M. Neu

X-ray solution scattering reveals conformational changes upon iron uptake in lactoferrin, serum and ovo-transferrins.

X-ray solution scattering has been used for studying the structural changes that take place upon uptake and release of iron from serum and chicken ovo-transferrin and human lactoferrin. In the case of chicken ovo-transferrin, data have been obtained for both the intact protein and the isolated N and C-lobes with and without iron. These studies reveal that both lobes undergo a change that is consistent with an opening of the inter-domain cleft when iron is removed from the protein. We suggest that the conformational change of the protein increases the specificity of receptor binding and that the closed configuration of the iron-loaded protein is one, or perhaps the, decisive step in the mechanism for receptor-mediated endocytosis.

The crystal and molecular structures of diferric porcine and rabbit serum transferrins at resolutions of 2.15 and 2.60 Å, respectively

The serum transferrins are monomeric proteins with a molecular mass of around 80 kDa and are responsible for the transport of iron in vertebrates. The three-dimensional structures of diferric porcine and rabbit serum transferrin have been refined against X-ray diffraction data extending to 2.15 and 2.60 A, respectively. Data for both proteins were collected using synchrotron radiation at temperatures of 277 K. The porcine protein crystallizes in the space group C2, with unit-cell parameters a = 223.8, b = 44.9, c = 78.9 A, beta = 105.4 degrees with one molecule in the asymmetric unit. The structure was solved by molecular-replacement methods using rabbit serum transferrin as the search model. The structure was refined using REFMAC, with a final residual of 13.8% (R(free) = 18.2% for a 5% data sample) for all data to 2.15 A. The final model comprises 5254 protein atoms, two Fe(3+) cations and two CO(3)(2-) anions, one N-acetyl glucosamine moiety and 494 water molecules. The rabbit protein crystallizes in space group P4(3)2(1)2, with unit-cell parameters a = 127.2, c = 144.9 A and one molecule per asymmetric unit. The structure was solved using the method of multiple isomorphous replacement and refined using REFMAC to give a final residual of 18.6% (R(free) = 22.2% for a 5% data sample) for all data to 2.60 A. The final model comprises 5216 protein atoms, two Fe(3+) cations and two CO(3)(2-) anions, a Cl(-) anion and 206 solvent molecules; there is no clear indication of the carbohydrate moiety attached to Asn490 (rabbit serum numbering). Both molecules adopt a bilobal structure typical for members of the transferrin family. Each of the structurally homologous lobes contains two dissimilar domains with a single iron-binding site buried within the interdomain cleft. The porcine serum protein lacks an interdomain disulfide bridge close to the connecting peptide between the lobes, but this seems to have little effect on the overall orientation of the lobes. The N-lobes of both proteins possess lysine residues, one from each of the two domains, that lie in close proximity to one another to form the so-called dilysine trigger. The more acid-labile release of iron from serum transferrins than from lactoferrins is discussed.

The mechanism of iron uptake by transferrins: the structure of an 18 kDa NII-domain fragment from duck ovotransferrin at 2.3 A resolution.

The molecular structure of an iron-containing 18 kDa fragment of duck ovotransferrin, obtained by proteolysis of the intact protein, has been elucidated by protein crystallographic techniques at 2.3 A resolution. This structure supports a mechanism of iron uptake in the intact protein whereby the binding of the synergistic (bi)carbonate anion is followed by binding of the metal with the lobe in the open configuration. These stages are then followed by domain closure in which the aspartic acid residue plays a further key role, by forming an interdomain hydrogen-bond interaction in addition to serving as a ligand to the iron. This essential dual role is highlighted by model building studies on the C-terminal lobe of a known human variant. In this variant a mutation of a glycine by an arginine residue enables the aspartic acid to form an ion pair and reduce its effectiveness for both metal binding and domain closure. The X-ray structure of the 18 kDa fragment strongly suggests that the histidine residue present at the iron binding site of the intact protein and arising from the second interdomain connecting strand has been removed during the preparative proteolysis.

The mechanism of iron uptake by transferrins: the X-ray structures of the 18 kDa NII domain fragment of duck ovotransferrin and its nitrilotriacetate complex.

In a previous paper [Lindley et al. (1993), Acta Cryst. D49, 292-304], the X-ray structure analysis of the 18 kDa fragment of duck ovotransferrin, corresponding to the NII domain of the intact protein, was reported at a resolution of 2.3 A. In this structure, the Fe(III) cation binds to two tyrosine residues and the synergistic carbonate anion in an identical manner to that found in the intact protein. However, the aspartate and histidine residues, normally involved in iron binding in transferrins, are absent in the fragment and it was not possible to unequivocally define what had replaced them. The electron density was tentatively assigned to be a mixture of peptides, presumably resulting from the proteolytic preparation of the fragment, binding to the iron through their amino and carboxylate termini. A more recent X-ray analysis of the fragment, from a different preparation, has resulted in a structure at 1.95 A, in which glycine appears to be the predominant residue bound to the cation. In an alternative attempt to clarify the binding of iron to the 18 kDa fragment, the metal was removed by dialysis and replaced in the form of ferric nitrilotriacetate. Crystallization of this complex has resulted in an X-ray structure at 1.90 A in which the Fe(III) is bound to the synergistic carbonate anion and only one tyrosine residue in a manner almost identical to the intact protein. The carboxylate groups and the tertiary amino group of the nitrilotriacetate occupy the remaining coordination sites. The second tyrosine residue, Tyr95, is not bound directly to the iron. The implication of these structures with respect to the mechanism of iron binding by the transferrins is addressed.

Quick Fluorescence XAFS: A Technique for Recording Fluorescence X-ray Absorption Spectra during Chemical/Biochemical Reactions

The Quick EXAFS (QuEXAFS) technique (1,2) provides an alternative way of recording X-ray absorption fine structure (XAFS) data where scan time is minimised by continuous scanning of the monochromator. In contrast to the dispersive technique, QuEXAFS is capable of obtaining data in fluorescence mode as well as in transmission and is therefore suitable for dilute samples. The reduction in data collection time makes it feasible to follow some biochemical or chemical reactions at room temperature with the XAFS technique. We have recently commissioned a QuEXAFS experimental set up on station 9.3 on the 5T wiggler at Daresbury SRS. This station is equipped with a vertically focussing mirror, which in addition to providing extra flux also performs harmonic rejection due to the critical angle cutoff off the mirror reflectivity. The use of the mirror for some harmonic rejection is complemented by the use of a very stable two crystal water cooled monochromator. This provides additional harmonic rejection yet is stable enough to be used without a dynamic harmonic rejection servo, an important consideration for QuEXAFS scans. The monochromator Bragg angle is driven by a dc motor system encoded by a 0.1mdeg encoder. This system can be optimised for constant angular velocity scans and thus is well suited for QuEXAFS and avoids vibrations associated with stepper motors driven at speed. The high flux on the sample together with the use of a high count rate 13 element solid state fluorescence detector (3) have allowed us to obtain quality data on a 5mM solution in less than a minute. Studies have also been carried out on a dilute solution of a transferrin intermediate. We note that this is the first fluorescence QuEXAFS reported and is the lowest concentration for which a XAFS spectrum has been collected in less than a minute. Improvements iii the sample geometry and in the scanning protocol will allow faster data collection on more dilute systems.

Influence of protein dynamics on the metal-sites of ovotransferrin.

Using the perturbed angular correlations (PAC) technique, the formation of hafnium-ovotransferrin complexes has been studied. Two binding configurations at each of the two specific binding-sites of the protein have been observed. They are characterized by well-defined electric quadrupole frequencies. Information about the dynamics of the protein was derived from temperature dependent measurements of the relaxation constant. The well-resolved spectra taken with fast BaF2-detectors allow a precise determination of the relaxation behaviour of the protein. The results are compared with the predictions from a hydrodynamic model for the reorientation of macromolecules. Thus the hydrodynamic volume of ovotransferrin and its N-terminal half-molecule were determined. The ovotransferrin volume is in agreement with a value derived for human serum transferrin from small angle neutron scattering. From experiments with immobilized protein material there is evidence for internal protein dynamics which is probed by the Hf-ion bound to the specific metal-sites.

Evidence of internal protein dynamics in transferrins from TDPAC experiments

The relaxation in liquid transferrin and ovotransferrin samples has been studied at different temperatures using the TDPAC method. Information about reorientation and internal dynamics has been obtained from immobilized protein samples. Characteristic differences between the two proteins will be discussed.

Hafnium binding to comparison: comparison between lactoferrin and other transferrins

The TDPAC method was used to study the electric field gradients at the metal sites of human and bovine lactoferrin. Two specific binding configurations were observed. The distribution between these configurations depends on the phosphate content, the pH, and the temperature of the samples. The electric field gradients are compared with the results of previous studies for human and rat serum transferrin, and hen ovotransferrin.