Structural reorganization of the transferrin C-lobe and transferrin receptor upon complex formation: the C-lobe binds to the receptor helical domain

Human transferrin, a bilobal protein, with each lobe bearing a single iron-binding site, functions to transport iron into cells. While the N-terminal lobe alone does not measurably bind cellular transferrin receptors or serve as an iron donor for cells, the C-lobe is capable of both functions. We used hydroxyl radical-mediated protein footprinting and mass spectrometry to reveal the conformational changes that occur upon complex formation for the human transferrin C-lobe (residues 334-679) bound to the ectodomain of human transferrin receptor 1 (residues 121-760). Oxidation rates for proteolytic peptides in the C-lobe, the receptor, and their complex have been measured by mass spectrometry; upon formation of the complex, a dramatic decrease in modification rates, indicating protection of specific side chain groups, can be seen in C-lobe sequences corresponding to residues 381-401, 415-433, and 457-470. Peptide sequences experiencing modification rate decreases in the transferrin receptor upon C-lobe binding include residues 232-240, 365-371, 496-508, 580 and 581, 614-623, 634-646, 647-681, and 733-760. In addition, several peptides in the receptor exhibit enhancements in the rate of modification consistent with allosteric effects of complex formation. Using tandem mass spectrometry, the sites of modification with altered reactivity in the complex include Met382, Met389, Trp460, Met464, and Phe427 in the C-lobe and Tyr503, Pro581, Tyr611, Leu619, Met635, Phe650, Trp740, Trp754, and Phe760 within the transferrin receptor. Using available genetic, biochemical, and structural data, we confirm that the conserved RGD sequence (residues 646-648) in the helical domain of the transferrin receptor, including residues from Leu619 to Phe650, is a primary binding site for the transferrin C-lobe.     

Characterization of transferrin binding and specificity of the placental transferrin receptor

This study systematically examined the characteristics of specific binding of adult diferric transferrin to its receptor using a Triton X-100 solubilized preparation from human placentas as the receptor source. The following information was obtained. The ionic strength for maximal binding is in the range of 0.1-0.3 M NaCl. The pH optimum for specific binding extends over the range, from pH 6.0-10.0. Specific binding of diferric transferrin is not affected by 2.5 approximately 50 mM CaCl2 or by 10 mM EDTA. Triton X-100 in the concentration range of 0.02-3.0% does not affect specific binding. Specific binding is saturated within 10 min at 25 or 37 degrees C in the presence of excess amounts of diferric transferrin. The binding is reversible and the dissociation of diferric transferrin from the transferrin receptor is complete within 40 min at 25 degrees C. Apotransferrin, both adult and fetal, showed less binding than the holotransferrin species by competitive binding assay in the presence of 10 mM EDTA independent of up to 20 mM CaCl2. A 1500-fold molar excess of adult and fetal apotransferrin is required to give 40% inhibition for 125I-labeled diferric transferrin binding. Since calcium ion is not a factor, and since apotransferrin has such high binding affinity for iron (Ka = 1 X 10(24], this experiment suggests that the EDTA was necessary to prevent conversion of apotransferrin to holotransferrin from available iron in the reaction system. The specificity of the transferrin receptor for transferrin was examined by competitive binding studies in which 125I-diferric transferrin binding was measured in the presence of a series of other proteins. The proteins tested in the competitive binding studies were classified into three groups; in the first group were human serum albumin and ovalbumin; in the second group were proteins containing iron ions, such as hemoglobin, hemoglobin-haptoglobin complex, heme-hemopexin complex, ferritin, and diferric lactoferrin; in the third group were the metal-binding serum proteins, ceruloplasmin and metallothionein. None of these proteins except ferritin showed inhibition of diferric transferrin binding to the receptor. The effect of ferritin was small since a 700- to 1500-fold molar excess of ferritin is required for 50% inhibition of binding of diferric transferrin to the receptor.     

Structural characteristics of the mouse transferrin receptor

Rat monoclonal antibodies against mouse transferrin receptor have been used to isolate and characterize the mouse receptor molecule. The molecule is a dimeric glycoprotein of Mr 200 000 resembling its human homolog of Mr 190 000. Receptor molecules prepared from different lymphoid cell populations show structural differences which can be explained by variations in the carbohydrate moiety of the molecule. Both the antibody-binding site and the transferrin-binding site are located on tryptic fragments of Mr 80 000 on the extracellular part of the molecule. After trypsin treatment, these fragments are partially retained at the cell surface, probably non-covalently bound to one intact receptor subunit, but they are released at higher trypsin concentrations. The soluble fragments retain their ability to bind transferrin and appear to exist as dimers. In this fragment, there are no disulfide bonds present. Disulfide bonds are located near the plasma membrane. Studies using a cleavable cross-linker indicated the presence of cross-linking sites at the intramembranous or the cytoplasmic part of the molecule.     

Iodination significantly influences the binding of human transferrin to the transferrin receptor

The human transferrin receptor (TfR) and its ligand, the serum iron carrier transferrin, serve as a model system for endocytic receptors. Although the complete structure of the receptor's ectodomain and a partial structure of the ligand have been published, conflicting results still exist about the magnitude of equilibrium binding constants, possibly due to different labeling techniques. In the present study, we determined the equilibrium binding constant of purified human TfR and transferrin. The results were compared to those obtained with either iodinated TfR or transferrin. Using an enzyme-linked assay for receptor-ligand interactions based on the published direct calibration ELISA technique, we determined an equilibrium constant of Kd=0.22 nM for the binding of unmodified human Tf to surface-immobilized human TfR. In a reciprocal experiment using soluble receptor and surface-bound transferrin, a similar constant of Kd=0.23 nM was measured. In contrast, covalent labeling of either TfR or transferrin with 125I reduced the affinity 3-5-fold to Kd=0.66 nM and Kd=1.01 nM, respectively. The decrease in affinity upon iodination of transferrin is contrasted by an only 1.9-fold decrease in the association rate constant, suggesting that the iodination affects rather the dissociation than the association kinetics. These results indicate that precautions should be taken when interpreting equilibrium and rate constants determined with covalently labeled components.     

Mutational analysis of the transferrin receptor reveals overlapping HFE and transferrin binding sites.

The transferrin receptor (TfR) binds two proteins critical for iron metabolism: transferrin (Tf) and HFE, the protein mutated in hereditary hemochromatosis. Previous results demonstrated that Tf and HFE compete for binding to TfR, suggesting that Tf and HFE bind to the same or an overlapping site on TfR. TfR is a homodimer that binds one Tf per polypeptide chain (2:2, TfR/Tf stoichiometry), whereas both 2:1 and 2:2 TfR/HFE stoichiometries have been observed. In order to more fully characterize the interaction between HFE and TfR, we determined the binding stoichiometry using equilibrium gel-filtration and analytical ultracentrifugation. Both techniques indicate that a 2:2 TfR/HFE complex can form at submicromolar concentrations in solution, consistent with the hypothesis that HFE competes for Tf binding to TfR by blocking the Tf binding site rather than by exerting an allosteric effect. To determine whether the Tf and HFE binding sites on TfR overlap, residues at the HFE binding site on TfR were identified from the 2.8 A resolution HFE-TfR co-crystal structure, then mutated and tested for their effects on HFE and Tf binding. The binding affinities of soluble TfR mutants for HFE and Tf were determined using a surface plasmon resonance assay. Substitutions of five TfR residues at the HFE binding site (L619A, R629A, Y643A, G647A and F650Q) resulted in significant reductions in Tf binding affinity. The findings that both HFE and Tf form 2:2 complexes with TfR and that mutations at the HFE binding site affect Tf binding support a model in which HFE and Tf compete for overlapping binding sites on TfR.     

Structural and functional studies on the stalk of the transferrin receptor

Transferrin (Tf) is an iron carrier protein that consists of two lobes, the N- and C-lobes, which can each bind a Fe³⁺ ion. Tf binds to its receptor (TfR), which mediates iron delivery to cells through an endocytotic pathway. Receptor binding facilitates iron release from the Tf C-lobe, but impedes iron release from the N-lobe. An atomic model of the Tf-TfR complex based on single particle electron microscopy (EM) indicated that receptor binding is indeed likely to hinder opening of the N-lobe, thus interfering with its iron release. The atomic model also suggested that the TfR stalks could form additional contacts with the Tf N-lobes, thus potentially further slowing down its iron release. Here, we show that the TfR stalks are unlikely to make strong interactions with the Tf N-lobes and that the stalks have no effect on iron release from the N-lobes of receptor-bound Tf.     

Ionic residues of human serum transferrin affect binding to the transferrin receptor and iron release

Efficient delivery of iron is critically dependent on the binding of diferric human serum transferrin (hTF) to its specific receptor (TFR) on the surface of actively dividing cells. Internalization of the complex into an endosome precedes iron removal. The return of hTF to the blood to continue the iron delivery cycle relies on the maintenance of the interaction between apohTF and the TFR after exposure to endosomal pH (≤6.0). Identification of the specific residues accounting for the pH-sensitive nanomolar affinity with which hTF binds to TFR throughout the cycle is important to fully understand the iron delivery process. Alanine substitution of 11 charged hTF residues identified by available structures and modeling studies allowed evaluation of the role of each in (1) binding of hTF to the TFR and (2) TFR-mediated iron release. Six hTF mutants (R50A, R352A, D356A, E357A, E367A, and K511A) competed poorly with biotinylated diferric hTF for binding to TFR. In particular, we show that Asp356 in the C-lobe of hTF is essential to the formation of a stable hTF-TFR complex: mutation of Asp356 in the monoferric C-lobe hTF background prevented the formation of the stoichiometric 2:2 (hTF:TFR monomer) complex. Moreover, mutation of three residues (Asp356, Glu367, and Lys511), whether in the diferric or monoferric C-lobe hTF, significantly affected iron release when in complex with the TFR. Thus, mutagenesis of charged hTF residues has allowed identification of a number of residues that are critical to formation of and release of iron from the hTF-TFR complex.     

Influence of transferrin glycans on receptor binding and iron-donation

Human bi-bi-antennary transferrin (Tf) was partially deglycosylated by subsequently incubating with one or more of the following exoglycosidases: neuraminidase, β-galactosidase or N-Acetyl-β-D-glucosaminidase. Aglyco-Tf obtained from serum of a patient suffering from the Carbohydrate Deficient Glycoprotein syndrome was isolated. Receptor binding and the Tf and iron uptake capacities of the fully glycosylated-, partially deglycosylated- and aglyco-Tf were compared using the human hepatoma cell line PLC/PRF/5. No difference in binding capacity between the iso-Tf fractions could be demonstrated, however, the Tf and iron uptake capacity of aglyco-Tf was clearly reduced compared with the other Tf fractions. This revised version was published online in November 2006 with corrections to the Cover Date.     

The hemochromatosis protein HFE competes with transferrin for binding to the transferrin receptor.

HFE is a class I major histocompatibility complex (MHC)-related protein that is mutated in patients with the iron overload disease hereditary hemochromatosis. HFE binds to transferrin receptor (TfR), the receptor used by cells to obtain iron in the form of diferric transferrin (Fe-Tf). Previous studies demonstrated that HFE and Fe-Tf can bind simultaneously to TfR to form a ternary complex, and that membrane-bound or soluble HFE binding to cell surface TfR results in a reduction in the affinity of TfR for Fe-Tf. We studied the inhibition by soluble HFE of the interaction between soluble TfR and Fe-Tf using radioactivity-based and biosensor-based assays. The results demonstrate that HFE inhibits the TfR:Fe-Tf interaction by binding at or near the Fe-Tf binding site on TfR, and that the Fe-Tf:TfR:HFE ternary complex consists of one Fe-Tf and one HFE bound to a TfR homodimer.     

Mechanism of formation of the complex between transferrin and bismuth, and interaction with transferrin receptor 1

The kinetics and thermodynamics of Bi(III) exchange between bismuth mononitrilotriacetate (BiL) and human serum transferrin as well as those of the interaction between bismuth-loaded transferrin and transferrin receptor 1 (TFR) were investigated at pH 7.4-8.9. Bismuth is rapidly exchanged between BiL and the C-site of human serum apotransferrin in interaction with bicarbonate to yield an intermediate complex with an effective equilibrium constant K(1) of 6 +/- 4, a direct second-order rate constant k(1) of (2.45 +/- 0.20) x 10(5) M(-1) s(-1), and a reverse second-order rate constant k(-1) of (1.5 +/- 0.5) x 10(6) M(-1) s(-1). The intermediate complex loses a single proton with a proton dissociation constant K(1a) of 2.4 +/- 1 nM to yield a first kinetic product. This product then undergoes a modification in its conformation followed by two proton losses with a first-order rate constant k(2) = 25 +/- 1.5 s(-1) to produce a second kinetic intermediate, which in turn undergoes a last modification in the conformation to yield the bismuth-saturated transferrin in its final state. This last process rate-controls Bi(III) uptake by the N-site of the protein and is independent of the experimental parameters with a constant reciprocal relaxation time tau(3)(-1) of (3 +/- 1) x 10(-2) s(-1). The mechanism of bismuth uptake differs from that of iron and probably does not involve the same transition in conformation from open to closed upon iron uptake. The interaction of bismuth-loaded transferrin with TFR occurs in a single very fast kinetic step with a dissociation constant K(d) of 4 +/- 0.4 microM, a second-order rate constant k(d) of (2.2 +/- 1.5) x 10(8) M(-1) s(-1), and a first-order rate constant k(-d) of 900 +/- 400 s(-1). This mechanism is different from that observed with the ferric holotransferrin and implies that the interaction between TFR and bismuth-loaded transferrin probably takes place on the helical domain of the receptor which is specific for the C-site of transferrin and HFE. The relevance of bismuth incorporation by the transferrin receptor-mediated iron acquisition pathway is discussed.     

Structural and functional stability of the mature transferrin receptor from human placenta

The transferrin receptor (TfR) is a N- and O-glycosylated transmembrane protein mediating the cellular iron uptake by binding and internalization of diferric transferrin. In this study, rate constants and dissociation constants of 125I-ferri-transferrin binding to the human TfR were examined dependent on receptor glycan composition, pH, bivalent cations, and temperature. To do so, purified human placental TfR was noncovalently immobilized to polystyrene surfaces and subjected to alterations in various parameters. We found that transferrin binding was clearly dependent on a receptor pretreatment with buffers of various pH in that most of the TfR molecules irreversibly lost transferrin binding activity below pH 6.5. However, the dissociation constant of the remaining active binding sites was not affected. Similarly, we were able to define the thermal stability of the receptor as a function of transferrin binding ability. Binding of transferrin was completely lost provided that the receptor was pretreated at temperatures of at least 65 degrees C. Treatment with EDTA also caused an irreversible loss of transferrin binding activity, indicating that the functionally active conformation of the mature TfR depends on bivalent cations. In order to examine the role of the receptor glycans, we enzymatically removed the sialic acid residues, the hybrid and oligomannosidic N-glycans, or all types of N-glycans. In contrast to the parameters described above, all desialylated and N-deglycosylated TfR variants had exactly the same transferrin binding properties as the native TfR. To assess changes in the secondary structure of the receptor, circular dichroic spectra were recorded from TfR at pH 5.0, from heat pretreated receptor and from deglycosylated TfR. Since the receptor did not exhibit detectable changes in the CD spectrum of the deglycosylated receptor, it can be concluded that the N-linked carbohydrates of the mature, fully processed TfR are not essential for transferrin binding and conformational stability.     

Reconstitution of the transferrin receptor in lipid vesicles. Effect of cholesterol on the binding of transferrin

Purified rabbit reticulocyte transferrin receptors were incorporated into phosphatidylcholine vesicles containing varying amounts of cholesterol. The binding of transferrin to the receptor in the reconstituted vesicles had three distinct characteristics: (1) The binding of transferrin exhibited the two components characteristic of transferrin binding to erythroid cells, a saturable, specific component and a nonsaturable, nonspecific component. (2) Transferrin binding exhibited positive cooperativity at low cholesterol/phospholipid (C/P) molar ratios. However, the cooperativity diminished and then disappeared as the C/P molar ratios were increased to the levels found in circulating red blood cells. (3) The amount of specific transferrin binding to the reconstituted vesicles also decreased as the C/P molar ratio was increased. These results indicate that in the reconstituted system the lipid environment plays a significant role in the expression of transferrin receptors.     

Ligand binding and allostery can emerge simultaneously

A heterotropic allosteric effect involves an effector molecule that is distinct from the substrate or ligand of the protein. How heterotropic allostery originates is an unanswered question. We have previously created several heterotropic allosteric enzymes by recombining the genes for TEM1 beta-lactamase (BLA) and maltose binding protein (MBP) to create BLAs that are positively or negatively regulated by maltose. We show here that one of these engineered enzymes has approximately 10(6) M(-1) affinity for Zn(2+), a property that neither of the parental proteins possesses. Furthermore, Zn(2+) is a negative effector that noncompetitively switches off beta-lactam hydrolysis activity. Mutagenesis experiments indicate that the Zn(2+)-binding site does not involve a histidine or a cysteine, which is atypical of natural Zn(2+)-binding sites. These studies also implicate helices 1 and 12 of the BLA domain in allosteric signal propagation. These results support a model for the evolution of heterotropic allostery in which effector affinity and allosteric signaling emerge simultaneously.     

Detection of a molecular complex between ras proteins and transferrin receptor

Immunoprecipitation of extracts of human carcinoma cell lines with three different monoclonal antibodies generated against ras proteins revealed the coprecipitation of a 90,000 dalton protein. The coprecipitated protein was identified as the transferrin receptor by comigration in both reducing and nonreducing SDS-polyacrylamide gels, by absorption with a monoclonal antibody directed against transferrin receptor, and by analysis of partial proteolysis products. Coprecipitation of the transferrin receptor with three monoclonal antibodies with differing specificities to ras proteins, as well as the inability to coprecipitate the transferrin receptor from cell extracts from which ras proteins were depleted by preabsorption, indicates that ras proteins and the transferrin receptor form a molecular complex. This complex is disrupted by addition of transferrin to cell extracts. These findings suggest that ras proteins function in regulation of cell growth via interaction with the cell surface receptor for transferrin.     

Structural requirements for high efficiency endocytosis of the human transferrin receptor.

Wild-type and mutant human transferrin receptors (TR) have been expressed in chicken embryo fibroblasts using a helper-independent retroviral vector. By functional studies of the mutant TRs, we have identified the tetrapeptide sequence, YXRF, in the cytoplasmic tail of the receptor as the internalization signal required for high efficiency endocytosis and shown that transplanted internalization signals from the low density lipoprotein receptor (LDLR) and the cation-independent mannose-6-phosphate receptor (Man-6-PR) are able to promote rapid internalization of the human TR. A six-residue LDLR signal, FDNPVY, is required for activity in TR, whereas a four-residue Man-6-PR signal, YSKV, is sufficient. These data indicate that internalization signals are interchangeable self-determined structural motifs and that signals from type I membrane proteins are active in a type II receptor. Putative internalization signals in the cytoplasmic tails of other receptors and membrane proteins can be identified based on the sequence patterns of the LDLR, Man-6-PR, and TR signals. Two such putative four-residue internalization signals, one from the poly-Ig receptor and one from the asialoglycoprotein receptor, were tested for activity by transplantation into TR and were found to promote high efficiency internalization. These results suggest that an exposed tight turn is the conformational motif for high efficiency endocytosis.     

Identification of the segment for binding of transferrin using a cellular receptor

Methods of proteolysis, radio-immunoblotting and affinity chromatography were used for identifying the human transferrin molecular binding site with cellular receptor. Monoclonal antibody HTF-14 which inhibits binding of the transferrin molecule with the receptor was employed. We showed that this monoclonal antibody has an antigenic determinant of the conformational type which is localized on the COOH-sublobe of the NH2-lobe of the molecule of the transferrin.     

Comparison of the interactions of transferrin receptor and transferrin receptor 2 with transferrin and the hereditary hemochromatosis protein HFE

The transferrin receptor (TfR) interacts with two proteins important for iron metabolism, transferrin (Tf) and HFE, the protein mutated in hereditary hemochromatosis. A second receptor for Tf, TfR2, was recently identified and found to be functional for iron uptake in transfected cells (Kawabata, H., Germain, R. S., Vuong, P. T., Nakamaki, T., Said, J. W., and Koeffler, H. P. (2000) J. Biol. Chem. 275, 16618-16625). TfR2 has a pattern of expression and regulation that is distinct from TfR, and mutations in TfR2 have been recognized as the cause of a non-HFE linked form of hemochromatosis (Camaschella, C., Roetto, A., Cali, A., De Gobbi, M., Garozzo, G., Carella, M., Majorano, N., Totaro, A., and Gasparini, P. (2000) Nat. Genet. 25, 14-15). To investigate the relationship between TfR, TfR2, Tf, and HFE, we performed a series of binding experiments using soluble forms of these proteins. We find no detectable binding between TfR2 and HFE by co-immunoprecipitation or using a surface plasmon resonance-based assay. The affinity of TfR2 for iron-loaded Tf was determined to be 27 nm, 25-fold lower than the affinity of TfR for Tf. These results imply that HFE regulates Tf-mediated iron uptake only from the classical TfR and that TfR2 does not compete for HFE binding in cells expressing both forms of TfR.     

Demonstration of the specific binding of bovine transferrin to the human transferrin receptor in K562 cells: evidence for interspecies transferrin internalization

Specific binding of ferric bovine transferrin to the human transferrin receptor was investigated using K562 cells propagated in serum-free medium without transferrin supplemented with 10(-5) elemental iron. Affinity chromatography of solubilized extracts of K562 cells surface-labeled with 125I was performed using bovine transferrin- and human transferrin-Sepharose 4B resins. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of resin eluates reveal that bovine transferrin specifically binds a Mr = 188,000 protein which dissociates into a Mr = 94,000 protein under reducing conditions, a finding identical to what is seen with human transferrin. The Mr = 94,000 reduced protein isolated by bovine transferrin resin shows an identical one-dimensional partial proteolytic digestion map with that of the human transferrin receptor. Unlabeled bovine transferrin was shown to specifically compete 125I-labeled human transferrin from the human transferrin receptor on the surface of K562 cells at 4 degrees C in a similar manner as unlabeled human transferrin; however, approximately a 2,000-fold higher concentration of bovine ligand was required to achieve comparable competition (50% inhibition of binding). Indirect immunofluorescence cytolocalization of bovine transferrin in K562 cells grown in serum-free medium supplemented with ferric bovine transferrin reveal patterns similar to those seen for human transferrin (both focal perinuclear and diffuse cytoplasmic fluorescence). Monensin treatment results in a dramatic accumulation of bovine ligand in perinuclear aggregates, suggesting that it is recycled through the Golgi, as is human transferrin. K562 cells grown in serum-free medium supplemented with either 300 micrograms/ml of ferric human or ferric bovine transferrin were found to demonstrate superimposable growth curves.     

Identification of the epitope of a monoclonal antibody that disrupts binding of human transferrin to the human transferrin receptor

The molecular basis of the transferrin (TF)-transferrin receptor (TFR) interaction is not known. The C-lobe of TF is required to facilitate binding to the TFR and both the N- and C-lobes are necessary for maximal binding. Several mAb have been raised against human transferrin (hTF). One of these, designated F11, is specific to the C-lobe of hTF and does not recognize mouse or pig TF. Furthermore, mAb F11 inhibits the binding of TF to TFR on HeLa cells. To map the epitope for mAb F11, constructs spanning various regions of hTF were expressed as glutathione S-transferase (GST) fusion proteins in Escherichia coli. The recombinant fusion proteins were analysed in an iterative fashion by immunoblotting using mAb F11 as the probe. This process resulted in the localization of the F11 epitope to the C1 domain (residues 365-401) of hTF. Subsequent computer modelling suggested that the epitope is probably restricted to a surface patch of hTF consisting of residues 365-385. Mutagenesis of the F11 epitope of hTF to the sequence of either mouse or pig TF confirmed the identity of the epitope as immunoreactivity was diminished or lost. In agreement with other studies, these epitope mapping studies support a role for residues in the C1 domain of hTF in receptor binding.