designed to complement the 27 previously constructed Cys substitution mutants near the ends of TMD5 and -6 (29) to provide a mutational analysis of the ends of all seven TMDs (Fig. 1). The corresponding residues were replaced with Cys instead of the more commonly used Ala to take advantage of the unique chemical properties of the thiol group in the Cys side chain for the analysis of Ste2p structure. The mutations were therefore created in a version of the α-factor receptor gene (STE2) (12) in which the endogenous Cys codons at positions 59 and 252 were mutated to other residues so that the substituted Cys provided a unique thiol group. This version of STE2 also has a triple HA epitope tag fused to the C terminus to facilitate analysis of the α-factor receptor protein (Ste2p). Plasmids containing the Cys mutants were then introduced into a yeast strain lacking the chromosomal copy of the receptor gene (ste2Δ strain yLG123) for analysis. Western immunoblots showed that the Cys substitution mutant receptors were produced at a level that was similar to the that of wild-type control, Ste2p (data not shown).

• Open in new tab
• Download powerpoint

FIG. 1.

Residues targeted for Cys substitution mutagenesis near the ends of TMDs in Ste2p. Residues with Cys substituted as part of this study are boxed with solid lines. The residues boxed with a dashed line were mutated to Cys in a previous study (29). The shading indicates Cys substitution mutations that caused a significant decrease in receptor function. Red indicates residues affected by dominant-negative mutations (9, 25, 55). This topographical representation of Ste2p was the working model for the TMDs used at the start of these studies.

Abilities of Cys mutants to respond to α-factor.All 46 of the new receptor mutant strains were able to conjugate with a MATα strain, indicating that they retained at least partial receptor activity. However, eight of the new mutants displayed reduced signaling capacity in halo assays for α-factor-induced cell division arrest (Fig. 2). Halo assays are a more sensitive test of receptor function that assesses the abilities of the mutants to maintain a high level of pheromone response for 2 days. These assays measure the zone of cell division arrest (halo) surrounding a disk containing α-factor placed on a lawn of cells (Fig. 2). The T50C and L277C mutants produced smaller halos of division arrest, indicating decreased receptor signaling activity. Mutations at six other positions (45, 46, 52, 102, 105, and 281) caused the formation of filled-in halos or a failure to form a detectable halo. Previous studies indicate that this phenotype can be due either to a decrease in the ability of the receptors to transduce a signal or to a decrease in the number of cell surface receptors.

• Open in new tab
• Download powerpoint

FIG. 2.

Cys substitution mutants defective in undergoing cell division arrest in response to α-factor. Yeast strain yLG123 carrying the wild-type receptor plasmid or versions containing the indicated Cys substitution mutants were assayed for the ability to undergo cell division arrest in response to α-factor. Cell division arrest leads to the formation of a zone of growth inhibition (halo) surrounding a filter disk containing α-factor (200 or 1,200 ng) applied to a lawn of cells on the surface of an agar plate.

To help distinguish between these possibilities, the mutants were assayed for the ability to induce the pheromone-responsive FUS1-lacZ reporter gene and for the ability to bind radiolabeled α-factor. Analysis of FUS1-lacZ induction after a 2-h treatment with α-factor is less sensitive to changes in the levels of cell surface receptors than are halo assays. Similar FUS1-lacZ responses can be detected in cells with receptor levels that vary over a broad range (7, 45).

The T50C and L277C mutants that formed slightly smaller halos activated the pheromone-responsive FUS1-lacZ reporter gene to the same maximal level as the wild type (Fig. 3B). Equilibrium binding assays carried out using ³⁵S-α-factor showed that these mutant receptors displayed ∼2-fold lower affinity for α-factor than the wild-type control (Table 1). The number of receptor binding sites present on the surfaces of these cells was about the same or greater than the number of receptor sites on wild-type cells. This indicates that the signaling defects caused by the T50C and L277C substitutions are due to decreased affinity for α-factor.

• Open in new tab
• Download powerpoint

FIG. 3.

Basal and α-factor-induced levels of the pheromone-responsive FUS1-lacZ reporter gene. Shown are PMY1 cells carrying either wild-type STE2 or the indicated Cys substitution mutants on a plasmid in the absence (A) or the presence (B) of 10⁻⁶ M α-factor for 2 h. FUS1-lacZ reporter gene activity was determined by assaying β-galactosidase activity as described in Materials and Methods, and the values were normalized to the basal or maximal induced value of the wild-type cells as indicated. The results represent the averages of two to four independent experiments, each done in duplicate. The error bars represent standard deviations. A t test indicated significant differences between the basal levels of the wild type and A281C (P < 0.0001) and between the induced levels of the wild type and A52C (P < 0.005), L102C (P < 0.005), N105C (P < 0.001), and A281C (P < 0.01).

Two of the mutants that gave filled-in halos (V45C and N46C) induced FUS1-lacZ to maximal levels (Fig. 3B). Previous studies have shown that mutants with similar properties—maximal signaling in a short 2-h FUS1-lacZ induction but failure to maintain cell division arrest in a 2-day halo assay—are typically defective in receptor trafficking to the plasma membrane and not in signal transduction per se (13, 19). Consistent with this, the V45C and N46C mutants formed clear halos similar to those of cells carrying the wild-type receptor plasmid when assayed in an stp22Δ strain (data not shown) that was shown previously to increase the number of mutant Ste2-3p receptors at the cell surface by preventing their trafficking to the vacuole (28). Thus, the V45C and N46C mutants were not studied further in order to focus on mutants with clear defects in receptor signaling.

The A52C, L102C, and N105C mutants also gave filled-in halos but were unable to activate FUS1-lacZ to maximal levels (Fig. 3B). A somewhat stronger signaling defect for the L102C and N105C mutations was reported as part of a previous scanning mutagenesis of residues in extracellular loop 1 (1). The differences are likely due to the variations in experimental strategies, given that different plasmids, strains, and epitope tags were used. The A52C and N105C mutants produced receptors with essentially the same affinity as the wild type, and the L102C receptors showed a slight decrease in affinity (K_d = 12.1 nM) (Table 1). These mutants also displayed fewer cell surface receptors than wild-type control cells, but the decreased number of receptors did not appear to be so great as to account for all of their signaling defects. Some mutants with as little as 2% of the wild-type level of receptors can still induce maximal levels in a short-term FUS1-lacZ assay (13, 45). Furthermore, since the STE2 genes in this study were introduced into cells on high-copy-number plasmid vectors, the levels of receptors are comparable to those of cells with a single copy of STE2 in the genome (∼5,000 to 10,000 receptors per cell) (10). Thus, these mutants appear to be defective in the abilities of ligand-activated receptors to promote G protein signaling.

The A281C mutant which did not form a detectable halo of division arrest could induce FUS1-lacZ to only 83% of the maximum level (Fig. 3B). Interestingly, the A281C mutant displayed a 4.7-fold-elevated level of basal FUS1-lacZ, indicating that it signals constitutively in a ligand-independent manner (Fig. 3A). The A281C mutant receptors showed a small decrease in affinity for α-factor (1.8-fold) and a 24-fold decrease in cell surface receptors relative to the wild type. This decreased number of cell surface receptors is typical of other constitutive receptors, perhaps because their ability to mimic the ligand activated state causes altered membrane trafficking or an increase in the rate of endocytosis, as seen for ligand-bound receptors (13, 24, 49). The A281C mutant is distinct from most of the previously studied constitutive forms of Ste2p that are still capable of being induced to maximal levels.

For comparison, α-factor binding activity was analyzed for two Cys substitution mutants constructed previously near TMD5 that caused strong defects in signaling. The F204C mutant receptors showed no detectable binding activity (Table 1). This was not due to a trafficking defect, since the F204C mutant receptors could be detected by Western blot analysis of purified plasma membrane fraction (data not shown). The F204C receptors also retain function at the plasma membrane, as they are able to dominantly interfere with the activities of wild-type receptors (29) and are also able to respond to an alternative agonist (see below). The N205C receptors displayed detectable binding activity, but the affinity was too low to be measured accurately (K_d > 30 nM). The strong ligand binding defects of these mutants contrast with that of a previously studied mutant near the end of TMD6 (Y266C) that was also strongly defective in signaling yet showed only a slight (∼2.5-fold) decrease in affinity (9).

The scanning mutagenesis results were also interesting in that none of the substitutions near the ends of TMD3 and -4 caused obvious phenotypes. This was unexpected, since several mutations near the extracellular ends of TMD3 and TMD4 were isolated previously as dominant-negative mutants (9, 25, 55) (Fig. 1). However, these mutations resulted in very nonconservative amino acid substitutions (see Discussion). Altogether, the scanning mutagenesis of all seven TMDs suggests that residues near the ends of TMD1, -2, -5, -6, and -7 are important for proper receptor function.

Responses of mutants to alternative agonist (novobiocin).The mutant receptors were examined further by testing their responses to novobiocin, a small nonpeptide molecule that can stimulate Ste2p. Novobiocin was identified as an inhibitor of bacterial DNA gyrase and was subsequently found to act as an agonist for the α-factor receptor (39). Cells respond weakly to novobiocin, so these assays were carried out in a strain that is hypersensitive to pheromone due to mutation of the RGS protein (Sst2p) that negatively regulates G protein signaling. Interestingly, two mutants with slight defects (A50C and L277C) and two mutants with strong defects (F204C and N205C) in binding α-factor responded to novobiocin as well as the wild type, as judged by their abilities to form a halo of cell division arrest in response to novobiocin (Table 2). In fact, L277C mutants appeared to be slightly supersensitive to novobiocin. These results demonstrate that the A50C, F204C, N205C, and L277C mutants are specifically defective in responding to α-factor. The other mutants were defective in responding to both novobiocin and α-factor (Table 2). This is consistent with their being generally defective in receptor activation but could also be due to their being defective in binding novobiocin.

Accessibility of Cys residues to MTSEA-biotin.To better understand how the ends of TMDs in Ste2p contribute to receptor function, the Cys-substituted residues were investigated for accessibility to the extracellular environment. This was carried out by assaying the Cys residues in the mutant receptors for the ability to react with MTSEA-biotin. MTSEA is a thiol-specific agent that does not react with membrane-embedded Cys residues (21). The biotin group attached to MTSEA was used to facilitate analysis of the reactivities of the Cys residues. Accessibility assays were carried out by treating membrane fractions with MTSEA-biotin, solubilizing them in detergent buffer, collecting the biotin-labeled receptors on streptavidin beads, and then quantifying the degree of receptor biotinylation on Western immunoblots as described in Materials and Methods. Among the advantages of this approach is that a wide range of Cys mutants could be screened in their native membrane environment without prior purification procedures that may alter receptor structure or membrane integrity. In each set of assays, the reactivities of the Cys substitution mutants were compared to that of a positive control mutant containing an accessible Cys residue near the end of TMD5 (T199C), and as a negative control, they were compared to a wild-type receptor lacking Cys residues.

Mutants containing Cys residues near the ends of TMD3, -4, and -7 showed a clear trend of higher reactivity to MTSEA-biotin for Cys residues positioned toward the predicted extracellular side and decreased reactivity for residues in the predicted membrane domain (Fig. 4). The TMD3 mutants showed a drop in reactivity with MTSEA-biotin after position 129 for three residues (positions 130 to 132) and then trailed off for the next three residues that were tested. The TMD4 mutants showed a drop in reactivity with MTSEA-biotin for four residues (positions 188 to 185) and then trailed off over the next seven residues. For TMD7, only the subset of residues in the predicted membrane domain were analyzed in this study because the residues on the extracellular side that comprise the third extracellular loop were analyzed previously (29) and found to be very reactive with MTSEA-biotin through position 275 (Fig. 5). The results for TMD7 show that reactivity with MTSEA-biotin was decreased for three residues (positions 276 to 278) and then trailed off for the next three residues. This pattern of decreased reactivity as residues are positioned more closely to the center of the predicted transmembrane segment is similar to what was observed previously for Cys mutants near TMD5 and -6 (29).

• Open in new tab
• Download powerpoint

FIG. 4.

Reactivities of Cys residues with MTSEA-biotin. Accessibility assays were carried out by treating membrane fractions from the indicated Cys substitution mutants with MTSEA-biotin. The proteins were solubilized in detergent buffer, and then the biotin-labeled receptors were collected on streptavidin beads. The degree of receptor biotinylation was quantified on Western immunoblots and is reported as the percentage of labeling relative to the T199C receptors. In each experiment, control samples confirmed that the wild-type receptors lacking Cys residues did not react significantly and that the T199C receptors, which contain a Cys in the second extracellular loop, reacted well with MTSEA-biotin (38%). This level of biotinylation is likely to be an underestimate, since stringent washing conditions were used to prevent nonspecific sticking to the streptavidin beads. The results represent the averages of at least three independent experiments plus standard errors. Analysis of variance carried out using the ANOVA program (GraphPad Software, San Diego, Calif.) indicated that the means differed significantly (P < 0.01) and that there was a significant trend of decreasing accessibility for TMD1, -3, -4, and -7 (P < 0.001).

• Open in new tab
• Download powerpoint

FIG. 5.

Surface accessibility maps of the α-factor receptor. (A) Reactivities of Cys residues substituted into the α-factor receptor are color-coded as indicated. This topology map was adjusted to place the membrane boundary at the start of the drop in reactivity to MTSEA-biotin. The dotted lines indicate the positions of previous predictions for the TMD boundaries, as shown in Fig. 1. (B) Comparison of topology predictions by the indicated computer programs. The bars represent the ends of the TMDs as predicted by the programs that were accessed at the ExPASy molecular biology server (http://ca.expasy.org/ ). Sites of Cys substitutions that altered the binding affinity are circled in blue, as are positions 47 and 48, which were implicated previously in binding α-factor using different approaches (27). Shaded in red are the sites of Cys substitutions that prevented maximal signaling activity in this study and also previously studied mutations at positions 102, 105, 108, and 111 (1). Residues affected by dominant-negative mutations are shaded gray (9, 25, 55).

The pattern for TMD1 appeared more complicated in that Cys residues at positions 45, 46, and 47 showed slightly lower reactivities than expected. While this could represent a special structural feature of TMD1, it is also possible that it may be due in part to the adverse effects of some Cys substitutions on Ste2p structure. The V45C and N46C receptors displayed defects in signaling (Fig. 2 and 3) and increased aggregation on protein gels (data not shown). In spite of this, the observation that reactivity was intermediate at positions 49 to 51 and then trailed off to low levels for the next two residues suggests that this represents the boundary of TMD1.

Cys residues near the end of TMD2 gave an unexpected pattern in that they displayed a prolonged zone of primarily intermediate accessibility (positions 101 to 107). Reactivity was very low for position 100 and 103. This unusual pattern of accessibility is interesting because it has been proposed that this region forms a special compact structure with sequences from extracellular loop 1 (1). The addition of 1 M NaCl to the assay to attempt to disrupt the structure did not improve the accessibility of the Cys residues in this region to MTSEA-biotin (data not shown). Consistent with this region forming a special structure that is important for receptor function, the L102C and N105C mutations that affect residues in this region caused significant defects in response to α-factor (Fig. 2) (1).

Accessibility map of Ste2p.To examine the overall implications of the accessibility data for receptor structure, the reactivities of the various Cys mutants were summarized by color coding residues in Fig. 5A. Data from the previous analysis of TMD5 and -6 are included to provide a comparison of all seven TMDs. A clear pattern emerged in which the most efficiently biotinylated Cys residues mapped to the predicted extracellular domains. For TMD3 to -7, a steep drop in accessibility to MTSEA-biotin generally occurred over two to four residues. This transition most likely corresponds to the membrane interface region. Although the accessibility of any particular residue can be influenced by a number of different factors, the pattern of decreasing accessibility over a series of residues is expected for the extracellular end of a TMD. The topology map of Ste2p was therefore adjusted in Fig. 5A to place the ends of the TMDs at the start of the steep drop in reactivity with MTSEA-biotin. The new topology based on experimental data had significant effects on the positioning of TMD2, -3, and -4 relative to our previous working model for Ste2p topology.

The new topology map was then compared with TMD predictions obtained from four different computer programs available at the ExPASy molecular biology server (http://www.expasy.org ) (Fig. 5B). The predictions of three other available programs (HMM Top, TMpred, and TMAP) are not shown because they varied widely and in some cases did not predict all seven TMDs. The TMHMM (33) and SOSUI (17) programs most closely matched the boundaries inferred from the accessibility data. TMHMM was reported to be the best among a group of programs whose abilities to predict the TMDs in a set of membrane proteins for which topology information is available were examined (35). TMHMM and SOSUI were also the best programs at predicting the seven TMDs in a large group of GPCR sequences (35). However, it was interesting that the predictions by TMHMM and SOSUI for the end of TMD4 differed by six residues and that the accessibility boundary fell in between them. In addition, both programs were shifted by three or four residues relative to the accessibility boundary in predicting the end of TMD5. Thus, the experimental data on the accessibility boundaries of the TMDs are expected to significantly improve the interpretation of the functional roles of residues identified by genetic analysis (see Discussion). In addition, these data are expected to improve three-dimensional models of Ste2p, since placing the TMDs in the proper register is an important requirement for developing a good molecular model (51).

Accessibility boundaries.To gain insight into how the ends of the TMDs contribute to receptor function, Ste2p topology was investigated by assaying Cys residues in the mutant receptors for the ability to react with MTSEA-biotin. MTSEA is a thiol-specific agent that does not react with membrane-embedded Cys residues (21). Various MTSEA derivatives have been used to identify extracellular loops in polytopic membrane proteins (6), to investigate effects on ligand binding caused by modifying Cys residues (18, 47), and to define an accessible pore within a helix bundle (17). The goal of this study was to identify which residues are accessible to the extracellular environment and are likely to be able to contact ligands.

Cys residues substituted near the ends of the TMDs in Ste2p showed an obvious drop in reactivity with MTSEA-biotin that typically occurred over several consecutive residues. Residues that failed to react efficiently with MTSEA-biotin could be either embedded in the membrane or sterically blocked by other receptor domains. Although it is difficult to distinguish between these possibilities by the analysis of a single residue, the observed decrease in reactivity over a series of residues is most consistent with expectations for the membrane boundary of an α-helical transmembrane domain. The accessibility boundaries defined by this analysis are likely to reflect the properties of the endogenous residues, since most Cys substitution mutants displayed wild-type properties.

Cys mutants near TMD2 showed an unusual zone of mainly intermediate accessibility (positions 101 to 107) that could represent a special structure for this region. Interestingly, Akal-Strader et al. proposed that the region forms a compact helical structure based on mutational and biochemical analysis of the first extracellular loop of Ste2p (1). This region also plays a special role in receptor activation, since Cys substitutions at positions 102, 105, 108, and 111 caused decreased signaling (this study and reference 1). Furthermore, other mutations near the end of TMD2 cause constitutive activity (38, 48) or a loss-of-function phenotype (9). The cluster of mutations in this area suggests that extracellular loop 1 may form a compact structure that is in close proximity to the ends of the TMDs. Extracellular loop 1 of Ste2p may therefore function in a manner analogous to extracellular loop 2 of rhodopsin that folds back into the central region of the helix bundle (1, 37).

α-Factor receptor activation.A comparison of accessibility data and mutant phenotypes suggests that different regions near the ends of the TMDs carry out distinct functions (Fig. 5B). Residues that primarily influence ligand binding affinity were situated in zones of high (Phe²⁰⁴ and Asn²⁰⁵) or intermediate (Thr⁵⁰ and Leu²⁷⁷) accessibility to MTSEA-biotin. Two other residues near TMD1 (Ser⁴⁷ and Thr⁴⁸) that were previously implicated in ligand specificity by other approaches (27) also map to the accessible zone. In contrast, residues implicated in receptor activation were found deeper in the intermediate zones (Leu¹⁰² and Asn¹⁰⁵) or inaccessible regions (Ala⁵², Tyr²⁶⁶, and Ala²⁸¹). Residues at positions 108 and 111 have also been implicated in receptor activation but appear to be out in extracellular loop 1 (1). However, as described above, this region may fold into a compact structure near the ends of the TMDs. Altogether, this pattern suggests that accessible residues play key roles in binding α-factor and that more deeply buried residues may mediate the transition to the activated state.

To visualize the potential relationship of the implicated residues in three dimensions, they were mapped onto the ends of the TMDs in the high-resolution structure of rhodopsin (Fig. 6). Although rhodopsin and Ste2p do not show significant sequence homology, the accessibility data enabled the ends of their TMDs to be aligned relative to the membrane interface. As shown in Fig. 6, residues implicated in binding and activation were clustered near the ends of TMD1, -2, -5, -6, and -7. Interestingly, these residues all reside near regions where α-factor has been predicted to dock with Ste2p (16, 29). Furthermore, clustering of the residues affected by the Cys substitutions on TMD1, -2, and -7 may identify a network important for receptor activation. These results suggest that the ends of TMD1, -2, -5, -6, and -7 form a microdomain that promotes ligand binding and receptor activation.

• Open in new tab
• Download powerpoint

FIG. 6.

Model for spatial relationshipsof residues implicated in ligand binding and activation of Ste2p.(A) Three-dimensional model of the extracellular ends of the TMDs based on the crystal structure of rhodopsin (7). The MTSEA-biotin accessibility data were used to map Ste2p residues near the ends of the TMDs onto the corresponding residues in rhodopsin. The side chains colored blue are implicated in binding α-factor; those colored red are implicated in receptor activation. The residues include those described in this study and previous studies, as explained in the legend to Fig. 5 (7 , 27, 29). (B) Schematic representation of TMD ends with Ste2p residue numbers labeled. A prediction for the orientation of α-factor in the binding domain is shown as a dashed green line. This prediction for how α-factor docks with the receptor was inferred from functional data and cross-linking experiments that indicate that Phe²⁰⁴ may interact with the C terminus of α-factor and Tyr²⁶⁶ may interact with the N terminus (9). Since α-factor contains a β-bend structure around a central Pro-Gly bond (7), it was proposed to dock onto a model of the receptor as viewed from the extracellular side of the plasma membrane.

This model also suggests a possible explanation for why the ends of TMD3 and TMD4 did not appear to be critical for receptor signaling. TMD4 is in the least contact with other TMDs in rhodopsin, and it is not predicted to be adjacent to α-factor in the model. The lack of phenotypes for Cys substitutions near the end of TMD3 was more unexpected, since this TMD is thought to play a key role in the activation of Ste2p and other GPCRs (36). However, TMD3 is tilted so that the extracellular end moves away from the central helix bundle in rhodopsin (37). Thus, if residues near the extracellular end of TMD3 in Ste2p behave in a similar manner, they are not expected to be in a close context with α-factor. In agreement with this, the known mutations near the ends of TMD3 and -4 that strongly affect function result in substitutions that radically change the character of the side chain (i.e., N132I, N132Y, Q135P, M180R, S184R, and A185P). In contrast, Cys is usually tolerated well at many positions (12, 15, 21). Therefore, dramatic substitutions, rather than the loss of the wild-type side chain, appear to be needed to alter the receptor function of TMD3 and -4.

The new topology map of Ste2p based on the MTSEA-biotin accessibility data also provides an important framework for the interpretation of the functions of residues in other regions of the receptor. For example, Gln¹⁴⁹, a residue that causes constitutive activity when mutated, now maps closer to the cytoplasmic side of TMD3. This new positioning further strengthens the suggestion that Gln¹⁴⁹ contributes to receptor activation in a manner analogous to the E/DRY motif found at the cytoplasmic end of TMD3 in many mammalian GPCRs (38). These data also further implicate the intracellular ends of the TMDs in functioning in an analogous manner as a microdomain that mediates G protein activation.

Analysis of TMD boundaries could also provide similar insight into other polytopic membrane proteins, including transporters and channels. Although many proteins contain essential Cys residues, protease digestion strategies could be used similarly to the way that they have facilitated the analysis of intramolecular disulfide bond formation (12, 50, 56). Furthermore, the results for Ste2p show that this analysis can be very helpful even if it is targeted only to the subset of TMDs that are most difficult to predict.

General features of GPCR activation.A microdomain formed by the ends of the TMDs is also likely to promote activation of other GPCRs, particularly those that bind larger peptide or polypeptide ligands. These ligands may have a correspondingly larger domain of interaction with their receptors than small ligands, such as catecholamines or the retinal group, that bind in a pocket formed by the helix bundle (20). Analysis of α-factor binding suggests that this microdomain may include residues that face away from the helix bundle, as well as residues that face the interior. Interestingly, residues near the ends of the TMDs have been implicated in binding other peptide ligands (3, 20, 31). Although in some cases the extracellular loops have also been implicated, they may fold in such a manner that they contact the ends of the TMDs, as occurs in family A GPCRs (the rhodopsin family), where extracellular loop 2 is connected via a disulfide bond to the extracellular end of TMD3. The binding sites of somewhat larger polypeptide ligands, such as chemokines, include the extracellular loops, but the N terminus of the chemokine must interact with the receptor helix bundle to trigger receptor activation (4). Similarly, although many class B family GPCRs, such as the glucagon receptor, bind ligands via a large N-terminal receptor domain, it appears that ligand interaction with the TMD helix bundle is required for receptor activation (42).

A general advantage of a ligand binding domain that includes the ends of several TMDs is that it could be an efficient mechanism to stabilize the active receptor conformation. Ligand binding to the ends of the TMDs could directly affect the overall structure of the receptor helix bundle instead of transmitting the effects of ligand binding indirectly through the loops. Thus, the effects of ligand binding could be relayed directly to the intracellular ends of the TMDs to promote G protein activation. Conversely, inhibitors that bind near this microdomain can be used to effectively block receptor signaling. In this regard, it is interesting that two small-molecule inhibitors that prevent human immunodeficiency virus type 1 infection by blocking the function of either the CXCR4 or CXCR5 coreceptor, members of the GPCR family, bind to residues near the extracellular ends of the TMDs (11, 41).