Journal of Structural Biology 173 (2011) 99–109 Contents lists available at ScienceDirect Journal of Structural Biology journal homepage: www.elsevier.com/locate/yjsbi Calcareous sponge biomineralization: Ultrastructural and compositional heterogeneity of spicules in Leuconia johnstoni Carter, 1871 Christophe Kopp a, Anders Meibom a, Olivier Beyssac b, Jarosław Stolarski c, Shakib Djediat d, Jakub Szlachetko e, Isabelle Domart-Coulon f,* a Muséum National d’Histoire Naturelle, Laboratoire de Minéralogie et Cosmochimie du Muséum (LMCM), UMR 7202, Case Postale 52, 61, rue Buffon, 75005 Paris, France Institut de Minéralogie et de Physique des Milieux Condensés (IMPMC), UMR 7590 CNRS-IPGP-Universités Paris 6 & 7, Campus Boucicaut, 140 rue de Lourmel, 75015 Paris, France Institute of Paleobiology, Polish Academy of Sciences, Twarda 51/55, 00-818 Warsaw, Poland d Muséum National d’Histoire Naturelle, Département RDDM, USM 505 Case Postale 39, 57 rue Cuvier, 75005 Paris, France e European Synchrotron Radiation Facility, X-Ray Microscopy Beamline ID21, B.P. 220, 38043 Grenoble Cedex, France f Muséum National d’Histoire Naturelle, Département Milieux et Peuplements Aquatiques, UMR 7208 MNHN-CNRS-IRD-UPMC Case Postale 26, Biologie des Organismes et des Ecosystèmes Aquatiques (BOREA), 43 rue Cuvier, 75005 Paris, France b c a r t i c l e i n f o Article history: Received 12 May 2010 Received in revised form 12 July 2010 Accepted 17 July 2010 Available online 23 July 2010 Keywords: Calcareous sponge Biomineralization Calcite Ultrastructure Composition Spicule a b s t r a c t In contrast to siliceous sponge spicules, the biomineralization in calcareous sponges is poorly understood. In particular, the existence of a differentiated central core in calcareous spicules is still controversial. Here we combine high-spatial resolution analyses, including NanoSIMS, Raman, SXM, AFM, SEM and TEM to investigate the composition, mineralogy and ultrastructure of the giant tetractines of Leuconia johnstoni Carter, 1871 (Baeriidae, Calcaronea) and the organization of surrounding cells. A compositionally distinct core is present in these spicule types. The core measures 3.5–10 lm in diameter and is significantly depleted in Mg and lightly enriched in S compared with the adjacent outer layer in the spicule. Measured Mg/Ca ratios in the core range from 70 to 90 mmol/mol compared to 125–130 mmol/mol in the adjacent calcite envelope. However, this heterogeneous distribution of Mg and S is not reflected in the mineralogy and the microstructure. Raman spectroscopy demonstrates a purely calcitic mineralogy. SEM examination of slightly etched spicules indicates an ultrastructure organized hierarchically in a concentric pattern, with layers less than 250 nm in width inside layers averaging 535 ± 260 nm. No change in structural pattern corresponds to the Mg/Ca variation observed. AFM and TEM observations show a nanogranular organization of the spicules with a network of intraspicular organic material intercalated between nanograins 60–130 nm in diameter. Observations of sclerocyte cells in the process of spiculogenesis suggest that the compositionally distinct core is produced by a sub-apical sclerocyte ‘‘founder cell” that controls axial growth, while the envelope is secreted by lateral sclerocytes ‘‘thickener cells”, which control radial growth. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Calcium carbonate spicules provide skeletal support for marine calcareous sponges (Calcarea, Porifera). Although the shape, size and spatial arrangement of spicules in the sponge body have traditionally been a very important taxonomic tool (Manuel et al., 2003), the biomineralization processes by which they form are still poorly understood. As a first step towards a better understanding of calcareous sponge spicules formation, it is necessary to characterize potential spatial variations in their composition and ultrastructure. Such observations could provide indications on the * Corresponding author. Fax: +33 1 40 79 37 71. E-mail address: icoulon@mnhn.fr (I. Domart-Coulon). 1047-8477/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2010.07.006 biological control of the organism, at the cellular level, during the biomineralization process. Calcareous sponge spicules are extracellular biominerals formed in the mesohyl of the sponge body between the pinacoderm surface and the choanocyte filtration chambers. Scenarios of formation have been proposed since the 1970s based on ultrastructural observations. It is currently believed that spicules form in an intercellular cavity between several specialized spiculesecreting cells, the sclerocytes (Ledger and Jones, 1977), sealed by septate junctions (Ledger, 1975). Growing spicules are enveloped by a thin organic sheath (Jones, 1967; Ledger, 1974; Ledger and Jones, 1977), which later becomes enclosed in the network of collagen connective fibrils positioning mature spicules in the sponge mesohyl (Simpson, 1984; Sethmann and Wörheide, 2008). Rates of spiculogenesis have been assessed based on 45Ca labeling and the incorporation of tetracycline fluorochrome, 100 C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 suggesting a 1–3% spicule turn-over in Clathrina cerebrum over 24 h in short-term aquarium experiments (Bavestrello et al., 1993, 1994). Dynamics and growth pattern of calcareous spicules were also studied in Sycon sp with calcein fluorescence labeling (Ilan et al., 1996), revealing marked differences in growth rates depending on spicule type, with slender monaxon deposition rate increased by a factor 6 compared to curved monaxon. Crystallographic studies and bulk analyses of their chemical composition have revealed low-Mg calcite (Jones and Jenkins, 1970) behaving as a single-crystal (reviewed by Sethmann and Wörheide, 2008). Ultrastructural observations of intact and fractured spicules showed a polycrystalline organization into nanogranular clusters and suggested organic matter intercalation between small crystal domains (Sethmann et al., 2006). Aizenberg et al. (1996b) have detected small amounts of intraspicular proteinaceous macromolecules (0.07–0.1 wt.%) characterized by a content rich in asparagine and/or aspartic acid. At the microscale, spicules are characterized by a pattern of successive concentric layers (traditionally called ‘stratification’), which are observed mostly in the larger triactine and tetractine types by alkaline or acidic etching (e.g. Jones and James, 1972; Jones, 1970; Ledger and Jones, 1991; Von Ebner, 1887). However, the existence of a differentiated central core in the calcareous sponge spicules has always been controversial. By comparison, it is well established that a central organic axial filament is present in the siliceous spicules of Demospongiae and Hexactinellida sponge species (Uriz et al., 2003; Uriz, 2006). Building on Von Ebner’s 1887 spicule heating experiments and Bütschli’s 1901 acid etching experiments, Minchin (1898) and then Minchin and Reid (1908) obtained an axial residue stainable with nigrosin or indulin from spicules decalcified with diverse acids, suggesting the existence of an organic axial filament. However, Jones (1967) (later confirmed by Ledger and Jones (1991)) observed that this organic axial filament was an artifact corresponding to remnants of the spicule organic sheath contracted during the decalcification process. Etched spicules of e.g. giant triactines of Leuconia nivea sometimes revealed central hollows (Jones and James, 1972; Ledger and Jones, 1991), which could suggest a differentiated axial core. According to Von Ebner (1887), the axial portion of the spicules was ‘a region of less pure calcite compared to the peripheral portion, incorporating impurities such as magnesium, sodium and sulfate ions’ and the stratification of spicule was due to the ‘periodic deposition of less pure and more pure calcite’. Jones (1970) also suggested that varying Mg content might explain the concentric lamination pattern. However, no significant variations in the spatial distribution of Ca, Mg, Sr and S could at that time be detected with the electron micro-probe along profiles through transverse and longitudinal sections of rays of giant triactines of L. nivea, probably due to low spatial resolution (>1 lm) and sensitivity (Jones and James, 1969). Based on the fact that the smaller types of spicules of L. nivea and Amphiute paulini contained less Mg than the larger ones (Jones and Jenkins, 1970), Jones (1970) proposed the hypothesis of a potential heterogeneous distribution of Mg in spicules, with the first secreted spicular material depleted in Mg compared to the peripheral deposits. Recently Sethmann et al. (2006) have used high-resolution TEM–EDX to detect nanometer-scale heterogeneity of Mg distribution in fractured spicules. However, these observations were not related to micrometric features like core or concentric layers or to differences in cellular activity. Mineralogical heterogeneity was reported by Aizenberg et al. (1996a, 2003) in the spicules of Clathrina sp (collected off Atlit, Israel) with a calcitic core enveloped by a thick layer of ACC, which was covered by a thin calcitic layer. However, it was not clear if such proposed mineralogical stratification was species specific or widely distributed among calcareous sponges. In this work we have mapped at high, sub-micrometric, spatial resolution with SEM,1 AFM, NanoSIMS, SXM and Raman mapping the structure and composition of giant tetractines of Leuconia johnstoni Carter, 1871. The main purpose is to examine compositional variations of calcareous spicules in relation to their ultrastructure. The ultrastructure of sclerocytes (spicule-secreting cells) and spicules was imaged with TEM and SEM, and a cellular model is proposed for axial growth (elongation) versus radial growth (thickening) to explain the spatial variations in trace element distribution detected in the calcitic spicules. 2. Materials and methods 2.1. Biological material Specimens of the calcareous sponge L. johnstoni Carter, 1871 (Baeriidae, Calcaronea, Calcispongia) were collected in the subtidal zone under granite boulders (Fig. 1a) in South Finistère off Concarneau, France, between May 2008 and October 2009 (47°52N, 3°55W). L. johnstoni is a littoral marine species living on sub-vertical rock surfaces in wave exposed sites, with a distribution ranging from the British Isles to the Channel coasts of France and the Gulf of Biscay (Picton and Morrow, 2007). This small-sized encrusting sponge (less than 50 mm diameter and 15 mm height) is composed of compressed lobes fused at their base and bearing apical openings. The color may be white to beige and silt is often trapped between the lobes on the surface. The specimens were fixed either with paraformaldehyde 3% or ethanol 70%, or were directly frozen at 20 °C. Taxonomic determination of the specimens was confirmed based on examination of spicule nature and arrangement, and by comparison to reference L. johnstoni from the MNHN collection of Porifera (microscopic preparations from the Borojevic Roscoff collection voucher C-1968-426, and from the Haeckel collection voucher C-1968-678). 2.2. Morphological analysis of spicule types and orientation To isolate spicules, small sponge fragments were bleached at room temperature in 5% commercial sodium hypochloride solution (NaClO) on a rocking table for 1 h to digest the organic matter. The dissociated spicule suspension was then briefly rinsed three times in tap water, rinsed once in deionized water and transferred to absolute ethanol. The morphology of isolated spicules was observed in reflection with a stereomicroscope LEICA MZFLIII or in transmission (in permanent araldite slide mounts) with a microscope LEICA DMRB. Images were acquired with a LEICA DC300F camera (Leica, France) and processed with IM50 software. For SEM observations, isolated spicules were deposited on a stub, coated by gold (5 nm thick) and examined at 15 keV with SEM TESCAN (model VEGA II LSU) or with SEM JEOL JSM-840 at the Muséum National d’Histoire Naturelle (MNHN) (Paris, France). Alternatively, sponge fragments were cryofractured in absolute ethanol, and either critical-point dried by CO2 substitution or lyophilized and then mounted on SEM stubs. Spicular arrangement and cell–spicule interface were observed in the fracture area at 15 keV with the SEM JEOL JSM-840 A at the MNHN. Calcareous sponge spicules are classified according to the number of rays (actines) that they possess. Thus, the different spicule categories are monaxon – a term including monactine (one ray) and diactine (two rays) – triactine (three rays) and tetractine (four 1 Abbreviations used: SEM, scanning electron microscopy; AFM, atomic force microscopy; NanoSIMS, secondary ion mass spectrometer; TEM, transmission electron microscopy; EDX, energy dispersive X-ray spectroscopy; SXM, scanning X-ray microscope; ACC, amorphous calcium carbonate. C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 101 Fig. 1. Macromorphology of Leuconia johnstoni. (a) Live specimen encrusting granite boulders in the tidal zone. (b) SEM micrograph of a sponge fragment cryofractured near the apex of a lobe, indicating that the sponge is composed of an outer ectosome (ect) rich in large spicules and inhalant canals, covering the choanosome containing the choanocyte filtration chambers (ch), the exhalant canals and a small atrial central cavity. Triactines and the basal triradiate system of giant tetractines (white arrows) are disposed tangentially to the ectosome, whereas the apical ray of giant tetractines is directed internally, perpendicular to the sponge surface. rays). Tetractines are built from three rays disposed like a triactine, referred to as the basal triradiate system, with an additional ray called the apical actine (Boury-Esnault and Rützler, 1997). 2.3. Compositional analysis Small fragments of L. johnstoni were dried and embedded in Körapox epoxy resin. They were then polished with diamond suspensions (particle sizes = 3 lm, 1 lm and 0.25 lm), revealing multiple sectional planes through rays. During this process, some fracturing of the rays occurred, especially for the smallest types of spicules, which are very fragile. Following established procedures (Meibom et al., 2004, 2008), the spatial distribution of Mg was mapped out in transverse and longitudinal sections of rays of giant tetractines of L. johnstoni with the Cameca NanoSIMS ion micro-probe at the Laboratoire de Minéralogie et de Cosmochimie du Muséum (LMCM) of the MNHN. Briefly, with a primary beam of O , focused to a spot-size of 200 nm on the gold-coated surface of the sample, secondary ions of 24Mg+ and 44Ca+ were sputtered from the sample surface and detected simultaneously in multi-collector mode with electron multipliers at a mass-resolving power of 4500. At this mass-resolving power, the measured secondary ions are resolved from potentially problematic interferences. Maps were obtained by rastering the primary beam across a pre-sputtered surface. Magnesium concentrations were calibrated against carbonate standard of known composition. Sulfur content was mapped at high-spatial resolution in crosssection of a giant tetractine spicule with SXM operating in the Xray fluorescence mode at the X-ray Microscopy beamline ID21 of European Synchrotron Radiation Facility (Grenoble, France) following procedures described by Cuif et al. (2003). The X-ray beam, monochromatized by means of double-crystal (Si(1 1 1)) fixed exit monochromator, was tuned to an energy just above the K absorption of sulfur. The Kirkpatrick–Baez mirror arrangement was employed to focus the X-ray beam down to size of 0.3 0.8 lm2. The X-ray fluorescence spectra were recorded by HpGe detector placed at 90° scattering angle. A 2D image was obtained by point-by-point scanning of the sample across the focal point of the beam, with typical exposure time of 150–300 ms per point. 2.4. Micro- and nanostructural analysis Following NanoSIMS analysis, the gold was delicately removed by polishing with a diamond paste of 0.25 lm particle size. The samples were then slightly etched, during approximately 2 min 15 s, with solution of 0.1% formic acid (pH 3–4) and 2.5% glutaraldehyde, then rinsed in tap water and ethanol 70%, dried and coated with gold (15 nm). Examinations of etched sections of giant tetractines were performed with SEM Tescan (model VEGA II LSU) at the MNHN. Giant tetractines were recognized by their diameter and/or the length of their rays. Microstructural organization was not characterized for the other categories of spicules, especially the smallest ones, because they were too fragile to resist polishing or completely dissolved by the etching. Nanostructural analysis was investigated with AFM performed with a MultiMode Nanoscope IIIa (Digital Instruments, Veeco), following procedures described by Stolarski and Mazur (2005). Standard silicone nitride cantilevers were used for measurements in tapping mode. Sections of giant tetractines polished as previously described were examined after etching in 1% ammonium persulfate in McIlvain buffer (pH 8) for 10 min, followed by rinsing in deionized water and drying. 2.5. Mineralogical analysis For mineralogical analysis, spicules were prepared from sponge specimens fixed by freezing at 20 °C and stored frozen at 20 °C until use, in order to avoid crystallization of potential unstable ACC phase. Isolated spicules were thawed just prior to Raman microspectroscopy. Raman maps were obtained from the surface of apical rays of giant tetractines freshly broken in a transversal way before analysis. Spot analyses were carried out in three different places on the surface of this ray: the center, the medium and the extreme periphery (respectively noted (1), (2) and (3) on Fig. 6a). Raman spectra were also acquired from the surface of minute tetractines, microdiactines, pugioles and triactines (data not shown). All the spicules were placed on an aluminium foil-covered glass slide. Raman maps and spectra were acquired using a Renishaw InVia Reflex micro-spectrometer at the Institut de Minéralogie et de Physique des Milieux Condensés (IMPMC) (Paris, France). A 785 nm near-infrared diode laser (Renishaw) was focused on the sample by a DMLM Leica microscope with a 50 (Numerical Aperture = 0.55) or a 100 (Numerical Aperture = 0.70) objective. The Rayleigh diffusion was eliminated by edge filters and the signal was finally dispersed using a 1800 lines/mm grating and analyzed by a Peltier cooled RENCAM CCD detector. Before each session, the spectrometer was calibrated with a silicon standard. For Raman mapping we used the dynamic line-scanning mapping device Streamline. See Bernard et al. (2008) for more details. 2.6. Histological and ultrastructural analysis of cell structures For TEM, fragments of L. johnstoni sponges were fixed in 2% glutaraldehyde in modified Sörensen phosphate buffer (NaH2PO4 and 102 C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 2H2O–Na2HPO4 0.1 M adjusted to pH 7.6 and sucrose 0.6 M) with 20 mM CaCl2 and 0.055% w/v Ruthenium Red. Fragments were cut into smaller pieces, which were partly decalcified by ascorbic acid 2% for 24 h at room temperature. They were postfixed in 1% OsO4 in modified Sörensen buffer and dehydrated through a graded ethanol series. The pieces were embedded in Spürr resin. Sections were cut with a Diatome 35° diamond (Ultracut microtome). Semi-thin sections were stained with solution of 1% toluidine blue, 1% borax in 70% ethanol and observed with a microscope LEICA DMRB equipped with a LEICA DC300F camera (Leica, France). Ultra-thin sections were counterstained with uranyl acetate 2% in 50% ethanol and were observed at 75 keV with a Hitachi H7100 transmission electron microscope (TEM) equipped with a digital CCD Hamamatsu camera at the MNHN. 3. Results 3.1. The different types of spicules of L. johnstoni The skeleton of L. johnstoni (76% of the sponge total dry weight) is composed of seven different categories of spicules: (1) large sagittal (giant) tetractines (Fig. 2a) with paired rays 220–820 lm long by 25–100 lm maximum width, unpaired ray 180–480 lm long by 25–100 lm maximum width and apical ray 160–725 lm long by 20–100 lm maximum width; (2) sagittal triactines (Fig. 2b), with paired rays 80–480 lm long by 10–35 lm maximum width and unpaired ray 85–360 lm long by 10–30 lm maximum width; (3) dagger-shaped small sagittal tetractines (Fig. 2c), termed pugioles and characteristic of the order Baerida (Borojevic et al., 2002), with paired rays 25–55 lm long by 10 lm maximum width, unpaired ray 65–100 lm long by 10 lm maximum width and apical ray 75–95 lm long by 10 lm maximum width; (4) long, smooth, and slightly curved diactines (Fig. 2d), 250–670 lm long by 15– 25 lm maximum width; (5) microdiactines 45–65 lm long by 3– 5 lm maximum width (Fig. 2e) ornamented by rows of minute spines (Fig. 4d); (6) regular, minute sized, tetractines (Fig. 2e and g) with rays of the basal triradiate system 10–20 lm long by 1– 2 lm maximum width and a short apical ray 2–3 lm long; (7) long, very fine and straight monaxons located on the external surface near the oscules (Fig. 2f). These observations match the original description of the spicule types in the L. johnstoni species by Carter (1871), with additional measurements. Each spicule ray behaves as a single-crystal in polarized light. In cryofractured specimens, the surface ectosome is differentiated from the internal choanosome (Fig. 1b). The ectosomal skeleton of the sponge is composed of tangential triactines and of the tangential basal trira- diate system of giant tetractines (Fig. 1b). Their unpaired rays seem to be all oriented towards the sponge substrate (data not shown). The apical ray of giant tetractines is directed internally, perpendicular to the sponge surface (Fig. 1b). The choanosome is rich in microdiactines, pugioles and a few minute tetractines. In this work, we have focused our analyses on the giant tetractines of the sponge ectosome, because they were less fragile than the other spicules and presented a surface more adequate for spatial analysis. 3.2. Microstructural concentric internal layers and nanostructural composite organization of the spicules Fig. 3 illustrates the microstructure of giant tetractines revealed by slight acidic etching of polished sections with 0.1% formic acid during 2 min 15 s. Giant tetractines are composed of fine concentric bands (Fig. 3a–c). Here, we define a band by the succession of a furrow of etched material and a ridge of acid-resistant material (inset of Fig. 3b). The lamination is more pronounced in certain regions of the rays and these differences may result of a potentially heterogeneous efficiency of the etching solution (artifact). In transverse sections layers less than 250 nm wide, referred to as primary bands, are detected inside larger layers, referred to as secondary bands. This is especially visible in the spicule center, where the layers are clearly ellipsoid (Fig. 3b). Average width of secondary bands is estimated to 535 ± 260 nm (n = 136 measures taken from different regions of transverse and longitudinal sections of 14 different rays). In longitudinal sections bands seem wider near the apex of the ray than laterally (Fig. 3c). A near rectilinear ridge crosses systematically the central region of the rays in transverse sections and could define a symmetric axis for the axial ellipsoid bands (Fig. 3b). This ridge extends in the range of 12–33 lm. At the nanostructural level, AFM results show that spicules consist entirely of an assemblage of nanograins ca. 60–130 nm in diameter, separated from each other by spaces of a few nanometers (Fig. 4a and b). This nanogranular organization is also supported by observations with SEM of the conchoidal fracture area of rays (Fig. 4c), in every spicule type. Complementary TEM examination of partially decalcified microdiactines (Fig. 4d–f) reveals two kinds of organic material associated with spicules, both stained by Ruthenium Red, indicating a content rich in proteoglycans: a very thin organic sheath enveloping the spicules (no more than few dozen nanometers wide); a networked intraspicular organic material delineating small compartments of approximately the same size as nanograins detected with AFM. These small com- Fig. 2. The seven types of spicules composing the skeleton of Leuconia johnstoni. Optical images (a–e) illustrate various spicules at the same scale, after dissociation with bleach from the sponge. SEM micrographs (f, g) of spicules in situ in the sponge body fragment near an oscule. (a) Giant sagittal tetractine, with indication of the terminology for each ray. (b) Sagittal triactine. (c) Small tetractine, termed pugiole. (d) Long, smooth, thick and slightly curved diactine. (e) Small microdiactine (left) and minute tetractine (right). (f) Long, very fine and straight monaxons located on the external surface (black arrow) between the abundant giant tetractines and large triactines. (g) Minute tetractine embedded in the wall of an exhalant canal. C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 103 Fig. 3. Microstructural organization of rays from giant tetractines of Leuconia johnstoni, revealed by slight acidic etching with 0.1% formic acid (2 min, 15 s) indicating a concentric lamination pattern. Large rectilinear scratches across the sections are polishing arctifacts. (a) Transverse section of a ray. (b) Detail of (a). The inset is a higher magnification of the axial region of the ray. Bands, defined by a doublet furrow plus ridge, are clearly visible. White arrows indicate the position of a systematically centrally located near rectilinear ridge. A primary banding is visible (black arrowhead) within the secondary bands of the axial region of the ray. (c) Longitudinal section of a ray. The inset represents the whole ray for orientation and the square indicates the general location of the enlarged view. Fig. 4. Nanostructural organization of spicules of Leuconia johnstoni. (a) AFM height mode and (b) amplitude mode image of the surface of a spicule in transverse section. Spicules consist of nanograins ca. 60–130 nm in diameter separated from each other by a thin space of few nanometers. (c) SEM micrograph of the surface of a fracture through a triactine ray revealing similar nanograins (inset) to those observed with AFM; organic remnants of the unbleached spicule envelope are visible. (d) SEM image of a dissociated microdiactine ornamented by rows of minute spines (inset). (e and f) TEM micrographs of a partially decalcified microdiactine showing organic material stained by Ruthenium Red proteoglycan fixative, forming the outer sheath and a networked intraspicular organic material delineating nanograins similar in size to those observed with AFM. These nanograins may form a peripheral layer 100 to 250 nm wide (inset (e)). partments can be seen forming a peripheral layer 100 to 250 nm wide (Fig. 4e), which may correspond to the primary banding increments reported for giant tetractines. 3.3. Trace element heterogeneity in the giant tetractines at ultrastructural length scale With the high-spatial resolution (200 nm) of the NanoSIMS the distribution of Mg was mapped in transverse sections of five rays, each belonging to a different giant tetractine (Fig. 5a illustrates two examples). Mg was also mapped in longitudinal section of a ray from a giant tetractine (Fig. 5b). In all spicules analyzed in transverse section, the core of the ray is systematically depleted in Mg compared to the envelope of the ray (Fig. 5a). The diameter of this Mg-depleted core, more or less cylindrical, ranges from 3.5 to 8 lm and corresponds in longitudinal section to a central band (Fig. 5b). The Mg/Ca ratio in the ray envelope averages approximately 125–130 mmol/mol, with fluctuations less than 10% and no periodicity (Fig. 5c and d). This ratio drops substantially in the ray core to reach values between 70 and 90 mmol/mol, which represent a diminution by 30% to 45% compared to the envelope (Fig. 5c and d). With SXM operating in the X-ray fluorescence mode a map of S distribution was obtained from a transverse section of a ray from a giant tetractine (Fig. 5e) showing an ellipsoid central ray core, 4.5 to 10 lm in diameter, lightly enriched in S compared to the ray envelope. There is a strong spatial correlation between the Mg-depleted area and the S-enriched area in the core of spicular rays (compare Fig. 5a and e). The extreme periphery of rays also appears to be enriched in Mg and S (Fig. 5a and e), conceivably due to remnants of organic sheath and cells covering the spicule (Fig. 5f). Spatial variations in Mg/Ca ratio and S content are thus tightly regulated in spicular rays and define a compositionally differentiated spicule core. 104 C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 Fig. 5. Spatially heterogeneous distribution of Mg (a–d) and S (e) in the rays of giant tetractines of Leuconia johnstoni. The axial region of the rays is depleted in Mg and enriched in S. (a) NanoSIMS maps of Mg distribution from two different rays cut transversely. The inset is a map centered on the axial region of a ray. Red color code indicates regions of high Mg/Ca ratios versus blue for low-Mg/Ca. (b) NanoSIMS map of Mg distribution in the axial region of a ray cut longitudinally. (c) Fluctuations of Mg/Ca ratio, quantitatively expressed along the AB transect of (a). (d) Fluctuations of Mg/Ca ratio along the CD transect of (b). (e) SXM map of S distribution in a ray cut transversely. The red arrow indicates the position of the sulfur enriched axial region (light blue color code). (f) SEM micrograph of a ray showing a conchoïdal fracture pattern. The ray is enveloped by an organic sheath and by sponge cells (red arrow). 3.4. Uniform calcitic mineralogical composition of giant tetractines at microstructural length scale Raman spectra were acquired with a spot-size of 1 lm whereas Raman maps were obtained with a step width of 1 lm directly from the surface of an apical ray of a giant tetractine freshly broken along a transversal fracture (Fig. 6a). Three representative spectra obtained from spot analysis of the center, the medium part and the extreme periphery of this ray (respectively noted (1), (2) and (3) on Fig. 6a) exhibit bands characteristic of magnesian–calcite (Fig. 6b): two lattice modes at 156 cm 1 and 283 cm 1, t1 CO3 symmetric stretching at 1090 cm 1 and t4 antisymmetric stretching at 716 cm 1. These values are globally similar to other synthetic and some biogenic magnesian–calcites analyzed with Raman spectroscopy with similar Mg composition in the range 70–130 mmol/mol (or 7 to 13 mol% MgCO3) (Bischoff et al., 1985; Urmos et al., 1991). Raman mapping confirm these observations and highlight the remarkable constancy of the frequency and full width at half maximum (FWHM) of these four bands over the whole section of the ray (Fig. 6c–e). In addition, the two lattice modes are systematically clearly observed over the whole section. Altogether, this suggests that giant tetractines of L. johnstoni are entirely composed of magnesian–calcite. No amorphous phase, like ACC was detected. Similar observations performed for some smaller types of spicules: minute tetractines, microdiactines, triactines and pugioles gave exactly similar spectra of magnesian–calcite (data not shown). This strongly suggests that the seven forms of spicules present in L. johnstoni are entirely composed of magnesian–calcite. The weak broad band observed near 1010 cm 1 (Fig. 6b) may be attributable to bicarbonate (HCO3 ) ions (Bischoff et al., 1985). It is interesting to note that the fluctuation of the Mg content between the core (70 to 90 mmol/mol, or 7 to 9 mol% MgCO3) and the envelope (125 to 130 mmol/mol, or 12.5 to 13 mol% MgCO3) of rays from giant tetractines was not detected by Raman spectroscopy. Indeed, no significant differences in the Raman band frequencies or FWHM were observed from the core to the outer rim of the ray analyzed (Fig. 6c–e). It has been shown that fluctuations of 7 to 13 mol% MgCO3 are also not clearly detected with Raman between samples of both synthetic and biogenic magnesian–calcites (Bischoff et al., 1985). Thus, Raman spectroscopy may not be sensitive enough to detect such relatively small fluctuations of Mg concentration in calcite. 3.5. Cellular observations at micro- and ultrastructural level In semi-thin (ca. 1 lm) sections through the ectosome outer part of the sponge body wall, stained by toluidine blue-Borax in ethanol 70% (Fig. 7a and b), growing spicules are associated with formative sclerocyte cells localized in the mesohyl, sometimes almost in contact with the surface pinacoderm (Fig. 7a). Their shape, orientation and localization in the ectosome suggest that they probably correspond to immature giant tetractines or triactines. At this stage of spicular growth, one sclerocyte is positioned subapically at the tip of each ray (Fig. 7a and b). The maximum width of the ray to which they are associated varies between 4.5 and 10 lm (values obtained from six different rays covered with a sub-apical sclerocyte). SEM examination of the choanosome inner part of a cryofractured sponge body wall (Fig. 1b) also reveals some sclerocytes associated with growing spicules (Fig. 7c and d). A short, straight and very thin (less than 500 nm wide) microdiactine probably at its initial stage of growth is lodged in an intercellular cavity bounded by two formative sclerocytes (Fig. 7c). At a later stage of spicular growth, two sclerocytes are positioned sub-apically, each associated to one of each tip of the microdiactine (Fig. 7d). The morphology of sclerocytes varies from cuboidal to flattened, more or less elongated and with some pseudopods (Fig. 7a–d). In semi-thin sections, they display a nucleus approximately 5 lm in diameter and intracellular granules stained with toluidine blue (Fig. 7a and b). At the ultrastructural level, they are characterized by numerous mitochondria and a great abundance of small vesicles with homogeneous electron-translucent C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 105 Fig. 6. Raman microspectroscopy analysis of an apical ray from a giant tetractine of Leuconia johnstoni revealing that this spicule consists entirely of magnesian–calcite. (a) Optical image of a transverse section of the apical ray. (b) Raman spectra were recorded from the center, the medium and the extreme periphery (respectively noted (1), (2) and (3)) of the surface of the section. The three spectra exhibit bands characteristic of magnesian–calcite: two lattice modes at 156 cm 1 and 283 cm 1, t1 CO3 symmetric stretching mode at 1091 cm 1 and t4 antisymmetric stretching mode at 716 cm 1. The weak broad band observed near 1010 cm 1 may be attributable to bicarbonate (HCO3 ) ions (Bischoff et al., 1985). (c) Raman map of the section showing the correlation index of each spectrum with the reference spectrum of lattice vibrations in crystalline calcite. A value of 1 indicates a perfect similarity of the sampled spot and the reference and any value above 0.95 an excellent similarity. (d) Raman image of the Raman shift of the t1 CO3 symmetric stretching mode. (e) Raman image of the full width at half maximum (FWHM) of the t1 CO3 symmetric stretching mode. For (c, d), the signal obtained on the outside of the ray is due to the basal triradiate system situated at the bottom. granular content (Fig. 7e and inset). Sclerocytes also contain a few electron-dense granules (Fig. 7e and inset). A very thin organic sheath is sometimes detected enveloping growing and mature spicules (Fig. 7a and b). After spiculogenesis this sheath may be thickened by addition of mesohyl-derived collagenous fibrils secreted by collencytes that were detected very close to the spicules (data not shown). Mature, cell-free spicules are observed to be positioned and anchored into the collagenous matrix of the mesohyl through an oriented network of collagenous fibrils (Fig. 7f and inset) with protein complexes closely associated to the tip of each spicule ray and to the collagenous fibrils. 4. Discussion 4.1. Hierarchical structure of calcareous sponge spicules Microstructural organization of calcareous spicules, characterized by successive concentric layers, strongly suggests that biomineralization in calcareous sponges is a cyclical process: during spicular growth, successive layers are secreted by sclerocytes and accreted around the axis of actines. At micrometric length scale, acidic etching of polished transverse and longitudinal spicule sections revealed that the giant tetractines of L. johnstoni consist of concentric layers of calcareous material surrounding a central actine axis. Some primary bands less than 250 nm wide were detected inside larger secondary bands averaging 535 ± 260 nm in width. The occurrence of this concentric banding pattern confirms previous SEM investigations made with or without etching in spicular sections of other calcareous sponge species, like L. nivea (Jones and James, 1972; Ledger and Jones, 1991), Sycon ciliatum (Ledger and Jones, 1991) and Pericharax heteroraphis (Sethmann and Wörheide, 2008). However, the sizes of secondary bands (550–6660 nm wide) and primary bands (100–900 nm wide) reported in giant triactines of L. nivea (Jones and James, 1972; Ledger and Jones, 1991) were larger than those we observe for giant tetractines of L. johnstoni. Banding pattern may be specific to the species and to the spicule type, but reflects an incremental growth process general to all calcareous sponge spicules. The factors responsible for this banding periodicity are certainly intrinsic to sclerocyte activity but still need to be discovered. As concluded by Ledger and Jones (1991) for small spicules of Sycon sp, the number of growth layers in an actine is too high to correlate with tidal or circadian rhythms, taking into account that one spicule takes no more than two days to fully grow. This rapid rate of skeletal deposition was confirmed for Sycon sp curved monaxon by in vivo fluorescent calcein-labeling studies (Ilan et al., 1996), establishing a 12 ± 3 lm/h growth-rate and full secretion within about 24 h, although growth rate of other spicules types varied. The near rectilinear ridge we observed in cross-section from giant tetractines of L. johnstoni could indicate a crystallographic continuity from one layer to another. Thus, this suggests that a new layer may crystallize in continuity to the precedent deposit. At the nanostructural level, combined AFM and SEM results showed that spicules of L. johnstoni are polycrystalline and consist of a multitude of nanograins commonly ca. 60–130 nm in diameter, intercalated with spaces of a few nanometers. Nanogranular organization, with nanograin size ca. 10–50 nm, has also been reported in spicules of the calcareous sponge P. heteroraphis by Sethmann et al. (2006) based on AFM observations. Each nanograin was interpreted as being composed of several crystal domains of a few nanometers, visible with high-resolution TEM, and possibly embedded in organic matter (Sethmann et al., 2006). This nanoorganization of calcareous sponge spicules may be responsible for the conchoidal fracture pattern observed when they are broken (Figs. 4c, 5f and 6a), which is different from the cleavage pattern characteristic of abiogenic calcite. Indeed, this fracture property 106 C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 Fig. 7. Spicule-secreting sponge sclerocytes. (a) Optical image of a semi-thin section of the sponge ectosome region stained with 1% ethanolic toluidine blue-borax showing three growing rays (outlined with lines) of an immature giant tetractine or triactine: each ray tip is associated with a sub-apical sclerocyte. (b) Enlargement of a sclerocyte on another spicule growing in the ectosome. (c) SEM micrograph of the cryofractured choanosome region of the sponge with a short, straight and very thin (less than 500 nm wide) microdiactine at its initial stage of growth, lodged in an intercellular cavity bounded by two formative sclerocytes, in the ceiling of a choanocyte chamber. (d) SEM micrograph of a microdiactine at a later stage of growth with an associated sub-apical sclerocyte. (e) TEM ultra-thin micrograph of a sclerocyte showing a great abundance of mitochondria and small electron-translucent vesicles and a few electron-dense granules. Inset is higher magnification view of the cell ultrastructure. Holes at the bottom of section correspond to spicule fragments fallen during sectioning. (f) TEM image of an ultra-thin section through a mature spicule positioned and anchored into the collagenous matrix of the mesohyl through an oriented network of collagenous fibrils associated with protein complexes at the tip of each spicule ray. edg = electron-dense granule; me = mesohyl; pc = protein complexes; pi = pinacoderm; ps = pseudopod; scl = sclerocyte; sh = organic sheath; sv = small vesicles; sp = spicule; sw = sea-water. might be the result of the deflection of cracks propagating at the boundaries of cluster-nanograins or crystal domains as suggested by Sethmann et al. (2006). The nanogranular organization of calcareous sponge spicules thus confirms that nanogranular components are widely distributed among organisms producing calcium carbonate biocrystals under biological control, suggesting that they are a universal component of biogenic carbonates (Stolarski and Mazur, 2005). Interestingly, TEM observations of partially decalcified microdiactines revealed a network of intraspicular organics, delineating small compartments of approximately the same size as nanograins observed with AFM. Therefore, taking into account their positive staining with Ruthenium Red, indicative of glycoproteins, we can suppose that each nanograin may be surrounded by glycopro- tein-rich intraspicular organic material. Aizenberg et al. (1996b) characterized the intraspicular proteinaceous macromolecule fraction representing 0.07–0.1 wt.% of the spicules of Sycon sp, Kebira sp and Clathrina sp, revealing a content rich in asparagine and/or aspartic acid. Based on synchrotron X-ray diffraction analysis and in vitro precipitation experiments, Aizenberg et al. (1995a, 1995b, 1996b) suggested that unidirectional elongation of the spicule rays was controlled by stereochemical interactions of the growing crystals with intraspicular specialized proteins. As acknowledged in many other biomineralization models (Weiner and Dove, 2003), intraspicular organic components of calcareous sponge spicules may function as mineralizing organic matrix serving as a template for the nucleation and subsequent growth of the fundamental spicule biomineralization units. C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 4.2. Existence of a compositionally distinct core in the giant tetractines of L. johnstoni The occurrence and the nature of a differentiated core in calcareous sponge spicules have been debated for a long time (e.g. Minchin, 1909). In this work, we used high-spatial resolution NanoSIMS analysis to show that the core of rays from giant tetractines of L. johnstoni is significantly depleted in Mg compared to the adjacent envelope. The measured Mg/Ca ratio in the 3.5–8 lm wide core is 70–90 mmol/mol compared to 125 to 130 mmol/ mol in the envelope. This compositional span is in agreement with the range of values obtained for spicules of others calcareous sponge species (Table 1). Moreover, preliminary SXM analyses indicate that the core of rays of giant tetractines is also enriched in S compared with the envelope. The S detected may be in the form of sulfated polysaccharides embedded in the organic matrix of the biomineral, as it is often the case for other biocarbonates like the aragonitic skeleton of scleractinian corals (Cuif et al., 2003). We did not detect any periodic variation in the Mg and S concentration in the concentric layers, which might have been expected considering the banding pattern observed at the microstructural scale. These results contrast with those of Jones and James (1969) who did not detect fluctuations in the distribution of Ca, Mg, Sr and S from the center towards the periphery of transverse and longitudinal sections of giant triactines of L. nivea. The relatively low spatial resolution and sensitivity of the electronic micro-probe they used likely provide an explanation for this discrepancy. Using energy-filtering TEM on crushed triactines of P. heteroraphis, Sethmann et al. (2006) observed a heterogeneous distribution of Mg. However, because the sample was crushed into powder and the analysis was conducted at the nanoscale, it was not possible to correlate these Mg fluctuations with potential microstructural features like a core or a banding pattern. Our results confirm the pioneer hypothesis of Von Ebner (1887), suggesting that the central part of rays may be compositionally distinct from the peripheral part of rays. Our results also support the hypothesis of Jones (1970) who suggested, based on the fact that the relatively small spicules of L. nivea and A. paulini contained slightly less Mg than larger spicule types (Jones and Jenkins, 1970), that the material first secreted (which we refer to as the core) was depleted in Mg compared to the later deposits (which we refer to as the envelope). Table 1 Compared magnesium concentration in calcareous sponge spicules. Species Mg/Ca (mmol/mol) in the spicules References Leuconia johnstoni 70–90 in the core and 125–130 in the envelope 104–114 This work, measured in giant tetractines with the NanoSIMS Jones and Jenkins (1970) Leuconia nivea Leucandra (=Leuconia) pumila Clathrina coriacea Sycon ciliatum Grantia compressa Amphiute paulini Leucosolenia complicata Leucosolenia eleanor Leucilla (=Rhabdodermella) nuttingi Clathrina contorta 98 Sycon sp Kebira uteoides Pericharax heteroraphis Leucetta sp 125–135 140–180 100–110 90–120 129 52–54 78 79–90 82 65 69 160 107 The existence of a compositionally distinct core may also be suggested for spicules of other calcareous sponge species, based on reports of central etching pits sometimes observed on ray sections, for example in giant triactines of L. nivea treated with 10% D-tartaric acid for 30 s to 3 min (Jones and James, 1972; Ledger and Jones, 1991). We did not see such corrosion figure in the giant tetractines of L. johnstoni etched with 0.1% formic acid during 2 min 15 s: in fact, no distinct microstructure was observed in the spicule center, which could be related to the size of the compositionally differentiated core, corresponding to the 2–4 first layers around the axis of actines. As emphasized by Jones and James (1972), the etching pattern may differ depending upon the etching agent, its concentration and its time of application. Minchin (1898) and Minchin and Reid (1908) observed in decalcified calcareous sponge spicules from several species (treated with picric, nitric, acetic or hydrochloric acid), the presence of an axial filament stained with nigrosin or indulin, suggesting an organic phase. However, Jones (1967) and then Ledger and Jones (1991) established that this putative organic axial filament was an artifact caused by contracted remnants of the decalcified spicule organic sheath. In cryofractured or etched sections of spicules of L. johnstoni examined with SEM, we did not observe any organic axial filament. Today, the general consensus is the absence of an organic axial filament in calcareous sponge spicules (Sethmann and Wörheide, 2008), unlike the proteinaceous axial filament observed in all siliceous sponge spicules (Uriz et al., 2003; Uriz, 2006). Raman mapping results show that the giant tetractines of L. johnstoni, as well as its smaller minute tetractines, microdiactines, pugioles and triactines, consist entirely of magnesian–calcite. There are no differences in terms of crystallography between the core and the envelope of the spicules. In contrast, the spicules of Clathrina sp (from Atlit, Israel) are reported to be composed of a calcitic core, enveloped by a thick layer of ACC, which may be covered by a thin calcitic layer (Aizenberg et al., 1996a, 2003). In our work, no ACC was detected with Raman microspectroscopy. It is possible that the presence of ACC in calcareous sponge spicules may be species specific, revealing different patterns of biomineralization among the spicule types in the Calcarea. We can also not rule out that a potential amorphous phase may have accidentally crystallized into calcite during the protocol of spicule preparation or during the Raman analyses. Based on our results it is not clear whether an unstable, transient, amorphous precursor phase of calcite might exist in calcareous sponges, as has been reported in several Metazoan groups like Echinoderms, Crustacea and Annelida (reviewed by Weiner et al., 2009). The higher Mg content we have detected in the spicule envelope as compared to its core could be a signature of transient ACC during spicule formation (Loste et al., 2003). The abundance of Mg may be regulated by a spatial difference in the composition of the intraspicular organic matrix. Wang et al. (2009) recently reported experimental in vitro regulation of Mg/Ca in carbonates with carboxylated acidic molecules, which are known to exist in sponge spicules (Aizenberg et al., 1996b). In any case, the tightly controlled spatial fluctuations we have observed in the spicule geochemical composition indicates a strong biological control exerted by the sponge organism on its skeletal biomineralization and can only be explained by differences in spicule-secreting sclerocyte cell types or metabolic activities, suggesting a two-step cellular growth process. Chave (1954) Aizenberg et al. (1995a) Sethmann et al. (2006) Uriz (2006) 4.3. Two-step cellular model of spiculogenesis for the giant tetractines of L. johnstoni Earlier optical and electron microscopy observations have provided the basis for a cellular model of calcareous sponge spicule formation for small monaxons, triactines and tetractines of sponge species of the genus Leucosolenia, Clathrina and Sycon (Minchin, 108 C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 Longitudinal section Growth direction Apical sclerocyte Lateral sclerocyte Concentric band = Core of ray (depleted in Mg and enriched in S) Envelope of ray (enriched in = Mg and depleted in S) Fig. 8. Cellular model of spiculogenesis of the giant tetractines of Leuconia johnstoni. Spicule actines grow by successive addition of concentric layers secreted by two types of formative sclerocytes: one apical sclerocyte ‘Founder-Cell’ depositing the core of ray which is depleted in Mg and enriched in S (axial growth), and lateral sclerocytes ‘Thickener-Cells’ depositing the envelope of ray, enriched in Mg and depleted in S (radial growth). 1909; Jones, 1970; Ledger and Jones, 1977). Two formative sclerocytes are required for the development of monaxons and for each ray of a triactine or the basal triradiate system of a tetractine. Therefore, the formation of triactines or of the basal triradiate system of tetractines begins within a group of three paired cells, termed a ‘‘sextet” by Minchin (1909). The fourth ray of a tetractine is added to the growing triradiate system after it has reached a certain size. Minchin (1909) stated that the extension in length of small monaxons and of each ray of small triactines or the basal triradiate system of a tetractine is assured by one of the two sclerocytes, named the ‘Founder-Cell’, in the course of its migration through the mesohyl. The extension in width is provided by the second sclerocyte, named the ‘Thickener-Cell’. Fluorescence studies with calcein-labeling have shown that spicule growth rates and mechanisms are different for each spicule type of calcareous sponges: triradiate spicules are deposited from the center towards the rays, whereas curved monaxons elongate unidirectionally, with separate, additional, lateral calcification (Aizenberg et al., 1996b; Ilan et al., 1996). We observed two formative sclerocytes associated with a microdiactine at an early stage of growth in the cryofractured choanosome of L. johnstoni. We also observed that only one sub-apical sclerocyte was associated per ray of a growing triactine or giant tetractine spicule in semi-thin sections of the sponge ectosome. Taking into account the huge size of rays of the L. johnstoni giant tetractines and their concentric growth pattern, we propose that each tetractine ray is generated by incremental growth involving more than two formative sclerocytes. A cellular model of giant tetractine formation is proposed in Fig. 8. The spatial heterogeneity we detected in spicule trace element composition is compatible with a spatial heterogeneity in cell type or activity. Indeed, the maximum width of immature rays of growing triactines or giant tetractines measured in the contact zone under the sub-apical sclerocytes varies between 4.5 and 10 lm, which corresponds to the size of the compositionally differentiated core of the giant tetractines (3.5 to 10 lm). We suggest that the depletion in Mg and enrichment in S observed in the core of rays reflects the activity of a sub-apical sclerocyte ‘founder cell’ controlling axial growth, working in a distinct mode compared to several lateral sclerocytes ‘thickener cells’ controlling radial growth and accretion of the ray envelope relatively enriched in Mg and depleted in S (Fig. 8). This model is supported by the two-stage model proposed for Sycon sp curved monaxon deposition based on fluorescence studies, with an apical cell forming a primary nucleation site and a thickener cell spatially constraining lateral growth (Aizenberg et al., 1996b; Ilan et al., 1996). Our results indicate that sclerocyte cells would not only control skeletal morphology via the shape of the space and macromolecule content of the spicule microenvironment, but that they also control trace element incorporation into the calcareous spicule. Similar mechanisms of biological control of calcium carbonate biomineralization have been proposed for other phyla with calcified skeleton, such as the scleractinian corals (e.g. Meibom et al., 2008). Future studies may involve detailed ultrastructural observations of the sclerocyte-spicule interface and dynamic studies with stable isotope labeling of Mg trace element incorporation into the calcitic skeleton in order to provide information on the fundamental mechanisms of the biological control of marine carbonate biomineralization. Acknowledgments We thank the director and staff of the Concarneau Marine Station of the MNHN for help during field collection of sponges and for providing access to aquarium facilities. A number of colleagues are thanked for inspiring discussions and technical help: Pr. Claude Lévi, Karim Benzerara, Chloé Brahmi, Maciej Mazur, Sylvain Pont (Direction des Collections du MNHN) and Gérard Mascarell (Plateforme de Microscopie Electronique du MNHN). The National NanoSIMS facility at the Muséum National d’Histoire Naturelle was established by funds from the CNRS, Région Île de France, Ministère délégué à l’Enseignement supérieur et à la Recherche, and the Muséum itself. J. Stolarski’s research was supported by funds from the Direction des Collections du MNHN and from the Polish Ministry of Science and Higher Education Project N307-015733. This work was funded in part by the Action Transversale du Muséum ‘Biomineralizations’. The Raman micro-spectrometer at IMPMC was funded by an ANR jeune chercheur grant (GeoCARBONS) to O. Beyssac. References Aizenberg, J., Hanson, J., Koetzle, T.F., Leiserowitz, L., Weiner, S., Addadi, L., 1995a. Biologically induced reduction in symmetry: a study of crystal texture of calcitic sponge spicules. Chem. Eur. J. 1, 414–422. C. Kopp et al. / Journal of Structural Biology 173 (2011) 99–109 Aizenberg, J., Hanson, J., Ilan, M., Leiserowitz, L., Koetzle, T.F., Addadi, L., Weiner, S., 1995b. Morphogenesis of calcitic sponge spicules: a role for specialized proteins interacting with growing crystals. FASEB J. 9, 262–268. Aizenberg, J., Addadi, L., Weiner, S., Lambert, G., 1996a. Stabilization of amorphous calcium carbonate by specialized macromolecules in biological and synthetic precipitates. Adv. Mater. 8, 222–226. Aizenberg, J., Ilan, M., Weiner, S., Addadi, L., 1996b. Intracrystalline macromolecules are involved in the morphogenesis of calcitic sponge spicules. Connect. Tissue Res. 34, 255–261. Aizenberg, J., Weiner, S., Addadi, L., 2003. Coexistence of amorphous and crystalline calcium carbonate in skeletal tissues. Connect. Tissue Res. 44, 20–25. Bavestrello, G., Cattaneo-Vietti, R., Cerrano, C., Sarà, M., 1993. Rate of spiculogenesis in Clathrina cerebrum (Porifera: Calcispongiae) using tetracycline marking. J. Mar. Biol. Assoc. UK 73, 457–460. Bavestrello, G., Cattaneo-Vietti, R., Cerrano, C., Giovine, M., Sarà, M., 1994. Rate of spiculogenesis in some common Mediterranean Calcispongiae: a tetracycline and 45Ca2+ labelling study. Ital. J. Zool. 61, 197–201. Bernard, S., Beyssac, O., Benzerara, K., 2008. Raman mapping using advanced linescanning systems: geological applications. Appl. Spectrosc. 62, 1180–1188. Bischoff, W.D., Sharma, S.K., MacKenzie, F.T., 1985. Carbonate ion disorder in synthetic and biogenic magnesian calcites: a Raman spectral study. Am. Mineral. 70, 581–589. Borojevic, R., Boury-Esnault, N., Manuel, M., Vacelet, J., 2002. Order Baerida Borojevic, Boury-Esnault & Vacelet, 2000. In: Hooper, J.N.A., Van Soest, R.W.M. (Eds.), Systema Porifera: A Guide to the Classification of Sponges. Springer, New York, pp. 1193–1200. Boury-Esnault, N., Rützler, K., 1997. Thesaurus of sponge morphology. Smithsonian Contrib. Zool. 596, 1–55. Bütschli, O., 1901. Einige Beobachtungen über Kiesel-und Kalknadeln von Spongien. Z. Wiss. Zool. 59, 235–286. Carter, H.J., 1871. A description of two new Calcispongiae (Trichogypsia, Leuconia). Ann. Mag. Nat. Hist. 8, 1–28. Chave, K.E., 1954. Aspects of the biogeochemistry of magnesium 1. Calcareous marine organisms. J. Geol. 62, 266–283. Cuif, J.P., Dauphin, Y., Doucet, J., Salome, M., Susini, J., 2003. XANES mapping of organic sulfate in three scleractinian coral skeletons. Geochim. Cosmochim. Acta 67, 75–83. Ilan, M., Aizenberg, J., Gilor, O., 1996. Dynamics and growth patterns of calcareous sponge spicules. Proc. R. Soc. Lond. 263, 133–139. Jones, W.C., 1967. Sheath and axial filament of calcareous sponge spicules. Nature 214, 365–368. Jones, W.C., James, D.W.F., 1969. An investigation of some calcareous sponge spicules by means of electron probe micro-analysis. Micron 1, 34–39. Jones, W.C., 1970. The composition, development, form and orientation of calcareous sponge spicules. Symp. Zool. Soc. Lond. 25, 91–123. Jones, W.C., Jenkins, D.A., 1970. Calcareous sponge spicules: a study of magnesian calcites. Calcified Tissue Int. 4, 314–329. Jones, W.C., James, D.W.F., 1972. Examination of the large triacts of the calcareous sponge Leuconia nivea Grant by scanning electron microscopy. Micron 3, 196– 210. Ledger, P.W., 1974. Types of collagen fibres in the calcareous sponges Sycon and Leucandra. Tissue Cell 6, 385–389. Ledger, P.W., 1975. Septate junctions in the calcareous sponge Sycon ciliatum. Tissue Cell 7, 13–18. 109 Ledger, P.W., Jones, W.C., 1977. Spicule formation in the calcareous sponge Sycon ciliatum. Cell Tissue Res. 181, 553–567. Ledger, P.W., Jones, W.C., 1991. On the structure of calcareous sponge spicules. In: Reitner, J., Keupp, H. (Eds.), Fossil and Recent Sponges. Springer, Berlin, pp. 341– 359. Loste, E., Wilson, R.M., Seshadri, R., Meldrum, F.C., 2003. The role of magnesium in stabilising amorphous calcium carbonate and controlling calcite morphologies. J. Cryst. Growth 254, 206–218. Manuel, M., Borchiellini, C., Alivon, E., Parco, Y., Boury-Esnault, J., 2003. Phylogeny and evolution of calcareous sponges: monophyly of calcinea and calcaronea, high level of morphological homoplasy, and the primitive nature of axial symmetry. Syst. Biol. 52, 311–333. Meibom, A., Cuif, J.P., Hillion, F., Constantz, B.R., Juillet-Leclerc, A., Dauphin, Y., Watanabe, T., Dunbar, R.B., 2004. Distribution of magnesium in coral skeleton. Geophys. Res. Lett., 31. doi:10.1029/2004GL021313. Meibom, A., Cuif, J.P., Houlbreque, F., Mostefaoui, S., Dauphin, Y., Meibom, K.L., Dunbar, R., 2008. Compositional variations at ultra-structure length scales in coral skeleton. Geochim. Cosmochim. Acta 72, 1555–1569. Minchin, E.A., 1898. Materials for a monograph of the ascons. I. On the origin and growth of the triradiate and quadriradiate spicules in the family Clathrinidae. Quart. J. Microsc. Sci. 40, 469–588. Minchin, E.A., Reid, D.J., 1908. Observations on the minute structure of the spicules of calcareous sponges. Proc. Zool. Soc. Lond. 78, 661–676. Minchin, E.A., 1909. Sponge-spicules. A summary of present knowledge. Ergeb. Fortschr. Zool. 2, 171–274. Picton, B.E., Morrow, C.C., 2007. Encyclopedia of Marine Life of Britain and Ireland. <http://www.habitas.org.uk/marinelife/species.asp?item=C570> (visited the 10/05/10). Sethmann, I., Hinrichs, R., Wörheide, G., Putnis, A., 2006. Nano-cluster composite structure of calcitic sponge spicules – a case study of basic characteristics of biominerals. J. Inorg. Biochem. 100, 88–96. Sethmann, I., Wörheide, G., 2008. Structure and composition of calcareous sponge spicules: a review and comparison to structurally related biominerals. Micron 39, 209–228. Simpson, T.L., 1984. The Cell Biology of Sponges. Springer, New York. Stolarski, J., Mazur, M., 2005. Nanostructure of biogenic versus abiogenic calcium carbonate crystals. Acta Palaeontol. Pol. 50, 847–865. Uriz, M.J., Turon, X., Becerro, M.A., Agell, G., 2003. Siliceous spicules and skeleton frameworks in sponges: origin, diversity, ultrastructural patterns, and biological functions. Microsc. Res. Tech. 62, 279–299. Uriz, M.J., 2006. Mineral skeletogenesis in sponges. Can. J. Zool. 84, 322–356. Urmos, J., Sharma, S.K., Mackenzie, F.T., 1991. Characterization of some biogenic carbonates with Raman spectroscopy. Am. Mineral. 76, 641–646. Von Ebner, V., 1887. Über den feineren Bau der Skelettheile der Kalkschwämme nebst Bemerkung uber Kalkskelete überhaupt. Sber. Akad. Wiss. Wien. Abt. 1, 55–149. Wang, D., Wallace, A.F., De Yoreo, J., Dove, P.M., 2009. Carboxylated molecules regulate magnesium content of amorphous calcium carbonates during calcification. PNAS 106, 21511–21516. Weiner, S., Dove, P.M., 2003. An overview of biomineralization processes and the problem of the vital effect. Rev. Mineral. Geochem. 54, 1–29. Weiner, S., Mahamid, J., Politi, Y., Addadi, L., 2009. Overview of the amorphous precursor phase strategy in biomineralization. Front. Mater. Sci. China 3, 104– 108.