The proton gradient acts as the driving force for the transport of many metabolites across fungal and plant plasma membranes. Identifying the mechanism of proton relay is critical for understanding the mechanism of transport mediated by these transporters. We investigated two strategies for identifying residues critical for proton-dependent substrate transport in the yeast glutathione transporter, Hgt1p, a member of the poorly understood oligopeptide transporter family of transporters. In the first strategy, we tried to identify the pH-independent mutants that could grow at higher pH when dependant on glutathione transport. Screening a library of 269 alanine mutants of the transmembrane domains (TMDs) along with a random mutagenesis strategy yielded two residues (E135K on the cusp of TMD2 and N710S on TMD12) that permitted growth on glutathione at pH 8.0. Further analysis revealed that these residues were not involved in proton symport even though they conferred better transport at a higher pH. The second strategy involved a knowledge-driven approach, targeting 31 potential residues based on charge, conservation and location. Mutation of these residues followed by functional and biochemical characterization revealed E177A, Y193A, D335A, Y374A, H445A and R554A as being defective in proton transport. Further analysis enabled possible roles of these residues to be assigned in proton relay. The implications of these findings in relation to Hgt1p and the suitability of these strategic approaches for identifying such residues are discussed.
The electrochemical proton gradient acts as the driving force for the uptake and efflux of many metabolites and ions across the plasma membrane of fungi and plants [1,2]. A large number of transporters in these organisms are, therefore, proton coupled . However, the mechanisms by which proton coupling to substrate transport occurs differs in the different transporters. The Saccharomyces cerevisiae plasma membrane high affinity glutathione (GSH) transporter, Hgt1p (ScOpt1p) is one such proton-coupled transporter [4,5]. It uses the inwardly directed proton electrochemical gradient to drive the uphill transport of GSH against a concentration gradient [2,4]. Hgt1p belongs to the oligopeptide transporter (OPT) family. Members of this family fall into two broad clades that mediate the uptake of GSH, peptides, modified peptides and metal-binding secondary amino acid derivatives: the peptide transporter (PT) clade and the yellow stripe-like (YSL) clade [6–8]. The OPT family is poorly studied and no crystal structures are available for any members to guide the analysis.
Hgt1p remains the best-studied member of this OPT family. Extensive mapping of the residues in the transmembrane domains (TMDs) has led to significant insights into the TMDs and their residues important for GSH binding and translocation [9–12]. However, the residues important for proton binding and proton relay have not been investigated. Identification of amino acid residues involved in proton transport is difficult even in transporters whose crystal structures are available [13,14]. It is the combination of crystal structure analysis and different biochemical, biophysical and molecular genetic approaches that have been instrumental in providing clues for determining how protons may bind and how they are transported. Mutants defective in proton translocation cannot catalyze active transport of substrate in such transporters . Hence, active transport in combination with exchange or counterflow transport experiments are generally used to delineate the steps in the substrate translocation cycle since the latter processes remain unaffected by the proton gradient and do not involve net proton transport [15,16]. Lactose permease (lactose/H+ symporter, LacY) serves as a paradigm for studying proton-coupled transporters and the residues that are involved in the proton coupling; proton transport and relay and the proton turnover have been clearly defined in this transporter [16–20].
In the absence of any structural information for any of the OPT family members, we have proceeded to investigate Hgt1p protein largely through molecular genetic approaches. In the present study, with the goal of identifying the residues involved in proton binding and transport by Hgt1p, we carried out an exhaustive analysis to screen for Hgt1p mutants showing pH-independent growth using a pH-based plate assay. Evaluation of all 269 alanine mutants of the thirteen predicted TMDs in combination with in vitro random mutagenesis identified E135A and N710A to be important for pH-dependent transport. Biochemical characterization of these mutants also showed increased uptake at higher pH as compared with the WT Hgt1p but the transport was still proton-dependent. We, therefore, extended the search by using a knowledge-driven approach where we targeted conserved and charged residues and, through this, we were able to identify several residues including E177, Y193, D335, Y374 and H445 and R554 to be important for proton-dependent substrate transport by Hgt1p. Determination of kinetic parameters and cytosolic acidification assays has allowed us to assign a possible role of these residues in proton translocation.
Materials and methods
Chemicals and reagents
Chemicals used in this study were of analytical grade and procured from commercial sources. Growth medium components were obtained from Sigma Aldrich (St. Louis, MO), Difco (Detroit, MI) and HiMedia, (Mumbai, India). [35S] GSH, 100 µCi (3.7 MBq) was obtained from Perkin Elmer (U.S.A). Oligonucleotides were obtained from Integrated DNA Technologies (IDT) India. Vent DNA polymerase, Phusion®DNA polymerase, restriction enzymes and DNA-modifying enzymes were purchased from New England BioLabs (NEB, Beverly, MA). Plasmid mini-prep and gel-extraction kits were obtained from Thermo Scientific. Hybridization nitrocellulose membrane (filter type 0.45 µm) and LuminataTM Forte Western HRP substrate were purchased from Millipore (Billercia, MA). HA-tag mouse monoclonal antibody and horse anti-mouse HRP-linked antibody, FLAG-tag rabbit monoclonal antibody, anti-rabbit IgG Alexa Fluor® 647-conjugated antibody were procured from Cell Signalling Technology (Danvers, MA). Alexa Fluor® 488-conjugated goat anti-mouse antibody was purchased from Molecular Probes.
Strains, medium and growth conditions
The Escherichiacoli strain DH5α was used as a cloning host and grown at 37°C. The S. cerevisiae strain ABC817 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 hgt1Δ::LEU2) carrying an HGT1 disruption (i.e. deficient in GSH uptake ability) and an organic sulphur auxotroph was used throughout the study . Yeast strains were maintained on yeast extract, peptone and dextrose (YPD) medium and grown at 30°C. The yeast transformants were selected and maintained on synthetic defined (SD) minimal medium containing yeast nitrogen base (YNB), ammonium sulphate and glucose supplemented with leucine, methionine, histidine, lysine and adenine at 100 mg/litre. The pH of the media was adjusted with NaOH or HCl. GSH was added in place of methionine when required. Growth and handling of yeast and bacteria and all the molecular biology techniques used in this study were carried out according to standard protocols .
In vitro hydroxylamine mutagenesis
The protocol for in vitro random mutagenesis was adopted from Rose and Fink (1987) . Briefly, 10 µg purified plasmid (p416TEF-His-HGT1-HA) was mixed in 0.5 ml freshly made hydroxylamine solution (350 mg hydroxylamine HCl (Sigma Aldrich), 90 mg NaOH in 5 ml ice-cold autoclaved Milli-Q water, pH 6.5). The mixture was then incubated at 37°C for 22 h and the plasmid was purified using DNA binding column. This pool of mutagenized plasmid was transformed into the appropriate yeast strain background.
The plasmid construct p416TEF-His-HGT1-HA  (i.e., HGT1), cloned downstream of the TEF promoter (BamH1 and EcoR1 site) of p416TEF vector, with a hexahistidine epitope at the N-terminus and a hemagglutinin tag at the C-terminus was used as a template for the creation of different site-directed mutants of HGT1 using the splice overlap extension strategy. The PCR products generated with the desired oligonucleotides were subcloned back into the empty TEF vector background using appropriate restriction sites. To confirm the desired nucleotide changes and rule out the possibility of any unwanted mutations introduced during the mutagenic procedure, the resulting mutants were sequenced.
Dual complementation-cum-toxicity assay
The assay has previously been described in great detail . Briefly, the met15Δhgt1Δ S. cerevisiae strain, ABC817 was transformed with an empty vector or a vector overexpressing WT or different mutants of HGT1. Yeast transformants were selected and grown overnight in minimal media containing methionine as a sulphur source along with other supplements without uracil. Cultures were reinoculated in fresh media and allowed to grow until optical density (OD600) of 0.5–0.8. Cells were harvested, washed twice and resuspended in sterile water at an OD600 of 0.2. These suspensions were serially diluted to attain 1:10, 1:100, and 1:1000 dilutions in 1 ml and 10 µl of these cell resuspensions were spotted on minimal medium plates containing methionine (200 µM, as control) or different pH plates containing a range of GSH concentrations (15, 30, 50, 100 and 200 µM) as the sole source of organic sulphur. The plates were incubated at 30°C for 2–3 days and photographs were taken. HGT1 under the strong TEF promoter is able to grow on low (15 µM) GSH concentrations. However, growth at higher concentrations (>50 µM) leads to excessive uptake of GSH, causing toxicity . Using this growth phenotype (i.e., complementation of the hgt1Δ defect at low GSH concentrations, and toxicity at higher GSH concentrations), we have been able to grade the mutants depending upon the transport defect, as described previously .
GSH transport assay
This assay has previously been described in great detail . Briefly, the yeast ABC817 strain transformed with WT or HGT1 mutant plasmid constructs under TEF promoter was grown in minimal media containing methionine and other supplements, without uracil, until an OD600 of 0.6–0.8 was reached. Cells were harvested, washed twice with ice-cold water and, finally, cells were resuspended into resuspension buffers, including 20 mM MES/KOH buffer (pH 5–6) or 20 mM HEPES/KOH (pH 7–8). In each case, the buffer contained 0.5 mM CaCl2, 0.25 mM MgCl2 and 2 glucose. 200 µl of cells at OD600 of 3.75 ml−1 were aliquoted for each sample and after a pre-incubation of cells for 5 min at 30 °C, GSH uptake was initiated by the addition of 200 µl of assay medium containing radiolabelled GSH (35S-GSH). After 1 min and 3 min time points, the uptake was stopped by diluting the medium 20-fold with ice-cold water and cells were collected on the glass fiber filter using vacuum filtration. The harvested cells were washed with ice-cold water and the filters containing cells were immersed in scintillation fluid (Sigma-Fluor Universal LSC cocktail, Sigma). Radioactivity was measured using a liquid scintillation counter (Perkin Elmer). For saturation kinetics, the initial rate of GSH uptake was measured at a range (12.5 µM to 800 µM) of GSH concentrations. The initial rate of GSH uptake in cells containing test plasmids was determined after subtracting the corresponding values from cells containing an empty vector. Total protein was estimated using the Bradford reagent (Sigma Aldrich), using bovine serum albumin (BSA) as standard.
The HGT1 protein sequence was retrieved from the Saccharomyces genome database (SGD), whereas, the sequence for other OPT members was retrieved from Entrez at NCBI website. The multiple sequence alignment of the protein sequences was generated using the MUSCLE program with default parameters .
Western blot analysis
Western blot analysis using crude cell preparation was carried out as described previously [9,25]. To quantify the protein expression levels in different mutants, Scion Image software was used for the densitometry analysis of the unsaturated band signals. The resulting signal intensity was normalized with respect to the band surface area and expressed in arbitrary units. Finally the data are represented as percentage protein expression levels normalized to the wild-type expression levels and are the means of three independent experiments.
Cellular localization of the mutants
To check localization of Hgt1p and its different mutants, indirect immunofluorescence was performed using a modified published protocol [9,11]. Briefly, yeast cells with a plasma membrane H+-ATPase , genomically FLAG-tagged at the C-terminus were transformed with WT-HGT1 or mutant plasmid construct under a TEF promoter. Cells were grown in minimal media containing methionine and other supplements without uracil and harvested at OD600 of 0.6–0.8. Cells were fixed and spheroplasting was carried out using lyticase enzyme (Sigma Aldrich), washed using sorbitol buffer and, finally, adhered on the polylysine coated coverslips. Permeabilization was carried out by treatment of 0.4% Triton X-100 in PBS (pH 7.4) followed by blocking with 1 % BSA in PBS. Overnight incubation with rabbit anti-FLAG and mouse anti-HA primary antibodies at 4°C, followed by washing and treatment with secondary antibodies (goat anti-mouse IgG-Alexa 488 and mouse anti-rabbit IgG Alexa 647) for 4 h at room temperature (RT). Coverslips were washed with blocking buffer and inverted onto a slide using vectashield antifade mounting medium (Vector Laboratories). Images were obtained with an inverted LSM780 laser scanning confocal microscope (Carl Zeiss) with a Plan-Apochromat X63, oil immersion objective with numerical aperture of 1.4 at RT. The 488 nm argon ion laser and 633 nm of He–Ne ion laser line was directed over an MBS 488/561/633 beam splitter, and fluorescence was detected. Images obtained were processed using ImageJ software.
Cytosolic acidification measurement
The yeast ABC817 strain expressing pHluorin was transformed with WT-HGT1 or mutant plasmid constructs under the TEF promoter, and was grown overnight in minimal media containing methionine. The cultures were then reinoculated in a fresh media and further allowed to grow until they reached OD600 of 0.6–0.8. Cells were harvested, washed and finally resuspended (3.5 OD600/ml) into a resuspension buffer (20 mM MES/KOH, 0.5 mM CaCl2, 0.25 mM MgCl2, 2% glucose, pH 5.5). After a 10 min incubation at 30°C, cell resuspension was transferred to cuvettes and fluorescence intensity was monitored for 30 or 60 min using FluoroMax 4 spectrofluorometer (Horiba Jobin Yvon Inc.) at 508 nm with excitation at 405 nm and 485 nm every 30 sec at 30°C with constant stirring (1 mM GSH was added after 3 min). Background signals (cells without pHluorin) were subtracted and the fluorescence ratio R405/485 was calculated and plotted vs. time using GraphPad Prism 5.0 software.
Isolation of Hgt1p mutants showing pH-independent growth phenotype on GSH
Based on a previously described strategy to isolate mutants conferring pH-independent uptake as residues that are important in proton coupling , we used the property of low or no transport at high pH (Figure 1a) as an initial strategy towards identifying the residues involved in proton binding and transport in Hgt1p. Using this strategy, mutants that allowed growth at higher pH were identified. The logical basis for isolating such mutants has been that those residues that, upon mutation, can lead to growth at a high pH would be likely to play a role in proton binding/proton transport. To determine at what pH the screen should be performed, the organic sulphur auxotroph strain of S. cerevisiae (met15Δ hgt1Δ) was transformed with empty or WT-HGT1 plasmid constructs under the TEF promoter and subjected to a growth-based assay on minimal media of different pHs (5.5–8.0) containing different concentrations of GSH (15–200 µM) as the sole organic sulphur source (Figure 1b) (only growth at 15 µM GSH, pH 5.5, pH 7.0 and pH 8.0 is shown). We observed that, at pH 8.0, the cells were not able to grow at concentrations of 15–30 µM GSH, although they were able to survive at higher GSH concentrations (>50 µM). These results are consistent with the [35S] GSH uptake assay, where the WT Hgt1p shows an almost negligible amount of transport at pH 8.0. Based on the lack of growth at 15 µM GSH at pH 8.0, we embarked on an effort to identify mutants able to grow at pH 8.0 on GSH.
Effect of pH on glutathione transport mediated by Hgt1p.
As the proton binding residues are expected to lie within or in the periphery of the TMDs , we first focused on the TMD residues. Hgt1p is predicted to have 13 TMDs consisting of 269 amino acid residues. All these residues have been previously subjected to alanine scanning mutagenesis that identified residues important for GSH binding and transport . This collection of 269 alanine mutants were subjected to an evaluation of growth at 15–200 µM GSH at pH 5.5 and pH 8.0 (data not shown). This allowed us to identify two mutant residues, E135A (TMD2) and N710A (TMD12), which showed a better growth compared with WT at pH 8.0.
Although a screen of the 269 TMD mutant residues yielded two candidate residues, the search was restricted to the predicted TMDs (and also limited to Ala mutations). To enlarge the scope of the screen beyond the TMDs, we carried out a random mutant hunt in order to identify any other residues that might also lead to a pH-independent growth phenotype. We performed in vitro hydroxylamine-based random mutagenesis of the TEF-HGT1 plasmid and, after transformation of the mutated plasmid in the ABC817 strain (met15Δhgt1Δ), we selected for mutants that were capable of growing on 15 µM GSH at pH 8.0. Four mutants were isolated, out of which one contained two mutations—E135K/D157N—whereas the other three carried just the E135K mutation. Separation of E135K and D157N revealed that E135K was in fact the critical mutant that allowed growth at a higher pH (see Supplementary Figure S1).
Analysis of pH-independence of amino acid substitutions at E135 and N710
To further elucidate the role of E135 and N710 residues, these residues were mutated to amino acids with different properties and functionally evaluated. Growth of the mutants was compared with WT-HGT1 at pH 5.5 and pH 8.0, and comparative [35S] GSH uptake was assessed.
We first evaluated the behavior of E135 mutants. Mutation of E135 (Glu) to a similarly charged Asp (E135D), retained the carboxylate group at this position. This mutant was found to behave like WT at pH 5.5 and pH 8.0 (Figure 2a). By contrast, the mutants E135K, E135Q, E135N, E135S, E135A, E135G and E135H retained significant growth at pH 8.0. E135W appeared to be a general defect since the activity was also lost at low pH. These mutants were further analyzed by biochemical uptake using [35S] GSH at the two different pHs (Figure 2b). At pH 5.5, the E135D mutant showed uptake comparable to the WT, whereas E135Q, E135N, E135A, E135K, E135H, E135G, E135S and E135W demonstrated transport of 30–70% of that of the WT protein. However, at pH 8.0, WT-HGT1 and E135D showed a drastic reduction in uptake. Although a reduction was also seen with other mutants, other mutants showed transport that was ≥ two-fold higher compared with the WT. Thus, in addition to the original E135, only E135D showed a lack of growth at 15 µM GSH, pH 8.0, suggesting that an acidic residue at this position hampered growth at a higher pH.
Analysis of pH dependence of E135 mutants.
In the case of the N710 residue with a polar uncharged side chain, when replaced with a charged residue, N710R and N710K showed severe growth defects at pH 5.5. In addition, they demonstrated a failure to grow at any GSH concentation at pH 8.0 (Figure 3a), suggesting that these charged residues were not well tolerated. Only N710S along with N710A showed an ability to grow at pH 8.0. These results were also reflected in the radioactive GSH uptake (Figure 3b). At pH 5.5, N710A and N710G showed similar uptakes compared with the WT, whereas N710E, N710K, N710H, N710R showed lesser uptakes. However, at pH 8.0, only N710S showed a better uptake, suggesting that mutation to Ser was leading to pH-independent uptake by Hgt1p.
Analysis of pH dependence of N710 mutants.
Since E135 and N710 mutants demonstrated a similar ability to grow at pH 8.0, we attempted to determine whether or not the two residues were interacting with each other. We created a double mutant, E135A N710A. However, this double mutant did not demonstrate an increase in the phenotype, suggesting that the residues were probably functioning in the same pathway, or were interacting elsewhere (see Supplementary Figure S2). To examine the latter possibility, we carried out a residue swap whereby E135 was changed to N, and N710 was changed to E. Although the double mutant was functional at pH 5.5, the swapped mutant did not show growth at pH 8.0, suggesting a possible interaction between the two residues. However, the single N710E mutant also showed no growth at the higher pH, contradicting the theory of the interaction between the two residues.
Kinetic characterization of E135 and N710 mutants at pH 5.5 and pH 8.0
To further analyze these mutants, uptake kinetics for [35S] GSH were determined at pH 5.5 and pH 8.0 over a range of concentrations (12.5–800 µM) in cells transformed with an empty vector, or a vector expressing WT-HGT1, E135 (E135A and E135N) or N710 (N710A and N710S) mutants.
At pH 5.5, the apparent Km of WT-HGT1 for GSH was estimated to be 27.8 ± 1.2 µM and the Vmax was found to be 57.0 ± 0.8 nmol of GSH per mg of protein per minute . The Km values for E135A (28.4 ± 2.3 µM) and E135N (28.7 ± 1.9 µM) were found to be similar to WT-HGT1 suggesting that the mutation did not affect substrate binding. However, the Vmax for E135A (40.0 ± 3.8 nmol of GSH per mg of protein per minute) and E135N (45.5 ± 2.1 nmol of GSH per mg of protein per minute) decreased in both of the mutants, suggesting that the moderate defect as seen in the plate-based assay is due to a slight defect in the carrier translocation step. At pH 8.0, Vmax for WT was found to be 9.0 ± 3.3 nmol of GSH per mg of protein per minute suggesting a six-fold decrease in the translocation rate compared with pH 5.5. However, for E135A and E135N, at pH 8.0, Vmax only decreased three-fold (Table 1). Hence, as the pH was increased from 5.5 to 8.0, there was a decrease in the translocation rate for both the WT and E135 mutants. However, the decrease was less than that observed for WT-Hgt1p, suggesting that the E135 mutants are not completely independent of pH, but retain some pH sensitivity.
|Mutants||pH 5.5||pH 8.0|
|WT||27.8 ± 1.2||57.0 ± 0.8||305.6 ± 60.1||9.0 ± 3.3|
|E135A||28.4 ± 2.3||40.0 ± 3.8||189.6 ± 54.3||13.9 ± 2.5|
|E135N||28.7 ± 1.9||45.5 ± 2.1||194.8 ± 38.7||20.4 ± 3.4|
|N710A||24.6 ± 3.6||47.7 ± 5.1||227.4 ± 67.8||9.6 ± 3.4|
|N710S||22.5 ± 3.8||43.8 ± 4.9||154.8 ± 49.3||15.4 ± 1.1|
|Mutants||pH 5.5||pH 8.0|
|WT||27.8 ± 1.2||57.0 ± 0.8||305.6 ± 60.1||9.0 ± 3.3|
|E135A||28.4 ± 2.3||40.0 ± 3.8||189.6 ± 54.3||13.9 ± 2.5|
|E135N||28.7 ± 1.9||45.5 ± 2.1||194.8 ± 38.7||20.4 ± 3.4|
|N710A||24.6 ± 3.6||47.7 ± 5.1||227.4 ± 67.8||9.6 ± 3.4|
|N710S||22.5 ± 3.8||43.8 ± 4.9||154.8 ± 49.3||15.4 ± 1.1|
On the other hand, both N710A and N710S showed an increase in substrate affinity (Km = 24.6 ± 3.6 and 22.5 ± 3.8 µM) at pH 5.5 and only a slight defect in Vmax (47.7 ± 5.1 and 43.8 ± 4.9 nmol of GSH per mg of protein per minute) (Table 1). Whereas at pH 8.0, the N710A mutant was found to have a similar Vmax compared with WT Hgt1p but a lower Km (227.2 ± 67.8 µM). However, in the case of the N710S mutant, Km further decreased to 154.8 ± 49.3µM and Vmax increased to 15.4 ± 1.1 nmol of GSH per mg of protein per minute suggesting a better translocation rate compared with the WT.
GSH transport by E135N and N710S mutant does not lead to defect in cytosolic acidification
Hgt1p is a proton-dependent symporter and GSH transport by Hgt1p will result in acidification of the cytosol. A proton-independent transport, on the other hand, will not change the intracellular pH. This could happen in a mutant for which significant substrate binding occurs even in the absence of prior proton binding. Since E135N and N710S showed a loss in pH sensitivity of the transporter, it is possible that they may be directly or indirectly affecting proton-dependent GSH uptake by Hgt1p. To check the possibility that the mutants may be involved in proton-dependent transport, we used yeast cells expressing ratiometric pHluorin in the cytosol [29,30]. Ratiometric pHluorin is a GFP variant that displays a dual excitation spectrum with peaks at 405 and 485 nm and an emission maximum at 509 nm. As the pH decreases, the fluorescence intensity decreases at an excitation of 405 nm and increases at 485 nm. Therefore, a reduction in the excitation ratio of 405/485 nm represents acidification of the cytosol. We transformed yeast cells expressing pHluorin with an empty vector and with vectors overexpressing WT Hgt1p or the mutants (E135N and N710S) independently, and measured the change in the fluorescence ratio in response to GSH transport for 27 min (Figure 4). For yeast cells with an empty vector, there was a negligible change in the fluorescence ratio (≈0.1) upon GSH addition, suggesting that the pH inside the cell remained unchanged. For WT Hgt1p, there was a drastic decrease in the ratio from 1.78 to 1.22 after 12 min of GSH addition, suggesting rapid acidification of the cytosol in response to GSH transport, after which the ratio starts recovering. However, the mutants E135N and N710S showing pH-independent growth showed a similar decrease in the ratio upon GSH addition, suggesting acidification of the cytosol in response to GSH transport. Hence, these mutants show proton-dependent GSH transport.
Cytosolic acidification in response to glutathione transport by Hgt1p and its mutants.
Evaluation of conserved charged residues for their role in proton translocation
Although the mutants E135N and N710S led to pH-independent growth, they were not defective in cellular acidification during GSH transport. This suggests that they are not defective in proton uptake. We therefore embarked on an additional screening approach whereby we targeted the conserved charged residues (Glu, Asp, His, Lys and Arg) and evaluated them for growth and proton uptake. Owing to the large number of such residues in Hgt1p and the difficulty in subjecting them all to mutational analysis, the search was limited by two criteria. Asp/Glu/His, which are at least partially conserved in the PT clade of the OPT family and are present in the predicted TMDs or interconnecting loops (ICLs), were considered for the analysis (Figure 5a). In addition, partially conserved Arg/Lys/Tyr present in the predicted TMDs were also included. This represents a total of 31 residues to be explored for their involvement in proton binding, or transport, as shown in the 2D predicted topology model of Hgt1p (Figure 5b).
Conservation pattern and position of residues selected for mutagenesis.
Among the 31 residues selected, 17 residues were present in the predicted TMDs and have already been mutated to alanine and studied for their effect on GSH transport in previous studies ; however, the probability of their involvement in proton transport was never explored. The remaining 14 were therefore mutated to alanine, and all the 31 residues were functionally evaluated using the ‘dual complementation-cum toxicity' assay as described in materials and methods section. Based on this assay, the majority of the mutants (D335A, D408A, D467A, D504A, D578A, E177A, E530A, E544A, E744A, R554A, K562A, Y193A, Y226A, Y289A Y374A, Y449A and Y485A) were found to be severely defective in GSH transport (Figure 6). Of these severely defective mutants, Y289A and Y485A did not show a consistent phenotype. Further, to confirm the functionality of the mutants, we measured the initial rate of [35S] GSH uptake in met15Δ hgt1Δ yeast strain transformed with WT or different alanine mutants of Hgt1p and found that the genetic plate-based assay was clearly reflected in the uptake assay (Figure 7). The initial rate of [35S]GSH uptake was 0.1–18% of the WT uptake in severely defective mutants, 22–40% in moderately defective mutants and 50–75% higher in minor defective mutants.
Functional characterization of mutants based on the plate-based assay.
Comparative uptake of [35S] GSH by alanine mutants of Hgt1p.
Since a defect in proton translocation will severely impede GSH transport by Hgt1p, only severely and moderately defective mutants showing less than 40% transport comprising of 19 mutants were studied further.
Analysis of protein expression levels and cell surface trafficking of mutants
The decreased transport activity of the HGT1 mutants as observed in the plate-based growth assay or comparative uptake assay can be due to either decreased protein expression levels, a defect in trafficking to the plasma membrane or due to loss of the activity of the protein. Previous studies on expression and localization of mutants Y193A, Y226A, Y374A, H445A, Y449A, E544A, R554A, K562A and E744A suggested that, except Y449A, K562A and E744A, all other mutants are properly expressed and localized to the plasma membrane . Hence, these mutants were not evaluated for expression and localization studies.
Steady-state protein expression analysis of the remaining 10 mutants revealed that the majority of Hgt1p mutants (D335A, D482A, D578A, D778A, E177A, E530A and H258A) expressed protein ranging from 70 to 100% of the total WT expression (see Supplementary Figure S3), thus suggesting that the mutation did not perturb the expression levels. By contrast, mutants D408A, D467A, and D504A showed a significant fall in protein expression levels, accounting for low transport levels in these mutants. The seven mutants expressing significant levels of protein were further analyzed for a defect in trafficking and localization, if any, by indirect immunofluorescence and were compared with the localization of genomically FLAG-tagged plasma membrane marker protein PMA1. A very bright fluorescence signal at the cellular periphery was observed for WT-Hgt1p and all the mutants co-localizing with PMA1 marker protein, suggesting that these mutants are correctly localized on the plasma membrane (see Supplementary Figure S4).
Cytosolic acidification monitoring by pHluorin based assay
The thirteen mutants showing proper expression and localization to the plasma membrane were further checked for proton-dependent GSH transport using pHluorin. In the case of mutants D482A, D578A, D778A, E530A, E544A, H258A and Y226A, there was a significant change in the ratio (405/485), although this was less than the change for WT upon GSH addition suggesting co-transport of the protons with substrate (Figure 8). Interestingly, for mutants E177A, Y193A, D335A, H445A and R554A, almost no acidification of cytosol was observed. Y374A showed only a slight decrease in the ratio, suggesting very weak cytosolic acidification in response to GSH transport as compared with mutants showing similar levels of [35S] GSH transport. Hence, these mutants are likely to be involved in substrate/proton coupling.
Cytosolic acidification monitoring of mutants using pHluorin.
The mutants defective in acidification of cytosol in response to GSH transport were further assessed for transport at different pHs (5, 6, 7 and 8) (see Supplementary Figure S5). E177A and Y193A showed an almost negligible amount of transport compared with WT at all tested pHs. By contrast, mutants D335A, Y374A, H445A and R554A showed pH-dependent uptakes similar to WT, although the transport rate was less in these mutants. This result suggest that the initial substrate binding is still proton dependent, and these residues may participate in the latter stages of the proton translocation.
Saturation kinetics studies of mutants
In the majority of proton-coupled symporters, proton binds first and this protonation assists the substrate binding [31,32]. In this model, if the initial protonation of Hgt1p is prevented, then substrate binding will be severely affected, leading to a decrease in substrate affinity and also a decreased translocation rate. However, in this model, if the initial protonation is unaffected but the subsequent proton translocation is perturbed, this will lead to a significant change in the translocation rate without affecting the substrate affinity. By contrast, if substrate binding does not depend significantly on the prior proton binding, one would expect a change in the translocation rates without affecting the Km. To evaluate these possibilities in Hgt1p, we carried out the saturation kinetics of the mutants.
E177A, Y193A, D335A, Y374A, H445A and R554A did not cause a significant defect on either protein expression or its cell surface localization. However, these mutants showed almost no acidification of the cytosol in response to GSH transport, suggesting a specific defect in proton-coupled GSH uptake. The mutants E177A and Y193A were completely defective in GSH transport, since the initial rate of [35S] GSH uptake ranged between 1–3% of the total WT transport, and thus the kinetic parameters could not be detected for these mutants. The Km for D335A and H445A did not change significantly compared with WT, but there was a drastic decrease in Vmax (see Supplementary Table S1). The Vmax value for D335A was found to be 12.6 ± 0.3 and, for H445A, was 25.2 ± 0.7 nmol of GSH per mg of protein per minute, suggesting that the mutation did not affect the substrate binding but led to a marked decrease (of two- to five-fold) in the carrier translocation rate. Y374A displayed an approximately two-fold increase in the Km value (55.0 ± 5.2 µM) and a four-fold reduction in Vmax (12.9 ± 1.2 nmol of GSH per mg of protein per minute), suggesting its possible role in both substrate binding and in controlling the translocation rate . Further kinetic analysis of R554A showed a drastic six-fold increase in the Km value (193.7 ± 43.7 µM) and a two-fold reduction in the Vmax value (23.9 ± 2.5 nmol of GSH per mg of protein per minute). The decreased Vmax values obtained for these residues, which could not be attributed to the decreased protein levels for the mutants, suggest that these residues might play some critical role in the proton transport or conformational changes associated with translocation of the substrate by the protein.
To identify the residues critical for proton-coupled GSH transport in Hgt1p in the absence of purified protein or crystal structure has been a difficult task. We have explored two different strategies to address this aspect of GSH translocation. The first strategy attempted to identify mutants that permit pH-independent growth. The rationale behind this approach was that pH-independent growth would yield a pH-independent transporter that was likely to point to the critical proton-binding residues. With this in mind, screening 269 alanine mutants present in the predicted TMDs along with a random mutagenesis strategy to isolate mutants showing pH-independent growth demonstrated two residues, E135 (in TMD 2) and N710 (in TMD 12) to be important for pH-dependent transport by Hgt1p. The observation that E135 repeatedly came up in independent mutagenic screens suggested that this residue was important for this phenotype. Biochemical analysis of these mutants revealed increased uptake of GSH compared with the WT Hgt1p at higher pHs. This appeared to be similar to the mutants of the proton-binding E185 residue of the proton-coupled folate transporter (PCFT), where a defect in transport was observed at lower pHs but there was no change in function at higher pHs . This explains the pH-independent phenotype that was observed. However, surprisingly, the cytosolic acidification studies revealed that the mutants did not render the transporter defective in proton symport. A possible explanation for the phenotypes is that the protonated residues E/D135 or N710 prevented the substrate from binding at higher pHs, and mutation of these residues allowed the substrate to bind in a proton-dependent manner through the involvement of the other residues identified. The findings indicate that the two processes are separable and act as a caution against linking the pH-independent growth phenotype with defective in proton transport.
The second approach explored a more ‘knowledge-based approach’ where information on previous proton-binding residues on similar proton symporters led to evaluating conserved Asp, Glu or His residues, as they have low pKa values and hence can rapidly undergo protonation and deprotonation in these carriers. We also included Tyr residues, and the positively charged Arg and Lys residues, which have been suggested to participate in proton movement through the transporter by pairing with other negatively charged residues to facilitate deprotonation of carboxylic acid during turnover [33,34]. We targeted these amino acid residues depending on their conservation pattern and their location in the predicted topology model. This strategy enabled us to identify E177A (cusp of TMD3), Y193A (TMD3), D335A (ICL5), Y374A (TMD6), H445A (TMD7) and R554 (ICL9) as specifically defective in proton transport since almost no cytoplasmic acidification was observed in response to GSH transport, even at high substrate concentrations (1 mM) or increased time intervals (1 h) as compared with mutants showing similar or low levels of transport, suggesting a defect in proton transport. Although all the residues tested in this strategy for their role in proton binding and transport are based on the conservation pattern in the PT clade, interestingly, only E177 was conserved in the entire OPT family and was present just outside the predicted TMD3 (A179-A200) facing the extracellular side. Hence, E177 is a strong candidate for involvement in the initial proton recognition/substrate binding. Y193, also in TMD3, is predicted to be more towards the cytoplasmic side and hence it may not be the initial proton-binding residue but it is likely to be involved in its relay. However, for both of these residues, the mutation was also completely defective in GSH transport, and thus it was difficult to determine whether the primary defect was in proton binding or in substrate binding. The mutant R554A showed a drastic increase in the Km with a two-fold decrease in the Vmax. This residue is predicted to be present in the long interconnecting loop facing the cytoplasm, and thus is unlikely to be directly involved in binding of a proton. This residue either affects initial proton binding, or may have long-range secondary effects on substrate or proton binding. Residues D335A, Y374A and H445A, which showed no significant change in the Km compared with WT Hgt1p, were severely compromised in the translocation rate suggesting that the mutant transporters are possibly protonated. This would enable the substrate to bind with high affinity and hence these residues may not be involved in direct protonation but are probably required for its release. Thus, while it is clear that these different residues are defective in proton transport, their precise involvement in proton coupling requires more detailed studies on the purified protein and/or into crystal structure determination.
M.Z. and A.K.B. designed the research; M.Z. performed all the experiments, M.Z. and A.K.B. analyzed the data; M.Z. and A.K.B. wrote the paper.
M.Z. acknowledges the Council of Scientific and Industrial Research for a senior research fellowship. A.K.B. is a JC Bose National Fellow. Partial financial support was received from Department of Science and Technology: project no. SB/SO/BB-017/2014 is acknowledged.
The Authors declare that there are no competing interests associated with the manuscript.