Delivery of Folates to the Cytoplasm of MA104 Cells Is Mediated by a Surface Membrane Receptor That Recycles*

MA104 cells, as well as several other rapidly divid- ing tissue culture cells, have a folate-binding protein associated with their cell surface. The protein has the properties of a membrane receptor: (a) 5-methyl[’H] tetrahydrofolic acid binds with high affinity (Kd = 3 nM); (b) the protein is an integral membrane protein; (c) it appears to deliver physiological concentrations of 5-methyl[’H]tetrahydrofolic acid to the inside of the cell; (d) binding activity is regulated by the concentration of folate within the cell. To better understand the mechanism of action of this receptor, we have studied the pathway of folate internalization. We present evidence that during internalization: (a) folate binds to the membrane receptor; (b) the ligand-receptor com-plex moves into the cell; (c) the ligand is released from the receptor in an acidic intracellular compartment and moves into the cytoplasm; and (d) the unoccupied receptor returns to the cell surface. plasma a formidable barrier to the movement of charged molecules, and particulate material from the extracellular space into the cell. and shifted 37 there was a reciprocal disappearance of surface-bound folic acid and an increase in the number of unoccupied receptors at the cell surface. These results suggest that the membrane fraction contains two populations of receptors:

The plasma membrane presents a formidable barrier to the movement of charged molecules, macromolecules, and particulate material from the extracellular space into the cell. Important metabolic ions and molecules that are taken up by cells depend upon specific transport systems present at the plasma membrane. These transport systems fall into two general categories: ( a ) membrane channels or carriers that vectorially move specific ions and low molecular weight solutes (1) and ( b ) endocytic vesicles that internalize macromolecules and particulate material (2). Each system utilizes a specific set of membrane proteins that bind with high affinity and specificity the molecule or ion that is being transported. For each type of transport activity, there is only a rudimentary understanding of the cellular organelles involved in the process.
One class of charged molecules that animal cells must obtained from their environment are the folates (3). Tetrahydrofolic acid is the fully reduced physiological form of folate utilized by cells for the biosynthesis of methionine, serine, deoxythymidylic acid, and purines (3). Cells obtain this folate from 5-methyltetrahydrofolic acid, which crosses the plasma membrane, donates the 5-methyl group to homocysteine during methionine synthesis, and acquires multiple glutamic acid residues. The last step is catalyzed by folylpolyglutamate synthetase, a cytoplasmic enzyme (4). Several studies have shown that in the presence of 1-10 p~ folate tissue culture cells will rapidly accumulate folate and that eventually it becomes polyglutamated. These studies suggest that folates utilize a membrane carrier to cross the membrane (5). This carrier is saturable (Kd = 1-10 p~) and internalizes folate with a VmaX of 1-12 nmol/min/g dry weight (5). The molecular basis of this carrier activity has not been determined.
Kolhouse and co-workers (6) as well as Kamen and Capdevila (7) have found that folate-depleted tissue culture cells express a folate-binding protein on their cell surface that is immunologically related to the soluble folate binder found in plasma and milk of many animals (8)(9)(10)(11)(12). Like the soluble protein, the membrane folate binder has a high affinity for 5methyltetrahydrofolic acid but at least a 10-fold higher affinity for folic acid and a 5-10-fold lower affinity for methotrexate (7). In addition, folates are readily released by acid treatment (13).
Folate binding activity on the surface of MA104 cells is maximal when cells are grown in the absence of folate. Whereas at 4 "C folate-depleted cells accumulate 1 pmol/106 cells, which can be released by brief acid treatment, at 37 "C cells will accumulate 5-7 pmol/106 cells (7). Folate accumulation at 37 "C is linear for approximately 4 h before gradually leveling off; by 24 h of incubation, accumulation virtually ceases (7). 5-Methyltetrahydrofolic acid binding as well as uptake is inhibited by antibodies that inhibit binding to the soluble folate-binding protein (7).
The membrane-bound folate-binding protein is a likely candidate for a membrane receptor that mediates the passage of physiological concentrations of folate into the cytoplasm of the cell. In this report we present evidence that this receptor is responsible for delivering 5-methyltetrahydrofolic acid to the cytoplasm of the cell and that it does so by cyclicly moving from the cell surface into an intracellular membrane compartment and back to the cell surface. (20 Ci/mmol, MT783) was from Moravek Biochemicals (City of Industry, CA). The reduced folates, L-5-methyltetrahydrofolic acid and ~-5-methyl[~H]tetrahydrofolic acid (20 Ci/mmol) were synthesized and purified as previously described (7,14). Monensin (475895) and nigericin (481990) were purchased from Behring Diagnostics. Culture flasks (T-25 and T-75) were from Costar (Cambridge, MA). Ammonium chloride (A-4514), chloroquine (C-6628), folic acid (F-7876), DL-5-methyltetrahydfolic acid (M-0132), tetrahydrofolic acid (T3125), folinic acid (F7878), p-aminobenzoyl-1-glutamic acid (A- (15). Cell Culture-MA104 cells, a monkey kidney epithelial cell line, was grown continuously as a monolayer in M199 supplemented with 5% (v/v) fetal calf serum. This medium contained approximately 20 nM folate and 0.68 M glutamine. Cells for each experiment were set up according to a standard format. On day zero, either 2.5 X lo6 cells or 1.5 X lo6 cells were seeded into a T-25 or T-75 culture flask, respectively, and grown for 4 or 5 days in folate-free M199 supplemented with 0.68 mM glutamine and 5% (v/v) charcoal-treated fetal calf serum (medium A). This medium contained less than 1 nM folate as determined by radiolabeled binding assay for folate (7,16). For uptake studies, medium was replaced with M199 without folate, containing 20 mM Hepes and 0.68 mM glutamine (medium B), and additions were made directly to the culture flask.

Materials-Medium
Measurement of Folate Binding and Accumulation-To measure folate receptor activity, on either day 3 or 4 (see figure legends) of cell growth, the medium was removed by aspiration and 1.5 ml of medium B were added to each T-25 flask. The indicated type of radiolabeled folate was added to the dish in the presence and absence of 100-fold excess unlabeled folate, and the cells were incubated as indicated in the figure legends. At the end of the incubation, cells were chilled on ice, and the medium was removed by aspiration. After rinsing with 5 ml of ice-cold DPBS, folate was released from the cells by washing rapidly for 30 s with 2 ml of ice-cold acid saline (0.15 M NaC1, adjusted to pH 3 with glacial acetic acid) followed by a rinse with 1 ml of cold DPBS. Increasing the time of exposure to acid from 0.5 to 10 min did not increase the amount of 3H released. The radiolabeled folate in the acid saline plus the 1 ml of cold DPBS wash represented acid-releasable folate. The cells were released from the culture flask by incubating with 1 ml of the trypsin-EDTA for 5 min at 37 "C and then rinsing twice with 1 ml of DPBS (7). The trypsin-EDTA suspension and washes were combined and the amount of tritium, which corresponded to acid-resistant folate, was determined. Radioactivity was measured by liquid scintillation counting using a Packard Tri-Carb counter that had an approximate efficiency of 40% in a scintillation fluid described previously (7). All values represent specific binding or accumulation, which was calculated by subtracting the value for 13H]folate in the presence of unlabeled folate (nonspecific) from the value in the absence of unlabeled folate (total). Each value shown represents the average of duplicate incubations. Nonspecific binding was never greater than 5-10% of specific binding. Where indicated, the cell number in the trypsin-EDTA suspension was determined with a hemocytometer. Subcellular Fractionation-At the end of the incubation with radiolabeled folate, cells were washed three times with 10 ml of ice-cold DPBS, released from the T-75 flask by incubation with 3 ml of trypsin-EDTA for 5 min at 37 "C, and pelleted in a clinical centrifuge at 500 X g for 3-5 min at room temperature. Cells were hypotonically swelled and homogenized according to the procedure of Atkinson (17). The pellet was suspended in 20-25 times its volume of ice-cold 10 mM Tris-HC1, pH 8, containing 15 mM iodoacetic acid and the cells allowed to swell on ice for 7 min. The cells were homogenized by 25 (up and down) slow strokes of a borosilicate glass Dounce homogenizer with a Teflon pestle. After homogenization, the suspension was adjusted to 3 mM MgC12 and 10 mM NaCl to stabilize the nuclei. Each homogenate was examined by light microscopy for the presence of unbroken cells. The homogenate (approximately 2 ml) The abbreviations used are: Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; DPBS, Dulbecco's phosphate-buffered saline; HPLC, high performance liquid chromatography. was layered on 8 ml of Percoll/sucrose (1.055 g/ml) prepared by combining 14.1 ml of Percoll with 10 ml of 2.5 M sucrose in water to a total volume of 100 ml and centrifuged at 60,000 X g for 45 min in a 70.1 TI rotor (Beckman Instruments). Fractions (0.5 ml) were collected from the bottom of the gradient (bottom fraction = 20; top fraction = 1) by inserting into the gradient a 2.5-inch, 18-gauge needle and aspirating with a 1-ml syringe. Before each fraction was counted by scintillation counting, a sample was removed and the refractive index measured with a Bausch and Lomb refractometer. The refractive index was used to calculate the density of the fraction. Acid and neutral 5'-nucleotidase activity was measured by the spectrophometric methods of Sabbadini and Okamoto (18).
Preparation of Cell Extracts and Cytoplasm for Analysis of Radiolabeled Folate-To analyze radiolabeled folate in cells, cells were grown in T-75 flasks, depleted of intracellular folate, and then incubated in the presence of either 40 nM 5-methyl[3H]tetrahydr~folic acid or 5 nM [3H]folic acid for the time indicated. Cells were then harvested by washing the monolayers twice with ice-cold DPBS, adding 2.0 ml of buffer A (25 mM ammonium acetate, pH 4.5, 10 mM EDTA), and subjecting the sample to one round of freeze-thawing at -80 "C. The lysed cells were withdrawn, with flasks washed with 2 ml of buffer A, and the two samples pooled. The pooled samples were adjusted to 1% (v/v) j3-mercaptoethanol and sonicated for 30 s with a Bronson sonicator (model 200) using a microtip and a 50% of maximum power setting. The sonicate was centrifuged at 20,000 X g for 15 min at 4 "C in a Beckman 52-21 centrifuge using a 5-20 fixed angle rotor. The supernatant fraction was saved and the pellet washed by suspending it in buffer A containing 1% j3-mercaptoethanol and centrifuging once again. The two supernatant fractions were pooled and lyophilized.
To analyze radiolabeled folates contained in the cytoplasm of cells, fractions 1-4 from the Percoll gradient ( Fig. 2) were pooled (2 ml) and adjusted to 1% (v/v) (3-mercaptoethanol. The sample was sonicated for 2 min as described above, and 1 ml of 1% 0-mercaptoethanol solution was adjusted to pH 3 with glacial acetic acid was added per 2.5 ml of sample. The sample was then centrifuged at 20,000 X g as described above. The supernatant fraction was saved and the pellet washed by suspending in 1 ml of 0.1 M potassium phosphate, pH 7.0, and centrifuging once again. The supernatant fractions were pooled and prepared for HPLC analysis. Preparation for HPLC Analysis-The lyophilized cell lysate was dissolved in 0.1-0.3 ml of 1% (v/v) j3-mercaptoethanol to which was added 9 volumes of ice-cold acetonitrile. The sample was centrifuged for 3 min in a Beckman model B microfuge at 4 "C to pellet precipitated protein. The acetonitrile was evaporated from the supernatant fraction by incubating for 30-40 min at 4 'C under a stream of nitrogen. The sample was subjected to a second microfuge centrifugation and the supernatant fraction filtered through a 0.45-pm filter before analysis by HPLC.
To the pooled supernatant fractions prepared from the cytoplasmic fractions were added 9 volumes of ice-cold acetonitrile. Following centrifugation in a microfuge as above, the acetonitrile was removed with a nitrogen stream and the sample lyophilized. The lyophilized sample was dissolved in 0.3 ml of 1% (v/v) P-mercaptoethanol and centrifuged in the microfuge. The supernatant fraction was saved and the pellet suspended in 0.2 ml of 1% j3-mercaptoethanol and centrifuged again. The supernatant fractions were pooled and filtered through a 0.45-pm filter before analysis.
Analysis of Folylpolyglutamates-HPLC analysis of polyglutamates was performed as previously described for methotrexate polyglutamates (19). Analysis was done by injecting 10,000-20,000 3H-labeled counts/min of sample (maximum volume was 200 pl) into a Waters CIS pBondapak column (3.9 mm X 15 cm) and eluting with a 20-40% acetonitrile gradient containing 5 mM Pic-A using a Waters automated gradient controller (model 680) on a program setting of 1. This setting gives a more convex gradient than used previously (19). Before each sample was run on the column, a set of polyglutamate standards was first run to mark the elution position of p t e r o y l -G l~'~~-~*~ and ' .
Analysis of Radiolabeled Folates-To identify the type of folate present in either the cell extract or cytoplasmic fraction, the sample was treated with human plasma y-glutamylcarboxypeptidase (conjugase) so that all folates would be in the monoglutamate form. Specifically, to a 0.15-ml sample (15,000-50,000 3H-labeled counts/min) was added 0.24 ml of potassium phosphate (pH 4.5), 0.1 ml of human plasma, 7 p1 of 14 M 8-mercaptoethanol, and 10.5 nmol of DL& methyltetrahydrofolic acid (in 2.7 pl) as an internal standard. The sample was incubated for 60 min at 37 "C in the dark. The reaction was stopped by placing on ice and adding 0.45 ml of ice-cold aceto-nitrile. The sample was centrifuged in a Beckman microfuge B at 4 "C and the supernatant fraction saved. The pellet was suspended in 0.2 ml of acetonitrile and centrifuged again. The two supernatant fractions were pooled, the acetonitrile was removed and passed through a 0.45-pm filter prior to analysis.
Folates were analyzed on HPLC by the method of Duch et al. (ZO), using a Waters Cle pBondapak column (3.9 mm X 15 cm) and an isocratic elution (1 ml/min) with buffer that contained 20% methanol, 5 mM Pic-A, and 10 mM sodium phosphate, pH 4.5. Prior to running each sample, the column was calibrated by running known folate standards in the same buffer system (see Fig. 3B). Each sample contained 5-10 nmol of either folic acid or 5-methyltetrahydrofolic acid as standards. We found that under the reaction conditions described, spectral quantities (5-10 nmol) of folic acid pentaglutamate were completely reduced to the monoglutamate form.
The overall recovery of 3H from the time the cells were extracted to recovery from either HPLC analysis was 275%.

RESULTS
Cellular Accumulation of pH]Folate(s)-Folic acid inhibits 5-methyl[3H]tetrahydr~folic acid accumulation in MA104 cells and also competes for binding of this ligand to the folate receptor (7). To see if cells would take up folic acid, we compared the time-dependent uptake of [3H]folic acid with 5methyl[3H]tetrahydrofolic acid. As shown in Fig. 1 To be sure that the radioactivity accumulated by the cells was not a contaminant or breakdown product, we analyzed by HPLC the [3H]f~lic acid and 5-methyl[3H]tetrahydr~folic acid used in the incubation media and the cellular 3H taken up by cells incubated with each ligand for 17 h at 37 "C. Although the radiolabeled folate added to the dish was >95% pure, at the end of the incubation the [3H]folic acid and 5methyl[3H]tetrahydrofolic acid in the media were 95 and 88% chromatographically pure, respectively. On the other hand, analysis of the cell extract from either set of cells revealed that 80% of the radioactivity co-eluted with a known folate standard. (Of the remaining 20%, one-half co-eluted with p - aminobenzoyl-1-glutamic acid and the remainder did not coelute with any standard.) We next used a cell fractionation procedure to analyze the subcellular distribution of both [3H]foli~ acid and 5methyl[3H]tetrahydrofolic acid (Fig. 2) in cells that had been incubated with the radiolabel at 37 "C for 17 h. Virtually all of the [3H]folic acid taken up by cells migrated as a single peak on a Percoll gradient at a density of 1.044 g/ml (fractions 7-10). This peak co-migrated with 5'-nucleotidase activity, a plasma membrane marker (data not shown). When we analyzed cells that had been incubated with 5-meth~l[~H]tetrahydrofolic acid, the 1.044 g/ml peak contained the same amount of folate (2 pmol/106 cells) but in addition, there was 4 pmol/106 cells in a lighter fraction (fractions 1-4). When we measured acid 5'-nucleotidase activity, a lysosomal marker (18), peak activity was in fractions 13-16. According to the Percoll manufacturer's handbook (Pharmacia LKB Biotechnology Inc.), the lighter fraction (density of 1.00-1.040) corresponds to the cell cytoplasm.
The radioactivity in the cytoplasmic fraction (fractions 1-4, Fig. 2) of cells incubated with 5-methyl[3H]tetrahydr~folic acid for 17 h at 37 "C was assayed for the presence of polyglutamates by gradient HPLC (Fig. 3A). Approximately 47% of the radioactivity eluted from the column at a position between the folylpentaglutamate and the folylheptaglutamate standards, indicating that the predominant form was a folylhexaglutamate. Five percent co-eluted with a folyltriglutamate standard and 30% with a folylmonoglutamate standard. The remaining 17% of the radioactivity (primarily fraction 3) did not co-elute with any folate standard and, therefore, most At the end of the incubation the cells were rinsed with ice-cold DPBS, released from the dish, and homogenized as detailed under "Experimental Procedures." The homogenate was layered on a solution of Percoll (1.05 g/ml) and centrifuged at 60,000 X g for 45 min. Each fraction (0.5 ml) was analyzed for 3H content. Fraction 8 had a density of 1.044 g/ml, corresponding to the known density of plasma membrane. In addition, it contained the peak activity for the plasma membrane marker, 5"nucleotidase measured at neutral pH. Fractions 1-4 correspond to the cytoplasmic fraction. The specific activity for [3H]folic acid and 5-methyl[3H]tetrahydr~folate was 9,500 and 8,800 cpm/pmol, respectively. Each of the membrane fractions (fractions 7-10) contained 2 pmol/106 cells of folate. In the presence of a 100fold excess unlabeled folate during the incubation period, cellular radioactivity was reduced by 94%. Cells were grown in T-75 flasks, which at the time the experiment was done contained approximately 3.5 x lo6 cells.  likely was a degradation product of 5-methyl[3H]tetrahydrofolic acid. If this non-folate radioactivity is subtracted from the total, then 83% of the tritium can be attributed to an authentic folate. Of this, 63% was a folylpolyglutamate. Another sample of the cytoplasmic fraction was treated with plasma conjugase and analyzed on isocratic HPLC to identify the type of folate (Fig. 3B). Approximately 66% of the radioactivity co-chromatographed with a 5-methyltetrahydrofolic acid standard, 13% with a 10-formyltetrahydrofolic acid/tetrahydrofolic acid standard, and 21% with a p-aminobenzoylglutamate standard. The latter compound is a known degradation product of folate and most likely corresponds to the radioactivity found in fraction 3, Fig. 3A.
We also analyzed the cytoplasmic fraction to determine if the radioactivity was tightly bound to a macromolecule or particulate material. All of the radioactivity was removed by dialysis and migrated on Sephadex G-100 column with folate standards (data not shown). In separate experiments, we also assayed this fraction from folate-depleted cells for the pres- ence of folate binding activity and found none to be present.
Movement of Surface-bound PHlFolic Acid into an Acidresistant Compartment-Even though [3H]folic acid remained associated with the membrane fraction and did not significantly enter the cytoplasm at 37 "C ( Figs. 1 and 2)) only 50% could be released by a brief acid wash. The acid-resistant fraction most likely corresponds to [3H]folic acid in a membrane compartment that was protected from the acid treatment. To determine if [3H]folic acid could actively move from an acid-releasable to an acid-resistant compartment, we analyzed the effect of temperature on the acid releasability of [3H]folic acid initially bound to cells at 4 "C. Cells were incubated with [3H]folic acid at 4 "C for 0.5 h to saturate all available receptors, rinsed with folic acid-free buffer, and warmed to 37 "C for various times before assaying for amount of [3H]folic acid that was acid-releasable (Fig. 4). Without any warming, all of the radioactivity was released; however, with time there was a progressive loss of acid-releasable and a corresponding increase in acid-resistant [3H]folic acid. Equilibrium was reached after 2-4 h when approximately 50% was in either the acid-resistant or acid-releasable form; this ratio changed very little when cells were incubated for up to 12 h at 37 "C.
During the warm up period, there was not any t3H]folic acid released into the media; therefore, under these conditions all of the ligand remained membrane-bound. This allowed us to examine cells for the presence of any unoccupied receptors that might appear during the temperature-dependent transfer of folic acid to an acid-resistant compartment. Cells were chilled to 4 "C and incubated in the presence of 5 nM unlabeled folic acid to saturate all available surface binding sites. The cells were then washed and incubated at 37 "C for various times before measuring [3H]folic acid binding to the cell surface (at 4 "C). As seen in Fig. 5 , before warming to 37 "C, there was very little [3H]folic acid binding detected, indicating that all external receptors were occupied with unlabeled folic acid. With time at 37 "C, however, there was a progressive increase in surface binding, which reached ~5 0 % of maximal binding by 4 h. The kinetics of appearance of unoccupied receptor on the cell surface was similar to the kinetics of appearance of acid-resistant [3H]f~lic acid (compare Fig. 5 with Fig. 4). During the course of this experiment, the total number of [3H]folic acid binding sites did not change (Fig. 5).
The appearance of an unoccupied receptor on the cell surface at the same time as folic acid moved into an acidresistant compartment suggested that both occupied and unoccupied receptors could shuttle between the acid-resistant and acid-releasable compartments at 37 "C. Additional evidence was provided by a third experiment (Fig. 6). A set of cells was incubated with [3H]folic acid at 37 "C for 4 h, a time when binding to the membrane fraction is maximal. The same set of cells was then subjected to repeated cycles of acid washing at 4 "C followed by warming to 37 "C for 0.5 h in TABLE I Effect of nigericin and weak bases on folate accumulation MA104 cells were grown in the absence of folate for 4 days. The medium was replaced with medium B containing the indicated concentration of the drug and incubated for 30 min at 37 "C. Total cellular 5-methyl[3H]tetrahydrofoli~ acid was measured as described.
Value is the mean of two experiments (+lo%) where each experiment had duplicate flasks per treatment. Monensin Inhibits Delivery of 5-Methyltetrahydrofolic Acid to the Cytoplasm-Many receptors that internalize ligands enter a low pH endosomal compartment (21). The H' gradient across the membrane of these compartments is rapidly dissipated with monensin (22). To see if this ionophore affected folate accumulation, MA104 cells were incubated with 5methyl[3H]tetrahydr~folic acid at 37 "C, in the presence or absence of 25 FM monensin for various times. As shown in Fig. 7, after 4 h, untreated cells accumulated normal amounts of 5-methyl[3H]tetrahydr~folic acid. Monensin-treated cells, by contrast, only accumulated one-third this amount, which was equal to the amount of [3H]folic acid that cells could accumulate (see Fig. 1). When the 4 h monensin-treated cells were chilled and washed with acid buffer, 1 pmol/106 cells of 5-methyl[3H]tetrahydrofolic acid (50% of the total 3H associated with the cells) was released. In the same experiment, monensin did not have any effect on either the binding or uptake of [3H]folic acid (data not shown).
Monensin had its effect by altering the subcellular distribution of 5-methyl[3H]tetrahydr~folic acid (Fig. 8). Cells were incubated for 3 h at 37 "C in the presence of either [3H]folic acid, 5-methyl[3H]tetrahydrofolic acid, or 5-meth~l[~H]tetrahydrofolic acid plus monensin before homogenization and fractionation on a Percoll gradient. As shown previously (Fig.  2), virtually all of the [3H]folic acid was present in the membrane fraction (Fig. 8A). Likewise in the presence of monensin, all of the 5-methyl[3H]tetrahydr~folic acid was also in this fraction (Fig. €8). In the absence of monensin, although the amount of folate bound to the membrane fraction was equal to that found in the first two conditions, there was a 6-fold increase in the 5-methyl[3H]tetrahydr~folic acid present in the cytoplasmic fraction (Fig. 8C, fractions 1-4).
Both 50 FM chloroquine and 20 mM ammonium chloride, two weak bases that neutralize acidic compartments (23), also inhibited 5-methyl[3H]tetrahydrofolic acid accumulation (Table I) but not as effectively as monensin. However, when the chloroquine concentration was raised to 500 FM, 5-meth~l[~H] tetrahydrofolic acid accumulation was reduced to the same level as seen with monensin. In addition, 5 p~ nigericin, another ionophore that neutralizes acidic compartments, mimicked the effect of monensin on 5-methyl[3H]tetrahydrofolic acid uptake.

DISCUSSION
The most fundamental question about the folate receptor is whether or not it mediates the cellular accumulation of physiological folate. Cellular accumulation can be divided into two steps: (a) binding to the cell surface receptor followed by ( b ) delivery to an intracellular site. Therefore, to establish a function, we must be able to distinguish between these two steps. We used acid releasability at 4 "C to identify folates on the cell surface, and cell fractionation together with measuring folylpolyglutamation to identify folate that reached the cytoplasm.
Many ligand-receptor interactions are sensitive to pH (2). This property has been exploited by investigators to identify ligands that are bound to receptors at the surface of cells (2,24). Since acid treatment of the soluble folate binder is a standard method for releasing folates (13), we reasoned that low pH should release folate bound at the cell surface of MA104 cells and leave behind any folate that had entered an intracellular compartment, To verify this method, we measured the amount of folate that was acid-releasable from cells incubated with the ligand at 4 "C, conditions that prevent intracellular accumulation (7). We found that virtually all of the folate could be released. Moreover, when cells were incubated with 5-methyl[3H]tetrahydrofolic acid at 37 "C for up to 24 h, chilled to 4 "C, and treated with acid saline, the same amount of ligand was released. However, in this case there was always a substantial amount that remained associated with the cell. As judged from cell fractionation experiments, much of this 5-methyl[3H]tetrahydrofolic acid had been delivered to the cytoplasm of the cell (Fig. 2 A ) .
Folylpolyglutamate synthetase, the enzyme that adds glutamic acid residues to folate by a y-carboxyl linkage, is a cytoplasmic enzyme (4). Therefore, polyglutamation of folate should serve as a measure of the ligand that reached the cytoplasm of the cell. HPLC analysis of the cytoplasmic fraction showed that 63% of the folate radioactivity was in the form of folylpolyglutamate.
The Folate Receptor Mediates 5-Methyltetrahydrofolic Acid Accumulation-If the folate receptor is responsible for the uptake of 5-methyl[3H]tetrahydr~folic acid into folate-depleted cells, then inactivation of the receptor should prevent cellular accumulation. One way to inactivate the receptor is with antibodies. This was done by Antony et al. (25), who showed that antibodies capable of inhibiting folate binding to the soluble folate binder completely block the uptake of 5methyl[3H]tetrahydrofolic acid at 37 "C in KB cells. Another way is with a high affinity, competitive ligand such as folic acid. Previously we showed that folic acid blocks accumulation of 5-methyl[3H]tetrahydrofolic acid, even when present in the medium at one-eighth the concentration of the radiolabeled folate (7). The current results show that folic acid has this effect by inactivating the surface receptor.
[3H]Folic acid was only found tightly associated with the plasma membrane fraction and the amount bound was equal to the number of 5-methyl[3H]tetrahydrofolic acid binding sites. Therefore, under conditions where folic acid blocks 5-meth~l[~H]tetrahydrofolic acid accumulation, we detected only one population of folate binding sites on the membrane.
Recycling of the Folate Receptor-Even though at 37 "C [3H] folic acid never appeared in the cytoplasmic fraction nor was it polyglutamated (data not shown), the membrane fraction contained twice as much ligand as could bind to cells at 4 "C.
In addition, 50% of [3H]folic acid taken up at 37 "C was released by acid treatment, indicating that only half of the [3H]folic acid was exposed at the cell surface. When cells incubated with [3H]folic acid at 4 "C were washed and shifted to 37 "C, there was a reciprocal disappearance of surfacebound folic acid and an increase in the number of unoccupied receptors at the cell surface. These results suggest that the membrane fraction contains two populations of receptors: 50% are at the cell surface exposed to the extracellular space and the other 50% are in an internal membrane compartment. Moreover, these two receptor pools can exchange with each other at 37 "C.
Folic acid, then, is a high affinity ligand for the folate receptor that does not enter the cytoplasm. Instead, it remains bound to the receptor as the receptor cyclically moves in and out of the cell. Therefore, this ligand serves as a convenient way to tag the receptor to monitor receptor recycling. The finding that a maximum of 50% of the 13H]folic acid bound at 4 "C was internalized at 37 "C indicates that ordinarily both occupied and unoccupied receptors can shuttle in and out of the cell at 37 "C.
We consistently found that at 37 "C [3H]folic acid was taken up more slowly than 5-methyl[3H]tetrahydr~folic acid (Fig.  1). In this experiment there was a 4-fold difference in the initial rate of uptake. Therefore, the normal rate of receptor recycling may be much faster than was detected by following t3H]folic acid movement. Attempts to follow the internalization of 5-methyl[3H]tetrahydrofolic acid prebound at 4 "C were unsuccessful because when the cells were warmed to 37 "C, >90% of the ligand was released into the surrounding medium.
Release of Internalized 5-Methyltetrahydrofolic Acid-All of the drugs that are known to dissipate proton gradients within intracellular membrane-bound compartments (22,23) interfered with the uptake of 5-methyltetrahydrofolic acid. Ionophores such as monensin and nigericin had the most profound effect. Monensin most likely had its effect by inhibiting the release of 5-methyl[3H]tetrahydr~folic acid from internalized receptors because it did not measurably affect receptor recycling. Thus at equilibrium monensin had no effect on the amount of membrane-associated 5-methyl[3H]tetrahydr~folic acid but caused a 6-fold inhibition of accumulation in the cytoplasm. These results suggest that the release of physiological folate takes place in an acidic intracellular compartment.
The inhibitory effect of monensin on 5-meth~l[~H]tetrahydrofolic acid accumulation is strong evidence that the receptor enters a membrane-bound compartment, e.g. an endocytic vesicle, during recycling. However, we cannot rule out the possibility that ionophores and weak bases are affecting some other cell process that is required for delivery of 5methyltetrahydrofolic acid to the cytoplasm.
Mechanism of Receptor-mediated Znternalization-Collectively the data we have gathered suggest a novel mechanism for folate internalization by the folate receptor. Extracellular 5-methyltetrahydrofolic acid binds to a population of externally oriented membrane receptors. The receptor-linked complex moves into the cell, most likely by becoming entrapped in a vesicular compartment that cannot be readily distinguished from the plasma membrane on Percoll gradients. The interior of the compartment becomes acidic, causing the ligand to dissociate from the receptor. The acidic environment of this compartment may facilitate the movement of 5-methyltetrahydrofolic acid into the cytoplasm. The unoccupied receptor then moves back to the cell surface to participate in another round of internalization.
The relationship between receptor-mediated internalization and carrier-mediated internalization (5) remains to be clarified. Most likely these are two separate internalization processes that operate in different tissues to serve the special metabolic demands of individual cells. There is an intriguing possibility, however, that the two processes are coupled in some fashion. For example, the transmembrane movement of folate that is internalized by the folate receptor might be mediated by a carrier mechanism. The receptor, which is maximally expressed in folate-deficient cells, seems adapted to concentrate folates at the surface membrane, which may be required for the carrier to transport in a cellular environment that is low in folate.
Physiological Significance-We think that the folate-depleted tissue culture cell model used in our studies (7) and those of Kolhouse and co-workers (6,25) as well as McHugh and Cheng (26) has revealed a physiologically important folate transport pathway that cannot be detected in cells grown in standard folic acid-containing media ( 10-6-10-5 M), because receptor activity is regulated by intracellular folate concentration (7). We found that, in the presence of only 0.04 p~ 5methyl[3H]tetrahydr~folic acid, within 4 h the cellular folate concentration increases from 0.5-1 pM to 5-6 pM, which is a 100-fold higher concentration of [3H]folate in the cell than is in the media. During this time, the cells accumulate 10-15% of the total 5-methyl[3H]tetrahydr~folic acid in the media, resulting in 70-80% of the cellular folate being radiolabeled (50-60 x lo3 cpm/106 cells in a typical experiment). This high proportion of labeled folate facilitates the chemical characterization of the intracellular folate (Fig. 3), which is important for detecting any metabolic modifications that have occurred within the cell. Most importantly, this uptake process occurs at near physiological concent,rations of folate.
Whether the folate receptor functions the same way in uiuo remains to be established. The recent evidence that the soluble folate binder found in plasma may be derived from the membrane folate receptor (6) together with the finding that human placenta (27) and rat kidney (28) are rich sources of the membrane receptor suggests that the receptor is expressed on the surface of tissue cells.

a Membrane Receptor
That Recycles