Structure and composition of ferritin cores isolated from human spleen, limpet (Patella vulgata) hemolymph and bacterial (Pseudomonas aeruginosa) cells

Ferritin cores isolated from human spleen, limpet (Patella vulgata) hemolymph and bacterial (Pseudomonas aeruginosa) cells have been investigated by high resolution transmission electron microscopy, electron diffraction and chemical analysis. Hemosiderin particles isolated from thalassemic spleens also have been studied. The results show that there is a marked difference in structure and composition of the biomineral phases. Human ferritin and hemosiderin particles are single domain crystals of hydrated iron (III) oxide (ferrihydrite). Lattice fringes were low in contrast and often discontinuous within the central regions of the core. Heat treatment of human ferritins results in a 5 A shrinkage in particle size and an increase in the single crystalline nature of the core. In contrast, lattice images and electron diffraction of limpet and bacterial cores show no evidence of long-range crystallographic order. Chemical analysis indicates a high inorganic phosphate (Pi) (Fe/Pi = 1.71) content in bacterial ferritin compared with human ferritin (thalassemic) (Fe/Pi = 21.0). The high Pi content of bacterial ferritin suggests a hydrated amorphous iron (III) phosphate mineral core. Structural disorder within the limpet and bacterial cores may be associated with increased Pi content and increased oxidation in Fe(II), resulting in rapid mineral deposition. Growth of the iron (III) oxide cores in human ferritin is discussed on the basis of high resolution electron microscopy results.     

The ferroxidase center is essential for ferritin iron loading in the presence of phosphate and minimizes side reactions that form Fe(III)-phosphate colloids

Ferritin iron loading was studied in the presence of physiological serum phosphate concentrations (1 mM), elevated serum concentrations (2–5 mM), and intracellular phosphate concentrations (10 mM). Experiments compared iron loading into homopolymers of H and L ferritin with horse spleen ferritin. Prior to studying the reactions with ferritin, a series of control reactions were performed to study the solution chemistry of Fe²⁺ and phosphate. In the absence of ferritin, phosphate catalyzed Fe²⁺ oxidation and formed soluble polymeric Fe(III)-phosphate complexes. The Fe(III)-phosphate complexes were characterized by electron microscopy and atomic force microscopy, which revealed spherical nanoparticles with diameters of 10–20 nm. The soluble Fe(III)-phosphate complexes also formed as competing reactions during iron loading into ferritin. Elemental analysis on ferritin samples separated from the Fe(III)-phosphate complexes showed that as the phosphate concentration increased, the iron loading into horse ferritin decreased. The composition of the mineral that does form inside horse ferritin has a higher iron/phosphate ratio (~1:1) than ferritin purified from tissue (~10:1). Phosphate significantly inhibited iron loading into L ferritin, due to the lack of the ferroxidase center in this homopolymer. Spectrophotometric assays of iron loading into H ferritin showed identical iron loading curves in the presence of phosphate, indicating that the ferroxidase center of H ferritin efficiently competes with phosphate for the binding and oxidation of Fe²⁺. Additional studies demonstrated that H ferritin ferroxidase activity could be used to oxidize Fe²⁺ and facilitate the transfer of the Fe³⁺ into apo transferrin in the presence of phosphate.     

Single crystals of dextran: 1. Low temperature polymorph

Lamellar single crystals of a synthetic, linear dextran and a slightly branched, bacterial dextran were grown at 95°C, by seeding their metastable solutions in mixtures of polyethylene glycol and water. The crystals gave well-resolved electron diffraction diagrams from which the unit cell parameters a ～ 2.563 nm and b ～ 1.021 nm were determined. The patterns displayed Pgg symmetry in a–b projection. A comparison of the elctron and X-ray powder diffraction diagrams indicated either 0.81 or 1.62 nm as the probable value for the c parameter of the unit cell.     

Three-dimensional structure of NADH: ubiquinone reductase (complex I) from Neurospora mitochondria determined by electron microscopy of membrane crystals

NADH: ubiquinone reductase (electron transfer complex I) has been isolated from Neurospora crassa mitochondria as a monodisperse protein-phospholipid-Triton X-100 complex (1:0.04:0.15, by weight). The enzyme is in the monomeric state, has a protein molecular weight of 610,000 and consists of about 25 different subunits. Membrane crystals of the enzyme complex have been prepared by adding mixed phospholipid-Triton X-100 micelles and then removing the Triton by dialysis. Diffraction patterns of the negatively stained membrane crystals extend to about 3.9 nm, with a unit cell size of 19 nm X 38 nm and gamma = 90 degrees. The two-sided plane group packing corresponding to pgg is p22(1)2(1). By combining four sets of tilted views, a low-resolution three-dimensional structure of the protein has been calculated. The structure shows that NADH: ubiquinone reductase extends 15 nm across the membrane, projecting 9 nm from one membrane side and 1 nm from the opposite side. Only about one-third of the total protein mass is located in the membrane. The structure of NADH: ubiquinone reductase is compared with that of ubiquinol: cytochrome c reductase determined by electron microscopy of membrane crystals.     

Crystallization and preliminary X-ray crystallographic analysis of plant ferritin from Glycine max

The iron storage protein ferritin from soybean (Glycine max) was expressed in E. coli and crystallized using the hanging drop vapor diffusion method with sodium tartrate as the precipitant. The crystals belong to the tetragonal I4(1)22 space group, with unit cell parameters a=b=324.0, c=182.7 A. The diffraction data were collected up to a resolution of 3.0 A with a multi-wire area detector.     

Crystal structure and morphology of poly(12-dodecalactone)

Poly(12-dodecalactone) (PDDL) crystals in the form of chain-folded lamellae were prepared by isothermal crystallization from a 1-hexanol solution. The lozenge-shaped crystals with and without spiral growth have been studied by transmission electron microscopy and atomic force microscopy. Wide-angle X-ray diffraction data, obtained from PDDL lamellae sedimented to form oriented mats and annealed solvent-cast film, were supplemented with morphological and structural data from electron microscopy. PDDL crystallizes as an orthorhombic form with a P2(1)2(1)2(1) space group and lattice constants of a = 0.746 +/- 0.001 nm, b = 0.500 +/- 0.001 nm, and c (chain axis) = 3.281 +/- 0.003 nm. There are two chains per unit cell, which existed in an antiparallel arrangement. The fiber repeat distance is appropriate for an all-trans backbone conformation for the straight stems. Molecular packing of this structure has been studied in detail, taking into account both diffraction data and energy calculations. The setting angles, with respect to the a axis, were +/-43 degrees for the corner and center chains according to intensity measurements and structure factor calculations. The optimized shift along the crystallographic c axis is 0.1c (0.328 nm). A final model was obtained to yield R = 0.180 with X-ray diffraction data and R = 0.162 with electron diffraction data. A brief comparison is also made with related polymer structures.     

Electron microscopic studies on microcrystals of D-ribulose-1,5-biphosphate carboxylase from Dasycladus clavaeformis roth (Ag.)

Crystals of D-ribulose-1,5-biphosphate carboxylase (E.C. 4.1.1.39), naturally occurring in the extraplasmatic space of the unicellular green algae $\text{Dasycladus clavaeformis}$ (Dasycladaceae), were studied by means of electron microscopy and optical diffraction. Optical diffraction patterns were obtained from thin sections. It is shown that the crystals are composed of cubic unit cells with α ～ 31.5 nm. The density of the crystals was estimated as 1.07 ± 0.005 g/ml, a value that gives evidence of the presence of 12 enzyme molecules per unit cell. Optical diffraction studies of the thin sectioned crystals revealed 4mm -symmetry with four 2-fold rotation axes, resulting in at least a 222-symmetry.     

Iron mobilization from ferritin by chelating agents

The release of iron from horse spleen ferritin by the chelating agents desferrioxamine B, rhodotorulic acid, 2,3-dihydroxybenzoate, 2,2′-bipyridyl and pyridine-2-aldehyde-2-pyridyl hydrazone (Paphy) has been studied in vitro. Ferritin prepared by classical procedures involving thermal denaturation releases its iron less effectively than ferritin isolated by a modified procedure that avoids this step. Desferrioxamine B and rhodotorulic acid are the most effective in releasing iron from both preparations of ferritin. When FMN is added, iron release by desferrioxamine B, rhodotorulic acid, and 2,3-dihydroxybenzoate was effectively blocked, whereas both bipyridyl and Paphy showed a marked simulation. A substantial increase in iron release was also observed for bipyridyl and Paphy with ascorbate; a less important increase was noted for rhodotorulic acid. EDTA exerted a marked inhibition of iron release from ferritin with rhodotorulic acid, 2,3-dihydroxybenzoate, bipyridyl, and Paphy. The effects of citrate and oxalate on iron release by the chelators was small. The effect of the concentration of flavin on iron release from ferritin by bipyridyl and desferrioxamine B have been studied. Desferrioxamine is unable to mobilize FeII from ferritin following reduction by reduced FMN, whereas bipyridyl can rapidly complex the ferrous iron. The results are discussed in the context of our current concepts of storage iron mobilization in the treatment of iron overload.     

Crosslinks between intramolecular pairs of ferritin subunits: effects on both H and L subunits and on immunoreactivity of sheep spleen ferritin

Ferritin is a multisubunit protein, controlling iron storage, with a protein coat composed of 24 subunits (up to three distinct types) in different proportions depending on cell type. Little is known about the subunit interactions in ferritin protein coats composed of heterologous subunits, despite the relevance to ferritin structure and ferritin function (iron uptake and release). Synthetic crosslinking is a convenient way to probe subunit contacts. Crosslinks between subunit pairs in ferritin protein coats are also a natural post-translational modification which coincides with different iron content in ferritin from sheep spleen; ferritin from sheep spleen also contains H and L subunits. Crosslinks synthesized by the reaction of ferritin low in natural crosslinks with difluorodinitrobenzene (F2DNB) reproduced the effects of the natural crosslinks on iron uptake and release. We now extend our observations on the structural effects of natural and synthetic crosslinks to include immunoreactivity of the assembled protein, with monoclonal antibodies as a probe. We also demonstrate, for the first time, ferritin peptides involved in an apparent H- and L-subunit contact: two peptides decreased 4X in cyanogen bromide peptide maps after F2DNB crosslinking were residues L-96-138 and H-66-96; the major DNP-dipeptide was Lys-DNP-Lys. Using the structure of an all L-subunit ferritin as a model, the most likely site for the H-L DNP crosslink is L-Lys 104 (C helix) and H-Lys 67 (B helix). The B helix forms the internal subunit dimer interface, a putative site of iron core nucleation. Alteration by crosslinks of the B helix could, therefore, explain the effect of crosslinks on ferritin iron uptake, release, and iron content.     

Neutron spin echo studies on ferritin: free-particle diffusion and interacting solutions

The dynamics of proteins are often studied by means of quasielastic neutron scattering (QENS), for example by time-of-flight methods. The spatial dimensions (10-20 nm) present in protein solutions are accessible by neutron scattering. In this article, a systematic study of diffusive dynamics of ferritin and apoferritin (=ferritin without iron core) is presented. Apoferritin consists of a spherical shell built of 24 protein units and carries net negative charge at pH 5. We have studied diffusive dynamics of ferritin solutions by neutron spin echo (NSE). We pay attention to an important feature of this technique compared to other QENS methods, which being the usage of a broad wavelength band. Using a more sophisticated fit function than usually used in NSE, we find as expected in low concentrated systems that the diffusion coefficient approaches the free-particle value of apoferritin and coincides with the diameter of the apoferritin shell (12.2 nm). In interacting solutions, the NSE results reveal that the dynamic picture of this complex liquid is dominated by slowing down of the dynamics. In low-salt solutions, a structure factor peak appears due to ordering of the ferritin molecules on the length scale of several intermolecular distances. We discuss the usage of different NSE fit functions for interacting solutions near the structure factor peak. Comparison of the dependence of elastic and dynamic data on the scattering vector value shows the influence of indirect interactions on the dynamic picture, irrespective of the way of data analysis, which being necessary due to the broad wavelength spectrum.     

Stoichiometry of Fe(II) oxidation during ceruloplasmin-catalyzed loading of ferritin.

Ceruloplasmin catalyzed the incorporation of iron into apoferritin with a stoichiometry of 3.8 Fe(II)/O2. This value remained the same when ferritin containing varying amounts of iron was used. Contrary to the "crystal growth" model for ferritin formation, no iron incorporation into holoferritin was observed in the absence of ceruloplasmin. Fe(II)/O2 ratios close to 2 were obtained for iron incorporation into apo- and holoferritin in Hepes buffer, in the absence of ceruloplasmin, indicating the formation of reduced oxygen species. Sequential loading of ferritin in this buffer resulted in increasing oxidation of the protein as measured by carbonyl formation. Sequential loading of ferritin using ceruloplasmin did not result in protein oxidation and a maximum of about 2300 atoms of iron were incorporated into rat liver ferritin. This corresponded to the maximum amount of iron found in rat liver ferritin in vivo after injection with iron. These results provide evidence for ceruloplasmin as an effective catalyst for the incorporation of iron into both apo- and holoferritin. The possibility that these findings may have physiological significance is discussed.     

Effect of chaotropes on the kinetics of iron release from ferritin by flavin nucleotides

BackgroundFerritins are ubiquitous multi-subunit iron storage and detoxification proteins that play a critical role in iron homeostasis. Ferrous ions that enter the protein's shell through hydrophilic channels are rapidly oxidized at dinuclear centers on the H-subunit before transfer to the protein's cavity for storage. The mechanisms of iron loading have been extensively studied, but little is known about iron mobilization. Fe(III) reduction can occur via rapid reduction by suitable reducing agents followed by chelation of Fe(II) ions or via direct and slow Fe(III) chelation. Here, the iron release kinetics from ferritin by FMNH₂ in the presence of various chaotropic agents are studied and their in-vivo physiological significance discussed.MethodsThe iron release kinetics from horse and human ferritins by FMNH₂ were monitored at 522 nm where the Fe(II)–bipyridine complex absorbs. The experiments were performed in the presence of different concentrations of three chaotropic agents, urea, guanidine HCl, and triton.Results and conclusionsUnder our experimental conditions, iron reductive mobilization by the non-enzymatic FMN/NAD(P)H system is limited by the concentration of FMNH₂ and is independent on the type or amount of chaotropes present. Diffusion of FMNH₂ through the ferritin pores is an unlikely mechanism for ferritin iron reduction. An iron mobilization mechanism involving rapid electron transfer through the protein shell is discussed.General significanceCaution must be exercised when interpreting the kinetics of iron mobilization from ferritin using the FMN/NAD(P)H system. The kinetics are highly dependent on the amount of dissolved oxygen and the concentration of reagents used.     

Molecular replacement studies on crystal structure of allophycocyanin from red algae Porphyra yezoensis

Using the crystal structure of allophycocyanin from cyanobacterium Spirulina platensis (APC-SP) as a search model,the crystal structure of allophycocyanin from red algae Porphyra yezoensis (APC-PY) has been studied by molecular replacement methods.The APC-PY crystals (Form 3) belong to the space group of R32,cell dimensions a=b= 10.53 nm,c=18.94 nm,α= β= 90°,γ=120°;there is one αβ monomer in each crystallographic asymmetric unit in the cell.The translation function search gave a unique peak with a correlation coefficient (Cc) of 67.0% and an R-factor of 36.1% for reflection data from 1.0 to 0.4 nm.Using the results by molecular replacement,the initial model of APC-PY was built,and the coincidence of the chromophore in APC-PY initial model with its 2Fo-Fc OMIT map further confirms the results by molecular replacement.     

"Opening" the ferritin pore for iron release by mutation of conserved amino acids at interhelix and loop sites.

Ferritin concentrates, stores, and detoxifies iron in most organisms. The iron is a solid, ferric oxide mineral (< or =4500 Fe) inside the protein shell. Eight pores are formed by subunit trimers of the 24 subunit protein. A role for the protein in controlling reduction and dissolution of the iron mineral was suggested in preliminary experiments [Takagi et al. (1998) J. Biol. Chem. 273, 18685-18688] with a proline/leucine substitution near the pore. Localized pore disorder in frog L134P crystals coincided with enhanced iron exit, triggered by reduction. In this report, nine additional substitutions of conserved amino acids near L134 were studied for effects on iron release. Alterations of a conserved hydrophobic pair, a conserved ion pair, and a loop at the ferritin pores all increased iron exit (3-30-fold). Protein assembly was unchanged, except for a slight decrease in volume (measured by gel filtration); ferroxidase activity was still in the millisecond range, but a small decrease indicates slight alteration of the channel from the pore to the oxidation site. The sensitivity of reductive iron exit rates to changes in conserved residues near the ferritin pores, associated with localized unfolding, suggests that the structure around the ferritin pores is a target for regulated protein unfolding and iron release.     

Three-dimensional structure of phenylalanyl-transfer RNA synthetase from Thermus thermophilus HB8 at 0.6-nm resolution.

The three-dimensional structure of the heterodimeric alpha 2 beta 2 enzyme phenylalanyl-tRNA synthetase from Thermus thermophilus HB8 has been determined by X-ray crystallography, using the multiple-isomorphous-replacement method at 0.6 nm resolution. Trigonal crystals of space group P3(2)21 have cell dimensions a = b = 17.6 nm and c = 14.2 nm. Assuming one heterodimeric molecule/asymmetric unit, the ratio of unit cell volume/molecular mass was V = 0.00244 nm3/Da, which is in the middle of the range normally observed. However, after a rotation-function calculation and measurement of the density of the native crystals, we postulate the existence of only the alpha beta dimer in the asymmetric units. This implies 73% solvent content in the unit cell. Three heavy-atom derivatives [K2PtCl4, KAu(CN)2 and Hg(CH3COO)2] and the solvent-flattening procedure were used for electron-density-map calculations. This map confirmed our hypothesis and revealed a remarkably large space filled by solvent, with alpha beta dimer only in the asymmetric unit. The phenylalanyl-tRNA synthetase from T. thermophilus molecule has a 'quasi-linear' subunit organization. As can be concluded at this level of resolution, there is no contact between small alpha subunits in the functional heterodimer.     

Iron mobilization from cultured rat bone marrow macrophages

The reticuloendothelial system is responsible for removing old and damaged erythrocytes from the circulation, allowing iron to return to bone marrow for hemoglobin synthesis. Cultured bone marrow macrophages were loaded with ⁵⁹Fe-labelled erythroblasts and iron mobilization was studied. After erythroblast digestion, iron taken up by macrophages was found in ferritin as well as in a low-molecular-weight fraction. The analysis of iron mobilization from macrophages shows: (1) the iron was mobilized as ferritin. (2) A higher mobilization was observed when apotransferrin was present in the culture medium. (3) In the presence of apotransferrin in the culture medium, part of the iron was found as transferrin iron. (4) Iron transfer from ferritin to apotransferrin was observed in a cell-free culture medium and this process was temperature independent. The results indicate that after phagocytosis of ⁵⁹Fe-labelled erythroblasts by macrophages, iron is mobilized as ferritin. In the plasma, this iron can be transferred to apotransferrin.     

Electron microscopy of alpha-latrotoxin from the venom of the black widow spider Latrodectus mactans tredecimguttatus]

Two-dimensional crystals of alpha-latrotoxin from the venom of black widow spider (Latrodectus mactans tredecimguttatus) were studied by the negative staining electron microscopy. Two-dimensional crystals were obtained by adsorption of the protein solution with a high Mg2+ concentration on carbon-coated electron microscopy grids. The crystals were about 0.4 mkm in size, had the unit cell parameters: a = b = 15.55 nm, gamma = 90 degrees, p4 plane group symmetry. The contour map of a stain-excluding region of such crystals was calculated by the Fourier-filtering procedure at about 4 nm resolution. The calculation of molecular weight of the unit cell, with the symmetry p4 taken into account, showed that alpha-latrotoxin particles, revealed by negative staining, consisted of 4 or 8 protomers.     

Purification and characterization of an iron-induced ferritin from soybean (Glycine max) cell suspensions.

Ferric citrate induces ferritin synthesis and accumulation in soybean (Glycine max) cell suspension cultures [Proudhon, Briat & Lescure (1989) Plant Physiol. 90, 586-590]. This iron-induced ferritin has been purified from cells grown for 72 h in the presence of either 100 microM- or 500 microM-ferric citrate. It has a molecular mass of about 600 kDa and is built up from a 28 kDa subunit which is recognized by antibodies raised against pea (Pisum sativum) seed ferritin and it has the same N-terminal sequence as this latter, except for residue number 3, which is alanine in pea seed ferritin instead of valine in iron-induced soybean cell ferritin. It contains an average of 1800 atoms of iron per molecule whatever the ferric citrate concentration used to induce its synthesis. It is shown that the presence of 100 microM- or 500 microM-ferric citrate in the culture medium leads respectively to an 11- and 28-fold increase in the total intracellular iron concentration and to a 30- and 60-fold increase in the ferritin concentration. However, the percentage of iron stored in the mineral core of ferritin remains constant whatever the ferric citrate concentration used and represents only 5-6% of cellular iron.     

Ferritin contains less iron (59Fe) in cells when the protein pores are unfolded by mutation

Ferric minerals in ferritins are protected from cytoplasmic reductants and Fe2+ release by the protein nanocage until iron need is signaled. Deletion of ferritin genes is lethal; two critical ferritin functions are concentrating iron and oxidant protection (consuming cytoplasmic iron and oxygen in the mineral). In solution, opening/closing (gating) of eight ferritin protein pores controls reactions between external reductant and the ferritin mineral; pore gating is altered by mutation, low heat, and physiological urea (1 mm) and monitored by CD spectroscopy, protein crystallography, and Fe2+ release rates. To study the effects of a ferritin pore gating mutation in living cells, we cloned/expressed human ferritin H and H L138P, homologous to the frog open pore model that was unexpressable in human cells. Human ferritin H L138P behaved like the open pore ferritin model in vitro as follows: (i) normal protein cage assembly and mineralization, (ii) increased iron release (t1/2) decreased 17-fold), and (iii) decreased alpha-helix (8%). Overexpression (> 4-fold), in HeLa cells, showed for ferritin H L138P equal protein expression and total cell 59Fe but increased chelatable iron, 16%, p < 0.01 (59Fe in the deferoxamine-containing medium), and decreased 59Fe in ferritin, 28%, p < 0.01, compared with wild type. The coincidence of decreased 59Fe in open pore ferritin with increased chelatable 59Fe in cells expressing the ferritin open pore mutation suggests that ferritin pore gating influences to the amount of iron (59Fe) in ferritin in vivo.     

Crystallization of the cpn60/cpn10 complex ('holo-chaperonin') from Thermus thermophilus.

A stable complex of the chaperonins, cpn60 and cpn10 (Escherichia coli GroEL and GroES homologues), from the extremely thermophilic bacterium Thermus thermophilus has been isolated and crystallized. The crystals have dimensions up to 30 x 200 x 200 microns. Ultra-thin sections of the crystals estimated by electron microscopy showed a rectangular lattice with unit cell parameters of a = 17 nm, b = 27 nm, gamma = 90 degrees.