Search

Rat placenta contains virtually no unsaturated (i.e., apo-form) folate binding protein. However, by lowering the pH of a solubilized membrane preparation of this tissue to 3.5, the endogenous bound folate was dissociated from the protein and adsorbed to charcoal. The apo-form of the folate binding protein thus obtained was purified by affinity chromatography using pteroylglutamic acid covalently coupled to Sepharose 4B. A single protein band with an apparent M(r) of 36,000 was observed by SDS-polyacrylamide gel electrophoresis of the eluate from the affinity matrix. Western blot of this preparation using a rabbit antiserum raised with the affinity eluate also identified a single 36 kDa protein band. However, peptide sequencing of the N-terminal region of the proteins in the affinity eluate established that it contained two homologous proteins. Computer alignment of the first 22 N-terminal amino acids of each rat placental protein with human, bovine milk and mouse folate binding proteins showed 50-64% identical homology and 27% homology when the eight proteins were aligned together. The affinity of both rat proteins is highest for pteroylglutamic acid (Ka = 1.6.10(9) l/mol) lower for N5-methyltetrahydrofolate and substantially lower for N5-formyltetrahydrofolate. In the dose-response range studied there was no apparent affinity for methotrexate. The folate binding proteins could be released from a preparation of placental membranes using phospholipase C indicating that these proteins belong to the class of proteins anchored to the plasma membrane by a glycosyl phosphatidylinositol adduct.

Antiserum raised in rabbits against the FBP obtained from CML cells, and the purified binder labeled with 125I, have been used for an RIA which can measure an immunologically similar protein in human serum. The concentration of the binding protein in normal serums ranged from 1.2 to 9.3 ng/ml, with a mean +/- S.E.M. of 3.8 +/- 0.4 ng/ml. Elevated values of the binder protein were measured in the serums from patients with folate deficiency, vitamin B12 deficiency, liver disease, uremia, myeloproliferative disease, chronic lymphocytic leukemia, and various types of cancer and in the serum from pregnant women. The concentration of the binder protein and the capacity of the serum to specifically bind isotopically labeled PGA correlated poorly, indicating that the binding protein concentration and degree of saturation by endogenous serum folate vary independently in many instances.

Previous studies have demonstrated that some chronic myelogenous leukemia cells contain a macromolecular binding factor for folic acid. This binder, which previously was believed to be a single factor, has now been resolved into two distinct binding proteins. Separation of each binder was obtained by DEAE chromatography of the partially purified lysate of chronic myelogenous leukemia cells. One binder has a molecular weight of 30;000-35,000, and the second binder has a molecular weight of 40,000-45,000. Both proteins bind the mono-, di-, and triglutamates of folic acid, N10-methyl-folate, dihydro-folate, and N5-methyltetrahydrofolate. Neither binder has determinants for N5-formyltetrahydrofolate or methotrexate. The preferred substrates for both binders appear to be the fully oxidized and partially reduced folates rather than the fully reduced folates. The lower-molecular-weight folate binding protein shows reversible binding with partially and fully reduced folates but irreversible binding with oxidized folates. This property suggests that this binder may have some function in the transport and storage of folate. The higher-molecular-weight folate binding protein, however, has only slight reversibility of binding with the partially and fully reduced folates, and it is therefore more difficult to postulate a physiologic function for this binding factor.

The blast transformation of human lymphocytes exposed to phytohaemagglutinin was associated with an increase in the intracellular concentration of a folate-binding protein. The protein was measured by a radioimmunoassay and by immunofluorescent microscopy using an antiserum raised against purified folate-binding protein prepared from chronic myelogenous leukaemia cells. Because it has been previously shown that protein-bound folate cannot function as a cofactor or substrate, the synthesis of this binder during some phase of the replicative cycle may serve to modulate folate-dependent RNA and DNA synthesis by controlling the concentration of free folate.

A new matrix for affinity chromatography using pteroylglutamic acid coupled to an epoxy-activated matrix via hexanediamine resulted in negligible ligand leakage and permitted the purification of soluble and membrane-associated folate-binding proteins from human leukemia cells contained in a human spleen. Two species of membrane-associated folate-binding proteins were purified from the solubilized membrane fraction of the tissue using 2 M guanidine-HCl to elute the proteins from the affinity matrix. The higher molecular weight binding protein had an Mr of approximately 310,000 and the smaller species had an Mr of approximately 28,000 by gel filtration. By SDS-polyacrylamide gel electrophoresis the smaller species of membrane-associated protein had a molecular weight of 35,500, but the molecular weight of the larger membrane-associated species could not be determined by this method because of the high concentration of residual Triton X-100 in the sample which interfered with the silver staining of the gel. Two folate-binding proteins, which by SDS-polyacrylamide gel electrophoresis had molecular weights of 34,500 and 32,000, were purified from the 44,000 X g supernatant fraction of the tissue homogenate by acid elution from the affinity matrix. Despite the different cell components from which the soluble and membrane-associated folate-binding proteins were purified, the amino acid compositions were similar, especially with respect to the apolar amino acids. All these forms of folate-binding proteins had higher affinity for oxidized than for reduced folates, and very low affinity for 5-formyltetrahydrofolate and methotrexate. Although these proteins cross-react with one antiserum raised previously to a folate-binding protein from other human leukemia cells, they do not cross-react with the folate-binding proteins purified from two other sources of human leukemia cells, from human placenta, or from the human KB cell line.

Chronic granulocytic leukaemia (CGL) cells which contained a high concentration of unsaturated folate binding protein were incubated in suspension culture for a period of 5 h. Cell samples were periodically assayed for binder and these demonstrated active synthesis which was inhibited by puromycin, cyclo heximide, N-ethylmaleimide, and by incubation at 4 degrees C, but not by actinomycin D. Folate binding activity could also be demonstrated in the culture medium and this increased with the duration of incubation. This release of binder was inhibited by culturing the cells at 4 degrees C and by the addition of N-ethylmaleimide, but not by actinomycin D, puromycin, or cycloheximide. When the pre- and post-culture cell lysates were saturated with tritiated folic acid ([3H]PteGlu) and subjected to chromatography on DEAE-agrarose, approximately half of the bound folate eluted with 0.001 M phosphate buffer at pH 6.0 and the other half eluted with 0.2 M buffer at pH 7.2. The culture medium and plasma from this patient with CGL was well as serum from two normal subjects saturated with [3H]PteGlu and similarly chromatographed contained primarily the acidic binder and much less of the binder eluting with the low molarity buffer. Since a folate binding protein immunochemically similar to the binder in CGL cells has been identified in the serum of non-leukaemic subjects, these experiments suggest that the source of circulating folate binding protein may be the immature granulocyte.

Folate-binding protein(s) from chronic myelogenous leukemia cells have been purified using acid dialysis, ammonium sulfate fractionation and affinity chromatography. The purified preparation which migrates as a single band on disc electrophoresis could be separated by DEAE agarose chromatography into two folate-binding proteins (binders I and II) which bind molar equivalents of folic acid. One binder (I) eluted from DEAE at 1 mM sodium phosphate, pH 6.0, and the other (II) at 100 mM sodium phosphate, pH 7.4. Analysis of the purified mixture, which contained more than 90% binder II, by sedimentation equilibrium centrifugation indicated a homogeneous protein with a calculated molecular weight of 44000. Antiserum raised against the purified mixture gave a single precipitin line by immunodiffusion against a preparation of partially purified cell lysate. Hydrolysis of the more acidic binder (II) with neuraminidase converted it to a weakly acidic protein similar to binder I, suggesting that these binders are glycoproteins which differ in sialic acid content. With isoelectric focusing, the binding of folic acid could be demonstrated at pH 6.7, 7.3, 7.8 and 8.2 for binder I, and at pH 5.1, 5.8, and 6.5 for binder II. Binders I and II had equally high affinity for folic acid and dihydrofolate, lower affinity for N5-methyl-tetrahydrofolate, and no apparent affinity for N5-formyltetrahydrofolate or methotrexate.

A radiochemical procedure is described for specific determination of pteroylglutamate in serum and urine. This method depends on denaturation of methyltetrahydrofolate with peroxide and measurement of the residual folate by a ligand-binding radioassay. The binding determinant for the radioassay is a folate-binding protein, partially purified from chronic myelogenous lekemia cells, that has low affinity for the reduced folates and thus will preferentially measure residual pteroylglutamate rather than any nondenatured residual methyltetrahydrofolate. We used this assay to measure the clearance from plasma and the urinary excretion of pteroylglutamate and a small fraction of serum folate that is stable to this oxidation procedure. The plasma clearance after intravenous injection is characterized by an initial rapid distribution phase followed by a second, slower metabolic phase; after about 2 h all of the administered pteroylglutamate has been cleared from the blood. The peak concentration of total folate in serum 1--2 min after administration of pteroylglutamate exceeded the sum of the endogenous stable and baseline serum folate, indicating that a reduced labile folate was released from the liver and perhaps from other tissues. This reduced folate had a slower metabolic clearance rate and was excreted to some extent in urine. Only 2.3 and 7.9% of the pteroylglutamate administered to two normal subjects was excreted as stable folate.

Unsaturated folate-binding proteins (i.e., apo forms) have been identified with the plasma membranes of rat liver by the binding of [3H]pteroylglutamic acid. Normal rat liver contains very little of the folate-binding apoproteins, but the folate-binding capacity increases substantially when the rats are made folate-deficient. This increase appears to be due to unsaturation of the folate-binding holoproteins rather than to synthesis of additional protein, because the binding capacity of the plasma membranes from normal rat liver following dissociation of the bound folate is equivalent to the binding capacity of the preparation from folate-deficient liver. Two molecular forms of folate-binding protein were identified by gel filtration of the solubilized plasma membrane fraction, a high-molecular-weight form (Mr less than 100,000), representing 25% of the binding capacity, and a smaller protein (Mr approximately equal to 55,000), representing 75% of the binding capacity. Whereas the larger species can be solubilized only with a detergent, the smaller form appears to be hydrophilic and dissociates spontaneously from the membrane preparation. The binding of [3H]pteroylglutamic acid by the membrane preparation was specific, saturable, and pH- and temperature-dependent. Scatchard analysis of the binding could be fitted to a curvo-linear plot, indicating at least two orders of binding sites which probably correspond to the two molecular forms identified by gel filtration. Competitive inhibition by folate analogues demonstrated that the apoproteins have higher affinity for oxidized folate than for N5-methyltetrahydrofolate and virtually no affinity for N5-formyltetrahydrofolate or methotrexate.