Structural analysis of trisialylated biantennary glycans isolated from mouse serum transferrin. Characterization of the sequence Neu5Gc(alpha 2-3)Gal(beta 1-3)[Neu5Gc(alpha 2-6)]GlcNAc(beta 1-2)Man.

Five variants of mouse serum transferrin (mTf, designated mTf-I to mTf-V) with respect to carbohydrate composition have been isolated by DEAE-cellulose chromatography in the following relative percentages: mTf-I: 0.55; mTf-II: 0.79; mTf-III: 71.80; mTf-VI: 21. 90 and mTf-V: 4.96. The primary structures of the major glycans from mTf-III and mTf-IV were determined by methylation analysis and 1H-nuclear magnetic resonance (NMR) spectroscopy. All glycans possessed a common trimannosyl-N,N'-diacetylchitobiose core. From the glycovariant mTf-III two isomers of a conventional biantennary N-acetyllactosamine type were isolated, in which two N-glycolylneuraminic acid (Neu5Gc) residues are linked to galactose either by a (alpha 2-6) or (alpha 2-3) linkage. A subpopulation of this glycovariant contains a fucose residue (alpha 1-6)-linked to GlcNAc-1. The structure of the major glycan found in variant mTf-IV contained an additional Neu5Gc and possessed the following new type of linkage: Neu5Gc(alpha 2-3)Gal(beta 1-3)[Neu5Gc(alpha 2-6)]GlcNAc(beta 1-2 )Man(alpha 1-3). In addition to this glycan, a minor compound contained the same antennae linked to Man(alpha 1-6). In fraction mTf-V, which was found to be very heterogeneous by (1)H NMR analysis, carbohydrate composition and methylation analysis suggested the presence of tri'-antennary glycans sialylated by Neu5Gc alpha-2,6- and alpha-2, 3-linked to the terminal galactose residues. In summary, mTf glycans differed from those of other analyzed mammalian transferrins by the presence of Neu5Gc and by a Neu5Gc(alpha 2-6)GlcNAc linkage in trisialylated biantennary structures, reflecting in mouse liver, a high activity of CMP-Neu5Ac hydroxylase and (alpha 2-6)GlcNAc sialyltransferase.

Translocation of different forms of transferrin from blood to bile in the rat.

Five different forms of transferrin (rat apo [iron-free], rat diferric, diferric rat asialo, human diferric, and diferric human asialotransferrin type 3) were used to monitor the passage of this protein and its metal to the bile. Cumulative biliary excretion of the dose over 3 hours was determined. In addition, an excretion profile was constructed from the concentration of tracer in bile samples collected over 10-minute intervals. The profile obtained with apotransferrin was very similar to that found in an earlier study with albumin, the implication being that the apo form is transferred passively (e.g., by diffusion). Behavior of rat diferric transferrin, however, was consistent with the assumption that this form is transferred both passively and actively (i.e., in vesicles). The three other transferrins were investigated with the intent of broadening the spectrum of ligand affinities for the plasmalemma of hepatocyte. The higher this attraction was, the larger fraction of the dose appeared in bile. When transferrin was targeted to lysosomes, the bile contained several intermediate proteolytic fragments. Double-labeled (125I, 59Fe) transferrin was used to measure recovery of iron (Fe) relative to the protein (P) in bile. With rat diferric transferrin, the Fe/P ratio was 0.72. Lower values were recorded with transferrins (human or asialo) that had higher affinities for the plasmalemma and therefore were expected to be transported to a larger extent in vesicles. Of the biliary 59Fe, 85% to 92% was protein bound. The proportion of the protein-bound fraction was essentially independent of the magnitude of Fe/P ratios.

Interaction of transferrin with rat alveolar macrophages.

Binding of rat transferrin to isolated alveolar macrophages was investigated in the 0.125 nM to 2 microM range. Computer analysis of the data revealed two classes of binding sites, a small number (< 1000 exposed/cell) having high affinity (dissociation constant (Kd), 3.4 nM) and a large number (approximately 4 x 10(6)/cell) having low affinity (Kd 48 microM). Measurements with a monoclonal antibody to the rat transferrin (rTf) receptor yielded values in the same range as the high-affinity sites derived from studies of ligand binding. Binding to the low-affinity sites at pH 5.8 was nearly one order of magnitude stronger than that at pH 7.3. Bovine lactoferrin (12 microM), cationized bovine serum albumin (14 microM), L-arginine (50 mM), and L-lysine (50 mM) did not compete against rTf binding to the low-affinity sites. Removal of an average of 2.6 x 10(8) sialyl residues from each cell did not affect binding. Heparan sulphate proteoglycan purified from alveolar macrophages bound strongly to immobilized rTf, thus raising the possibility that the low-affinity interaction of transferrin with these cells may be mediated, at least in part, by this glycosaminoglycan.

Effect of transferrin on the degradation of glycoproteins bearing a hybrid or high-mannose glycan by alveolar macrophages.

A subfraction of hen ovalbumin and a special form of rat transferrin, possessing a hybrid glycan, were studied with respect to their binding to, and degradation by macrophages isolated from the rat lung. Both ligands were found to be degraded partly through the mannose receptor pathway and partly by another mechanism, presumably adsorptive pinocytosis. Catabolism of both proteins was markedly increased by adding standard (i.e., normally glycosylated) transferrin from various species to the medium. This increase was not diminished, or even augmented, when Ca2+ was depleted in the medium, thus implying involvement of the pinocytic pathway rather than the mannose receptor. Binding to the cell surface of both ligands was altered in the presence of transferrin. A hypothesis is advanced to suggest that when transferrin is bound to a component of macrophage plasmalemma, thought to be the glycosaminoglycan of heparan sulfate, its conformation may change in such a way that it attracts other proteins. Protein molecules temporarily captured by adsorbed transferrin would then be taken up by a "piggyback" process and degraded.

Observations with a residualizing label in use to map the catabolic sites of plasma proteins.

Rat albumin, immunoglobulin G, transferrin, and aglycotransferrin were prepared for the comparison of their sites of degradation in rats. Iodotyramine-cellobiose was used as the residualizing label and a tyrosine-iodinated portion of the corresponding protein was used as the marker of extracellular undegraded protein. Each protein yielded a distinct distribution (or map) of catabolic activity throughout the body when expressed as percent dose accumulated per gram of tissue. The maps for albumin and transferrin were broadly comparable, whereas those for immunoglobulin G and aglycotransferrin were markedly different. As a whole entity, the liver appeared to top the list of organs/tissues contributing to the degradation of albumin and transferrin. Additional experiments aimed at facilitating the interpretation of results with residualizing labels were carried out with denatured albumin, asialofetuin, and human asialotransferrin type 3. These showed that various types of cells retained the label for markedly different periods of time. We feel, therefore, that the technique is more suited for making comparative measurements than for obtaining degradation rates as absolute values in a given anatomical location.

Transport of lactoferrin from blood to bile in the rat.

The bile contains small quantities of lactoferrin, the origin of which is uncertain. For this reason, we studied the liver's capability of transferring lactoferrin from the plasma to the bile by injecting a dose (10 to 20 micrograms/100 gm) of labeled bovine lactoferrin intravenously and following its appearance in bile over 3 hr. Whether diferric or iron free, lactoferrin peaked in the bile 35 min after administration (i.e., the same time as bovine lactoperoxidase and diferric rat transferrin). However, only a small portion of the lactoferrin dose (approximately 1%) was recovered with the bile in 3 hr. On the basis of autoradiographic evidence, the excreted lactoferrin appeared intact. The biliary excretion profile of albumin, a protein thought to reach the canaliculus by paracellular diffusion, was notably devoid of a peak. This, together with competition observed between lactoferrin and lactoperoxidase on one hand and 2Fe-transferrin and lactoferrin on the other for transfer to bile, suggests that lactoferrin is routed through the hepatocyte in vesicles. The process is initiated by binding to a plasma membrane component to which lactoperoxidase and 2Fe-transferrin can also bind. Most 59Fe bound to lactoferrin accompanied the protein carrier to the bile. We conclude that under normal circumstances (i.e., when concentration of lactoferrin in the plasma is very low), lactoferrin transferred from plasma by the liver is probably not the major source of this protein in bile.

Interaction of transferrin and its iron-binding fragments with heparin.

The interaction of heparin with transferrin (Tf; bovine and rat) and the isolated iron-binding lobes of bovine Tf were investigated. Affinity chromatography of rat Tf on heparin-agarose showed that interaction depended on both the iron content of Tf and the pH of the medium. Both the iron-free and iron-saturated forms of Tf were strongly bound by the column at pH 5.6, but only the iron-free form revealed significant affinity at pH 7.4. Desialylation of Tf moderately promoted interaction, treatment with cyclohexanedione moderately reduced interaction, and succinylation abolished it altogether. In the presence of heparin, iron release from the N-terminal lobe of native bovine Tf was accelerated and from the C-terminal lobe it was slightly reduced. The heparin effect remained qualitatively the same on each lobe after their separation by tryptic digestion and DEAE-cellulose chromatography. The affinity of native bovine Tf for heparin was very close to that of its isolated N-terminal lobe, thus suggesting that it is this portion of the molecule that binds to the glycosaminoglycan. It is concluded that the consequences for iron-binding strength of the two transferrin lobes are diagonally opposite when Tf is bound to heparin as opposed to its natural cell-surface receptor.

Isolation of the four forms of transferrin with respect to iron by high-performance liquid chromatography: comparison of three mammalian species.

A method, based on anion exchange combined with high-performance liquid chromatography, is presented for the separation of the four different forms of transferrin with respect to iron, i.e. diferric, monoferric with Fe in the C-terminal lobe, monoferric with Fe in the N-terminal lobe, and apo (iron-free). Depending on the size of column applied, these forms can be obtained on a preparative scale amounting to milligrams. The procedure was also found to free transferrin from loosely bound iron and protein aggregates. Comparative studies with human, murine and rat transferrins showed that optimal resolution of these proteins depended on establishing a distinct elution programme for each species. The order in which the four forms referred to above were displaced from the exchanger differed from species to species.

Lactoferrin interferes with uptake of iron from transferrin and asialotransferrin by the rat liver.

Intravenous injection of bovine or human lactoferrin (6.25 x 10(-2) mumol/100 g body wt) in rats resulted in marked reduction of hepatic iron uptake from transferrin and asialotransferrin. The effect was dose dependent, saturable at approximately 5 mg/100 g body wt, and independent of lactoferrin's iron content. At this dose level, iron uptake from transferrin was reduced by 28% and from asialotransferrin by 43% in experiments lasting 90 min. Bovine lactoperoxidase, another basic protein, was similarly effective. The clearance of asialofetuin and pinocytosis of polyvinylpyrrolidone remained unaffected. Perfusion of isolated rat livers at 4 degrees C showed a strong reduction in asialotransferrin binding in the presence of lactoferrin. Chromatography of hepatic heparan sulfate proteoglycan on immobilized lactoferrin, lactoperoxidase, asialotransferrin, and transferrin showed that it possessed affinity for each of these proteins, more for the first two than the latter two. Heparan sulfate proteoglycan binding and efficacy in reducing hepatic iron uptake were also studied after selective modifications of positively charged amino acids in these proteins. The data obtained are compatible with the hypothesis that lactoferrin and other proteins with similarly high affinity for hepatic heparan sulfate exert their negative effect on iron uptake by preventing transferrin binding to the proteoglycan. The possibility is thus raised that the large number of low-affinity transferrin binding sites reported by earlier investigators for the liver may be heparan sulfate molecules.

The perfused liver is capable of producing all transferrin glycan variants found in the sera of intact rats.

The single oligosaccharide attachment in rat transferrin exhibits marked structural microheterogeneity. In this study we examined whether all microheterogeneous forms of rat transferrin found in plasma are derived from a single organ, such as the liver. To this end we analyzed the glycans of rat transferrin synthesized by the isolated perfused rat liver by a method established earlier for rat transferrin isolated from rat plasma. Our observations provide evidence that the liver can and does produce all variant rat transferrin glycans present in plasma. However, this discovery does not preclude the possibility that extrahepatic sources with an active rat transferrin gene may contribute to the circulation rat transferrin molecules, which bear glycan variants identical to those made by the liver. The glycan spectra of rat transferrin in plasma and in liver perfusate compared closely with each other in a quantitative sense. Nevertheless, rat transferrin in the perfusate was sialylated to a lesser extent and fucosylated to a greater extent than rat transferrin in plasma. These differences could not be eliminated by supplementation of the medium with insulin, dexamethasone, pyruvate and adenine or adenosine either alone or in combinations, nor could it be eliminated by use of a fluorocarbon O2 carrier. In contrast, epidermal growth factor normalized both parameters. The pH of the perfusing medium also influenced sialylation and fucosylation in such a way that higher pH brought these parameters closer to their values in plasma rat transferrin. Lower pH, on the other hand, reduced sialylation and left the fucosylation index unchanged.

In vivo behaviour of rat transferrin bearing a hybrid glycan and its interaction with macrophages.

Production of rat transferrin containing a single hybrid glycan was induced by treating rats with swainsonine, an inhibitor of alpha-mannosidase II. The principal component of this variant transferrin containing one sialic acid residue per mole of protein was separated from other forms of transferrin by anion-exchange chromatography, followed by lectin affinity chromatography. Transferrin bearing the hybrid glycan was degraded in vivo with a half-life of 14 h as compared with 40 h for transferrin containing a standard diantennary glycan. By using 125I-labelled tyramine-cellobiose, a label whose discharge from lysosomes is strongly retarded, organs rich in reticuloendothelial elements (liver, bone marrow, lungs, and spleen) were identified as the major sites of catabolism of the transferrin variant. The liver took up more 59Fe from the variant (26% of the dose in 90 min) than from control rat transferrin (12%). The excess iron uptake was reduced by the intravenous injection of either human transferrin or ovalbumin, and it was abolished by administering both. Macrophages from bone marrow and lungs degraded the transferrin variant in vitro. The degradation was significantly enhanced when transferrin receptors were blocked by human transferrin, and it was significantly reduced by ovalbumin and methyl glucopyranoside.

Hepatic heparan sulphate proteoglycan and the recycling of transferrin.

Heparan sulphate proteoglycan, labelled with [35S]sulphate, was prepared from rat livers for studies of its interaction with purified rat transferrin. Affinity chromatography of the preparation on columns of immobilized differic transferrin and apotransferrin showed that the proteoglycan possessed affinity for both types of matrices at pH 7.3 and that this affinity significantly increased at pH 5.6. The glycosaminoglycan chains liberated from the proteoglycan by heparan sulphate lyase also bound to apotransferrin, albeit less strongly, whereas the deglycosylated core protein exhibited virtually no interaction with this matrix. In the presence of the proteoglycan at pH 5.6, the release of iron from the N-lobe of transferrin was accelerated. These observations suggest that heparan sulphate proteoglycan from the liver can mimick some of the known functions of bona fide transferrin receptors and, hence, interaction with the proteoglycan may provide an alternative nondegradative pathway for transferrin through hepatic cells.

Transferrin glycosylation in hypoxia.

The aim of this study was to examine the effect of reduced O2 tension on the glycosylation of transferrin. Rats were placed in a hypobaric chamber (380 mmHg) that corresponded to an altitude of 5486 m above sea level for 21 days. The animals responded with marked increases in hematocrit (from 44 to 76%) and cardiac weight, and with reductions in the concentration of plasma transferrin averaging 15%. Analyses of their plasma transferrin by serial anion-exchange and lectin affinity chromatography revealed no changes in the extent of glycan branching. However, there was a moderate rise in the proportion of fucosylated transferrin molecules (fucosylation index) and a slight decrease in the transferrin fraction bearing a tetrasialylated biantennary glycan. The fucosylation index correlated positively with plasma transferrin concentrations in the test animals, but not in the controls. In contradistinction to the situation with transferrin, hypoxic rats exhibited a reduced fucosylation index of immunoglobulin G.

Relationship between pinocytic rate and uptake of transferrin by suspended rat hepatocytes.

The aim of this study was to compare quantitatively the capacity to transcytose (i.e. to uptake and release) transferrin (Tf) with the pinocytic activity of suspended adult rat hepatocytes. An oligodisperse preparation of 131I-polyvinylpyrrolidone (PVP; Mr 36,000) was used to measure the inward and outward aspects of the pinocytic process in separate experiments. Cell association of rat 125I-Tf was measured at Tf concentrations approaching physiological, where 59Fe uptake obeyed first-order kinetics. Release studies with both PVP and Tf were carried out under conditions which minimized the probability of de novo endocytosis of a molecule already released. Sets of experimental points representing cell-associated radioactivities were converted into continuous algebraic functions by fitting with two-term (release studies) or three-term (uptake studies) exponential equations. Transport of PVP and Tf through the cells was computed from these equations by deconvolution. This analysis showed that, under the present experimental conditions, the fractional transcytosis rates of Tf and PVP by hepatocytes were in the ratio of 1:0.77. These values imply that, in the physiological range of Tf concentrations, about 75% of the Fe taken up by hepatocytes may be due to a pinocytic mechanism (fluid-phase or mixed). Inclusion of chloroquine (1 mM) in the suspending medium, both in uptake and release experiments, resulted in more PVP and Tf passing through the cells, while Fe uptake was reduced. It is suggested that the base probably exerted its enhancing effect on transcytosis by shunting the subcellular transport of PVP and Tf to the outward leg through a shorter circuit.

Carbohydrate microheterogeneity of rat serotransferrin. Determination of glycan primary structures and characterization of a new type of trisialylated diantennary glycan.

A previously established procedure [Regoeczi, E., Chindemi, P.A., Rudolph, J. R., Spik, G. & Montreuil, J. (1987) Biochem. Cell Biol. 65, 948-954] was used to isolate from three DEAE-cellulose chromatographic fractions of diferric rat serotransferrin (rTf) subpopulations having discernible affinities for concanavalin A (ConA). These entities are designated rTf-1 (not retarded by ConA column), rTf-2 (retarded) and rTf-3 (bound). Each rTf type was found to be endowed with carbohydrate sufficient to account for a single diantennary glycan/protein molecule. Glycan structures were determined on the glycopeptides by employing GLC/MS and 400-MHz 1H-NMR spectroscopy. All glycans possessed a common, trimannosyl-N,N'-diacetylchitobiose core with or without one L-fucose alpha-1,6-linked to the Asn-linked GlcNAc. However, there were differences in the antennae. Thus, in rTf-3, both antennae were of the disialylated diantennary N-acetyllactosamine type which is frequently encountered in other plasma glycoproteins. However, the alpha-1,3-Man-linked antenna in rTf-1 as well as rTf-2 had the sequence: Neu5Ac(alpha 2-3)Gal(beta 1-3)[Neu5Ac(alpha 2-6)]GlcNAc(beta 1-2)Man. In addition, the alpha-1,6-Man-linked antenna deviated in rTf-2 from the standard structure by having the sequence: Neu5Ac(alpha 2-3)Gal(beta 1-3)GlcNAc(beta 1-2)Man. The possible relevance of the above structures to the ConA binding of rTf is discussed. A further preparation, obtained from the most anionic DEAE-cellulose fraction (peak V) or rTf contained several tetrasialylated diantennary glycans whose precise structures remain to be established in future studies.

Reduced hepatic iron uptake from rat aglycotransferrin.

Rat aglycotransferrin (rAgTf) was produced from the disialosyl diantennary fraction of rat transferrin (rTf) by treatment with peptide: N-glycosidase F. Following removal of the enzyme by gel filtration and isolation of the deglycosylated protein by lectin chromatography, rAgTf was compared to rTf both in vitro and in vivo. No significant differences were found between the two proteins with respect to affinity for iron and kinetics of Fe release from the N-lobe and C-lobe. The fluorescence emission spectrum of apo-rTf was red-shifted by approximately 3 nm relative to diferric rTf; however, no spectral difference was detected between rTf and rAgTf when the analogous forms (apo or diferric) were compared. Plasma clearance of radioactive iron administered to rats as either rTf or rAgTf was comparable. Reticulocytes took up iron from rAgTf slightly faster than from rTf. In contrast, Fe acquisition by the liver from rAgTf was significantly reduced relative to rTf. This finding contrasts sharply with earlier observations with asialotransferrin (rAsTf) and provides a basis for discounting charge loss as the mechanism of enhanced hepatic Fe uptake from rAsTf. It is suggested that the glycan complement of rTf, while unimportant for interaction of the protein with specific receptors, probably plays a role in the interaction with low-affinity hepatic binding sites.

Absorption of plasma proteins from peritoneal cavity of normal rats.

The present study was undertaken to examine whether the uptake of plasma proteins from the peritoneal cavity is quantitative so that tracers could be introduced that way for measuring their turnover. To this end, the metabolic behavior of seven homologous plasma proteins, labeled with 125I, was compared in rats after intravenous or intraperitoneal administration. The animals were maintained under physiological conditions. Total body radiation measurements showed that the degradation rates of albumin, immunoglobulins A and G, alpha 1-macroglobulin, and transferrin were the same regardless of the route of injection. This implies that these proteins are quantitatively absorbed from the peritoneum without undergoing modifications. The half-life of intraperitoneally injected alpha 1-acid glycoprotein was consistently shorter by an average 9%, thus suggesting that this protein becomes slightly altered if introduced that way. Only one-half of intraperitoneally injected fibrinogen survived normally, whereas the other underwent rapid degradation. The surviving molecules had the same half-life as fibrinogen injected intravenously. The fraction of surviving fibrinogen could be augmented by mixing the dose with serum. Within a wide range of concentrations and quantities injected, the degradation rate of transferrin remained the same. Analysis by deconvolution of the plasma curves of albumin and alpha 1-macroglobulin absorbed from the peritoneum showed that the transport process was independent of protein size and, at least up to 35 mg, of the amount injected. According to the same technique, intraperitoneally administered diferric transferrin retained its iron during passage into the circulation.

Rat aglycotransferrin and human monoglycotransferrin: production and metabolic properties.

Rat transferrin (rTf), containing one complex glycan, and human transferrin (hTf), containing two complex glycans, were treated with peptide:N-glycosidase F (PNGase) under nondenaturing conditions. Apo rTf with a nonfucosylated standard biantennary glycan, but not its diferric counterpart, yielded satisfactory amounts (approximately 55% in 7 h) of aglyco Tf (AgTf). The presence of a chitobiose core fucose reduced yields to approximately 30%, whereas an additional NeuAc on the GlcNAc in the Man(alpha 1-3) branch had no adverse effect. Triton X-100 impaired deglycosylation. The main product (approximately 65%) obtained from apo hTf was monoglyco Tf (MgTf). Analysis of the cyanogen bromide fragments of MgTf revealed that PNGase did not discriminate between the two glycosylation sites of hTf. A negligible portion (2-4%) of AgTf, that was also obtained during the reaction, probably resulted from PNGase action on denatured hTf molecules. Modified Tfs were separated by affinity chromatography, radiolabeled, mixed with another preparation of interest, and injected in rats. Total-body radiation measurements showed that the half-life of rat AgTf was 19-20% shorter than that of rTf but 9% longer than that of asialo Tf. This suggests that close to 76% of the change in the degradative rate observed after desialylating rTf is referable to charge loss rather than the exposure of Gal residues. Human MgTf was catabolized by rats 8-9% faster than the parent protein, while human AgTf behaved in vivo like a denatured protein. It is concluded that sialyl carboxyls are a codeterminant of the normal lifetime of transferrins.

Interaction of rat asialotransferrin with adult rat hepatocytes: its relevance for iron uptake and protein degradation.

Comparative studies with rat transferrin (rTf) and asialotransferrin (asialo-rTf) were performed on suspended adult rat hepatocytes with the aim of elucidating the mechanism of enhanced hepatic Fe acquisition from asialo-rTf observed previously in vivo. At low ligand concentrations (0.05-20 micrograms/ml), the cells bound more asialo-rTf than rTf. However, the excess binding was abolished by incubation either in the presence of 1.55 mg/ml of diferric rTf or of 1 mg/ml of asialomucin. Following either treatment, asialo-rTf and rTf were bound to comparable extents. These findings indicate that both transferrin receptors and the hepatic galactose recognition system (lectin) are essential for preferential binding of asialo-rTf by hepatocytes. The possibility is considered that the lectin facilitates capture of asialo-rTf by the same binding sites that are normally available for rTf rather than that it functions as an alternative pathway. In agreement with this view, asialo-rTf could not be channeled into the lectin-mediated degradative pathway by blocking Tf receptors with human Tf. Enhanced Fe uptake from asialo-rTf was fully prevented by asialomucin and partially prevented by human Tf. The incomplete efficacy of human Tf in this regard supports reports in the literature about Fe uptake by the liver in a manner that is independent of Tf receptors. Rhesus asialo-Tf was deployed to show that no recognition mechanism exists for heterologous asialo-Tf in rat hepatocytes. The importance of using undenatured labeled proteins for studies with cells is demonstrated.