Advertisement

Calcite Formation in Soft Coral Sclerites Is Determined by a Single Reactive Extracellular Protein*

  • M. Azizur Rahman
    Correspondence
    To whom correspondence should be addressed.
    Affiliations
    Department of Earth and Environmental Sciences, Palaeontology and Geobiology, 80333 Munich, Germany
    Search for articles by this author
  • Tamotsu Oomori
    Affiliations
    Department of Chemistry, Faculty of Science, University of the Ryukyus, 1 Senbaru, Nishihara-cho, Okinawa 903-0213, Japan
    Search for articles by this author
  • Gert Wörheide
    Footnotes
    Affiliations
    Department of Earth and Environmental Sciences, Palaeontology and Geobiology, 80333 Munich, Germany

    Bavarian State Collections of Palaeontology and Geology, 80333 Munich, Germany

    GeoBio-Center, Ludwig-Maximilians-Universität, 80333 Munich, Germany
    Search for articles by this author
  • Author Footnotes
    * This work was supported by grants from the Japan Society for the Promotion of Science (JSPS) and the Alexander von Humboldt Foundation.
    The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1–S6 and Tables S1 and S2.
    2 Supported by the German Science Foundation (DFG).
Open AccessPublished:July 15, 2011DOI:https://doi.org/10.1074/jbc.M109.070185
      Calcium carbonate exists in two main forms, calcite and aragonite, in the skeletons of marine organisms. The primary mineralogy of marine carbonates has changed over the history of the earth depending on the magnesium/calcium ratio in seawater during the periods of the so-called “calcite and aragonite seas.” Organisms that prefer certain mineralogy appear to flourish when their preferred mineralogy is favored by seawater chemistry. However, this rule is not without exceptions. For example, some octocorals produce calcite despite living in an aragonite sea. Here, we address the unresolved question of how organisms such as soft corals are able to form calcitic skeletal elements in an aragonite sea. We show that an extracellular protein called ECMP-67 isolated from soft coral sclerites induces calcite formation in vitro even when the composition of the calcifying solution favors aragonite precipitation. Structural details of both the surface and the interior of single crystals generated upon interaction with ECMP-67 were analyzed with an apertureless-type near-field IR microscope with high spatial resolution. The results show that this protein is the main determining factor for driving the production of calcite instead of aragonite in the biocalcification process and that –OH, secondary structures (e.g. α-helices and amides), and other necessary chemical groups are distributed over the center of the calcite crystals. Using an atomic force microscope, we also explored how this extracellular protein significantly affects the molecular-scale kinetics of crystal formation. We anticipate that a more thorough investigation of the proteinaceous skeleton content of different calcite-producing marine organisms will reveal similar components that determine the mineralogy of the organisms. These findings have significant implications for future models of the crystal structure of calcite in nature.

      Introduction

      The primary mineralogy of marine carbonates has changed over geological history depending on the magnesium/calcium ratio in seawater, including during the so-called “aragonite sea” (
      • Stanley S.M.
      • Hardie L.A.
      ) period and two periods of “calcite seas” (
      • Sandberg P.A.
      ). This ratio is apparently driven by changes in spreading rates along mid-ocean ridges (
      • Stanley S.M.
      • Hardie L.A.
      ). The calcification process of aragonite and calcite mineralogy in mollusk shells (
      • Falini G.
      • Albeck S.
      • Weiner S.
      • Addadi L.
      ,
      • Addadi L.
      • Joester D.
      • Nudelman F.
      • Weiner S.
      ,
      • Thompson J.B.
      • Paloczi G.T.
      • Kindt J.H.
      • Michenfelder M.
      • Smith B.L.
      • Stucky G.
      • Morse D.E.
      • Hansma P.K.
      ,
      • Metzler R.A.
      • Evans J.S.
      • Killian C.E.
      • Zhou D.
      • Churchill T.H.
      • Appathurai N.P.
      • Coppersmith S.N.
      • Gilbert P.U.P.A.
      ,
      • Amos F.F.
      • Destine E.
      • Ponce C.B.
      • Evans J.S.
      ,
      • Belcher A.M.
      • Wu X.H.
      • Christensen R.J.
      • Hansma P.K.
      • Stucky G.D.
      • Morse D.E.
      ) and some information about calcite (
      • Stephenson A.E.
      • DeYoreo J.J.
      • Wu L.
      • Wu K.J.
      • Hoyer J.
      • Dove P.M.
      ,
      • D'Souza S.M.
      • Alexander C.
      • Carr S.W.
      • Waller A.M.
      • Whitcombe M.J.
      • Vulfson E.N.
      ) have been reported, but our knowledge of the mechanism of the direct biological formation of calcite in marine organisms, especially in corals, remains incomplete. To develop a more complete understanding of calcite formation, detailed information concerning how biomolecules contribute to the kinetics of crystal formation, as well as the structural details of both the surface and the interior of single crystals in the submicrometer to nanometer scale must be analyzed. The mechanism of calcite formation in soft coral sclerites that we report here is completely different from the mechanisms used by other calcifying marine organisms, featuring new chemical groups and different types of single crystals in the biomineralization process.
      By investigating the novel functions of extracellular proteins, we have gained new insight into biomineralization strategies in the soft coral endoskeleton. Calcite and aragonite are two polymorphs of CaCO3 that are commonly observed in biominerals. They differ from one other in lattice structure and stability. Stony corals form needle-like aragonite crystals, whereas soft corals form only calcite crystals. An important unresolved question is how soft corals, unlike stony corals, form calcite without forming aragonite. To understand these in vivo mineralization processes, both the extraction fluid and the ionic composition should be considered. It is well known that sea water contains a high concentration of Mg2+ relative to Ca2+ (5:1 ratio). In solution, Mg2+ poisons calcite formation, whereas having only a minimal effect on aragonite precipitation (
      • Davis K.J.
      • Dove P.M.
      • De Yoreo J.J.
      ). During crystallization in vivo, Mg2+ inhibits calcite formation, resulting in a predominance of aragonite crystals. For this reason, stony coral skeletons exclusively form aragonite crystals during calcification. A special case is that of soft corals, which form only calcite crystals. In this study, we explore the interesting, as yet uncharacterized phenomenon of how calcite crystals form in soft corals. We hypothesized that some biological processes, especially those involving matrix proteins (
      • Weiner S.
      • Hood L.
      ,
      • Rahman M.A.
      • Isa Y.
      • Uehara T.
      ) and catalyzed by certain pivotal enzymes (e.g. carbonic anhydrase (
      • Jackson D.J.
      • Macis L.
      • Reitner J.
      • Degnan B.M.
      • Wörheide G.
      )), strongly influence and override the formation of calcite crystals. To gain an understanding of the details of this interesting phenomenon, we have characterized the proteins from a high abundance octocorallian soft coral, Lobophytum crassum, and we found that a reactive extracellular matrix protein (ECMP)
      The abbreviations used are: ECMP
      extracellular matrix protein
      NFIR
      near-field infrared
      AFM
      atomic force microscopy
      XRD
      x-ray diffraction
      EDX
      energy dispersive x-ray
      Tricine
      N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine.
      is mainly responsible for calcite formation in the biocalcification process of soft coral despite the aragonite sea environment. However, the origin of ECMPs in soft corals is unknown. To address this question, we isolated proteins from the calcified endoskeletons (sclerites) of L. crassum (supplemental Fig. S1). Soft corals contain small spicules of calcium carbonate called sclerites, which are biomineralized structures composed of an organic matrix and a mineral (calcite) fraction. Calcite crystals, as shown in this study, are secreted on an organic matrix and then are transported to the outside of the cell for subsequent extracellular calcification in a process similar to sclerite calcification in the gorgonians (
      • Goldberg W.M.
      • Benayahu Y.
      ,
      • Kingsley R.J.
      • Watabe N.
      ). Mature sclerites are completely free of cellular materials and ultimately become extracellular structures.
      Detailed information regarding the function of ECMPs in the kinetics of crystal formation in endoskeletal sclerites has not been reported. One reason is that the purification of proteins from the organic matrix of soft coral sclerites is difficult due to the possibility of contamination by soft tissues and the high sensitivity of these matrices to handling. Key parameters of nucleating biomineralized chemical structures, which are formed by ECMPs, are still uncharacterized due to a lack of effective tools. Vibrational IR spectroscopy is a well established tool in materials and life sciences research; however, obtaining complete chemical information for protein-formed crystals by vibrational spectroscopy at IR wavelengths has been problematic because the diffraction limits of conventional IR microscopy result in low spatial resolution. The spatial resolutions of IR and Raman spectroscopy are ∼10 μm (
      • Kebukawa Y.
      • Nakashima S.
      • Nakamura-Messenger K.
      • Zolensky M.E.
      ) and 1 μm, respectively. Therefore, submicrometer-to-nanometer-scale chemical structures are difficult to analyze by conventional IR spectroscopy. Specifically, scientists have been hampered by the difficulty of separating protein-complexed small molecules from biomineralized organisms. We have resolved this limitation by applying a newly developed near-field optical technique that permits IR spectral mapping below the optical diffraction limit with a spatial resolution of several hundred nanometers (
      • Kebukawa Y.
      • Nakashima S.
      • Nakamura-Messenger K.
      • Zolensky M.E.
      ), a resolution that cannot be achieved by conventional IR spectroscopy. The principles of near-field IR (NFIR) microspectroscopy and a comparison between conventional Fourier transform IR spectroscopy and NFIR microspectroscopy have been recently reported (
      • Koga S.
      • Yakushiji T.
      • Matsuda M.
      • Yamamoto K.
      • Sakai K.
      ), and this study showed that NFIR microspectroscopy is a useful technique for obtaining submicroscale-to-nanoscale chemical information. However, the present study is the first that utilizes NFIR microspectroscopy to analyze mineral-based CaCO3 crystals. NFIR microspectroscopy is a powerful tool for functional group analysis of nanoscale sample surfaces. This non-destructive method allows us to measure the distribution of organic polar functional groups, including hydrogen-bearing groups, together with those of hydrous minerals. In this study, the protein structures and chemical information in nucleated particles were also assessed using this approach.
      We also introduce a new strategy for understanding the surface structure and related parameters of the protein-complexed single crystal kinetics involved in the calcification system. We employ techniques including atomic force microscopy (AFM), Raman spectroscopy, x-ray diffraction (XRD), scanning electron microscopy (SEM), and an electron probe microanalyzer.

      RESULTS

      In this study, we analyzed several aspects of mineral-protein interaction to gain a more complete understanding of the process that determines whether calcite or aragonite is formed. We used several techniques to elucidate the structural details of both the surface and the interior of single calcitic crystals and to examine the interaction of the crystals with sclerite-derived proteins that are thought to be involved in the biocalcification process. We purified four ECMPs from sclerites with apparent molecular masses of 102, 67, 48, and 37 kDa. We named the four proteins ECMP-102, ECMP-67, ECMP-48, and ECMP-37, respectively, due to their molecular sizes (see supplemental Fig. S2) and putative extracellular origin (
      • Weiner S.
      • Hood L.
      ,
      • Rahman M.A.
      • Isa Y.
      • Uehara T.
      ). Among the four proteins, ECMP-102 and ECMP-67 appeared to be calcium-binding proteins as they were detected as radioactive bands by 45Ca autoradiography (
      • Maruyama K.
      • Mikawa T.
      • Ebashi S.
      ) (supplemental Fig. S3). ECMP-102 and ECMP-67 also appeared to be glycosylated (supplemental Fig. S3, lane 3). Interestingly, both of these proteins showed carbonic anhydrase activity (see supplemental Table S1). Supplemental Table S1 shows that both ECMP-102 and ECMP-67 possessed specific carbonic anhydrase activity and that ECMP-67 showed higher activity than ECMP-102. To understand the biomineralization of sclerites, the four proteins were tested singularly, in pairs, and all together. We found complete crystallization in sclerites with the formation of different shapes and phases of single calcite crystals when we introduced either all of the proteins together or ECMP-67 alone. ECMP-67 was the most potent of the four proteins in forming calcite crystals in the biomineralization process. This protein has multiple strong functional properties, including being a Ca2+-binding glycoprotein with enzymatic activity and having an N-terminal amino acid sequence with 44 acidic residues out of the 175 determined (
      • Rahman M.A.
      • Isa Y.
      • Uehara T.
      ). These characteristics are highly favorable for mineral formation in the sea. The primary N-terminal sequence of ECMP-67 was found to be NH2···Asp-Glu-Leu-Asn-Lys-Lys-Val-Asp-Ser-Asp-Glu-Thr-Ile-Ser-Asp-Asp-Gly-Val-Val-Ala···COOH.
      We also analyzed amino acid composition with the extracellular matrix protein (ECMP-67) of sclerites. The composition of the matrix protein was characterized by a predominance of aspartic acid, comprising greater than 37% of all residues. Next in abundance was alanine, making up ∼14%, followed by glycine at about 11%, and then glutamate at about 8%. The total acidic residues (Asp + Glu = 45.44%) comprised almost half of the protein (supplemental Table S2).
      We identified single crystals on the basis of unit cell symmetry and the shape of the observed morphologies. Structural details of both the surface and the interior of single crystals on the submicrometer to nanometer scale were analyzed to gain a complete understanding of the morphology. In the first step, very small single crystals (∼25 μm) growing in a hard endoskeleton were investigated using an apertureless-type NFIR microscope with high spatial resolution. We investigated crystals (∼50% of which were single crystals) that were 25 μm in diameter (on average) to determine whether all had the same structural composition. NFIR spectra from a small single crystal particle, formed with and without the interaction of ECMP-67 in solution, are shown in Fig. 1. As shown in Fig. 1A, in the presence of ECMP-67 (1.4 μg/ml), a strong absorption peak is observed at 3400 cm−1, corresponding to the –OH group of a single crystal; however, this absorption is not observed for small crystals formed in the absence of proteins (Fig. 1B). Two strong bands representing calcite were detected at 1086.9 cm−1 (v1) and 1415 cm−1 (v3) in the presence of ECMP-67 (Fig. 1A). Additionally, the topographic NFIR image of single crystals grown with ECMP-67 clearly showed a rough surface with corresponding different heights for the bulk composition (Fig. 1A). In contrast, when ECMP-67 was absent, a planar surface was evident (Fig. 1B). The characteristic absorption bands of CO3 at 1800 cm−1 were observed in samples both with and without protein induction. However, in the absence of ECMP-67, only a weak signal is observed for the 1800 cm−1 band (Fig. 1B). The structures of the proteins and important chemical groups in the nucleating particles were also assessed. The NFIR spectra show the following important structural protein amide bands in a single crystal (Fig. 1A): amide I (1650–1700 cm−1), amide II (1550–1600 cm−1) and amide III (1300–1350 cm−1). A secondary structure, appearing to be an α-helix (generally associated with absorbance in the range from 1650 to 1658 cm−1 (
      • Dauphin Y.
      )) was observed at 1652 cm−1. The presence of a band arising from an amino acid side chain at 1574 cm−1 indicates an aspartic acid-rich protein based on the shape of amide II (
      • Venyaminov SYu.
      • Kalnin N.N.
      ). Two absorption peaks to the right of the C–H stretch near 2850 and 2750 cm−1 that are caused by the C–H bond that is part of the CHO (aldehyde) functional group were also assigned to this particle. In addition, a band at 2900–2960 cm−1 corresponding to an alkane (–CH) was found. The weak absorption band at 1250–1260 cm−1 may be attributed to sulfate (
      • Prange A.
      • Arzberger I.
      • Engemann C.
      • Modrow H.
      • Schumann O.
      • Trüper H.G.
      • Steudel R.
      • Dahl C.
      • Hormes J.
      ), which is most likely the O-sulfate group. The bands at 1050 and 970–990 cm−1 are indicative of a carbohydrate and a trans-vinylene CH (
      • Kupka T.
      • Lin H.M.
      • Stobinski L.
      • Chen C.H.
      • Liou W.J.
      • Wrzalik R.
      • Flisak Z.
      ), respectively (Fig. 1A).
      Figure thumbnail gr1
      FIGURE 1NFIR analyses of small single crystals obtained by the in vitro biocalcification process. A, an 8 × 8-μm topographic image of a protein-induced small single crystal and its complete bulk composition. The arrow on the topographic image indicates an –OH bond produced by the interaction with ECMP-67. The arrows on the NFIR absorption (Abs) spectrum indicate the complete bulk composition in the particle produced by ECMP-67 (see legend for for details). Two positions on the crystal (blue and green spectra) were analyzed. B, depiction of the results of the near-field IR spectrum of a single particle that was taken from the calcitic solution. The experimental design was the same as shown in A, but the crystal was grown without ECMP-67. In the absence of the protein interaction, only CO3 bands were produced. This figure reveals that the calcifying organisms cannot produce an –OH bond (arrow on topographic image) and other necessary components for biocalcification without protein interactions. Three positions on the crystal were analyzed, as indicated by the red, blue, and green spectra.
      The topographic images of relatively large single crystals (∼150 μm) and their NFIR spectra were investigated so that we might gain a structural understanding of nucleating large crystals and be able to compare this information with that obtained for small single crystals during the biocalcification process. The large single crystal produced in the presence of ECMP-67 (1.4 μg/ml) exhibits a relatively rough surface on which the height of the “mound” was measured to be ∼1.8 μm (see supplemental Fig. S4, A2). Typical CO3 bands in nucleated crystals obtained from an in vitro crystallization system in the presence of ECMP-67 (1.4 μg/ml) were investigated (see supplemental Fig. S4, A3). The topographic image is completely different in the absence of protein (see supplemental Fig. S4, B2). CO3 bands grow in supersaturated solutions in either the presence or the absence of protein; however, in the absence of protein, the bands are weaker (supplemental Fig. S4, B3). Therefore, the topographic view of the surface and the complete bulk composition of nucleating crystals obtained by the NFIR technique could be one of the most effective tools for analyzing the regulation of mineralization in calcified organisms.
      To further study the role of ECMP-67 in mineral formation, in vitro crystallization experiments were conducted with both calcitic and aragonitic solutions. First, the influence of Mg2+ on CaCO3 polymorphism was studied. In the absence of both Mg2+ and protein, typical rhombohedral calcite crystals were generated (Fig. 2, A, B, L, and P). In the presence of Mg2+ but without any protein, large needle-like aragonite crystals were formed (Fig. 2, C, D, M, and Q). When the solution was supplemented with ECMP-67, all of the crystals formed were calcite instead of aragonite (Fig. 2, H–K). Fig. 2, E–K, shows the SEM images of CaCO3 crystals grown in the presence of ECMP-67 at concentrations of 0.7 and 1.4 μg/ml. At a concentration of 0.7 μg/ml, a number of needle-like aragonite crystals remained (Fig. 2, E, arrow, and F), and some rhombohedral and round calcite crystals formed (Fig. 2E, arrowhead, and G). The Raman measurements (Fig. 2N) demonstrate that the former crystals (Fig. 2F) consisted of aragonite, and the latter consisted of calcite (Fig. 2G). Both crystallization mineral phases were confirmed by XRD (Fig. 2R). When ECMP-67 was present at a higher concentration (1.4 μg/ml), the calcite crystals (Fig. 2, H–K) formed different shapes, and no aragonite was observed. Although crystal growth was inhibited at high protein concentrations, all of the remaining aragonites formed by Mg2+ (50 mm) were transformed into calcites (Fig. 2H, enlarged view indicated by arrows in I, J, and K). Raman measurements showed that all single crystals formed under these conditions were calcites at 1086, 720, 288, and 162 cm−1 (Fig. 2O). At least 10 crystals with the same morphology were observed under an optical microscope, and their Raman spectra were individually measured. All crystals were of similar shape and polymorphism. The transition of minerals from aragonites to calcites was confirmed by XRD, which showed that all phases of the crystals were calcites and that no aragonites were present (Fig. 2S). These observations strongly suggest that ECMP-67 is the key component in calcite crystal formation during biocalcification and that an ECMP-67 concentration of 1.4 μg/ml is conducive to growing calcite crystals during the biocalcification process of soft corals. We also utilized commercial bovine serum albumin (BSA) and introduced it into the crystallization solution instead of ECMP-67 to confirm whether any other protein can affect crystal formation, but BSA failed to change the crystal formation from aragonite or calcite (see supplemental Fig. S5).
      Figure thumbnail gr2
      FIGURE 2SEM images of crystals grown in vitro and identification of their polymorphisms by XRD and Raman spectra. A and B, crystals grown without protein and in the absence of Mg2+ show a rhombohedral calcite morphology. C and D, crystals grown without protein and in the presence of Mg2+ appear as needle-like crystals (aragonite). E–G, crystals grown in the presence of protein (0.7 μg/ml) and Mg2+ (50 mm). H–K, crystals grown in the presence of protein (1.4 μg/ml) and Mg2+ (50 mm). L–O, verification of single crystals formed in these experiments by micro-Raman spectra. P–S, verification of the crystal polymorphisms formed in these experiments by XRD. The 2 θ scan diffraction angle identifies the calcite and aragonite minerals shown in SEM panels A, C, E, and H.
      An EDX was used to analyze the chemical composition (i.e. calcium, magnesium, carbon, and oxygen content) of single crystals grown in the presence of ECMP-67. EDX showed that single calcite crystals contained low Mg2+ concentrations, although the solution contained a high concentration of Mg2+ (see supplemental Fig. S6 for details). The EDX analysis of nucleated crystals formed in the presence of ECMP-67 revealed novel functions of ECMP-67 in controlling the chemical composition of the crystals (see supplemental Fig. S6). The average elemental weight composition (atomic percentage) in a single crystal with 1.4 μg/ml protein (supplemental Fig. S6B, bottom) was found to be the following: calcium, 53.33; carbon, 15.75; oxygen, 29.26; and magnesium, 1.66. The protein-induced flower-shaped crystal exhibited a magnesium concentration of 1.66, indicating that this crystal is low magnesium calcite. In addition, the 2 θ diffraction angle scan identified this crystal as low magnesium calcite (Fig. 2S). The crystals formed in the control crystallization processes without protein or magnesium exhibited no magnesium signal. In contrast, the crystals formed in the control crystallization processes without protein but with magnesium showed a weak signal (0.58%), whereas crystal formation by protein induction showed a 3-fold stronger signal (1.66%), which is similar to the magnesium content of the sclerite (supplemental Fig. S6A, right-hand panel). These results indicate that ECMP-67 is highly influential in forming low magnesium calcite in the sea environment.
      Finally, AFM was used to investigate whether the molecular-scale kinetics of crystal formation are indeed influenced by ECMP-67. Different sizes of rhombohedral (104) crystal seeds were used as substrates for in vitro crystallization experiments. The AFM image in Fig. 3A shows an example of a blank crystal that was used for this study. Crystallization was evaluated with or without the influence of Mg2+ ions. Because Mg2+ ions may influence production of aragonite crystals in the marine environment, we examined crystallization under the same environmental ionic conditions present in the sea and in the presence or absence of protein. As shown in Fig. 3B, crystal growth on the surface of the crystal was not observed by AFM in the absence of protein and Mg2+ ions. When we added 50 mm Mg2+, we observed that nanostructured aragonites were formed (Fig. 3C). Fig. 3D depicts a single aragonite crystal grown separately in 50 mm Mg2+. Although the substrate was a calcite, the AFM image clearly shows that the Mg2+ ions influence the experiment by causing the production of needle-like aragonite crystals.
      Figure thumbnail gr3
      FIGURE 3Atomic force microscopy topographic images of crystals grown in vitro without ECMP-67 on a calcite (104) substrate. A, AFM of a blank crystal surface. B, a single crystal in the absence of protein and Mg2+ ions. No growth was observed on the surface of the crystal under these conditions. C, a single crystal without protein, but with 50 mm Mg2+. Under these conditions, growth of needle-like aragonite crystals was observed on the surface, indicating the influence of Mg2+ on the growth of aragonite during calcification. D, topography of a single needle-like aragonite crystal grown with 50 mm Mg2+ and without protein. The AFM image clearly shows how the Mg2+ ions influence the crystallization process by producing needle-like aragonite crystals even with calcite (104) as the substrate.
      When 1.4 μg/ml protein was added to solutions under the same conditions as described for Fig. 3, B–D, and the crystals were collected during the initial stage of biocalcification after 7 of 21 total days, crystal growth processes were changed, and protein attachments to the surface (shown in white) were clearly visible (Fig. 4A). A total of five single crystals were checked at this initial stage of the crystal growth process, and all showed the same results. Again, we examined the crystal growth process using the same protein concentration (1.4 μg/ml) and experimental conditions, and crystals were collected during biocalcification after 14 of 21 total days. These crystals showed growth patterns that were different from the crystals in Fig. 3 (no protein present) and an earlier time point in Fig. 4A (Fig. 4B, arrows; see Fig. 5 for growth spacing details), indicating the potential influence of proteins during the biocalcification process (Fig. 4B).
      Figure thumbnail gr4
      FIGURE 4Topographic AFM images of crystals grown in the presence of ECMP-67 and 50 mm Mg2+ ions in vitro for 21 days on calcite (104) substrates. A, topography of a single crystal grown with protein (1.4 μg/ml) and 50 mm Mg2+. The crystals were collected during the initial stage of biocalcification (7 of 21 total days). B, topography of a single crystal grown with the same protein concentration (1.4 μg/ml) and experimental conditions as those shown in , B–D, but after 14 days of biocalcification. This figure shows the initiation of calcite steps with larger edges (arrows), indicating the potential influence of proteins during the biocalcification process. C, a dynamic force plot of the surface of a single crystal with protein (0.7 μg/ml). The calcite formation step was initiated but not completed due to the low concentration of protein. D, a dynamic force plot of the surface of a single crystal with a standard concentration of protein (1.4 μg/ml). The calcite step edges were overgrown with protein attachments (arrows). E, a dynamic force plot of the substrate with 1.4 μg/ml protein. The step edges were smooth, and a rhombohedral calcite (104) crystal nanostructure (∼500 nm) was clearly visible on the surface (arrow). F, an AFM image of the surface structure in the presence of protein (1.4 μg/ml). A complete calcite structure was apparent and showed the same structural design as a rhombohedral calcite. G, a computer model of a rhombohedral calcite.
      Figure thumbnail gr5
      FIGURE 5Upper panel, X, molecular-scale AFM image of crystals formed in the presence of ECMP-67 (1. 4 μg/ml) after 14 of 21 total days of biocalcification showing initiation of calcite steps with larger edges (arrows). A and B, the molecular-scale full steps with spacing from starting point to end point. C and D, the molecular-scale specific points with spacing. The distance (μm), height (cm), and angle (°) of growth steps are clearly shown, as indicated by red, green, and yellow colors. Lower panel, Y, AFM image of the surface in the presence of proteins (1.4 μg/ml) showing a complete calcite structure. Left panel, the extended view of the crystal defined by the box reveals the same structural design as an ideal rhombohedral calcite.
      When a low concentration of protein (0.7 μg/ml) was added to solutions under the same conditions, as shown in Fig. 3, A and B, calcite growth was initiated (Fig. 4C, arrows) but was incomplete. In one experiment, when we added 1.4 μg/ml protein under the same conditions, the calcite crystals grown were clearly visible by protein attachment (Fig. 4D, arrows). The mineral and protein attachments on the same crystals were examined by a Raman microprobe (Fig. 6). The Raman microprobe clearly showed the protein attachments on the surface of single crystals (Fig. 6, protein peaks) and that the crystals produced by the influence of protein were completely calcite (see Fig. 6 for details). In Fig. 4D, the arrows on the topography plot depict crystal growth with fluffy white proteins on top. After completion of this 21-day experiment, the nanostructure (∼500 nm) of a rhombohedral calcite (104) crystal was clearly visible (Fig. 4E, arrow). An AFM image of the surface structure in the presence of 1.4 μg/ml protein shows a complete calcite shape (Figs. 4F and 5Y), and this shape was consistent with the original rhombohedral calcite structure shown in Fig. 4G (a computer model).
      Figure thumbnail gr6
      FIGURE 6Raman spectroscopy of protein attachment to crystals formed in the presence of ECMP-67 (1. 4 μg/ml) during the biocalcification process. The Raman spectra show that all crystals were calcite.

      DISCUSSION

      Biologically controlled calcification is almost always controlled by polysaccharides or proteins, which act as templates or so-called matrix proteins (
      • Weiner S.
      • Hood L.
      ) or catalyze certain pivotal reactions (e.g. carbonic anhydrase (
      • Jackson D.J.
      • Macis L.
      • Reitner J.
      • Degnan B.M.
      • Wörheide G.
      )). Our results demonstrate that only a single reactive protein, ECMP-67, is necessary to regulate mineralogy in vitro. This protein provides crucial insight into how some marine animals might be able to control and maintain their preferred skeletal mineralogy even when the ocean chemistry changes to favor a different calcium carbonate variety. During the history of the earth, this chemistry change has likely occurred quite often, and evidence of these changes is seen in the preferred skeletal mineralogy adopted by different taxa according to the sea water chemistry at the time of their origin (
      • Zhuravlev A.Y.
      • Wood R.A.
      ).
      A high magnesium/calcium ratio in seawater strongly favors aragonite over calcite precipitation (
      • Stanley S.M.
      • Hardie L.A.
      ,
      • Davis K.J.
      • Dove P.M.
      • De Yoreo J.J.
      ). The means by which organisms can precisely control their skeletal mineralogy despite what would be favored by ocean chemistry, however, remained elusive. Using several different approaches, we established that the ECMP-67 protein extracted from soft coral sclerites controls the formation of calcite crystals in vitro even when aragonite is favored by the chemistry of the solution from which the crystals are precipitated. Furthermore, the –OH group (either carboxylic or alcoholic) and other structural features detected in the protein-containing calcite may play a key role in the process of forming low magnesium calcite in soft corals. As demonstrated above, the near-field IR spectrum of the small protein-induced calcite particle contains a stronger –OH bond than the larger particle. We thus assumed that although the adsorption of proteins onto the surface may inhibit the aging of the crystals, it strongly influences the formation of a crystal face; however, a weak protein-crystal surface interaction may not be a strong inhibitor and may permit the growth of a relatively large crystal. SEM analysis clearly showed that protein-induced crystals are much smaller than control (uninduced) crystals (Fig. 2, A–K). In this study, we established that the protein ECMP-67 in soft corals only influences the production of calcite crystals and, therefore, the –OH group obtained in proteinaceous calcite may be a key factor in its formation. The topographic image of the single crystal formed in the presence of ECMP-67 clearly depicts a rough surface with different heights for the bulk composition. In contrast, a topographic image of a single crystal in the absence of any proteins depicts a planar surface. Near-field IR analysis suggests that ECMP-67 in soft corals is potentially involved in the formation of the coral body structure through small chemical compounds via sclerites. This new technique enables the chemical structure of submicrometer-to-nanometer-scale samples to be optically analyzed. Our results constitute the first practical implementation of near-field IR spectroscopy for biomineralization applications.
      Protein attachments on the surfaces of crystals (Fig. 4, A and D, white) were clearly visible, and these attachments functioned as templates for the growing calcite. Furthermore, protein attachment on the crystal surface during the biocalcification process was confirmed by a Raman microprobe (Fig. 6). Several experiments using different time scales in the biocalcification process were conducted in the presence of varying concentrations of protein to obtain a complete characterization of the transition from aragonite to calcite. A protein concentration of 1.4 μg/ml was found to be conducive for growing a calcite crystal (Figs. 4, B–G, 5, and 6).
      The solution from which crystals were precipitated contained 50 mm Mg2+ and 10 mm Ca2+ (i.e. a 5:1 ratio), which would strongly favor precipitation of only aragonite (
      • Rivadeneyra M.A.
      • Delgado R.
      • Párraga J.
      • Ramos-Cormenza A.
      • Delgado G.
      ,
      • Weber J.N.
      ). However, two strong bands at 1086.9 cm−1 (v1) and 1415 cm−1 (v3) in the NFIR spectrum, corresponding to calcite, were detected (Fig. 1A) when the crystals were precipitated from a solution that contained 1.4 μg/ml ECMP-67 (Fig. 2). As predicted (
      • Davis K.J.
      • Dove P.M.
      • De Yoreo J.J.
      ), aragonite was formed when ECMP-67 was absent from the solution (Fig. 2C), indicating that ECMP-67 can control the formation of the mineral phase.
      In coral mineralization, transporters such as proton and Ca2+ pumps, components involved in morphology control, and enzymes involved in hydration of CO2 are all considered to play essential roles. Soluble anionic macromolecules, including glycoproteins and calcium-binding proteins, are known to function as inhibitors or stabilizers by binding precursor amorphous calcium carbonate, which acts as a template for the transition of amorphous calcium carbonate to structured crystals on an insoluble or soluble organic matrix (
      • Cölfen H.
      • Mann S.
      ). However, triple functionality is necessary for complete crystallization in corals, especially in the sclerites of soft corals. Isolated ECMP-67 possesses carbonic anhydrase and calcium binding activity, as well as glycosylation, all of which allow it to function in calcium carbonate crystal formation. In the soft coral sclerites, aragonite precipitation is the expected initial state because of the high concentration of magnesium ions (magnesium/calcium = 5:1) in the sea. The growing crystals are then excluded from the cell, and some biological processes, especially those involving extracellular proteins (here, ECMP-67), strongly influence and override the formation of calcite crystals. The multifunctionality of ECMP-67 indicates that it is not only a structural protein, but also a catalyst and a potential template for single crystal formation in soft coral sclerites. These results demonstrate that a single reactive protein is likely to function as a determinant of skeletal mineralogy in soft coral sclerites, thereby providing an essential clue to how these organisms can determine their skeletal mineralogy.
      Based on our findings, we propose a model of how the calcified sclerites of soft coral are formed (Fig. 7). CO2 produced by respiration is converted to HCO3 by the carbonic anhydrase enzyme in the presence of H2O. After binding of calcium ions by Ca2+-binding proteins, ECMP-67 modulates the stereochemical position of carbonate groups dissociated from HCO3 to generate calcite crystals (Fig. 7). Because calcium carbonate is soluble at an acidic pH, the protons produced by the enzymatic reaction may be diluted in seawater or removed with a proton pump. In this scheme, ECMP-67 is a potential regulator of skeletal mineralogy in soft coral and greatly facilitates sclerite formation. However, further precise functional analyses of the effect of this protein on nucleation and growth of calcite crystals may lead to a deeper understanding of various biomineralization processes in many organisms.
      Figure thumbnail gr7
      FIGURE 7A model for sclerite formation by ECMP-67 in soft coral. Lane 1 shows the triple functional protein (ECMP-67) after purification by electroelution treatment. The primary sequence of ECMP-67 is shown at the top of the figure. CA enzyme, carbonic anhydrase enzyme.
      The mechanisms underlying the formation of single crystals (calcites) that have been shown here, as well as the postulated mechanisms, may have direct biological relevance and broad implications in environmental sciences and in the synthesis of many types of materials, including catalysis, tissue engineering, biomaterials, and drug design.

      Acknowledgments

      We thank Chihiro Jin, JASCO Corp., for technical assistance in the near-field IR analysis.

      REFERENCES

        • Stanley S.M.
        • Hardie L.A.
        Palaeogeogr. Palaeocl. 1998; 144: 3-19
        • Sandberg P.A.
        Nature. 1983; 305: 19-22
        • Falini G.
        • Albeck S.
        • Weiner S.
        • Addadi L.
        Science. 1996; 271: 67-69
        • Addadi L.
        • Joester D.
        • Nudelman F.
        • Weiner S.
        Chemistry. 2006; 12: 980-987
        • Thompson J.B.
        • Paloczi G.T.
        • Kindt J.H.
        • Michenfelder M.
        • Smith B.L.
        • Stucky G.
        • Morse D.E.
        • Hansma P.K.
        Biophys. J. 2000; 79: 3307-3312
        • Metzler R.A.
        • Evans J.S.
        • Killian C.E.
        • Zhou D.
        • Churchill T.H.
        • Appathurai N.P.
        • Coppersmith S.N.
        • Gilbert P.U.P.A.
        J. Am. Chem. Soc. 2010; 132: 6329-6334
        • Amos F.F.
        • Destine E.
        • Ponce C.B.
        • Evans J.S.
        Crystallogr. Growth Des. 2010; 10: 4211-4216
        • Belcher A.M.
        • Wu X.H.
        • Christensen R.J.
        • Hansma P.K.
        • Stucky G.D.
        • Morse D.E.
        Nature. 1996; 381: 56-58
        • Stephenson A.E.
        • DeYoreo J.J.
        • Wu L.
        • Wu K.J.
        • Hoyer J.
        • Dove P.M.
        Science. 2008; 322: 724-727
        • D'Souza S.M.
        • Alexander C.
        • Carr S.W.
        • Waller A.M.
        • Whitcombe M.J.
        • Vulfson E.N.
        Nature. 1999; 398: 312-316
        • Davis K.J.
        • Dove P.M.
        • De Yoreo J.J.
        Science. 2000; 290: 1134-1137
        • Weiner S.
        • Hood L.
        Science. 1975; 190: 987-989
        • Rahman M.A.
        • Isa Y.
        • Uehara T.
        Proteomics. 2005; 5: 885-893
        • Jackson D.J.
        • Macis L.
        • Reitner J.
        • Degnan B.M.
        • Wörheide G.
        Science. 2007; 316: 1893-1895
        • Goldberg W.M.
        • Benayahu Y.
        B. Mar. Sci. 1987; 40: 287-303
        • Kingsley R.J.
        • Watabe N.
        Cell Tissue Res. 1982; 223: 325-334
        • Kebukawa Y.
        • Nakashima S.
        • Nakamura-Messenger K.
        • Zolensky M.E.
        Chem. Lett. 2009; 38: 22-23
        • Koga S.
        • Yakushiji T.
        • Matsuda M.
        • Yamamoto K.
        • Sakai K.
        J. Membr. Sci. 2010; 355: 208-213
        • Lowry O.H.
        • Rosebrough N.J.
        • Farr A.L.
        • Randall R.J.
        J. Biol. Chem. 1951; 193: 265-275
        • Laemmli U.K.
        Nature. 1970; 227: 680-685
        • Maruyama K.
        • Mikawa T.
        • Ebashi S.
        J. Biochem. Tokyo. 1984; 95: 511-519
        • Segrest J.P.
        • Jackson R.L.
        Methods Enzymol. 1972; 28: 54-63
        • Yang S.Y.
        • Tsuzuki M.
        • Miyachi S.
        Plant Cell Physiol. 1985; 26: 25-34
        • Narita Y.
        • Kimura S.
        Anal. Sci. 2001; 17: i685
        • Rahman M.A.
        • Isa Y.
        • Uehara T.
        Mar. Biotechnol. 2006; 8: 415-424
        • Dauphin Y.
        Comp. Biochem. Physiol. A. Mol. Integr. Physiol. 2006; 145: 54-64
        • Venyaminov SYu.
        • Kalnin N.N.
        Biopolymers. 1990; 30: 1243-1257
        • Prange A.
        • Arzberger I.
        • Engemann C.
        • Modrow H.
        • Schumann O.
        • Trüper H.G.
        • Steudel R.
        • Dahl C.
        • Hormes J.
        Biochim. Biophys. Acta. 1999; 1428: 446-454
        • Kupka T.
        • Lin H.M.
        • Stobinski L.
        • Chen C.H.
        • Liou W.J.
        • Wrzalik R.
        • Flisak Z.
        J. Raman. Spectrosc. 2010; 41: 651-658
        • Zhuravlev A.Y.
        • Wood R.A.
        Geology. 2008; 36: 923-926
        • Rivadeneyra M.A.
        • Delgado R.
        • Párraga J.
        • Ramos-Cormenza A.
        • Delgado G.
        Folia Microbiol. (Praha). 2006; 51: 445-453
        • Weber J.N.
        Am. J. Sci. 1974; 274: 84-93
        • Cölfen H.
        • Mann S.
        Angew. Chem. Int. Ed. Engl. 2003; 42: 2350-2365