Ferritin, a multimeric cage-like enzyme, is integral to iron metabolism across all phyla through the sequestration and storage of iron through efficient ferroxidase activity. While ferritin sequences from ∼900 species have been identified, crystal structures from only 50 species have been reported, the majority from bacterial origin. We recently isolated a secreted ferritin from the marine invertebrate Chaetopterus sp. (parchment tube worm), which resides in muddy coastal seafloors. Here, we present the first ferritin from a marine invertebrate to be crystallized and its biochemical characterization. The initial ferroxidase reaction rate of recombinant Chaetopterus ferritin (ChF) is 8-fold faster than that of recombinant human heavy-chain ferritin (HuHF). To our knowledge, this protein exhibits the fastest catalytic performance ever described for a ferritin variant. In addition to the high-velocity ferroxidase activity, ChF is unique in that it is secreted by Chaetopterus in a bioluminescent mucus. Previous work has linked the availability of Fe2+ to this long-lived bioluminescence, suggesting a potential function for the secreted ferritin. Comparative biochemical analyses indicated that both ChF and HuHF showed similar behavior toward changes in pH, temperature, and salt concentration. Comparison of their crystal structures shows no significant differences in the catalytic sites. Notable differences were found in the residues that line both 3-fold and 4-fold pores, potentially leading to increased flexibility, reduced steric hindrance, or a more efficient pathway for Fe2+ transportation to the ferroxidase site. These suggested residues could contribute to the understanding of iron translocation through the ferritin shell to the ferroxidase site.

Introduction

Bioluminescence is the ability of living organisms to produce visible light, often for communication, preying, or protection. It usually occurs in eukaryotes (except fungi) as series of bright short-lived flashes (milliseconds to seconds) [1,2]. In contrast, bioluminescence in bacteria and fungi is much longer lived (hours) and usually associated with intracellular processes coupled with the metabolic machinery [3,4]. Uniquely, long-lived bioluminescence is also observed for up to 72 h in vitro in a mucus secreted by the marine parchment tubeworm Chaetopterus sp. [5]. Unlike other long-lived bioluminescence, the light production in this secreted mucus cannot be reliant on intracellular processes. Initial characterization of this bioluminescence showed dependence on Fe2+ [57]. Subsequent studies identified abnormally high Fe2+ concentrations in the mucus and provided evidence that a ferritin-like protein was secreted along with the mucus [5]. Because of the light-stimulating effect of Fe2+, the presence of ferritin in the mucus was considered of interest for understanding the control mechanisms of this unusual light-production pattern in the worm. Recently, many mechanisms were hypothesized in which ferritin manages Fe3+/Fe2+ balance in the mucus to build an electronic source for light production, which would be a novel function of ferritin in nature [5].

Ferritins (E.C. 1.16.3) are found across almost all kingdoms of life, where they play an important role in iron storage and metabolism, while also preventing the formation of harmful reactive oxygen species [810]. The availability of Fe2+ is of high importance for many metabolic processes, such as the formation of pigments like porphyrins, hemoglobin, and chlorophyll, which are involved in oxygen delivery and photosynthesis [1114]. Ferritins exist as cytosolic, mitochondrial, membrane-bound, and secreted variants. These proteins self-assemble into hollow cage-like structures composed of 24 subunits, with an octahedral (432) symmetry, providing storage capacity for ferric iron. In vertebrates, the 24-meric cage consists of a mixture of two of the three classes of ferritins: light chain (L chain, ∼20 kDa), middle chain (M chain, ∼21 kDa), and heavy chain (H chain, ∼23 kDa). Only the H and M chains contain a catalytic site for ferroxidase activity. Ferritins take up their substrate ferrous iron (Fe2+) from solution, oxidize it, and store the iron as a compact ferric mineral (primarily ferrihydrite) inside the 24-meric cage, enabling the ferritin cage to store up to 4500 Fe atoms per cage [15]. The internalized iron mineral can be reduced and released as ferrous iron for incorporation as cofactors in protein scaffolds and porphyrins [1619].

The iron uptake, oxidation, and mineralization pathways in ferritins have been extensively studied, but the corresponding mechanistic details are not yet fully elucidated [2025]. The release of mineralized iron back into the metabolic cycle as Fe2+ is even less well understood. Three key components of the ferritin structure have been identified as important for overall ferritin activity, representing three steps of the iron chemistry:

  • (1) The iron ion channel for Fe2+ uptake. The pore at the 3-fold axis of ferritin is thought to be the primary point of entry for Fe2+. It is lined with metal-binding residues (glutamate and histidine) from three subunits and allows for shuttling of Fe2+ to the interior of the cage [20,21,25].

  • (2) The ferroxidase di-iron center for oxidation of Fe2+ to Fe3+. After uptake, the Fe2+ ions are transported by a series of metal-binding residues toward the ferroxidase di-iron center [20,26,27], which is situated between the four α-helices of each ferritin subunit. At this site, two Fe2+ ions react with either dioxygen or peroxide to form an oxidized diferric peroxo species (DFP, or blue di-iron intermediate) with a characteristic absorbance near 650 nm [19,26,28,29].

  • (3) The ferrihydrite nucleation sites allow for deposition of Fe3+ as a mineral inside the ferritin cage. After oxidation, the newly formed Fe3+ moves toward a ferrihydrite nucleation site as oxo/hydroxo multimers to be integrated into the mineral core. The resulting diferric oxo/hydroxo species [DFO(H) or mineral] has a typical absorbance ∼350 nm. The absorption peaks of the DFP and DFO(H) species are commonly used to follow the separate steps of conversion of free Fe2+ into sequestered iron mineral [2934].

Over 7000 ferritin sequences and fragments from over 900 species have been deposited in the gene database [35]. However, crystal structures have been published for ∼50 different species, most of which are bacterioferritins [36]. Of the available non-bacterial ferritin structures (human, horse, mouse, bullfrog, the insect cabbage looper, the soybean plant, the green marine seaweed Ulva pertusa and freshwater alga Cyanophora paradoxa, and the planktonic marine diatom Pseudo-nitzschia), none are from marine invertebrates. Most of these ferritins have also been characterized biochemically [30,32,3740], and through comparative studies, only mouse ferritin was found to display higher catalytic activity (2-fold) than human heavy-chain ferritin (HuHF).

The ferritin-like protein found in mucus secreted by the Chaetopterus worm was initially identified through the 82% sequence similarity to HuHF [5]. Here, we report detailed characterization of recombinantly expressed Chaetopterus ferritin (ChF) and compare the ferroxidase activity of recombinantly expressed ChF with HuHF. Since members of the Chaetopterus worm family (Chaetopterids) are found in all kinds of marine environment, including near hydrothermal vents, around mud volcanoes, at cold seep sites, and around whale falls [41,42], iron uptake was tested under varying temperatures, salt concentrations, and pH conditions to probe for adaptations to their specific environment. We ultimately present the first crystal structure of a marine invertebrate ferritin and propose key residues in the structures that may be important for accelerating iron uptake in ferritin.

Materials and methods

expression, and purification of HuHF and ChF

The ChF gene was obtained through amplification of Chaetopterus cDNA [5] and cloned into a pET24b vector between the NdeI and BamHI restriction sites. The ChF insert contained the stop codon, making the transcript an untagged protein (using forward primer: CACAAGATCATATGGCCCAGACTCAGCCG and reverse primer: GTCGTGGATCCTTAGCTGCTCAGGCTCTCCTTGT). The HuHF wild-type gene was obtained through site-directed mutagenesis of a HuHF ΔC* mutant in a pJexpress414 vector from DNA 2.0 (Menlo Park, CA) as previously described [43]. The HuHF gene is codon optimized for Escherichia coli expression. BL21 star (Invitrogen) cultures were grown at 37°C to an OD600nm of 0.7–0.8, and the proteins were expressed for 8–10 h after induction with IPTG and the cells were harvested by centrifugation at 4000×g for 25 min at 4°C. The cell pellet was resuspended in a 50 mM phosphate buffer at pH 7.5 with 150 mM NaCl and lysed using lysozyme and benzonase nuclease at 37°C for 30 min each, followed by needle sonication on ice for 3 min with 0.5 s intervals. The resulting lysate was incubated in a water bath at 75°C for 20 min, followed by centrifugation at 4000×g for 15 min to remove denatured protein. The remaining lysate solution was further purified through gel filtration on a Superdex 200 column in a 20 mM MES buffer with 150 mM NaCl at pH 6.5. Fractions with absorption at 280 and 340 nm (characteristic of the ferrihydrite mineral absorbance) were selected and combined, and ferroxidase activity was assessed using the assay described below. When necessary, the active cages were further concentrated using spin concentrators with a molecular mass cutoff of 30 kDa (Pierce).

Sample analysis for protein and iron content

SDS–PAGE was used to assess the purity of the samples. Protein purity was assessed by SDS–PAGE (Supplementary Figure S1). HuHF and ChF samples were added to a reducing loading buffer, heated to 98°C for 3 min, and loaded onto a 12% Tris–glycine gel along with an AccuRuler prestained protein ladder (Lamda Biotech). The gel was run for 50 min at 200 V and stained with Coomassie Blue to visualize the proteins. Full absorption spectra from 230 to 1000 nm were taken in a SpectraMax i3 multimode plate reader (Molecular Devices) using 200 µl samples in a 96-well plate with UV-clear bottom (Corning). More detailed, fast spectra of the blue di-iron intermediate (500–900 nm) were recorded using an in-house built setup with a tungsten halogen lamp, optic fibers to and from a (closable) cuvette holder, a Low Light Coupled Intensified SE200 Digital Spectrograph (Catalina Scientific, Tucson, AZ, U.S.A.) coupled to a camera (Princeton Instruments 7467-0002) and an amplifier (Princeton Instruments 7513-0002). Iron content was measured using inductively coupled plasma mass spectrometry (ICP-MS), for which samples were prepared at 0.02 µg/ml in 2% nitric acid. The Fe content was measured on an iCAPQc Single Quadrupole ICP-MS instrument (Thermo Fisher Scientific) with a 0.05 s dwell time.

Transmission electron microscopy

A 3.5-µl aliquot of a solution containing 10 µM ChF (by subunit; see sections below), 15 mM TRIS [Tris(hydroxymethyl)aminomethane; pH 7.4], and 150 mM NaCl was drop cast onto a negatively glow-discharged carbon-coated Cu grid (Ted Pella, Inc.). After allowing the sample to adhere to the grid surface for 1 min, excess fluid was removed by blotting with filter paper (Whatman). The sample was then washed with deionized water (18 mΩ) and stained with 1% uranyl acetate. Sample imaging was performed on an FEI Sphera transmission electron microscope equipped with a LaB6 electron gun at 200 keV and imaged on Gatan 2K2 CCD.

Ferroxidase assay (FerroZine and Ferene S)

The ferroxidase activity of ferritin was assessed by a method adapted from refs [44,45], using Fe2+ specific chelators to measure the remaining Fe2+ in solution throughout the reaction. The general reaction mixture contained 10 µl protein sample [0.1–0.5 mg/ml or 5–25 µM (subunit) stock concentration], 160 µl buffer [100 mM MES (2-(N-morpholino)ethanesulfonic acid) or MOPS (3-morpholinopropane-1-sulfonic acid) with 250 mM NaCl, pH 6.5]. The assay was initialized by the addition of 30 µl FeCl2 solution (1 mM stock concentration). From this reaction, 20 µl samples were removed and analyzed in 180 µl of a Fe2+ specific indicator. The indicators of choice were FerroZine™ [Acros Organics; 3-(2-Pyridyl)-5,6-diphenyl-1,2,4-triazine-4′,4″-disulfonic acid; 1 mM Abs at 562 nm] and FereneS [3-(2-Pyridyl)-5,6-di(2-furyl)-1,2,4-triazine-5′,5″-disulfonic acid disodium salt, Sigma–Aldrich, 20 mM Abs at 600 nm]. The assays were carried out at 40°C. To account for the background Fe2+ oxidation, a sample without protein was also measured. A sample containing EDTA (10 mM final concentration) instead of protein was used as a negative control for background absorbance. To convert absorption to Fe2+ concentrations, standard curves were made for each specific combination of chelator and experimental conditions.

This protocol was adapted for the measurements of ferroxidase activity at different pH values [50 mM buffers based on MES, MOPS, HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), and TRIS with pH varying from 3.0 to 8.6 and containing 250 mM NaCl], measurement of activity at different temperatures (20–97°C, in a buffered solution with a pH of 6.0), measurement of saturation kinetics (FeCl2 concentration varied from 0 to 15 mM, FerroZine concentration increased to 10 mM), and measurement of activity in varying ionic strengths (NaCl concentration varied from 0 to 1.5 M).

kinetics

The stopped-flow experiments used a protocol adapted from Pozzi et al. [30]. The ferritin samples were buffer-exchanged into a buffered solution (200 mM MOPS with 200 mM NaCl at pH 7.0) and concentrated to 200 µM (by subunit). FeSO4 stock solutions were made fresh by dissolving FeSO4 in 1 mM HCl. FeSO4 working concentrations were expanded to 66.7, 100, 200, 300, 400, 600, 800, 1000, and 4000 µM. This is, respectively, 1 Fe2+ ion per 3-fold ion channel, 1 Fe2+ ion per subunit, 2, 3, 4, 6, 8, 10, and 40 Fe2+ per subunit. During symmetric mixing of 100 µl of ferritin and 100 µl FeSO4 at room temperature, 4000 data points were collected during the first 10 s, logarithmically spread over time (0.25 ms intervals at the start, 23 ms at the end) with an initial delay of 0.25 ms. Absorption was monitored at 650 nm for the formation of the DFP species (the blue intermediate) and at 350 nm to measure the formation of the DFO(H) species (the iron oxide mineral). Each measurement was repeated five times and averaged for further analysis.

To determine the initial velocities, the data (all data for A350 nm, the first 0.01 s for A650 nm) were fitted with a monoexponential decay curve y = B1 * exp(−x/t1) + y0. The slopes were calculated as y0/t1, and these values were then plotted against the total added Fe2+ concentration. Since co-operativity of the ferritin ferroxidase activity occurs [4649], the plots were fitted with the Hill equation.

Crystallization and X-ray diffraction

The N81D mutation was installed on ChF to facilitate the formation of highly ordered crystals through a Ca-mediated crystal contact identical with the one used by Lawson et al. [50] to crystallize HuHF. Single crystals of N81DChF were obtained through sitting drop vapor diffusion. Crystallization conditions are shown in Supplementary Table S1. Crystals suitable for sc-XRD were removed from the well and soaked for 30 min in 10 µl of a buffered solution containing 10 mM Zn, 20 mM Tris (pH 8.5), 20 mM CaCl2, and 3% (v/v) PEG 400. After soaking, crystals were cryoprotected with polyfluoropolyether and frozen in liquid N2. sc-XRD data were collected at 100 K at Beamline 12-2 of the Stanford Synchrotron Radiation Laboratory (SSRL) using 0.98 Å radiation. The data were integrated using iMosflm [51] and scaled using Aimless [52] (see Supplementary Table S2 for statistics). The structure for the Zn-bound N81DChF was solved to 1.57. Molecular replacement was performed with Phaser [53] using the homology model derived from the I-TASSER server [54] as the search model. All residues were truncated at Cβ in the initial search model. Rigid-body, positional, and thermal refinements were carried out using Phenix [55]. Coot [56] was used for iterative manual model building. All figures were produced with Pymol [57].

Results and discussion

Protein purification

The spectral analysis and SDS–PAGE of the purified samples revealed no noticeable impurities (Supplementary Figures S1 and S2). The iron content of the purified protein was determined at ∼90 ± 2 atoms per cage for both HuHF and ChF by means of ICP-MS. An initial comparison with apoferritins was performed to exclude the effect of iron load. In these tests, Apo–ChF showed the same reaction rate increase over Apo–HuHF as observed in the non-Apo ferritins (data not shown).

Transmission electron microscopy: ChF forms ferritin-like cages

Transmission electron microscopy (TEM) images of purified ChF show donut-shaped structures with a diameter of 11.9 ± 0.9 nm (Figure 1), which is consistent with other characterized ferritin variants [43,58]. Based on their similarity, it is likely that ChF forms a 24-meric cage.

Negative-stained TEM image of recombinant ChF.

Figure 1.
Negative-stained TEM image of recombinant ChF.

The 10 µM sample shows structures highly similar to TEM images of HuHF available in the literature.

Figure 1.
Negative-stained TEM image of recombinant ChF.

The 10 µM sample shows structures highly similar to TEM images of HuHF available in the literature.

Kinetics of the ferroxidase activity: ChF performs faster than HuHF

The sequence similarity of ChF to other H-chain ferritins suggests that ChF likely possesses similar ferroxidase activity. Biochemical characterization was performed using established iron uptake assays and stopped-flow experiments. Throughout all experiments, the most remarkable difference between ChF and HuHF was that the (net) iron uptake and/or ferroxidase activity of ChF occurred much more rapidly than that of HuHF. This difference can be observed at both short (milliseconds) and long (hours) time scales. In our standard ferroxidase assay, lasting up to 3 h, we measured a half time of 8.2 ± 0.6 min for ChF, compared with 59.7 ± 4.3 min for HuHF in the same conditions, corresponding to a 7.3-fold rate increase (Figure 2 and Table 1).

Amount of Fe2+ converted over time by 1.5 µM ferritin, fitted to a monoexponential curve.

Figure 2.
Amount of Fe2+ converted over time by 1.5 µM ferritin, fitted to a monoexponential curve.

Aliquots of 50 µl were taken from 4 ml reaction mixtures at different time points and analyzed by adding the fraction to 150 µl Ferene S (4 mM). A standard curve was used to convert the measured absorbance (600 nm) to Fe2+ concentration.

Figure 2.
Amount of Fe2+ converted over time by 1.5 µM ferritin, fitted to a monoexponential curve.

Aliquots of 50 µl were taken from 4 ml reaction mixtures at different time points and analyzed by adding the fraction to 150 µl Ferene S (4 mM). A standard curve was used to convert the measured absorbance (600 nm) to Fe2+ concentration.

Table 1
Optimal conditions and kinetic parameters of ChF in comparison with HuHF

t1/2 is the half time expressed in minutes. Topt is the temperature at which optimal activity was measured. Vmax is the maximal velocity of the reaction. Km is the concentration at which half Vmax is reached. λmax is the wavelength of peak absorbance of the DFP intermediate.

 ChF HuHF 
t1/2 (min) 8.2 ± 0.6 59.7 ± 4.3 
Topt (°C) 58.1 ± 0.3 62.1 ± 0.4 
Optimal pH 6.77 ± 0.01 6.72 ± 0.03 
Parameters DFO(H) formation 
Vmax (substrate per subunit s−133.9 ± 1.5 4.16 ± 0.25 
Km (µM) 283 ± 23 336 ± 38 
n 1.58 ± 0.14 1.49 ± 0.17 
Parameters DFP formation 
Vmax (a.u. s−16.0 ± 0.12 1.21 ± 0.08 
Km (µM) 430 ± 25 121 ± 19 
n 1.1 ± 0.1 0.9 ± 0.2 
λmax 619 nm 653 nm 
 ChF HuHF 
t1/2 (min) 8.2 ± 0.6 59.7 ± 4.3 
Topt (°C) 58.1 ± 0.3 62.1 ± 0.4 
Optimal pH 6.77 ± 0.01 6.72 ± 0.03 
Parameters DFO(H) formation 
Vmax (substrate per subunit s−133.9 ± 1.5 4.16 ± 0.25 
Km (µM) 283 ± 23 336 ± 38 
n 1.58 ± 0.14 1.49 ± 0.17 
Parameters DFP formation 
Vmax (a.u. s−16.0 ± 0.12 1.21 ± 0.08 
Km (µM) 430 ± 25 121 ± 19 
n 1.1 ± 0.1 0.9 ± 0.2 
λmax 619 nm 653 nm 

The initial kinetics at the millisecond time scale, as derived from stopped-flow experiments, show a similarly elevated reaction rate for ChF. The reaction was followed at 350 nm to show the formation of the Fe3+ mineral [DFO(H) species] inside the ferritin cage, as well as at 650 nm to follow the formation of the so-called blue intermediate (the DFP species). Both species are formed 7.7 times faster in ChF than those in HuHF (Figure 3 and Supplementary Figure S3).

Initial iron uptake kinetics showing ∼8 times faster kinetics for ChF compared to HuHF.

Figure 3.
Initial iron uptake kinetics showing ∼8 times faster kinetics for ChF compared to HuHF.

(a) Initial iron uptake kinetics were monitored in stopped-flow setup over 10 s by measuring the DFP formation (650 nm) and DFO(H) formation (350 nm), here shown for 100 mM ferritins with 300 mM FeSO4 . (b) The linear ranges (fitted in black) in the first few milliseconds were used to determine the initial rate of formation for both species. Both graphs illustrate the increased reaction rate of ChF over HuHF. Additional ferritin:Fe2+ ratios can be found in Supplementary Figure S3.

Figure 3.
Initial iron uptake kinetics showing ∼8 times faster kinetics for ChF compared to HuHF.

(a) Initial iron uptake kinetics were monitored in stopped-flow setup over 10 s by measuring the DFP formation (650 nm) and DFO(H) formation (350 nm), here shown for 100 mM ferritins with 300 mM FeSO4 . (b) The linear ranges (fitted in black) in the first few milliseconds were used to determine the initial rate of formation for both species. Both graphs illustrate the increased reaction rate of ChF over HuHF. Additional ferritin:Fe2+ ratios can be found in Supplementary Figure S3.

The co-operativity of the initial reaction rate for the formation of both DFO(H) and DFP was assessed by plotting their formation as a function of initial Fe2+ concentration and applying the Hill equation. We found a Hill coefficient of ∼1.5 for the formation of DFO(H), but no co-operative effect for the formation of the DFP (Table 1) was observed. Several research groups have previously shown co-operativity in the ferroxidase activity of ferritin, with a Hill coefficient of ∼1.5 [28,46,48]. The Vmax for ChF of 33.9 Fe2+ ions converted per subunit per second was over 8-fold faster than the Vmax of HuHF of 4.2 Fe2+ ions per subunit per second. Interestingly, the DFP formation rate in HuHF showed signs of early saturation at a concentration as low as 2 Fe2+ ions per ferritin subunit, while the reaction in ChF did not reach saturation at 20 Fe2+ per subunit (Figure 4). The formation of the DFP (blue intermediate) does not appear to be the rate-determining step for the HuHF as the formation rate of the mineral never reached saturation and had a Km value higher than that of ChF (Table 1). A reduced Km or increased affinity of ChF for Fe2+ would be an interesting property for a ferritin that operates in seawater, where Fe2+ concentrations are particularly low. Considering the worm biology however, the ChF is most likely iron-loaded when secreted from the body, the load being dependent on the amount of iron available from the worm's diet. The bioluminescence produced by the secreted mucus from Chaetopterus is regulated or potentially even driven by the presence of free Fe2+. In that case, the iron release properties might be more important than uptake, and a lower affinity would be preferred.

Hill plot representation of DFP and DFO(H) formation of ChF and HuHF.

Figure 4.
Hill plot representation of DFP and DFO(H) formation of ChF and HuHF.

Hill plot of (a) DFP formation and (b) DFO(H) formation measured immediately after mixing in a stopped-flow chamber. All measurements were performed by symmetrical mixing of 200 µM ferritin solutions (final concentration of 100 µM) and increasing FeSO4 concentrations. Both graphs show that ChF is ∼8 times faster in both DFP and DFO(H) formation than recombinant HuHF.

Figure 4.
Hill plot representation of DFP and DFO(H) formation of ChF and HuHF.

Hill plot of (a) DFP formation and (b) DFO(H) formation measured immediately after mixing in a stopped-flow chamber. All measurements were performed by symmetrical mixing of 200 µM ferritin solutions (final concentration of 100 µM) and increasing FeSO4 concentrations. Both graphs show that ChF is ∼8 times faster in both DFP and DFO(H) formation than recombinant HuHF.

While the higher absolute absorption at 650 nm in the ChF reaction could potentially be caused by a difference in extinction coefficient or spectral shift between the DFP intermediates of the two variants, this is unlikely since the sequences of the ferroxidase sites are nearly identical. Experimental spectra of the DFP intermediates can only be compared on the basis of the position of their absorption peak, as the concentration of the DFP species during the reaction is unknown. Their peak positions are nearly identical (Figure 5), which leads us to believe that the differences in absorbance in the stopped-flow experiments are, indeed, indicative of the formation of more DFP species per millisecond. This indicates that the formation of both DFO(H) and DFP species in ChF is accelerated in comparison with HuHF, and that both steps of the ferritin activity are limited by a different rate-determining step, potentially the accessibility of the ferroxidase site for the Fe2+ ions.

Absorbance spectrum of the DFP species of ChF and HuHF.

Figure 5.
Absorbance spectrum of the DFP species of ChF and HuHF.

The spectra were captured using the Low Light Coupled Intensified SE200 Digital Spectrograph during the addition of 4 mM FeCl2 to 1 mM solutions of ChF and HuHF.

Figure 5.
Absorbance spectrum of the DFP species of ChF and HuHF.

The spectra were captured using the Low Light Coupled Intensified SE200 Digital Spectrograph during the addition of 4 mM FeCl2 to 1 mM solutions of ChF and HuHF.

The comparative interpretation of these results may also depend heavily on how different the optimal reaction conditions are for ChF with respect to HuHF. ChF and HuHF occur naturally in very different environments and may have evolved to increase their effectiveness. To rule out the chance that the increased velocity would be purely due to a large difference in optimal conditions, iron uptake by both ferritins was tested at varying temperatures, pH, and salt concentrations.

Ferroxidase activity was affected similarly by temperature, pH, and ionic strength in both ChF and HuHF

Determining the optimal reaction conditions for ferroxidase activity by ChF in comparison with HuHF was important for two reasons: first, because the Chaetopterus worm lives in seawater with a temperature range of roughly 10–20°C, while HuHF is found at human body temperature (37°C); second, because genealogy of the Chaetopterus sp. shows that its ancestors were living near hydrothermal vents as well as near cold seeps [41,42], raising the question about adaptation to a large range of environmental conditions. No experimental data assessing the optimal temperature of HuHF activity were found in the published literature. Our experiments showed very comparable optimal temperatures for ChF (58.4°C) and HuHF (62.1°C) (see Figure 6 and Table 1), an interesting observation given the sequence similarity of the proteins, while the sequence divergence does not seem to bring any adaptation to the temperature of their specific environments. ChF is consistently faster than HuHF except at very high temperatures where the analysis is more complicated by both the Fe2+ auto-oxidation and potential denaturation of ferritin.

Amount of Fe2+ converted in 15 min by the same amount of ChF vs. HuHF at different temperatures.

Figure 6.
Amount of Fe2+ converted in 15 min by the same amount of ChF vs. HuHF at different temperatures.

20 µl aliquots were taken from 300 µl reaction mixtures incubating in a thermocycler with temperature gradient. A standard curve was used to convert the measured absorbance (562 nm) into Fe2+ concentration.

Figure 6.
Amount of Fe2+ converted in 15 min by the same amount of ChF vs. HuHF at different temperatures.

20 µl aliquots were taken from 300 µl reaction mixtures incubating in a thermocycler with temperature gradient. A standard curve was used to convert the measured absorbance (562 nm) into Fe2+ concentration.

Different reports on optimal pH can be found for HuHF, generally agreeing that its activity is lost below pH 6.0–6.5 [40,59]. Here, we tested ferroxidase activity in pH ranging from 5.6 to 7.4. We found that optimal activity was pH ∼6.7–6.8 for both HuHF and ChF ferritins (Figure 7 and Table 1). These results confirm the loss in activity below pH 6.0–6.5, but also show reduced activity relative to auto-oxidation of free Fe2+ at pH values higher than 7. The pH of sea water is usually higher than 7.4, even up to 8.3, but under these conditions the auto-oxidation of Fe2+in vitro happens too fast to perform any ferroxidase measurements.

Amount of Fe2+ converted by identical amounts of ChF vs. HuHF at different pH.

Figure 7.
Amount of Fe2+ converted by identical amounts of ChF vs. HuHF at different pH.

For each condition, three reaction mixtures of 200 µl were halted after 5 min by adding 10 µl Ferene S and analyzed by measuring Abs (600 nm) in a microplate reader.

Figure 7.
Amount of Fe2+ converted by identical amounts of ChF vs. HuHF at different pH.

For each condition, three reaction mixtures of 200 µl were halted after 5 min by adding 10 µl Ferene S and analyzed by measuring Abs (600 nm) in a microplate reader.

Dependence of the ferroxidase activity on ionic strength was tested by varying the NaCl concentration between 0 and 1.5 M. Neither of the ferritins showed any preference for salt concentration between 0.3 and 1.5 M (Supplementary Figure S4). Unfortunately, it was impossible to obtain reasonable experimental half times for the ferroxidase reaction alone in filtered sea water or artificial sea water because auto-oxidation of Fe2+ in the negative controls was happening too fast.

Based on these initial findings, we decided to perform all kinetics experiments in an MOPS buffer with 250 mM NaCl at pH 6.7 and at 40°C, since there was no moderate temperature at which the reaction rates appeared the same. Reaction rates were very close at high temperatures (>90°C; Figure 6), but at that point the reaction is compromised by water evaporation, Fe2+ auto-oxidation, and potential denaturation of the ferritin cages. The ratio of activity (ChF/HuHF) was large at low temperatures, but stabilized ∼36°C, slowly but steadily decreasing to 1 around 95°C, as the measured reaction rates are the same for ChF and HuHF at that temperature.

Crystallography: same overall structure, near-identical active sites

To facilitate the formation of highly ordered crystals, a Ca2+ crystal contact was installed through the N81D mutation. As expected, this mutation co-ordinated Ca2+ and forms ChF crystals with face-centered cubic symmetry. The structure of Zn-bound N81DChF was solved to 1.57 Å. Like other mammalian H-chain ferritins, ChF forms a hollow octahedral 24-meric cage (Supplementary Figure S5a). Each subunit of the protein cage consists of a four-helix bundle with a short fifth helix that extends into the 4-fold pore. The backbone of ChF overlays nearly perfectly (RMSD = 0.236 Å) with that of HuHF (PDB ID: 2CEI [60]). The only notable difference between the backbones of these two ferritins is the absence of three residues (HuHF: Ala160, Glu162, Ser163; see Supplementary Figure S5c) on ChF that comprise the loop between the fourth and fifth helices, as can also be seen in the sequence alignment (Figure 8). This shorter loop on ChF allows P157 to adopt a trans conformation. There is no indication that a shorter loop on ChF would affect ferroxidase activity, especially since it does not perturb the location of the fifth helix in the 4-fold pore (relative to HuHF). However, this could potentially affect the flexibility in the loop and open the 4-fold channel for easier access for Fe2+ ions.

Protein sequence comparison between HuHF and ChF.

Figure 8.
Protein sequence comparison between HuHF and ChF.

Important residues for the three sites identified in HuHF are indicated with triangles and residues near the active sites that would be interesting for further studies through mutagenesis are highlighted with rectangles.

Figure 8.
Protein sequence comparison between HuHF and ChF.

Important residues for the three sites identified in HuHF are indicated with triangles and residues near the active sites that would be interesting for further studies through mutagenesis are highlighted with rectangles.

Since the overall topology of these two ferritins is essentially identical, their considerably different ferroxidase activities must stem from specific differences in the residue composition of the two ferritins. Despite originating from vastly different organisms, ChF and HuHF have a sequence identity of 66.3%. Comparison of the ferroxidase sites of the Zn-bound structures for ChF and HuHF reveals that these active sites are nearly identical (Figure 9), thus implicating residues outside the ferroxidase site to account for the increase in ferroxidase activity of ChF. Initial uptake and internalization of Fe2+ are thought to occur primarily through the 3-fold pores of H-chain ferritins [20,21,25]. Much of the 3-fold channel is highly conserved between these two ferritins, including the three residues that are observed bound to Zn2+ (ChF: His115, Cys127, and Glu131). On ChF, the outer surface of the 3-fold pore has an alanine (ChF: Ala119) instead of the threonine found in HuHF (HuHF: Thr122, Figure 8). This would allow for a slightly more open entrance to the pore that could help facilitate more rapid Fe2+ uptake (Figure 10). The 3-fold pore of ChF also has an alanine residue (Ala124) where HuHF has a proline (Pro127), which could add flexibility to ChF that HuHF does not exhibit (Supplementary Figure S5b). The metal co-ordination site in the 4-fold channels also differs between the ferritins, but both have metal-binding residues in the right place (ChF Glu167 vs. HuHF His173, see Supplementary Figure S5d). The importance of the 4-fold channel for Fe2+ uptake is under debate [20,21,25], but it is unlikely that this difference could lead to the significant increase in activity.

Ferroxidase sites of the Zn-bound ferritins are shown for HuHF (green, PDB ID: 2CEI [61]) and ChF (purple, PDB ID: 5WPN) structures.

Figure 9.
Ferroxidase sites of the Zn-bound ferritins are shown for HuHF (green, PDB ID: 2CEI [61]) and ChF (purple, PDB ID: 5WPN) structures.

Important residues are highlighted as sticks and Zn atoms are shown as spheres. The two ferroxidase sites for ChF and HuHF are nearly identical.

Figure 9.
Ferroxidase sites of the Zn-bound ferritins are shown for HuHF (green, PDB ID: 2CEI [61]) and ChF (purple, PDB ID: 5WPN) structures.

Important residues are highlighted as sticks and Zn atoms are shown as spheres. The two ferroxidase sites for ChF and HuHF are nearly identical.

Proposed residues that may contribute to increased velocity of ChF (PDB ID: 5WPN) over HuHF (PDB ID: 2CEI [61]).

Figure 10.
Proposed residues that may contribute to increased velocity of ChF (PDB ID: 5WPN) over HuHF (PDB ID: 2CEI [61]).

(a) ChF A120 (marked with ) leads to a more open 3-fold pore which may reduce steric hindrance and facilitate Fe2+ uptake relative to the T122 in HuHF (b). (c) ChF E133 and E137 provide additional acidic residues between ion channel and ferroxidase site that may account for the faster translocation of the Fe2+ ions than the combination of T135 and N139 in HuHF (d). The acidic O atoms are shown as red spheres to emphasize the more complete path of Fe2+ transportation in ChF when compared with HuHF. The in (c) and (d) marks the same residue as in (a) and (b), respectively.

Figure 10.
Proposed residues that may contribute to increased velocity of ChF (PDB ID: 5WPN) over HuHF (PDB ID: 2CEI [61]).

(a) ChF A120 (marked with ) leads to a more open 3-fold pore which may reduce steric hindrance and facilitate Fe2+ uptake relative to the T122 in HuHF (b). (c) ChF E133 and E137 provide additional acidic residues between ion channel and ferroxidase site that may account for the faster translocation of the Fe2+ ions than the combination of T135 and N139 in HuHF (d). The acidic O atoms are shown as red spheres to emphasize the more complete path of Fe2+ transportation in ChF when compared with HuHF. The in (c) and (d) marks the same residue as in (a) and (b), respectively.

The most profound difference between the two structures is found near the interior of the 3-fold channel. ChF has two additional acidic residues between the 3-fold pore and the ferroxidase site (ChF Glu132 and Glu136) that are strategically placed one α-helical turn apart and are located on both sides of a commonly observed anomalous bend in α-helix 4 of the ferritin (Figure 10). In comparison with the threonine and glutamine duo in HuHF (Thr135 and Gln139), these glutamate residues could better facilitate Fe2+ transport path between the 3-fold pore and the ferroxidase site. While the crystal structure does not show evidence of these residues binding metal ions, the analog of HuHF Thr135 (Thr168 in soy bean) has been identified as a metal-binding site in the crystal structure of soy bean ferritin [61]. To our knowledge, none of these proposed residues has been the target of mutagenesis studies yet. Studying the effects of these residues through site-directed mutagenesis will be the scope of further research.

Conclusion

We have characterized and crystallized the first example of a marine invertebrate ferritin, originating from the parchment tubeworm Chaetopterus sp. The formation of a 24-meric assembly, analogous to other mammalian ferritins, was confirmed through TEM and XRD experiments. Through colorimetric and spectroscopic measurements, we determined that ChF mineralizes iron ∼8-fold faster than HuHF. Since both ferritins show only minor differences in optimal conditions of temperature, salt concentration, and pH on the ferroxidase activity, we believe that the ChF is performing faster because of differences in essential residues near active sites. The crystal structure of ChF is also similar to other available eukaryotic ferritin structures, including the highly conserved ferroxidase site. A select few residues outside the active sites have been suggested that may affect the ferroxidase velocity of ChF. Many of the proposed residues are situated around the 3-fold pore and may affect the pore opening on the outside of the cage (ChF Ala119) or provide a more efficient pathway for movement of Fe2+ through the inside of the cage and toward the ferroxidase site (ChF Glu132 and Glu136). Other sites may be affected by increased flexibility in the 3-fold loop (Ala124). Further studies will be needed to address if any of these residues positively affect the rate of ferroxidase activity.

Abbreviations

     
  • ChF

    Chaetopterus ferritin

  •  
  • DFO(H)

    diferric oxo/hydroxo

  •  
  • DFP

    diferric peroxo

  •  
  • H chain

    heavy chain

  •  
  • HEPES

    4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

  •  
  • HuHF

    human heavy-chain ferritin

  •  
  • ICP-MS

    inductively coupled plasma mass spectrometry

  •  
  • L chain

    light chain

  •  
  • M chain

    middle chain

  •  
  • MES

    2-(N-morpholino)ethanesulfonic acid

  •  
  • MOPS

    3-morpholinopropane-1-sulfonic acid

  •  
  • sc-XRD

    single crystal X-ray diffraction

  •  
  • TEM

    transmission electron microscopy

  •  
  • TRIS

    Tris(hydroxymethyl)aminomethane

Author Contribution

E.D.M. and D.D.D. formulated the research design and are responsible for all cloning and enzymatic characterization. J.B.B. and F.A.T. performed all crystallography related work. Both laboratories expressed and purified the ferritins. E.D.M. wrote the first draft of the manuscript. All authors provided suggestions and comments on the text; all authors read and agreed with the submitted version of the text.

Funding

E.D.M. and D.D.D. are supported by the Air Force Office of Scientific Research [grant no. FA9550-17-0189 to D.D.D.]. J.B.B. and F.A.T. were supported by the National Science Foundation [protein crystallography, Grant DMR-1602537 to F.A.T.].

Acknowledgments

The authors wish to thank Dr K.L. Hailey and Prof. P.A. Jennings at the Department of Chemistry and Biochemistry at UC San Diego for access to and support with the stopped-flow instrument. We acknowledge the Scripps Isotope Geochemistry Laboratory for access to perform ICP-MS measurements. We thank Phil Zerofski, the MBRD Invertebrate Collector from the Scripps Institution of Oceanography at UC San Diego, for collection of the marine worms in the field.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Hastings
,
W.J.
and
Morin
,
J.G.
(
1991
) Bioluminescence. In
Neural and Integrative Animal Physiology-Animal Physiology
,
4th edn
. (
Prosser
,
C.L.
, ed.).
John Wiley & Sons
,
New York
, pp.
131
170
2
Haddock
,
S.H.D.
,
Moline
,
M.A.
and
Case
,
J.F.
(
2010
)
Bioluminescence in the sea
.
Ann. Rev. Mar. Sci.
2
,
443
493
3
Desjardin
,
D.E.
,
Oliveira
,
A.G.
and
Stevani
,
C.V.
(
2008
)
Fungi bioluminescence revisited
.
Photochem. Photobiol. Sci.
7
,
170
182
4
Tu
,
S.-C.
and
Mager
,
H.I.X.
(
1995
)
Biochemistry of bacterial bioluminescence
.
Photochem. Photobiol.
62
,
615
624
5
Rawat
,
R.
and
Deheyn
,
D.D.
(
2016
)
Evidence that ferritin is associated with light production in the mucus of the marine worm Chaetopterus
.
Sci. Rep.
6
,
36854
6
Shimomura
,
O.
and
Johnson
,
F.H.
(
1968
)
Chaetopterus photoprotein: crystallization and cofactor requirements for bioluminescence
.
Science
159
,
1239
1240
7
Shimomura
,
O.
and
Johnson
,
F.H.
(
1966
) Partial purification and properties of the Chaetopterus luminescence system. In
Bioluminescence in Progress
(
Johnson
,
F.H.
and
Haneda
,
Y.
, eds).
Princeton University Press
,
Princeton, NJ
, pp.
495
521
8
Arosio
,
P.
and
Levi
,
S.
(
2002
)
Ferritin, iron homeostasis, and oxidative damage
.
Free Radical Biol. Med.
33
,
457
463
9
Orino
,
K.
,
Lehman
,
L.
,
Tsuji
,
Y.
,
Ayaki
,
H.
,
Torti
,
S.V.
and
Torti
,
F.M.
(
2001
)
Ferritin and the response to oxidative stress
.
Biochem. J.
357
,
241
247
10
Smith
,
J.L.
(
2004
)
The physiological role of ferritin-like compounds in bacteria
.
Crit. Rev. Microbiol.
30
,
173
185
11
Schümann
,
K.
(
2001
)
Safety aspects of iron in food
.
Ann. Nutr. Metab.
45
,
91
101
12
Dev
,
S.
and
Babitt
,
J.L.
(
2017
)
Overview of iron metabolism in health and disease
.
Hemodial. Int.
21
(
suppl 1
),
S6
S20
13
Carrier
,
M.-C.
,
Bourassa
,
J.-S.
and
Massé
,
E.
(
2017
)
Cellular homeostasis: a small RNA at the crossroads of iron and photosynthesis
.
Curr. Biol.
27
,
R380
R383
14
Camacho
,
A.
,
Walter
,
X.A.
,
Picazo
,
A.
and
Zopfi
,
J.
(
2017
)
Photoferrotrophy: remains of an ancient photosynthesis in modern environments
.
Front. Microbiol.
8
,
323
15
Granick
,
S.
(
1946
)
Ferritin: its properties and significance for iron metabolism
.
Chem. Rev.
38
,
379
403
16
Theil
,
E.C.
,
Matzapetakis
,
M.
and
Liu
,
X.
(
2006
)
Ferritins: iron/oxygen biominerals in protein nanocages
.
J. Biol. Inorg. Chem.
11
,
803
810
17
Hintze
,
K.J.
and
Theil
,
E.C.
(
2006
)
Cellular regulation and molecular interactions of the ferritins
.
Cell. Mol. Life Sci.
63
,
591
600
18
Yoshida
,
T.
,
Tanaka
,
M.
,
Sotomatsu
,
A.
and
Hirai
,
S.
(
1995
)
Activated microglia cause superoxide-mediated release of iron from ferritin
.
Neurosci. Lett.
190
,
21
24
19
Ebrahimi
,
K.H.
,
Hagedoorn
,
P.-L.
and
Hagen
,
W.R.
(
2015
)
Unity in the biochemistry of the iron-storage proteins ferritin and bacterioferritin
.
Chem. Rev.
115
,
295
326
20
Bou-Abdallah
,
F.
(
2010
)
The iron redox and hydrolysis chemistry of the ferritins
.
Biochim. Biophys. Acta Gen. Subj.
1800
,
719
731
21
Ebrahimi
,
K.H.
,
Bill
,
E.
,
Hagedoorn
,
P.-L.
and
Hagen
,
W.R.
(
2012
)
The catalytic center of ferritin regulates iron storage via Fe(II)-Fe(III) displacement
.
Nat. Chem. Biol.
8
,
941
948
22
Masuda
,
T.
,
Goto
,
F.
,
Yoshihara
,
T.
and
Mikami
,
B.
(
2010
)
The universal mechanism for iron translocation to the ferroxidase site in ferritin, which is mediated by the well conserved transit site
.
Biochem. Biophys. Res. Commun.
400
,
94
99
23
Haldar
,
S.
,
Bevers
,
L.E.
,
Tosha
,
T.
and
Theil
,
E.C.
(
2011
)
Moving iron through ferritin protein nanocages depends on residues throughout each four α-helix bundle subunit
.
J. Biol. Chem.
286
,
25620
25627
24
Bradley
,
J.M.
,
Moore
,
G.R.
and
Le Brun
,
N.E.
(
2014
)
Mechanisms of iron mineralization in ferritins: one size does not fit all
.
J. Biol. Inorg. Chem.
19
,
775
785
25
Hagen
,
W.R.
,
Hagedoorn
,
P.-L.
and
Ebrahimi
,
K.H.
(
2017
)
The workings of ferritin: a crossroad of opinions
.
Metallomics
9
,
595
605
26
Behera
,
R.K.
and
Theil
,
E.C.
(
2014
)
Moving Fe2+ from ferritin ion channels to catalytic OH centers depends on conserved protein cage carboxylates
.
Proc. Natl Acad. Sci. U.S.A.
111
,
7925
7930
27
Tosha
,
T.
,
Hasan
,
M.R.
and
Theil
,
E.C.
(
2008
)
The ferritin Fe2 site at the diiron catalytic center controls the reaction with O2 in the rapid mineralization pathway
.
Proc. Natl Acad. Sci. U.S.A.
105
,
18182
18187
28
Fetter
,
J.
,
Cohen
,
J.
,
Danger
,
D.
,
Sanders-Loehr
,
J.
and
Theil
,
E.C.
(
1997
)
The influence of conserved tyrosine 30 and tissue-dependent differences in sequence on ferritin function: use of blue and purple Fe(III) species as reporters of ferroxidation
.
J. Biol. Inorg. Chem.
2
,
652
661
29
Pereira
,
A.S.
,
Small
,
W.
,
Krebs
,
C.
,
Tavares
,
P.
,
Edmondson
,
D.E.
,
Theil
,
E.C.
et al. 
(
1998
)
Direct spectroscopic and kinetic evidence for the involvement of a peroxodiferric intermediate during the ferroxidase reaction in fast ferritin mineralization
.
Biochemistry
37
,
9871
9876
30
Pozzi
,
C.
,
Di Pisa
,
F.
,
Bernacchioni
,
C.
,
Ciambellotti
,
S.
,
Turano
,
P.
and
Mangani
,
S.
(
2015
)
Iron binding to human heavy-chain ferritin
.
Acta Crystallogr. D Struct. Biol.
71
,
1909
1920
31
Zhang
,
B.
and
Watt
,
G.D.
(
2007
)
Anaerobic iron deposition into horse spleen, recombinant human heavy and light and bacteria ferritins by large oxidants
.
J. Inorg. Biochem.
101
,
1676
1685
32
Bou-Abdallah
,
F.
,
Santambrogio
,
P.
,
Levi
,
S.
,
Arosio
,
P.
and
Chasteen
,
N.D.
(
2005
)
Unique iron binding and oxidation properties of human mitochondrial ferritin: a comparative analysis with human H-chain ferritin
.
J. Mol. Biol.
347
,
543
554
33
Ebrahimi
,
K.H.
,
Bill
,
E.
,
Hagedoorn
,
P.-L.
and
Hagen
,
W.R.
(
2016
)
Spectroscopic evidence for the role of a site of the di-iron catalytic center of ferritins in tuning the kinetics of Fe(ii) oxidation
.
Mol. Biosyst.
12
,
3576
3588
34
Bernacchioni
,
C.
,
Ciambellotti
,
S.
,
Theil
,
E.C.
and
Turano
,
P.
(
2015
)
Is His54 a gating residue for the ferritin ferroxidase site?
Biochim. Biophys. Acta Proteins Proteomics
1854
,
1118
1122
35
Gene [Internet]
.
National Library of Medicine (US), National Center for Biotechnology Information
,
Bethesda, MD
;
2004
-
[cited 2017 Feb 10]. Available from:
https://www.ncbi.nlm.nih.gov/gene/
36
Berman
,
H.M.
,
Westbrook
,
J.
,
Feng
,
Z.
,
Gilliland
,
G.
,
Bhat
,
T.N.
,
Weissig
,
H.
et al. 
(
2000
)
The protein data bank
.
Nucleic Acids Res.
28
,
235
242
37
Proulx-Curry
,
P.M.
and
Chasteen
,
N.D.
(
1995
)
Molecular aspects of iron uptake and storage in ferritin
.
Coordin. Chem. Rev.
144
,
347
368
38
Cavallo
,
S.
,
Mei
,
G.
,
Stefanini
,
S.
,
Rosato
,
N.
,
Finazzi-Agró
,
A.
and
Chiancone
,
E.
(
1998
)
Formation and movement of Fe(III) in horse spleen, H- and L-recombinant ferritins. A fluorescence study
.
Protein Sci.
7
,
427
432
39
Orino
,
K.
,
Harada
,
S.
,
Natsuhori
,
M.
,
Takehara
,
K.
and
Watanabe
,
K.
(
2004
)
Kinetic analysis of bovine spleen apoferritin and recombinant H and L chain homopolymers: iron uptake suggests early stage H chain ferroxidase activity and second stage L chain cooperation
.
BioMetals
17
,
129
134
40
Sun
,
S.
,
Arosio
,
P.
,
Levi
,
S.
and
Chasteen
,
N.D.
(
1993
)
Ferroxidase kinetics of human liver apoferritin, recombinant H-chain apoferritin, and site-directed mutants
.
Biochemistry
32
,
9362
9369
41
Kiel
,
S.
and
Dando
,
P.R.
(
2009
)
Chaetopterid tubes from vent and seep sites: implications for fossil record and evolutionary history of vent and seep annelids
.
Acta Palaeontol. Pol.
54
,
443
448
42
Shah
,
D.U.
,
Vollrath
,
F.
,
Stires
,
J.
and
Deheyn
,
D.D.
(
2015
)
The biocomposite tube of a chaetopterid marine worm constructed with highly-controlled orientation of nanofilaments
.
Mater. Sci. Eng. C
48
,
408
415
43
Huard
,
D.J.E.
,
Kane
,
K.M.
and
Tezcan
,
F.A.
(
2013
)
Re-engineering protein interfaces yields copper-inducible ferritin cage assembly
.
Nat. Chem. Biol.
9
,
169
176
44
Erel
,
O.
(
1998
)
Automated measurement of serum ferroxidase activity
.
Clin. Chem.
44
,
2313
2319
PMID:
[PubMed]
45
Viollier
,
E.
,
Inglett
,
P.W.
,
Hunter
,
K.
,
Roychoudhury
,
A.N.
and
Van Cappellen
,
P.
(
2000
)
The ferrozine method revisited: Fe(II)/Fe(III) determination in natural waters
.
Appl. Geochem.
15
,
785
790
46
Liu
,
X.
and
Theil
,
E.C.
(
2004
)
Ferritin reactions: direct identification of the site for the diferric peroxide reaction intermediate
.
Proc. Natl Acad. Sci. U.S.A.
101
,
8557
8562
47
Ebrahimi
,
K.H.
,
Bill
,
E.
,
Hagedoorn
,
P.-L.
and
Hagen
,
W.R.
(
2013
)
The catalytic center of ferritin regulates iron storage via Fe(II)-Fe(III) displacement
.
Nat. Chem. Biol.
9
,
406
48
Theil
,
E.C.
and
Behera
,
R.K.
(
2013
) The chemistry of nature's iron biominerals in ferritin protein nanocages. In
Coordination Chemistry in Protein Cages: Principles, Design, and Applications
,
1st edn
(
Ueno
,
T.
and
Watanabe
,
Y.
, eds).
John Wiley & Sons, Inc.
,
Hoboken, NJ
49
Schwartz
,
J.K.
,
Liu
,
X.S.
,
Tosha
,
T.
,
Theil
,
E.C.
and
Solomon
,
E.I.
(
2008
)
Spectroscopic definition of the ferroxidase site in M ferritin: comparison of binuclear substrate vs cofactor active sites
.
J. Am. Chem. Soc.
130
,
9441
9450
50
Lawson
,
D.M.
,
Artymiuk
,
P.J.
,
Yewdall
,
S.J.
,
Smith
,
J.M.A.
,
Livingstone
,
J.C.
,
Treffry
,
A.
et al. 
(
1991
)
Solving the structure of human H ferritin by genetically engineering intermolecular crystal contacts
.
Nature
349
,
541
544
51
Battye
,
T.G.G.
,
Kontogiannis
,
L.
,
Johnson
,
O.
,
Powell
,
H.R.
and
Leslie
,
A.G.W.
(
2011
)
iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM
.
Acta Crystallogr. D Biol. Crystallogr.
D67
,
271
281
52
Evans
,
P.R.
and
Murshudov
,
G.N.
(
2013
)
How good are my data and what is the resolution?
Acta Crystallogr. D Biol. Crystallogr.
D69
,
1204
1214
53
McCoy
,
A.J.
,
Grosse-Kunstleve
,
R.W.
,
Adams
,
P.D.
,
Winn
,
M.D.
,
Storoni
,
L.C.
and
Read
,
R.J.
(
2007
)
Phaser crystallographic software
.
J. Appl. Crystallogr.
40
,
658
674
54
Zhang
,
Y.
(
2008
)
I-TASSER server for protein 3D structure prediction
.
BMC Bioinf
9
,
40
55
Adams
,
P.D.
,
Afonine
,
P.V.
,
Bunkóczi
,
G.
,
Chen
,
V.B.
,
Davis
,
I.W.
,
Echols
,
N.
et al. 
(
2010
)
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
213
221
56
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Coot: model-building tools for molecular graphics
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
486
501
57
The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC
.
58
Meldrum
,
F.C.
,
Douglas
,
T.
,
Levi
,
S.
,
Arosio
,
P.
and
Mann
,
S.
(
1995
)
Reconstitution of manganese oxide cores in horse spleen and recombinant ferritins
.
J. Inorg. Biochem.
58
,
59
68
59
Bou-Abdallah
,
F.
,
Arosio
,
P.
,
Santambrogio
,
P.
,
Yang
,
X.
,
Janus-Chandler
,
C.
and
Chasteen
,
N.D.
(
2002
)
Ferrous ion binding to recombinant human H-chain ferritin. An isothermal titration calorimetry study
.
Biochemistry
41
,
11184
11191
60
Toussaint
,
L.
,
Bertrand
,
L.
,
Hue
,
L.
,
Crichton
,
R.R.
and
Declercq
,
J.-P.
(
2007
)
High-resolution X-ray structures of human apoferritin H-chain mutants correlated with their activity and metal-binding sites
.
J. Mol. Biol.
365
,
440
452
61
Masuda
,
T.
,
Goto
,
F.
,
Yoshihara
,
T.
and
Mikami
,
B.
(
2010
)
Crystal structure of plant ferritin reveals a novel metal binding site that functions as a transit site for metal transfer in ferritin
.
J. Biol. Chem.
285
,
4049
4059