JOURNAL OF
Inorganic Biochemistry Journal of Inorganic Biochemistry 100 (2006) 88–96 www.elsevier.com/locate/jinorgbio
Nano-cluster composite structure of calcitic sponge spicules—A case study of basic characteristics of biominerals Ingo Sethmann
a,*
, Ruth Hinrichs
a,b
, Gert Wo¨rheide c, Andrew Putnis
a
a
Institut fu¨r Mineralogie, Universita¨t Mu¨nster, D-48149 Mu¨nster, Germany Instituto de Geocieˆncias, Universidade Federal do Rio Grande do Sul, 91501-970 Porto Alegre, RS, Brazil Geowissenschaftliches Zentrum Go¨ttingen, Abteilung Geobiologie, Universita¨t Go¨ttingen, D-37077 Go¨ttingen, Germany b
c
Received 19 July 2005; received in revised form 23 September 2005; accepted 16 October 2005 Available online 29 November 2005
Abstract Spicules of calcareous sponges are elaborately shaped skeletal elements that nonetheless show characteristics of calcite single-crystals. Our atomic force microscopic and transmission electron microscopic investigation of the triradiate spicules of the sponge Pericharax heteroraphis reveals a nano-cluster structure with mostly well-aligned small crystal domains and pockets with accumulated domain misalignments. Combined high-resolution and energy-filtering transmission electron microscopy revealed carbon enrichments located in between crystal domain boundaries, which strongly suggests an intercalated network-like proteinaceous organic matrix. This matrix is proposed to be involved in the nano-clustered calcite precipitation via a transient phase that may enable a Ôbrick-by-brickÕ formation of composite and yet single-crystalline spicules with elaborate morphologies. This composite cluster structure reduces the brittleness of the material by dissipating strain energy and deflecting crack propagation from the calcite cleavage planes, but the lattice symmetry and anisotropic growth properties of calcite still play a major role in the morphogenesis of these unusual calcite single-crystals. Our structural, crystallographic, textural, and chemical analysis of sponge spicules corroborates the view that nano-clustered crystal growth, induced by organic matrices, is a basic characteristic of biomineralisation that enables the production of composite materials with elaborate morphologies. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Sponge spicules; Biomineralization; Calcite; Organic matrix; Nano-structure
1. Introduction Sponges are the most primitive metazoans with merely a loose multi-cellular organisation that lacks real tissues and organs comparable to those of higher organised, triploblastic animals [1]. Despite their simple organisation, sponges of the class Calcarea exercise control over CaCO3 precipitation, which is demonstrated by characteristic morphologies of their skeletal spicules [1–3]. Rod-shaped, triradiate and quadriradiate sponge spicules have been described to behave as single-crystals of calcite, optically [4–6] and by X-ray diffraction [2], with their morphology strictly related
*
Corresponding author. Tel.: +49 251 83 33487; fax: +49 251 83 38397. E-mail address:
[email protected] (I. Sethmann).
0162-0134/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2005.10.005
to certain crystallographic orientations [7]. The material of different calcareous sponge spicules has been characterised as Mg-calcite [1,7,8] that may be associated with amorphous calcium carbonate (ACC), as shown for Clathrina spicules by Aizenberg et al. [9]. Also small amounts of proteinaceous organic materials, mostly less than 0.1 wt% [3,7,9], have been analysed as spicule constituents, but organic matter has already been presumed to be a component of the spicule material by earlier authors, e.g., Refs. [10–13]. The spicules form in intercellular cavities lined by sclerocytes that control the calcite crystal nucleation and the fluid conditions for further CaCO3 precipitation [12,14]. Spicule growth takes place by crystal elongation as rays, termed actines, in certain crystallographic directions, as well as by concentric thickening of the spicule actines [3,12,15].
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
Proteins, as found in the spicules, have been proposed to play a role in the spicule formation process by modifying the crystal morphology through interaction with specific crystal planes and inhibition of crystal growth in specific directions. It has further been assumed that subsequently the macromolecules become intercalated as a component of the mineralized skeletal elements [3]. However, it seems impossible for such elaborately tailored structures to grow by the same process considered as normal for non-biogenic calcite crystals, namely an advancement of monomolecular crystal steps over the rhombohedral f1 0 1 4g crystal faces [16]. In addition to influencing the spicule morphology, the interaction and intimate association of organic molecules with the mineral phase also interferes with the material texture, which may cause the nearly isotropic fracture behaviour instead of the usual calcite cleavage parallel to the f1 0 1 4g crystal planes [7,12,13]. Stimulated by the long-standing controversial discussion of the composite crystal structure of calcareous sponge spicules and so-called single-crystalline biominerals in general, we carried out a detailed structural and textural investigation of sponge spicules. An analysis of the crystal texture and the spatial distribution of incorporated organic material in relation to structural subunits of the mineral phase on a nanometre-scale allows conclusions on the biomineralization processes and the role of soluble proteinaceous organic matter therein. 2. Materials and methods The investigation was carried out on triactine spicules of the calcareous sponge Pericharax heteroraphis Pole´jaeff, 1884, collected by SCUBA diving by G.W. at Northwest Island, Great Barrier Reef, Australia. Immediately after the sponge had been collected it was frozen at 20 °C and subsequently freeze-dried before we performed our analyses on the dry samples. Field-emission scanning electron microscopy (FESEM): Examination of spicule surfaces and fracture patterns was performed with an FESEM (JSM 6300F, JEOL), equipped with an energy dispersive X-ray analysis (EDX) system for element analyses. Intact and fractured spicules were glued to a C-tab and sputtered with carbon before the investigation. Atomic force microscopy (AFM): Structural investigation of large triradiate sponge spicules was carried out with a Multimode scanning probe microscope and a Dimension 3000 scanning probe microscope, both with a Nanoscope IIIa controller (Digital Instruments). The naturally grown spicule surfaces were examined with an AFM-integrated optical microscope camera to locate smooth light-reflecting surface areas, where the external membraneous organic sheath had peeled off, probably during the removal of the spicules from the dried sponge body. This ensured that organic tissue-free pristine spicule surfaces were investigated with AFM. The internal structure of the spicule
89
actines was investigated on fresh fracture surfaces. In dissolution experiments spicules have been immersed in a few droplets of demineralised water for 2 h, dried on a tissue and examined with an AFM afterwards. All AFM images were obtained by slow scanning (0.5–1.5 Hz) in tapping mode for most accurate results. X-ray diffraction (XRD): For single-crystal X-ray examination a single triradiate spicule was mounted on a glass fibre. The investigation was performed with a precession camera (Huber) with a molybdenum anode using white X-ray radiation produced at 40 kV and 30 mA. The distance between the spicule and the photo film was 60 mm. The Laue pattern was obtained after adjusting the spicule orientation by the precession method. Optical microscopy: The optical investigation was performed with a Leica DMRX microscope, equipped with a CCD camera (Kappa CF 15/4). Large spicules were separated and examined in reflected and transmitted light. Bits of sponge tissue comprising the spicules were embedded in epoxy resin, sectioned and polished on one surface, then glued facedown to a glass slide with wax before polishing the sample to a thickness of 30 lm. The crystallographic properties of the spicules were investigated in transmitted polarised light. Transmission electron microscopy (TEM): From the thin section (see above) a sample of 3 mm in diameter was drilled out with an ultrasonic drill. A carbon-coated TEM copper grid was glued to the surface of the sample before the sample was detached from the glass slide through gentle heating and melting of the wax. The sample was thinned to electron transparency by ion beammilling, and the surface was coated with an evaporated carbon film. Additional samples have been prepared by gentle crushing of spicules in an agate mortar and transferring the smallest fraction with acetone onto lacy carbon-coated TEM copper grids. No further preparation was performed on these samples. The investigation was performed with a JEOL 3010 TEM equipped with a Gatan post-column energy filter (Gatan imaging filter) and an Oxford link ISIS EDX system. High-resolution TEM (HRTEM) was performed on electron transparent areas, and the structural information was correlated to chemical information from nanometre-scale element maps for C, O, Mg, and Ca of the same areas, obtained through energy-filtering TEM (EFTEM) imaging with the Gatan imaging filter. The energy-filters have been adjusted to the respective element edges of the electron energy-loss spectra (EELS) of the samples. Care was taken to select sample areas with a low ratio of sample thickness to mean free path of electrons, t/k < 0.5, corresponding to sample thicknesses, t < 50 nm, which is appropriate for CaCO3 samples to avoid signal interferences by substantial electron energy loss through inelastic scattering. For the textural–chemical investigation crushed samples remained uncoated and sample areas freestanding over holes of the lacy carbon glue were selected to obtain a pristine carbon-signal of the sample.
90
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
Crystallographic setting: Calcite shows a trigonal structural symmetry, 32=m, space group R 3c. All given Miller– Bravais indices for lattice planes (h k i l) and direction indices [u v t w] refer to the hexagonal setting and the ˚ and structural unit cell with the parameters, a = 4.989 A ˚ c = 17.062 A. 3. Results An arrangement of spicules as skeletal elements prevents the sponge body from collapsing (Fig. 1(a)). The three actines of each triactine spicule are equal in length (between 150 and 1500 lm) and thickness (depending on length, between 20 and 180 lm near the spicule centre), round in cross section, and symmetrically arranged in one plane with angles of 120° between them (Fig. 1(b)) (see also Wo¨rheide and Hooper [17]). Each triactine spicule behaved as a single-
Fig. 1. (a) Sponge skeleton, consisting of an array of spicules; (b) small calcareous triactine spicule of the sponge Pericharax heteroraphis (optical micrograph, reflected light); (c) and (d) thin section of variously oriented spicules embedded in resin. In transmitted light with crossed polarisers the spicules behave as single-crystals. The large spicule in the centre, embedded and cut obliquely to the plane of the actines, is shown in bright orientation (c) and rotated by 45° in extinction orientation (d), with the extinction being incomplete; (e) X-ray diffraction (Laue pattern) confirms the calcitic single-crystalline behaviour of the spicules (spicule orientation as in (b), irradiation parallel to the c-axis) and (f) drawing of a spicule in relation to the crystallographic axes, according to (e), and to a calcite f1 0 1 4g rhombohedron in different view directions.
crystal of calcite in polarised light (Fig. 1(c) and (d)) and in X-ray diffraction (Fig. 1(e)). The optical axis (i.e., the crystallographic c-axis, [0 0 0 1]) was oriented perpendicular to the plane of the actines (i.e., the (0 0 0 1) lattice plane), and the actines were elongated parallel to the three equivalent a*-axes (i.e., the h1 0 1 0i set of directions) (Fig. 1(f)). The incomplete extinction of spicules in polarised light (Fig. 1(d)) may have been due to dispersed misaligned crystal domains and dispersed light refraction through lattice strain at domain boundaries. In optical microscopy, focussed below the spicule surface, actines appeared finely segmented perpendicular to the respective actine long axis (Fig. 2(a)). In reflected light and in FESEM images the naturally grown spicule surfaces appeared very smooth, but AFM images with superior height resolution revealed a nano-granular cluster structure (Fig. 2(b)) that is absent in normal calcite crystals. The size of the granules was about 30–50 nm in diameter. Dissolution experiments in water, carried out on natural (0 0 0 1) spicule surfaces (i.e., the plane of the actines), produced etch pits in which also nano-granular structures were recognisable (Fig. 2(c) and (d)). Despite this granular structure etch figures developed in regular triangular shape, with edge directions nearly parallel to the long axes of the actines (i.e., the a*-axes) (Fig. 2(c)–(e)), which emphasises the single-crystalline character of the spicules. The morphology of the cylindrical spicule actines may be approximated by a set of crystal faces that comprise the relatively common {0 0 0 1} and f1 1 2 0g forms (Fig. 2(f); cf. Goldschmidt [18]) of which the respective a*-axis represents the zone axis. Despite the largely single-crystalline appearance of the spicules the calcitic material fractured with a conchoidal, glass-like pattern (Fig. 3(a)) instead of showing the brittleness parallel to the crystallographic f1 0 1 4g set of rhombohedral cleavage planes typical of normal calcite. Fracture surfaces showed similar nano-cluster structures, with a size range of the cluster subunits of about 10–30 nm (Fig. 3(b)). HRTEM reveals that the fracture pattern is strongly affected by the crystal texture in that cracks propagate and change direction according to domain boundaries (Fig. 3(c)). TEM investigation was performed to further characterise the texture of the spicule material. Corresponding to the results from XRD and optical microscopy the spicules behave predominantly as calcite single-crystals. Fig. 4(a) shows an HRTEM image of a well-aligned area of the calcitic crystal, with the electron beam direction close to [0 0 0 1], as confirmed by the fast Fourier transform (FFT) of the image (Fig. 4(b)). Filtering the image of Fig. 4(a) by selecting the predominant frequencies of the FFT (Fig. 4(c)) and subsequently inverting the transform function clearly revealed numerous minute imperfections of the crystal (Fig. 4(d)). On a larger scale, irregular structures were recognisable in TEM bright field images (Fig. 5(a)). Selected area electron diffraction patterns demonstrated the presence of a
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
91
Fig. 2. (a) Centre of a large triactine spicule imaged in reflected light with the image focus beneath the spicule surface. The actines appear finely segmented perpendicular to their long axes, the a*-axes (optical micrograph), (b) nano-granular structure of a spicule surface in the plane of the actines (0 0 0 1) (AFM height image). (c) and (d) regular triangular etch figures in the nano-cluster material developed in etch pits on the (0 0 0 1) surface (AFM images of the same area, (c) height image with triangle drawn as a guide to the eye, (d) phase image for enhanced small-scale morphology contrast), (e) relationship between orientations of etch figures as in (c) and (d), the spicule morphology, the crystallographic axes, the calcite f1 0 1 4g rhombohedron, and the prismatic f1 1 2 0g set of calcite crystal faces and (f) schematic cross section through a spicule actine (grey) in relation to the crystallographic axes and crystal faces that approximate the cylindrical surface.
minor constituent of distinctly misaligned crystal domains in addition to the predominant crystal orientation, with all the diffraction spots and rings being assignable to calcite lattice spacings. In contrast to the periodic main diffraction pattern, the misaligned domains were not of one single orientation, but had a number of preferential orientations, as indicated by additional streaked spots, some of them even broadened to complete diffraction rings (Fig. 5(b)). Dark field imaging by selecting one diffraction spot of the periodic pattern in Fig. 5(b) highlighted a part of the sample area that contributed to the predominant single-crystalline appearance (Fig. 5(c)). A dark field image obtained by
Fig. 3. (a) Fractured spicule actine with conchoidal fracture pattern (FESEM), (b) AFM height image of a conchoidal fracture surface, revealing the nano-cluster structure of the spicule material, and (c) HRTEM image of the fracture pattern in relation to the crystal texture. The trace of fracture is clearly influenced by crystal domain boundaries (arrows). The black line represents a rounded fracture trace, which may clarify the structural relation between crystal domains and cluster units as displayed in (b).
92
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
Fig. 5. (a) TEM bright field image of a Pericharax spicule (polished, ionmilled, and carbon coated); (b) selected area electron diffraction pattern of (a), with the electron beam direction almost perpendicular to the ð1 0 1 4Þ cleavage plane of the predominating calcite Ôsingle-crystalÕ, producing the periodic spot pattern. Additional streaked spots and rings indicate misaligned calcite domains. The encircled diffraction spots C, 024 of the periodic calcite pattern, and D, a streaked 206 spot of a secondary misaligned calcite pattern, have been selected for dark field images (c) and (d), respectively; (c) TEM dark field image of the same area as in (a), obtained by using spot C of the periodic pattern in (b), highlighting corresponding areas of the predominant Ôsingle-crystalÕ and (d) TEM dark field image of the same area as in (a), obtained by using the streaked spot D of the secondary pattern in (b), highlighting areas with crystal domains of a certain misaligned orientation with respect to the predominant ÔsinglecrystalÕ.
Fig. 4. (a) HRTEM image of an area with semicoherent and well-aligned crystal domains (crushed sample); (b) FFT of (a), indicating three equivalent crystal planes, the f1 1 2 0g planes, oriented parallel to the electron beam; (c) selecting the signals of the f1 1 2 0g planes from the FFT pattern and inverting the FFT function produces a filtered image (d) of the original (a), revealing the spicule crystal structure viewed perpendicular to the (0 0 0 1) plane, and showing the slightly distorted trigonal calcite symmetry, as indicated by the inset equivalent a*-axes.
selecting a streaked spot, not belonging to the periodic pattern, highlighted minor, spherical pockets in the sample with distinctly accumulated textural misalignments (Fig. 5(d)). But electron diffraction patterns only from these ÔpocketÕ areas were still essentially the same as in Fig. 5(b), with the predominant periodic pattern in the same orientation and the independent secondary pattern of streaked spots and rings. HRTEM images with distinctly
misaligned domains clearly demonstrated crystal domain sizes of only about 5 nm in cross section (Figs. 3(c) and 6(a)). In the HRTEM image of Fig. 6(a) the crystal domains appeared to be tilted randomly with respect to each other, but a FFT (inset of Fig. 6(a)) clearly reveals certain preferential crystallographic orientations, confirming the information obtained from electron diffraction patterns (Fig. 5(b)). We assume that the pocket-like local accumulations of misalignments in the spicule crystals are pristine naturally grown features and not preparation artefacts, since similar features were observed in differently prepared samples, in crushed samples as well as in polished and ion beam-milled samples. Additionally, the fact that the pocket areas still show the undisturbed periodic diffraction pattern dominating over the misalignment pattern disagrees with the idea of artificial misalignment structures. However, we cannot exclude with certainty that crushing of the sample and the polishing procedure may produce similar defect patterns induced by mechanical stress. If the observed misalignments are in fact artefacts, then these can be assumed to be part of the typical fracture pattern of the sponge spicules. Ion beam-milling cannot be held responsible for producing the misalignments, since it has not been applied to
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
93
all the investigated samples. Moreover, the ion irradiation is known to produce a surface amorphisation of completely different appearance. We can also exclude the TEM electron beam as the cause of the domain misalignments, since a long-term irradiation test yielded no noteworthy sample damage. Proteinaceous organic matter incorporated in the spicules can be clearly distinguished from the Mg-calcite by its much higher carbon content, estimated to 40 wt% C in proteins (based on amino acid compositions of sponge spicule proteins published by Aizenberg et al. [3]), in contrast to little more than 12 wt% C in the Mg-calcite. As an attempt to localise organic matter in relation to the crystal nano-texture imaged through HRTEM (Fig. 6(a)), element maps were obtained through EFTEM imaging of exactly the same sample areas for Ca (Fig. 6(b)), C (Fig. 6(c)), and Mg (Fig. 6(d)). This analytical combination provides directly related textural and chemical information spatially resolved on a nanometre scale. Concentrations of Ca and C were negatively correlated to each other. Carbonenriched and calcium-depleted networks intercalated in between the crystal domains were recognisable. While calcium-rich areas largely corresponded to the calcite crystal domains, the carbon-enriched network presumably represented intercalated organic matter. For oxygen no specific ultrastructure-related distribution pattern was discernable, probably due to less significant differences in the oxygen contents of CaCO3 and the proteins. Magnesium also did not show a distinctly ultrastructure-related distribution pattern (Fig. 6(d)), which means that the Mg content of the spicules may not be entirely incorporated in the calcite structure, but that some of it was associated with the intercalated organic matter. For the sample area in Fig. 6(a)–(d) the value of t/k was in the range of about 0.1–0.3 (Fig. 6(e)), corresponding to a sample thickness of about 10–30 nm. A series of TEM-EDX analyses showed varying Ca/Mg ratios of the Mg-calcite spicule material. An average Ca/Mg molar ratio of the carbonate mineral (cf. Fig. 6(f)) could roughly be determined as about 8–9, corresponding to a MgCO3 content of about 10–11 mol%, without taking into account the effect of non-crystalline Mg ions associated with the suggested intercalated organic matter. 4. Discussion
Fig. 6. (a) HRTEM image of a mosaic of misaligned crystal domains in a Pericharax spicule and the corresponding FFT (inset) (crushed sample without further preparation or coating). Element distribution maps of the same area as in (a) obtained with EFTEM, (b) Ca, (c) C, (d) Mg. ((a–d) Images are subdivided into four sectors for easier recognition of corresponding image structures; (b–d) lighter shadings correspond to higher relative concentrations.), (e) profile of t/k of the horizontal centre line for (a–d) and (f) TEM-EDX spectrum of spicule material with roughly an average Ca/Mg ratio (crushed sample on holey carbon-coated copper grid).
The exact crystallographic orientation of the symmetric Pericharax triactine spicules (Fig. 1(c)–(f)) requires strict control of the mineralization process, which seems improbable to be accomplished through the spatial arrangement of the spicule-forming cells alone (cf. Ledger and Jones [14]). Corresponding relations between morphology and crystallography have already been observed in spicules of different calcareous sponges [2,3,7], which is certainly not coincidental. The morphological symmetry of a spicule reflects crystallographic symmetry elements of a calcite crystal, the
94
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
material of which it is made. The formation of the elaborate actines instead of normal massive calcite rhombohedra is certainly a result of local cellular control over crystal growth. This control is exercised by the so-called Ôfounder cellsÕ or apical cells located at the tips of the actines, regulating their elongation by site-directed secretion of the spicule constituents, and by the Ôthickener cellsÕ that finish the formation of the lateral actine surfaces [3,12,19]. However, a peculiar feature of the spicules is their glasslike, isotropic fracture pattern (Fig. 3(a)), which already points to the fact that the spicules do not consist of normal single-crystalline calcite. Normal calcite would show a brittle cleavage parallel to the f1 0 1 4g crystallographic lattice planes that define the typical calcite rhombohedra. This modification of the material properties has already been reported for calcareous spicules of other sponge species [7,20] as well as for different single-crystalline biominerals, such as sea urchin skeletal elements [16,21–23]. The spicule ultrastructure of nano-cluster blocks (Figs. 2(b)–(d) and 3(b)), consisting of even smaller, more or less well-aligned crystal domains (Fig. 3(c)), is the key to understanding the unusual mechanical behaviour of the biogenic calcitic material. Similar observations of cluster-like nano-structures have been made for CaCO3 bio-crystals of sea urchin skeletons [16] and mollusc shells [24,25]. Deflection of crack propagation at the boundaries of cluster blocks or crystal domains seems to prevent cleavage (Fig. 3(b) and (c)) and to cause the observed conchoidal fracture pattern (Fig. 3(a)). The size of crystal domains observed with HRTEM (Figs. 3(c) and 6(a)) is distinctly smaller than the cluster units observed with AFM (Figs. 2(b)–(d) and 3(b)). The AFM-observed cluster units may consist of several crystal domains, as suggested by the fracture pattern viewed in HRTEM (Fig. 3(c)). Mechanically the spicule material behaves in a nearly isotropic way, which demonstrates enhanced toughness [26]. Our TEM investigation with selected area electron diffraction and dark field imaging confirmed a predominant crystal orientation of the entire spicule material (Fig. 5(b) and (c)), but also revealed pockets with abundant misaligned domains (Fig. 5(d)). A comparison of the HRTEM images of the Figs. 4 and 6(a) clearly demonstrates the textural difference in the degree of domain alignment in the two different types of texture in the spicule material. Provided that these imperfections in the single-crystalline structure are pristine growth features, their formation may be either due to an inability of the organism to perfectly control crystal growth, or they may be a welcome feature induced Ôon purposeÕ to further enhance the materials toughness through the resulting Ôplywood effectÕ. To some extent the organic matrix may prevent crumbling of the material under mechanical stress by having a cohesive effect on mineral fragments. It has already been assumed by other authors that organic macromolecules, e.g., proteins, become intercalated in between crystal domains of sponge spicules and other biominerals, where they might be involved in modify-
ing the mechanical material properties [3,13,22,23]. Element mapping on a nanometre-scale with EFTEM (Fig. 6(b)–(d)) in combination with HRTEM (Fig. 6(a)) enabled us to actually localise the intercalated organic matter and to determine its distribution in direct relation to the crystal domains. The enrichment of carbon and depletion of calcium indicates the locations of organic matter that forms a network-like matrix intercalated in between crystal domain boundaries (Fig. 6(a)–(c)). The almost uniform distribution of Mg (Fig. 6(d)) suggests that a minor amount of the magnesium has remained associated with the organic matter during crystal growth. The accumulated amount of Mg may have exceeded the critical capacity of calcite for thermodynamically stable Mg-uptake (Fig. 6(d)), which is estimated to about 7–10 mol% MgCO3 under ambient seawater conditions [27]. This value is in agreement with the average MgCO3 content of the spicules, roughly determined as 10–11 mol%. The organic matter must have been present during the spicule formation, since it is incorporated in the spicule material. The nano-clustered granular growth structure, as seen on the natural spicule surfaces (Fig. 2(b)), easily allows for intercalation of organic molecules. A similar internal nano-cluster structure, as seen in dissolution patterns (Fig. 2(c) and (d)) as well as on the conchoidal fracture surfaces (Fig. 3(b)), suggests that during the growth process these cluster blocks are formed as basic building block particles, similar to bricks in a brick wall. The fine segmentation of the actines (Fig. 2(a)) may be caused by spicule growth taking place in an oscillating manner rather than in a steady way. The fact that the final product is largely a calcite single-crystal, despite its nano-cluster structure, could be explained by two different scenarios: The clustered crystals may form by oriented aggregation of preformed calcite nano-particles (cf. [28]). Intercalated organic macromolecules may, however, obstruct the crystallographically aligned aggregation of crystal blocks, making it difficult to produce large structures with largely single-crystalline properties. Alternatively, spicule growth may take place by attachment of ACC particles associated with proteinaceous macromolecules to the pre-existing crystalline biomineral, or by the formation of gelatinous films of CaCO3 and proteins on the biomineral surface (cf. [16]). Subsequent calcitic crystallisation of the attached precursor material may take place, mostly in a substrate-conformable way, with protein intercalation in between the semicoherent crystal domains. On basis of the presented results the actual growth pattern cannot finally be determined. However, a growth process via a precursor phase may be supported by the occurrence of ACC in very similar spicules of the sponge Clathrina, although in that case the detected ACC represented a stable phase that occurred in an internal concentric layer of the actines [9]. According to our present high-resolution analyses, ACC is at most a minor phase in Pericharax spicules. Small amounts of ACC
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
may be directly associated with the organic material intercalated in between crystal domains (Fig. 6(a)–(c)). However, the techniques applied are not appropriate to exclude the presence of further stable ACC dispersed in the investigated spicules. The accumulation of inorganic and organic materials for spicule growth is likely to occur in specific vesicles of the adjacent cells from which they are released to the growth sites of the spicules, as implied by Ledger and Jones [14]. Organic polyanion-mediated growth processes via amorphous precursor phases have already been described or proposed for other CaCO3 biominerals formed by organisms very distantly related to sponges, such as sea urchin larval spicules [29,30] and adult skeletons [16,31], mollusc shells [32,33], and coccoliths [34]. Nano-cluster structures of biogenic single-crystals very similar to those described here have so far been reported from mollusc shells [24,25] and sea urchin skeletons [16], where they probably result from a growth mechanism similar to the one described above. This biomineralization process has been experimentally approached by polyaspartate-mediated calcite precipitation via organic–inorganic liquid or gelatinous precursor phases [16,35]. Precursor gels induce the development of single-crystalline calcite nano-clusters, presumably by acting as a mineral–ion source and porous crystal growth matrix at the same time [16]. While surface-adsorbed gelatinous proteins inhibit crystal growth through step advancement, substrate-conformable nucleation is favoured. We assume that in biomineralization a similar precipitation mechanism in confined microenvironments under cellular control enables single-crystal growth as a nano-cluster with elaborate morphologies. Site-directed attachment of nano-particulate amorphous precursor material facilitates subsequent substrateconformable crystallisation with rounded, microscopically smooth crystal morphologies through a nano-scale brick wall-like pattern. The morphology of the resulting nanocluster construction does not depend on layer-by-layer growth that inevitably would lead to the development of distinct crystal faces. The examined sponge spicules, formed in a controlled manner by one of the simplest animals on Earth, are a demonstration model for basic biomineral construction principles and emphasise the key-role of soluble organic matrices in biomineralization. 5. Abbreviations ACC AFM EDX EELS EFTEM FESEM FFT HRTEM
amorphous calcium carbonate atomic force microscopy energy-dispersive X-ray analysis electron energy-loss spectrum energy-filtering transmission electron microscopy field-emission scanning electron microscopy fast Fourier transform high-resolution transmission electron microscopy
TEM XRD
95
transmission electron microscopy X-ray diffraction
Acknowledgements We would like to thank Ju¨rgen Lo¨ns for support with XRD, Christine Putnis for FESEM imaging, and Lia Addadi (Weizmann Institute of Science, Israel) for comments that helped to improve this paper. This study was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG), priority program SPP 1117, ÔPrinciples of BiomineralizationÕ. R.H. thanks the Alexander von Humboldt Foundation for research support. G.W. would like to thank the Great Barrier Reef Marine Park authority for permitting fieldwork (Permit Nos. G98/142, G98/022) as well as the Australian Biological Resources Study (ABRS) and AstraZeneca R&D Griffith University (Brisbane, Australia) for funding. This work is part of the doctoral thesis of I.S. at the Department of Geosciences of the Westfa¨lische Wilhelms-Universita¨t Mu¨nster. References [1] J.N.A. Hooper, R.W.M. Van Soest, Systema Porifera. Guide to the Supraspecific Classification of Sponges and Spongiomorphs (Porifera), Plenum, New York, 2002. [2] J. Aizenberg, J. Hanson, T.F. Koetzle, L. Leiserowitz, S. Weiner, L. Addadi, Chem. Eur. J. 1 (1995) 414–422. [3] J. Aizenberg, M. Ilan, S. Weiner, L. Addadi, Connect. Tiss. Res. 34 (1996) 255–261. [4] W.J. Sollas, Sci. Proc. R. Dubl. Soc. (NS) 4 (1885) 374–392. [5] V. von Ebner, Sber. Akad. Wiss. Wien, 1. Abt. 95 (1887) 55–149. [6] W.C. Jones, Quart. J. Micr. Sci. 96 (1955) 129–149. [7] J. Aizenberg, J. Hanson, M. Ilan, L. Leiserowitz, T.F. Koetzle, L. Addadi, S. Weiner, FASEB J. 9 (1995) 262–268. [8] W.C. Jones, D.A. Jenkins, Calc. Tiss. Res. 4 (1970) 314–329. [9] J. Aizenberg, G. Lambert, L. Addadi, S. Weiner, Adv. Mater. 8 (1996) 222–226. [10] E. Weinschenk, Centralbl. Mineral. Geol. Pala¨ont. 19 (1905) 581–588. [11] R.T. Paine, Ecology 45 (1964) 384–387. [12] W.C. Jones, Symp. Zool. Soc. London 25 (1970) 91–123. [13] A. Berman, J. Hanson, L. Leiserowitz, T.F. Koetzle, S. Weiner, L. Addadi, Science 259 (1993) 776–779. [14] P.W. Ledger, W.C. Jones, Cell Tiss. Res. 181 (1977) 553–567. [15] P.W. Ledger, W.C. Jones, in: J. Reitner, H. Keupp (Eds.), Fossil and Recent Sponges, Springer-Verlag, Berlin, Heidelberg, 1991, pp. 341– 359. [16] I. Sethmann, A. Putnis, O. Grassmann, P. Lo¨bmann, Am. Min. 90 (2005) 1213–1217. [17] G. Wo¨rheide, J.N.A. Hooper, Mem. Queensland Mus. 43 (1999) 859– 891. [18] V. Goldschmidt, Atlas der Kristallformen Band II, Carl Winters Universita¨tsbuchhandlung, Heidelberg, 1913. [19] E.A. Minchin, Quart. J. Micr. Sci. (NS) 40 (1898) 469–587. [20] W.C. Jones, D.W.F. James, Micron 3 (1972) 196–210. [21] K.M. Towe, Science 157 (1967) 1048–1050. [22] A. Berman, L. Addadi, S. Weiner, Nature 331 (1988) 546–548. [23] J. Aizenberg, J. Hanson, T.F. Koetzle, S. Weiner, L. Addadi, J. Am. Chem. Soc. 119 (1997) 881–886. [24] Y. Dauphin, Pala¨ont. Z. 75 (2001) 113–122. [25] Y. Dauphin, J. Biol. Chem. 278 (2003) 15168–15177.
96
I. Sethmann et al. / Journal of Inorganic Biochemistry 100 (2006) 88–96
[26] S. Weiner, L. Addadi, H.D. Wagner, Mater. Sci. Eng. C 11 (2000) 1–8. [27] R.A. Berner, Geochim. Cosmochim. Acta 39 (1975) 489–504. [28] H. Co¨lfen, M. Antonietti, Angew. Chem. Int. Ed. 44 (2005) 5576– 5591. [29] E. Beniash, L. Addadi, S. Weiner, J. Struct. Biol. 125 (1999) 50–62. [30] S. Raz, P.C. Hamilton, F.H. Wilt, S. Weiner, L. Addadi, Adv. Funct. Mater. 13 (2003) 480–486.
[31] Y. Politi, T. Arad, E. Klein, S. Weiner, L. Addadi, Science 306 (2004) 1161–1164. [32] M.E. Marsh, R.L. Sass, J. Exp. Zool. 226 (1983) 193–203. [33] S.J. Hattan, T.M. Laue, N.D. Chasteen, J. Biol. Chem. 276 (2001) 4461–4468. [34] M.E. Marsh, Protoplasma 177 (1994) 108–122. [35] L.B. Gower, D.J. Odom, J. Cryst. Growth 210 (2000) 719–734.
