Selective crystallization of horse isoferritins

A crystallization screen based on alternative polymeric precipitants

Most commercially available crystallization screens are sparse-matrix screens with a predominance of inorganic salts and polyethylene glycols (PEGs) as precipitants. It was noted that commercially available screens are largely unsatisfactory for the purpose of the crystallization of multimeric protein and protein-nucleic acid complexes. This was reasoned to be a consequence of the redundancy in screening crystallization parameter space by the predominance of PEG as a precipitant in standard screens and it was suggested that this limitation could be overcome by introducing a variety of other organic polymers. Here, a set of 288 crystallization conditions was devised based on alternative polymeric precipitants and tested against a set of 20 different proteins/complexes; finally, a screen comprising the 96 most promising conditions designed to complement PEG- and salt-based commercial screens was proposed.

Purification, electrophoretic behavior, and kinetics of iron release of liver ferritin of Dasyatis akajei

From the liver of fish Dasyatis akajei, ferritin has been isolated by thermal denaturation and ammonium sulfate fractionation and then further purified by anion exchange chromatography and gel exclusion chromatography. The molecular weight of the liver ferritin of D. akajei (DALF) was measured to be 400 kDa by PAGE. Moreover, SDS-PAGE experimentation indicates that protein shell of DALF consists of the H and L subunits with molecular weight of 18 and 13 kDa, respectively. Using isoelectric focusing with pH ranging from 5.0 to 6.0, the ferritin purified by the PAGE exhibited three bands with different pI values in the gel slab. Diameters of the protein shell and iron core were also investigated by transmission electron microscope and determined to be 10-12 nm and 5-8 nm, respectively. A kinetic study of DALF reveals that the rate of self-regulation of the protein shell rather than the complex surface of the iron core plays an important role in forming a process for iron release with mixed orders.

Sequencing of cDNA clones that encode bovine ferritin H and L chains

The molecular weight of the liver-type subunit (L) of bovine ferritin is much larger than that of the heart-type subunit (H) as determined by SDS-PAGE (L, 20.5 kDa; H, 18.4 kDa). The migration of these two subunits on SDS-PAGE gels, relative to each other, is opposite to that reported for ferritin L and H subunits in other mammalian species (L, 19 kDa; H, 21 kDa). To determine the cause of this anomaly, full-length cDNA clones of the bovine L and H chains were isolated from a bovine spleen gamma gt11 cDNA library and sequenced. The amino acid sequences of the L and H chains of bovine ferritin, deduced from their cDNA sequences, contained open reading frames coding for 174 and 180 amino acid residues with calculated molecular weights of 19,856 and 20,920 Da, respectively. The deduced amino acid sequence of the L chain shows 86%, 84%, 87%, 83% and 83% homology with the amino acid sequences of horse, human, rabbit, rat and mouse L chains, respectively. The H chain displays a higher homology with the human, rat and mouse H chains (91%, 92% and 93%, respectively). In addition, the bovine L chain did not contain the extra octapeptide present in rodent L chains, and bovine, L and H chains did not react with concanavalin A. The bovine L and H chains expressed using a baculovirus expression system showed almost the same mobilities as those of bovine spleen ferritin, respectively, by SDS-PAGE. These results suggest that the much slower mobility of the bovine L chain compared with other mammalian L chains on SDS-PAGE cannot be attributed to insertion(s) of amino acid(s) or peptide(s) into the L chain, to the deletion(s) of them of it or to the addition of carbohydrate chains(s) but may result from significant differences in the binding affinity of SDS for bovine ferritin L chains.

Control of crystal forms of apoferritin by site-directed mutagenesis

Surface charges of protein molecules are not only important to biological functions but also crucial to the molecular assembly responsible for crystallization. Appropriate alteration in the surface charge distribution of a protein molecule induces new molecular alignment in the proper direction in the crystal and, hence, controls the crystal form. Apoferritin molecules are known to crystallize in two- and three-dimensional forms in the presence of cadmium ions, which bridge neighboring protein molecules. Here we report a controlled transformation of the apoferritin 2-D crystal by site-directed mutagenesis. In mutant apoferritin, two amino acid residues binding a cadmium-ion through their negative charge, were replaced by one type of nonionic amino acid residues. The amino acid residues, Asp-84 and Gln-86 in the sequence of recombinant (i.e., wild-type) horse L-apoferritin, were replaced by Ser. The wild-type apoferritin yielded a hexagonal lattice 2-D crystal in the presence of cadmium ions. In contrast, the mutant apoferritin yielded two types of oblique crystals independent of the presence of cadmium ions. Image reconstruction of electron micrographs of the mutant crystals made clear that the mutant apoferritin molecules oriented themselves with the 2-fold symmetry axis perpendicular to the crystal plane in both crystals, while the wild-type apoferritin molecules oriented themselves with the 3-fold symmetry axis perpendicular to the crystal plane. The changes of crystal forms and molecular orientation in the 2-D crystals were well explained by a change of the electrostatic interactions induced by the mutagenesis.

Solving the structure of human H ferritin by genetically engineering intermolecular crystal contacts

Ferritin is important in iron homeostasis. Its twenty-four chains of two types, H and L, assemble as a hollow shell providing an iron-storage cavity. Ferritin molecules in cells containing high levels of iron tend to be rich in L chains, and may have a long-term storage function, whereas H-rich ferritins are more active in iron metabolism. The molecular basis for the greater activity of H-rich ferritins has until now been obscure, largely because the structure of H-chain ferritin has remained unknown owing to the difficulties in obtaining crystals ordered enough for X-ray crystallographic analysis. Here we report the three-dimensional structure of a human ferritin H-chain homopolymer. By genetically engineering a change in the sequence of the intermolecular contact region, we obtained crystals isomorphous with the homologous rat L ferritin and of high enough quality for X-ray diffraction analysis. The X-ray structure of human H ferritin shows a novel metal site embedded within each of its four-helix bundles and we suggest that ferroxidase activity associated with this site accounts for its rapid uptake of iron.

Characterization of human ferritin H chain synthetized in Escherichia coli

We have inserted the coding region of the cDNA for human ferritin H chain into the expression vector pEMBLex2. The plasmid obtained is able to direct the synthesis of the ferritin H chain in Escherichia coli up to a concentration of 15% of total soluble proteins. All expressed subunits are found correctly assembled in the complete ferritin molecule, which can be easily purified. We have shown that the ferritin synthesized in E. coli has an Mr, electrophoretic mobility, and thermal stability similar to natural human isoferritins and is recognized by monoclonal antibodies specific for the H, but not by those for the L human ferritin chains.

Immunochemical characterization of human liver and heart ferritins with monoclonal antibodies

A library of 27 murine monoclonal antibodies was obtained by using human liver and heart ferritins as immunogens. The specificity of the antibodies for the two ferritins and their subunits was studied with five different methods. The antibodies elicited by the liver ferritin bound preferentially the immunogen and were specific for the L subunit. Some antibodies elicited by the heart ferritin had characteristics similar to the anti-liver antibodies, other ones bound preferentially the heart over the liver ferritin and were specific for the H subunit. Only two antibodies were able to bind both ferritins and subunits. Some anti-H and anti-L chain antibodies were used to develop and compare four types of immunoassay to quantitate isoferritins. The results indicate that heart ferritin is immunologically more heterogeneous than liver, the H and L subunits having large immunological differences with few, if any, identical epitopes; and that that the architecture of the immunoassays have a strong influence on the crossreactivity of the antibodies with the two isoferritins, probably because H and L chains are not arranged randomly in the assembled protein.