Role of the Amino Acid 45 Residue in Reduced Folate Carrier Function and Ion-Dependent Transport as Characterized by Site-Directed Mutagenesis

Materials and Methods

Chemicals.

[3′,5′,7-³H](6S)-5-CHO-THF and [3′, 5′,7-³H](6S)-5-methyltetrahydrofolate (5-CH₃THF), obtained from Moravek Biochemicals (Brea, CA), and [3′,5′,7-³H]MTX, obtained from Amersham (Arlington Heights, IL), were purified by HPLC. Unlabeled MTX and 5-CHO-THF, provided by Lederle (Carolina, Puerto Rico), as well as folic acid and 5-CH₃-THF, obtained from Sigma (St. Louis, MO), were also purified by HPLC. Silicon and mineral oil used for the specific surface binding assay were purchased from Aldrich Chemical (Milwaukee, WI). All other reagents were of the highest purity available from commercial sources.

Cell Culture Conditions.

L1210 leukemia cells and the MTX^rA line, which lacks endogenous RFC1 function (Schuetz et al., 1989; Brigle et al., 1995), were grown in RPMI 1640 medium containing 2.3 μM folic acid, supplemented with 5% bovine calf serum (HyClone, Logan, UT), 2 mM glutamine, 20 μM 2-mercaptoethanol, penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C in a humidified atmosphere of 5% CO₂.

Site-Directed Mutagenesis of RFC1.

Site-directed mutagenesis was carried out according to the QuickChange protocol from Strategene (La Jolla, CA). Wild-type RFC1 cDNA was isolated from L1210 cells and directionally cloned into the mammalian expression vector pcDNA3.1(+) (Invitrogen, San Diego, CA) in the same way as reported for the mutated RFC1-E45K (Zhao et al., 1998b). The resulting vector was used as the template, and two completely complementary 24-mer oligonucleotides, which carry the targeted nucleotide changes in the middle region, were used as the primers for the QuickChange protocol. Sense primers containing the degenerate codons for amino acid 45 are listed in Table1. The codons for the amino acids introduced were chosen so that the single HindIII restriction site in the RFC1 coding region was eliminated to facilitate screening for positive clones. The RFC1 expression vector (30 ng), a pair of primers (each 150 ng), dNTP (each 10 ng), and Pfupolymerase (2.5 U) were dissolved in 50 μl of reaction buffer. The temperature was cycled 17 times for 30 s at 95°C, 1 min at 55°C, and 14 min at 68°C. DpnI (10 U) was added and incubated at 37°C for 1 h to digest the parental supercoiled dsDNA. The mixture (1 μl) was transformed into 50 μl DH5α competent cells (Life Technologies, Rockville, MD). Plasmids harboring the mutations were identified by restriction analysis withHindIII and then confirmed by DNA automated sequencing, in the Albert Einstein Comprehensive Cancer Center DNA Sequencing Facility, to be free of any mutations in the RFC1 coding region that might be introduced by Pfu polymerase. The plasmids were then ready for transfection.

Transfections.

MTX^rA (1 × 10⁷ cells) were electroporated (250 V, 200 μf) with 50 μg of nonlinearized pcDNA3.1(+) harboring the mutated RFC1 cDNA in a final volume of 800 μl of serum-free RPMI 1640 medium. Cells were then diluted in 20 ml of complete RPMI 1640 medium, allowed to recover for 48 h, adjusted to 2 × 10⁵ cells/ml in medium containing G418 (750 μg/ml of active drug), and then distributed into 96-well plates at ∼4 × 10⁴ cells/well. About 20 surviving clones were picked for each transfection and expanded in the presence of G418. After initial screening for MTX growth inhibition, one clone with the lowest MTX IC₅₀ was chosen for each mutation and subjected to further study. The transfectants were maintained in RPMI 1640 medium containing 750 μg/ml of G418.

Northern Analyses.

Total RNA was isolated from the transfectants, MTX^rA, and L1210 cells using the TRIzol reagent (Life Technologies). RNA (20 μg) was resolved by electrophoresis on 1% agarose gels containing formaldehyde. Transfers and hybridizations were performed as described previously (Zhao et al., 1997). Transcripts were quantitated by PhosphorImager analysis of the hybridization signals and normalized to β-actin.

Analysis of Specific 5-CH₃-THF Binding to the Cell Surface as a Measure of RFC1 Membrane Expression.

Cells were harvested, washed twice with, and resuspended in, HEPES (190 mM HEPES, 5 mM KCl, 2 mM MgCl₂, 5 mM glucose, pH 7.4) to a density of 5 × 10⁷ cells/ml. Portions of 200 μl were incubated with 1 μM [³H]5-CH₃-THF (specific activity of 800 dpm/pmol) at 0°C for 10 min in the presence or absence of a 10-fold excess of unlabeled 5-CH₃-THF. Bound and free ligands were then separated by centrifugation of a 180-μl aliquot through 150 μl of a mixture of silicon and mineral oil (9:1) in a 5.8 × 47.5-mm plastic (0.4 ml) microfuge tube at 14,000 rpm for 15 s. The tip of the tube, which contained the cell pellet, was cut off, dropped into a glass scintillation vial, and the contents digested with 0.5 ml 1 N KOH following which scintillation fluid was added and radioactivity determined. Other tared tubes, containing pellets with an equal number of cells, were weighed after drying to determine dry weight so that surface binding could be expressed as nmol/g dry wt. of cells.

For 5-CH₃-THF binding kinetics, six substrate concentrations, 0.05, 0.1, 0.2, 0.5, 0.75, and 1.0 μM, were used in the absence and presence of 10 μM unlabeled 5-CH₃-THF. Under these conditions, the binding constant (K _b) for 5-CH₃-THF in MTX^rA-R16 cells, which overexpress wild-type RFC1 (Brigle et al., 1995), was 0.11 ± 0.023 μM (n = 3), one-ninth the substrate concentration used in the assay, a value similar to what had been reported previously (Henderson et al., 1980). The binding assay provides an accurate comparison of RFC1 expression in MTX^rA-E45K, MTX^rA-E45Q, and MTX^rA-E45R cells in which the influxK _i, a reflection of substrate binding, is unchanged (see result sections). For E45L, the binding K_d is 0.39 ± 0.06 μM (n = 3). Hence, the assay underestimates carrier expression by ∼30%.

Transport Studies.

Influx measurements for all the folate compounds were performed by methods described in detail previously in HEPES-buffered saline (HBS) (20mM HEPES, 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 5 mM glucose, pH 7.4) or HEPES buffer (190 mM HEPES, 5 mM KCl, 2 mM MgCl₂, 5 mM glucose, pH 7.4) (Zhao et al., 1997). K⁺-HEPES-sucrose buffer (20 mM HEPES, 235 mM sucrose, pH 7.4) and Mg²⁺-free HBS buffer (20 mM HEPES, 140 mM NaCl, 5 mM glucose, pH 7.4) were also used in some experiments. Before transport studies with the transfectants, cells were expanded for only three to seven doublings without further addition of G418 to ensure that expression of the mutated RFC1 was maintained.

Results

Expression of E45 Site-Directed Mutants in MTX^rA Cells.

The cDNAs of the mutated RFC1s were transfected into the murine L1210 leukemia cell line, MTX^rA, which lacks endogenous RFC1 function due to a proline for alanine substitution at amino acid 130 within the fourth transmembrane domain (Schuetz et al., 1989; Brigle et al., 1995). One transfected clone, with the greatest restoration of sensitivity to MTX for each mutant, was identified and used for further detailed studies. The nomenclature for these cell lines is MTX^rA followed by the specific amino acid substitution. Figure1 shows a representative Northern blot analysis. The radioactive blots were quantitated directly by PhorphorImage analysis, and RFC1 mRNA levels were normalized to β-actin and a wild-type L1210 cell level of 1. The amount of RFC1 transcript in the transfectants derived from the expression vector was higher than that of L1210 cells by factors of 2.6 to 8.0. Consistent with previous determinations, RFC1 mRNA in recipient MTX^rA cells was about half that of L1210 cells (Brigle et al., 1995; Zhao et al., 1998a, 1998b).

Specific 5-CH₃-THF binding to membrane RFC1 of L1210, MTX^rA, and transfected cells was assessed in HEPES buffer along with MTX^rA-R16 cells. The latter, transfected with wild-type RFC1 cDNA, overexpresses carrier (Zhao et al., 1997). 5-CH₃-THF specifically bound to MTX^rA-R16 cells was nearly 3.5 times greater than that of L1210 cells (Fig. 2); binding to MTX^rA cells (0.19 ±0.06 nmol/g dry wt.) was half that of L1210 cells, consistent with the results reported previously (Brigle et al., 1995). Specific binding was detected in all lines transfected with mutated carriers in excess of the level in the recipient MTX^rA line. When the background level in MTX^rA cells is subtracted from the level of specific binding in the transfectants as indicated in Fig. 2, the expression of mutated carriers ranged between levels about one-half to nearly equal that of L1210 cells.

Folate Influx in Transfectants.

Table2 summarizes initial uptake rates for MTX, 5-CH₃-THF, and 5-CHO-THF in all cell lines studied. The most active of the mutated carriers was E45Q. In fact, transport of 5-CHO-THF and 5-CH₃-THF in MTX^rA-E45Q cells was nearly three times faster than that in L1210 cells and far exceeded the rate of MTX transport. For the E45L carrier, influx of the reduced folates was comparable with that mediated by wild-type RFC1; MTX influx was ∼50% less. RFC1 expression in both E45Q and E45L transfectants was comparable with the wild-type level in L1210 cells. For other mutated carriers, influx was decreased and residual activity of each of the folate substrates among the mutants varied considerably. The lowest activities were observed for the asparate and tryptophan substitutions. Although the membrane binding assay underestimates the level of E45W expression due to its high influx K _i (see Table3 below), residual transport activity measured was <10% that mediated by wild-type RFC1

The Inhibitory Effect of 5-CHO-THF, 5-CH₃-THF and Folic Acid on MTX Influx Mediated by the Mutated Carriers.

Inhibition of MTX influx by reduced folates and folic acid was assessed as a prelude to the determination of the affinity of the mutated carriers for these substrates. Three different concentrations of each inhibitor were used: the wild-type influx K _t, one-fifth the influx K _t, and five times the influxK _t, along with a control in which no inhibitor was added. As shown in the upper panel of Fig.3, the potency of inhibition of MTX influx by 5-CHO-THF was markedly altered in the transfectants as compared with L1210 cells. Inhibition of MTX influx was 1) not abolished even at a 5-CHO-THF concentration of 25 μM in MTX^rA-E45W cells, 2) markedly decreased in MTX^rA-E45D cells, 3) markedly increased in MTX^rA-E45Q and MTX^rA-E45R cells whereas 4) unchanged in MTX^rA-E45L cells. The pattern of inhibition of MTX influx by 5-CH₃-THF resembled that of 5-CHO-THF, but 5-CH₃THF inhibition appeared greater in MTX^rA-E45W and weaker in MTX^rA-E45R relative to L1210 cells (middle panel of Fig. 3). Folic acid inhibition of MTX influx in the transfectants differed from that observed with either 5-CHO-THF or 5-CH₃-THF (lower panel). At all concentrations, inhibition was markedly increased in MTX^rA-E45Q, MTX^rA-E45R, and MTX^rA-E45L cells as compared with that of L1210 cells, whereas it was roughly comparable in MTX^rA-E45D and MTX^rA-E45W cells. These data were then used to identify appropriate concentrations for determination of inhibitory constants for these folates as indicated below.

Influx Kinetics.

As indicated in Table 3, the influxK _t for MTX was increased in all mutated carriers except E45R; the greatest increase inK _t was observed with E45D. Influx kinetics could not be determined for E45W due to the very low level of transport activity. The influx V _max was increased by a factor of 2 in E45Q, but influx was minimally decreased (Table 2), attributed to the accompanying 2-fold increase in influxK _t. The V _maxwas unchanged in E45L and was decreased by 50–60% in the E45R and E45D transfectants as compared with that of L1210 cells; the latter can be explained, at least in part, by a comparable decrease in carrier protein expression (Fig. 2, Table 2).

Influx K _i for the other folates were then assessed by Dixon analysis and compared with wild-type RFC1 (Table 3). The influx K _i for 5-CH₃-THF in E45K was also determined, and together with K _t for 5-CHO-THF influx andK _i for folic acid, which were reported previously (Zhao et al., 1998b), are listed for the purpose of comparison. As observed earlier for the E45K, theK _i for 5-CHO-THF influx was decreased by a factor of ∼4, the 5-CH₃-THF influxK _i was essentially unchanged, and the folic acid K _i was decreased by a factor of 6 to 10 in the E45Q and E45R. The E45L mutation resulted in a 2- and 4-fold decrease in carrier affinity for 5-CHO-THF and 5-CH₃-THF, respectively. For the remaining mutant carriers, E45D and E45W, the influx K _i was increased for both reduced folates, although the loss of affinity was much greater for 5-CHO-THF (9- and 29-fold, respectively) than 5-CH₃-THF (3- and 10-fold, respectively). TheK _i for folic acid was greater for E45D than wild-type RFC1 but was approximately one-half the wild-typeK _i for the E45W and E45L mutants. Hence, amino acid substitutions at the 45 position are associated with profound changes in binding of the natural folates to RFC1.

Although the high affinity of E45Q is associated with high influx rates for 5-CHO-THF and 5-CH₃-THF, influx mediated by E45R is less than that of L1210 cells (Table 2) despite the fact that the influx K _t is comparable with E45Q, suggesting that the mobility of the mutant carrier-substrate complex is less than that for the wild-type carrier complex. In contrast, despite a 2- to 4-fold increase in K _t for 5-CHO-THF and 5-CH₃-THF for the E45L mutant, rates of transport are comparable with L1210 cells (Table 2), consistent with an increase in mobility of the mutated carrier-substrate complex.

The Anion Dependence of MTX Influx in E45 Mutants.

Folate transport in L1210 cells is inhibited by virtually all inorganic and organic anions and influx is stimulated by the removal of chloride (Goldman, 1971; Henderson and Zevely, 1988). In contrast, activity of the E45K mutant was markedly reduced in the absence of chloride (Zhao et al., 1998b). To determine the effect of the other amino acid substitutions on the anion-dependence of transport, all but 9 mM extracellular chloride was replaced by HEPES and MTX influx was assessed. MTX influx in HEPES buffer was increased ∼150% in L1210 cells and all other transfectants except for E45Q in which the increase was ∼50% and E45R in which the effect was reversed with a 60% decrease in influx (Fig. 4). The latter was qualitatively the same as observed for E45K, and the increase in MTX influx in MTX^rA-E45R as extracellular chloride was increased, similar to what was observed for 5-CHO-THF influx in MTX^rA-E45K cells (Fig.5) (Zhao et al., 1998b).

Discussion

Previous studies from this and other laboratories have begun to elucidate some of the structure-function properties of the reduced folate carrier. These insights have derived largely from analyses of the functional consequences of mutations in RFC1 in cell lines obtained under antifolate selective pressure augmented by chemical mutagenesis (Zhao et al., 1999). The first transmembrane domain has been identified as one important site for mutations that result in antifolate drug resistance with highly selective changes in the spectrum of carrier affinities for folate substrates and/or mobilities of the carrier-substrate complex (Tse et al., 1998; Zhao et al., 1998a,1998b). The current study was undertaken to further clarify the structural basis for the functional changes that occurred when lysine was substituted for glutamate at amino acid 45—increased affinity for folic acid and 5-CHO-THF, decreased affinity for MTX, and the induction of an obligatory requirement for small inorganic anions to allow carrier function (Zhao et al., 1998b).

Based on the substitution of glutamate with amino acids of different charge, polarity, and size, the data indicate that preservation of the negative charge at amino acid 45 is not an essential element in RFC1 function. In fact, replacement of glutamate with the uncharged glutamine, or positive-charged arginine, resulted in increased affinity for 5-CHO-THF, a similar or greater affinity for 5-CH₃-THF, and a markedly increased affinity for folic acid as compared with wild-type carrier, as was observed also for the lysine-substituted RFC1 (Zhao et al., 1998b). It is of interest that the loss of the negative charge has the most profound salutary effect on folic acid binding with up to a 10-fold decrease in influxK _t. Indeed, the only substituted amino acid that produced a carrier with a slightly lower affinity for folic acid than wild-type RFC1 was aspartate, and increased binding occurred even with the tryptophan substitution despite its bulk. Another bulky substitution of phenylalanine for isoleucine at amino acid residue 48 without a change in polarity produced an even greater increase in affinity for folic acid with a 20-fold decline in influxK _t (Tse et al., 1998; Tse and Moran, 1998). Hence, it would appear that this region in the first transmembrane domain plays an important role in determining the very low affinity of wild-type carrier for this oxidized folate.

Carrier functional activity is determined by affinity for the substrate, carrier mobility, and the level of carrier protein expressed. Carrier protein levels at the cell surface were assessed using a specific binding assay for 5-CH₃-THF. All of the mutated carriers were expressed and processed to the cell membrane and the measured levels were within a factor of ∼2 of the wild-type carrier. In the absence of murine RFC1 antibody, the efficacy of processing and extent to which carrier might be sequestered along trafficking route cannot be assessed. In general, the level of mutated protein present at the cell membrane (Fig. 2) was far lower than expected based on a comparison of RFC1 mRNA (Fig. 1) in the transfectants relative to L1210 cells. Discrepancies were also observed for human RFC1 transfectants in which the amount of RFC1 mRNA or protein far exceeded relative levels of RFC1 activity (Wong et al., 1997, 1998). However, these studies go beyond protein identification to the assessment of specific folate binding to RFC1 at the cell surface.

Based on the level of protein detected and the measured influxV _max for MTX, there were some substrate-dependent differences in the mobility of the carrier-substrate complexes, less or greater than wild-type RFC1, that accompanied changes in influx K _t. This is in contrast to the functional changes associated with the S→N mutation, at the adjacent amino acid 46, in which there was a highly selective decline in carrier mobility that allowed a much greater rate of translocation of carrier loaded with reduced folates than with MTX without any change in the spectrum of affinities for these substrates (Zhao et al., 1998a).

The further demonstration that only arginine, of all the substituted amino acids, reproduces the anion dependence of MTX influx observed with lysine clarifies the basis for this phenomenon (Zhao et al., 1998b). Induction of anion-dependent transport in the E45K mutant was associated with a decrease in influx V _max, consistent with the loss of carrier mobility in the absence of chloride. The data here indicate that it is the introduction of the positive charge at amino acid 45, and not the loss of the negative charge at this site, that results in the loss of function. Therefore, the impact of chloride is likely due to neutralization of the positive charge associated with the lysine or arginine substitution.

The very marked decrease in influx of MTX and the other folates associated with the aspartate substitution was of particular interest because this amino acid is of the same charge as, and is close in size to, glutamate. This resulted in a fall in affinity for all the folates. The loss of a methylene group in the side chain by this substitution relocates the carboxyl group by about 1.5 Å, a small change that is apparently sufficient to alter the interaction among this residue, its neighboring amino acids, and folates. The impact of the aspartate for glutamate substitution on carrier transport function has been well documented in other systems. For instance, the same substitution at amino acid 121 in the fourth transmembrane domain of the myo-inositol/H symporter nearly abolished transport activity (Seyfang et al., 1997). When this mutation was present in the glutamate transporter at residue 404, potassium-coupled transport was abolished, and glutamate exchange became obligatory for carrier function (Kavanaugh et al., 1997).