Monomeric cobalt-oxygen complexes.

Some Schiff base complexes of cobalt(II) form stable cobalt-oxygen species in solution instead of the usual cobalt-oxygen-cobalt bridged complexes. These one-to-one adducts were isolated as crystalline solids and characterized by means of elemental analysis, magnetic properties, and molecular weight determinations.

Comparison of colloidal gold electrode fabrication methods: the preparation of a horseradish peroxidase enzyme electrode.

In order to prepare biosensing electrodes which respond to hydrogen peroxide, horseradish peroxidase has been adsorbed to colloidal gold sols and electrodes prepared by deposition of these enzyme-gold sols onto glassy carbon using three methods: evaporation, electrodeposition and electrolyte deposition. In the latter method the enzyme-gold sol is applied to the surface of a glassy carbon disk electrode followed by an equal volume of 2 mM CaCl2. The electrolyte causes the sol to precipitate on the electrode surface, producing an immobilized enzyme electrode. Satisfactory electrodes which gave an electrochemical response to hydrogen peroxide in the presence of the electron transfer mediator ferrocenecarboxylic acid were produced by all three methods. Evaporation of horseradish peroxidase-gold sols produced electrodes with the best reproducibility and the widest linear amperometric response range. These electrodes can also easily be stored in a dry state. Although not as good as evaporation, electrodeposition also produced satisfactory electrodes. Electro-deposition provides the added advantage that it lends itself to the preparation of multi-enzyme/multi-analyte electrodes by the adsorption of different enzymes to separate gold sols, followed by sequential electrodeposition onto discrete areas of a multichannel electrode.

A carrageenan hydrogel stabilized colloidal gold multi-enzyme biosensor electrode utilizing immobilized horseradish peroxidase and cholesterol oxidase/cholesterol esterase to detect cholesterol in serum and whole blood.

The preparation of two immobilized enzyme electrodes is described. One electrode contains horseradish peroxidase absorbed to colloidal gold and deposited on a glassy carbon electrode along with cholesterol oxidase entrapped in a carrageenan hydrogel. The second electrode also includes cholesterol esterase entrapped in the carrageenan. The incorporation of ferrocene or ferrocenecarboxylic acid mediator is brought about by either evaporation on the glassy carbon electrode or, in the latter case, entrapment in the carrageenan hydrogel. Amperometric signal generation results from the HRP catalyzed turnover of H2O2, a secondary product of the cholesterol oxidase catalyzed oxidation of cholesterol. Use of these enzyme electrodes makes cholesterol detection possible in human serum, low density lipoprotein, and whole blood.

Detection thresholds of potassium salts are related to the molar conductivity of the anion.

In a previous study, we found that human taste detection thresholds for Na+ salts were linearly correlated with molar conductivity values at infinite dilution of their anions. In the present study, detection threshold concentrations for potassium salts were also found to be linearly correlated (r = -0.92) with the molar conductivity of the anion of the salt. Detection thresholds were determined here for nine potassium salts with the same anions as the sodium salts previously tested. The mean detection thresholds for these potassium salts were found to be: K Acetate (0.00311 M), K Carbonate (0.00286 M), K Chloride (0.00242 M), K Citrate (0.000300 M), K Phosphate (0.00196 M), K Sulfate (0.00090 M), K Tartrate (0.00164 M), K Glutamate (0.00153 M), and K Ascorbate (0.00375 M). The rank order correlation between the detection threshold values for sodium and potassium salts was 0.88. This finding suggests that detection thresholds for both Na+ and K+ salts are determined by the charge mobility of the anion.

Mediator-free amperometric determination of toxic substances based on their inhibition of immobilized horseradish peroxidase.

We have demonstrated the feasibility of quantitatively detecting selected toxic materials through their inhibitory effect on an enzyme electrode that utilizes colloidal gold-immobilized horseradish peroxidase and does not require a mediator. Quantitative detection of azide, cyanide, thiourea, sulfide, and dichromate is demonstrated. The sensitivity and inhibition kinetics for this immobilized enzyme electrode are found to be different from those observed previously for homogeneous horseradish peroxidase. Possible reasons for this difference are discussed. Due to the availability of a large number of enzymes and their toxic inhibitors, this work based on immobilized enzyme inhibition coupled to an electrode surface significantly broadens the possible applications of biosensors and offers alternative methods for toxic substance determination.

Factors that influence siderophoremediated iron bioavailability: catalysis of interligand iron (III) transfer from ferrioxamine B to EDTA by hydroxamic acids.

Deferriferrioxamine B (H3DFB) is a linear trihydroxamic acid siderophore with molecular formula NH2(CH2)5[N(OH)C(O)(CH2)2C(O)NH(CH2)5]2N(OH)C(O)CH3 that forms a kinetically and thermodynamically stable complex with iron(III), ferrioxamine B. Under the conditions of our study (pH = 4.30, 25 degrees C), ferrioxamine B, Fe(HDFB)+, is hexacoordinated and the terminal amine group is protonated. Addition of simple hydroxamic acids, R1C(O)N(OH)R2 (R1 = CH3, R2 = H; R1 = C6H5, R2 = H; R1 = R2 = CH3), to an aqueous solution of ferrioxamine B at pH = 4.30, 25.0 degrees C, I = 2.0, results in the formation of ternary complexes Fe(H2DFB)A+ and Fe(H3DFB)A2+, and tris complexes FeA3, where A- represents the bidendate hydroxamate anion R1C(O)N(O)R2-. The addition of a molar excess of ethylenediaminetetraacetic acid (EDTA) to an aqueous solution of ferrioxamine B at pH 4.30 results in a slow exchange of iron(III) to eventually completely form Fe(EDTA)- and H4DFB+. The addition of a hydroxamic acid, HA, catalyzes the rate of this iron exchange reaction: (formula; see text) A four parallel path mechanism is proposed for reaction (1) in which catalysis occurs via transient formation of the ternary and tris complexes Fe(H2DFB) A+, Fe(H3DFB)A2+, and FeA3. Rate and equilibrium constants for the various reaction paths to products were obtained and the influence of hydroxamic acid structure on catalytic efficiency is discussed. The importance of a low energy pathway for iron dissociation from a siderophore complex in influencing microbial iron bio-availability is discussed. The system represented by reaction (1) is proposed as a possible model for in vivo catalyzed release of iron from its siderophore complex at the cell wall or interior, where EDTA represents the intracellular storage depot or membrane-bound carrier and HA represents a low molecular weight hydroxamate-based metabolite capable of catalyzing interligand iron exchange.

Sulfate as a synergistic anion facilitating iron binding by the bacterial transferrin FbpA: the origins and effects of anion promiscuity.

The ferric binding protein, FbpA, has been demonstrated to facilitate the transport of naked Fe3+ across the periplasmic space of several Gram-negative bacteria. The sequestration of iron by FbpA is facilitated by the presence of a synergistic anion, such as phosphate or sulfate. Here we report the sequestration of Fe3+ by FbpA in the presence of sulfate, at an assumed periplasmic pH of 6.5 to form FeFbpA-SO4 with K'(eff) = 1.7 x 10(16) M(-1) (at 20 degrees C, 50 mM MES, 200 mM KCl). The iron affinity of the FeFbpA-SO4 protein assembly is 2 orders of magnitude lower than when bound with phosphate and is the lowest of any of the FeFbpA-X assemblies yet reported. Iron reduction at the cytosolic membrane receptor may be an essential aspect of the periplasmic iron-transport process, and with an E(1/2) of -158 mV (NHE), FeFbpA-SO4 is the most easily reduced of all FeFbpA-X assemblies yet studied. The variation of FeFbpA-X assembly stability (K'(eff)) and ease of reduction (E(1/2)) with differing synergistic anions X(n-) are correlated over a range of 14 kJ, suggesting that the variations in redox potentials are due to stabilization of Fe3+ in FeFbpA-X by X(n-). Anion promiscuity of FbpA in the diverse composition of the periplasmic space is illustrated by the ex vivo exchange kinetics of FeFbpA-SO4 with phosphate and arsenate, where first-order kinetics with respect to FeFbpA-SO4 (k = 30 s(-1)) are observed at pH 6.5, independent of entering anion concentration and identity. Anion lability and influence on the iron affinity and reduction potential for FeFbpA-X support the hypothesis that synergistic anion exchange may be an important regulator in iron delivery to the cytosol. This structural and thermodynamic analysis of anion binding in FeFbpA-X provides additional insight into anion promiscuity and importance.

Multiple-path dissociation mechanism for mono- and dinuclear tris(hydroxamato)iron(III) complexes with dihydroxamic acid ligands in aqueous solution.

Linear synthetic dihydroxamic acids ([CH3N(OH)C=O)]2(CH2)n; H2Ln) with short (n = 2) and long (n = 8) hydrocarbon-connecting chains form mono- and dinuclear complexes with Fe(III) in aqueous solution. At conditions where the formation of Fe2(Ln)3 is favored, complexes with each of the two ligand systems undergo [H+]-induced ligand dissociation processes via multiple sequential and parallel paths, some of which are common and some of which are different for the two ligands. The pH jump induced ligand dissociation proceeds in two major stages (I and II) where each stage is shown to be comprised of multiple components (Ix, where x = 1-3 for L2 and L8, and IIy, where y = 1-3 for L2 and y = 1-4 for L8). A reaction scheme consistent with kinetic and independent ESI-MS data is proposed that includes the tris-chelated complexes (coordinated H2O omitted for clarity) (Fe2(Ln)3, Fe2(L2)2(L2H)2, Fe(LnH)3, Fe(L8)(L8H)), bis-chelated complexes (Fe2(Ln)2(2+), Fe(LnH)2+, Fe(L8)+), and monochelated complexes (Fe(LnH)2+). Analysis of kinetic data for ligand dissociation from Fe2(Ln)(LnH)3+ (n = 2, 4, 6, 8) allows us to estimate the dielectric constant at the reactive dinuclear Fe(III) site. The existence of multiple ligand dissociation paths for the dihydroxamic acid complexes of Fe(III) is a feature that distinguishes these systems from their bidentate monohydroxamic acid and hexadentate trihydroxamic acid counterparts and may be a reason for the biosynthesis of dihydroxamic acid siderophores, despite higher environmental molar concentrations necessary to completely chelate Fe(III).

The effect of an applied potential on the activity of carbonic anhydrase immobilized on graphite rods.

The effect of both a positive and a negative applied potential on the p-NPA hydrolysis activity of bovine carbonic anhydrase (BCA) immobilized on graphite rods has been investigated. Background experiments show that the pH-activity profile for BCA free in solution is not affected by either a negative or a positive potential applied to graphite rods placed in the same solution. However, the activity of BCA immobilized by covalent attachment to a graphite rod is influenced by a potential externally applied to the graphite rod. An overall increase in activity (as determined by the initial rate of the p-NPA hydrolysis reaction) is observed in the presence of a -0.2 V (Ag/AgCl) applied potential, while decreased activity is evident at +0.6 V (Ag/AgCl). This is indicative of an electrolyte anion effect rather than a local pH effect. In the presence of the specific anion inhibitors Cl(-) and SCN(-), the relative BCA activity increases at -0.2 V (Ag/AgCl) and decreases at +0.6 V (Ag/AgCl) are consistent with the different BCA inhibition constants for Cl(-) and SCN(-). Accelerated loss of immobilized BCA activity also accompanies the application of the external potentials, particularly at +0.6 V (Ag/AgCl). Results described here represent an early example of potentiostatic control of nonredox enzyme activity. Several possible mechanisms are discussed including specific anion inhibition, enzyme surface charge/charged support material interactions, and charged product inhibition. It is likely that a combination of such mechanisms is operational in this system. The implications of external potentials affecting the activity of immobilized enzymes in the design of stable immobilized enzyme electrodes are also discussed.

Treatment of iron deficiency anemia: are monomeric iron compounds suitable for parenteral administration?

Iron deficiency is the most common nutritional problem worldwide, especially in the developing countries. Oral iron supplementation programs have failed because of noncompliance and gastrointestinal toxicity, thereby necessitating parenteral administration of iron. For parenteral administration, only iron-carbohydrate complexes are currently used, because monomeric iron salts release free iron, thereby causing oxidant injury. However, iron-carbohydrate complexes also have significant toxicity, and they are expensive. We have proposed the hypothesis that monomeric iron salts can be safely administered by the parenteral route if iron is tightly complexed to the ligand, thereby causing clinically insignificant release of free iron, and the kinetic properties of the compound allow rapid transfer of iron to plasma transferrin. A detailed analysis of the physicochemical and kinetic properties reveals that ferric iron complexed to pyrophosphate or acetohydroxamate anions may be suitable for parenteral administration. We have demonstrated that infusion of ferric pyrophosphate into the circulation via the dialysate is safe and effective in maintaining iron balance in patients undergoing maintenance hemodialysis. Parenteral administration of monomeric iron compounds is a promising approach to the treatment of iron deficiency in the general population and merits further investigation.

Kinetics and mechanism of a catalytic chloride ion effect on the dissociation of model siderophore hydroxamate-iron(III) complexes.

Proton-driven ligand dissociation kinetics in the presence of chloride, bromide, and nitrate ions have been investigated for model siderophore complexes of Fe(III) with the mono- and dihydroxamic acid ligands R(1)C(=O)N(OH)R(2) (R(1) = CH(3), R(2) = H; R(1) = CH(3), R(2) = CH(3); R(1) = C(6)H(5), R(2) = H; R(1) = C(6)H(5), R(2) = C(6)H(5)) and CH(3)N(OH)C(=O)[CH(2)](n)C(=O)N(OH)CH(3) (H(2)L(n); n = 2, 4, 6). Significant rate acceleration in the presence of chloride ion is observed for ligand dissociation from the bis(hydroxamate)- and mono(hydroxamate)-bound complexes. Rate acceleration was also observed in the presence of bromide and nitrate ions but to a lesser extent. A mechanism for chloride ion catalysis of ligand dissociation is proposed which involves chloride ion dependent parallel paths with transient Cl(-) coordination to Fe(III). The labilizing effect of Cl(-) results in an increase in microscopic rate constants on the order of 10(2)-10(3). Second-order rate constants for the proton driven dissociation of dinuclear Fe(III) complexes formed with H(2)L(n)() were found to vary with Fe-Fe distance. An analysis of these data permits us to propose a reactive intermediate of the structure (H(2)O)(4)Fe(L(n)())Fe(HL(n))(Cl)(OH(2))(2+) for the chloride ion dependent ligand dissociation path. Environmental and biological implications of chloride ion enhancement of Fe(III)-ligand dissociation reactions are presented.

Microcalorimetric determination of thermodynamic parameters for ionophore-siderophore host-guest complex formation.

Thermodynamic parameters (delta H, delta S, and delta G) were determined by microcalorimetry in wet chloroform for host-guest assembly formation involving second-sphere complexation of the siderophore ferrioxamine B by crown ether (18-crown-6, cis-dicyclohexano-18-crown-6, benzo-18-crown-6) and cryptand (2.2.2 cryptand) hosts. Similar data were also collected for the same hosts with the pentylammonium ion guest, which corresponds to the pendant pentylamine side chain of ferroxamine B. Host-guest assembly formation constants (Ka) obtained from microcalorimetry agree with values obtained indirectly from chloroform/water extraction studies in those cases where comparable data are available. On the basis of a trend established by the pentylammonium guest, an enhanced stability relative to the crown ethers is observed for the assembly composed of ferrioxamine B and 2.2.2 cryptand that is due to entropic effects. Trends in delta H and delta S with changes in host and guest structure are discussed and attributed directly to host-guest complex formation, as solvation effects were determined to be insignificant (delta Cp = 0).

Hypothesis: iron chelation plays a vital role in neutrophilic inflammation.

Neutrophil influx into tissues occurs in many diverse diseases and can be associated with both beneficial and injurious effects. We hypothesize that the stimulus for certain neutrophilic inflammatory responses can be reduced to a series of competing reactions for iron, with either a labile or reactive coordination site available, between host chelators and chelators not indigenous to that specific living system. The iron focuses the transport of host phagocytic cells through a metal catalyzed generation of oxidant sensitive mediators including cytokines and eicosanoids. Many of these products are chemotactic for neutrophils. We also postulate that the iron increases the activity of the phagocyte associated NADPH oxidoreductase in the neutrophil. The function of this enzyme is likely to be the generation of superoxide in the host's attempt to chemically reduce and dislodge the iron from its chelate complex. After the reoxidation of Fe2+ in an aerobic environment, Fe3+ will be coordinated by host lactoferrin released by the neutrophil. When complexed by this glycoprotein, the metal does not readily undergo oxidation/reduction and is safely transported to the macrophages of the reticuloendothelial system where it is stored in ferritin. Finally, we propose that the neutrophil will attempt to destroy the chelator not indigenous to the host by releasing granular contents other than lactoferrin. Inability to eliminate the chelator allows this sequence to repeat itself, which can lead to tissue injury. Such persistence of a metal chelate in the host may be associated with biomineralization, fibrosis, and cancer.

New insights into the proton-dependent oxygen affinity of Root effect haemoglobins.

A long-standing puzzle with regard to protein structure/function relationships is the proton-dependent modification of haemoglobin (Hb) structure that causes oxygen to be unloaded from Root effect Hbs into the swim bladders and eyes of fish even against high oxygen pressure gradients. Although oxygen unloading in Root effect Hbs has generally been attributed to proton-dependent stabilization of the T-state, protonation of Root effect Hbs can alter their ligand affinities in both R- and T-state conformations and either stabilize the T-state or destabilize the R-state. The C-terminal residues that are so important in the Bohr effect of human Hb appear to be involved in the Root effects of some fish Hbs and not in others, indicating that several evolutionary pathways have resulted in expression of highly pH-dependent Hbs. New data are presented that show surprising similarities in the pH- and anion-dependence of sulfhydryl group reactivity and anaerobic oxidation of human and fish Hbs. The available evidence supports the concept that in both Bohr effect and Root effect Hbs a large steric component acts in addition to quaternary shifts between R and T conformations to regulate ligand affinity. Allosteric effectors moderate these steric effects within both R- and T-state conformations and allow for an elegant match between Hb function and the wide-ranging physiological needs of diverse organisms.

Crystal structure of ferrioxamine B: a comparative analysis and implications for molecular recognition.

Ferrioxamine B was successfully co-crystallized with ethanolpentaaquomagnesium(II) and perchlorate ions as counter ions, C27H62Cl3FeMgN6O26, and the crystal structure has been determined by single-crystal X-ray diffraction. The crystals are monoclinic, space group P2(1)/n, four molecules per unit cell with dimensions a=21.1945(7) A, b=10.0034(3) A, c=106.560(1) A, and beta=106.560(1) degrees. The crystal structure contains a racemic mixture of Lambda-N-cis,cis and Delta-N-cis,cis coordination isomers. The structural parameters and the conformational features of ferrioxamine B compare very well with those of ferrioxamines D1 and E, with an exception of the orientation of the pendant protonated amine, which is pointing away from the connecting amide chains and towards the carbonyl face of the inner coordination shell distorted octahedron. This pendant protonated amine, in conjunction with the carbonyl face of the Fe(III) coordination shell, is proposed to play an important role in the recognition and membrane transport processes.

A spectroelectrochemical method for differentiation of steric and electronic effects in hemoglobins and myoglobins.

Spectroelectrochemical techniques are described which enable us to compare anion effects on redox curves of structurally distinct hemoglobins with oxygenation curves obtained under equivalent conditions. Nernst plots for tetrameric vertebrate Hbs show evidence of cooperativity, with the T state conformation more resistant to oxidation than the R state. Anions shift the conformation toward the T state and decrease the ease of oxidation, with variations in anion sensitivity similar to those observed in oxygen equilibria. Oxygen binding, unlike electron exchange, is known to be subject to steric constraints that vary considerably in natural and engineered hemoglobins that have differences in the distal residues of the heme pocket. Since oxidation curves are not subject to steric hindrance, anion-induced differences between the oxidation and oxygenation curves can be indicative of anion-induced alterations in the stereochemistry of the heme pocket that alters the ease of ligand entry or exit. Addition of inositol hexaphosphate to solutions of Hb A in 0.2 M nitrate generates such differences: the ease of electron abstraction from deoxy (T state) Hb A is unaffected, while, as previously reported, the oxygenation of deoxy (T state) Hb A is greatly hindered. The difference between inositol hexaphosphate effects on initial stages of oxidation and oxygenation indicates that the explanation for "multiple T states" in oxygen binding lies in the ability of the polyanion to greatly increase steric hindrance to ligand entry, without appreciable changes in the electronic features of the heme environment.

Spectroelectrochemistry of heme proteins: effects of active-site heterogeneity on Nernst plots.

In order to detect and model the effect of functional chain heterogeneity on Nernst plots for heme proteins, we examined the redox properties of various myoglobins (Mbs) and their mixtures using an improved spectroelectrochemical method. Specific redox responses and formal half potentials (E1/2) were obtained for Aplysia, horse, and sperm whale Mbs, as well as 1:1 mixtures of Mbs consisting of Aplysia/sperm whale, sperm whale/horse, and horse/Aplysia. Linear Nernst plots with slopes near unity were observed for horse, sperm whale, and Aplysia Mbs, with E1/2 values of 14, 19, and 96 mV (vs. NHE) respectively, consistent with previous reports using indirect methods. The Nernst plot responses for mixtures of some of these Mbs allowed us to evaluate and model the non-Nernstian behavior that results from intrinsically different values of E1/2 and from incomplete spectral overlap. The data demonstrate that increasing the E1/2 differences between the components of a Mb mixture increases the changes in shape of the resulting Nernst plots, the dominant effect being a decrease in the observed Nernst coefficient (nNernst). Comparison of Nernst plots for redox data with Hill plots for O2 binding data shows that the redox process is more affected by the structural differences in the distal heme pockets of the Mbs studied than is O2 binding. Similar effects of chain heterogeneity may give rise to disproportionate reductions in the slopes of Nernst and Hill plots for hemoglobins (Hbs). This possibility is discussed in relation to Hbs investigated for redox and O2 binding activity in our laboratories where we find nNernst to be commonly less than nHill over a range of experimental conditions.

Preparation and activity of carbonic anhydrase immobilized on porous silica beads and graphite rods.

Techniques for the immobilization of bovine carbonic anhydrase (BCA) on porous silica beads and graphite are presented. Surface coverage on porous silica beads was found to be 1.5 x 10(-5) mmol BCA/m(2), and on graphite it was 1.7 x 10(-3) mmol BCA/m(2) nominal surface area. Greater than 97% (silica support) and 85% (graphite support) enzyme activity was maintained upon storage of the immobilized enzyme for 50 days in pH 8 buffer at 4 degrees C. After 500 days storage, the porous silica bead immobilized enzyme exhibited over 70% activity. Operational stability of the enzyme on silica at 23 degrees C and pH 8 was found to be 50% after 30 days. Catalytic activity expressed as an apparent second-order rate constant K'(Enz) for the hydrolysis of p-nitrophenyl acetate (p-NPA) catalyzed by BCA immobilized on silica beads and graphite at pH 8 and 25 degrees C is 2.6 x 10(2) and 5.6 x 10(2) M(-1)s(-1) respectively. The corresponding K(ENZ) value for the free enzyme is 9.1 x 10(2) M(-1)s(-1). Activity of the immobilized enzyme was found to vary with pH in such a manner that the active site pK, on the porous silica bead support is 6.75, and on graphite it is 7.41. Possible reasons for a microenvironmental influence on carbonic anhydrase pK(a), are discussed. Comparison with literature data shows that the enzyme surface coverage on silica beads reported here is superior to previously reported data on silica beads and polyacrylamide gels and is comparable to an organic matrix support. Shifts in BCA-active site pK(a) values with support material, a lack of pH dependent activity studies in the literature, and differing criteria for reporting enzyme activity complicate literature comparisons of activity; however, immobilized BCA reported here generally exhibits comparable or greater activity than previous reports for immobilized BCA.

Fe3+ coordination and redox properties of a bacterial transferrin.

The Fe(3+) binding site of recombinant nFbp, a ferric-binding protein found in the periplasmic space of pathogenic Neisseria, has been characterized by physicochemical techniques. An effective Fe(3+) binding constant in the presence of 350 microm phosphate at pH 6.5 and 25 degrees C was determined as 2.4 x 10(18) m(-1). EPR spectra for the recombinant Fe(3+)nFbp gave g' = 4.3 and 9 signals characteristic of high spin Fe(3+) in a strong ligand field of low (orthorhombic) symmetry. (31)P NMR experiments demonstrated the presence of bound phosphate in the holo form of nFbp and showed that phosphate can be dialyzed away in the absence of Fe(3+) in apo-nFbp. Finally, an uncorrected Fe(3+/2+) redox potential for Fe-nFbp was determined to be -290 mV (NHE) at pH 6.5, 20 degrees C. Whereas our findings show that nFbp and mammalian transferrin have similar Fe(3+) binding constants and EPR spectra, they differ greatly in their redox potentials. This has implications for the mechanism of Fe transport across the periplasmic space of Gram-negative bacteria.