The Crystal Structure of Apo Human Serum Transferrin Provides Insight Into Inter-Lobe

02/03/06

The Crystal Structure of Apo Human Serum Transferrin Provides Insight into Inter-Lobe Communication and Receptor Binding

Jeremy Wally1*, Peter J. Halbrooks2*, Clemens Vonrhein3, Mark A. Rould4, Stephen J. Everse2, Anne B. Mason2 and Susan K. Buchanan1

1National Institute of Diabetes, Digestion and Diseases of the Kidney, National Institutes of Health, 50 South Drive, Bethesda, MD 20892 USA

2Department of Biochemistry, University of Vermont, College of Medicine, 89 Beaumont Avenue, Burlington, VT 05405 USA

3Global Phasing Ltd., Sheraton House, CastlePark, Cambridge, CB3 0AX, UK

4Department of Molecular Physiology and Biophysics, University of Vermont, College of Medicine, 89 Beaumont Avenue, Burlington, VT 05405 USA

*These authors contributed equally. Additionally the Mason and Buchanan laboratories shared equally in solving the structures and generating the manuscript.

Correspondence should be addressed to:

Dr. Susan K. Buchanan

Laboratory of Molecular Biology

NIDDK, NIH

50 South Drive, Room 4503

Bethesda, MD 20892-8030

Phone: 301-594-9222

Fax: (+1) 301-480-0597

Coordinates- The atomic coordinates and structure-factor amplitudes of hTF have been deposited in the Protein Data Bank (accession code XXXX).

Abstract

The function of serum transferrin is to reversibly bind Fe (III) in each of two lobes and to deliver it to actively dividing cells by a receptor mediated pH dependant process. The binding and release of iron from transferrin is accompanied by a large conformational change in which two subdomains in each lobe close or open with a rigid twisting motion around a hinge in the bottom of each. The crystal structure of recombinant non-glycosylated human serum transferrin (hTF) lacking iron (apo-hTF) has been solved at a resolution of 2.7 Å using a MAD phasing strategy in which the nine methionines in hTF were replaced by selenomethionine. A baby hamster kidney (BHK) cell expression system was used to produce the recombinant hTF. This is the first report of using BHK cells to incorporate selenomethionine into a recombinant protein. The structure of commercially available glycosylated apo-hTF (from human serum) has also been determined to a resolution of 2.7 Å by molecular replacement using the structures of the human apo-N-lobe and the rabbit holo-C1-subdomain as search models. The two solutions are identical. The structure of apo-hTF represents the first published model for human TF and reveals that, in contrast to family members, lactoferrin and ovotransferrin, both lobes of hTF are almost equally open (62° and 52° rotations about the hinge are required to open the N-and C-lobe, respectively). Seven molecules of citrate are found in apo-hTF, six of which reside in the iron binding sites. Availability of this structure is critical to a complete understanding of the process by which each lobe of hTF binds and releases iron. In particular the relationship of apo-hTF to the hTF receptor is evaluated.
Introduction

The transferrins are a family of bilobal iron-binding proteins that includesserum transferrin (TF), ovotransferrin (oTF), and lactoferrin (LTF). These proteins play the crucial biological role of binding ferric iron, and keeping it in solution, thereby controlling the levels of this important metal in bodily fluids {1196}{303}. Human serum transferrin (hTF), is made up of 679 amino acids (~80 kDa), and functions in the transport of iron. Synthesized in the liver and secreted into the bloodstream, hTF acquires Fe(III) from cells in the intestine and delivers the iron to cells by binding to specific transferrin receptors (TFR) on the cell surface. The entire hTF/TFR complex is taken up by receptor-mediated endocytosis and iron is released within the endosome in a process that involves both a reduction in pH and the receptor itself {276}. Critical to the efficient delivery of iron, hTF lacking iron stays bound to the TFR at low pH and is returned to the cell surface where it is released to acquire more iron.

Strong homologies exist, both between TF family members, and between the two lobes of any given TF. The N- and C-lobes are linked by a short peptide chain. Each lobe is further divided into two subdomains (designated N1 and N2 and C1 and C2), separated by a hinge and giving rise to a deep cleft in which the iron-binding sites reside.Structures of diferric LTF {5}{996}, oTF {3523}, as well as rabbit {181} {5055} and porcine {5055} serum TF reveal that iron is coordinated in each lobe by four highly conserved amino acid residues:an aspartic acid (the sole ligand from the N1 or C1 domain), a tyrosine in the hinge at the edge of the N2 or C2 domain, a second tyrosine within the N2 or C2 domain and a histidine at the hinge bordering the N1 or C1 domain. In addition, the iron atom is coordinated by two oxygen atoms from a synergistic carbonate anion which is itself stabilized by a hydrogen bond from conserved arginineresidue in the N2 or C2 domain. Further stabilization of the carbonate is provided by a conserved threonine residue and a main chain amide at the end of an α-helix {6259}.

Iron release from hTF depends upon a number of factors, including pH, a chelator, and salt, as well as the specific TFR {4942}{3621}{3982}. ADD CHELATOR AND ANION As clearly revealed by the crystal structure of apo-oTF{4425} in comparison to diferric oTF {3523}, a key feature of TF family members is the large conformational change accompanying iron release involving the opening and a twist of the two subdomains in each lobe. Also available are structures of the recombinant hTF N-lobe (hTF/2N) yielding models of both the apo- and ferric-conformations {4118}{4161}. Although, as mentioned above, the N- and C-lobes have considerable sequence similarity, substantial differences in their properties exist and have been the subject of extensive investigation {836}{1804}{5071}. Specifically, many studies show that the rate of iron release from the C-lobe is considerably slower than the rate of release from the N-lobe particularly at the putative endosomal pH of ~5.6{5310}{5256}{5512}. In one such study from our laboratory a 200-fold difference is found{6159}. At least some of the difference is attributed to the presence of a “dilysine trigger” in the N-lobe replaced by a triad of residues in the C-lobe {5512}. Thus the N-lobe of diferric hTF features two lysines, Lys206 and Lys296, residing on opposite sides of the iron binding cleft and sharing a hydrogen bond in the closed conformation {3155}{4377} In the C-lobe a triad of residues, Lys534, Arg632 and Asp634, takes the place of the lysine pair {5512}{6157}. These unique features play a large role in opening the cleft by a “triggering” action in which protonation leads to repulsion of two like charged residues residing on opposite sides of the binding cleft.

A 7.5 Å structuredetermined by cryo-electron microscopy of diferric hTF bound to its specific receptor involved docking the human N-lobe and rabbit C-lobe onto an electron density map of TFR {5817}. This work offersa preliminary view of the regions of hTF and TFR which interact; it is suggested that both the N1 and N2 subdomains of the N-lobe contact the TFR while only the C1-domain of the C-lobe of hTF is involved in the interaction. Independently, a number of the residues on the TFR which contact hTF had been previously identified by the extensive TFR mutagenesis and binding studies of Giannetti et al. {5820}. Of great interest is the finding that a ~9 Å movement of the ferric N-lobe of hTF relative to the ferricC-lobe isrequiredtodock the two lobes onto the TFR structure. Although it is well established that the TFR discriminates between diferric, the two monoferric species and apo-TF, favoring the conformation with iron in each lobe at pH 7.4, the basis of this discrimination is notunderstood {87}{4072}{3291}. Significantly, our recent work with authentic monoferric hTF constructs establishes that each lobe contributes equally (and non-additively) to the binding energy of the interaction {6159}.A structure of human apo-hTF is required to determine whether the change in orientation of the two lobesthat resultsfrom iron release could provide further insight into the receptor interaction.

The structure of full-length apo-hTF has been derived from two sets of data; both a non-glycosylated recombinant form of hTF (pH 6.5)and a glycosylated form of hTF (pH 7.0) were solved to a resolution of 2.7 Å In thesetwo structures which are identical within the limits of the resolution, both the N- and C-lobes are in the open conformation. This work represents the first mammalian TF structure with an apo C-lobe, the first published structure of hTF and the first report of use of a BHK expression system to replace the nine methionines residues in hTF with selenomethionine. The apo-hTF structure allows comparisons to other relevant structures including those for diferric porcine (2.15 Å) and rabbit TF (2.6 Å) {5055}, and an unpublished model for a monoferric hTF with iron in the C-lobe (3.3 Å).{3611} which unfortunately remains unrefinedand incomplete.

Experimental Procedures

Production of hTF-To produce recombinant non-glycoylated hTF (hTF-NG)with selenomethionine (SeMet)substituted for methionine, BHK cells transfected with the pNUT plasmid containing the sequence of the N-His tagged hTF-NG were placed into four expanded surface roller bottles {5256}. Addition of culture media containing SeMet results in a significant deterioration of the BHK cells within a 24-48 hour period. To maximize the incorporation, media containing SeMetwas added when production of recombinant hTF was at a maximum as determined by a competitive immunoassay {4908}. Briefly, the adherent BHK cells were grown in DMEM-F12 containing 10% FBS. This medium was changed twice at two day intervals, followed by addition of DMEM-F12 containing 1% Ultroser G and 1 mM butyric acid (BA). After one or two changes in this medium, 200 mL of SeMet containing DMEM-F12 (lacking normal methionine), with BA and Ultroser G was added to each roller bottle. Following a four hour wash-in period this media was discarded and replaced with 250 mL of the same mediumfor an additional 48 hours of incubation. The recombinant hTF-NG was purified from the medium as described elsewhere in detail {6025}. In two production runs between 8 and 16 mg of Se Met containing N-His hTF-NG was produced, of which approximately half was recovered. Electrospray mass spectrometry analysis indicated a mass consistent with incorporation of 8-9 SeMet residues (data not shown).

Human serum TF lacking iron was obtained from Sigma-Aldrich and applied to a 10 mL Q-Sepharose High Performance column (GE Heathcare) equilibrated with 20 mM Tris-HCl, pH 7.5. It was eluted using a linear gradient from 0 to 150 mM NaCl over X column volumes. Peak fractions were pooled and dialyzed overnight into 20mM Tris-HCl, pH 8.0, 20mM sodium carbonate and 200 mM sodium chloride.

Crystallization- Recombinant hTF-NG (with or without a His tag) at a concentration 15 mg/mL in 0.1 M NH4HCO3, was mixed with an equal volume of reservoir solution composed of 0.3 M ammonium citrate (pH 6.5) and 16-18% PEG 3350 at 20 C°. The SeMet N-His hTF-NG required slightly lower levels of PEG 3350 and streak seeding using the non-SeMet labeled hTF-NG crystals. Clear crystals (0.2 mm x 0.3 mm x 0.4 mm) formed in 3-5 weeks. Crystals of the SeMet-labeled protein were essentially isomorphous with those of the native protein, showing similar cell dimensions and crystallizing in the orthorhombic space group P212121 with two molecules in each asymmetric unit (Table 1). All crystals were cryoprotected by addition of 0.5 µL of a solution of 50% PEG 3350 and 60% ethylene glycol to the hanging drop.

Apo-hTF-Gly was concentrated to 30 mg/ml and screened using a Mosquito crystallization robot (TTP Labtech) with a hanging drop format. Conditions yielding the best crystals were then further refined using 24 well VDX plates (Hampton Research). The crystals used for data collection were obtained from hanging drops with a well solution of 0.2 M tri-ammonium citrate pH 7.0, 20% PEG-3350, and 15% glycerol, incubated at 21°C. Drop sizes varied from 2 to 16 μl and consisted of equal parts of protein and well solution. Crystals grew in ~24 hours and were flash frozen in propane cooled to -170°C.

Data collection and refinement- For the SeMet labeled hTF samples, data was collected at 100 K on beamlines X26C (peak and remote wavelengths for Se) and X25 (peak, inflection, and remote wavelengths) at Brookhaven National Laboratory. The data was initially processed using DENZO {5304} and intensities were reduced and scaled using SCALEPACK {5304}. In order to find the selenium sites, MAD data was prepared with XPREP and analyzed with ShelxD {6057}. The data sets were combined and refinement of the selenium sites was carried out using SHARP {6058}. Profess {5300} was used to find NCS, which revealed two hTF molecules in the asymmetric unit. Following a round of density modification, the structure of apo N-lobe (PDB_1BP5) was used as a search model for a phased translation and rotation function. Subsequent model building of the C-lobe was done using O {2987} in a stepwise manner by incorporating fragments of porcine TF (converted to the human sequence). Iterative rounds of density modification and phase recombination were performed with autoSHARP {6058}. Refinement was accomplished using both CNS {5299} and CCP4 {5300}; final rounds of refinement were carried out using BUSTER/TNT {6059}{6060}{6061}.

X-ray diffraction data from the apo-hTF-Gly crystals was collected at the SER-CAT beamline 22ID (APS). The crystal belonged to the orthorhombic space group P212121, with cell constants a = 88.3, b = 103.3, c = 200.4, and two molecules in the asymmetric unit with a solvent content of 59.3%. The images were processed, scaled and merged using the program HKL2000 {5304}(Table 1), and the structure was solved using the molecular replacement program Phaser {6428}. A search model consisting of the human apo-N-lobe structure (residues 4 to 331) and the rabbit holo-C1-domain (residues 342 to 424 and 579 to 676) was used, leading to a single solution containing two copies of each domain. After a single round of refinement using the program REFMAC {6430} with medium NCS restraints over residues 4 to 679, the rabbit holo-C2-domain (residues 428 to 576) was manually fitted into the electron density map. The rabbit C1- and C2-domains were then mutated to the human sequence in O {6429} and the linker regions between the N- and C-lobes and between the C1- and C2-domains were built into the electron density.

RESULTS AND DISCUSSION

Quality of Final Model- The structure of apo-hTF is shown in Figure 1A. Asin all other TF structures, no electron density for residues 1-3 is visible. Even the His- tagged recombinant hTF with 14 extra amino acids at the N-terminus showed no density in this region implying that the amino terminus is very mobile. Details of the refinements and final models are summarized in Table 1. Each solution has two molecules per asymmetric unit and because the two solutions had an r.m.s. deviation of only 0.64 Å with 676 equivalent Cα positions we are presenting and discussing them as a single structure. Additionally we will focus on the C-lobe since the structures of the isolated human N-lobe with and without iron have been discussed ingreat detail {4161} {4118}. No breaks in the main chain densityare observed in either molecule of apo-hTF and, as documented in Table 1, the geometry is good. Additionally, no density in the vicinity of Asn413 and Asn611 in the glycosylated hTF is observed indicating that the carbohydrate moieties are flexible and/or present in multiple conformations. Significantly the presence or absence of carbohydrate is without effect on iron release or interaction with the TFR {3110} as further confirmed by mass spectrometry studies{5604}. The Ramachandran plot of the main chain torsion angles shows 97.6% of the residues lie inmost favored or allowed regions Twenty two residues fall into the generously allowed regions and 6 residues reside in the disallowed region.Of these Leu 294 and 630 in each of the two molecules are comprise the central residues in classic γ -turns with and anglesofapproximately 77and -46°, respectively in the four sites {6056}. This structural feature was first noted in the hTF N-lobe {4161}, and subsequently has been observed in each lobe all TF structures to date.The Leu-Leu-Phe sequence which makes up the turn is found in both lobes of all mammalian and avian TF molecules and the C-lobe of all fish. The fish TF N-lobe sequences havea methionine residue substituted for the leucine at the middle position which would probably not interfere with formation of the turn.Interestingly, the Leu-Leu-Phe sequence is not conserved in the any insect TF N-lobeand is only partially conserved in the insect C-lobe sequences {6158}{6259}. While the role of the γ -turn has not been definitively established, it seems reasonable to suggest that it adds rigidity to this region of the structure {4161}. We believe that it is significant that theγ- turn immediately precedes dilysine residue Lys296 in the N-lobe and triad residue Arg632 in the C-lobe. Within this context it may help to stabilize the orientation of these two residues to provide better repulsion or triggering actionaiding in opening of the cleft.

Overall organization of the structure- As with all TF family members, the structure of hTF illustrates the bilobal nature of the molecule with the amino-terminal lobe (N-lobe, residues 1-331) and a carboxyl-terminal lobe (C-lobe, residues 340-679). The lobes are connected by a linker peptide (residues 332-340) that is unstructured but visible in ourmodel (although missing in the pig Tf structure {5055}). Each of the lobes is further separated into two domains: the N1 (1-92, 247-331) and C1 (339-425, 573-679) domains are each composed of two discontinuous sections of the polypeptide chain, while the N2 (93-246) and C2 (426-572) domains are composed of a single region of continuous polypeptide.