Polar residues are present in TM (transmembrane) helices and may influence the folding or association of membrane proteins. In the present study, we use an in vivo approach to analyse the functional and structural roles for amino acids in membrane-spanning motifs using the Rot1 (reversal of Tor2 lethality 1) protein as a model. Rot1 is an essential membrane protein in Saccharomyces cerevisiae and it contains a single TM domain. An alanine insertion scanning analysis of this TM helix revealed that the integrity of the central domain is essential for protein function. We identified a critical serine residue inside the helix that plays an essential role in maintaining cell viability in S. cerevisiae. Replacement of the serine residue at position 250 with a broad variety of amino acids did not affect protein targeting and location, but completely disrupted protein function causing cell death. Interestingly, substitution of the serine residue by threonine resulted in sustained cell viability, demonstrating that the hydroxy group of the TM serine side chain plays a critical role in protein function. The results of the present study indicate that Rot1 needs the TM Ser250 to interact with other membrane components and exert its functional role, avoiding exposure of the serine hydrogen-bonding group at the lipid-exposed surface.
Membrane proteins represent more than one-fourth of all genome-encoded proteins  and play an important role in cell biology and physiology, carrying out many essential cell functions. As membrane protein functions have to be tightly co-ordinated to ensure the survival of every cell and every organism, mutations that cause membrane protein misfolding or impair interactions with other membrane components may compromise cell survival. Most membrane proteins span the membrane with an α-helical conformation of the polypeptide chain and are often organized into functional complexes, forming homo-oligomeric and/or hetero-oligomeric assemblies . In these proteins, TM (transmembrane) α-helices are characterized by long stretches of hydrophobic residues with few or no polar side chains. Rather than serving merely as featureless hydrophobic stretches required to anchor proteins to membranes, these TM domains have structural and/or functional roles well beyond this canonical capacity, such as facilitating interactions with other membrane proteins. The driving forces that stabilize these interactions are generally packing effects (van der Waals interactions) and interhelical polar interactions (hydrogen bonds and ion pairs)  between closely packed helices.
Considering the individual TM α-helices as independently folded units allows the study of membrane protein assembly. Micelles and lipid bilayers have been used together with synthetic peptides and chimaeric proteins containing designed TM helices to study the influence of different residues or motifs in the structure and function of these TM domains [4,5]. These studies revealed that polar amino acids in the TM helices play critical roles in the association and folding of membrane proteins through the formation of side chain–side chain interhelical hydrogen bonds within the membrane environment . Sequence analysis of high-resolution membrane protein structures has shown that serine and threonine residues are found in TM helices more frequently than other polar residues (i.e. asparagine, glutamine, histidine, aspartic acid, glutamine, arginine and lysine) , partially because of their potential to form hydrogen bonds with main-chain carbonyl oxygens . Such hydrogen bonding may have important consequences for the function, conformational specificity and thermodynamic stability of membrane proteins. Using randomized libraries of TM interfaces, it has been found that serine/threonine motifs exhibit an ability to drive the association of TM helices, stabilizing the formation of helical oligomers [9,10]. Recently, the presence of serine/threonine motifs in TM helices has been shown in MD simulations and bioinformatic analyses to have significant implications on local membrane protein structure and dynamics .
Despite the information obtained using in vitro approaches in structure-based analyses of TM helices, studies in complex living cells, such as eukaryotic cells, are sparse. In the present study, we analysed the amino acid sequence requirements of the single C-terminal hydrophobic domain of the yeast Rot1 (reversal of Tor2 lethality 1) protein in order to identify the amino acids responsible for TM helix function using an in vivo model system. Rot1 is an essential protein in the budding yeast Saccharomyces cerevisiae. This protein is involved in several cellular functions, including protein folding, cell wall biosynthesis, actin cytoskeleton dynamics and cell-cycle control . Rot1 is located primarily at the ER (endoplasmic reticulum)–nuclear membrane, facing the lumen, and is anchored to the membrane by a single C-terminal hydrophobic region . In the present study, we analysed the role of the residues included in this hydrophobic region. A systematic mutagenesis study of the TM domain was performed. Although a single polar serine residue at position 250 appears to be critical for cell viability, other polar residues, such as Thr234, Asn236, Ser235 and Ser242, do not affect cell survival. Therefore Ser250 is critical for proper protein function and crucial for cell viability.
MATERIALS AND METHODS
Yeast strains, genetic methods and plasmids
To obtain the yeast strains used in the present study, the ROT1 promoter was substituted with the tetO7 promoter in the strain JCY216 by integrating a DNA fragment amplified from the plasmid pCM225 . A centromeric plasmid containing ROT1 was constructed by cloning a PCR-amplified fragment from positions −430 to +955 of the ROT1 gene into an EcoRI-BamHI-cleaved pRS314 vector. Construction of the plasmid pROT1-HA-CEN, which contains a sequence encoding three copies of the HA (haemagglutinin) epitope, was described previously . A PCR-amplified fragment (using the pROT1-HA-CEN plasmid as a template) from positions −430 to +684 of the ROT1 gene was cloned into EcoRI-XhoI-cleaved pRS314 to obtain the plasmid pROT1-HAΔ229–256-CEN. In contrast, pROT1-HAΔ229–256–18L was constructed in two steps. First, a PCR-amplified fragment from positions −430 to +684 of the ROT1 gene was cloned into the YCplac50 vector. Secondly, a DNA fragment covering positions +722 to +955 of the ROT1 gene, which was obtained by PCR amplification using a forward primer containing the sequence coding for the polyleucine tail (18L), was cloned in-frame in the previous plasmid.
The plasmid encoding the TM domain sequence of ROT1 was constructed by subcloning the TM domain into the pGEM1 vector between the KpnI and SpeI restriction sites using the PCR-amplified ROT1 sequence containing positions +709 to +768 and primers containing the appropriate restriction sites. The PCR product was digested and ligated to the Lep (leader peptidase) vector to obtain a construct containing the P2 domain of the Escherichia coli Lep fused in-frame at the C-terminus.
The TM domain corresponding to residues 237–256 of Rot1 were PCR-amplified using primers containing positions +709 to +768 and digested with the BamHI and NheI restriction enzymes. The fragment was cloned into the BamHI-NheI sites of the pccKAN vector to obtain the construct in-frame with the MBP (maltose-binding protein) domain at the N-terminus and ToxR at the C-terminus.
Cells were grown in standard YPD medium [1% (w/v) yeast extract/2% (w/v) peptone/2% (w/v) glucose] or SD (synthetic dextrose) medium [0.67% yeast nitrogen base, synthetic complete mixture (drop out) and 2% glucose] and supplemented as required. To fully repress the tetO7 promoter, doxycycline (Sigma) was added to a final concentration of 10 μg/ml. For growth assays, 10-fold serial dilutions in growth medium were prepared from an exponentially growing culture (5×106 cells/ml). A total of 5 μl of each dilution was spotted on to the appropriate medium and the plates were incubated at 25°C.
Protein expression in vitro
Wild-type and ROT1 constructs were transcribed and translated in the TNT® SP6 Quick Coupled system (Promega). DNA template (75 ng), [35S]methionine/cysteine (5 μCi) and 1 μl of dog pancreas RMs (rough microsomal membranes; tRNA Probes) were added to 5 μl of lysate and the samples were incubated for 90 min at 30°C. The translation reaction mixture was diluted in 5 volumes of PBS (pH 7.4). The membranes were collected by layering the supernatant on to a 50 μl sucrose cushion and centrifugation at 100000 g for 20 min at 4°C in a Beckman table top ultracentrifuge with a TLA-55 rotor. The pellets were analysed by SDS/PAGE (12% gel) and the gels visualized on a Fuji FLA3000 Phosphor Imager using the ImageGauge.
Western blot analysis
Approximately 108 cells were collected and resuspended in 100 μl of water. After adding 100 μl of 0.2 M NaOH, the cells were incubated for 10 min at room temperature (25°C). Cells were collected by centrifugation (9000 g for 1 min at 25°C), resuspended in 50 μl of sample buffer and incubated for 5 min at 95°C. Extracts were clarified by centrifugation (800 g for 10 min at 4°C) and equivalent amounts of protein resolved by SDS/PAGE (10% gel). After transferring on to nitrocellulose filters, tagged proteins were detected using the corresponding antibody (anti-HA 3F10 monoclonal antibody from Roche) and the ECL Advance Western blotting detection kit (GE Healthcare) according to the manufacturer's instructions.
Approximately 108 cells were collected and fixed in buffer A [0.5 mM MgCl2 and 100 mM KH2PO4 (pH 6.4)] with 37% formaldehyde (1:10 dilution) for 120 min with agitation. The cells were then collected, washed with buffer B (buffer A plus 1.2 M sorbitol) and incubated in 0.5 ml of buffer B containing 0.1% 2-mercaptoethanol and 200 μg of Zymolyase® (20-T/ml) at 37°C for 15 min. Spheroplasts were collected and washed with buffer B. An aliquot (5–10 μl) was applied to multitest slide wells, blocked with PBS containing 0.1% BSA for 20 min at 4°C, resuspended in 10 μl of a solution of 0.4 μg/ml anti-HA antibody 3F10 and incubated for 2 h at 4°C. After washing with PBS/BSA, the secondary antibody (Alexa Fluor™ 546-labelled anti-rat antibody; Molecular Probes) was added. DNA was stained at the end of the process by incubating spheroplasts in PBS containing 1 μg/ml of DAPI (Sigma). Samples were analysed using an Axioskop 2 fluorescence microscope (Zeiss) and pictures were taken with a SPOT digital camera (Diagnostic Instruments).
Approximately 108 cells were resuspended in 50 mM Tris/HCl (pH 7.5) supplemented with a mixture of protease inhibitors and broken with vigorous shaking in the presence of glass beads. The lysate was removed from the intact cells and the debris by centrifuged 2000 g for 5 min at 4°C. The supernatant was denatured by adding 1 volume of denaturation buffer [1 mM Tris/HCl (pH 7.5), 40 mM DTT and 0.5% SDS] and incubated for 5 min at 95°C. Samples were diluted 2-fold (final concentration 0.25% SDS) and adjusted to a final concentration of 0.2 mM ammonium acetate (pH 5.5). The sample was divided into two aliquots: one was treated with 50 milliunits of endo H (endoglycosidase H; Roche) and the other incubated with buffer at 37°C for 2 h. Proteins were resolved by SDS/PAGE (10% gel) and immunoblotted.
A crude cell extract was prepared in 0.1 M sorbitol, 50 mM potassium acetate, 20 mM Hepes (pH 7.5), 2 mM EDTA, 1 mM DTT and protease inhibitors with vigorous shaking of the cells in the presence of glass beads. Cell debris was pelleted at 700 g for 5 min and the supernatant divided into separate tubes, which were then subjected to one of the following treatments for 1 h on ice: buffer, 0.5 M NaCl, 0.1 M Na2CO3 (pH 11.5), urea (2.5 M) or 1% Triton X-100 plus 0.5 M NaCl. After incubation, the samples were separated into supernatant and pellet fractions by centrifugation at 4°C for 1 h at 100000 g (Beckman rotor TLA-55). Equivalent amounts of supernatant and pellet fractions were subjected to SDS/PAGE (10% gel) and detected by immunoblotting with anti-HA antibodies.
Approximately 6.0 A420 units of the E. coli NT326 strain cells (from Professor D.M. Engelman, Yale University, New Haven, CT, U.S.A.) were harvested by centrifugation and washed with 0.4 ml of sonication buffer [25 mM Tris/HCl and 2 mM EDTA (pH 8.0)]. The cells were resuspended in 2 ml of sonication buffer and lysed by probe sonication. After removing an aliquot (20 μl) for Western blot analysis, the remaining lysate was clarified by centrifugation at 13000 g for 15 min and the supernatant stored on ice until the spectrophotometric assay was performed. All constructs conferred the ability to grow on maltose plates to the malE-deficient strain NT326, which indicates that proper membrane insertion of the ToxR–Rot1 TM–MBP fusion protein occurred . For maltose complementation assays, E. coli NT326 cells expressing ToxR–Rot1 TM–MBP constructs were streaked on M9 minimal media plates containing 0.4% maltose as the only carbon source and incubated for 3 days at 37°C. The self-association ability of the TM domain triggers the expression of a CAT (chloramphenicol acetyltransferase) gene reporter, and production of the CAT protein can be quantified using a CAT-ELISA kit (Roche Diagnostics). CAT measurements and construct expression measurements were performed in at least triplicate and normalized for the relative expression level of each construct by Western blotting.
The hydrophobic C-terminal region of Rot1 spans ER-derived microsomal membranes
We described previously that the Rot1 protein is attached to the ER membrane through a C-terminal hydrophobic domain  (Figure 1A). To differentiate between the peripheral and TM association of this protein region, we assayed the membrane insertion capability of this region using an in vitro experimental system based on the inner membrane protein Lep of E. coli, which accurately determines the integration of TM helices into ER membranes . Lep consists of two TM segments connected by a cytoplasmic loop (P1) and a large C-terminal domain (P2), and inserts into ER-derived RMs with both termini located in the lumen (Figure 1B). The analysed segment [HR (hydrophobic region)-tested] was engineered into the luminal P2 domain and flanked by two acceptor sites (G1 and G2) for N-linked glycosylation. Both engineered glycosylation sites were used as membrane insertion reporters because G1 will always be glycosylated due to its native luminal localization, but G2 will be glycosylated only upon translocation of the tested region through the microsomal membrane. Therefore single glycosylation (i.e. membrane integration) results in an increase in molecular mass of ~2.5 kDa relative to the observed molecular mass of Lep expressed in the absence of microsomes; the molecular mass shifts 5 kDa upon double glycosylation (i.e. membrane translocation of the HR-tested). PK (proteinase K) added to microsomal vesicles digested the cytoplasm-exposed non-glycosylated form of the P2 domain (Figure 1B, left-hand panel) to produce a protected, double-glycosylated P2 fragment when the P2 domain is located in the lumen of the RMs (Figure 1B, right-hand panel).
Amino acid sequence and insertion of the transmembrane domain of Rot1
The membrane-spanning capacity of any protein region can be predicted using the apparent free energy (ΔGapp) for insertion into biological membranes using the ΔG Prediction Server v1.0  (http://dgpred.cbr.su.se/). Given the amino acid sequence of the Rot1 protein, the algorithm predicted only one TM helix with a computed ΔGapp value of −0.66 kcal/mol, which is indicative of a membrane-spanning disposition. On the basis of this prediction, residues 237–256 of Rot1 were selected for analysis using Lep chimaera membrane insertion. The translation of this construct mainly resulted in single glycosylated forms (Figure 1C, lane 5). PK treatment of this sample rendered a complete loss of detectable fragments (Figure 1C, lane 6), confirming the TM disposition of the predicted Rot1 sequence. A previously tested control construct [18,19] with a computer-designed non-integration sequence is shown for comparison (Figure 1C, lanes 1–3).
The Rot1 TM segment plays an essential role in cell viability
After demonstrating that the Rot1 C-terminal hydrophobic region spans biological membranes, we designed an experimental cell viability approach to study its functional role. To check the functionality of mutated forms of Rot1, we used a conditional mutant strain that expresses the ROT1 gene under the control of the tetracycline-regulated tetO7 promoter. As expected, the tetO7:ROT1 cells were non-viable in YPD medium containing 10 μg/ml doxycycline (one of the tetracycline antibiotics). This lethality was suppressed by the presence of a centromeric plasmid-containing ROT1 (pROT1) (Figure 2A). Thus the functionality of mutants can be evaluated by their ability to recover cell growth in the presence of doxycycline. Furthermore, in order to detect the Rot1 protein, a triple HA epitope sequence was fused in the middle region of the protein (between amino acids 181 and 182) without perturbing the essential C-terminus domain and with maintenance of the fully functional protein (pROT1–HA) (Figure 2A).
Essential role of the Rot1 TM domain in cell viability
We demonstrated previously that deletion of the C-terminal TM helix of the protein results in cell death . To determine whether the loss of function of the protein lacking the TM segment was due to the mislocation of the protein, to the C-terminus-truncated Rot1 protein (Rot1∆229–256) we fused a polyleucine stretch (Rot1∆229–256–18L; 18 consecutive leucine residues) that can restore the membrane-insertion capacity through this chimaeric C-terminal domain. We first tested the membrane association of this chimaera by subcellular fractionation. As shown in Figure 2(B), only the chimaeric protein was extracted from membranes after treatment with detergent and salt, a condition that has been proven to solubilize integral membrane proteins . Immunofluorescence assays of cells expressing Rot1 wild-type or Rot1∆229–256–18L proteins revealed a similar nuclear–peripheral signal (Figure 2C), a pattern reminiscent of nuclear envelope membrane staining, largely different from the diffuse pattern observed for cells expressing Rot1∆229–256. Thus the chimaeric Rot1∆229–256–18L protein localized correctly in yeast cells. However, the protein was not functional because it was unable to suppress the lethality of the tetO7:ROT1 strain in the presence of doxycycline (Figure 2D). The loss of functionality of this Rot1∆229–256–18L protein could have been caused by incorrect positioning of the mutant protein in the ER membrane. The fact that Rot1 is a glycosylated protein at Asn103, Asn107 and Asn139  is useful for unravelling the correct orientation of the protein. In eukaryotic cells, N-glycosylation occurs only in the lumen of the ER, where the active site of oligosaccharyl transferase is located . The glycosylation status of a protein can be assessed by digesting cell extracts with the glycan-removing enzyme endo H because the glycosylated and non-glycosylated forms are easily detected by their different electrophoretic mobilities. The Rot1∆229–256–18L protein was properly glycosylated (Figure 2E), indicating that the polyleucine-anchored chimaeric protein was properly inserted and oriented in the cellular endomembrane system, but it was not functional as shown by its incapacity to restore cell viability. Taken together, these results suggest that the TM domain of Rot1 not only anchors the protein to the membrane, but also plays an important role in protein function, probably through its association with other membrane proteins.
Scanning analysis of amino acids in the TM domain of Rot1
To rapidly identify the most critical segment of the TM helix of Rot1, we used an insertion scanning mutagenesis approach . The rationale was that insertion of an alanine residue (i.e. a hydrophobic residue with a high helix propensity) into a TM helix will displace the residues on the N-terminal side of the insertion by 100° relative to those on the C-terminal side of the insertion. This turn might effectively disrupt a putative helix–helix packing interface involving residues on both sides of the insertion site. Alternatively, no detrimental effects should be observed if the insertion is outside the critical interface residues. To test the structural consequences of the insertions, six alanine insertions were designed throughout the Rot1 TM helix (Table 1). The ability of the different Rot1 mutant proteins to support growth was tested using the tetO7:ROT1 strain in the presence of doxycycline. Only cells transformed with the plasmid-expressing mutants Rot1A232 and Rot1A251 were able to grow, whereas alanine residue insertions inside this region (Rot1A237, Rot1A242, Rot1A245 or Rot1A248) generated mutants that were unable to grow in the presence of doxycycline (Figure 3A). These results identified the central region of the TM C-terminal domain, between residues 231–251, as a critical region for cell viability.
Identification of the essential region within the TM domain of Rot1
|Plasmid||TM sequence (residues 229–256)|
|Plasmid||TM sequence (residues 229–256)|
Interestingly, when the extracts of cells expressing different alanine-inserted Rot1 mutants were treated with endo H, a clear shift in electrophoretic mobility was observed in all cases (Figure 3B). These results confirm that the alanine residue insertions did not affect the glycosylation pattern of the protein and, therefore, all Rot1 forms translocated across the yeast ER membrane similar to the native protein. We also tested the association of these mutant proteins with the membrane using subcellular fractionation. All of the alanine-inserted Rot1 mutants were recovered in the soluble fraction after incubation with detergent plus salt, which, as mentioned above, is a common feature for integral membrane proteins although some proteins also partially solubilized with Na2CO3 (pH 11.5) treatment (Figure 3C and Supplementary Figure S1 at http://www.biochemj.org/bj/458/bj4580239add.htm). These results indicate that all of the tested forms of the protein translocated correctly their large N-terminal region and inserted the hydrophobic C-terminal domain into the ER membrane. Therefore loss of viability in these alanine-inserted Rot1 mutants was not due to reverse orientation or mislocalization of the proteins. In conclusion, the results of the present study indicate that the region between residues 231–251 plays an essential role in cell viability, and clearly shows that the integrity of this domain is required for proper protein function.
Transmembrane Ser250 of Rot1 is a critical residue for protein function
Next, we further analysed the 231–251 region using a more focused saturation mutagenesis strategy. Within the past two decades, interactions between TM helices have been found to be based on different types of physical forces and to involve different amino acid types . Non-hydrophobic residues are involved in membrane protein stability and packing . Because replacement of the native TM segment by a polyleucine TM region did not restore cell viability, all of the non-hydrophobic residues in this region were independently mutated to leucine residues. Leucine is a highly hydrophobic residue, is the most prevalent residue in TM helices  and contributes significantly to TM helix stability.
The mutated Rot1 proteins (Table 2) were expressed in a centromeric plasmid, and its capacity to suppress the lethality of the tetO7:ROT1 strain in the presence of doxycycline was investigated. The ability of cells expressing the mutant proteins shown in Table 2 to support growth are shown in Figure 4(A). Almost all of the point mutant protein forms were able to maintain cell viability, indicating that these residues were not essential for protein functionality and/or helix association (Figure 4A). However, cells expressing Rot1 protein mutated at Ser250 (Rot1S250L) were clearly not viable in the presence of doxycycline. To rule out the possibility that this mutant protein was not properly expressed in yeast cells, we analysed the glycosylation status of the Rot1S250L mutant using endo H treatment and its association with the ER membrane using immunofluorescence assays (Figures 4B and 4C). No differences were found when comparing the wild-type protein and Rot1S250L. These results suggest strongly that Ser250 in the TM domain is not required for proper protein localization and orientation within the membrane, but that it is essential for protein function and cell viability.
Amino acid analysis of the TM domain of Rot1
|Plasmid||TM sequence (residues 232–252)|
|Plasmid||TM sequence (residues 232–252)|
To explore further the relevance of this single serine residue, we replaced Ser250 with other amino acid residues covering a broad range of side chain chemical groups, and confirmed the expression of all protein mutants in cell cultures (Supplementary Figure S2, upper panel at http://www.biochemj.org/bj/458/bj4580239add.htm). First, Ser250 was substituted with alanine in order to rule out the possibility that the loss of function was caused by the bulky leucine side chain. The Rot1S250A mutant was unable to support cell viability (Figure 4D), confirming the importance of the serine residue for Rot1 functionality. Residues containing hydroxy groups with strong hydrogen-bonding potential significantly change the local helical structure . In order to determine whether the essential role of the TM Ser250 is due to its polar hydroxy group, we analysed the effect of replacing Ser250 with a hydroxylated side chain residue, threonine or tyrosine. S250Y was not able to support the growth of the tetO7:ROT1 strain in the presence of doxycycline. In contrast, cells expressing the Rot1S250T mutant protein grew similarly to cells expressing the wild-type protein (Figure 4D), indicating that Rot1S250T is a functional protein. This finding demonstrates clearly that a side chain containing a hydroxy group is required for protein function, but this function can be prevented in the case of a bulkier side chain, as demonstrated for the S250Y mutant. Several studies have demonstrated a substantial contribution of hydrogen bonds formed by asparagine residues [3,8,27–29] and counter-charge aspartic acid residues  on the packing of TM helices. Therefore the replacement of Ser250 with asparagine (S250N) or aspartic acid (S250D) residues was investigated. As shown in Figure 4(D), neither S250N nor the S250D mutant restored cell viability, highlighting the relevance of the unique serine residue. As the S250D mutant was not viable, a negative charge, such as potential phosphorylation of the serine residue, was deemed as not essential for Rot1 function.
Next, we tested whether the hydroxy group has to be placed at a particular position within the TM helix. In order to test whether the position of Ser250 is critical for cell viability without altering the amino acid composition of the TM segment, double mutants were designed to locate the serine residue one position upstream (pROT1G249S/S250G) or downstream (pROT1S250L/L251S) of its usual position (Figure 5A). As shown in Figure 5(B), swapping Ser250 with one of its adjacent residues (Gly249 or Leu251) precludes cell survival. Next, we tested whether locating the critical serine residue just one helix turn upwards or downwards could rescue Rot1 function. Mutants swapping the leucine residues at positions 246 and 254 with Ser250 (Figure 5A) were designed, and none of these mutants recovered cell viability (Figure 5B). As before, double-mutant protein expressions were confirmed by Western blotting (Supplementary Figure S2, lower panel). Taken together, these results suggest that a precise geometric arrangement of the serine side chain in terms of azimuthal rotation (G249S/S250G and S250L/L251S) and depth of insertion (L246S/S250L and S250L/L254S) into the hydrophobic core of the membrane (Figure 5A) is required for cell viability. Therefore the serine residue hydroxy group must be placed at position 250 to achieve a functional protein in living cells.
Analysis of the Ser250 position in the TM domain and its conservation among yeast strains
The Rot1 TM segment does not homo-oligomerize in biological membranes
Integral membrane proteins are often organized into functional complexes and form homo- and/or hetero-oligomeric assemblies. To assess the propensity of the TM segment of Rot1 to homo-oligomerize in biological membranes, residues 237–256 were inserted into the ToxR–Rot1 TM–MBP chimaeric protein for use in the ToxCAT assay (Supplementary Figure S3 at http://www.biochemj.org/bj/458/bj4580239add.htm). In this assay [15,31], if the TM segment in the chimaeric protein drives homo-oligomerization in the inner membrane of E. coli, the ToxR domain is brought into close proximity with another ToxR domain, activating the transcription of a reporter gene encoding CAT. Thus the level of CAT expression in the cell can be used as an indicator of the extent of TM segment-mediated homo-oligomerization in the membrane . The malE complementation test showed that the chimaeric construct was inserted correctly into the inner membrane of E. coli strain NT326 and had the expected membrane topology (Supplementary Figure S3B). Western blotting of the E. coli cell lysates showed that the expression level of the Rot1-derived chimaeric protein was comparable with that of wild-type GpA and GpA G83I mutant, the well-documented control constructs for highly homo-oligomerizing and monomeric TM segments respectively [4,15]. However, the Rot1 TM construct did not induce significant CAT expression, indicating that the Rot1 TM domain does not homo-oligomerize in biological membranes (Supplementary Figure S3C).
Most studies that have characterized and analysed TM domains have been performed in vitro using artificial membranes with synthetic peptides. However, the environment in which a TM segment is studied can be critical and the detergent used to generate artificial membranes can significantly influence the arrangements of the TM helix, especially when this domain contains polar residues . The paucity of data obtained from in vivo approaches led us to investigate the role of the C-terminal hydrophobic region of the Rot1 protein. Rot1 is an essential yeast protein implicated in diverse processes, such as actin cytoskeleton dynamics, cell wall biosynthesis and molecular chaperoning. The Rot1 C-terminal domain provides an attractive model for testing whether this hydrophobic region functions as a mere membrane anchoring sequence or whether specific residues play a precise role in eukaryotic cells. An intrinsic advantage of using an essential protein is that the relevance of individual residues in protein function can be analysed in cell viability tests.
First, we demonstrated that the C-terminal hydrophobic region spans biological membranes and that this region is important for protein function. When the complete TM segment was substituted with a polyleucine stretch (18 consecutive leucine residues), no effect was observed on protein location, but cell viability was severely compromised. However, alanine insertion scanning and single leucine substitutions of most of the residues had no effect on protein function or cell growth. Only some mutations within the TM segment were found to be essential for cell survival, and the results of the present study identified a unique serine residue within the TM domain, Ser250, which appears to be critical for protein functionality. These results suggest that Ser250 may participate in homomeric and/or heteromeric interactions with other membrane components. Interestingly, Ser250 can be replaced by threonine (S250T mutant) without affecting protein function. This result strongly indicates that the hydroxy group present in serine and threonine is essential for a functional protein, probably by promoting TM helix association through polar interactions between similar residues or between these residues and backbone oxygens. The S250Y mutants were not viable, probably because the bulkier side chain of tyrosine precludes appropriate helix packing.
In principle, hydrogen-bonding interactions can be relevant at any position along the TM region. However, in the case of Rot1, the precise location of Ser250 in the helix is crucial. Relocating Ser250 one residue or one helix turn upstream or downstream compromised cell viability. In addition, mutation of the other serine residue present in the TM segment, Ser242, to leucine turned out to be irrelevant to protein function. These results indicate that an important hydrogen bond must be formed at the C-terminal portion of the helix in a particular position and/or orientation.
A single serine–serine interaction has been suggested to not provide a sufficient driving force for TM peptide association . In a model peptide that is entirely hydrophobic throughout a 20-residue segment containing a single central site for the introduction of various amino acid ‘guests’, asparagine, glutamine, aspartic acid and glutamic acid formed stable trimers in micelles; however, serine and threonine failed to direct trimer formation , indicating that the hydroxy group of serine plays a relatively modest role in stabilizing peptide association compared with other polar residues, such as asparagine and glutamine [3,5,6,33]. Instead, the interaction between multiple serine residues allows the formation of a ‘serine zipper’, which is postulated to promote TM helix association [9,10]. Thus first we studied the importance of Asn236 within the TM domain of Rot1, and secondly whether Ser250 function can be replaced by asparagine. The results with N236L and S250N showed clearly that the polar asparagine residue side chain is not involved in Rot1 TM function.
The essential role for the Rot1 TM segment beyond its anchoring function suggests that the protein must associate with other membrane components to exert its function. Notably, global studies using massive analysis techniques have not been able to identify any Rot1-associated protein in crude yeast extracts [34,35]. In a recent interaction landscape study specific for membrane protein complexes in S. cerevisiae no protein interaction was identified for Rot1 , probably due to the harsh extraction conditions that could disrupt the association of Rot1 with other proteins. ToxCAT studies indicated the absence of TM-driven homo-oligomerization in cell membranes (Supplementary Figure S3). We also studied other proteins that are known to interact with Rot1, such as the protein chaperone Kar2 (karyogamy 2)  and a component of the post-translational translocation complex Sec62 (secretory 62)  (Supplementary Figure S4 at http://www.biochemj.org/bj/458/bj4580239add.htm). In both cases, we did not find significant differences in the protein interactions between wild-type Rot1 and the Rot1S250L mutant using co-immunoprecipitation analysis.
The relevance of TM serine residues to protein function has been studied previously in other proteins. The mutation of some serine residues within the predicted TM helix 17 of MRP3 (multidrug-resistance protein 3) determined changes in substrate specificity . Replacing Ser319 of TM helix H7 of hamster β2-adrenergic receptor can influence receptor expression and function . In addition, substitution of Ser161 and Ser165 in TM helix H4 decreased the expression and activity of the β2-adrenergic receptor, but did not affect specific binding to the antagonist ligand . As far as we know, the present study of the Rot1 protein is the first time in which a single polar amino acid in a TM segment has been shown to be essential for cell viability. Rot1 is a conserved protein in yeast, although its function has only been studied in S. cerevisiae. Comparative analysis of the TM sequence showed that Ser250 is conserved in 21 of the 25 species compared (one substituted by threonine) (Figure 5C). In this sequence alignment, we observed two organisms that replaced Ser250 with glycine and in Candida tenuis Ser250 was replaced with alanine. Since substitution with alanine (S250A) has been proven to preclude cell growth, we further replaced Ser250 with glycine in the S. cerevisiae Rot1 protein. We performed the required amino acid change, S250G, but this mutant was not viable (Figure 5B). Therefore the function of Ser250 seems to be highly conserved among yeast.
To summarize, we used an in vivo system to describe how a unique hydroxy-containing residue inside a TM domain is essential for maintaining protein function and/or protein association. These findings add force to the emerging picture of the role of polar groups in membrane protein folding. Although interactions with other parts of the molecule cannot be discarded, on the basis of our biochemical and cell viability results, Ser250 might be involved in hetero-oligomeric interactions. Further studies are necessary to ascertain the specific role for Ser250 in these interactions and its partners within the cell membrane, and to investigate whether single serine residue-based motifs may provide a novel mechanism for membrane protein association/function.
J. Carlos Igual, Ismael Mingarro and M. Carmen Bañó developed the original concept. All authors conceived and designed the experiments. Carlos Martínez-Garay performed the experiments. Carlos Martínez-Garay, J. Carlos Igual, Ismael Mingarro and M. Carmen Bañó analysed the data. Ismael Mingarro and M. Carmen Bañó wrote the paper. Carlos Martínez-Garay, M. Angeles Juanes and J. Carlos Igual proofread drafts of the paper before submission.
We thank Professor D. M. Engelman for the ToxCAT vectors and E. coli NT326 cells.
This work was supported by the Ministerio de Ciencia e Innovación of the Spanish Governement (co-financed by European Regional Development Fund of the European Union) [grant numbers BFU2010-20927 and BFU2012-39482] and Generalitat Valenciana [grant numbers ACOMP/2012/219, GVACOMP/2013-175, PROMETEO/2010/005 and ISIC/2013/004].
Present address: Centre de Recherche en Biochimie Macromoléculaire, 1919 Route de Mende, 34293 Montpellier, France.