Introduction

Bone, as a biomineral, is composed of an organic matrix (mainly collagen, ≈25 % w/w), a mineral phase (Ca deposits, ≈65 % w/w), and 10 % water [1, 2]. The dominant component of the inorganic scaffold is calcium phosphate, more specifically calcium hydroxyapatite (Ca–HA, Ca5[PO4]3[OH]). However, bone mineral contains, in addition, around 5 % w/w carbonate, such as carbonate-fluorapatite (francolite) [3]. Earlier studies suggested that carbonate exists in the apatite crystal lattices as CO3 2− ions, perhaps substituting for PO4 3− and/or OH− ions [reviewed in 4]. Topographically, carbonate has been localized in vivo, even as crystalline calcium carbonate (CaCO3) within medullary bone [5]. It has been determined that the bone CO2 compartment consists of 30 % bicarbonate (HCO3 −) and 70 % carbonate (CO3 2−) [6]. In humans/mammals, besides Ca-phosphate-based skeletal elements, biomineral deposits predominantly formed of CaCO3 exist, e.g., the biomineralized otoliths in the vestibular labyrinth of the ear [7, 8]. In otoliths an organic matrix protein, covered with 90–95 % of CaCO3, exists that has been identified as otolin. In addition, otoconins, calcium-binding proteins, exist in the vestibular system, onto which CaCO3 deposition is initiated [9].

Based on the observation that in humans, as a result of metabolic acidosis occurring during aging, a counteraction to the retained endogenous acids proceeds that might result in anabolic processes from the skeleton, followed by a decrease in bone mass; thus, the diet of postmenopausal women was supplemented with potassium bicarbonate. The results revealed that those subject groups show improved calcium and phosphorus balance and reduced bone resorption, resulting in an increased rate of bone formation [10]. Already, bicarbonate supplementation in the diet had been described to improve the nutritional status of patients with chronic kidney disease [11].

Both CaCO3 and Ca-phosphate have been described to precipitate under physiological conditions [12]. Already in this early study it was reported that CaCO3 precipitation was prevented in the presence of low phosphate concentrations, and in turn, it has been suggested that in the biological system metabolic processes that locally raise the bicarbonate level and simultaneously lower the phosphate concentration are required for mineral deposition. HCO3 − accumulates on the surfaces of cell membranes in the vicinity of the Na+-driven Cl−/HCO3 − exchanger (NBC) [13], the Na+–HCO3 − cotransporter [14], and the Na+-independent Cl−/HCO3 − exchanger (AE) [15], all of them being (primarily) involved in the regulation of intracellular pH. Interestingly, some of those channels are connected with carbonic anhydrase (CA), like the AE3 Cl−/HCO3 − exchanger and CA XIV [16]. CA enzymes catalyze the rapid interconversion of one carbon dioxide molecule and water to HCO3 − and one proton. Under physiological conditions, the CA-driven HCO3 − formation is the rate-limiting reaction during carbonate deposition in animals, e.g., in mollusk shells [17]. Among mammalian CAs the most abundant isoforms are the cytosolic CA II, with about 75 %, and the cytosolic CA I, with about 20 % [18]. One considerable fraction of those CAs is cell membrane-associated, both for CA II [19] and for other isoforms [20].

Recently, more detailed data on the initiation of mineralization in osteoblasts, the bone-forming cells, have been elaborated. Ca deposition in osteoblasts starts intracellularly in calcium-containing vesicles under the formation of poorly crystalline carbonated apatite deposits [21]. Evidence has been presented showing that amorphous Ca-phosphate particles are formed intracellularly using phosphate formed from polyphosphate (polyp) after hydrolytic degradation of this polymer [22], a process that is under enzymatic control. On the other hand, it has been proposed that the extracellular fluid is sufficiently saturated with respect to Ca2+ and phosphate, allowing Ca-phosphate to precipitate [23]. As for (almost) any kind of biological reaction that proceeds within a living biological system, enzymes are involved that allow also exergonic reaction to run under the formation of a chemical bond. In addition, in contrast to chemical reactions, which are usually performed under saturating conditions, biological/enzymatic processes proceed under unsaturated substrate conditions, where a subtle regulation of reaction velocity, by slight changes of the substrate, can be obtained. A first example that even inorganic biological skeletal elements, the siliceous spicules from sponges, are formed enzymatically is silicatein, the major proteinaceous component of spicules [reviewed in 24, 25]. Surely, not only enzymes but also interfaces between two phases can lower the activation energy, allowing exergonic reactions to occur, e.g., in atomic/molecular transport processes [26].

CAs have been identified not only in the anabolic bone-forming osteoblasts but also, though at higher levels, in the catabolic bone-resorbing cells, the osteoclasts [27]. CA II deficiency in humans is a rare autosomal recessive disorder that is characterized by the triad of osteopetrosis, renal tubular acidosis, and cerebral calcifications [28]. In view of this fact it came as a surprise that in vivo administration of CA inhibitors does not affect the function of osteoclasts [29]. Only little information is available about the potential therapeutic effect of CA inhibitors on bone anabolism [30]. Until now only a few CA activators have been identified, but none has been tested for its potential in the treatment of bone disorders [reviewed in 31–33]. This view has changed to some extent by the findings that (1) CaCO3 deposits might function as bio-seeds for Ca-phosphate precipitation onto bone-forming cells [34] and (2) the CAs are decisively involved in CaCO3 deposition in vitro [35] and likely in vivo [36]. The latter studies were performed with the calcareous sponge Sycon raphanus, which has a calcareous CaCO3-based skeleton. Because all metazoan phyla, including the oldest animals, the sponges (phylum Porifera), originate from one ancestor, the Urmetazoa [37], it was tempting to speculate that even the inorganic skeletons of the more evolved animals, e.g., Ca-phosphate-based skeletons, might have components and structures in common with this basal animal taxon.

It is well established that mineralization of osteoblasts, bone mineral-forming cells, is significantly induced by polyP in vitro even in the presence of orthophosphate [38]. In addition, polyP turned out to be an inducer of osteoblast-specific alkaline phosphatase. This finding is interesting in view of published data indicating that intracellularly polyP might be formed in the vesicles of bone-forming cells as Ca2+ salt, which is supposed to act as a potential precursor of carbonated HA [39].

In the present study we show that CA-driven CaCO3 deposition can be stimulated by CA activators both in vitro and in intact cell system (SaOS-2). As activator(s) we chose extracts from the sponge Suberites domuncula and one component, isolated from these extracts, quinolinic acid (QA). This effect is amplified by polyP. Mesenchymal stem cells, which differentiate to osteoblasts, are known to release a factor that attenuates QA toxicity [40]. The rationale behind this study was the finding that even siliceous demosponges form CaCO3 deposits [41], suggesting that also the siliceous sponges might be provided with both anabolic and catabolic biocalcite-forming enzymes and inhibitors of them.

Materials and Methods

Cultivation of SaOS-2 Cells

SaOS-2 cells (human osteogenic sarcoma cells [42]) were cultured in McCoy’s medium (Biochrom-Seromed, Berlin, Germany), lacking Na-bicarbonate but containing 2 mM l-glutamine and 1 mM CaCl2 [43]. The medium was supplemented with 10 % heat-inactivated fetal calf serum (FCS) and 100 units mL−1 penicillin/100 μg mL−1 streptomycin. Where indicated 20 mM Na-bicarbonate (NaHCO3) was added as well. The medium/serum was buffered with 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) to a pH of 7.4. The cells were incubated in 25-cm2 flasks or in six-well plates (surface area 9.46 cm2; Orange Scientifique, Braine-l’Alleud, Belgium) in a humidified incubator at 37 °C [44, 45]. Routinely, 3 × 105 cells/well were added (total volume 3 mL). Where indicated, cultures were supplemented with a mineralization activation cocktail (MAC), composed of 5 mM β-glycerophosphate, 50 mM ascorbic acid, and 10 nM dexamethasone [46]. The MAC was added 3 days after starting the experiments. Medium was changed every 3 days, and new MAC was added.

Where indicated with the respective assay, either S. domuncula extract (SubDo-extr) or QA was added at the given concentration.

Mineralization by SaOS-2 Cells In Vitro

The extent of mineralization in assays with SaOS-2 cells was determined quantitatively using the alizarin red S spectrophotometric assay [47, 48]. The amount of bound alizarin red S is given in nanomoles. Values were normalized to total DNA in the samples.

In one series of experiments the cultures, growing in culture wells, were stained directly on the coverslips with 10 % alizarin red S, after fixation with ethanol [44].

Incubation of SaOS-2 Cells with PolyP

SaOS-2 cells were incubated, under otherwise identical conditions, with polyP. As source, Na-polyP (average chain of approximately 40 phosphate units), obtained from Chemische Fabrik Budenheim (Budenheim, Germany), was used for the studies. To compensate for any effect caused by the chelating activity to Ca2+ of polyP after hydrolysis of the polymer to monomeric phosphate by phosphatases in vitro, polyP was mixed with CaCl2 in a stoichiometric ratio of 2:1 (polyP:CaCl2), as described [38]. The salt, designated as polyP (Ca2+ salt), was added at a concentration of 50 μM to the assays.

CA-Mediated CaCO3 Precipitation Assay

For the CaCO3 precipitation assay the ammonium carbonate diffusion method was used, as described [36]. In a 3.4-L large desiccator CO2 vapor was generated from a 1-M NH4HCO3 (A6141; Sigma, Taufkirchen, Germany) solution, which had been placed into the lower compartment of the desiccator. CO2 diffused into the upper compartment, which contained in eight-well square chamber slides 50 mM CaCl2 in 30 mM Tris-HCl (pH 8.3). The slides with an edge length of 10 mm (Lab-Tek II slides 154534) came from Thermo Scientific (Dreieich, Germany). The surface area of the 500-μL assay was 0.7 cm2. Where indicated, the recombinant human CA2 enzyme, expressed in Escherichia coli (C6624, Sigma), with a specific activity of ≈5,000 units/mg [49, 50], was added at a concentration of 35 W-A units (10 μg)/500 μL of CaCl2 to the assays. Prior to the addition of the Tris/CaCl2 reaction vessels, the desiccator was preequilibrated with CO2. The formation of CaCO3 was quantitatively determined on the basis of the consumption of free Ca2+ using the EDTA titration procedure [36, 51].

Crystals were visualized using a light digital microscope (VHX-600 Digital Microscope) from KEYENCE (Neu-Isenburg, Germany).

Preparation of S. domuncula Extract

Freshly collected sponge specimens, dredged 5 km offshore from Rovinj (Adriatic Sea, Croatia), were cleaned and cut into small pieces. The material (1 kg) was homogenized in a Waring blender in the presence of 100 mL of ethanol; then, the slurry was extracted with ethyl acetate (3 × 250 mL) at room temperature. The combined ethyl acetate extracts were washed with water (3 × 200 mL), dried over Mg2SO4, filtered, and evaporated to dryness, yielding ≈4 g of dark tar-like residue. This material, termed SubDo-extr, was dissolved in dimethyl sulfoxide (DMSO) and subjected to testing in the cell culture. Control cultures received the carrier solvent alone (1.0 % DMSO).

Extraction and Identification of QA

QA was extracted as described [52]. The extract was shaken in a separation funnel with chloroform to eliminate nonpolar components. The residue was used for the determination of QA by high-performance liquid chromatography. The eluate was monitored with ultraviolet detection at 268 nm. At this wavelength the molar extinction coefficient (ε) is 4,000 M−1 cm−1 [53].

Scanning Electron Microscopy and Energy-Dispersive X-Ray Analysis

Scanning electron microscopic (SEM) analysis was performed with an SU 8000 (Hitachi High-Technologies Europe, Krefeld, Germany) microscope, employed at low voltage (<1 kV, analysis of near-surface organic surfaces) [38]. The SEM was coupled to an XFlash 5010 detector, an X-ray detector that allows simultaneous energy-dispersive X-ray (EDX)-based elemental analyses. Likewise, the same combination of devices was used for higher-voltage (10 kV) analysis, during which the XFlash 5010 detector was used for element mapping of the surfaces of the deposits. The HyperMap database was used for interpretation, as described [54].

Additional Methods

The results were statistically evaluated [55]. DNA content was determined by application of the PicoGreen method, as described [44], using calf thymus DNA as a standard.

Results

Effect of Sponge Extracts and Sponge QA on CA-Mediated CaCO3 Deposit Formation

Under the conditions described in “Materials and Methods,” 20 % of the soluble CaCl2 salt was converted to insoluble CaCO3 after an incubation period of 240 minutes using the ammonium carbonate diffusion assay, in the absence of CA (Fig. 1a). Addition of 10 μg × mL−1 of SubDo-extr or of 10 μM QA did not significantly change the extent of CaCO3 formation; the amount of CaCO3 formation was 32.5 ± 4.9 % in the presence of SubDo-extr and 28.4 ± 3.7 % in the presence of QA after the 240-minute incubation period.

Fig. 1
figure 1

Formation of CaCO3 in the ammonium carbonate diffusion assay in the absence a or presence b of 35 W-A units CA per 500-μL crystallization assay. The reaction assays either remained free of additional compound(s) (filled square) or were supplemented with 10 μg × mL−1 of S. domuncula extract (SubDo-extr, filled circle) or with 10 μM quinolinic acid (QA, filled triangle). At the indicated time points samples were taken and the free Ca2+ concentration was determined. The decrease in the concentration of free Ca2+ indicates an increase in deposited CaCO3. Samples of six parallel determinations were quantitated; means ± SD are given. *p < 0.05

Full size image

In contrast, in the presence of CA the rate of CaCO3 formation increased significantly to 38.4 ± 5.2 % over 240 minutes. Even more, if CA was administered together with QA, a significant increase of mineralization was seen, which amounts to 69.2 ± 7.8 %. In the presence of SubDo-extr, the increase was 83.1 ± 8.1 % (Fig. 1b).

Rate of CaCO3 Crystal Formation Depending on the Presence of CA, SubDo-extr, and QA

No crystals were seen at the beginning of the reaction in the diffusion/desiccator assays (Fig. 2a, b). In the absence of CA, only very scarce, 42 ± 3 μm (n = 10), large crystals were seen after 30 minutes (Fig. 2c); in contrast, in the presence of CA 105 ± 16 μm-sized crystals were regularly seen (Fig. 2d). After an incubation period of 60 minutes, the size of the crystals increased to 122 ± 25 μm in both assays (Fig. 2e, f), again with a much higher abundance in the assays with CA. After a further, extended incubation for 180 min (Fig. 2g, h), the crystals, especially in the assays containing CA, further increased in size. Especially in the assays containing CA two forms of calcitic crystals were seen: (1) round-shaped deposits and (2) rhombohedral prisms. The round-shaped deposits consisted of the metastable vaterite with the characteristic peaks of FTIR spectra at 875 and 744 cm−1, while the prisms consisted of calcite with FTIR peaks at 873 and 711 cm−1, as determined by us previously [35].

Fig. 2
figure 2

Intensified calcitic crystal formation in the carbonate diffusion assay in the presence of 35 W-A units CA/500 μL (b, d, f, h) compared to assays without the enzyme (a, c, e, g). Two different morphologies of the calcitic crystals can be distinguished, the round-shaped vaterite deposits (d) and the rhombohedral prisms of calcite (p). Light microscopic images

Full size image

In the presence of SubDo-extr, 10 μg × mL−1, the abundance of the calcitic crystals increased significantly in the assays. While at time zero again no crystals were seen (Fig. 3a, b), the number of crystals increased gradually during incubation from 30 minutes (Fig. 3c, d), over 60 min (Fig. 3e, f), to 180 minutes (Fig. 3g, h). The number of crystals in the CA-containing assays significantly higher (Fig. 3d, f, h) compared to the number in those lacking the enzyme (Fig. 3c, e, g).

Fig. 3
figure 3

Crystallization studies in the carbonate diffusion assay were performed again in the absence (a, c, e, and g) or presence (b, d, f, and h) of 35 W-A units/500 μL of CA. All assays contained, in addition, 10 μg × mL−1 S. domuncula extract. Some vaterite deposits (d) and rhombohedral prisms of calcite (p) are marked

Full size image

A similar effect on the number of crystals formed can be seen if the crystallization studies are performed with 10 μM QA. Again, the prevalence of the distinctly formed crystals increased during the 180-minute incubation period (Fig. 4), especially in the assays containing the CA (Fig. 4d, f, h) compared to the assays lacking CA (Fig. 4c, e, g). However, in addition to the crystals, small-sized deposits were seen together with the crystals that accumulate on the aqueous/air interphase.

Fig. 4
figure 4

In vitro crystallization assay in the absence (a, c, e, g) or presence (b, d, f, h) of 35 W-A units/500 μL of CA; assays were supplemented with 10 μM quinolinic acid. Some vaterite deposits (d) and rhombohedral prisms of calcite (p) are marked

Full size image

Effect of PolyP and CaCO3 on Mineralization of SaOS-2 Cells

As expected, addition of MAC to SaOS-2 cells caused an increase in mineralization, which became significant after an incubation period of 3 days. The extent of the MAC-caused increase in mineral deposition was lower in the absence of HCO3 − in the assays (Fig. 5a) compared with the increase measured in the presence of HCO3 − (Fig. 5b). At day 5 the ratio/increase of mineral deposition between differentiated (+MAC) and undifferentiated (−MAC) cells was significantly higher in the presence of HCO3 − (Fig. 5b, 5.3-fold increase) compared to the same ratio in the absence of HCO3 − (Fig. 5a, 1.3-fold increase). In contrast, if polyP (Ca2+ salt) was added to the cultures, the MAC-caused stimulation, e.g., at day 5, in the presence of polyP (Ca2+ salt) was significantly higher in the presence of HCO3 − (a calculated 3.6-fold increased ratio, Fig. 5b) compared to experiments in the absence of HCO3 − (1.5-fold, Fig. 5a). Even more, the polyP (Ca2+ salt)-mediated stimulatory effect was even a potentiating one: the increase caused by MAC was 4.8-fold and that by polyP (Ca2+ salt) in the presence of MAC was 17.6-fold, with respect to the effect seen in the absence of MAC. As measured, the basal mineral deposition in the controls (without MAC), cultured in the presence of HCO3 −, resulted in a decrease during the 5-day incubation period. At present, we attribute this finding to a dissolution of the initial amorphous Ca-carbonate, which is highly soluble [56].

Fig. 5
figure 5

Effect of polyP (Ca2+ salt) on the extent of mineralization of SaOS-2 cells in the absence a or presence b of HCO3 −; alizarin red S spectrophotometric assay. As indicated, experiments were performed in the absence (−MAC, open bars/left hatched in black) or presence (+MAC, gray-colored bars/right hatched in white) of MAC. As marked, polyP (Ca2+ salt) was added in the absence of MAC (−MAC + polyP, left hatched in black) or presence (+MAC + polyP, right hatched in white) of MAC. Standard errors of the means are shown (n = 6); *p < 0.05. MAC mineralization activation cocktail, polyP polyphosphate, AR alizarin red

Full size image

Effect of SubDo-extr and QA on Biomineralization of SaOS-2 Cells

In assays supplemented with 20 mM NaHCO3, both SubDo-extr and QA did not significantly change the low extent of mineralization in the absence of MAC, ≈0.1 nmol alizarin red S staining/1 μg DNA at days 1 and 3, or in medium lacking HCO3 −, ≈0.1 nmol/μg (data not shown). In the presence of MAC, SubDo-extr and QA caused only a small increase in mineralization of SaOS-2 cells, as demonstrated by staining with alizarin red S (Fig. 6). However, in the presence of both MAC and polyP (Ca2+ salt) the two natural components, SubDo-extr and QA, significantly increased the extent of mineral deposition onto SaOS-2, as determined by the quantitative alizarin red spectrophotometric assay (Fig. 7). This increase was especially pronounced at day 3; at this time point the extent of biomineralization, after addition of 5 μg SubDo-extr or of 10 μM QA per assay, was amplified to 2.3-fold or 2.1-fold, respectively. At day 5 of mineralization, the increase was still significant but less pronounced.

Fig. 6
figure 6

Alizarin red staining of cells incubated in medium supplemented with HCO3 − in either the absence (–MAC) or the presence (+MAC) of MAC in dependence on the S. domuncula compounds. Assays were incubated without the sponge compound(s) (−QA/−SubDo-extr) or with the compounds (+QA/+SubDo-extr). The incubation was terminated after 5 days, and coverslips were stained with 10 % alizarin red S. MAC mineralization activation cocktail, QA quinolinic acid, SubDo-extr S. domuncula extract

Full size image
Fig. 7
figure 7

Stimulatory effect of coincubation of SaOS-2 cells with MAC and polyP (Ca2+ salt) in medium supplemented with HCO3 − (BiCa), if administered together with 10 μM QA (stippled bars) or 5 μg per assay of SubDo-extr (filled bars). Control values without QA or SubDo-extr are given in comparison (right hatched in white). Means (n = 6) with *p < 0.05 are given. MAC mineralization activation cocktail, QA quinolinic acid, SubDo-extr S. domuncula extract, AR alizarin red, polyP, polyphosphate, BiCa bicarbonate

Full size image

Carbon-Rich Deposits Formed onto SaOS-2 Cells after HCO3 − Exposure: Element Mapping

The cells were incubated in the presence of HCO3 − and MAC for 5 days and then, after fixation, subjected to element mapping, using two-dimensional element mapping in the SEM/EDX setup adjustment (Fig. 8). As shown in Fig. 8a, the nodules are highlighted at the secondary electron image setting. Mapping studies revealed that within the crystallite nodules a strong increase of the impulses for Ca (Fig. 8b), P (Fig. 8c), and C (Fig. 8d) lights up.

Fig. 8
figure 8

Semiquantitative determinations of the accumulation of calcium (Ca), phosphorus (P), and carbon (C) around growing nodules (no), which have been distinguished by secondary electron imaging (SEI). Element mapping was performed by scanning electron microscopy-based energy-dispersive X-ray mapping. a Secondary electron image. Mapping for Ca (b), P (c), and C (d). The intensity scale for the pseudocolor documentation is given below the images; dark/blue low intensities, red/white high intensities. SaOS-2 cells were incubated in the presence of HCO3 − and mineralization activation cocktail for 5 days (Color figure online)

Full size image

The same series of experiments were performed for SaOS-2 cells that had been grown in the presence of HCO3 − and MAC, again for 5 days, if to this experimental setup 10 μM QA is added as well (Fig. 9). The growing nodules can be distinguished by secondary electron imaging (Fig. 9a). Within the areas of nodules a strong increase of the impulses for Ca (Fig. 9b), P (Fig. 9c), and especially C (Fig. 9d) is seen. It should be particularly mentioned that the regions with increased C signals expand those regions highlighting the nodules (secondary electron imaging) and Ca or P.

Fig. 9
figure 9

Mapping studies for the elements calcium (Ca), phosphorus (P), and carbon (C) around growing nodules (no) onto SaOS-2 cells. Cells were grown for 5 days in the presence of HCO3 − and mineralization activation cocktail, as well as 10 μM quinolinic acid. a Secondary electron image. Mapping for Ca (b), P (c), and C (d). It is obvious that the areas of increased impulses for C (d) extend the areas of increased signals for Ca (b) and P (c)

Full size image

Discussion

The biomineralization process in osteoblasts starts with the intracellular accumulation of the respective mineralizing ions, HCO3 −, CO3 2−, PO4 3−, and Ca2+, most likely in vesicles, from where they are sequestered to the sites where mineral deposition occurs. Basically, three mechanisms have been proposed to explain early bone mineral formation [21]: (1) a cell-independent process, which suggests that the ions are deposited extracellularly onto charged noncollagenous proteins; (2) a cell-controlled mechanism, via which the intracellular vesicles bud from the plasma membrane and subsequently form Ca-phosphate deposits; and (3) already intracellularly the first amorphous Ca deposits are formed, which are deposited in a further step onto collagen fibrils, where they are transformed to more crystalline apatite platelets. Recent studies favor the last proposal, implying that amorphous mineral precursors act as bio-seeds for bone mineralization [57] followed by a crystallization process [58].

The process of bone mineral formation proceeds on the surface of the bone-forming cells, either in close spatial connection with collagen [59] or on other distinct nucleating sites on the cells [60]. In the latter view it is proposed that bio-seed formation/nucleation starts with the formation of calcium phosphate aggregates that subsequently become densified. After a transient generation of an amorphous precursor phase, the deposits might transform into the crystalline apatite. However, if the biomimetic process is performed under physiological conditions, the mineral deposits are formed of oriented nanocrystallites of carbonated HA. In the amorphous, nanostructured state a stoichiometric replacement of the ionic substitutions can occur, e.g., CO3 2− replacing PO4 3− and Mg2+ replacing Ca2+ [61, 62]. Based on chemical and spectroscopic analyses it has been proposed that bone formation starts from collagen fibrils around which poorly crystalline carbonated apatite crystals, carbonate apatite aggregates, are deposited [21, 63]. These results revealed that under in vitro conditions osteoblasts contain low concentrations of CO3 2− ions within their Ca2+/PO4 3− mineral phase.

As intracellular precursor for Ca-phosphate formation, PO4 3−, deposited as polyP polymer, has been proposed [39]. This polymer has previously been shown to act anabolically on Ca-phosphate formation in SaOS-2 cells in vitro [38, 64]. Even more, this polymer acts as an inducer for the expression of the osteoblast-specific alkaline phosphatase [38]. The experiments suggested that polyP is deposited in vesicles as insoluble Ca2+-polyP salt [39].

Since both Ca-phosphate formation and CaCO3 formation are exergonic reactions and, in analogy to (most) metabolic reactions proceeding in living systems, run at nonsaturating conditions, we postulated that mineralization processes of bone material are also enzyme-driven. Focusing on CaCO3 CA is assumed to be the prime enzyme involved in CaCO3 deposition. Recently, we described that HCO3 − causes an induction of CA in SaOS-2 cells and a subsequent increase in mineralization [34]. Therefore, we screened for CA activators and used sponges as a biological starting material. Control of CA in these animals is crucial for the maintenance of their intermediate metabolism as well as their survival since they live in an environment that is poor in CO2 [65]. While an intensive and successful search for CA inhibitors has been launched, only comparably minor activities have been performed to screen for CA activators [32]. We screened organic extracts from sponges, and among them the ethyl-acetate extract from the demosponge S. domuncula showed a distinct activation of the CA enzyme during in vitro CaCO3 formation. In the presence of 10 μg mL−1 a severalfold increase of the transformation of free Ca2+ into insoluble CaCO3 was found. No distinct differences in the crystal form and state could be detected. During the 240-minute incubation period the enzymatically formed CaCO3 deposits exist in the calcite form, based on the FTIR spectra. Functionally important was the finding that in the presence of HCO3 − and polyP SubDo-extr caused a significant increase in the extent of mineralization in the in vitro assay to 232 % compared to the controls (100 %).

Among the secondary metabolites isolated from S. domuncula [66, 67], we selected QA since this S. domuncula bioactive compound [52] comprises, with its N-heteroatom in the pyridine backbone as well as the dicarboxylic acid side chains, two potential interacting groups with the Zn-containing CA. The zinc prosthetic group in the CA is coordinated in three positions by histidine side chains, by His94, His96, and His119 [68]. The fourth coordination position at Zn is occupied by one water molecule and causes there a polarization of the hydrogen–oxygen bond, resulting in an increased negativity of the oxygen and a weakening of the bond to the Zn atom. A fourth histidine, His64, is placed close to the substrate of water and accepts a proton from the water molecule through general acid-base catalysis under formation of a hydroxide bound to the zinc (Fig. 10). These studies have been performed with human crystallized CA. In analogy to amino acid derivatives [69], the formation of a hydrogen bridge bond can be postulated between the N-heteroatom of QA and the water130 that facilitates the rate-limiting proton transfer step. The dicarboxylic acid side chains of QA could interact, via their conjugated side chains, with the enzyme-bound Zn ion, like in the bipyridyl Zn chelators [70] (Fig. 10). In the future, we plan to modify the position of the carboxylic acid side chain at the pyridine backbone to modulate the basic heterocyclic organic compound through an induction of the delocalization of the π electron within the ring structure.

Fig. 10
figure 10

Proposed interaction of quinolinic acid (QA, in red) with the active center of the human carbonic anhydrase. Three His residues are interacting with the central Zn2+ ion. The three-hydrogen bond network, formed by Wat-150, Wat-129, and Wat-130, is indicated [modified after 75]. It is proposed that QA interacts with its N-heteroatom to Wat-130 and its one carboxylic acid group to Zn2+ (Color figure online)

Full size image

The mineralization studies with alizarin red also revealed that QA causes a significant upregulation of the biomineralization process by SaOS-2 cells in vitro. This effect is especially pronounced during the initial phase of mineralization, during the first 3 days. At this time point the increase caused by QA is 210 % compared to the 100 % of the controls. Studies using SEM coupled with EDX analysis confirmed our previous findings [34] that carbon is strongly accumulated in the mineral nodules on the surface of the SaOS-2 cells, whether incubated with SubDo-extr or with QA.

The experimental data given here show that at the concentration of 10 μM (0.85 μg/500 μL assay) chosen, QA causes a similarly strong activation compared to the 10 μg × mL−1 of SubDo-extr. This concentration of QA is very low and does not cause apoptotic signs to neuronal cells in vitro [71]. The higher efficacy of QA, on the basis of weight, indicates that it is one of the major constituents of the extract, activating the CA.

As outlined, polyP, likely as polyP (Ca2+ salt), is stored in bone cells [39]. Released from the vesicles, polyP undergoes enzymatic degradation into Ca2+ and PO4 3−, two precursors required for Ca-phosphate formation in SaOS-2 cells [38], via alkaline phosphatase [64]. In a second approach the effect of polyP on the mineralization of SaOS-2 cells in the absence and presence of HCO3 − was tested. We demonstrated for the first time that an enhancement of mineralization of polyP (Ca2+ salt) by HCO3 − in the presence of MAC occurs. This finding is remarkable since polyP causes this effect in the presence of orthophosphate (Fig. 11), present in the culture medium. This finding implies that polyP, as such, causes an enhancing effect on mineral deposit formation in the polymeric state. The most likely explanation for this amplifying effect of polyP is a local extrusion and sequestration of Ca2+-polyP from cells adjacent to the bone-forming sites of the osteoblasts, which appears to be very likely [39]. The alkaline phosphatase activity is spatially localized in mitochondria [72] on the apical region of the (secretory) cell membranes [73]. As schematically outlined in Fig. 11, the extracellularly localized alkaline phosphatase, induced by polyP, in a positive auto-circle increasingly degrades Ca2+-polyP to Ca2+ and PO 3-4 , perhaps allowing Ca-carbonated phosphate to form onto the CaCO3 bio-seeds. The exchange of ions is by far energetically favored in amorphous carbonated phosphate; such a state is likely to exist for this mineral. Interestingly enough, PO4 3− enzymatically released from polyP negatively controls CaCO3 formation by inhibiting the CA [34].

Fig. 11
figure 11

Schematic representation of the early stages of the enzymatic synthesis of the bio-seeds formed during bone mineralization. Two osteoblasts are depicted, one of them (upper) being involved in polyP metabolism. After enzymatic polyP formation and its compartmentalization into vesicles, the charges of the polymer are neutralized by cations, e.g., Ca2+. Those vesicles are extruded and polyP (Ca2+ salt) is discharged. Driven by the membrane-bound CA, HCO3 − and subsequently CaCO3 deposits are formed, which act as bio-seeds for calcium carbonate/phosphate and hydroxyapatite deposition. PolyP activates alkaline phosphatase. The released orthophosphate, in turn, inhibits the CA. The CA on the other side is activated by HCO3 − and by S. domuncula extract and quinolinic acid. PolyP polyphosphate, ATP adenosine triphosphate, ADP adenosine diphosphate, ALP alkaline phosphatase, CA carbonic anhydrase, NBC Na+-driven Cl−/HCO3 − exchanger, AE Na+-independent Cl−/HCO3 − exchanger

Full size image

The data are schematically summarized in Fig. 11. For reasons of simplicity, the polyP (Ca2+ salt) storing cell is separated from the biomineralizing bone cell. PolyP is synthesized/degraded by a series of enzymes [74] and stored in vesicles as polyP (Ca2+ salt). The vesicles are sequestered into the extracellular space, where polyP induces the gene expression of alkaline phosphatase. This enzyme in turn accelerates the hydrolysis of polyP (Ca2+ salt), resulting in a spatial upregulation of the concentration of PO4 3− and Ca2+, the building blocks for Ca-phosphate formation. The adjacent bone mineralizing cell again spatially and locally synthesizes CaCO3 bio-seeds in close association of the cell membrane, either extracellularly or intracellularly under the control of CA. This enzyme catalyzes the rate-limiting step of HCO3 − formation. The exchange of HCO3 − between the extracellular and intracellular compartments is made possible by transporters and exchangers, e.g., NBC or AE. After formation of the CaCO3 bio-seeds, the activity of CA is downregulated by PO4 3−, as outlined before [38]. In the subsequent step the HCO3 − bio-seeds are partially transformed to Ca-carbonated phosphate, followed by deposition of Ca-phosphate. The activity of the two regulator enzymes, alkaline phosphatase and CA, is under feedback control, offering new targets for pharmacological intervention of bone diseases. While the alkaline phosphatase is activated by polyP, CA is inhibited by PO4 3−. Besides the physiological control of the enzyme activities, CA can be stimulated by secondary metabolites, produced by sponges, e.g., QA. Future studies will be performed to resolve the regulatory cross-talk between these two enzymes, CA and alkaline phosphatase, during the initial phase of bio-seed formation.