Formation mechanism of iron–catechol complexes in the colored periostracum of Corbicula spp.

Abstract

The shell color of Corbicula clams, which are globally distributed, is roughly divided into yellowish and blackish depending on the environmental conditions of the sediment. The formation of an iron–L-3,4-dihydroxyphenylalanine (DOPA) complex in a thin organic layer, called the periostracum, on a calcareous layer causes the blackening of the clamshell. However, the iron–DOPA complex formation mechanism is unclear. To reveal how the iron is transported from the aquatic environment to the periostracum, cross-sectional analyses of the shell were conducted using an electron probe microanalyzer and Raman spectroscopy to investigate the distribution and structure of iron in the shell. Iron was only present in the periostracum, excluding deposition, and all iron was in the form of an iron–DOPA complex. Attenuated total reflection infrared spectroscopy and oxygen K-edge X-ray absorption fine structure spectroscopy revealed that the molecular structure of the native periostracum is independent of shell color. These results indicate that dissolved iron–organic complexes diffuse from the aqueous environment to the periostracum, forming iron–DOPA complex through ligand exchange. Because the iron–DOPA complex color depends on the pH, the shell color can serve as a historical indicator of the shell's growth environment.

Graphical Abstract

Open in new tabDownload slide

Introduction

Corbicula spp. are distributed worldwide, including native and invaded ranges. Japan has three native species of Corbicula spp.: C. japonica, C. leana, and C. sandai, which inhabit brackish water, freshwater, and the Lake Biwa–Yodo River system, respectively.

Shell color and pattern are nearly constant within species, with each species showing distinctive colors and patterns. However, diet and habitat can affect color and pattern [1]. Two predominant hues are associated with the shells of Corbicula spp.: yellowish and blackish (Fig. 1A1 and B1). Yellowish and blackish clams grow in sandy and sandy-muddy sediments, respectively [2–5]. Thus, the shell color depends on the environmental conditions of the sediment.

A clamshell contains a thin organic layer called the periostracum and calcareous layers composed of calcium carbonate (CaCO₃). Proteins are the principal constituents of the periostracum, and macromolecules are highly polymerized via a process that involves L-3,4-dihydroxyphenylalanine (DOPA)-containing proteins in bivalves [6]. In our previous paper, Raman spectroscopy and X-ray absorption fine structure (XAFS) spectroscopy revealed the formation of an iron–DOPA complex in the blackish periostracum of Corbicula spp. [7]. The coordination number of Fe–O was from five to six. The darker the shell color, the more abundant the tris-ferric–DOPA complex. The native membrane was yellow (Fig. 1A2 and B2), and the complex was purple [8]. Therefore, the periostracum was dark and blackish due to the superposition of complementary colors during the formation of the iron–DOPA complex and not the presence of black substances such as iron sulfide and melanin [7].

The periostracum is a substrate for the growth of CaCO₃ crystals and a sealing membrane for extrapallial fluid [9, 10]. It is possible to create complex shell morphology and ornamentation with fine and flexible periostracum [11]. The periostracum protects the shell from dissolution [11–13], acts as a chemical defense against fouling [14], and camouflages against predators [15, 16].

The periostracum is a complex structure, and marine and freshwater bivalves typically contain a layered periostracum [9, 17]. The most supported model of periostracum formation is as follows: a thin membrane (10–50 nm), called the pellicle, is initially secreted from the basal cells at the base of the fold [16, 18]. Subsequently, secretions from epithelial cells at the inner surface of the outer mantle fold are progressively added beneath the pellicle [16]. When the periostracum reaches sufficient thickness, the membrane acts as a substrate for the growth of CaCO₃ crystals.

The iron–DOPA complex we observed is formed when a ferric ion supplied by natural water binds the DOPA residues of proteins in the periostracum. However, the iron–DOPA complex formation mechanism is unclear.

The aim of this study was to elucidate how the iron–DOPA complex is formed in the periostracum. Cross-sectional analysis of the shell was conducted using an electron probe microanalyzer (EPMA) and Raman spectroscopy to investigate the distribution of iron and the iron–DOPA complex. The molecular structure of the native periostracum was analyzed using attenuated total reflection infrared (ATR-IR) spectroscopy and oxygen (O) K-edge XAFS spectroscopy. These results showed that dissolved iron–organic complexes are transported to the periostracum by diffusion from the aqueous environment, and the iron–DOPA complex is formed through ligand exchange.

Materials and methods

Sample preparation

C. japonica was collected from Lake Shinji, Shimane and purchased at a market in Osaka, Japan. C. sandai was collected from Lake Biwa and purchased at a local market in Shiga, Japan. C. leana was collected from the Seta River, the only river flowing out of Lake Biwa, and purchased at the Seta-Cho Fisheries Cooperative Association in Shiga, Japan. Shells containing both yellow and black were selected for EPMA and Raman spectroscopy. Blackish and yellowish shells were selected for other analyses. After the soft body was removed, the shells were air-dried and stored in a desiccator until measurements were taken.

Cross-sectional samples were prepared for EPMA and Raman spectroscopy. The shells were embedded in acrylic resin and sectioned along the maximum growth axis from the umbo to the ventral margin. The cross-sectional surface was then mirror-polished. Carbon vapor deposition was performed on the surface to prepare EPMA samples. After EPMA, the cross-sectional surface was repolished for Raman spectroscopy.

Two types of periostracum membranes were prepared for the IR analysis. The first and second types were obtained by acid treatment of the entire shell and only the calcareous layer, respectively. Acid treatment was performed using 6 N HCl. When only the calcareous layer was acid-treated, the surface of the shell at the periostracum side was covered with Parafilm. After treatment, the periostracum membranes were rinsed with distilled water and air-dried.

DOPA was purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan) as the standard for XAFS spectroscopy.

EPMA

The distributions of iron, manganese, and calcium elements (elemental mapping) were recorded over the cross-sections of the shells. The distribution of iron and manganese was detected using EPMA. The calcium distribution was detected using an energy dispersive spectrometer. EPMA was performed using a field emission EPMA (JXA-8530F, JEOL Ltd, Japan) with scanning electron microscopy (SEM) and wavelength dispersive spectroscopy. The acceleration voltage was 15 kV, the sample current was 30 nA, and the electron beam size was 1.0 μm.

Micro Raman spectroscopy

Micro Raman spectra were acquired using a Renishaw inVia Raman spectrometer (Renishaw, Gloucestershire, UK). The light source was a 532 nm semiconductor laser with a power of 0.3 mW, 1800 line/mm grating, and 1.0 µm spot-size.

ATR-IR spectroscopy

ATR FT-IR spectra were collected on a Spectrum 65 FT-IR spectrometer (PerkinElmer Co., Ltd, MA, USA) equipped with a PerkinElmer universal ATR sampling accessory with a diamond crystal at 1 cm⁻¹ resolution.

XAFS spectroscopy

Oxygen K-edge (O K-edge) XAFS spectroscopy was performed on a soft X-ray grating beamline BL-11 at the SR Center, Ritsumeikan University, Shiga, Japan. Spectral data were collected in the partial electron yield mode. At the sample position, the beam size was approximately 2 mm (horizontal) × 3 mm (vertical). The Athena software was used to analyze the XANES data [19].

Results

Microscopy and EPMA

Figure 2 shows light microscopy images of a cross-section of C. japonica, C. leana, and C. sandai shells containing both yellow (Y) and black (B) parts. The periostracum was 10–20-μm thick and tended to thin toward the umbo. The yellow part of C. leana contained an approximately 10-μm thick deposit on the surface. All shell sections exhibited intense coloration in the calcareous layers, which comprise the inner and outer shell layers. In C. leana and C. sandai, the outer shell layers were purple along the growth line. In C. japonica, the outer shell layer was yellow along the growth line.

We analyzed the yellow (Y) and black (B) parts using SEM and EPMA (Fig. 2). It has been confirmed that aquatic organisms, such as mussels, combine the vanadium they take from their environment with catechol [20, 21]. As studied in our previous experiments, except for base elements, iron, calcium, silicon, and manganese were detected, but no vanadium [22]. Since iron was dominantly populated, we focused on iron as the primary target element and manganese as the second. In all Corbicula spp., the periostracum had a layered structure consisting of dense and sponge-like zones (Fig. 3), consistent with that of other bivalves [9, 17]. Sponge-like zones were clearly seen in the periostracum of C. leana from the black part and in C. japonica from the black and yellow parts. The interface between the calcareous layer and the periostracum of C. japonica was the roughest of the three Corbicula spp. The distribution of iron in the periostracum was color-dependent. In the periostracum of C. leana, iron accumulated near the surface and bottom of the black part (1B in Fig. 3). In contrast, iron was not detected in the yellow part (1Y in Fig. 3). The iron layer was thicker near the surface than at the bottom. In the periostracum of C. japonica and C. sandai, iron accumulated near the surfaces of the yellow (2Y and 3Y) and black (2B and 3B) parts (Fig. 3). The black part had a thicker iron layer and a higher iron concentration than the yellow part. Large amounts of iron and manganese were found on the deposit on the surface of the yellow part of C. leana (1Y in Fig. 3). Iron was not detected in the calcareous layer of any analyzed part of all three Corbicula spp. Manganese did not accumulate explicitly in the periostracum. Its concentration was significantly lower than iron's and at almost background levels in all three Corbicula spp.

Micro Raman spectroscopy

Micro Raman spectrum was recorded for black C. japonica (Fig. 4). This is a typical Raman spectrum of black Corbicula spp. The Raman spectra showed a pronounced peak at 1484 cm⁻¹ and structures at approximately 311, 532, 589, 635, 1271, 1323, and 1425 cm⁻¹. Peaks at 1271, 1323, 1425, and 1485 cm⁻¹ can be assigned to the vibrations of the catechol ring [20, 21, 23–26]. Peaks at 311, 532, 589, and 635, cm⁻¹ can be assigned to the interaction between iron and oxygen in catechol [20, 21, 23–26]. These peaks are used as indicators of metal–DOPA complexation [27]. In particular, since iron–DOPA gives rise to a characteristic broad peak near 330 cm⁻¹ in Raman spectra, the peak strengthens the argument that the metal bound to DOPA is not another metal, such as vanadium, but iron [21]. These results show that black C. japonica contains an iron–DOPA complex, which is responsible for the black color of periostracum.

Cross-sectional Raman spectroscopy provided information on the presence of chemical compounds in the shell. The relative Raman spectra collected from the periostracum of C. leana at the black part near the surface (i) and bottom (ii), where high iron accumulation, revealed characteristic signals of iron–DOPA (Fig. 5B (a)). However, the Raman spectrum collected from deposits in the yellow part (1) in Fig. 5Y (a) does not reveal iron–DOPA signals but indicates hematite [28, 29]. After EPMA, the cross-sectional surface was repolished to remove the carbon vapor deposition layer. Owing to the polishing, the cross-section analyzed by Raman spectroscopy is not precisely the same as that observed by SEM. Therefore, it should be noted that the marked areas in the SEM images indicate an approximate Raman analysis position, not a strict one. The marked areas in the image also indicate an approximate Raman analysis position.

CaCO₃ occurs naturally in three crystalline polymorphs: calcite, aragonite, and vaterite. Raman spectra are easily identifiable by their main peak due to the stretching vibration of carbonate and the low-to-medium-intensity bands arising from the translational and rotational modes of lattice vibration [30–33]. Both Raman spectra of the calcareous layer show a prominent peak at 1085 cm⁻¹ and structures at approximately 156, 208, and 703 cm⁻¹ (Fig. 5B (a3) and Y (a2). This spectral feature shows that the calcareous layer comprises aragonite crystals. Two signals were observed at approximately 1520 and 1130 cm⁻¹ in both calcareous layers. These Raman bands corresponded to the stretching vibrations of the C=C (ν1) and CC (ν2) bindings in polyene molecules [34].

Raw Raman spectra shown in Fig. 5B(b), B(c), Y(b), and Y(c) reveal characteristic signals such as iron–DOPA, aragonite, and polyene superimposed on the fluorescence background. In the black part, the fluorescence background was too high to visually identify the Raman spectra 4–7 μm from the surface. Similarly, in the yellow part, at 0–23 μm from the surface, the fluorescence intensity was too high to distinguish the Raman signal. Variations in the specific Raman band intensity (iron–DOPA: 1480 cm⁻¹, aragonite: 1080 cm⁻¹, and polyene: 1520 cm⁻¹) and fluorescence intensity (1850 cm⁻¹) along the cross-section are shown in Fig. 5(d) and Table S1. The iron–DOPA was detected only in the black part. Iron–DOPA distribution in the periostracum agreed with iron distribution. Iron–DOPA was not detected in the calcareous layer. These results agreed with EPMA results. In contrast, polyene was collected from the calcareous layers of both the black and yellow parts.

The trends of the relative and raw Raman spectra in the cross-sectional modes of C. japonica and C. sandai shells were the same as those of the C. leana shell (Figs. 6, 7 and Tables S2, S3).

ATR-IR spectroscopy

Figure 8 shows photographs and ATR-IR spectra of the periostracum membranes of C. leana. The color of the periostracum remained unchanged after removal of the calcareous layer (Fig. 8A). The ATR-IR spectra showed absorption bands at approximately 1640 cm⁻¹ (amide I), 1520 cm⁻¹ (amide II), 1230 cm⁻¹ (amide III), and 3200 cm⁻¹ (amide A). The absorption band at approximately 1020 cm⁻¹ was associated with carbohydrate bond vibration [35]. Additionally, the strong peaks at approximately 1230 and 1520 cm⁻¹ suggest the presence of tyrosine and its derivatives in the periostracum [35, 36]. Acid treatment of the periostracum removed carbohydrates along with iron (Fig. 8B). However, bands of amide I, amide II, amide III, and amide A remained, and the color of the periostracum changed to yellow (Fig. 8B). The absence of characteristic peaks of crystalline CaCO₃ in the periostracum, such as aragonite, in the ATR-IR and Raman spectra was consistent.

XAFS spectroscopy

Figure 9 shows the O K-edge spectra of the periostracum of yellowish and blackish C. leana and the standards. The spectra of the yellowish and blackish shells are similar. The O K-edge XAFS spectra of the shells and DOPA are characterized by a peak at approximately 532 eV (labeled A in Fig. 9) attributed to the 1s→π*(C=O) transition of carboxyl groups, esters, and amides [37, 38]. The peak energy of periostracum was 0.3 eV lower than that of DOPA. Typical ether and hydroxyl groups in DOPA appeared at approximately 534 eV [1s→π*(C−O)] and 535 eV [1s→π*(C−OH)], respectively (labeled as B1 and 2 in Fig. 9) [37–39]. The broad features at approximately 539 eV (labeled as C in Fig. 9) were attributed to the σ*(O−C) transition observed in amino acids containing hydroxyl group, such as serine, threonine, hydroxyproline, and tyrosine [39, 40]. Characteristic σ*(O=C) peaks of CaCO₃, such as aragonite and calcite, were not observed in the spectra [41]. Thus, periostracum does not contain CaCO₃. This result is consistent with the results obtained from EPMA analysis, Raman, and ATR-IR spectroscopy data.

Discussion

ATR-IR spectra showed that the native periostracum comprised proteins and carbohydrates. The color of the membrane remained yellow after the carbohydrates and inorganic substances were removed. Therefore, the yellow color originates from materials based on amino acids. Under wet conditions, the self-supported periostracum membranes were flexible and robust. However, the dry membranes were fragile. This result suggests that the matrices in periostracum have a flexible three-dimensional network structure that can trap water like a hydrogel. Consistent with this, Chi et al. reported that the P. viridis periostracum has a sandwich structure and contains a novel middle fibrous layer, with a thickness expansion ratio of approximately 75% after absorbing water [42].

XAFS spectra indicated that the molecular structure of the main component of the native periostracum is independent of shell color. IR and XAFS spectra indicated that periostracum contains tyrosine and its derivatives. High tyrosinase activity has been observed in the mantle margins of Corbicula spp. [7]. Tyrosine hydroxylase catalyzes the production of DOPA from tyrosine. The absence of CaCO₃ in the periostracum indicates that the secretions of epithelial cells do not contain calcium and carbonate ions, and no substance enters the periostracum from the calcareous layer.

In the presence of oxygen, the catechol groups of DOPA are oxidized and transformed into catechol quinones, which are highly reactive and covalently crosslink to form a solidified structure. This nature of the catechol leads to the hypothesis that the native periostracum formation by crosslinking is not simultaneous with the secretion of the liquid membrane on the pellicle. Instead, oxic water diffusion gradually crosslinks the liquid membrane. Initially, an acid solution containing catechol groups is secreted into the natural water. The effective pH for the onset of autoxidation of the catechol group is approximately 5, but the reaction rate increases markedly at pH >8 [43]. When the natural water environment is basic (pH ∼8) and dissolved oxygen is saturated, catechol is immediately oxidized to catechol quinone. This oxidization causes irreversible crosslinking, leading to the formation of the initial periostracum membrane (Fig. 10A). After the initial thin membrane (pellicle) is formed, the secretion containing soluble proteins is supplied on the pellicle (Fig. 10B). Some catechol groups within the acidic liquid membrane are oxidized to catechol quinone by diffusion of the natural water with dissolved oxygen, and protein crosslinking proceeds (Fig. 10, reaction I). Crosslinking continues until no more catechol quinone is present.

Two ways are suggested to transport iron from the aquatic environment to periostracum through (i) metabolites and (ii) diffusion. Iron was absent in the calcareous layer. The iron–DOPA complex tended to accumulate near the periostracum surface, which is constantly exposed to water. These observations show that iron is not transported to the periostracum by food and water through the tegument but by diffusion directly from the aquatic environment, similar to dissolved oxygen. Iron uptake through diffusion explains why young shells appear yellowish and mature shells turn shiny black [44]. Iron exists in ferric form in oxic natural water, and its solubility is dependent on pH. Because the solubility is very low at anything but acidic pH [8], the majority of dissolved iron exists as an iron complex bonded to organic matter such as humic substances [45–48]. Therefore, it is reasonable to assume that iron is transported to the periostracum as a dissolved iron–organic complex. Because the catechol group is capable of chelating metal ions, the iron–DOPA complex may have formed in the periostracum by ligand exchange between dissolved iron–organic complexes and the catechol group (Fig. 10, reactions II and III). Interestingly, these colored complexes were observed near the surface of the periostracum, as well as at the bottom of the periostracum (Fig. 31B). This result suggests that the formation of the iron–DOPA complex can also start immediately after the secretion solution is supplied if the conditions are adequate. In addition, when deposited on the periostracum, such as iron oxide, prevents the diffusion of dissolved iron, an iron–DOPA complex does not form (Fig. 31Y).

Catechol exhibits pH-dependent iron-chelating properties. The pH required to establish the bis- and tris-DOPA complexes is typically reported to be above pH 7 [23]. At pH <5.5, the mono-iron–DOPA complex with an absorption maximum around 760 nm dominates. The bis-iron complex with an absorption maximum around 580 nm dominates in the 5.5 < pH < 9 range. At pH >8, the formation of the tris-iron complex with an absorption maximum around 490 nm increases rapidly [8]. Color variations in the shell depend on the coordination states and quantity of the iron–DOPA complex. Thus, the pH of water in which Corbicula spp. live may be a significant factor in determining shell color. When the concentration of the dissolved iron–organic complex is high, Corbicula spp. with black shells live in sediment with interstitial water pH >8 (Fig.10, reaction III), whereas those with shells of other colors, such as dark red and brown, live in sediment with interstitial water of basic pH (<8) (Fig.10, reaction II). However, when the concentration of the dissolved iron–organic complex is low, the shell color remains yellow even when Corbicula spp. live in water with pH >8 (Fig.10, reaction I). Generally, the total organic carbon content in the sediment increases with the increasing concentrations of finer particles. Refractory low molecular weight dissolved organic carbon preferentially accumulates in sediment pore waters [49]. The notion is consistent with the observation that yellowish and blackish clams tend to grow in sandy and sandy-muddy sediments, respectively [2–5]. These results suggest that the shell color can indicate the history of the shell's growth environment.

Polyenes were observed in the colored calcareous layer. Polyenes are polyunsaturated organic compounds that exhibit pigmentation. Polyenes have been detected in the colored shells of several mollusks [35, 50, 51]. We used the method described by Schaffer et al. to identify approximately 10 C−C single bonds and C=C double bonds in the polyenes of the calcareous layer [52]. There was no difference in the number of bonds among the different colored calcareous layers. It is suggested that the pigments in the calcareous layer comprise polyenes. Color variation may be due to different functional groups of pigment.

The fluorescence background was too high to identify Raman spectra within the native periostracum. Note that this does not indicate the absence of polyene and other pigments in the native periostracum. In addition, hematite, known as orange/red rust, deposited on the surface of the shell affects the apparent color. However, it is not a coloring substance of the periostracum.

Conclusion

The iron–DOPA complexes formed in the yellow periostracum cause the blackening of Corbicula spp. The binding of iron to the DOPA residues of proteins in the periostracum results in the formation of the iron–DOPA complex. To elucidate the formation mechanism of the iron–DOPA complex in the periostracum, we performed cross-sectional EPMA and Raman spectroscopy of the shell, as well as ATR-IR spectroscopy and O K-edge XAFS spectroscopy of the periostracum. Our research shows that the periostracum comprises proteins and carbohydrates, which create a flexible three-dimensional structure that retains water. This structure is crosslinked with tyrosine derivatives through the diffusion of natural water. Iron was only present in the periostracum, excluding deposition. All the iron observed in the periostracum was present as an iron–DOPA complex, especially accumulating near the surface of the periostracum. Our findings show that the iron–DOPA complex is formed by the exchange of ligands of dissolved iron–organic complexes that are transported directly from the aqueous environment to the periostracum. Because the iron–DOPA complex color depends on the pH, the shell color can provide historical information about the environment in which the shell grew. In addition, our study suggests that the Corbicula clams can be helpful environmental indicators for tracking environmental change.

Acknowledgments

The authors thank the Fisheries Cooperative Association of Seta-Cho, Shiga, for providing samples. O K-XAFS experiments were conducted at the SR Center, Ritsumeikan University, Shiga, Japan (proposal no. S21004).

Author contributions

Conceptualization: K.T. Funding: K.T. Developing methods: K.T., M.M., Y.U., and H.Y. Data analysis: K.T., M.M., Y.U., and H.Y. Conducting the research: K.T., M.M., Y.U., and H.Y. Data interpretation K.T., M.M., Y.U., D.B., and H.Y. Preparation of figures and tables: K.T. Writing–original draft: K.T. Writing–reviewing & editing: K.T., M.M., Y.U., D.B., and H.Y.

Conflict of interest

The authors declare no conflict of interest.

Funding

This work was supported by JSPS KAKENHI Grant Number 21K05789 and Kansai Medical University, Molecular Imaging Center of Diseases.

Data availability

Data are available from the authors upon reasonable request.

References