Search

The interaction of an 18-base cis-element in the 5'-untranslated region of human folate receptor (FR)-alpha mRNA with a cytosolic trans-factor protein is critical for the translation of FR (Sun, X.-L., and Antony, A. C. (1996) J. Biol. Chem. 271, 25539-25547). This trans-factor was isolated to apparent homogeneity as a 43- and 38-kDa doublet from human placenta using poly(U)-Sepharose, followed by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electro-elution as major purification steps. Amino acid microsequencing of two cyanogen bromide-generated peptide fragments of the 43-kDa trans-factor revealed complete identity with 43-kDa heterogeneous nuclear ribonucleoprotein E1 (hnRNP E1). Purified specific rabbit anti-hnRNP E1 peptide antibodies (generated against a synthetic oligopeptide that was not represented in microsequenced peptides of the trans-factor) also recognized the purified trans-factor on Western blots. Conversely, the 18-base FR RNA cis-element also bound hnRNP E1 protein on Northwestern blots. Moreover, a 19-base RNA cis-element in the 3'-untranslated region of 15-lipoxygenase mRNA that is known to bind hnRNP E1 also interacted with placental 43-kDa trans-factor. In addition, several murine tissues containing a hnRNP E1-related protein (also known as alphaCP-1) readily interacted with the 18-base FR RNA cis-element. Finally, anti-hnRNP E1 antibodies specifically inhibited translation of FR in vitro in a dose-dependent manner, and the antibody effect could be reversed in a dose-dependent manner by either purified trans-factor or hnRNP E1. Collectively, the data favor identity of the FR mRNA-binding trans-factor and hnRNP E1, confirm its critical role in the translation of FR, and highlight yet another role of multifunctional hnRNP E1 in eukaryotic mRNA regulation.

Folate receptors (FR) are inversely regulated by the extracellular folate concentration at the translational level in cervical carcinoma cells. Accordingly, the potential for interaction of cis-elements in FR-alpha mRNA and trans-factors in these cells was determined. Gel-shift assays identified two signals that were specifically derived from the interaction of cytosolic proteins with the 5'-untranslated region of FR-alpha mRNA. RNase T1 mapping revealed that the RNA sequences interacting with these proteins were located between nucleotides -133 to -116 (18-bases) and -158 to -116 (43-bases), upstream of the translation start site. However, selective RNase H cleavage indicated that the 18-base RNA sequence was the cis-element. The RNA-protein interaction was competed by poly(C), but not by poly(U), homopolymers. UV cross-linking and Northwestern blot analysis confirmed that the trans-factors were 46-kDa proteins. An 18-base antisense oligodeoxynucleotide complementary to the cis-element specifically quenched the RNA-protein interaction and also completely inhibited translation of FR-alpha mRNA without changing its stability. Thus, the interaction of the 18-base cis-element and the 46-kDa trans-factors likely have an important role in translational regulation of FR. In addition, because the 46-kDa proteins were widely distributed in cells expressing little to no FR-alpha, these species probably have additional functions that are unrelated to translation of FR.

Glycosyl-phosphatidylinositol-anchored hydrophobic placental folate receptors (PFRs), which have an important functional role in maternal-to-fetal transplacental folate transport, can be converted to soluble hydrophilic forms by a placental metalloprotease. Using a Triton X-114 temperature-induced phase separation assay to monitor enzyme-mediated conversion of radiolabeled hydrophobic PFR into hydrophilic PFR, a metalloenzyme was isolated to apparent homogeneity from Triton X-114-solubilized human placenta using concanavalin A-Sepharose and reverse-phase high performance liquid chromatography (HPLC) as major purification steps. The purified hydrophobic enzyme eluted as a single protein peak on reverse-phase HPLC and SDS-polyacrylamide gel electrophoresis revealed a single 63,000 M(r) species, which was reduced to 58,000 M(r) following deglycosylation, findings comparable with amino acid analysis (M(r) approximately 59,000). The metalloenzyme was activated by Mg2+, Zn2+, Mn2+, and Ca2+, optimally at physiologic pH; it also exhibited EDTA-sensitive endoproteolytic cleavage of [3H]leucine-labeled full-length nascent PFR polypeptide generated in vitro in the absence of microsomes. Rabbit polyclonal anti-metalloprotease antiserum specifically immunoprecipitated 125I-metalloprotease and recognized cross-reacting moieties on plasma membranes of normal human hematopoietic progenitor cells and human cervical carcinoma cells, both of which also express FR.

Although normal human chorionic villi-associated hydrophobic placental folate receptors (PFR) are converted to hydrophilic forms by an endogenous, EDTA-sensitive, Mg(2+)-dependent protease under serum-free conditions (Verma, R. S., and Antony, A. C. (1991) J. Biol. Chem. 266, 12522-12535), it is not known whether hydrophobic PFR are also susceptible to conversion by endogenous phospholipases. We isolated and characterized hydrophobic PFR, and tested the hypothesis that purified, in situ, and de novo-synthesized native PFR were covalently linked to glycosyl-phosphatidylinositol (GPI) anchors. 125I-hydrophobic PFR, but not 125I-hydrophilic PFR, (i) separated into the Triton X-114 micellar phase at 30 degrees C, (ii) efficiently incorporated into phosphatidylcholine-cholesterol liposomes, and (iii) were covalently labeled by the hydrophobic probe 3-(trifluoromethyl)-3-(meta[125I]iodophenyl)diazirine, [125I]TID. (iv) [125I]TID-labeled- and [phenyl-3H]Triton X-100-bound hydrophobic PFR, as well as native PFR in situ, were released as hydrophilic forms by recombinant (r) GPI-specific phospholipase(PL) C (GPI-PLC), and GPI-PLD (but not by PLC), in the absence and presence of a concentration of EDTA known to inhibit endogenous Mg(2+)-dependent protease. (v) Nitrous acid deamination of [125I]TID-labeled hydrophobic PFR as well as (r)GPI-PLC cleavage of [phenyl-3H]Triton-X-100- and [125I] TID-labeled hydrophobic PFR, released hydrophobic radiolabeled moieties which comigrated on thin layer chromatography distinct from free radiolabel. Finally, (vi) biosynthetic studies on chorionic villi cultured in vitro revealed incorporation of radiolabeled precursors into the GPI anchor of hydrophobic PFR. We conclude that native hydrophobic PFR are linked to GPI anchors and are therefore potential substrates for three distinct endogenous enzymes (GPI-PLC, GPI-PLD, and specific Mg(2+)-dependent metalloprotease) in maternal serum and placenta in vivo.

Full term placental chorionic villi cultured for 7 days in serum-free medium released hydrophilic folate-binding proteins (S-FBP(PCM) into the conditioned medium; in contrast, hydrophobic FBPs were the only form recovered from chorionic villi. Kinetic studies revealed that (i) S-FBP(PCM) was maximally released by the 3rd day, and this was associated with a proportionate decrease in hydrophobic FBPs; (ii) although cycloheximide inhibited de novo synthesis of [35S]methionine-labeled hydrophobic FBPs and S-FBP(PCM) by greater than 90%, unlabeled net S-FBP(PCM) release was only inhibited by 50%; (iii) EDTA markedly inhibited release of S-FBP(PCM) which was accompanied by a proportionate increase in hydrophobic FBPs; (iv) EDTA effects were completely reversed by 5-fold molar excesses of Mg2+ which led to a 50-fold greater release of S-FBP(PCM) compared to EDTA alone; (v) whereas Mg2+ alone stimulated S-FBP(PCM) release 4-fold greater than basal conditions, addition of cycloheximide to Mg2+ suppressed (by 4-fold) the expected increase observed with Mg2+ alone. Biochemical analyses of isolated S-FBP(PCM) revealed similarities to hydrophobic FBPs with respect to the ligand-binding domain and epitopes but differed in detergent-binding characteristics; furthermore, amino acid and carbohydrate analysis revealed a lower Mr = 25,500 with 12% carbohydrate. Based on kinetic analysis of S-FBP(PCM) release from chorionic villi-associated hydrophobic FBPs as well as structural analysis of S-FBP(PCM), these data continue to support the hypothesis that (a) a significant amount of maternal and probably fetal serum hydrophilic FBPs originate from placental hydrophobic FBPs, and (b) the endogenous hydrophobic FBP-directed Mg(2+)-dependent placental protease plays a major role in their release.

Methotrexate accumulation, subcellular distribution, metabolism, and cytotoxicity were studied in human epidermoid carcinoma (KB) cells that were exposed to a low extracellular concentration of methotrexate (25 nM) following culture in widely differing concentrations of folic acid. KB cells cultured in standard medium with a high folic acid concentration (2.3 microM) had high levels of cellular folate (21.4 pmol/10(6) cells). Five passages through low folate (2.7 nM) medium reduced the level of cellular folate to near physiologic levels (0.4-1.0 pmol/10(6) cells). In contrast to KB cells cultured in standard medium, in KB cells cultured in low folate medium, 1) methotrexate inhibited growth; 2) methotrexate uptake was markedly increased; 3) methotrexate polyglutamation was almost complete; 4) methotrexate binding to dihydrofolate reductase was markedly enhanced; and 5) significant methotrexate binding to a previously undescribed membrane-associated protein occurred. The amount of methotrexate bound to the membrane-associated protein from KB cells cultured in low folate medium equaled the quantities bound by dihydrofolate reductase. Further characterization of this membrane-associated protein indicated that it was soluble in solutions containing Triton X-100, was capable of binding folic acid as well as methotrexate, had an apparent Mr of 160,000 by gel filtration in the presence of Triton X-100, and was precipitated by antiserum to human placental folate receptor. This membrane-associated protein may play an important role in the uptake and metabolism of methotrexate under physiologic conditions.

Human KB cells produce two immunologically cross-reactive folate-binding proteins: a particulate cell-associated protein which is solubilized by Triton X-100, and a soluble protein which is released into their growth medium. This compartmentation of these two folate-binding proteins provides a convenient system for studies of their biochemical relationship. The two folate-binding proteins behave similarly to the purified particulate and soluble folate-binding proteins of human milk in analysis by radioactive folate binding, Sephacryl S-200 gel filtration profiles, polyacrylamide gel electrophoresis in either Triton X-100 or sodium dodecyl sulfate, and in Triton X-100 binding based on sucrose density gradient ultracentrifugation in H2O and D2O. The two folate-binding proteins were endogenously labeled by pulsing methionine-starved KB cells with [35S]methionine, and each protein was purified to apparent homogeneity by affinity chromatography at different times during the chase with nonradioactive methionine. The time course of the changes in specific activity (moles of [35S]methionine per mole of folate-binding protein) revealed a more rapid initial rate of synthesis and an earlier maximum in specific activity for the cell-associated folate-binding protein than for the soluble folate-binding protein released into the growth medium. Differences in the levels and specific activities of the two folate-binding proteins of cells exposed to cycloheximide compared with simultaneous controls after pulsing with [35S]methionine suggest that, whereas the cell-associated folate-binding protein is probably produced by de novo protein synthesis, the soluble folate-binding protein seems to be produced from a cellular pool of an already synthesized protein. These results combined with the immunologic cross-reactivity of the two folate-binding proteins strongly suggest a precursor-product relationship between them.

The cobalamin-dependent enzyme, methionine synthetase, has been purified approximately 1000-fold to apparent homogeneity from human placenta with a 19% recovery. The final two steps of the purification utilized two different affinity columns. The first was a N5-methyltetrahydrofolate-cystamine-agarose column, and the second was a S-adenosylhomocysteine-agarose column. The enzyme was eluted from the first affinity column by buffer containing reducing agent which released the folate and the enzyme while elution from the second affinity column was accomplished with buffer containing 0.5 M sodium chloride. Criteria for purity were the observations that single peaks of enzyme activity, protein, and cobalamin with an apparent molecular weight of 160,000 were obtained by gel filtration and that holomethionine synthetase contained 1 mol of cobalamin/mol of protein. Furthermore, analysis by high performance liquid chromatography using a molecular weight sizing column demonstrated a single peak of protein with a corresponding cobalamin peak. This single peak of protein was progressively converted to a second protein peak that was enzymatically inactive, and this conversion was associated with a directly proportional loss of enzyme activity and cobalamin from the first peak. Methionine synthetase appeared to have a molecular weight of 160,000 on unreduced sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis and subunits of Mr 90,000, 45,000, and 35,000 on reduced sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis.

Hydrophobic folate-binding proteins (FBPs), which are only 5-10 kDa larger than 40-kDa hydrophilic FBPs, bind significant quantities of Triton X-100 micelles and elute as apparent 160-kDa species on Sephacryl S-200 gel filtration in Triton X-100. Detergent-solubilized placental membranes release a major (greater than 95%) 40-kDa hydrophilic FBP species as well as a minor apparent 160-kDa folate binding species when similarly analyzed. We tested the hypothesis that this recovery of predominantly hydrophilic FBPs was mediated by a putative hydrophobic FBP-specific placental protease. When placenta was solubilized in the presence of increasing concentrations of EDTA, there was a progressive increase in apparent 160-kDa folate binding moieties concomitant with a decrease in 40-kDa FBPs. At 20 mM EDTA, a single apparent 160-kDa folate binding species was recovered and the 40-kDa FBPs could not be detected by radioligand binding or specific radioimmunoassay. The apparent 160-kDa species specifically bound radiolabeled folates and were specifically immunoprecipitated by rabbit anti-40-kDa FBP antiserum. On 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by transfer to nitrocellulose and probing with anti-40-kDa FBP antiserum, the apparent 160-kDa FBPs electrophoresed as 45-kDa species. Detergent binding studies indicated that apparent 160-kDa FBPs were hydrophobic, thus accounting for the molecular weight discrepancy in gel filtration in Triton X-100 versus sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The EDTA-mediated inhibition of conversion of hydrophobic FBPs to hydrophilic FBPs by protease was reversed in a dose-dependent manner by Mg2+. If this protease is physiologically relevant, it could play an important regulatory role in folate transport by influencing the net number of hydrophobic FBPs on the cell surface.

While folate binding proteins have been described in serum and a variety of tissues, the function of these proteins is unknown. A particulate folate binding protein from human placenta has been isolated and characterized following solubilization with Triton X-100. The protein was purified 61,000-fold using affinity chromatography on pteroylglutamic acid-Sepharose as the major purification step. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis the purified protein gave a single band with a Mr = 38,500. Stoichiometry of binding indicated that 1 mol of folate was bound per mol of protein. The protein was a glycoprotein that contained 12% carbohydrate. Antiserum was raised in a rabbit, and on immunodiffusion, gave a single precipitin line with the purified placental folate binding protein. Immunoprecipitation studies using this antiserum indicated that the purified placental folate binding protein shared antigenic determinants with both the large Mr and small Mr folate binding proteins from human milk. Immunofluorescent studies with this antiserum and human erythrocytes revealed the presence of an immunologically similar protein on the plasma membrane of these cells suggesting that this protein may function as a folate receptor.

The characteristics of the uptake by human epidermoid carcinoma (KB) cells of 5-methyltetrahydrofolate at extracellular concentrations in the physiologic range and the possible role of a membrane-associated folate binder in folate uptake by KB cells have been investigated. Uptake of 5-methyltetrahydrofolate was specific, saturable, and time-, temperature-, and concentration-dependent. Trypsin treatment released 50% of the 5-methyltetrahydrofolate accumulated by KB cells at 4 degrees C, but only 12% at 37 degrees C, indicating that most of the accumulated ligand was intracellular at 37 degrees C, thus demonstrating transport. Accumulated 5-methyltetrahydrofolate was bound to a membrane-associated protein which required detergent for its solubilization, and a significant amount of which was oriented to the cell exterior as demonstrated by its release by trypsin treatment of intact KB cells. The membrane-associated folate binder was immunoprecipitated by antiserum to purified human placental folate receptor, and this antiserum inhibited 5-methyltetrahydrofolate uptake by intact KB cells in a concentration-dependent manner. These data support the hypothesis that the membrane-associated folate-binding protein of human cells participates in the transport of folates under physiologic conditions.