Substituted-cysteine accessibility and cross-linking identify an exofacial cleft in the 7th and 8th helices of the proton-coupled folate transporter (SLC46A1)

Abstract

The proton-coupled folate transporter (PCFT-SLC46A1) is required for folate transport across the apical membrane of the small intestine and across the choroid plexus. This study focuses on the structure/function of the 7th transmembrane domain (TMD), and its relationship to the 8th TMD as assessed by the substituted cysteine accessibility method (SCAM) and dicysteine cross-linking. Nine exofacial residues (I278C; H281C–L288C) of 23 residues in the 7th TMD were accessible to 2-((biotinoyl)amino)ethyl methanethiosulfonate (MTSEA-biotin). Pemetrexed, a high-affinity substrate for PCFT, decreased or abolished biotinylation of seven of these residues consistent with their location in or near the folate binding pocket. Homology models of PCFT based on Glut5 fructose transporter structures in both inward- and outward- open conformations were constructed and predicted that two pairs of residues (T289-I304C and Q285-Q311C) from the 7th and 8th TMDs should be in sufficiently close proximity to form a disulfide bond when substituted with cysteines. The single Cys-substituted mutants were accessible to MTSEA-biotin and functional with and without pretreatment with dithiotreitol. However, the double mutants were either not accessible at all, or accessibility was markedly reduced and function markedly impaired. This occurred spontaneously without inclusion of an oxidizing agent. Dithiotreitol restored accessibility and function consistent with disulfide bond disruption. The data establish the proximity of exofacial regions of the 7th and 8th TMDs and their role in defining the aqueous translocation pathway and suggest that these helices may be a component of an exofacial cleft through which substrates enter the protein binding pocket in its outward-open conformation.

INTRODUCTION

The proton-coupled folate transporter (PCFT-SLC46A1) mediates the intestinal absorption of folates at the apical membrane of the proximal small intestine (17). PCFT is also required for the transport of folates across the choroid plexus (27). In the absence of this transporter, as occurs in the autosomal recessive disorder, hereditary folate malabsorption, there is severe systemic folate deficiency and the absence of folate in the cerebrospinal fluid (7, 27). Antifolates are also substrates for PCFT; most notable are pemetrexed (32) and a novel class of purine synthesis inhibitors that have high affinities for this transporter (3, 9). A variety of studies have characterized the secondary structure of PCFT and residues that play a role in folate and proton binding, proton coupling, and oscillation of the carrier between its inward- and outward- open conformations (22, 23, 26–29).

Studies have identified transmembrane domains (TMDs) that come together to form an external gate during carrier cycling (31). This was established by cross-linkage of Cys-substituted exofacial residues involving the 1st (Gln45), 2nd (Asn90), 7th (Leu290), and 11th (Ser407 and Asn411) transmembrane domains. Most recently, using the substituted-cysteine accessibility method (SCAM) in an analysis of the 8th TMD, 14 contiguous exofacial residues (L303–L316), of the 24 total residues in this helix, were found to be accessible to the extracellular compartment consistent with their location in the aqueous translocation pathway (1). Further, multiple residues extending deep into the pathway were located within or near to, the folate binding pocket. In the current report, SCAM was used to characterize the 7th PCFT TMD demonstrating a similar but not as extensive contiguity of accessible exofacial residues many of which are located in or near to the folate binding pocket based on folate-substrate protection of residues located in the aqueous translocation pathway. Homology models of PCFT, based on the Glut5 fructose transporter structures in both the inward- and outward-open configurations (15), predicted that exofacial regions of the 7th and 8th transmembrane domains (TMDs) are in proximity and that in the inward-open conformation, residue pairs, if Cys-substituted, should be sufficiently close to allow the formation of a disulfide bond. This was confirmed by the cross-linkage of two Cys-substituted residue pairs between the two helices. These observations suggest that the exofacial regions of both helices likely participate in the formation of a deep extracellular cleft into, and line, the aqueous translocation pathway.

MATERIALS AND METHODS

Chemicals.

[3′,5′,7′-³H(N)]methotrexate (MTX) and generally labeled [³H]pemetrexed were purchased from Moravek Biochemicals (Brea, CA). Sulfosuccin-imidyl-6-(biotinamido) hexanoate (EZ-Link sulfo-NHS-LC biotin) and streptavidin agarose beads were obtained from Fischer Scientific (Pittsburgh, PA), and protease inhibitor cocktail from Roche Applied Science (Mannheim, Germany). N-biotinylaminoethyl methanethiosulfonate (MTSEA-biotin) was purchased from Biotium (Hayward, CA).

Cell lines and culture conditions.

HeLa-R1-11 cells, which lack PCFT and RFC expression because of epigenetic silencing of the former and a genomic deletion of the latter, were the transfection recipient (31) Cells were grown and maintained in RPMI 1640 medium containing 10% fetal bovine serum (Gemini Bio-Products, Irvine, CA), 100 U/ml penicillin, and 100 µg streptomycin under 5% CO₂. For transport studies, 3 × 10⁵ cells were seeded in 17-mm glass vials; 5 × 10⁵ cells/well were seeded in six-well plates for Western blot analyses. After 48 h, the cells were transfected with 0.8 µg DNA/vial of the various PCFT constructs for the former, and 1.6 µg DNA/well for the latter, studies. Two percent lipofectamine 2000 (Invitrogen, Carlsbad, CA) in serum- and antibiotic- free RPMI-1640 was employed for transfections.

Site-directed mutagenesis; utilization of a PCFT construct that lacks an extracellular dicysteine bond.

Residues in the 7th and 8th TMDs were individually substituted with Cys using the QuikChange II XL Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). An expression vector, pcDNA3.1(+), contains a COOH-terminal hemagglutinin (HA) tag on a PCFT template that lacks cysteine residues in the 1st (Cys66) and 4th (Cys298) TMDs, which spontaneously form a disulfide bond in wild-type PCFT (PCFT-WT). This construct retains nearly normal function (31) and allows the use of dithiothreitol (DTT) to reverse modifications of Cys-substituted residues. This is required for the cross-linking protocol in which it is necessary to break the disulfide bond that may have formed between Cys-substituted residues on the 7th and 8th TMDs to confirm the intrinsic function and accessibility of a PCFT construct which harbors both Cys substitutions. The presence of the disulfide bond in PCFT-WT would render uninterpretable studies in which accessibility of Cys-substituted pairs was assessed with MTSEA-biotin before and after treatment with DTT. PCFT is unique in that none of the remaining five cysteine residues in the transporter are accessible to MTSEA-biotin, obviating the requirement for use of a Cys-less scaffold that is fragile and has a limited tolerance to additional mutations (33). The open reading frame of all PCFT mutants was verified at the Albert Einstein Cancer Center Genomic Shared Resource (Bronx, NY).

Analysis of initial uptake rates.

Influx measurements were made in cells growing in monolayer at the bottom of glass vials. Forty-eight hours after transfection, the growth medium was aspirated followed by two washes with HBS buffer (20 mM HEPES, 140 mM NaCl, 5 mM KCL, 2 mM MgCl₂, and 5 mM dextrose, at pH 7.4). HBS (2 ml) was then added to each vial and the cells were preincubated for 20 min at 37°C. The buffer was then aspirated and 0.5 ml of MBS buffer (20 mM MES, 140 mM NaCl, 5 mM KCL, 2 mM MgCl2, and 5 mM dextrose) containing the tritiated substrates was added and influx continued for 1 min. Influx was halted by addition of 5 ml of ice-cold HBS buffer (pH 7.4), and the cells were washed thrice with the same buffer and then digested with 0.5 ml of 2 N NaOH at 65°C for 1 h. A portion of the digest (0.4 ml) was assayed for tritium on a liquid scintillation spectrometer; 20 µl of the digest was analyzed by the bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Waltham, MA). Cell MTX or pemetrexed influx is expressed as picomoles per mg of protein per minute and in some cases as percentage of PCFT-DSL. For transport studies following cysteine-cross-linking, cells were washed twice with HBS, incubated at 37°C for 1 min, and influx measurements were made as described above. In some experiments, cells were treated with 12 mM dithiotreitol (DTT) for 10 min at room temperature then washed twice followed by influx measurements.

Analysis of PCFT surface expression and MTSEA biotin accessibility of Cys-substituted residues.

The extent of PCFT expression at the cell surface was evaluated by biotinylation of accessible lysine residues at the external interface of the protein 48 h posttransfection. Cells were washed twice with HBS buffer (pH 7.4), treated with EZ-Link sulfo-NHS-LC biotin in HBS buffer for 30 min at room temperature, then washed twice with HBS buffer, followed by the addition of 0.7 ml hypotonic buffer (0.5 mM NaHPO₄ and 0.1 mM EDTA, pH 7.0) containing protease inhibitor cocktail. The cells were then scraped off the plates, centrifuged at 21,000 g for 20 min at 4°C and the pellet resuspended in 0.4 ml of lysis buffer (0.1% SDS, 1% Triton X-100, 1 mM EDTA, and 150 mM NaCl, and 20 mM Tris, pH 7.4). Fifty microliters of the supernatant was mixed with an equal volume of 2× Laemmli buffer containing DTT for analysis of PCFT expression in the crude membrane extract. The remainder of the supernatant was rotated on a rotamix for 1 h at 4°C then centrifuged at 21,000 g for 15 min at 4°C. The supernatant was collected and mixed with 50 µl streptavidin-agarose beads (prewashed in lysis buffer) then rotated overnight on a rotamix at 4°C. The beads were washed twice with 0.5 ml lysis buffer and twice with this buffer containing 2% SDS. Then, 75 µl of 2× Laemmli buffer (containing DTT) was added to the beads, heated at 95°C for 5 min, centrifuged at 21,000 g for 3 min, and the supernatant was collected. Fifteen microliters of the crude membrane preparation, and 15 µl of biotinylated pulldown sample, were then loaded onto 4–12% SDS-PAGE. Membrane-impermeable MTSEA-biotin was used to assess the accessibility of the Cys-substituted residues in the mutant PCFT constructs following a procedure similar to that described for biotinylation of the lysine residues described above. In this case, MTSEA-biotin was dissolved in 2 mg/100 µl DMSO and then diluted 1:100 with HBS buffer, pH 7.4 (final concentration, 0.5 mM). The cells were washed twice with HBS and incubated with 1 ml of MTSEA-biotin in HBS for 30 min at room temperature or at 4°C following which the cells were processed as described for lysine biotinylation. To evaluate whether Cys-substituted residues were located within or near to the folate binding pocket, the extent to which pemetrexed blocked biotinylation was assessed. This antifolate was used because its affinity for PCFT far exceeds that of the physiological folates and MTX and its activity is better retained as the pH is increased into the neutral range (32). Cells were washed twice with HBS and treated with either HBS or HBS containing 1 mM pemetrexed at 4°C before MTSEA-biotin labeling.

Gel electrophoresis and Western blot analysis.

Protein samples were resolved on 4–12% polyacrylamide gels (Bio-Rad, Hercules, CA) followed by electroblotting onto polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, MA). After electroblotting, PVDF membranes were blocked with 10% nonfat dry milk in Tris-buffered saline-Tween (20 mM Tris base and 135 mM NaCl, 0.1% Tween 20, pH 7.4). The membranes were washed in this buffer, probed with anti-HA antibody (Sigma, St. Louis, MO) at room temperature for 90 min followed by a 1 h treatment with horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology, Danvers, MA) under rocking conditions. Blots were developed with ECL Lightning Plus reagent (Perkin Elmer, Shelton, CT). Actin was used as an internal loading control and assessed with a rabbit anti β-actin antibody (Cell Signaling Technology).

Homology models.

A prior model of PCFT was based on the structure of the Escherichia coli glycerol-3-phosphate proton-coupled transporter, GlpT, in the inward-open conformation (10, 31). The recently reported structures of the bovine (inward-open) and rat (outward-open) Glut5 fructose transporters (15) [Protein Data Bank (PDB) codes 4YB9 and 4YBQ, respectively] provided an opportunity to develop homology models of PCFT in both conformations (1). The two Glut5 transporters share 88% sequence identity with each other and 13% sequence identity to PCFT similar to level of identity of GlpT (10, 31). The properties of these models were recently described in detail (1). Briefly, the optimal alignment between PCFT and the Glut5 template structures and the subsequent molecular model was obtained from MMM alignment optimization and comparative modeling (5, 18, 19). The quality of the model was verified through energetic analysis using statistical pair potentials implemented in Prosa (21). Optimal superposition of the two models was obtained using the Align3D program of Modeller (6).

Statistical analysis.

The data were analyzed by a two-tailed paired t-test using GraphPad Prism7.

RESULTS

Cysteine scanning mutagenesis of the 7th TMD.

Based on the established PCFT topology (2, 4, 30, 34), the 23 residues from His266 to Leu288 make up the 7th TMD (Fig. 1). To explore the accessibility of residues within this domain, single cysteine substitutions were made for each residue in a PCFT scaffold (PCFT-DSL) that lacked the Cys66 and Cys298 residues in the 1st and 4th extracellular loops, respectively, to eliminate the disulfide bond that forms in PCFT-WT, a necessary prerequisite for subsequent studies as described in materials and methods. As indicated in Fig. 2A, all of the Cys-substituted residues were expressed in the crude membrane fraction when the constructs were transfected into HeLa R1-11 cells, which lack endogenous PCFT or RFC. Likewise, all of the constructs trafficked to the cell membrane as assessed by biotinylation of lysine residues at the cell surface. However, the levels of expression were substantially decreased for the Y270C, V277C, V280C, F282C, and G283C mutants.

Fig. 1.

The secondary structure of PCFT as established by SCAM (2, 4, 34). Cys-substituted residues that span the 7th TMD are denoted in violet. Residues that span the 8th TMD are indicated in green (1).

Fig. 2.

The effect of cysteine substitutions on PCFT expression and function. A, top band: expression of Cys-substituted PCFT mutants at the cell membrane as assessed by biotinylation of native lysine residues. Middle band: expression of Cys-substituted PCFT mutants in the crude membrane extract. Bottom band: expression of β-actin, the loading control. The blot is representative of three independent experiments. The vertical line indicates two separate gels run at same time. B: influx mediated by the Cys-substituted mutants over 1 min at 0.5 µM and 50 µM [³H]MTX at pH 5.5 and 37°C. Data are percentage ± SE of PCFT-DSL from 3 independent experiments. *P < 0.05, **P < 0.01, ****P < 0.0001.

Transport function was assessed for all of the mutants at a concentration far below, (0.5 µM), and at least an order of magnitude above, 50 µM, the influx K_t for [³H]MTX at pH 5.5, the latter concentration approximating the influx V_max. All the Cys-substituted mutants that retained expression comparable to PCFT-DSL retained at least 50% of function at the low and/or high concentrations except for H281C and D286C, for which function was markedly decreased, (Fig. 2B). The His281 residue was previously reported to play an important role in proton binding (22) and a previous study showed that the D286K mutation was well tolerated (20). A decreased in function paralleled the decrease in expression for the Y270C, V277C, V280C, and F282C mutants. Transport activity of several of the mutants was modestly increased (≤2-fold) at the high saturating concentration, reaching 3-fold for the I287C mutant. This is also likely the case for the G283C mutant, for which influx was comparable to that of PCFT-WT at the high concentration but expression was markedly reduced. For most of these mutants, transport at the low MTX concentration was unchanged, which, in view of the high V_max, is consistent with an increase in the influx K_t as observed previously for a variety of PCFT mutants (1, 26). The G283C mutant V_max is increased when influx is corrected for the low level of expression.

Accessibility of Cys-substituted residues; impact of pemetrexed.

The accessibility of Cys-substituted residues to MTSEA-biotin (a lipid-impermeable reagent) was first assessed at room temperature to preserve oscillation of the carrier between its conformational states (Fig. 3A). Residues from His266-Val277, along with T279C and V280C, were not labeled. Nor was PCFT-DSL labeled consistent with the lack of accessibility of the remaining five Cys residues as is also the case for all the native Cys residues in PCFT-WT (31, 34). Residues from H281C-L288C were labeled along with I278C. The labeling pattern was not altered even after the cells were treated with DTT excluding modification of Cys residues by endogenous Cys-reacting moieties as a basis for the lack of accessibility, as previously described for some Cys-substituted residues (30, 31). Hence, eight contiguous exofacial residues along with I278C were accessible to the aqueous compartment. To investigate the role of the accessible residues in folate binding, the impact of pemetrexed (a high-affinity PCFT substrate) on MTSEA-biotin labeling was assessed at 4°C (Fig. 3B). Pemetrexed markedly suppressed biotinylation of all the residues except Q285C and I287C; labeling was modestly decreased for the L288C mutant more prominently visualized when the exposure time was decreased (not shown). Since D286C was not accessible at 4°C, unlike the finding at room temperature, studies were undertaken to assess pemetrexed protection of this residue at room temperature. Under these conditions, pemetrexed blocked MTSEA-biotin labeling (Fig. 3C). Taken together, pemetrexed alters the accessibility to I278C, H281C, F282C, G283C, A284C, D286, and likely L288C residues.

Fig. 3.

MTSEA-biotin accessibility of Cys-substituted residues in the 7th TMD; impact of pemetrexed. A: cells were treated with 0.5 mM MTSEA-biotin in HBS for 30 min at room temperature (RT). The pull-down fraction is indicated in the upper band; protein expression in the crude membrane extract is indicated in the bottom band. B: accessibility of Cys-substituted residues to MTSEA-biotin (0.5 mM) at 4°C in the presence or absence of 1 mM pemetrexed at pH 7.4. C: accessibility of the D286C mutant at both 4°C and RT in the presence or absence of pemetrexed as described in B. The Western blot is a representative of 3 independent experiments.

Homology modeling of the 7th TMD and its relationship to the 8th TMDs.

Nine residues, of which eight are contiguous; in the exofacial region of the 7th TMD are accessible to the aqueous compartment. An even greater degree of exofacial accessibility, 14 residues, was observed for the 8th TMD (1), suggesting that both helices are important components of the aqueous translocation pathway. To explore the relationship between these helices, a homology model was utilized based on the structures of the rat and bovine Glut5 fructose transporters in the inward-open and outward-open conformations, respectively (15). The properties of this model were recently reported (1) and are described briefly in materials and methods. The upper models of Fig. 4 illustrate the predicted conformations of the 7th TMD (highlighted) in the inward- open (blue) and outward-open (green) conformations with the Leu288 and Gln285 side-chains depicted from two different perspectives. It can be seen that there is a break in the helix in both conformations. There is a striking rotation of the exofacial region reflected by the substantial change in spatial orientation of the Gln285 and Leu288 residues that may account for the uninterrupted accessibility of eight Cys-substituted residues in this helix. The lower models of Fig. 4 illustrate two pairs of residues between the 7th and 8th TMDs predicted from homology modeling to be in sufficient proximity to allow the formation of a disulfide bond when substituted with cysteine residues: Gln285 (7th TMD) and Gln311 (8th TMD) and Thr289 (7th TMD) and Ile304 (8th TMD), the latter in proximity to the external interface of their respective helices. Both contacts were made in the inward-open conformation when the external gate is closed and the exofacial helices would be most closely aligned.

Fig. 4.

Homology modeling of PCFT in the inward- and outward-open conformations based on the reported structures of the Glut5 fructose transporter (15). A–C: illustration of the PCFT models in both conformations with the 7th TMD highlighted in dark blue (inward-open) and dark green (outward-open). Also shown is the change in spatial orientation of Gln285 and L288 residues in the 7th TMD in both conformations. The aqueous translocation pathway is indicated by an arrow. The two left models (A and B) are planar views into the membrane. The right model (C) is a view into the aqueous translocation pathway (indicated by the arrow) from the extracellular space. D and E: illustration of two planar views of the residues in the 7th and 8th TMDs in both conformations that are predicted to be in sufficiently close proximity to allow disulfide bond formation between the Q285C-Q311C and T289C-I304C residue pairs.

Cross-linking of Cys-substituted residues between the 7th and 8th TMDs.

The dicysteine mutant pairs (Q285C-Q311C and I304C-T289C), predicted to be in cross-linkage distances, were constructed in the PCFT-DSL scaffold and were assessed for accessibility and function. As indicated in Fig. 5A, all of the single and double mutant constructs were expressed at comparable levels in the crude membrane preparation. The extent of biotinylation of the Q285C mutant was less relative to the Q311C mutant. There was no effect of DTT on biotinylation of Q311C, and biotinylation of Q285C was only slightly increased with DTT indicating no, and negligible, modification by an endogenous sulfhydryl-reactive moiety, respectively (30, 31). However, biotinylation of the Q285C-Q311C double mutant was markedly decreased and was fully restored after treatment with DTT consistent with the spontaneous formation of a disulfide bond that was broken by DTT. Likewise, while there was a modest increase in biotinylation with DTT for the T289C and I304C mutants, suggesting some endogenous modification, biotinylation of the double mutant (T289C-I304C) was abolished and fully restored after treatment with DTT. These observations are consistent with disulfide bond formation involving both dicysteine pairs. The pattern of biotinylation is also consistent with the functional data indicated in Fig. 5B. DTT resulted in a small decrease in transport of the Q311C and Q285C mutants. There was a marked fall in transport activity of the double mutant which was reversed by DTT. Likewise, DTT had no effect on the activities of the T289C and I304C mutants; the activity of the double-mutant was nearly abolished but function was restored to ~70% of the single Cys-mutants by DTT.

Fig. 5.

Cross-linkage of dicysteine pairs in the 7th and 8th TMDs; impact of DTT. A: accessibility to MTSEA-biotin. Single and dicysteine PCFT mutants were treated with HBS or 10 mM DTT and then treated with 0.5 mM MTSEA-biotin for 30 min. The cells were then processed as described in materials and methods and Fig. 3. High- and low-exposure Western blots are indicated in top and middle bands, respectively. Expression in the crude membrane extract is indicated in the bottom band. The blots are representative of 3 independent experiments. B: function: influx was assessed over 1 min with 0.5 µM [³H]MTX at 37°C. The cells were treated with HBS or HBS containing 10 mM DTT for 10 min before transport measurements. Data are means ± SE from 3 independent experiments.

DISCUSSION

In a previous study, residues deep into the exofacial portion of the 8th TMD were found to be accessible to the extracellular compartment and the aqueous translocation pathway (1). This high degree of contiguous accessibility differed from the usual expected “helix wheel” accessibility pattern observed for solute transporters (8, 12–14) and went far beyond what would be expected for residues at the extracellular interface. Using homology models of PCFT, based on the inward- and outward- open conformations of the bovine and rodent Glut5 fructose transporter structures (15), respectively, the data suggested that conformational changes in the proximal helix during cycling of the carrier explain the accessibility of all residues in this region to the aqueous channel (1). The observation that residues deep within the helix appeared to be involved in folate substrate binding was consistent with an important role for this TMD in defining the aqueous pathway. This report extends these findings to establish the close proximity of the exofacial regions of both the 7th and 8th TMDs and the accessibility to biotinylation of eight contiguous exofacial residues in the 7th TMD.

PCFT provides a very useful model for SCAM analyses since none of the five Cys residues in the PCFT-DSL scaffold are accessible to MTSEA-biotin as is the case for the Cys-less PCFT (34). This avoids the use of the Cys-less scaffold, as is required for the application of SCAM for the characterization of most solute transporters. For PCFT, at least, Cys-less PCFT lacks the resilience to tolerate some second mutations (33) and perhaps results in subtle structural differences as compared with PCFT-WT or PCFT-DSL. Using the PCFT-DSL background for the SCAM analysis, all Cys-substituted residues in the 7th TMD were expressed although, for a few, expression was somewhat decreased. Likewise, activity was generally preserved and, beyond decreases commensurate with a decrease in expression, few of the Cys-substituted residues lost appreciable function. Transport activity of the H281C mutant was markedly decreased, but this residue was previously shown to play an important role as a determinant of proton and folate binding to the protein (22). While activity of D286C was decreased, function was preserved for the D286K mutant (20). Further, the addition of sulfhydryl reagents markedly augmented the activity of the D286C (not shown), suggesting that bulk at this site enhances function. The pattern of biotinylation for the 7th TMD was similar to, but did not extend into the membrane as deeply as observed with SCAM analysis of the 8th TMD. Eight contiguous Cys-substituted residues, starting with the most exofacial, were biotinylated (L288C–H281C) skipping two residues to label I278C. This uninterrupted series of accessible residues correlated with a substantial rotation of the exofacial region of the helix, even more profound than reported for the 8th TMD, based on homology modeling in both conformations (1).

On the basis of the homology models, the exofacial regions of the 7th and 8th TMDs are in proximity but most prominently in the inward-open conformation when the external gate is closed, allowing contacts between the residues in the exofacial regions of the helices. There is a break in the 8th TMD in the outward-open configuration, but there is little change in the orientation of the endofacial region of the helix. On the other hand, the 7th TMD is broken in both conformations and the endofacial portion of the helix is splayed into the membrane away from the aqueous channel, more so for the inward-facing conformation, consistent with the lack of accessibility of residues beyond this segment of the helix. The model predicted further that two pairs of Cys-substituted resides, T289C-I304C and Q285C-Q311C, should be in sufficient proximity to be cross-linked. This turned out to be the case.

Cross-linkage of these residues was established based on 1) the loss of accessibility to MTSEA-biotin and the loss of transport function when both residues were Cys-substituted in a single PCFT construct and 2) the restoration of biotinylation and function when the disulfide bond was disrupted by the treatment of cells with DTT. The restoration of function by DTT excluded the possibility that the loss of function of the double mutants was due to the functional consequences of two mutations in a single protein and not due to the formation of the disulfide bond itself. It is of interest that these bonds formed spontaneously, while in the establishment of an external PCFT gate, only one of the dicysteine pairs formed a disulfide bond spontaneously; the others required the use of an oxidizing agent for bond formation (31). Disulfide bond formation of Cys-substituted residues in helix-pairs has been employed frequently to establish proximity based on molecular modeling of solute transporters. In most cases, however, there was a requirement for an oxidative stimulus to achieve bond formation (24, 25, 31).

Taken together, the data indicate that the exofacial regions of the 7th and 8th TMDs play an important role in defining the aqueous translocation pathway. The data suggest that the region defined in part by the external portion of these two helices represents an exofacial cleft within the outward-open conformation of the carrier protein as observed for other solute transporters (11, 16). The data further suggest that both helices may play an important role in folate binding in that biotinylation of a large number of residues could be blocked or diminished by pemetrexed. In the case of the 8th TMD, biotinylation of six residues deep within the helix (A309C, A310C, H312C, L313C, P314C, and L316C) was suppressed with pemetrexed. For the 7th TMD, biotinylation of seven residues (I278C, H281C, F282C, G283C, A284C, A286C, and likely L288C) closer to the extracellular interface was markedly diminished by pemetrexed. Hence, between the two helices, this protection involved 14 residues (Fig. 6). Alternatively, it is possible that the suppression of accessibility induced by pemetrexed could be a result of a substrate-induced conformational change limiting access of the MTSEA-biotin to these residues, rather than due to a direct interaction between these residues and pemetrexed. However, these effects of pemetrexed were also detected at 4°C, conditions in which conformational changes should be minimized or absent.

Fig. 6.

Illustration depicting accessibility of residues in the 7th and 8th TMDs and protection of biotinylation by pemetrexed. Residues in the 7th TMD are denoted in violet, the 8th TMD in green. Residues accessible to MTSEA-biotin in both helices are highlighted in dark colors. Residues for which biotinylation is suppressed by pemetrexed are labeled in white. Residues that form cysteine cross-links, Gln285-Gln311 and Thr289-Iso304 are highlighted in light and dark orange, respectively. Data for the 8th TMD is from Ref. 1.

GRANTS

This work was supported by National Institutes of Health Grants CA082621 (I. D. Goldman) and GM118709 (A. Fiser).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

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