Functional roles of the A335 and G338 residues of the proton-coupled folate transporter (PCFT-SLC46A1) mutated in hereditary folate malabsorption

Abstract

The proton-coupled folate transporter (PCFT-SLC46A1) mediates intestinal folate absorption and folate transport across the choroid plexus, processes defective in hereditary folate malabsorption (HFM). This paper characterizes the functional defect, and the roles of two mutated PCFT residues, associated with HFM (G338R and A335D). The A335D-PCFT and other mutations at this residue result in an unstable protein; when expression of a mutant protein was preserved, function was always retained. The G338R and other charged mutants resulted in an unstable protein; substitutions with small neutral and polar amino acids preserved protein but with impaired function. Pemetrexed and methotrexate (MTX) influx kinetics mediated by the G338C mutant PCFT revealed marked (15- to 20-fold) decreases in K_t and V_max compared with wild-type PCFT. In contrast, there was only a small (∼2-fold) decrease in the MTX influx K_i and an increase (∼3-fold) in the pemetrexed influx K_i for the G338C-PCFT mutant. Neither a decrease in pH to 4.5, nor an increase to 7.4, restored function of the G338C mutant relative to wild-type PCFT excluding a role for this residue in proton binding or proton coupling. Homology modeling localized the A335 and G338 residues embedded in the 9th transmembrane, consistent with the inaccessibility of the A335C and G338C proteins to MTS reagents. Hence, the loss of intrinsic G338C-PCFT function was due solely to impaired oscillation of the carrier between its conformational states. The data illustrate how alterations in carrier cycling can impact influx K_t without comparable alterations in substrate binding to the carrier.

hereditary folate malabsorption (HFM) is an autosomal recessive disorder caused by loss-of-function mutations in the proton-coupled folate transporter (PCFT; SLC46A1) (2, 15, 34). The major pathophysiological consequences are impaired folate transport across the apical brush-border membrane of the proximal small intestine and impaired vectorial transport from blood to cerebrospinal fluid across the ependymal cells of the choroid plexus (31, 33). These defects result in severe systemic folate deficiency with anemia, immune deficiency, and central nervous system folate deficiency, resulting in developmental defects and, ultimately, seizures (2, 6, 33). The functional consequences of point mutations in PCFT in HFM that result in amino acid substitutions have been characterized with the demonstration of defects in protein stability and trafficking, folate substrate binding, and/or alterations in the rates of oscillation between the conformational states of the carrier (11, 14, 21, 22). These studies, along with site-directed mutagenesis at these and other residues, have provided insights into the structure/function of this transporter and its secondary structure (26, 27, 28, 35, 36, 37).

The current paper characterizes the basis for the loss of function of two PCFT residues mutated in subjects with HFM, located in the 9th of its 12 transmembrane domains (TMDs) (20), and provides information on the role each residue plays in PCFT function. One of the mutations (A335D) results in an unstable protein; when any other amino acid substitution at this residue results in a defect in function this is always associated with an unstable protein. The other mutant PCFT (G338R) protein is also unstable; however, other mutants that are expressed have impaired function. In one case (G338C), loss of function was due to a marked defect in the oscillation of the carrier between its conformational states. The latter studies reveal the discrepancies that can occur between the measured influx K_t (the concentration of substrate at which influx is one half the maximum rate of transport) and the actual binding constant, K_i, associated with a marked decrease in the rate of cycling of the carrier.

EXPERIMENTAL PROCEDURES

Chemicals.

[³H]methotrexate (MTX-disodium salt, [3′,5′,7-[³H](N)]) and [³H]pemetrexed were obtained from Moravek Biochemicals (Brea, CA). Purity was established and monitored by liquid chromatography (30). EZ-Link Sulfo-NHS-LC-Biotin [sulfosuccin-imidyl-6-(biotinamido) hexanoate] was purchased from Pierce Biotechnology (Rockford, IL), streptavidin-agarose beads were from Fischer Scientific (Pittsburgh, PA), and protease inhibitor cocktail was from Roche Applied Science (Mannheim, Germany). The sulfhydryl reactive reagent N-biotinylaminoethyl methanethiosulfonate (MTSEA-Biotin) was purchased from Biotium (Hayward, CA). Methanethiosulfonate-ethyltrimethylammonium (MTSET⁺) and methanethiosulfonate-ethylsulfonate (MTSES⁻) were obtained from Affymetrix (Santa Clara, CA). MG132 (N-CBZ-Leu-Leu-Leu-AL) was purchased from Sigma-Aldrich (St. Louis, MO), and bafilomycin A1 was from LC Laboratories (Boston, MA).

Construction of mutant plasmids by site-directed mutagenesis.

Mutations in pcft cDNA were generated with the QuikChange II XL Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). A PCFT pcDNA3.1(+) expression vector was used as template which encodes hemagglutinin (HA)-tagged PCFT at the COOH terminus. Mutant constructs were verified by DNA sequencing in the Albert Einstein Cancer Center Genomics Shared Resource.

Cell lines, cell culture conditions, and transient transfection.

HeLa-R1–11 cells, the transient transfection recipients in this study, were derived from the HeLa R1 cell line in which there is a genomic deletion of reduced folate carrier (RFC) and silencing of PCFT expression due to methylation of the promoter and a loss of gene copies (3, 32). R1–11 cells were maintained in RPMI-1640 medium supplemented with 5% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO₂. Cells (3.5 × 10⁵/vial) were seeded into 17-mm liquid scintillation vials in preparation for transport studies, or 6 × 10⁵ cells/well were seeded into 6-well plates in preparation for analyses of PCFT protein in the crude membrane preparation and biotinylation at the cell surface. Forty-eight hours later, these cells were transfected with PCFT constructs (0.8 μg/vial or 2 μg/well of a 6-well-plate) with lipofectamine 2000 (3 μl/vial for transport studies, 5 μl/well for Western blot analysis, Invitrogen, Carlsbad, CA). Two days later, the cells were processed for transport studies or Western blot analyses. The transport data are the average of three separate experiments performed on different days. In each experiment, points were taken in duplicate. Influx was always compared among R1–11 cells transiently transfected with wild-type PCFT, the mutant PCFT, or the vector alone.

Real-time PCR.

Total RNA was purified by TRIzol reagent (Invitrogen) from HeLa-R1–11 cells transiently transfected with wild-type PCFT, mutant PCFTs, or the vector alone. cDNA was synthesized by Superscript H-Reverse Transcriptase (Invitrogen) from 5 μg of RNA. Quantitative PCR was performed with primers reported previously (15) using SYBR green PCR Master Mix (Applied Biosystems, Warrington, UK) at the Albert Einstein College of Medicine Genomics Shared Resource.

Membrane transport analyses.

In preparation for transport studies, R1–11 transfectants were washed twice with 2 ml of HBS buffer (20 mM HEPES, 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, and 5 mM dextrose at pH 7.4) and incubated with the same HBS buffer (2 ml per vial) in a water bath (37°C) for 20 min. The buffer was then aspirated, 500 μl of transport buffer, either HBS or MBS (20 mM MES, 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, and 5 mM dextrose), was added containing tritiated antifolates. The HBS transport buffer was adjusted with 1 N NaOH for studies at pH ≥ 7.0, while MES was adjusted with 1 N HCl for studies at pH < 7.0. Transport was halted after 1 min by the addition of 10 volumes (5 ml) of ice-cold HBS buffer (pH 7.4). Over this interval, PCFT-mediated uptake was unidirectional (29). Cells were washed three times with 5 ml of ice-cold HBS, followed by the addition of 500 μl of 0.2 M NaOH; then the cells were digested for 1 h at 65°C. A portion of the hydrolysate (400 μl) was analyzed on a liquid scintillation spectrometer; the protein content of another portion of hydrolysate (10 μl) was determined using the Pierce kit (Thermo Scientific, Rockford, IL).

Influx is expressed as pmol tritiated substrate·mg protein⁻¹·min⁻¹ or percentage of wild-type activity. Influx kinetics were determined on the basis of the relationship between influx and extracellular tritiated substrate concentration by a nonlinear regression best fit to the Michaelis-Menten equation. Nonspecific uptake and adsorption to the cell surface was assessed in vector only-transfected cells and subtracted from total uptake. A Dixon analysis was employed for the measurement of influx K_i as described previously (22).

Mutant PCFT constructs were screened to obtain clues as to the basis for the functional defects. A defect in proton binding to the carrier which results in decreased binding of folate substrate can be corrected, in part, by a decrease in pH (21, 26). A defect in proton coupling results in a loss of activity at low pH, but retention of activity at neutral pH in the absence of a proton gradient (27). When there is a decrease in influx K_t, the ratio of influx mediated by the mutated PCFT to influx mediated by wild-type PCFT is decreased when the substrate concentration is increased. If the ratio remains low at the highest concentration, the influx V_max is decreased.

Sulfhydryl modification by MTS reagents.

As described above, RI-11 cells were seeded and transfected with the PCFT expression vectors. Two days later, the cells were washed with HBS buffer (2 ml) twice and treated with fresh MTSET (positive charge) or MTSES (negative charge) solution in HBS buffer (1 ml, pH 7.4, final concentration ≈3 mM) at room temperature. After 30 min, cells were washed twice with HBS buffer and [³H]MTX influx was measured as described above.

Cell surface and Cys biotinylation assays.

For analysis of PCFT accessibility at the cell surface, cells in HBS were treated for 30 min with 1 mg/ml EZ-Link Sulfo-NHS-LC-Biotin which reacts with lysine residues on the cell surface. For cysteine biotinylation assays, cells were treated with MTSEA-Biotin (2 mg/100 μl DMSO 1:100 diluted into HBS) for 30 min at room temperature. The cells were then washed twice with 2 ml of HBS and treated with hypotonic buffer (0.5 mM Na₂HPO₄ and 0.1 mM EDTA at pH 7.0) containing protease inhibitors and kept on ice for 30 min. Cells were then scraped from the plates with a cell lifter, centrifuged at 14,000 rpm for 15 min at 4°C, following which the pellet was resuspended in 400 μl of lysis buffer (50 mM Tris-base, 150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, pH 7.4) containing protease inhibitors and rotated in a cold room (4°C) for 30 to 120 min. A 25 μl portion, designated as the crude membrane protein fraction, was taken for Western blot analysis. The remaining sample was centrifuged at 14,000 rpm at 4°C for 15 min, and the supernatant was mixed with streptavidin-agarose beads (50 μl that had been prewashed three times with 100 μl of lysis buffer) overnight at 4°C on the rotator. The next day, the beads were washed twice with lysis buffer (500 μl) and an additional two times with lysis buffer (500 μl) containing 2% SDS. With each wash the cells were rotated for 20 min at room temperature. Protein bound to the beads was stripped by heating for 5 min at 95° in 2× SDS-PAGE loading buffer containing dithiothreitol (DTT).

Western blot analysis.

Protein samples were loaded directly onto 12% Tris·HCl polyacrylamide gels. The crude membrane fraction was mixed with 2× SDS-PAGE loading buffer (1:1) containing DTT at room temperature before Western blot analysis. After SDS-PAGE, proteins were transferred to Amersham Hybond membranes (GE Healthcare, Piscataway, NJ) at 4°C for 2 h (300 mAmp) 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 of crude membrane fractions from cells assessed for surface biotinylation (Figs. 1 and 2) were probed first with a rabbit β-actin antibody (Cell Signaling Technology, Danvers, MA) followed by stripping with buffer (100 mM 2-β-mercaptoethanol, 2% SDS, 62.5 mM Tris·HCl pH 6.7) and were reprobed with anti-HA antibody (Sigma, St. Louis, MO; 1:4,000 in TBST, 0.1% milk). The blots of crude membrane fractions of cysteine-substituted residues exposed to MTSEA-Biotin were probed directly with anti-HA antibody. For samples obtained from the beads, blots were probed directly with anti-HA antibody. After application of the first antibody, the blot was probed with an anti-rabbit IgG-horseradish peroxidase conjugate (1:5,000 in TBST; Cell Signaling Technology). Subsequently, blots were developed with the Amersham ECL Plus reagent (GE Healthcare). The levels of expression of wild-type PCFT and G338C protein were determined with ImageJ software (National Institutes of Health, Bethesda, MD; http://rsbweb.nih.gov/ij/download.html).

Fig. 1.

Expression and function of the A335 and G338 proton-coupled folate transporter (PCFT) mutants. A and C: [³H]methotrexate (MTX) influx was assessed at a concentration of 0.5 μM at pH 5.5 over 1 min for the A335 and G338 PCFT mutants, respectively. Data are means ± SE from three experiments performed on different days. WT, wild type. B and D: Western blot analyses of the crude membrane fraction (top bands) and biotinylated protein accessible at the cell surface (bottom bands) for the A335 and G338 PCFT mutants, respectively. Actin was the loading control (middle bands). The Western blots are representative of two experiments, performed on different days. Vertical lines were inserted on B and C to indicate repositioned gel lanes.

Fig. 2.

The effect of neutral or polar substitutions on the expression and function of the A335 and G338 residues. A and C: [³H]MTX influx was assessed at a concentration of 0.5 μM and pH 5.5 at 37°C over 1 min. Data are means ± SE from three experiments performed on different days. B and D: Western blot analyses of the crude membrane fraction (top bands) and biotinylated protein at the cell surface (bottom bands) for the A335 and G338 residues, respectively. Actin was the loading control (middle bands). The Western blots are representative of two experiments performed on different days. Vertical lines were inserted to indicate repositioned gel lanes.

Molecular modeling.

A homology model for PCFT has been employed by this laboratory as an adjunct to the interpretation of experimental observations within a three-dimensional context. A variety of homology models were built that differed in their target-to-template input alignments, which largely determines the quality of subsequent model building and optimization. Energetically, the most stable model according to Prosa energy score evaluation (23) was obtained using input from the HHpred (24), a Hidden Markov model-based fold-recognition and alignment method. The best scoring template remained the crystal structure of the glycerol-3-phosphate transporter from Escherichia coli [Protein Data Bank (PDB) code 1pw54] as in other studies (11, 21, 26). The optimal alignment between PCFT and 1pw54 served as input for the comparative protein structure modeling program Modeller (5, 19). Additional model quality verification was conducted by predicting the location of transmembrane segments using HMMTOP (25) and comparing it with the location of transmembrane helices in the three-dimensional model. Alternative models were obtained by other alignment techniques [such as Muscle (4), Clustalw (1), MMM (16, 17), Align2D (12, 13)]. The corresponding models resulted in a largely conserved core of transmembrane segments but substantial variations in the modeling of the extra and intracellular loop regions.

Statistical analysis.

Statistical comparisons were performed by the two-tailed Student's paired t-test, and statistical significance was determined at P ≤ 0.05 (GraphPad Prism, version 3.0 for Windows).

RESULTS

Expression and functional properties of the A335 and G338 PCFT mutants.

A335D and G338R substitutions were identified as loss-of-function PCFT mutations in two subjects with HFM (20). To understand the basis for the loss of activity, and the role these residues play in PCFT function, mutants were constructed to create these and other substitutions. The A335D and A335K mutant PCFTs were inactive; there was slight residual activity for the A335E and A335P mutants compared with the vector-transfected cells (P < 0.009; Fig. 1A). All mutations at the G338 residue resulted in a complete loss of activity (Fig. 1C).

As illustrated in Fig. 1, B and D, the loss of function observed with mutants of both residues could be attributed to the lack of protein in the crude membrane fraction and protein accessible for biotinylation at the cell surface. A trace of the A335P PCFT protein was detected at the cell surface but not in the crude membrane fraction, an observation related to the greater dilution of protein in the latter isolation and electrophoresis procedures. There was no apparent difference in mRNA stability between wild-type PCFT and the A335D or G338R mutants as the mRNA levels assessed by quantitative RT-PCR were not altered in the mutant compared with the wild-type PCFT transient transfectants (P = 0.1156 and 0.2144, respectively, based on the average of three experiments; data not shown).

To assess the structural and functional roles or requirements of the A335 and G335 residues (the initial four mutants all resulted in unstable proteins), further site-directed mutagenesis was performed by substituting these residues with amino acids representative of differences in size and polarity in a search for mutants in which PCFT protein was preserved. Accordingly, A335 was replaced with Gly, Leu, Asn, Cys, and Phe and G338 was replaced with Ala, Cys, Ser, and Phe. As indicated in Fig. 2A, the A335 residue tolerated substitutions with a wide variation in amino acid size and polarity except for Asn. The latter loss of function correlated with the decrease in expression of the mutant A335N PCFT protein. All of these mutant proteins were accessible to biotinylation at the cell surface except for A335N, which was barely detectable (Fig. 2B). Hence, all substitutions at this site that result in a loss of function are due to loss of protein stability. Among those that are expressed, function is retained irrespective of size and polarity except in the case of Asn.

Substitutions at the G338 position produced a different result. The Ala and polar substitutions, Cys and Ser, preserved protein and some function, ∼35%, ∼12%, ∼27%, respectively, of wild-type activity (Fig. 2C). The Phe substitution resulted in complete loss of activity due to a lack of protein expression (Fig. 2D). Further studies were performed with the G338C mutant PCFT since it was expressed at levels only slightly less (0.73) than wild-type PCFT, based on densitometry. A role in proton coupling for this residue was excluded by the observation that the loss of function relative to wild-type PCFT for G338C was not corrected at pH 7.4 in the absence of a proton gradient. A role in proton binding was excluded by the demonstration that MTX influx was not enhanced when the pH was decreased to 4.5 (data not shown). However, an analysis of influx at low versus high MTX concentrations for the G338A, -C, -S PCFT mutants suggested that there was an increase in affinity and decrease in V_max for these mutant carriers (higher activity relative to wild-type PCFT at lower MTX concentrations, data not shown). This was not observed for the A335G, -L, -C, -F and -T mutants, suggesting that there was no change in their intrinsic function. The G338C mutant was chosen for further study since its expression was comparable to that of wild-type PCFT, as described above.

MTX and pemetrexed influx kinetics mediated by the G338C PCFT mutant.

The kinetic basis for the loss of activity of the G338C PCFT mutant was assessed. Recent studies have identified differences in the impact of PCFT mutations on MTX and pemetrexed influx kinetics and binding to PCFT (22); hence, these parameters were assessed with both substrates. As indicated in Fig. 3, there was a marked decrease in the influx V_max for pemetrexed (top) and MTX (bottom) for the G338C mutant relative to wild-type PCFT, 20- and 15-fold, respectively. Because the velocities were so low for the G338C PCFT, the graphs were expanded over the zero-to-1 μM range for the mutant (insets), revealing clear saturation kinetics. As indicated in Table 1, there was a marked decrease in the influx K_t for pemetrexed and MTX, 15- and 20-fold, respectively. Differences in the magnitude of these changes between MTX and pemetrexed were not statistically significant (P = 0.5). The data indicated a marked decrease in the rate of cycling of the carrier while suggesting a marked increase in the affinity of the mutated carrier for its substrates.

Fig. 3.

Pemetrexed and MTX influx kinetics mediated by wild-type and G338C mutant PCFT. [³H]pemetrexed (PMX; top) and [³H]MTX (bottom) influx was measured over 1 min at pH 5.5. The insets are an expansion of the data over the 0-to-1 μM concentration range for the G338C mutant PCFT. Data are representative of three experiments performed on different days; measurements in each experiment were performed in duplicate.

Table 1.

Summary of MTX and pemetrexed influx kinetic values for wild-type and G338C mutant PCFT

				Vmax, pmol•mg protein•−1min−1
	Kt , μM	Mutant/Wild-Type	Ki, μM	Uncorrected	Relative expression at the cell surface (G338C/wild-type)	Corrected for expression	Mutant/wild-type; corrected for expression
Wild type
MTX	3.37 ± 1.09		2.3 ± 0.3	571.0 ± 76.4	1.0
PMX	0.59 ± 0.21		0.18 ± 0.04	147.6 ± 17.5
G338C
MTX	0.17 ± 0.05	0.05	1.1 ± 0.3	7.85 ± 0.53	0.73	10.7	0.02
PMX	0.04 ± 0.02	0.07	0.12 ± 0.02	7.65 ± 0.55		10.5	0.07

Values are means ± SE. A comparison of methotrexate (MTX) and pemetrexed (PMX) influx kinetics based on the experimental data of Figs. 3, 4, and 5. Relative expression was measured by ImageJ densitometry in two separate experiments and was used to normalize V_max. The P values for differences between the K_t (concentration of substrate at which influx is half the maximum rate of transport) and K_i (inhibition constant) for wild-type proton-coupled folate transporter (PCFT) were 0.08 for pemetrexed and 0.09 for MTX. The P values for differences between the K_t and K_i for the G338C-PCFT mutant for pemetrexed was 0.001 and for MTX was 0.03.

The K_t reflects the concentration at which the influx process is half the maximum rate and is not necessarily an indication of the affinity of the carrier for its substrates. On the other hand, quantification of the inhibition constant (K_i) of one substrate, derived from the kinetics of inhibition of the influx of another substrate, is based solely on an interaction at the level of substrate binding. To determine this parameter, the K_i, for pemetrexed was assessed based on inhibition of [³H]MTX influx (Fig. 4), and the K_i for MTX was assessed based on inhibition of [³H]pemetrexed influx (Fig. 5) using the Dixon analysis—the reciprocal of substrate influx as a function of inhibitor concentration, at two different substrate concentrations. In this analysis a perpendicular from the point of intersection of the two lines to the x-axis is the −K_i. On the basis of this analysis, as indicated in Table 1, the MTX K_i was 2.3 ± 0.3 μM for wild-type PCFT and 1.1 ± 0.3 μM for the G338C-PCFT mutant (P = 0.01), in contrast to the 20-fold difference in influx K_t values. When MTX influx K_t and K_i for wild-type PCFT were compared, the P-value was 0.08; the small decrease in the K_i relative to K_t for the G338C-PCFT reached a P-value of 0.03. The pemetrexed influx K_i at 0.18 ± 0.04 μM for wild-type PCFT was not different from the G338C mutant, 0.12 ± 0.02 μM, (P = 0.2), in contrast to the 15-fold difference in K_t values. The difference between the wild-type values for pemetrexed reached a P-value of 0.09, but the higher pemetrexed K_i than K_t for the G338C-PCFT was highly significant (P = 0.001). Hence, there was only a slight decrease in the influx K_i for MTX and pemetrexed, associated with the G338C mutation despite the marked decrease in K_t values.

Fig. 4.

Determination of the MTX influx inhibition constants (K_i) based on inhibition of [³H]pemetrexed influx as assessed by the Dixon analysis. [³H]pemetrexed influx was assessed in wild-type PCFT (left) and the G338C mutant PCFT (right) at two concentrations, 0.04 μM and 0.15 μM, in the absence and presence of nonlabeled MTX at the indicated concentrations. Transport was measured at pH 5.5 over 1 min at 37°C. The vertical line from the point of intersection of the two lines intersects the x-axis at the negative K_i. The K_i values indicated in Table 1 are based on the average of three experiments performed on different days, one of which is represented in this figure; measurements in each experiment were performed in duplicate.

Fig. 5.

Determination of the pemetrexed influx inhibition constants (K_i) based on inhibition of [³H]MTX influx as assessed by the Dixon analysis. [³H]MTX influx was assessed in wild-type PCFT (left) and the G338C mutant (right) at two concentrations, 0.15 μM and 0.5 μM in the absence and presence of nonlabeled pemetrexed at the indicated concentrations. Transport was measured at pH 5.5 over 1 min at 37°C. The vertical line from the point of intersection of the two lines intersects the x-axis at the negative K_i. The K_i values indicated in Table 1 are based on the average of three experiments, one of which is represented in this figure; measurements in each experiment were performed in duplicate.

Impact of sulfhydryl modification of Cys mutants by MTS reagents.

Cys-reactive MTS reagents were used to assess the accessibility of residues in transmembrane domains to the aqueous translocation pathway (9, 18). Of the two HFM-associated residues substituted with Cys in this study, A335C preserved ∼100% of wild-type PCFT activity, while G338C retained only 10% of wild-type activity. The two Cys-active water-soluble reagents used were MTSET and MTSES, with positive and negative charges, respectively, at concentrations of 1–10 mM (9). Neither reagent suppressed MTX influx mediated by the two mutant carriers (data not shown).

To exclude the possibility that these PCFT mutant residues are modified by MTS reagents, but without functional consequence, further experiments were performed with MTSEA-Biotin. Residues accessible to this reagent should form an MTSEA-PCFT biotin conjugate that can be pulled down with the streptavidin beads. This was not the case for either of the Cys mutants in contrast to the E292C mutant (on a Cys-less PCFT background) used as a positive control [Fig. 6 (37)]. Hence, neither the G338C nor the A335C residues appeared to be accessible to the aqueous translocation pathway or near the folate binding pocket.

Fig. 6.

Cysteine biotinylation by N-biotinylaminoethyl methanethiosulfonate (MTSEA-Biotin) for the A335C and G338C mutants. Biotinylation and immunoprecipitation were performed for A335C, G338C, wild-type, Cys-less PCFT and a positive control (E292C on Cys-less background) as previously reported (37). The blot is representative of two independent experiments; the vertical lines were inserted to indicate repositioned gel lanes.

DISCUSSION

Since the cloning of PCFT, several approaches have been utilized to elucidate the structure-function of this carrier. These have included systematic site-directed mutagenesis, random mutagenesis, and characterization of mutant residues identified in subjects with HFM (31, 33). These studies have defined residues critical to PCFT function. Eight point mutations have been identified in subjects with HFM. The R113 residue, located in the first intracellular loop between the 2nd and 3rd TMDs was mutated in two subjects with HFM (10, 34). Mutational analysis indicated that only preservation of the positive charge at this residue (R113H, R113K) could sustain slight residual function (10, 34). The D156 residue, located in the 4th TMD, is critical for protein stability (21). Two mutations associated with HFM allowed sufficient expression and activity to characterize the nature of the functional defect. A R376Q mutant PCFT resulted in impaired substrate binding and oscillation of the carrier; binding was better preserved for pemetrexed than for MTX and the reduced folates (14). A P425R mutation, at the junction between the 6th extracellular loop and the 12th TMD, also resulted in impaired substrate binding but in a highly selective way. There was complete preservation of the pemetrexed influx K_t, while there was a marked defect in binding of MTX. The current study provides information on two additional residues mutated in subjects with HFM, A335D and G338R, both in the 9th TMD. These residues are highly conserved; the only variance is G338S in Xenopus. However, this substitution resulted in preservation of ∼25% of function relative to wild-type PCFT in Hela cells; about half of this loss could be attributed to a lower level of protein expression at the cell surface (Fig. 2, C and D).

All of the PCFT point mutations reported to date in subjects with HFM involve drastic changes (charged to noncharged, or noncharged to charged, residues) in or adjacent to TMDs. Likewise, all the mutants in this report are drastic. Loss-of-function substitutions at the A335 residue were always associated with lack of protein in the crude membrane preparation and accessibility to biotinylation at the cell surface. Replacement of G338 residue with charged residues was also not tolerated. On the other hand, there was full preservation of expression and function when this residue was replaced with neutral amino acids irrespective of size. Polar amino acid replacement was less well tolerated but function was better preserved with replacement by the smaller Ser than Cys.

To obtain further insight into the basis for the loss of function of the G338C residue, influx kinetics were assessed for this mutant which was expressed at levels only slightly less than wild-type PCFT. The G338C-PCFT mutant manifested a marked decrease in influx V_max for both MTX and pemetrexed, far out of proportion to the small decrease in protein expression. Hence, it would appear that this residue plays a critical role in oscillation of the carrier between its conformational states. What was of particular interest was the observation that the marked decrease in influx K_t for both substrates was not accompanied by a comparable change in influx K_i. Indeed, the pemetrexed and MTX influx K_i values were only slightly decreased as compared with that of wild-type PCFT. This was a different pattern of change than recently reported by this laboratory in which there was a marked increase in the influx K_t for MTX associated with a P425R PCFT mutation identified in another subject with HFM. However, in that case, the increase in MTX influx K_i for the mutant PCFT was even greater despite the fact that the V_max was decreased (22). Also, in contrast to the current report, the influx K_t for pemetrexed decreased slightly for the P425R mutant compared with the wild-type carrier. In the current study, while the influx K_t and K_i were comparable for wild-type PCFT, this relationship deviated substantially when the residue was mutated G338C. While discrepancies between K_m and K_i are well recognized in enzymatic reactions, this is not a phenomena commonly looked for, nor recognized, for membrane transport processes. Indeed, most investigators describe influx kinetics in terms of a K_m as if this represents an affinity constant. The discrepancies found in this and an earlier study with PCFT mutants should not be surprising (22). In a proton-coupled process, the transport cycle is highly complex and dependent on many different parameters. Starting with the initiation of influx, proton binds to the carrier, then folate substrate binds, following by a carrier protein shift from an outward to an inward facing conformation, followed by the release of folate substrate, then the release of proton, and a return of the carrier protein to an outward facing conformation. Earlier studies indicate that both the latter step and the release of proton within the cell, if sufficiently slow, can be rate-limiting in the carrier cycle (27).

A homology model of the PCFT structure has been generated based on the known structure of the glycerol-6-phosphate bacterial transporter (8, 11, 26) and was recently refined (22). According to this model, the small G338 and A335 residues are located about one helix turn away, in close spatial proximity to one other, both buried within the 9th TMD and opposing the 10th TMD (Fig. 7). Both residues are remote from the aqueous translocation pathway and folate binding pocket consistent with the observation that neither of the Cys-substituted residues was accessible to MTS reagents or MTSEA-Biotin. Likewise, mutants at the A335 residue that were expressed were fully functional, and there were only minor alterations in the binding constant for the G338 mutant that was expressed. The location of these residues is also consistent with the loss of protein stability with charged substitutions that occurred in subjects with HFM that would disrupt the helix-helix contacts, predicted between the 9th and 10th TMDs (7). In the wild-type transporter the environment of these residues appears to be hydrophobic, the contacting residues from TMD10 appear to be L367 and V370. Therefore introduction of an uncompensated charge (Arg, Asp) or a strong hydrogen bond donor or acceptor residues (Asn) would be expected to thermodynamically destabilize the environment, consistent with the experimental results. In addition, replacing the G338 which does not have a side chain with much bulkier residues would also be expected to introduce a rigid body shift of the 9th and 10th TMDs to accommodate the larger volumes, as occurred with the Phe substitution. The Glu336 residue, located between A335 and G338, points away from the other TMDs, outward to the extramolecular space, approximately at the border of the intracellular and lipid bilayer environment. It does not require a compensating amino acid as it is most likely compensated by solute electrolytes and is expected to be more accommodating to substituted residues. Hence, as previously reported, when this residue was replaced by Leu (as occurs in Zebrafish), function was fully preserved (27).

Fig. 7.

A cartoon representation of a homology model of PCFT. Left: a side view of PCFT, where the top faces the extracellular space and the bottom faces the intracellular space. Ala335 is in green, Gly338 is in yellow, and Glu336 is in orange, in a space-filling representation. Transmembrane domains (TMDs) 9 and 10 are labeled; the opening of the aqueous translocation pathway into the cytosol (transport channel) is marked. The TMDs are color coded from NH₂ to COOH termini, from red to blue color shades, respectively. Right: a view looking into the PCFT aqueous translocation pathway from the cytoplasm. The colors and labels are as before.

It is of interest that the consequences of the substitutions at the A335 and G338 residues were quite different despite their close spatial proximity. In the former case, all mutations that affected function were due to a loss of protein stability. When protein was expressed, intrinsic function was preserved irrespective of size or polarity, save for Asn, a polar substitution that would require a compensating residue such as a hydrogen bond donor or acceptor, presumable in the 10th TMD, to sustain the integrity of the protein. On the other hand, for the G338C PCFT mutant, which would be expected to produce only a minor destabilization of the hydrophobic pocket, the bulk of the protein was preserved but there was a marked fall in influx V_max that can be attributed to a marked decrease in the mobility of the carrier due to the steric restrictions of the bulkier Cys side chain. Alternative models were also inspected and the conclusions were similar; this is attributed to the fact that this is the most conserved portion of the protein. Continued analyses of mutations associated with HFM will provide further insights into the structure-function of this transporter and its tertiary structure that will further inform the development, and test the validity, of a molecular model of PCFT. Findings in the current report warrant further study of the 9th and 10th TMDs to define their roles in PCFT function.

GRANTS

This work was supported by National Institutes of Health Grant CA-82621 (to I. D. Goldman).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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