Impact of posttranslational modifications of engineered cysteines on the substituted cysteine accessibility method: evidence for glutathionylation

Abstract

The substituted cysteine accessibility method (SCAM) is widely used to study the structure and function of channels, receptors and transporters. In its usual application, a cysteine residue is introduced into a protein which lacks native cysteines following which the accessibility of the residue to the aqueous compartment is assessed. Implicit, and generally assumed, is that if the cysteine-substituted residue is not available to react with sulfhydryl reagents it is not exposed to the extracellular compartment or within the aqueous translocation pathway. We demonstrate here, in a Hela-derived cell line, that some cysteine-substituted residues of the proton-coupled folate transporter (PCFT, SLC46A1) that are inaccessible to 2-((biotinoyl)amino)ethyl methanethiosulfonate are glutathionylated by biotinylated glutathione ethyl ester in the absence of an oxidizing agent. Intramolecular disulfide formation involving cysteine-substituted residues was also identified in some instances. These posttranslational modifications limit the accessibility of the cysteine residues to sulfhydryl-reactive reagents and can have a profound impact on the interpretation of SCAM but may not alter function. When a posttranslationally modified residue is used as a reference extracellular control, the high level of exposure required for detection on Western blot results in erroneous detection of otherwise inaccessible intracellular cysteine-substituted residues. The data indicate that in the application of SCAM, when a cysteine-substituted residue does not appear to be accessible to sulfhydryl-reactive reagents, the possibility of a posttranslational modification should be excluded. The data explain the discrepancies in the assessment, and confirm the localization, of the first intracellular loop of PCFT.

the substituted cysteine accessibility method (SCAM) has been widely utilized to study the structure and function of channels, receptors, and transporters (1). Usually, a cysteine residue is introduced into a protein in which endogenous cysteine residues have been substituted with serine, and the mutant protein is then evaluated for function and accessibility of the cysteine-substituted residue to sulfhydryl-reactive reagents. In particular, a biotinylated reagent allows for pull-down of the cysteine-substituted mutant and assessment of its accessibility by Western blotting. If a sulfhydryl-reactive reagent is used that is impermeable to the cell membrane, the accessibility of the residue to the extracellular compartment can be assessed. The methodology is limited by the requirement that the cysteine-less protein and its daughter cysteine-substituted protein are stable, traffic to the cell membrane and the protein retains sufficient function to allow the analysis of accessibility.

SCAM has been utilized for the characterization of the proton-coupled folate transporter (PCFT, SLC46A1). This transporter mediates the translocation of the family of B9 folate vitamins across the apical membrane of proximal small intestinal cells and is required for the transport of folates across the choroid plexus into the cerebrospinal fluid (20, 28). Failure of these processes, as occurs in the autosomal recessive disorder hereditary folate malabsorption, leads to severe systemic and cerebral folate deficiency (7). This and other laboratories have utilized SCAM to verify the predicted membrane topology of PCFT (Fig. 1), to identify residues that line the aqueous translocation pathway and that are required for function (5, 9, 17, 24, 29, 31, 32, 41). Paired cysteine insertions, with disulfide linkage of adjacent residues, was also utilized to define an external gate for this transporter (38).

Fig. 1.

Human proton-coupled folate transporter (PCFT) membrane topology indicating residues evaluated with the substituted cysteine accessibility method (SCAM) in the current or previous reports. There are six extracellular (EL), and five intracellular (IL) predicted loops. The seven native cysteine residues are highlighted in orange. The residues studied include those in blue predicted to reside in a cytoplasmic loop and those in green predicted to reside in an extracellular loop, with experimental confirmation based on several reports (5, 9, 41). Residues in red are shown, in this report, to be posttranslationally modified when replaced with cysteine. The rectangle encloses the 1st intracellular loop.

Relatively few PCFT mutants that are expressed and traffic to the cell membrane lack intrinsic function. One PCFT domain that is of particular interest in this regard is the first intracellular loop, between the second and third transmembrane helices where three fully conserved residues (Asp109, Gly112, and Arg113) are required for function (14, 23, 31). However, there are conflicting reports, utilizing SCAM, regarding the accessibility of residues in this domain to the aqueous pathway, with the suggestion that this may represent a re-entrant loop (31). Because of the importance of this loop in PCFT function, studies were undertaken to understand the basis for the discrepancy in these findings.

It is assumed for SCAM that if a cysteine-substituted residue is not accessible to membrane-impermeant sulfhydryl-reactive reagents, it is not exposed to the extracellular compartment or aqueous pathway. Similarly, if a cysteine mutant has decreased activity, it is usually attributed to the substitution. However, our previous findings indicated that both assumptions can be incorrect if the engineered cysteine residue is posttranslationally modified (17, 38).The current study extends these findings to demonstrate for the first time in a solute carrier, and in the absence of oxidizing agents, glutathionylation of cysteine-substituted PCFT residues. We demonstrate that modification of cysteine-substituted residues, likely accounted, in part, by glutathionylation, can have a profound impact on the interpretation of SCAM, particularly in topological analyses. These findings verify the topology of PCFT and the intracellular localization of the first intracellular loop as previously reported (5, 41).

MATERIALS AND METHODS

Chemicals.

Key chemicals used in the studies were obtained from following commercial sources: [3′,5′,7-³H(N)]methotrexate (Moravek Biochemicals, Brea, CA); (2-((biotinoyl)amino)ethyl methanethiosulfonate) (MTSEA-biotin; Biotium, Hayward, CA), sodium (2-sulfonatoethyl)methanethiosulfonate (MTSES) and [2-(trimethylammonium)ethyl]methanethiosulfonate bromide (MTSET) (Toronto Research Chemicals, Toronto, ON, Canada), l-ascorbic acid; EZ-Link Sulfo-NHS-LC-Biotin, BioGEE, glutathione oxidized (GSSG), tris(2-carboxyethyl)phosphine (TCEP), and 2-mercaptoethanol (BME) (Thermo Fisher Scientific, Waltham, MA); dl-dithiothreitol (DTT; MP Biochemicals, Solon, OH); glutathione (GSH), cysteine and N-ethylmaleimide (NEM) (Sigma-Aldrich, St. Louis, MO).

Cells and culture conditions.

R1-11 cells that do not express the reduced folate carrier and PCFT were used as recipients for all transient transfections (8, 36). This cell line was obtained from HeLa cells in two steps under methotrexate selective pressure. In the first step, the gene of the reduced folate carrier was deleted, while in the second step, expression of PCFT was silenced (35, 36). R1-11 cell growth is normal in RPMI-1640 medium, which contains a high concentration of folic acid (2.2 µM), despite the lack of these two major folate transporters. R1-11 cells were maintained in RPMI medium supplemented with 10% fetal bovine serum (Gemini Bio Products, West Sacramento, CA), 100 U/ml penicillin and 100 µg/ml streptomycin, and were thawed regularly (once every 3 months) from liquid nitrogen stocks to ensure that PCFT expression was absent (35).

PCFT mutants.

PCFT mutations were introduced in three different templates, all of which were tagged with hemagglutinin (HA) at the COOH-terminus. Mutants generated in wild-type PCFT and cysteine-less-PCFT were identified with the suffixes “WT” or “CL”, respectively. Mutants introduced into a PCFT variant (C66S/C298S-WT) which lacks the disulfide bond were identified by suffix “DSL” (disulfide-bond-less). The properties of S110C-CL, V141C-CL, S174C-CL, G207C-CL, T240C-CL, E261C-CL, E292C-CL, L329C-CL, L357C-CL, T386C-CL, T417C-CL were initially described in a study that defined the PCFT topology (41). Q45C-DSL, and L290C-DSL were utilized in a study to define the PCFT extracellular gate (38). W48C-DSL, W85C-WT, W107C-WT, W202-WT, W213C-WT, and W299C-DSL were studied within the context of the characterization of the role of Trp residues in PCFT function (17). G105C-CL, A106C-CL, S108C-CL, and V111C-CL were generated in the current study using the QuikChange II Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) as reported previously.

Transient transfection.

R-11 cells were seeded in 20 ml Low Background glass scintillation vials (Research Products International, Prospect, IL) for transport studies and seeded in six-well plates for biotinylation studies. Transient transfections were conducted with lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Both transport and biotinylation assays were performed 2 days after transfection unless specified. Cell growth medium was changed 1 day after transfection.

Membrane transport.

Transient transfectants were washed twice at room temperature with HBS (HEPES-buffered saline: 20 mM 4-[2-hydroxyethyl]-1-piperazineethanesulfonic acid, 5 mM dextrose, 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, pH 7.4). HBS was removed and the vials were placed in a 37°C water bath for 1 min before transport was initiated by the addition of 0.5 ml of prewarmed (37°C) MBS (MES-buffered saline: 20 mM 2-(4-morpholino) ethanesulfonic acid, 5 mM dextrose, 140 mM NaCl, 5 mM KCl, 2 mM MgCl_2, pH 5.5) containing 0.5 µM [³H]methotrexate. In this protocol, a 20 min preincubation step in HBS at 37°C usually utilized in transport studies was omitted. Influx was halted after 1 min by the addition of 5 ml of ice-cold HBS. Cells were washed three times with ice-cold HBS and digested in 0.5 ml of 0.2 M NaOH at 65°C for 1 h. Radioactivity in 0.4 ml of lysate was measured on a liquid scintillation spectrometer and normalized to the protein level obtained with the BCA Protein Assay (Pierce, Rockford, IL). All transport assays were conducted at least three times on separate days; each data point was performed in duplicate.

Treatment of transient transfectants with sulfhydryl-reactive and reducing reagents.

All solutions were made up immediately before use. Solutions of reducing reagents (DTT, BME, TCEP, GSH, cysteine, or ascorbate) were made directly in HBS. While MTSES and MTSET were dissolved directly in HBS, MTSEA-biotin and NEM were first dissolved in DMSO as a 100× stock before dilution in HBS. PCFT transient transfectants were incubated with these reagents at room temperature for a specified interval.

Analysis of PCFT at the cell surface and accessibility of PCFT Cys residues by biotinylation.

PCFT accessible at the cell surface or localized at the plasma membrane was assessed with EZ-Link Sulfo-NHS-LC-Biotin which targets Lys residues, while cysteine accessibility was probed with the membrane impermeant sulfhydryl-reactive MTSEA-biotin (38). Transient transfectants were incubated with Sulfo-NHS-LC-Biotin (0.5 mg/ml) or MTSEA-biotin (0.2 mg/ml) in HBS at room temperature for 30 min, and cells were processed in both assays in the same way. Cells were washed twice in HBS at room temperature then overlaid on ice with 0.7 ml ice-cold hypotonic buffer (0.5 mM Na₂HPO₄, 0.1 mM EDTA, pH 7.0) containing protease inhibitor cocktail (Roche, Indianapolis, IN). [We found that washing MTSEA-biotin-labeled cells with a buffer containing 14 mM BME, a step designed to remove residual MTSEA-biotin (31), can also reduce disulfide bond-containing biotinylation product, consequently decreasing the signal, resulting in erroneous MTSEA-biotin labeling.] The cells were then detached from the plates with a disposable cell lifter and centrifuged at 16,000 g for 10 min at 4°C. The pellet was then resuspended in 0.4 ml lysis buffer (50 mM Tris-base, 150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, pH 7.4) and a 50 µl portion was collected as a crude membrane sample for assay of total PCFT expression. The remaining suspension was mixed on a rotisserie for 0.5 to 1 h at 4°C and centrifuged again at 16,000 g for 15 min at 4°C. The supernatant was mixed on a rotisserie overnight at 4°C with 50 µl of streptavidin agarose beads (Thermo Fisher Scientific) that had been prewashed three times with the lysis buffer. The agarose beads were then washed twice with lysis buffer and an additional two times with the lysis buffer containing 2% SDS, each with a 20-min mix on a rotisserie, at room temperature. Protein bound to the beads was released by heating at 100°C for 5 min in 2× SDS-PAGE sample loading buffer (70 µl) with or without DTT.

Protein samples were resolved on 12% or 4–20% SDS-PAGE (Bio-Rad, Hercules, CA). The proteins released from the beads were loaded directly on gels while the crude membrane fractions were mixed (1:1) with the 2× SDS-PAGE sample loading buffer at room temperature before loading on the gels. After SDS-PAGE, proteins were transferred to an Immobilon-P Transfer Membrane (Millipore, Billerica, MA) and were blocked with 10% dry milk in TBST (20 mM Tris, 135 mM NaCl, 1% Tween 20, pH 7.6) overnight at 4°C. The blots were probed with polyclonal anti-HA antibody (Sigma, H6908, 1:4,000 in TBST, 0.1% milk) and a second antibody, anti-rabbit IgG-HRP conjugate (7074S, Cell Signaling Technology, Danvers, MD, 1:4,000 in TBST). The blots were developed with Western Lightning Plus-ECL (PerkinElmer, Waltham, MA).

Assay for glutathionylation.

Expression vectors encoding for the different PCFT mutants were transfected into R1-11 cells then grown in RPMI-1640 medium containing 22 µM BioGEE for 24 h. After being washed twice with HBS, transient transfectants were subjected to pull-down assay with streptavidin beads as described for analysis of PCFT at the cell surface and accessibility of PCFT cysteine residues by biotinylation. For assessment of the effect of DSSG on glutathionylation, cells were grown in medium containing 22 µM BioGEE in the presence or absence of 2 mM GSSG for 24 h.

Statistical analysis.

Statistical analyses were conducted with t-test, GraphPad Software (La Jolla, CA).

RESULTS

Posttranslational modification of the G207C-CL residue and its impact on the SCAM analysis; evidence for localization of PCFT predicted intracellular loops to the intracellular compartment.

Previously, we found that the Cys residues introduced into PCFT at positions Glu45, Trp48, Leu290, and Trp299 were posttranslationally modified in HeLa-derived cells and only reacted with MTSEA-biotin after removal of the modification with DTT reduction (17, 38). The modification resulted in a significant reduction of function of the L290C-DSL mutant (PCFT in which Cys66 and Cys298 were also replaced with serine) that could be fully reversed by DTT. The current study evaluated the status of the G207C-CL mutant employed earlier in a topological analysis of PCFT (41). As indicated in Fig. 2A, the G207C-CL mutant was barely labeled by MTSEA-biotin, a membrane-impermeant sulfhydryl-targeting reagent, under conditions in which labeling of E292C-CL, an established extracellular maker, was robust. DTT restored the labeling of G207C-CL by MTSEA-biotin, consistent with removal of a modifying molecule from the Cys residue. Transport activity of the G207C-CL mutant was only slightly increased after treatment with DTT (18%, P = 0.028). This was in contrast to a fourfold increase in transport function observed for the L290C-DSL mutant, with DTT treatment (Fig. 3). The activity of the DTT-treated G207C-CL mutant was not inhibited by treatment with cysteine-reactive MTSET, MTSES, or MTSEA-biotin. In comparison, the transport function of the DTT-treated L290C-DSL mutant was reduced with MTSEA-biotin, MTSES, or MTSET modification by 53%, 63%, and 84%, respectively. Hence, modification of G207C-CL markedly limits it accessibility to MTSEA-biotin but unlike the L290C-DSL mutant, has little or no functional consequence. Hence, modification of G207C-CL cannot be recognized based on a loss of transport activity.

Fig. 2.

MTSEA-biotin labeling of the substituted cysteine residues in human PCFT. A: MTSEA-biotin labeling of wild-type PCFT (PCFT-WT) and the S110C-CL (cysteine-less) mutant along with the negative control (PCFT-CL) and residues predicted to be extracellular (G207C-CL and E292C-CL) as positive, extracellular, controls. B: MTSEA-biotin labeling of residues previously shown to be extracellularly inaccessible and residues in the 1st intracellular loop. “Pull down” indicates proteins separated by streptavidin beads; proteins in crude membranes are indicated in the lower bands and serve as controls for PCFT expression. The films were subjected to low, medium, and high exposure times. The arrows indicate that G207C-CL transient transfectants were treated with 10 mM dl-dithiothreitol (DTT) for 10 min at room temperature before MTSEA-biotin labeling. The data are representative of at least three independent experiments.

Fig. 3.

The effects of DTT treatment as well as remodification by methanethiosulfonate reagents on transport function of the L290C-DSL and G207C-CL mutants. Transient transfectants were incubated with DTT to remove the modification from the substituted cysteine. After being washed, cells were treated with MTSEA-biotin (0.5 mM), sodium (2-sulfonatoethyl)methanethiosulfonate (MTSES; 3 mM), or [2-(trimethylammonium)ethyl]methanethiosulfonate bromide (MTSET; 3 mM) for 30 min at room temperature. The references set at 100% are the transport activities of the mutants after treatment with 10 mM DTT. Data are means ± SE from three independent measurements.

When the G207C-CL mutant was used as the reference “extracellular” positive control to assess accessibility of cysteine-substituted residues, in or adjacent to the first intracellular loop (G105 to V111), several residues were labeled. This was the basis for the conclusion that this region becomes accessible to the aqueous pathway during the transport cycle (31). WT-PCFT was also found to be labeled in this study. These observations were at variance with several reports in which WT-PCFT was not labeled and residues in this loop were localized solely to the intracellular compartment (5, 9, 17, 29, 39, 41). To explore the possibility that this discrepancy was due to the posttranslational modification of the G207C-PCFT mutant limiting its accessibility to MTSEA-biotin, studies were undertaken to compare labeling in the same cell model in the presence and absence of DTT which would restore the accessibility of the G207C residue to the labeling reagent. Accordingly, MTSEA-biotin labeling was performed in the same experiment for the PCFT-WT, S110C-CL, PCFT-CL, G207C-CL, and E292C-CL variants with the films for Western blot analysis subjected to “low”, “medium”, and “high” exposure times. As indicated in Fig. 2A, expression of each variant in the crude membrane fraction was comparable except for a lower level of expression of PCFT-CL. At the low exposure time, MTSEA-biotin labeling of PCFT-WT and the S110C-CL and PCFT-CL mutants was absent, although labeling of the E292C-CL mutant and the DTT-treated G207C-CL mutant was robust. In the absence of DTT, the G207C mutant was barely detected. When the exposure time was increased to “medium”, labeling of the PCFT-WT, S110C-CL was slight, G207-CL was evident, and DTT-treated G207C and E292-CL were overexposed. At the “high” exposure, both PCFT-WT and S110C-CL proteins were modestly labeled, while G207C-CL labeling was prominent even in the absence of DTT although less when compared with the very dense labeling of the E292C-CL mutant and DTT-treated G207C. Hence, both PCFT-WT and the S110C-CL mutant were not labeled when the reference control was either not modified, E292C-CL, or the modified G207C-CL mutant was treated with DTT. However, both were labeled if the G207C-CL mutant was not treated with DTT and the films were overexposed.

To further evaluate and confirm the topology of the first intracellular loop of PCFT, MTSEA-biotin labeling was assessed for six cysteine-substituted residues within the loop (G105C-CL, A106C-CL, W107C-WT, S108C-CL, S110C-CL, V111C-CL) and for residues previously localized by SCAM to each of the other five predicted intracellular loops comparing the E292C-CL mutant and the G207C mutant, with and without DTT treatment, as extracellular references. As indicated in Fig. 2B, 1) MTSEA-biotin labeling of the G207C-CL mutant was barely detectable in the absence of DTT, but in the presence of DTT it was equivalent to the robust labeling of the E292C-CL mutant at “low” exposure, as observed in Fig. 2A. 2) With the E292C-CL and DTT treated-G207C-CL mutants as references, at the “low” exposure none of Cys mutants located in the intracellular loops was labeled by MTSEA-biotin. 3) As the exposure time was increased, labeling of the G207C-CL mutant appeared. Only under conditions in which the reference bands were overexposed did MTSEA-biotin labeling become visible for all the cysteine-substituted mutants located in cytoplasmic loops. 4) MTSEA-biotin labeling of the E261C-CL, T386C-CL, and S110C-CL mutants was stronger than that of other mutants probably due to the higher level of expression of these mutants.

Further evaluation of the accessibility, and evidence for nonspecific labeling, of residues within the first intracellular loop.

Two additional experimental approaches addressed the location of Cys-substituted residues. 1) If these residues are accessible to the aqueous pathway, treatment first with another membrane-impermeant, water-soluble MTS reagent, such as MTSES, should block subsequent labeling by MTSEA-biotin. As indicated in Fig. 4A, pretreatment with MTSES markedly decreased MTSEA-biotin labeling of the E292C-CL mutant as expected (low exposure), but there was no effect on labeling of the PCFT-WT, S110C-CL, or W107C-WT mutants (high exposure), consistent with nonspecific labeling of these molecules. 2) Labeling was also performed after cells were permeabilized with digitonin so that the MTSEA-biotin could reach the Cys residues otherwise only accessible from the cytosol. Under this condition, the substituted Cys residues exposed to both the extracellular and intracellular aqueous compartments should be labeled with MTSEA-biotin. As indicated in Fig. 4B, the T240C-CL, E261C-CL, S110C-CL, and V111C-CL mutants were as heavily labeled as, if not more than, the G207C-CL and E292C-CL mutants. Interestingly, the G105C-CL, A106C-CL, and S108C-CL mutants were barely labeled above the background level of PCFT-CL, consistent with a location that restricts accessibility to this reagent and nonspecific labeling under conditions of extended exposure times. The PCFT-WT or PCFT-DSL transporters were also labeled owing to the presence of native, otherwise inaccessible, cysteine residues in the molecule, notably Cys21, as previously reported (5). The low level of labeling of PCFT-CL is further evidence that a component of the labeling is nonspecific. Taken together, these observations confirm the cytoplasmic location of the first intracellular loop.

Fig. 4.

Specificity of the weak labeling, and labeling after membrane permeabilization by MTSEA-biotin, of the substituted cysteine residues within the 1st intracellular loop. A: impact of treatment with membrane-impermeant MTSES on MTSEA-biotin labeling. Transient transfectants were treated with 3 mM MTSES, or buffer alone, for 30 min at room temperature before exposure to MTSEA-biotin. B: MTSEA-biotin labeling was performed after transient transfectants were permeabilized with digitonin, (50 µg/ml), for 5 min at room temperature. The vertical line indicates the repositioned lanes. Each panel is a representative image from two separate experiments.

Analysis of effects of reducing reagents on the restoration of L290C-DSL PCFT activity; impact of inclusion in the growth medium.

Since the markedly impaired transport mediated by the L290C-DSL mutant was fully restored after reduction by DTT, this mutant was used to evaluate the potential of other reducing reagents to restore methotrexate transport activity. As indicated in Fig. 5A, none of the reducing reagents at a concentration of 10 mM increased the activity of PCFT-DSL; rather, there was a slight decrease with TCEP (~20%, P = 0.044). In contrast, 10 mM BME, TCEP, restored L290C-DSL activity to a level equal to, or only slightly less than achieved with DTT. GSH was less potent with ~53% restoration of activity compared with DTT. When the concentration of DTT and TCEP was decreased to 1 mM, restoration was as potent as observed at 10 mM but transport activity was substantially decreased at 0.1 mM (Fig. 5B). There was a greater decrease in the ability of the other agents to restore activity as the concentration was decreased. Further studies explored the extent to which the addition of BME, cysteine or GSH to the growth medium preserved function. As indicated in Fig. 5C, the addition of 20 µM, 10 µM, or 5 µM BME restored 62%, 50%, and 33% of the activity of the L290C-DSL mutant. The addition of 150 µM cysteine or 10 µM GSH to the growth medium had no effect. While cysteine and GSH are both present in plasma of healthy adults at these levels (18), BME is not a physiological molecule.

Fig. 5.

Efficiency of reducing reagents in removing the posttranslational modification of the L290C-DSL mutant. A: effect of reducing reagents on restoration of the transport activity of the L290C-DSL mutant with PCFT-DSL (DSL indicates the template in which serine is substituted for Cys66 and Cys298 to eliminate the disulfide bond that would be broken with DTT, making the molecule accessible), as a negative control. Transient transfectants were incubated with 10 mM reducing agents for 10 min at room temperature before [³H]methotrexate influx was measured. The reference is the activity of a mutant treated with 10 mM DTT. B: activation of the L290C-DSL was assessed at lower concentrations (1 and 0.1 mM) of reducing agents. The reference is the activity after 10 mM DTT treatment. The dotted line indicates the activity of L290C-DSL without treatment with a reducing agent. C: the impact of the inclusion of reducing agents in the growth medium on transport activity. Freshly made reducing agents were diluted in growth medium and added to cells immediately after transfection. Since activity was determined 2 days after transfection, the incubation period with reducing agents was also 2 days with one change of medium after 1 day. The cells were treated either with 10 mM DTT or with buffer alone for 10 min before measurement of PCFT function with [³H]methotrexate. The control in this experiment was the transport activities of transfectants neither treated with a reducing reagent in the growth medium nor with DTT afterwards. For all panels, data are means ± SE from three separate measurements.

Nature of the posttranslational modification of Cys-substituted residues as analyzed by nonreduced SDS PAGE: evidence for disulfide bond formation.

A common posttranslational modification is the formation of a disulfide bond between two cysteine residues. Such is the case for the endogenous Cys66 and Cys298 residues in extracellular loops that form a cross-link in PCFT-WT (41). Since the G207C mutation was introduced into the PCFT-CL, the formation of an intramolecular disulfide bond was excluded. Intramolecular disulfide formation was also excluded for the L290C-DSL and Q45C-DSL mutants since the posttranslational modifications persisted even when the mutations were introduced into the Cys-less background (38). The possibility of formation of a disulfide bond between two monomers was probed utilizing a DTT-free sample loading buffer for the G207C-CL, L290C-DSL, and Q45C-DSL mutants. Two other mutants, W48C-DSL and W299C-DSL, were also included since they are also posttranslationally modified (17). Using a DTT-free sample loading buffer, or under a nonreducing condition, only one distinct band between molecular masses of 58 and 75 kDa was detected for the Q45C-DSL, W48C-DSL, L290C-DSL mutants along with the PCFT-DSL control. (Fig. 6A). However, a high-molecular band (close to 135 kDa) was detected under the same conditions for the G207C-DSL and W299C-DSL mutants, consistent with formation of a cross-link between two PCFT monomers. These bands were detected for both the pull-down fraction, representing PCFT expressed at the plasma membrane, and for the crude membrane fraction. Further experiments were also performed to determine if the formation of a dimer could be reversed with DTT reduction. This was achieved by mixing the protein sample in DTT-free sample loading buffer with DTT-containing sample loading buffer at room temperature followed by reexamination of the migration of proteins on the gel. For both the G207C-CL and W299C-DSL mutants, the formation of a dimer was partially reversible with DTT reduction in the crude membranes but not in the pull-down fraction.

Fig. 6.

SDS-PAGE analysis of dimer formation of the PCFT mutants posttranslationally modified. A: PCFT expressed on the plasma membrane in transient transfectants was labeled with lysine-targeted EZ-Link Sulfo-NHS-LC-Biotin and pulled down with streptavidin beads. DTT-free sample buffer was used to mix the crude membrane fractions at room temperature and to strip proteins from the beads at 100°C for 5 min. The protein samples were then mixed with equal volumes of either DTT-free sample buffer (DTT−) or DTT-containing sample buffer (DTT+) for at least 30 min at room temperature before being loaded on the gel. B: transient transfectants were treated successively with DTT, N-ethylmaleimide (NEM), and EZ-Link Sulfo-NHS-LC-Biotin before biotinylated PCFT was pulled down with streptavidin beads. DTT-containing or DTT-free sample buffer were used to process the crude membranes and the beads. Each panel is a representative image from at least two separate experiments.

Since the preparation of the pull-down samples included a step in which the streptavidin beads were heated in DTT-free sample buffer for 5 min at 100°C, this raised the possibility that heating may have resulted in the formation of irreversible cross-linkage between the two monomers.

As indicated in Fig. 6B, the dimer was observed in pull-down, but not in the crude membrane, samples for the G207C-CL and W299C-DSL mutants when DTT-containing sample loading buffer was utilized, consistent with the formation of an irreversible cross-link due to the heating. In further experiments with the G207C-CL and W299C-DSL mutants, the disulfide bond was first broken by DTT reduction in intact cells, and the newly generated sulfhydryl groups were blocked by NEM to prevent the linkages from reforming. As indicated in Fig. 6B, without any treatment the PCFT dimer band was clearly detected again for both G207C-CL and W299C-DSL mutants but was absent for PCFT-DSL. Treatment of cells with NEM in the absence of DTT reduction did not alter the pattern of migration, consistent with the preservation of the disulfide bond. Treatment with DTT alone only slightly decreased the intensity of the dimer band, suggesting that the disulfide bond was largely re-formed during sample processing after DTT was removed. On the other hand, successive treatment with DTT and NEM resulted in a marked reduction in dimer formation in both the pull-down and crude membrane samples. Hence, disruption of the disulfide bond not only breaks the cross-link between two monomers as expected but also prevents heating-related irreversible cross-linkage between two monomers for both G207C-CL and W299C-DSL mutants.

Glutathionylation of the PCFT mutants.

Lack of intermolecular disulfide formation for the Q45C-DSL, W48C-DSL and L290C-DSL and lack of possible intramolecular disulfide formation for the Q45C-DSL and L290C-DSL mutants suggested the possibility that these mutants were modified by endogenous sulfhydryl reacting molecule. Glutathione is the most abundant thio-containing molecule within cells. To evaluate the potential modification of Cys-substituted residues by glutathione, a biotinylated glutathione ester, BioGEE, was utilized for the detection of glutathionylation of PCFT Cys-mutants. BioGEE rapidly diffuses into cells but becomes impermeant and retained after the ester moiety is cleaved (25). If a PCFT mutant is glutathionylated, biotinylated glutathione should be covalently linked to PCFT and precipitated by streptavidin beads. As indicated in Fig. 7A, by this analysis substantial glutathionylation was identified for the Q45C-DSL, W48C-DSL, and L290C-DSL mutants, lesser glutathionylation was observed for the G207C-CL and W299C-DSL mutants, and no glutathionylation could be detected for PCFT-DSL. The signals obtained with BioGEE were much weaker than those obtained from biotinylation of PCFT at the plasma membrane (last lane). This is presumed to be due to dilution of the BioGEE (22 µM) by the much higher intracellular GSH concentration (1–10 mM) depending on the cell type and physiological setting (6).

Fig. 7.

Glutathionylation of the substituted cysteine residues in PCFT-mutants. Panel A: Glutathionylation of PCFT mutants that were found to be posttranslationally modified. After transient transfection, cells were grown in RPMI-1640 medium containing 22 µM BioGEE for 24 h before cells were processed for pull-down assay with streptavidin beads. The last lane of this panel serves as a quantitative control in which PCFT expressed on the plasma membrane (P.M.) was labeled with EZ-Link Sulfo-NHS-LC-Biotin. B: the effects of DTT on the removal of the glutathionylation modification from PCFT mutants. Transient transfectants were labeled with BioGEE for 24 h and then treated with DTT (DTT+) or buffer alone (DTT−) before pull-down assays were performed with streptavidin beads. C: the effects of addition of GSSG on biotinylation of PCFT by BioGEE in PCFT mutants. After transfection, cells were grown in medium containing 22 µM BioGEE in the presence or absence of 2 mM GSSG for 24 h. D: glutathionylation assays in a wider spectrum of PCFT mutants and PCFT-WT. For all panels, DTT-free sample buffer was used to process all samples and the images shown are representative of two independent analyses.

As indicated in Fig. 7B, following DTT treatment of transient transfectants, the glutathionylation signals were all markedly decreased except for the W48C-DSL mutant. It is unclear if modification by GSH or biotinylated GSH has the same vulnerability to DTT reduction. The addition of 2 mM GSSG, the physiological glutathionylation species, markedly suppressed biotinylation with BioGEE for all five PCFT mutants (Fig. 7C), confirming that the modification induced by BioGEE was glutathionylation. Because glutathionylation was observed in all PCFT mutants that were previously identified as posttranslationally modified, other mutants in or close to extracellular loops that were not found to be modified were studied. As indicated in Fig. 7D, a low level of glutathionylation was detected in only one mutant, W202C-WT. No glutathionylation was observed for PCFT-WT, reflecting the strength of the Cys66-Cys298 bond in the native molecule.

Lack of nitrosylation of the PCFT mutants.

Nitrosylation is another common modification of cysteine residues that can be removed by ascorbate, a weaker reducing agent than DTT (10, 12, 21). Incubation of L290C-DSL transient transfectants with 20 mM ascorbate at room temperature for 1 h did not significantly alter the transport activity of this mutant while treatment of the same cells with 10 mM DTT at room temperature for 10 min augmented the activity by 5.2 ± 0.2-fold (based on three separate experiments). This excludes a role for nitrosylation in the posttranslational modification of the L290C-DSL mutant.

DISCUSSION

This report provides evidence for glutathionylation as a likely basis for the modification of engineered cysteine-substituted residues in a solute carrier. This modification, along with dimerization, blocks the substituted cysteine sulfhydryl group eliminating its chemical reactivity with sulfhydryl-reactive reagents. This can occur with or without a change in transport function and, as described in this report, can lead to misleading results in SCAM analyses if it is not recognized. This turned out to the basis for the reported discrepancies in the localization of the first intracellular loop of PCFT (5, 31, 41).

The first intracellular loop, between the second and third transmembrane helices, plays an important role in PCFT function. Three fully conserved residues in this domain, Asp109, Gly112, and Arg113, are required for function (14, 23, 31). Loss-of-function mutations involving the Arg113 residue were identified in two subjects with hereditary folate malabsorption (15, 37). Studies in this laboratory with HeLa cells localized the Ser110 and Trp107 residues in this domain to the cytoplasmic compartment (17, 41). Although W107C-CL is not expressed, both W107C-WT and W107C-DSL are expressed but are not labeled with MTSEA-biotin (17). Likewise, studies in which PCFT was expressed in Xenopus oocytes localized residue(s) in this domain to the cytoplasm; the A106C-CL, V111C-CL PCFT mutants could only be labeled by MTSEA-biotin when PCFT-expressing oocytes were permeabilized (5). However, in another report, residues in this loop were found to be accessible to MTSEA-biotin, leading to the suggestion that this is a re-entrant loop that enters the aqueous channel during the transport cycle (31). That study utilized the G207C-CL mutant as the extracellular reference control. The current study demonstrates that 1) the G207C-CL mutant is posttranslationally modified owing to intramolecular dimerization and to a lesser extent glutathionylation and 2) when the G207C-CL mutant is used as the reference extracellular control, the duration of exposure required to detect the modified residue is so long that cysteine mutants in every intracellular loop are detected. This report also verifies that PCFT-WT is not labeled by MTSEA-biotin when the exposure time is appropriate so that SCAM can be performed on this scaffold, obviating the difficulty that occurs when mutations introduced in PCFT-CL result in the lack of protein expression (40).

The formation of a disulfide bond between two monomers, as demonstrated for the G207C-Cl and W299C-DSL PCFT mutants, is readily detected under nonreducing conditions on SDS-PAGE, since the protein migrates as a dimer under these conditions. Formation of an intramolecular disulfide bond between the substituted cysteine residue and a native cysteine residue is more difficult to detect because of the lack of an apparent change in mobility on the gel. Although this possibility can be excluded by introducing the cysteine mutation in a cysteine-less template, a cysteine-less protein can be vulnerable to further mutation as demonstrated for PCFT and other transporters (27, 40). Glutathionylation of the substituted residue is even more difficult to recognize since migration on SDS-PAGE is not altered and it occurs with a mutation introduced into a cysteine-less template. As demonstrated, both modifications can be excluded by treatment with DTT before probing the accessibility of the substituted cysteine with MTSEA-biotin. This requires the absence of the native disulfide bond within the molecule that would be broken by DTT, an analysis possible for PCFT by utilization of a mutant that lacks the two cysteine residues located in the first and fourth extracellular loops (38).

Evidence for glutathionylation come from results with BioGEE which has been utilized to document S-glutathionylation of various proteins, such as aquaporin-2 (26), K_ATP channel (33), histone H3 (11), annexin A2 (3), p21Ras (4), Rpn2 regulatory subunit (42), and NF-κB kinase-β (22). BioGEE rapidly enters cells and is retained intracellularly when the ester is hydrolyzed. It is incorporated into protein only as a consequence of cysteine thio-oxidation and is readily detected by pull-down of its biotin moiety on streptavidin beads (25). Additional evidence that the modification detected by BioGEE represents glutathionylation comes from the observations that it is inhibited by GSSG, the physiological glutathionylation species.

BioGEE-mediated glutathionylation recapitulated the pattern of modification of cysteine-substituted residues observed in the absence of this agent. All cysteine-substituted residues that were translationally modified in PCFT were found to be glutathionylated by BioGEE. Hence, only cysteine residues introduced at 45, 48, 207, 290, and 299, and to a lesser extent at 202, were modified by BioGEE; cysteine residues introduced at position 85, 141, 213, 292, 357, 417 were not. These glutathionylated residues are all located at, or in, a transmembrane helix in proximity to the extracellular interface. Nor was there glutathionylation of the PCFT templates (PCFT-WT, PCFT-DSL) used in these experiments although the native cysteine residues are present in these molecules. The basis for this selectivity is not clear but may be due to a more basic microenvironment with greater reactivity in the former locations (12). Glutathionylation is likely to occur in the oxidative environment of the endoplasmic reticulum where glutathione S-transferase is present (13, 34). The weak glutathionylation of the G207C-CL and W299C-DSL mutants is consistent with the partial intramolecular cross-link observed for these residues. The data indicate that modification of the Cys-substituted residues can be suppressed by the inclusion of a reducing agent (BME) in the growth medium. Hence, consideration should be given to incorporating this into the SCAM protocol.

S-glutathionylation is a common posttranslational modification of cysteine residues and occurs through the formation of a mixed disulfide bond between the target protein and glutathione (12). One key feature of this modification is its reversibility based on the cellular oxidative potential (2). S-glutathionylation reduces the functions of diverse transport proteins such as, vascular K_ATP channel (33), aquaporin-2 (26), cystic fibrosis transmembrane conductance regulator (CFTR) (30), and Na-K-ATPase (19). Modified cysteine residues were always accessible from the cytosol and thus in contact with intracellular glutathione and S-glutathionylation required an oxidative environment induced by oxidants. In contrast, S-glutathionylation of PCFT occurred in the absence of oxidants; indeed, this occurred even when cells were grown in low concentrations of BME, and involved cysteine-substituted residues exposed to the extracellular compartment. While a substituted cysteine residue in CFTR (T338C) was also modified in the absence of an oxidizing environment (16), the nature of the modification was not determined.

GRANTS

This work was supported by National Institutes of Health National Cancer Institute Grant CA-82621.

DISCLOSURES

Conflict of interest statement: No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

R.Z., M.N., S.A., and I.D.G. conceived and designed the research; R.Z., M.N., and S.A. performed experiments; R.Z., M.N., S.A., and I.D.G. analyzed data; R.Z., M.N., S.A., and I.D.G. interpreted results of experiments; R.Z. and M.N. prepared figures; R.Z. drafted the manuscript; R.Z. and I.D.G. edited and revised the manuscript; R.Z., M.N., S.A., and I.D.G. approved the final version of the manuscript.