Amino group in Leptothrix sheath skeleton is responsible for direct deposition of Fe(III) minerals onto the sheaths

Leptothrix species produce microtubular organic–inorganic materials that encase the bacterial cells. The skeleton of an immature sheath, consisting of organic exopolymer fibrils of bacterial origin, is formed first, then the sheath becomes encrusted with inorganic material. Functional carboxyl groups of polysaccharides in these fibrils are considered to attract and bind metal cations, including Fe(III) and Fe(III)-mineral phases onto the fibrils, but the detailed mechanism remains elusive. Here we show that NH2 of the amino-sugar-enriched exopolymer fibrils is involved in interactions with abiotically generated Fe(III) minerals. NH2-specific staining of L. cholodnii OUMS1 detected a terminal NH2 on its sheath skeleton. Masking NH2 with specific reagents abrogated deposition of Fe(III) minerals onto fibrils. Fe(III) minerals were adsorbed on chitosan and NH2-coated polystyrene beads but not on cellulose and beads coated with an acetamide group. X-ray photoelectron spectroscopy at the N1s edge revealed that the terminal NH2 of OUMS1 sheaths, chitosan and NH2-coated beads binds to Fe(III)-mineral phases, indicating interaction between the Fe(III) minerals and terminal NH2. Thus, the terminal NH2 in the exopolymer fibrils seems critical for Fe encrustation of Leptothrix sheaths. These insights should inform artificial synthesis of highly reactive NH2-rich polymers for use as absorbents, catalysts and so on.

The immature sheath skeleton of another L. cholodnii strain, OUMS1 (hereafter OUMS1-WT) becomes encrusted with metal cations and/or metal solid phases to eventually form uniquely structured microtubular sheaths comprising an organic-inorganic complex enriched with Fe, Si, P, and Ca [21][22][23] . Since an Fe(II)-oxidzing bacterium oxidizes Fe(II) and uses the generated electron as an energy source in the presence of low concentrations of oxygen [11][12][13] , the sheath probably possesses an ecological role in avoiding encrustation in Fe(III) oxyhydroxides. Fe(III) minerals (~50 nm diameter) that are generated abiotically in the culture medium (=complex of ferric oxyhydroxides as major components and inorganic components of the medium components as minors, Fig. S1) were reported to adhere directly to the sheath materials of SP-6 24 and OUMS1-WT 25 . The Fe(III) minerals also directly adhere to cell-and protein-free sheath remnants, indicating that living cells and their proteins conjugated to sheath fibrils are not inevitably required for the Fe(III) mineral encrustation 24,25 . The metal encrustation of the sheath skeleton is considered to result from the interactions between aquatic phase inorganic cations and the functional groups in the sheath skeleton 12 . Indeed, a strong correlation exists between the presence of acidic polysaccharides with carboxyl groups (COOH) and the distribution of iron oxyhydroxides in Leptothrix sheaths 6,26 , but little is known about the roles of NH 2 within the sheath fibrils on the metal encrustation.
In this study, we sought to ascertain the involvement of the terminal NH 2 within the constitutive molecules of the immature sheath skeleton of OUMS1-WT in the adsorption of Fe(III) minerals to the skeleton by differential interference contrast (DIC) and fluorescent microscopy, scanning, transmission, and scanning-transmission electron microscopy (SEM, TEM, and STEM, respectively), energy-dispersive x-ray microanalysis (EDX), x-ray fluorescence and photoelectron spectroscopy (XRF and XPS, respectively) ( Fig. 1).

Results
Importance of sheath fibrils for Fe encrustation. We examined the ultrastructure of OUMS1-WT grown in silicon-glucose-peptone medium (Table S1) (hereafter SGP) 27 for 2 days and then in SGP + Fe-plate medium (hereafter SGP + Fe plate) for 2 more days for comparison with previous TEM images of the cells incubated in SGP + Fe plate for 3 days 28 . For comparison, a sheathless mutant of OUMS1 (named hereafter OUMS1-SL) 28 was incubated and studied in similar conditions. After 2 days incubation in SGP, chains of OUMS1-WT cells (ca 3-4 μm long) were encased within a thin, immature sheath, while longitudinally extending chains of OUMS1-SL cells (ca 7-8 μm long) were not (Fig. 2a, left). After another 2 days incubation in SGP + Fe plate, a few aggregated fibrils were seen around OUMS1-SL cells, suggesting fibril excretion is enhanced in the presence of the Fe source, but they did not assemble into a sheath (Fig. 2, lower center) as reported previously 28 . Electon-dense particles were deposited on and near the immature sheath encasing OUMS1-WT cells (Fig. 2a, upper center) and on aggregated fibrils far from OUMS1-SL cells (Fig. 2a, lower center). To confirm the absence of sheaths in OUMS1-SL, the culturing period of both isolates in SGP + Fe plate was extended to 14 days. The OUMS1-WT cells were evidently encased within thick sheaths having lots of hairy fibrils extending outward onto which electron-dense particles deposited (Fig. 2a, upper right). Aggregated fibrils 10-150 μm from the OUMS1-SL cell became electron-dense but never assembled into a sheath (Fig. 2a, lower right).
SEM/EDX analyses of OUMS1-WT cells incubated in SGP + Fe plate as above detected a mass of aggregated sheaths and apparent deposition of Si, P, and Fe on sheaths (Fig. 2b), while chained OUMS1-SL cells aggregated without being encased in sheaths and lacked any distinguishable element depositon, even of Fe (Fig. 2c).
From these microscopic results, we confirmed that (i) OUMS1-WT formed immature thin sheaths within 2 days in SGP, while exopolymer fibrils were not excreted from OUMS1-SL cells, and no sheath formed; (ii) OUMS1-SL cells excreted fibrils but failed to form sheaths even after an additional 2 days incubation in SGP + Fe; and (iii) inorganics including Fe were deposited on OUMS1-WT sheaths within another 2 days in SGP + Fe plate, while Fe was not deposited on chained OUMS1-SL cells, which lacked a sheath. NH 2 functional groups in the amino sugars of the OUMS1-WT sheaths play a role in Fe(III) mineral deposition. We examined whether amino sugars were components of OUMS1-WT sheaths as found for SP-6 15, 16 by using protein-free sheath remnants that had been chemically prepared from immature sheaths of OUMS1-WT cultured in SGP for 2 days (Fig. 3a). The GC/MS analysis revealed that GalN and GlcN were major saccharic components of the sheath fibrils of OUMS1-WT, as NH 2 -holding materials in addition to amino acids, similar to those in the L. cholodnii SP-6 sheath fibrils 15,16 .   Whitish, immature sheaths encasing OUMS1-WT cells that had been incubated in SGP for 2 days were treated with allophycocyanin (APC)-conjugated NH 2 -reactive reagent (hereafter APC-reagent) (Fig. 3b). Within 3 h, the OUMS1-WT sheaths turned visibly blue and, viewed with fluorescence microscopy, fluoresced an intense red (Fig. 3c, lower left), demonstrating the presence of terminal NH 2 in the sheaths. In comparison, cells of OUM1-SL, incubated in SGP and similarly treated with the APC-reagent, remained white and did not fluoresce (Fig. 3c, lower right). The lack of excretion of exopolymer fibrils from OUMS1-SL cells within 2 days incubation in SGP (Fig. 2a) accounted for the lack of a positive response (change in color) or any notable deposition of elements such as Fe (Fig. 2c). Subsequently, the APC-reagent-treated OUMS1-WT sheaths were incubated with a suspension of abiotic Fe(III) minerals in SGP [hereafter Fe(III)SGP] to examine the influence of APC-masked NH 2 in the sheath skeleton on Fe(III) mineral encrustation (Fig. 3b). XRF analysis of washed and freeze-dried untreated control sheaths revealed that the ratio of Fe was nearly 50 times higher in Fe(III)SGP relative to an atomic (At) % of Fe = 1 in Fe-free SGP (Fig. 3d). However, in the APC-reagent-treated sheaths, the relative ratio was enhanced approximately 10-fold more than after the SGP incubation (Fig. 3d). Similarly, the increase in Fe percentage after the Fe(III)SGP incubation was suppressed by two other NH 2 -blocking reagents, sulfo-NHS-acetate or acetic anhydride (Figs 3d and S2a,b). Electron-dense particles were apparently deposited around the immature control sheaths after incubation in Fe(III)SGP (Fig. 3e, left two), while such particles were not seen on or around APC-reagent-, sulfo-NHS-acetate-, or acetic anhydride-treated OUMS1-WT cells (Fig. 3e, right two, S2a,b). To confirm whether the deposited electron-dense particles corresponded to Fe minerals, we incubated the SGP-precultured OUMS1-WT cells with Fe(III)SGP for 2 days. EDX analyses of cross-sectioned cells (Fig. 4a,b) showed that the electron-dense particles on or around the exopolymer fibrils were composed of Fe, bound to the medium components such as P and S, as judged by the similar distribution of Fe/O and Fe/P/S (Fig. 4c). The XRD analysis provided evidence that the Fe detected by EDX was actually Fe(III) oxyhydroxide (Fig. S1a-c).

Binding energy shift on the sheath surface caused by incubation with Fe(III)SGP. Since XPS
can determine the intrinsic binding energy of the atomic orbital, which shifts chemically with the surroundings of the atom, the electronic state of the material surface was analyzed using XPS 29 . From the XPS measurement of SGP-incubated OUMS1-WT sheaths, peaks of O1s (532 eV), N1s (399 eV), and C1s (284 eV) were detected, suggesting the organic nature of the sheaths (Fig. 5a). Additional peaks of Fe2p (710 and 723 eV) were detected in the sheaths incubated in Fe(III)SGP (Fig. 5b). The N1s peak yielded by Fe(III)SGP-incubated sheaths shifted toward a binding energy higher than that yielded by the SGP-incubated sheaths (Fig. 5c), suggesting that Fe(III) minerals affected the N-related functional group such as NH 2 30, 31 . Binding energy shift on the surface of NH 2 -holding C-polymer and NH 2 -coated polystyrene beads caused by incubation with Fe(III)SGP. The described results led us to consider that NH 2 in the OUMS1-WT sheath skeleton plausibly contributes to encrustation of the skeleton with Fe(III) minerals. We verified this possibility in model experiments to compare the affinity of cellulose (β-glucose polymer) and chitosan (GlcN polymer) and that of NH 2 -coated polystyrene beads and uncoated (plain) beads for Fe(III) minerals. The Fe ratio of chitosan increased drastically after incubation in Fe(III)SGP relative to At% of Fe = 1 in Fe-free SGP, while that of cellulose after the incubation in SGP was comparable to that in Fe(III)SGP (Fig. 6a). SEM and EDX indicated that aggregated Fe(III) minerals were attached to the chitosan fibrils after incubation, while the smooth surfaces of cellulose fibers were unchanged after incubation with SGP and Fe(III)SGP (Fig. 6b,c). In XPS, peaks of photoelectron O1s and C1s were detected for cellulose, while an additional N1s peak was detected for chitosan ( Fig. 6d-g). The incubation in Fe(III)SGP yielded additional peaks of Fe2p for chitosan, but not for cellulose ( Fig. 6f,g), suggesting the possible binding of Fe(III) minerals to NH 2 of chitosan. Similar to the results for OUMS1-WT sheaths, the binding energy of N1s in chitosan shifted to a higher level (Fig. 6h), suggesting that NH 2 of chitosan was influenced by Fe(III) minerals. To further confirm the involvement of NH 2 with the Fe encrustation, we examined the affinity of NH 2 -coated polystyrene beads (hereafter NH 2 beads) and uncoated beads (hereafter plain beads) for Fe(III) minerals. When NH 2 beads were incubated with Fe(III)SGP, a precipitate formed within 30 min, while the suspension of plain beads remained turbid (Fig. 7a). SEM revealed that the surfaces of the NH 2 -beads after Fe(III)SGP incubation were heavily coated with granular or rod-shaped Fe particles in contrast to the smooth surface of the plain and NH 2 beads after SGP incubation (Figs 7b and S3c). The Fe atomic percentage determined by XRF was ~6.0 for the NH 2 -coated beads and ~1.3 for the plain after incubation in Fe(III)SGP (Fig. 7b). XPS detected photoelectron C1s and O1s peaks derived from polystyrene in plain beads and an additional N1s peak from the NH 2 beads (Fig. 7c,e). The peak patterns from the plain beads in SGP and in Fe(III)SGP were comparable (Fig. 7c,d). In contrast, additional Fe2p peaks were detected in NH 2 beads after incubation in Fe(III)SGP (Fig. 7e,f). Notably, incubation of NH 2 beads in Fe(III)SGP shifted the N1s peak toward a higher binding energy (Fig. 7g), again suggesting an interaction between Fe(III) minerals and NH 2 . When acetylated NH 2 beads (NH-Ac beads) were incubated with Fe(III)SGP, lack of Fe attachment to these beads was confirmed by SEM, EDX, and XPS (Fig. S3b-e).

Discussion
The behavior of dissolved metals in natural bodies of water is strongly influenced by particular inorganic and organic materials 32 , suggesting complex interactions of various metal-complexing agents in aquatic systems with microbes and/or their constituent polymers. Generally, extracellular polymeric substances (EPS) of bacteria contain carbohydrates, proteins, lipids, extracellular DNA, and humic substances 33 , which possess various types of terminal functional groups in the molecules. The exopolymer fibrils of the Leptothrix sheaths are made of EPS 7,15,21,22 , as such as they can also bind metals. The Leptothrix sheaths have a high affinity for a variety of metal cations such as Fe, Zn, and Pb 6, 21-25, 27, 28, 34-36 . Negatively charged functional groups of saccharic and proteinaceous materials in the Leptothrix sheaths and Gallionella stalks contribute to attracting dissolved cations 6,7 . Several lines of evidence presented here indicate that NH 2 in amino sugars and amino acids in the Leptothrix sheath is involved at least in direct deposition of Fe(III) minerals onto the sheath skeleton, which is composed of organic exofibrils. The Fe(III) minerals generated in this study are a complex of Fe oxyhydroxides 37 and light elements such as P, S, K, Ca, and Cl from the surrounding medium (Fig. S1). Fe oxyhydroxides tend to be negatively charged below pH 7.0 38 . In addition, binding of Fe oxyhydroxides with the added P was proved to cause a more negative surface charge 39 . Therefore, we infer that the interaction between negatively and positively charged Fe(III) minerals and NH 2 , respectively, could be one of the driving forces for binding the minerals to Leptothrix sheath fibrils at pH 7.0 used in this study, although this assumption should be verified in the near future.
Note that we do not interpret our present results as meaning that NH 2 is the only functional group that interacts with Fe(III) minerals, because masking of NH 2 in sheaths with specific NH 2 -reactive reagents did not completely block the Fe(III) minerals deposition onto the immature sheaths (Fig. 3d), although the reagents appeared to block most of the terminal NH 2 in the sheath skeleton. This incomplete blocking suggests the possible involvement of other functional groups in the deposition. On the basis of earlier reports 6,21 , there is no reason to rule out possible synergistic reactions between NH 2 and other functional groups such as COOH. At present, we infer that polymer-directed mineralization is a general phenomenon that could occur in any system containing NH 2 -enriched polysaccharides originating from an organism.
The energy shift that was detected in the OUMS1-WT sheaths, chitosan, and NH 2 beads incubated with Fe(III) minerals (Figs 5-7) could be due to the influence of Fe(III) minerals as they approach the N-atomic orbital, which could readily shift chemically in the atom 29 . For example, the electron density around N is reduced when the covalent bond forms between NH 2 and Fe(III) minerals or from the intense attractive force of charged particles 30,31 . On the basis of earlier reports 30, 37 , we infer that in the aqueous phase the positive charge from NH 2 and the negative charge from ferric oxyhydroxides may interact and that the chemical shift of N2p may reflect  Mineral precipitation could be an important determinant of microbial activity levels in the environment, and thus, spatially resolved analyses of organic compound combined with high-resolution mineralogical information should enhance our understanding of biomineralization mechanisms and promote the development of templated models to use in fabricating new materials 15 . Detailed information on the synthetic mechanism for biological organic-inorganic complexes will help expand the use of these ingenious complexes to benefit life. A variety of industrial applications such as lithium-ion battery anode material, catalyst enhancer, and pigment have been developed for metal-encrusted Leptothrix sheaths 7-10 . Current information on the roles of NH 2 in Fe encrustation of the Leptothrix sheaths will greatly help in harnessing biomineralization mechanisms to create novel functional (e,f) XPS spectra of NH 2 beads incubated in SGP and Fe(III)SGP, respectively, detecting O1s, C1s, and N1s with SGP incubation and additional Fe2p peaks with Fe(III)SGP. (g) N1s peak shifted toward higher binding energy after incubation in Fe(III)SGP. materials 26 . Toward creating novel materials, we should examine whether NH 2 -blocking affects the potential of other functional groups to bind inorganics including Fe, how NH 2 -blocking influences previously reported the functioning of sheath fibrils, and how we can artificially synthesize NH 2 -rich polymers with high reactivity to inorganics for use as adsorbents and/or catalysts.
Organic polymers play important roles in ecosystems by adsorbing biologically important elements 26,40 . Fe cycling, mediated by microbiological oxido-reduction of the element, is a vital environmental process, both on the micro and global scales, and iron-oxidizing bacteria such as Leptothrix are significant facilitators of ferric iron reduction in ecosystems 6 . A step-by-step approach such as the present study should improve our understanding of the complex systems of Fe circulation.

Methods
Strains, medium, and culturing. Leptothrix cholodnii strain OUMS1 (NITE BP-860) (hereafter OUMS1-WT) and its sheathless mutant (OUMS1-SL) were used 27,28 . Cells of OUMS1-WT and OUMS1-SL, obtained from frozen stock cultures, were independently streaked onto SGP agar plates 27 and incubated at 20 °C for 7 days. Single colonies were transferred separately to 25 ml of SGP in 200 ml aluminum foil-capped Erlenmeyer flasks and cultured on a rotary shaker at 20 °C and 70 rpm. After 2 days, 1-5 ml of the cell suspension (adjusted to 10 cfu ml −1 by densitometry) was used to inoculate 25 ml of (i) SGP, and (ii) SGP containing 500 μM FeSO 4 (SGP + FeSO 4 ), or (iii) 100 ml SGP containing three small pieces of Fe plate (SGP + Fe plate). Samples were examined after 2 days of incubation. Viability of OUMS1-WT in these culture conditions was confirmed by live/ dead staining 41 after a 4-day culture (Fig. S1d). Specimens incubated in SGP + Fe plate were used for preliminary microscopic observations, and those in SGP or SGP + FeSO 4 were mainly used for the following affinity tests and electron microscopy, unless otherwise stated.

Abiotic preparation of Fe(III) minerals in SGP.
Our previous work provided evidence that in a shaken solution of SGP + FeSO 4 , Fe(II) was oxidized to Fe(III) almost completely within 6 h 24 . In the present study, uninoculated SGP + FeSO 4 was shaken similarly for 8-12 h to ensure generation of abiotic Fe(III) minerals. The suspension of Fe(III) minerals in SGP [hereafter Fe(III)SGP] was used for the affinity tests, as follows.
Determination of amino sugar composition in protein-free OUMS1-WT sheath remnants. Protein-free sheath remnants were prepared according to a previous protocol 24,25 . OUMS1-WT cells, pelletized from 4-6 l of a 2-day-old culture in SGP, were washed once with ultrapure water (H 2 O) and resuspended in 22.5 ml of the lysis solution containing 2.5 mM EDTA and 150-500 μg lysozyme (Sigma-Aldrich), and incubated at 37 °C for 0.5-2 h. Then 2.5 ml of 10 % w v −1 SDS was added, and the suspension was shaken for 0.5-2 h at room temperature. Thus-prepared sheath remnants were washed twice with H 2 O, then treated with proteinase K (50-100 μg ml −1 ; Nacalai Tesque, Kyoto, Japan) at 37 °C for 12 h. The final protein-free sheath remnants were washed with H 2 O six times and analyzed for sugar composition as follows.
The sheath remnants (20 mg) were vacuum-dehydrated at 50 °C for 2 h, followed by gas phase hydrazinolysis at 110 °C for 1 h with Hydraclub S-204 (J-OI Mills, Tokyo, Japan) to release oligosaccharides from the remnants. The resultant product was dissolved in 1 ml of 0.2 N sodium bicarbonate (NaHCO 3 , pH 6.7), followed by filtration (0.8 μm pore, Advantest, Tokyo, Japan) and subsequent overnight dialysis against 100 mM 0.2 N NaHCO 3 using a dialysis membrane (Molecular Weight Cut-off 12,000-14,000, Wako Pure Chemical, Osaka, Japan). For N-acetylation and demineralization, the dialyzed, freeze-dried sample was combined with 250 μl of 0.2 M ammonium acetate (pH 6.7) and 25 μl of acetic anhydride at room temperature for 30 min. This step was repeated for another 30 min. With the gradual addition of 100 % ethanol, the sample was freeze-dried to eliminate both reagents, then dissolved in 200 μl of H 2 O, demineralized using an acilyzer G0 (Asahi-kasei, Tokyo, Japan), and freeze-drying. The dried specimens were treated in 1 ml of 2 M trifluoroacetic acid at 100 °C for 2 h for acid hydrolysis. Subsequently, the hydrolyzed specimens were dissolved in 200 μl of H 2 O. Ten microliters of the solution was freeze-dried, followed by complete dissolution in 100 μl of pyridine and then combined wtih 20 μl of N,O-bis(trimethylsilyl) trifluoroacetamide (BSTFA) and 2 μl of trimethylchlorosilane (TMCS) at 60 °C for 30 min for trimethylsilylation. The resultant specimens were subjected to GC/MS analysis (Clarus SQ8T, Perkin Elmer, Waltham, MA, USA). The mass spectra were compared with those of a standard sugar mixture (Glc, Gal, and GalA: Wako Pure Chem.; GlcA: Sigma-Aldrich; GlcN: Yaizu Suisankagaku, Shizuoka, Japan). Fluorescent labeling to mask NH 2 in sheath skeletons. The fluorescent protein, allophycocyanin (APC)-conjugated reagent (Dojindo, Kumamoto, Japan) (APC-reagent) was used to confirm the existence of NH 2 in sheath skeletons of OUMS1-WT (Fig. 3b). In parallel, OUMS1-SL, which does not excrete exofibrils within 2 days after incubation in Fe-free SGP, was used as a negative control for exofibril absence. OUMS1-WT and OUMS1-SL, which had been precultured in SGP for 2 days, were separately mixed with this reagent at dilutions from 10 −2 to 10 −3 and incubated at 37 °C for 1-12 h. After four washes in SGP, the specimens were observed microscopically. For monitoring their affinity for Fe(III) minerals, they were incubated with Fe(III)SGP for 2 days before observation. For XRF analysis, the Fe(III)SGP-incubated specimens were washed six times in H 2 O and freeze-dried.
Two other reagents that can mask NH 2 42 were similarly tested: sulfo-NHS-acetate (1-10 mM, Tokyo Kasei, Tokyo, Japan) and acetic anhydride (200 mM, Nacalai Tesque), were used for masking NH 2 in the OUMS1-WT sheath skeleton. OUMS1-WT, precultured in SGP for 2 days, was incubated with either reagent at 37 °C for 1-12 h. After washing with SGP four times, the respective specimens were incubated in Fe(III)SGP for 2 days, then washed six times in H 2 O, freeze-dried, and analyzed by XRF. For TEM/EDX imaging, the incubated specimens were chemically fixed as described below. Freeze-dried sugar chains were similarly attached to carbon tape, as were ethanol-suspended polystyrene beads, then air-dried. Specimens were coated with platinum (ca 15 nm thick) using an ion-sputter (E-1030, Hitachi, Tokyo, Japan) and then observed with an SEM (S-4300, Hitachi) equipped with an energy dispersive X-ray spectrometer (EDX) (Genesis 2000, Amtek-Edax, Berwyn, PA, USA) at 15 kV.
For TEM, the above-treated specimens collected by centrifugation were fixed with a mixture of 2 % v v −1 glutaraldehyde and 2 % v v −1 paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 4 °C overnight. After a 30 min buffer wash, the specimens were embedded in 3 % agar in the buffer. Small pieces of the agar block were post-fixed with 2 % w v −1 OsO 4 in 0.1 M phosphate buffer (pH 7.4) for 1.5 h, then washed with the buffer. Then, the specimens were dehydrated in an increasing series of ethanol solutions (30 %, 50 %, 70 %, 95 %, and 100 %), treated with propylene oxide, then embedded in Spurr's resin. Sections (70-80 nm thick) were cut using a ultramicrotome (Leica, Wetzlar, Germany) equipped with a diamond knife (Ultra 45° 3.0 mm, Diatome, Hartfield, PA, USA) and stained with uranyl acetate and lead solutions and observed with a TEM (JEM-2100F, JEOL, Tokyo, Japan) equipped with EDX at 200 kV. The semiqualitative distribution maps were calculated by the Cliff-Lorimer method 43,44 to confirm their location.

X-ray fluorescence (XRF) analysis.
To determine the atomic percentage of Fe in the test specimens, freeze-dried OUMS1-WT sheaths and sugar polymers and air-dried polystyrene beads were packed into small aluminum pans for elemental analysis with an Orbis micro x-ray fluorescence (XRF) analyzer (Ametek, Berwyn, PA, USA). Atomic percentage of any detected element with a standard error was expressed as the mean (±SE) of 10 spots. X-ray photoelectron spectroscopy (XPS). For obtaining photoelectron spectra from OUMS1-WT sheaths, sugar chain polymers and polystyrene beads, XPS measurements were carried out using monochromatic A1-Kα radiation (hv = 1486.6 eV) as described previously 45,46 . Briefly, specimens were spread on carbon tape (Nisshin EM) and vacuum-dried. In the spectroscopy apparatus, surface charges on specimens were neutralized by using a low energy flood gun (~5 eV). Before evaluating the shifts in binding energy of the N1s in the respective specimens (Figs 5c, 6h and 7g), the detected N1s values in Fe(III)SGP-treated specimens (Figs 5b, 6g and 7f) were compensated on the basis of the value of C1s of the respective SGP-treated specimens (284.3 eV) (Figs 5a, 6e and 7e), because the binding energy levels of all elements varied slightly between Fe(III)SGP-and SGP-treated specimens. The photoelectron O1s spectrum acquired from all specimens was negligible, because it was apparently derived from contaminated carbonate within the x-ray photoelectron spectroscopy vacuum chamber of the XPS device.

X-ray diffraction (XRD) analysis.
To examine the crystallinity of the Fe(III) minerals obtained from SGP + Fe plate or SGP + FeSO 4 , XRD patterns of ethanol-washed and dried Fe(III) minerals were analyzed using an RINT-2500HF X-ray generator (Rigaku, Tokyo, Japan) with Cu-Kα radiation (voltage: 40 kV; current: 200 mA) as described previously 47 . The freeze-dried specimens were fixed on a zero background sample holder and scanned continuously from 10° to 90° (2θ value) at a rate of 3° min −1 . The XRD pattern of the zero background sample holder was also measured.