Membrane Folate Binding Protein Research Articles

Overexpression of folate binding protein α is one of the mechanism explaining the adaptation of HT29 cells to high concentration of methotrexate

The human colon adenocarcinoma cell line HT29 can be adapted to 10 −7- 10 −4 M concentrations of methotrexate (MTX). Cells adapted to 10 −4 M MTX have an enterocyte-like phenotype with DHFR gene amplification. Presently, we hypothetized that an increased expression of folate binding protein (FBP) may participate to the MTX resistance of 10 −4 MTX HT29 cells. The cDNA FBPα/β-actin ratio of amplified transcripts was 4.8- and 1.5- fold higher in 10 −4 and in 10 −7 M MTX HT29 respectively, than in standard type HT29 cells. An increase of transcript level was observed when decreasing folic acid concentration. PI-PLC cleaved 7.7 times more membrane FBP in 10 −4 M than in 10 −7 M MTX and wild type HT29 cells. In contrast to 10 −7 M MTX cells, growth of 10 −4 M MTX cells was dependent on folic acid concentration and abolished at a concentration lower than 0.9 μM. In conclusion, the adaptive mechanism of HT29 cells resistant to 10 −4 M MTX is the result of the synergistic overexpression of both DHFR and FBPα. Overexpression of FBPα may be related to the enterocyte-like phenotype of the cells.

Isolation of Complementary Deoxyribonucleic Acids Encoding Putative Secreted and Membrane-Bound Folate Binding Proteins from Endometrium of Swine1

Two distinct forms of endometrial folate binding protein (FBP) cDNAs were isolated using reverse transcription-polymerase chain reaction and 3' and 5' rapid amplification of cDNA ends (RACE) procedures. On the basis of the absence or presence of an intact glycophosphatidylinositol linkage site in the C terminus of the predicted amino acid sequences, the two forms appear to encode secreted and membrane-bound forms of FBP. The cDNAs for the putative secreted and membrane forms encoded 252- and 249-amino acid proteins, respectively, that were 73% identical with each other and were 66-82% identical with other known FBPs. However, the nucleotide sequences within the 5' untranslated region and from codons 224 and 223 of the secreted and membrane forms, respectively, to the 3' ends of each RNA, were divergent. The divergence in the 3' ends of the two cDNAs was exploited to determine changes in concentrations of each mRNA in the endometrium during the estrous cycle and early pregnancy. Northern blots of endometrial total RNA probed with a putative secreted FBP specific probe indicated that mRNA concentrations do not change during early pregnancy. In contrast, blots probed with a putative membrane FBP specific probe indicated that mRNA concentrations increase dramatically from Day 15 to Day 24 of pregnancy. Finally, N-terminal amino acid sequencing of FBP purified from Day 15 pregnant uterine flushings matched the secreted form of FBP mRNA. These data are consistent with a role for putative secreted and membrane-bound forms of FBPs in the transport of folate to the developing swine conceptus during early pregnancy.

Intracellular regulation of intestinal folate uptake: studies with cultured IEC-6 epithelial cells

Although the mechanism of folate uptake in the small intestine has been well characterized, very little is known about the intracellular regulation of the uptake process. Using mature confluent monolayers of the intestinal epithelial cell line IEC-6 as an in vitro intestinal epithelial cell model, we have found the uptake of folic acid to be similar to that of the native small intestine in that it is 1 ) temperature, energy, and pH dependent, 2) Na+ independent, 3) inhibited by structural analogs and anion transport inhibitors, and 4) saturable as a function of substrate concentration [apparent Michaelis constant (Km) = 0.45 +/- 0.06 microM; maximal velocity (Vmax) = 3.08 +/- 0.14 pmol x mg protein(-1) x 5 min(-1)]. Furthermore, IEC-6 cells were found by Northern blot analysis to lack the expression of the membrane folate-binding protein. Pretreatment of IEC-6 monolayers with specific protein tyrosine kinase (PTK) inhibitors genistein and tyrphostin A25 caused a significant inhibition in folic acid uptake. On the other hand, their negative controls, genistin and tyrphostin A1, respectively, had no effect. The inhibitory effect of genistein was mediated through inhibition in the Vmax of the folate uptake process with no change in the apparent Km. Pretreatment of IEC-6 monolayers with compounds that increase intracellular adenosine 3',5'-cyclic monophosphate (cAMP) level (e.g., dibutyryl cAMP) also resulted in a significant (though modest) inhibition in folic acid uptake; however, specific inhibitors of protein kinase A did not affect the uptake process. Specific modulators of protein kinase C and Ca2+/calmodulin-mediated pathways did not significantly affect folic acid uptake. These results demonstrate the suitability of IEC-6 monolayers as an intestinal epithelial model to study folate transport and demonstrate for the first time that uptake of folic acid is regulated by a PTK- and a cAMP-mediated pathway.

Role of membrane folate-binding protein in the cytotoxicity of 5,10-dideazatetrahydrofolic acid in human ovarian carcinoma cell lines in vitro

Lometrexol (5,10-dideazatetrahydrofolic acid; DDATHF), is a specific inhibitor of glycinamideribonucleosyl (GAR) transformylase with anti-tumour activity in murine and human carcinomas. The cytotoxicity activity of DDATHF was evaluated in vitro in NIH/3T3 cells transfected with human alpha-folate-binding protein (FBP) complementary DNA to examine the role of the receptor. In FBP-transfected NIH/3T3 (FBP-tNIH/3T3) cells, which internalised about three times more 5-methyltetrahydrofolic acid than the mock-transfected cells, the cytotoxtic potential of DDATHF showed a clear increase. Subsequently, we analysed four ovarian carcinoma cell lines (OVCAR3, IGROV1, SKOV3, and SW626) expressing different amounts of FBP. Cells were conditioned to grow in medium depleted of folic acid then tested by MOv18 and folic acid binding. Only SKOV3 and SW626 cells grown in folic acid-depleted medium showed increased FBP expression, about 3- and 8-fold respectively. The cytotoxic potential of DDATHF was evaluated by a standard clonogenic assay. In a medium containing 2.27 microM folic acid the DDATHF IC50 values were 50 nm on OVCAR3, 500 nM on SW626 and 1000 nM on IGROV1. In folic acid-free medium IC50 values were 2 nM on OVCAR3 and Sw626 and 40 nM on IGROV1. Only on SKOV3 cells was DDATHF cytotoxicity the same regardless of the amount of folic acid in the medium (IC50 8 nM). Thus, DDATHF did not inhibit the growth of IGROV1 cells depleted of folic acid after stripping FBP with phosphatidylinositol-phospholipase C, even at a dose toxic for cells constitutively expressing FBP. Although FBP expression is certainly one of the parameters affecting drug toxicity, taken alone it is not a sufficiently reliable predictor of cancer cell sensitivity to DDATHF.

Determinants of the disparate antitumor activities of (6 R)-5,10-dideaza-5,6,7,8-tetrahydrofolate and methotrexate toward human lymphoblastic leukemia cells, characterized by severely impaired antifolate membrane transport

We previously reported (Matherly et al., J Biol Chem 267: 23253–23260, 1992) that impaired methotrexate transport in a drug-resistant CCRF-CEM variant (CEM/MTX) involved the synthesis of a structurally altered isoform of the “classical” carrier for methotrexate and related derivatives. Although CEM/MTX cells were highly resistant (162- to 300-fold) to assorted antifolate substrates for the classical transporter, including methotrexate, aminopterin, 10-ethyl-10-deazaaminopterin, ICI D1694, and 1843U89, they were only 3.6-fold resistant to (6 R)-5,10-dideaza-5,6,7,8-tetrahydrofolate (DDATHF)> These divergent antifolate sensitivities were not associated with appreciable differences in the levels of dihydrofolate reductase, thymidylate synthase, and 5′-phosphoribosylglycinamide (GAR) transformylase, or the expression of a high affinity membrane folate binding protein receptor in either line. The initial rate of [ 14C]DDATHF influx was increased 2.9-fold over that for [ 3H]metrotrexate in parental cells (at 2 μM). Whereas [ 14C]DDATHF initial uptake was, likewise, increased over [ 3H]methotrexate in CEM/MTX cells (5.3-fold), influx of both compounds was impaired substantially (95–97%). For the parent, influx of [ 14C]DDATHF was inhibited by substrates for the classical transporter including unlabeled DDATHF, methotrexate, (6 R)-5-formyl tetrahydrofolate, 10-ethyl-10-deazaaminoproterin, ICI D1694, 1843U89, and folic acid. The synthesis of a modified transporter in CEM/MTX cells was accompanied by significant changes in the binding of all these transport substrates. Inspite of its impaired transport, [ 14C]DDATHF (at 2 μM), unlike methotrexate, continued to accumulate in CEM/MTX cells, eventually reaching 62% of the parental drug levels after 4 hr. At this time, 53% (parent) and 71% (CEM/MTX) of the intracellular radioactivity from [ 14C]DDATHF was identified as polyglutamates. DDATHF polyglutamates in CEM/MTX cells after 4 hr reached 90% of the levels measured in parental cells. While significant levels of methotrexate polyglutamates were detected in the parental line, methotrexate polyglutamylation was negligible in intact CEM/MTX cells. The specific activity of folypolyglutamate synthetase was measured in cell-free extracts from parental and CEM/MTX cells using aminopterin, methotrexate, and DDATHF as substrates; in each case, CEM/MTX cells showed 2-fold higher enzyme activity than parental cells. These data show that even for tumor cells with severely impaired antifolate transport, the extensive conversion of DDATHF to polyglutamyl forms required for GAR transformylase inhibition preserves high levels of antitumor activity.

Membrane folate-binding proteins are responsible for folate-protein conjugate endocytosis into cultured cells.

Folate-protein conjugates have been shown to bind to and enter HeLa and KB cells by receptor-mediated endocytosis [Leamon and Low (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 5572-5576]. Although these cells contain a membrane folate-binding protein (FBP) involved in the uptake of free folate, no studies have been conducted to evaluate whether the folate-protein conjugates enter cells via the same protein. To address this issue, HeLa cell monolayers were treated with folate-labelled 125I-RNAase under various conditions characteristic of FBP-mediated folate uptake. Folate-labelled 125I-RNAase was found to bind to cells with high affinity (Kd = 24 nM), and like the free vitamin, its binding could be competitively blocked by excess free folate. Furthermore, binding could be reversed by either washing the cells with acid/saline, pH 3.0, or by treating the cells with phosphatidyl-inositol-specific phospholipase C, an enzyme known to release FBP from cell surfaces. Because cells pretreated with anti-FBP serum were unable to bind folate conjugates, and since the same antiserum identified a single 65 kDa band reminiscent of FBPs found in many other tissues, we conclude that a classical FBP is responsible for the uptake of folate-protein conjugates by receptor-bearing cells.

Folate transport and binding by cultured human proximal tubule cells.

Because mechanisms for the renal regulation of folate excretion are poorly understood, a cell culture system representative of the human proximal tubule (HPT) was used for studies of renal folate transport. After confluent cultures of HPT cells were incubated with 3H-labeled folic acid (PteGlu), binding to the apical membrane was determined by an acid removal procedure, and transport was subsequently measured in solubilized cells. Although PteGlu binding was almost all specific (suppressed by excess unlabeled PteGlu), HPT cells transported PteGlu by specific and nonspecific processes. Specific PteGlu binding and transport were both saturable processes, reaching maxima of 0.5 and 0.1 pmol/mg protein, respectively, with half-maximal constants of 12 and 50 nM, respectively. The PteGlu analogues methotrexate and 5-methyltetrahydrofolic acid (5-CH3-H4PteGlu) inhibited both the binding and transport of PteGlu, with 5-CH3-H4PteGlu being more potent (lower half-maximal inhibitory concentration). In contrast, 5-formyltetrahydrofolic acid significantly reduced PteGlu transport at concentrations (100-250 nM) that had no effect on binding. These data suggest that the HPT cells will serve as a good model for studies of renal folate reabsorption. Initial characterization of the transport of folate by HPT cells suggests two distinct processes, binding to a high-affinity membrane folate-binding protein followed by a structurally specific transfer into the cell.

Synthesis and biological evaluation of a fluorescent analog of folic acid

A fluorescein derivative of the lysine analogue of folic acid, N alpha-pteroyl-N epilson-(4'-fluoresceinthiocarbamoyl)-L-lysine (PLF), was synthesized as a probe for dihydrofolate reductase (DHFR) and a membrane folate binding protein (m-FBP). Excitation of PLF at 282 nm and at 497 nm gave a fluorescence emission maximum at 518 nm. Binding of PLF to human DHFR or human placental m-FBP results in approximately a 20-fold enhancement in the magnitude of the fluorescence emission, suggesting that the ligand interacts with a hydrophobic region on these proteins. Additional evidence suggests that an energy transfer may occur between the pteridine and the fluorescein moieties. PLF binds to the active site of human DHFR since methotrexate (MTX) competes stoichiometrically and the denatured enzyme in the presence of PLF did not exhibit fluorescent enhancement. The dissociation constant for the fluorescein derivative with respect to human DHFR is 115 nM as compared to 111 nM for folic acid. The Ki value for the competitive inhibition of human DHFR by the fluorescent analogue of folic acid is 2.0 microM compared to 0.48 microM for folic acid. PLF was reduced to N alpha-(7,8-dihydropteroyl)-N epilson-(4'-fluoresceinthiocarbamoyl)-L-lysine (H2PLF) and assayed by the enzymatic conversion to the tetrahydro derivative. The Km value for human DHFR for the dihydrofolate analogue is 2.0 microM. The KD value for H2PLF to human DHFR is 47 nM as compared to 44 nM for dihydrofolate. The KD values for both H2PLF and PLF indicate that the fluorescein moiety does not significantly affect folate binding in enzyme binary complexes.(ABSTRACT TRUNCATED AT 250 WORDS)

The glycophospholipid-linked folate receptor internalizes folate without entering the clathrin-coated pit endocytic pathway.

The folate receptor, also known as the membrane folate-binding protein, is maximally expressed on the surface of folate-depleted tissue culture cells and mediates the high affinity accumulation of 5-methyltetrahydrofolic acid in the cytoplasm of these cells. Recent evidence suggests that this receptor recycles during folate internalization and that it is anchored in the membrane by a glycosyl-phosphatidylinositol linkage. Using quantitative immunocytochemistry, we now show that (a) this receptor is highly clustered on the cell surface; (b) these clusters are preferentially associated with uncoated membrane invaginations rather than clathrin-coated pits; and (c) the receptor is not present in endosomes or lysosomes. This receptor appears to physically move in and out of the cell using a novel uncoated pit pathway that does not merge with the clathrin-coated pit endocytic machinery.

A human membrane-associated folate binding protein is anchored by a glycosyl-phosphatidylinositol tail

Membrane-associated and soluble forms of folate binding protein (FBP) have been identified in mammalian tissues and biological fluids. Despite their solubility differences, these two forms are functionally similar, immunologically cross-reacting, and have the same apparent molecular weights. In this study we demonstrate, for the first time, that the membrane FBP of cultured human KB cells contains a glycosyl-phosphatidylinositol (GPI) tail which is responsible for its hydrophobic properties and distinguishes it from the soluble FBP released into the medium. Treatment of the purified membrane FBP with phospholipase C specific for phosphatidylinositol (PI-PLC) removed the GPI tail and converted it to the soluble form without a change in apparent Mr by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. In addition, virtually all of the folate binding sites on the plasma membrane of the intact cells were released as soluble, functional FBP following treatment with PI-PLC. The GPI tail contained 1-O-alkyl-2-O-acylglycerol as a mixture of fatty alcohols in ether linkage at C1 of the glycerol backbone and almost exclusively docosanoic acid (22:0) as the fatty acid on C2. The inositol also contained a mixture of fatty acids (16:0, 18:0, 18:1, 20:4, 22:0) located on a site other than the C2 position since the FBP was susceptible to PI-PLC cleavage. After nitrous acid deamination, the aqueous portion of the FBP contained covalently bound fatty acids, predominantly palmitate (16:0) and stearate (18:0), indicating the presence of additional acyl groups attached to the peptide in the form of amide, ester, or thioester linkage.

The Complete Amino Acid Sequence of a Human Folate Binding Protein from KB Cells Determined from the cDNA

Two species of folate binding protein (FBP), an integral membrane-associated form and a soluble secreted form, have been previously purified from cultured human KB cells. The complete nucleotide sequence of the complementary DNA (cDNA) clone for the coding region of the mature membrane-associated FBP has now been determined, and the deduced amino acid sequence has been computer-analyzed for a prediction of the secondary structure of the protein. The clone has 857 nucleotides of which 678 comprise the coding region for 226 amino acids. The deduced amino sequence contains the identical sequence of the published 18 NH2-terminal amino acids of the purified FBP from KB cells and the published partial amino acid sequence of the human milk FBP except for 1 residue. There was also over 90% homology with the published amino acid sequence of the bovine milk FBP. A total of 16 cysteine residues has been conserved in the three proteins indicating that this amino acid may provide a tertiary structure which is required for its ligand binding function. Northern blot analysis using the cDNA probe identified a single band of 1.28-kilobase pair mRNA in KB cells which was 4.7-fold more intense in folate-depleted cells than in normal cells. These results indicate that the membrane FBP and the soluble FBP in the medium are translation products of the same gene. Computer analysis of the deduced amino acid sequence indicates that there is only one stretch of amino acids of sufficient hydrophobicity and length to span the lipid bilayer of the plasma membrane, but it lacked a predictable helical structure. Those regions of the sequence which did have a predictable helical structure lacked sufficient hydrophobicity required for a membrane anchor. Thus, it is likely that the fatty acids previously reported to be present in the membrane-associated FBP from these cells rather than a peptide sequence provide an important membrane anchoring function.

Folate coenzyme and antifolate transport proteins in normal and neoplastic cells

The transport systems for folate coenzymes and antifolate compounds into various types of normal and neoplastic cells display considerable diversity in their pharmacokinetics and also in terms of the apparent molecular weights of the proteins involved. Further, several uptake routes may exist in a given cell type. A variety of neoplastic tissues have been reported to rely upon a single major transport system that has a relatively high affinity for the reduced form of folate compounds and for antifolates such as methotrexate. Using a photoaffinity analogue of methotrexate, we have identified the involvement of a 48 kDa membrane protein and a 38 kDa cytosolic or peripheral membrane protein in the transport of this compound into murine L1210 leukemia cells. Such an uptake is absent in mutant L1210 cells that are defective in methotrexate transport. We propose a model for the uptake of reduced folate coenzymes in L1210 cells in which the compound is initially transported across the cell membrane by the 48 kDa protein and delivered on the cytoplasmic surface to the 38 kDa protein; the 38 kDa protein then carries the folate compound to a specific enzyme of folate metabolism. Antibodies to the membrane folate binding protein from human placenta cross-react with the 48 kDa protein in L1210 cell membranes indicating an immunological relationship between these two proteins. Comparison of the amino acid sequences of peptides of the placental receptor obtained by digestion with S. aureus V8 protease indicate the presence of two homologous forms of the folate binding protein in placenta; one of these forms appears to have an identical sequence to the soluble and membrane associated folate binding proteins in human epidermoid carcinoma (KB) cells, which in turn share the primary structure of the soluble and membrane associated folate binders in human milk in the regions that have been sequenced. These results indicate that the folate coenzyme transport proteins in various tissues may be structurally related and in several instances may even be identical. In the latter cases the observed differences in apparent molecular weights may be due to differences in glycosylation and/or proteolysis. In support of this view is our observation that the deglycosylated and/or partially proteolyzed placental receptor retains the ability to bind folate.

Purified membrane and soluble folate binding proteins from cultured KB cells have similar amino acid compositions and molecular weights but differ in fatty acid acylation.

A membrane-associated folate binding protein (FBP) and a soluble FBP, which is released into the culture medium, have been purified from human KB cells using affinity chromatography. By NaDodSO4/PAGE, both proteins have an apparent Mr of approximately 42,000. However, in the presence of Triton X-100, the soluble FBP eluted from a Sephadex G-150 column with an apparent Mr of approximately 40,000 (similar to NaDodSO4/PAGE) but the membrane-associated FBP eluted with an apparent Mr of approximately 160,000, indicating that this species contains a hydrophobic domain that interacts with the detergent micelles. The amino acid compositions of both forms of FBP were similar, especially with respect to the apolar amino acids. In addition, the 18 amino acids at the amino termini of both proteins were identical. The membrane FBP, following delipidation with chloroform/methanol, contained 7.1 mol of fatty acid per mol of protein, of which 4.7 mol was amide-linked and 2.4 mol was ester-linked. The soluble FBP contained only 0.05 mol of fatty acid per mol of protein. These studies indicate that the membrane FBP of KB cells contains covalently bound fatty acids that may serve to anchor the protein in the cell membrane.