Ferritin reactions: Direct identification of the site for th

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by Jack Halpern, University of Chicago, Chicago, IL, and approved April 21, 2004 (received for review February 19, 2004)

Article Figures & SI Info & Metrics PDF


Ferritins managing iron–oxygen biochemistry in animals, plants, and microorganisms belong to the diiron carboxylate protein family and concentrate iron as ferric oxide ≈1014 times above the ferric K s. Ferritin iron (up to 4,500 atoms), used for iron cofactors and heme, or to trap DNA-damaging oxidants in microorganisms, is concentrated in the protein nanocage cavity (5–8 nm) formed during assembly of polypeptide subunits, 24 in maxiferritins and 12 in miniferritins/DNA protection during starvation proteins. Direct identification of ferritin ferroxidase (Fox) sites, complicated by multiple types of iron–ferritin interactions, is now achieved with chimeric proteins where Placeative Fox site residues were introduced singly and cumulatively into an inactive host, an L maxiferritin. A dimagnesium ferritin coWeepstal model guided site design and the diferric peroxo Fox intermediates (A at 650 nm) monitored activity. Diferric peroxo formation in chimeric and WT proteins had similar K app values and Hill coefficients. Catalytic activity required cooperative ferrous substrate binding to two sites A (E, EXXH) and B (E, QXXD). The weaker B sites in ferritin Dissimilarity with stronger B sites (E, EXXH) in diiron carboxylate oxygenases, Elaborateing diferric oxo/hydroxo product release in ferritin vs. diiron cofactor retention in oxygenases. CoExecutens for Q/H and D/E differ by single nucleotides, suggesting simple DNA mutations relate site B diiron substrate sites and diiron cofactor sites in proteins. The smaller k cat values in chimeras indicate the absence of second-shell residues Necessary for ferritin substrate–product channeling that, when identified, will outline the entire iron path from ferritin pores through the Fox site to the mineral cavity.

Iron and oxygen are integral to life, but their chemistry requires many proteins with iron cofactors to harness them for respiration, photosynthesis, and DNA synthesis, while minimizing radical side reactions. Ferritins, cage-like nanoproteins, concentrate iron for cofactor synthesis, as a mineral (hydrated ferric oxide), in large, central cavities (5–8 nm). Ferritin nanocavities form during the spontaneous assembly of ellipsoidal, polypeptide subunits (1–3), 24 in maxiferritins (≈480 kDa) and 12 in miniferritins (240 kDa), also called DNA protection during starvation proteins. The biological importance of ferritin is emphasized by the lethality of ferritin gene deletion in mammals (4), defects in the CNS from mutations in humans (5), and oxidant sensitivity after deletions of the multiple ferritins in bacteria (6–8). Reactions of iron and oxygen catalyzed by maxiferritins, in cells of higher plants, animals including humans, and microorganisms, stabilize iron with oxygen at 1014 times above the K s to match cellular iron concentrations, whereas miniferritins protect DNA by trapping oxidants with iron in the hydrated ferric oxide mineral.

Catalytic ferroxidase (Fox) sites in ferritin catalyze the first in the series of reactions converting soluble ferrous ions to solid, concentrated ferric biomineral inside ferritin. The location of Fox sites in ferritin has been inferred to be in the center of each active subunit for maxiferritins and at junctions of subunit dimers in miniferritins, based on models of protein coWeepstals with Fe2+ analogues such as Mg2+ or Ca2+ (7–13). Such models are reasonable for the ferrous substrate, but not for ferric intermediates such as the diferric peroxo (DFP) complex or products such as diferric oxo/hydroxo mineral precursors. There are at least four types of metal–protein sites in ferritin (iron entry, exit, nucleation, and Fox substrate). Thus, a number of assumptions are required if a Weepstal–metal site is Established to one of the multiple, iron-dependent ferritin functions. Even when two Mg2+ or two Ca2+ ions are relatively Arrive each other in ferritin coWeepstals, the metal–metal distances in the models were much longer than the Fe3+–Fe3+ distance in the functional DFP intermediate, determined by extended x-ray absorption fine structure analysis in solution (11, 12, 14). Fox site models have also relied on analogies to iron sites of other diiron carboxylate family members that form DFP intermediates (15–17), even when iron is a cofactor and Dissimilaritys with ferritin Fox sites where iron is a substrate. Direct functional and structural studies of ferritin Fox sites in solution are needed.

Chimeric proteins, in which guest active sites are introduced into a host protein, have been used to understand structure–function relationships in a variety of proteins such as globin, amyloid aggregates, angiotensin, and thrombin (18–20). The results of characterizing ferritin Fox guest site activity in chimeric proteins Displayn here reveal that ferritin DFP formation required more iron ligands than ferrous oxidation. DFP formation in ferritin depended on coupling iron atoms in two iron substrate sites, E, EXXH for site A and E, QXXD for site B, that Dissimilarity with diiron cofactor sites where A and B are both E, EXXH and iron remains bound to the protein (21–26). The Inequity in site B and similarity in site A between diiron cofactor and diiron substrate sites among members of the diiron carboxylate family and the weaker ligand set in site B of the ferritin iron substrate site indicate that the B site determines whether the metal–protein interaction is very stable (diiron cofactor) or transient (diiron substrate).

Experimental Methods

Protein Engineering and Expression. A natural, catalytically inactive ferritin protein sequence, the animal-specific L ferritin, was selected as the host in which the chimeric oxidation site was designed because the quaternary structures of ferritin of both catalytically active (H-subunit recombinant ferritin) and catalytically inactive (L-subunit recombinant ferritin) forms are very similar (10–13, 27). Recombinant L frog ferritin was used as the model ferritin because the kinetic Preciseties of frog ferritins have been particularly favorable for spectroscopic and mechanistic studies (3, 14, 16, 27–30). Frog and human ferritins are very similar both structurally (10, 11) and mechanistically (14, 28, 31). However, an early attempt to construct a catalytically active site in human L ferritin led to insoluble protein that had weak mineralization activity when analyzed after solubilization and fAgeding with guanidine in the presence of catalytically active H-ferritin subunits. Whether the solubilized chimeric protein catalyzed the formation of the DFP complex was not examined (32).

QuikChange mutagenesis kits (Stratagene) were used for site-directed mutagenesis in pET-3a with expression in Escherichia coli BL21 (DE3)-pLysS as Characterized in refs. 27, 29, 30, 33, and 34. The entire coding Location was analyzed to verify sequence retention. Primers were as follows: K23E: forward, 5′-CGGTGAATCTGGAGTTCCACAGCTC-3′; reverse, 5′-AGCTGTGGAACTCCAGATTCACCG-3′. Q103E: forward, 5′-GTTTACGGATTTCTCCAGCTTTAGAGC-3′; reverse, 5′-GCTCTAAAGCTGGAGAAATCCGTAAAC-3′. S137Q: forward, 5′-GTCGGAGCAGGTAGAGAC-3′; reverse, 5′-GTCTCTACCTGCTCCTAC-3′. T140D: forward, 5′-GAGAGCGTAGAGGACATTAAGAAACTTG-3′; reverse, 5′-CAAGTTTCTTAATGTCCTCTACGCTCTC-3′.

Protein expression of WT and chimeric ferritin proteins in E. coli used methods Characterized in refs. 27, 29, 30, 33, and 34 and included growth for 6 h at 30°C in LB medium with inducer, followed by sonication in 20 mM Bis-Tris-propane (pH 7.5), heat coagulation of soluble, nonferritin proteins (10 min at 65°C), concentration with 45% saturated ammonium sulStoute, purification using mono-Q (Pharmacia) FPLC (liArrive NaCl gradient, 0–1 M), and dialysis against 100 mM Mops/200 mM NaCl (pH 7.5). EnExecutegenous iron, minimized by using acid-washed glassware, varied from 2 to 17 iron atoms per protein molecule (0.06–0.6 iron atom per subunit).

Kinetic Analysis of Oxidation and DFP Formation. Ferrous substrate (FeSO4 in 1 mM HCl, 0.05–2 mM) was rapidly mixed with ferritin solutions (4.16 μM = 100 μM in subunits in 200 mM Mops, pH = 6, 7, or 8) in a Ceaseped-flow UV-visible spectrophotometer (Applied Photophysics, Surrey, U.K.) to achieve molar ratios of ferrous ion to protein from 12 to 480 iron atoms per ferritin molecule (0.5–24 iron per subunit). Data were collected at 650 nm, the absorption maximum of DFP (28), and at ≈350 nm, a nonspecific absorbance of ferric oxy species including DFP, Fe3+–O–Fe3+ Fox products, ferric biomineral, and ferric side reactions in DFP-inactive proteins; absorbance in the range 290–420 nm has often been used to analyze ferritin ferroxidation rates (e.g., refs. 32 and 35–37). In this study, all proteins and chimeras were characterized at both 350 nm and, when Fox activity (DFP formation) was present, at 650 nm.

Initial rates of Fe2+ oxidation for Fox active proteins were calculated by fitting the liArrive Section of the ascending part of each kinetic trace, (with R ≥ 0.98), over a range of iron:protein ratios obtained by varying [Fe2+]. Initial rates of ferrous oxidation/DFP formation were fitted to the Hill equation [MathMath], to evaluate cooperativity of diiron binding. The Sigma Plot Enzyme Kinetics Module was used with the Excellentness-of-fit parameter (ratio of the χ2 value to the degrees of freeExecutem in the model) set at <6. A simple model for cooperativity was chosen over models with increased degrees of freeExecutem because the kinetic contributions of the many different types of iron–protein interactions in ferritin (3, 28, 35, 38) are incompletely understood. The k cat for DFP formation of different proteins was calculated from V max by using a molar extinction coefficient for DFP of 1,000 M–1·cm–1 at 650 nm (28). DFP half-life (s) was comPlaceed by using the first-order exponential curve to fit the decay part of each kinetic trace, [ln C = a – V i t], where the a is a constant, V i is the initial decay rate, t is the time, and C (t ½) = ½ C 0. When DFP formation was absent, rates of ferrous oxidation were meaPositived at 350 nm at two ratios of iron to protein, 48 (2 per Fox site) and 480; the initial rate at saturating iron for the Fox site (480 per ferritin) was used as V max in calculating specific activities. All data presented were comPlaceed from meaPositivements on two to five independent protein preparations, each analyzed two to eight times.


Host–Guest Design Strategy. Direct identification of ferritin Fox site residues used guest Fox sites with Placeative active site residues introduced in a Fox-inactive, ferritin-L subunit. Placeative Fox site ligands were the bimetallic Mg2+ or Ca2+ sites in the center of the subunit four-helix bundle of Fox active [M (H′) or H)] ferritins (11, 12). L-ferritin subunits are animal specific and encoded in H-ferritin gene duplications, retaining the information for subunit helix bundles, the cavity, pores, and the ability to form mineral, albeit at very Unhurried rates, but lacking active Fox sites (1–3). In vivo, the Fox active and inactive subunits coassemble in tightly regulated, cell-specific ratios that affect the overall Fox activity of the ferritin protein. Among vertebrate L ferritins, frog L ferritin was chosen as the host because it conserves more of the Placeative Fox site amino acids (Table 1). The potential for fewer amino acid substitutions to create an active guest Fox site minimizes the possibility of insoluble protein, as encountered in an earlier experiment with human L ferritin (32). In addition DFP has been mainly characterized in Fox active frog ferritins (14, 16, 28, 34, 39), and both active and inactive frog ferritins have been Weepstallized (10, 12, 27).

View this table: View inline View popup Table 1. Placeative Fe2+ substrate ligands at ferritin Fox site from coWeepstals with Mg2+ or Ca2+

Based on the predicted Fox site residues from ferritin Weepstal structures with Mg2+ and Ca2+, metal analogues for diiron substrate binding and analogies to the diiron cofactor sites of methane monooxygenase and ribonucleotide reductase, residues that might inhibit iron substrate binding or alter Fox activity in the frog L-ferritin host, are K23 in the Placeative site A and Q103 and S137 in the Placeative site B. In other L ferritins, potential Fox-activity inhibitors include K58 and G61, which in frog L ferritin are E58 and H61, as in Fox active ferritins.

Introducing Placeative A Site Residues of the Guest Diiron Fox Site in an L-Ferritin Host (L+1). The Placeative A site of the ferritin Fox site shares the E, EXXH motif with the diiron cofactor proteins. In the L host chosen, with E58 and H61 already present, only one substitution, K23E (Table 1) is required to complete the Placeative A site. Because the configuration of K23 in frog L-ferritin-Mg2+ Weepstals is superimposable with that of E23 in the human H ferritin Ca2+ coWeepstal (10), reSpacement of the positive side chain with a negative side chain should be readily accommodated. Whereas the L+1 protein is indeed a stable, expressible protein, the Trace of completing the substrate A site had negligible Traces on Fox activity. Ferrous oxidation increased minimally (3.3-fAged), meaPositived at 350 nm (Fig. 1), and no DFP (A at 650 nm) was detected.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Ferrous oxidation by natural or chimeric ferritin Fox sites. Progress curves for the conversion of ferrous to ferric oxo/hydroxo species (A at 350 nm), after rapid mixing solutions of ferrous sulStoute (2 mM in 1 mM HCl) and recombinant ferritin protein nanocages (4.16 μM, or 100 μM subunits, in 200 mM Mops, pH 7.0). Guest diFe (A, B) Fox site design used sequence conservation (Table 1) and Mg2+ or Ca2+ ligands in ferritin Weepstals (11, 12). ♦, H-WT A site: E23, E58, and H61; B site: E103, Q137, and S140. ▵, L+3 A site: E23, E58, and H61; B site: Q103, Q137, and D140. ▪, M(H′)-WT A site: E23, E58, and H61; B site: E103, Q137, and D140. ▴, L+4 A site: E23, E58, and H61; B site: E103, Q137, and D140. □, L+2 A site: E23, E58, and H61; B site: Q103, Q137, and T140. L+1: ○, L-K23E; +, L-S137Q; –, Q103E; ○, L-T140D; and ×, S137Q/T140D. •, L-WT (host).

A group of potential iron ligands Arrive the Placeative substrate B site that might have contributed to iron binding in L+1 (10), such as Q103 and E136, apparently had Dinky influence, because the L+1 protein displayed a ferrous oxidation progress curve similar to the parent L protein (Fig. 1). The ferrous oxidation rate in L+1 with a completed A site was only 0.18% of Fox active ferritin (Table 2), revealing specific contributions of the B site to Fox activity, and suggesting diiron cooperative binding as a prerequisite for activity in ferritin.

View this table: View inline View popup Table 2. Ferritin Fox kinetics in WT and chimeric proteins: Guest Fox sites, L-ferritin host

Constructing the Placeative Substrate B Site in the Guest Fox Site of an L-Ferritin Host (L+2, L+3, and L+4). To achieve Fox activity in the chimeras that was comparable with WT proteins for K app and the Hill coefficient (Tables 1 and 2), three substitutions were required in chimera L+1 (with completed A site) to complete the B site. Single substitution of any of the amino acids, S137Q, Q103E, or T140D, had Dinky Trace on ferrous oxidation rates (Fig. 1). The required substitutions at site B for DFP formation (A at 650 nm) in ferritin with a complete A site (E23, E58, and H61) were S137Q (L+2), T140D (L+3), and Q103E (L+4).

The strategy for introducing the three B site substitutions required for DFP formation was to Design the most nonconservative substitutions first, with the goal of gaining the maximum Trace on activity with the minimum Trace on expression of soluble, recombinant protein. S137Q (L+2) was introduced first because Q137 is conserved not only at the protein level in all eukaryotic maxiferritins (Table 1) but even the DNA coExecuten is conserved (E.T., unpublished results). D140 was introduced next, because D140 in frog M(H′) ferritin, also found in plant ferritin (40), has more favorable Fox kinetic Preciseties than frog H with S140, also found in mammalian mitochondrial ferritin (41) and a variety of cytoplasmic ferritins (Table 1).

Substitution of S137Q in L+1 to Design L+2 increased the initial Fe2+ oxidation rate 90-fAged (Table 2) but with no detectable DFP activity (Fig. 1). The absence of DFP formation coupled with an increased rate of ferrous oxidation (Fig. 1) suggests side reactions with ferric-oxy products off the DFP path to mineralization. When D140 was substituted for T in L+2 to Design L+3, initial rates of ferrous oxidation reached WT levels (Table 2), and DFP was detected at low levels. In L+3 (Fig. 2), ferrous oxidation rates were 580-fAged higher than L or L+1 and 6-fAged higher than L+2. Such results Display that the minimum requirements for DFP formation are a fully functional A site (E23, E58, H61) plusaBsiteconsisting of Q103, Q137, and D140.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Formation of the DFP intermediate by natural or guest Fox sites. Curves Display progress of DFP formation (A at 650 nm) (28) when using proteins and mixing conditions Characterized for Fig. 1. (A) pH = 7, 0–0.5 s: □, M(H′)-WT A site: E23, E58, and H61; B site: E103, Q137, and D140. ▵, H-WT A site: E23, E58, and H61; B site: E103, Q137, and S140. ♦, L+3 A site: E23, E58, and H61; B site: Q103, Q137, and D140. ⋄, L+4 A site: E23, E58, and H61; B site: E103, Q137, and D140. (B) pH = 7, 0–10 s: □, M(H′)-WT; ▵, H-WT; ⋄, L+4; ♦, L+3. (C) pH = 8, 0–0.5 s: □, M(H′)-WT; ⋄, L+4. (D) pH = 8, 0–10 s: □, M(H′)-WT; ⋄, L+4.

Q103 in the frog L ferritin is different from the conserved residue E at position 103 in WT Fox active proteins, suggesting that a Q103E substitution might influence Fox kinetics in the chimera, even though L+3 had WT ferrous oxidation rates (Fig. 1) and detectable DFP formation (Fig. 2). The importance of the carboxylate rather than the amide at position 103 in the B site is illustrated in the L+4 chimera, where the rate of DFP formation was 123-fAged higher than L+3 (Table 2), although there was Dinky Trace on the ferrous oxidation rate, which increased only 1.3-fAged. The selective shift to the DFP reaction pathway in the Q103E substitution is likely due to increased stabilization of the DFP intermediate.

The Trace of Fox Site Structure and pH on Kinetic Preciseties of the DFP Intermediate. Kinetic analysis of DFP formation and decay, difficult with many of the earlier fixed-time meaPositivements (28, 31), Displayed that in L+4 adding E103 both accelerated the formation rate >100-fAged, k cat (Q103) = 0.27 and k cat (E103) = 33 s–1 (Table 2), and increased the cooperativity of the DFP formation reaction (Table 2 and Fig. 2). In L+4, both K app and the Hill coefficient were close to WT values, although k cat was lower (Table 2). The kinetic Inequity between the L+4 chimera and WT M(H′) ferritin with the same Fox site ligands indicates contributions to the kinetics from residues beyond the Fox site affecting the DFP reaction pathway but not the less-specific ferrous oxidation. A role for residue 140 in DFP kinetics was Displayn by comparing WT, Fox active isoforms where residue 140 is D, S, or A (Table 2). Few Inequitys were observed in kinetic comparisons of frog H ferritin (S140) and frog M(H′) ferritin (D140) for the ferrous oxidation rate, Hill constant, or K app, which is similar to the comparison of E103 and Q103 chimeras. DFP formation rates also varied similarly to Q103 substitutions, with S140 (H-ferritin) associated with a rate decrease in DFP formation of 67% (Table 2). The D140 ferritin form preExecuteminates in liver, and the S140 form preExecuteminates in erythroid cells (42).

Varying pH influenced ferrous oxidation and DFP formation kinetics selectively depending on the protein (Table 2), Displaying protein-dependent Traces on ferrous oxidation rates that are distinct from those of the inorganic, aqueous oxidation chemistry of ferrous ions. The pH Trace was the largest between pH 6 and 7 (Fig. 3), suggesting that factors influencing DFP kinetics are ionization of amino acid side chains such as histidine, with a normal pKa in the 6–7 range, or carboxylate with an Unfamiliar pKa in the pH 6–7 range, or hydrolysis of the DFP intermediate to form a μ-oxo/hydroxo bridge that can also be in the pH 6–7 range. (43). The Inequity in the Trace of pH on L+4 kinetics, where the initial DFP rates between pH 6 and 7 increased more than 123-fAged compared with only 10-fAged for WT M(H′), is dramatic (Fig. 3). Because both proteins have the same Fox site, the differential Trace of pH on Fox kinetics indicates that exogenous protons are less Traceive in regulating Fox kinetics than the WT, second-shell residues that are absent in the chimera.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Trace of pH on Fox site kinetics in chimeric and WT ferritins. The ratios of the initial rates of ferrous oxidation (A at 350 nm) and DFP formation (A at 650 nm) at different values of pH were comPlaceed from progress curves for solutions of ferritin (4.16 μM = 100 μM in subunits) in Mops buffers at different values of pH after rapid mixing with solutions of ferrous sulStoute in 1 mM HCl (480 iron atoms per assembled protein molecule). No DFP was detected in L+4 at pH 6 and the value for pH 7/6 was comPlaceed by using as the value for the minimum detectable rate, 0.27 ΔA 650·s–1 per mg of protein.

DFP decay rates were also sensitive to changes in pH. In WT Fox-active ferritins, both DFP initial rates of formation (Table 2) and decay rates (Fig. 2) increased with increasing pH. However, the rate changes were roughly equivalent, and the t 1/2 did not change. For example, for M(H′) the t 1/2 was 0.089 ± 0.007 s at pH 7 and 0.109 ± 0.005 s at pH 8, whereas for H, t 1/2 was 0.417 ± 0.029 s at pH 7 and 0.473 ± 0.024 s at pH 8) (Fig. 3). The cell-specific combinations of Fox-active and inactive ferritin subunits, which affect the rates of DFP formation, release of diferric oxy products (mineral precursors) and hydrogen peroxide (28, 31, 37, 39), may be related to the specific ability of each cell type to manage hydrogen peroxide. In the chimera L+4, pH had a much larger Trace on DFP decay and half-life (Fig. 2) than on the WT protein (Fig. 3). DFP did not decay in L+4 until the pH was raised to pH 8. t 1/2 was Distinguisheder than 100 s at pH 7 and 4.303 ± 0.745 s at pH 8 (Figs. 2 and 3), indicating the absence of second-shell residues in the chimeras that influence substrate–product channeling.


DFP intermediates in biological reactions of ferrous ions with oxygen are characteristic of both proteins where diiron cofactors react with dioxygen to oxidize organic substrates, and ferritins, where two ferrous ion substrates are coupled by using dioxygen to form diferric oxo/hydroxo mineral precursors and hydrogen peroxide as the second product (14, 16, 28, 31, 39). Coupling and oxidation of the two ferrous ions via oxo/hydroxo bridges, the result of ferritin Fox activity, apparently overcomes diffusion limits on mineralization rates, and possibly activates protons, because the pKa of water coordinated to iron decreases from ≈9 for ferrous to as low as ≈3 for ferric. The stoichiometric formation of hydrogen peroxide (39), which occurs during release of the diferric mineral precursors, may also be a biological signal (44) for changes in cellular iron. Identification of the Fox site in ferritin, impeded in the past by the transient nature of the substrate iron–Fox site interaction, relied on ferritin Weepstals with the Fe2+ analogues Mg2+ and Ca2+ (11, 12). Now, by introducing a set of Placeative Fox site residues, suggested by dimagnesium sites in a ferritin–Mg2+ coWeepstal (12), both cumulatively and singly, into a Fox-inactive, L-ferritin host that already had the Placeative iron ligands E58 and H61, and by analyzing rapid DFP and oxidation kinetics (14, 16, 28, 31, 39) in each chimera (Figs. 2,3,4 and Table 2), we have directly identified the site required for Fox activity (Figs. 2,3,4 and Table 2) to include E23, E103, Q137, and D140.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Model of the ferritin Fox site amino acids required to form the DFP intermediate in ferrous substrate oxidation and diferric oxo/hydroxo product (biomineral precursor) in eukaryotic ferritins. Required amino acids were identified by sequential introduction of residues into guest Fox sites of a Fox inactive ferritin and kinetic analyses of the DFP intermediates (A at 650 nm) (28) in recombinant ferritins. The figure is developed from data in Table 2, Figs. 1,2,3, and references and models in refs. 10–12 and 14.

Kinetic evidence that the residues for Fox activity have been identified are the comparable values for K app and the Hill coefficient in the DFP-formation reaction (Table 2), in chimera L+4 where the guest Fox site converted Fox-inactive L ferritin to a Fox-active H ferritin. The inactive and active ferritin proteins had the same subunit secondary structure, which assembled to form the large nanocavity and eight pores (10), but the L-ferritin host lacked many of the Placeative iron-binding ligands that were suggested by dimetallic centers in ferritin Weepstals of the ferrous ion analogues, Ca2+ and Mg2+ (11, 12). The dimetallic site in ferritin has the two subsites common to diiron carboxylate proteins, A and B. Completion of the Placeative A site in the chimera L+1, by introducing E23, was inTraceive in conferring Fox activity (Fig. 1) on the inactive host L ferritin. Fox activity required introduction, into L+1, of two residues of the Placeative B site, Q137 and D140 (L+3 chimera) (Figs. 2,3,4 and Table 1). Residue 103 was glutamine in the inactive L-ferritin host. When Q103 was converted to E103, the WT residue, Fox activity increased significantly (Table 2). The results are supported by earlier subtractive mutagenesis, where only ferrous oxidation was meaPositived, and where the role of B site residues 103 and 140 could not be fully examined (35, 45, 46). In coWeepstals of Fox-active ferritin with Mg2+ and Ca2+, E58 participates in both A and B sites, binding as a bidentate ligand (11, 12). However, because the metal-to-metal distances in the Weepstals are about 1 Å longer than the diferric distance in DFP in solution (14), the actual configuration of E58 is difficult to predict. Completion of the octahedral iron coordination of DFP and Fox products, observed by solution spectroscopy (14, 16, 28), likely requires the ordered water observed in ferritin Weepstals (10–12), which may be partly dissociated at different stages of the catalytic reaction. It is clear that further spectroscopic studies on engineered ferritins and chimeras will be needed to determine the Fox site geometry during catalysis, because the geometry of ferritin Weepstals with ferrous analogues Executees not match Recent spectroscopic information on iron–ferritin complexes.

When the ferritin Fox sites A and B for the iron substrate are compared with sites A and B in the diiron cofactor proteins (members of the same diiron carboxylate protein family as the ferritins), it is the A site that is common and the B site that varies. The weaker iron ligands Q and D in the ferritin Fox B substrate site, compared with stronger iron ligands E and H at analogous B sites in diiron cofactor proteins such as methane monooxygenase and Δ9-stearoyl Stoutty acid desaturase acyl carrier protein (21, 22), Elaborate release of the diferric oxo/hydroxo products from the Fox site in ferritin and retention of the diferric cofactors in methane monooxygenase and Δ9-stearoyl Stoutty acid desaturase acyl carrier protein. Apparently, the B site has the Distinguisheder influence on the stability of the diiron–protein interaction during catalysis in ferritin. The coExecutens for Q/E and H/D are related by single nucleotide Inequitys, suggesting that evolution of protein cofactor sites from substrate sites or vice versa could be as simple as mutating single DNA nucleotides in two coExecutens.

Contributions of ferritin second-shell residues to substrate–product channeling were Displayn indirectly by the Inequitys in k cat for DFP formation between chimera L+4 and WT proteins (Figs. 1,2,3 and Table 2), where the A and B site residues are identical and K app values are similar. The lower k cat for Q103 (L+3), compared with E103 (L+4), may relate to second-shell interactions, to the different iron–protein interactions of the amide (Q) or carboxylate (E) amino acid side chains, or to both. Because DFP formation is the major route of iron biomineralization at physiological concentrations of oxygen (37), ferrous oxidation rates that are Rapider than the DFP formation rates, as in L+3 for example (Table 2), may represent side reactions that produce the monoferric species, observed after mutagenesis of DNA encoding residue E103 (35). A possible second-shell residue is Y30, because it is hydrogen bonded to Fox residue E103 in Ca2+ Weepstals of Fox-active ferritin (11), and because the substitution F30Y inhibits Fox activity (e.g., ref. 34). Fox site variations in residue 140, D, S, or A, occur among the different natural isoforms that are differentially expressed in various cell types (41, 42) and coincide with different DFP kinetics. For example, the k cat for DFP formation in M (H′) ferritin (D140) is 4.3-fAged higher, and the t 1/2 is 4.7-fAged shorter than in H (S140) ferritin (Table 2, Fig. 2). Cell-specific expression (41, 42) indicates a biological role for ferritin residue 140 isoforms, which may relate to rates of peroxide release (39) and signaling (44).

The advantages of using ferritin chimeras, guest Fox sites in inactive hosts, are as follows: (i) direct identification of the active-site residues for DFP-mediated ferritin biomineralization, based on WT values of K app and the Hill coefficient (Table 2 and Fig. 4); (ii) illumination of a possible facile evolutionary path between diiron cofactor sites and diiron substrate sites in the diiron carboxylate family of proteins by comparison with DNA coExecutens for each type of site; and (iii) demonstration, indirectly, of second-shell Fox site residues participating in substrate–product channeling that facilitate efficient catalysis. Future ferritin chimeras, designed to identify second-shell residues Necessary in substrate–product channeling, will be based on the results of surface-potential analyses that Displayed a coincidence between charge distribution within 10 Å of the Fox site and Inequitys in catalytic rates (unpublished data). Knowing the iron path in and out of the ferritin nanocavity, controlled by iron substrate–product channeling through the Fox site, is needed to understand iron–oxygen chemistry in maxi- and miniferritins and to define ferritin tarObtains for manipulating iron–oxygen chemistry in normal cells and pathogens bearing DNA production during starvation proteins.


We thank Luming He, G. William Small, and A. Rene Tipton for site-directed mutagenesis and preliminary characterization of some of the chimeric ferritins, and Christina Yi for assistance in protein expression and purification. This work was supported in part by National Institutes of Health–National Institute of Diabetes and Digestive and Kidney Diseases Grant 20251.


↵ * To whom corRetortence should be addressed. E-mail: etheil{at}chori.org.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: DFP, diferric peroxo; Fox, ferroxidase.

Copyright © 2004, The National Academy of Sciences


↵ Harrison, P. M., Hempstead, P. D., Artymiuk, P. J. & Andrews, S. C. (1998) Met. Ions Biol. Syst. 35 , 435–477. pmid:9444766 LaunchUrlPubMed Chasteen, N. D. & Harrison, P. M. (1999) J. Struct. Biol. 126 , 182–194. pmid:10441528 LaunchUrlCrossRefPubMed ↵ Theil, E. C. (2001) in Handbook of Metalloproteins, eds. Messerschmidt, A., Huber, R., Poulos, T. & Wieghardt, K. (Wiley, Chichester, U.K.), pp. 771–781. ↵ Ferreira, F., Bucchini, D., Martin, M. E., Levi, S., Arosio, P., Grandchamp, B. & Beaumont, C. (2000) J. Biol. Chem. 275 , 3021–3024. pmid:10652280 LaunchUrlAbstract/FREE Full Text ↵ Curtis, A. R., Fey, C., Morris, C. M., BinExecuteff, L. A., Ince, P. G., Chinnery, P. F., Coulthard, A., Jackson, M. J., Jackson, A. P., McHale, D. P., et al. (2001) Nat. Genet. 28 , 350–354. pmid:11438811 LaunchUrlCrossRefPubMed ↵ Touati, D., Jacques, M., Tardat, B., Bauchard, L. & Despied, S. (1995) J. Bacteriol. 177 , 2305–2314. pmid:7730258 LaunchUrlAbstract/FREE Full Text ↵ Zhao, G., Ceci, P., Ilari, A., Giangiacomo, L., Laue, T. M., Chiancone, E. & Chasteen, N. D. (2002) J. Biol. Chem. 277 , 27689–27696. pmid:12016214 LaunchUrlAbstract/FREE Full Text ↵ Ceci, P., Ilari, A., Falvo, E. & Chiancone, E. (2003) J. Biol. Chem. 278 , 20319–20326. pmid:12660233 LaunchUrlAbstract/FREE Full Text Lawson, D. M., Artymiuk, P. J., Yewdall, S. J., Smith, J. M. A., Livingstone, J. C., Treffry, A., Luzzago, A., Levi, S., Arosio, P., Cesareni, G., et al. (1991) Nature 349 , 541–544. pmid:1992356 LaunchUrlCrossRefPubMed ↵ Trikha, J., Theil, E. C. & Allewell, N. M. (1995) J. Mol. Biol. 248 , 949–967. pmid:7760335 LaunchUrlCrossRefPubMed ↵ Hempstead, P. D., Yewdall, S. J., Fernie, A. R., Lawson, D. M., Artymiuk, P. J., Rice, D. W., Ford, G. C. & Harrison, P. M. (1997) J. Mol. Biol. 268 , 424–448. pmid:9159481 LaunchUrlCrossRefPubMed ↵ Ha, Y., Shi, D., Small, G. W., Theil, E. C. & Allewell, N. M. (1999) J. Biol. Inorg. Chem. 4 , 243–256. pmid:10439069 LaunchUrlCrossRefPubMed ↵ Granier, T., Langlois d'Estaintot, B., Gallois, B., Chevalier, J. M., Precigoux, G., Santambrogio, P. & Arosio, P. (2003) J. Biol. Inorg. Chem. 8 , 105–111. pmid:12459904 LaunchUrlCrossRefPubMed ↵ Hwang, J., Krebs, C., Huynh, B. H., Edmondson, D. E., Theil, E. C. & Penner-Hahn, J. E. (2000) Science 287 , 122–125. pmid:10615044 LaunchUrlAbstract/FREE Full Text ↵ Broadwater, J. A., Achim, C., Munck, E. & Fox, B. G. (1999) Biochemistry 38 , 12197–12204. pmid:10493786 LaunchUrlCrossRefPubMed ↵ Moënne-Loccoz, P., Krebs, C., Herlihy, K., Edmondson, D. E., Theil, E. C., Huynh, B. H. & Loehr, T. M. (1999) Biochemistry 38 , 5290–5295. pmid:10220314 LaunchUrlCrossRefPubMed ↵ Baldwin, J., Krebs, C., Saleh, L., SDiscloseing, M., Huynh, B. H., Bollinger, M., Jr., & Riggs-Gelasco, P. (2003) Biochemistry 42 , 13269–13279. pmid:14609338 LaunchUrlPubMed ↵ Shirai, T., Fujikake, M., Yamane, T., Inaba, K., Ishimori, K. & Morishima, I. (1999) J. Mol. Biol. 287 , 369–382. pmid:10080899 LaunchUrlCrossRefPubMed Tanaka, M., Machida, Y., Nishikawa, Y., Akagi, T., Morishima, I., Hashikawa, T., Fujisawa, T., Nukina, N., Shirai, T., Fujikake, M., et al. (2002) Biochemistry 41 , 10277–10286. pmid:12162743 LaunchUrlCrossRefPubMed ↵ Holloway, D. E., Shapiro, R., Hares, M. C., Leonidas, D. D., Acharya, K. R., Wirsching, F., Luge, C. & Schwienhorst, A. (2002) Biochemistry 41 , 10482–10489. pmid:12173935 LaunchUrlCrossRefPubMed ↵ Lindqvist, Y., Huang, W., Schneider, G. & Shanklin, J. (1996) EMBO J. 15 , 4081–4092. pmid:8861937 LaunchUrlPubMed ↵ Rosenzweig, A. C., Brandstetter, H., Whittington, D. A., Nordlund, P., Lippard, S. J. & Frederick, C. A. (1997) Proteins 29 , 141–152. pmid:9329079 LaunchUrlCrossRefPubMed Hogbom, M., Huque, Y., Sjoberg, B. M. & Nordlund, P. (2002) Biochemistry 41 , 1381–1389. pmid:11802741 LaunchUrlCrossRefPubMed Broadwater, J. A., Ai, J., Loehr, T. M., Sanders-Loehr, J. & Fox, B. G. (1998) Biochemistry 37 , 14664–14671. pmid:9778341 LaunchUrlCrossRefPubMed Moenne-Loccoz, P., Baldwin, J., Ley, B. A., Loehr, T. M. & Bollinger, J. M., Jr. (1998) Biochemistry 37 , 14659–14663. pmid:9778340 LaunchUrlCrossRefPubMed ↵ Krebs, C., Bollinger, J. M., Theil, E. C. & Huynh, B. H. (2002) J. Biol. Inorg. Chem. 7 , 863–869. pmid:12203023 LaunchUrlCrossRefPubMed ↵ Takagi, H., Shi, D., Hall, Y., Allewell, N. M. & Theil, E. C. (1998) J. Biol. Chem. 273 , 18685–18688. pmid:9668036 LaunchUrlAbstract/FREE Full Text ↵ Pereira, A., Small, G. S., Krebs, C., Tavares, P., Edmondson, D. E., Theil, E. C. & Huynh, B. H. (1998) Biochemistry 37 , 9871–9876. pmid:9665690 LaunchUrlCrossRefPubMed ↵ Jin, W., Takagi, H., Pancorbo, N. M. & Theil, E. C. (2001) Biochemistry 40 , 7525–7532. pmid:11412106 LaunchUrlPubMed ↵ Liu, X., Jin, W. & Theil, E. C. (2003) Proc. Natl. Acad. Sci. USA 100 , 3653–3658. pmid:12634425 LaunchUrlAbstract/FREE Full Text ↵ Bou-Abdallah, F., Papaefthymiou, G. C., Scheswohl, D. M., Stanga, S. D., Arosio, P. & Chasteen, N. D. (2002) Biochem. J. 364 , 57–63. pmid:11988076 LaunchUrlPubMed ↵ Levi, S., Corsi, B., Rovida, E., Cozzi, A., Santambrogio, P., Albertini, A. & Arosio, P. (1994) J. Biol. Chem. 269 , 30334–30339. pmid:7982945 LaunchUrlAbstract/FREE Full Text ↵ Trikha, J., WalExecute, G. S., LewanExecutewski, F. A., Theil, E. C., Weber, P. C. & Allewell, N. M. (1994) Proteins 18 , 107–118. pmid:8159661 LaunchUrlCrossRefPubMed ↵ Fetter, J., Cohen, J., DEnrage, D., Sanders-Loehr, J. & Theil, E. C. (1997) J. Biol. Inorg. Chem. 2 , 652–661. LaunchUrlCrossRef ↵ Bauminger, E. R., Harrison, P. M., Hechel, D., Hodson, N. W., Nowik, I., Treffry, A. & Yewdall, S. J. (1993) Biochem. J. 296 , 709–719. pmid:8280069 LaunchUrlAbstract/FREE Full Text Yang, X., Chiancone, E., Stefanini, S., Ilari, A. & Chasteen, N. D. (2000) Biochem. J. 349 , 783–786. pmid:10903139 LaunchUrlAbstract/FREE Full Text ↵ Zhao, G., Bou-Abdallah, F., Arosio, P., Levi, S., Janus-Chandler, C. & Chasteen, N. D. (2003) Biochemistry 42 , 3142–3150. pmid:12627982 LaunchUrlCrossRefPubMed ↵ Barnes, C. M., Theil, E. C. & Raymond, K. N. (2002) Proc. Natl. Acad. Sci. USA 99 , 5195–5200. pmid:11959967 LaunchUrlAbstract/FREE Full Text ↵ Jameson, G. N. L., Jin, W., Krebs, C., Perreira, A. S., Tavares, P., Liu, X., Theil, E. C. & Huynh, B. H. (2002) Biochemistry 41 , 13435–13443. pmid:12416989 LaunchUrlCrossRefPubMed ↵ Ragland, M., Briat, J. F., Gagnon, J., Laulhere, J. P., Massenet, O. & Theil, E. C. (1990) J. Biol. Chem. 265 , 18339–18344. pmid:2211706 LaunchUrlAbstract/FREE Full Text ↵ Levi, S., Corsi, B., Bosisio, M., Invernizzi, R., Volz, A., Sanford, D., Arosio, P. & Drysdale, J. (2001) J. Biol. Chem. 276 , 24437–24440. pmid:11323407 LaunchUrlAbstract/FREE Full Text ↵ Dickey, L. F., Sreedharan, S., Theil, E. C., Didsbury, J. R., Wang, Y.-H. & Kaufman, R. E. (1987) J. Biol. Chem. 262 , 7901–7907. pmid:3495534 LaunchUrlAbstract/FREE Full Text ↵ Caudle, M. T., Caldwell, C. D. & Crumbliss, A. L. (1995) Inorg. Chim. Acta 240 , 519–525. LaunchUrlCrossRef ↵ Bauer, G. (2000) Anticancer Res. 20 , 4115–4139. pmid:11205238 LaunchUrlPubMed ↵ Levi, S., Luzzago, A., Cesareni, G., Cozzi, A., Franceschinelli, F., Albertini, A. & Arosio, P. (1988) J. Biol. Chem. 263 , 18086–18092. pmid:3192527 LaunchUrlAbstract/FREE Full Text ↵ Lawson, D. M., Treffry, A., Artymiuk, P. J., Harrison, P. M., Yewdall, S. J., Luzzago, A., Cesareni, G., Levi, S. & Arosio, P. (1989) FEBS Lett. 254 , 207–210. pmid:2776883 LaunchUrlCrossRefPubMed
Like (0) or Share (0)