Electron and ion microprobe analysis of calcium distribution and transport in coral tissues

Calcium is involved in numerous cellular functions and, in the ionised state, is usually maintained at an intracellular concentration of less than 1μmol l^–1 within the cytosol. The majority of intracellular Ca is probably bound to calcium-binding proteins and sequestered in intracellular compartments such as the endoplasmic reticulum(Pozzan et al., 1994). Although much is known about calcium signalling(Brini and Carafoli, 2000) and membrane transport systems for calcium(Hoenderop et al., 2005),there is a paucity of information on intracellular Ca transport in relation to calcification and biomineralisation. Relatively little information is available on total intracellular Ca concentrations in skeletogenic tissues even though it may be expected that intracellular Ca concentration would be related to skeletal Ca deposition (e.g. Bordat et al., 2004).

Scleractinian corals form calcium carbonate skeletons and show extremely high rates of skeletal deposition and calcium transport. The processes involved in the formation of the CaCO₃ skeleton are not well understood (Cohen and McConnaughy,2003; Allemand et al.,2004). The nature of the organism (essentially four cell layers closely overlying a massive exoskeleton) renders obtaining physiological information difficult. Investigations of calcification in corals using metabolic and enzyme inhibitors (Marshall,1996; Tambutté et al.,1996) have indicated the involvement of active Ca²⁺transport in coral epithelia. It is generally accepted that active Ca²⁺ transport occurs in the aboral epithelia immediately adjacent to the skeleton (McConnaughey,1994). However, the mechanism of Ca²⁺ transport across the outer, or oral, epithelium is controversial, being reported as active(Wright and Marshall, 1991; Clode and Marshall, 2002a) and passive (Benazet-Tambutte et al.,1996).

Intracellular Ca concentrations have been obtained from tissues of coral larvae by electron microprobe (x-ray microanalysis)(Clode and Marshall, 2004) and we have applied this method in the present investigation to mature coral polyps. As shown in settled coral larvae by Clode and Marshall(Clode and Marshall, 2004), we show that in mature polyps intracellular total Ca is very high and increases from the outer to inner cell layers. We have also used the ion microprobe at low and high spatial resolution (see Guerquin-Kern et al., 2005; Lechene et al., 2006; Clode et al., 2007) to follow the transport of Ca²⁺ across coral epithelia using the stable isotope ⁴⁴Ca as a tracer. With ⁴⁴Ca in the external seawater it was possible to follow the exchange of ⁴⁴Ca for the endogenous ⁴⁰Ca in cells and seawater-filled compartments. This allowed direct visualisation of Ca²⁺ influx across epithelia and into the external coelenteron, i.e. the seawater-filled compartment between the oral and aboral epithelia. Our observations are consistent with the suggestion that Ca²⁺ accumulates across the oral epithelium against a concentration gradient and that Ca²⁺ transport is transcellular and involves some sort of active process.

Colonies of Galaxea fascicularis L. were collected from the reef flat at Heron Reef, Great Barrier Reef, Australia and transported in buckets of seawater (SW) to Heron Island Research Station. Colonies were maintained in semi-shaded outdoor flow-through aquaria [photosynthetic photon flux density(PPFD) 500–1500 mmol photons s^–1 m^–1;23–25°C] and allowed to recover for 2 days. Polyps were easily separated using forceps and placed in trays of running SW (PPFD 50–150 mmol photons s^–1 m^–1; 23–25°C) to recover for a further 2 days. Small separated polyps were incubated for 2 h in jars containing 200 ml filtered SW (0.25 mm) that were partially submerged in shallow, flow-through aquaria in full sunlight (PPFD 800–1900 mmol photons s^–1 m^–1; 23–25°C). Polyps destined for secondary ion mass spectrometry (SIMS) analysis were incubated in vials containing 10 ml of artificial seawater (ASW)(Benazet-Tambutté et al.,1996) in which ⁴⁴CaHCO₃ replaced ⁴⁰CaHCO₃. Incubation was carried out under the same conditions of light and temperature for 1 min or 8 min. Polyps were then frozen at approximately midday in liquid propane (–190°C) that had been cooled by liquid nitrogen (LN₂), as previously described(Marshall and Wright, 1991). Polyps were gently blotted on seawater-soaked filter paper to remove excess adhering seawater prior to freezing. This was necessary to achieve reasonable freezing rates to minimize intracellular ice crystal damage. All samples were transferred to La Trobe University, Melbourne, in a CryoPak dry shipper(Taylor-Wharton Australia Pty Ltd, Albury, Australia) at –180°C and stored in LN₂ until required.

For quantitative x-ray microanalysis, frozen polyps were freeze-substituted in 10% acrolein in diethyl ether, essentially as described by Marshall(Marshall, 1980) and Marshall and Wright (Marshall and Wright,1991), infiltrated in increasing concentrations of ether and Araldite™ mixtures and embedded in Araldite™. Araldite™ was the preferred embedding medium as it contains negligible levels of elements detectable by energy dispersive spectrometry(Pålsgård et al.,1994). All solutions were anhydrous, with processing conducted in a dry box at a relative humidity of 10%. The embedded polyps were cut into slices approximately 0.5 mm thick with a diamond saw (Buehler Ltd, Lake Bluff,IL, USA). Embedded tissue was dissected from the skeleton and re-embedded under anhydrous conditions. Dry cut sections 1.5 μm thick were mounted on Formvar®-filmed copper grids, coated in 100 Å aluminium and analysed by energy dispersive spectrometry. Briefly, x-ray mapping was performed using a JEOL 1200EX analytical scanning transmission electron microscope (STEM)(JEOL Australasia Pty Ltd, Sydney, Australia) with a Link Atmospheric Thin Window energy dispersive detector (Oxford Instruments, High Wycombe, UK). The detector was interfaced to a 4pi Spectral Engine (4pi Analysis Inc., Durham,NC, USA) and a Quadra 700 Apple Macintosh computer. The microscope was fitted with custom-made LN₂-cooled anticontaminators and a LaB₆filament. Analyses were carried out by elemental imaging using the multispectral analysis program ImagNspect(Ingram et al., 1999), at 120 kV with a beam diameter of <90 nm and a beam current of 5×10^–10 A. Peak integral and quantitative images were obtained with a resolution of 128×128 pixels and a dwell time of 3 s pixel^–1. Quantitative numerical data, based upon the Hall peak/continuum model (Hall and Gupta,1979), were extracted directly from the elemental maps by selecting areas of interest. Individual spectra for each pixel in the selected regions were summed and processed to yield concentrations for every element(LeFurgey et al., 1992). Elemental concentrations are given in mmol kg^–1 embedded tissue.

For analysis by ion microprobe (SIMS) 1–2 μm-thick dry cut sections were flattened on thin aluminium discs and coated with a thin layer of gold. Analysis was carried out in either a Cameca ims5f SIMS (Cameca,Gernevilliers-Cedex, France) operated in the microprobe mode using a duoplasmatron source (oxygen primary ion beam) at 15 kV and 0.5 nA beam current or a Cameca NanoSIMS (Cameca N50). The imaging spatial resolution of the ims5f was <2 μm while the resolution of the N50 is approximately 200 nm. All NanoSIMS analyses were conducted using a 16 kV 16O^–primary beam with a probe current of approximately 5 pA (D1-5) to 23 pA(D1-3). Ion maps were acquired at a resolution of 512×512 pixels, with a typical dwell time of between 3 and 7 ms pixel^–1.

Isotopic images of masses 12, 23, 24, 39, 40, 44 and 88 were recorded to reveal cell and tissue distributions of ¹²C, ²³Na, ²⁴Mg, ³⁹K, ⁴⁰Ca, ⁴⁴Ca and ⁸⁸Sr, respectively. The purity of secondary ion signals was over 90%.

Elemental isotope images were processed using ImageJ(http://rsb.info.nih.gov/ij/)to obtain ⁴⁴Ca/⁴⁰Ca ratios, line scans and pixel intensities of selected regions. Statistical analysis of pixel intensities was carried out by nonparametric tests using the computer program JMP (SAS Institute Inc., Cary, NC, USA). Statistical analysis of calcium concentrations obtained by x-ray microanalysis was carried out by a one-way analysis of variance (ANOVA) with post hoc Tukey-Kramer HSD tests using JMP.

All x-ray and ion microprobe (SIMS) analyses were carried out on 1–2μm-thick sections of tissues that covered the external wall or theca of the polyp. These tissues can be seen in thick (1 mm) transverse slices of freeze-substituted polyps (Fig. 1A) and visualised in more detail by fluorescence microscopy of slices stained with Acridine Orange (Fig. 1B). The oral epithelia, consisting of oral ectoderm and oral gastrodermis, are well defined. Numerous orange-staining mucocytes are present in both cell layers and numerous yellow-staining symbiotic algae(zooxanthellae) are present in the gastrodermis. The aboral epithelia are much thinner and are hard to distinguish in dissected thin-sectioned preparations for x-ray microanalysis and ion microprobe analysis. Further structural details are apparent in confocal images of slices(Fig. 2A,B). The oral ectoderm comprises non-specialised ectodermal cells, mucocytes and cnidocytes. This is separated from the oral gastrodermis by a well-defined acellular mesogloea. The oral gastrodermis consists of non-specialised gastrodermis cells,mucocytes and host cells containing symbiotic algae (zooxanthellae). The aboral gastrodermis contains relatively few zooxanthellae and consists primarily of non-specialised gastrodermis cells and mucocytes. A thin mesogloea separates the aboral gastrodermis from the tenuous calicoblastic ectoderm that closely adheres to the skeleton. The aboral gastrodermis contains few zooxanthellae and consists primarily of mucocytes and non-specialised gastrodermis cells. The calicoblastic ectoderm is a thin epithelium containing some mucocytes but is primarily composed of skeletogenic cells containing numerous vesicles.

Measurements of total Ca concentration were obtained from quantitative x-ray images derived from 1–2 μm-thick sections of freeze-substituted coral tissues (Fig. 3A,B). Potassium, Na, Cl, P and S images were obtained simultaneously to assist in the interpretation of Ca images and the identification of cellular regions and seawater compartments. The elemental concentrations differ slightly in the two sets of images but are within the range measured in this study.

In Fig. 3A,B it can be seen that high concentrations of Na and Cl were present in the seawater-filled space of the extrathecal coelenteron, indicating the retention of diffusible ions during the freeze-substitution process. Calcium concentration was high in cnidocysts and was clearly higher in the mesogloea and extrathecal coelenteron than in the coral cells and zooxanthellae. Calcium was frequently also in high concentration in mucocytes, particularly in the oral gastrodermis(Fig. 4).

In the aboral epithelium, the mesogloea was too thin to permit accurate analytical information to be extracted from x-ray images. Calcium concentrations were obtained from the aboral gastrodermis and calicoblastic ectoderm (Fig. 5). On the skeletal side of the calicoblastic ectoderm, loci of high Ca concentration are believed to represent nucleating calcium carbonate deposits on the organic matrix (Fig. 5). Fig. 5 also confirms the presence of NaCl in the sub-skeletal space.

Regions of interest were applied to images such as those in Fig. 3 to extract Ca concentrations from non-specialised epithelial cells and seawater-filled compartments (Fig. 6A,B). The Ca concentration in the mucus-containing external seawater layer was 21±5 mmol l^–1 (mean ± s.e.m., n=3 where n is the number of measurements from one preparation). Because of the few data points, the calcium concentration of the latter compartment is not included in the statistical analysis in Fig. 6A.

The distribution of C in secondary ion images of ¹²C was homogeneous across the resin-embedded tissue and pure resin, indicating that local matrix effects on ion sputtering were minimal. The distribution of an elemental isotope in the SIMS images may thus be taken as an indication of relative concentration. The natural ⁴⁴Ca/⁴⁰Ca ratio(i.e. not enriched) is 0.02, thus any measure above this level is indicative of ⁴⁴Ca enrichment.

After 1 min exposure to ⁴⁴Ca, the tracer can be seen to have penetrated the oral epithelium and entered the extrathecal coelenteron(Fig. 7). However, the tracer did not appear to have entered the cells of the aboral epithelia. The ⁴⁴Ca/⁴⁰Ca ratio image and a line plot across this image(Fig. 7C,E) show that ⁴⁴Ca exceeds ⁴⁰Ca to a significant extent only in the external seawater layer after 1 min exposure.

After 8 min exposure to ⁴⁴Ca in the light, the tracer was present at higher levels in the mesogloea and extrathecal coelenteron than in the external seawater (Fig. 8B)and this is also reflected in the ⁴⁴Ca/⁴⁰Ca image and line scan (Fig. 8C,D). Very little ⁴⁴Ca appeared to have entered the aboral epithelia after 8 min and it was not present in the skeletal fragments adhering to the aboral epithelium.

After incubation for 8 min in the dark, a similar distribution of ⁴⁴Ca was apparent (Fig. 9) but the amount of ⁴⁴Ca relative to ⁴⁰Ca in the extrathecal coelenteron was lower in the dark than in the light [light,1.38±0.14 (mean ± s.e.m.); dark, 0.75±0.29; N=3, P<0.05, where N represents the number of polyps].

Samples exposed to artificial SW containing ⁴⁴Ca for 1 min (i.e. from the same specimen as in Fig. 7) were analysed at the higher resolution offered by NanoSIMS. It can be seen (Fig. 10) that ⁴⁴Ca had entered the oral ectodermal cells and replaced a significant fraction of the original intracellular ⁴⁰Ca(Table 1). However, very little exchange had occurred in the mucocytes or cnidocytes, which contained considerably higher concentrations of ⁴⁰Ca than the unspecialised ectodermal cells. The ⁴⁴Ca/⁴⁰Ca ratio was higher in the mesogloea and slightly lower in the oral gastrodermal cells(Table 1), compared with the unspecialised ectodermal cells, and was extremely low in the zooxanthellae in the oral gastrodermis.

Analysis by NanoSIMS after 8 min exposure to ⁴⁴Ca showed that the ⁴⁴Ca/⁴⁰Ca ratio in the unspecialised ectodermal cells was only slightly higher than in the 1 min samples, but the ratios in the mesogloea and gastrodermal cells were considerably higher(Fig. 11; Table 1). Again, little exchange had occurred in the cnidocytes, mucocytes or zooxanthellae(Fig. 11).

X-ray microanalysis showed that there is a gradient in total calcium,increasing from the outer oral epithelia to the inner aboral epithelia. The concentration rises from approximately 11 mmol kg^–1 in the oral ectodermal cells to approximately 19 mmol kg^–1 in the calicoblastic ectodermal cells. Furthermore, calcium concentration in the mesogloea of the oral epithelium approaches 18 mmol kg^–1. These values are similar to those observed in settled larvae of Pocillopora damicornis (Clode and Marshall, 2004). The concentrations are recorded as mmol kg^–1 embedded mass. This approximates to wet mass if the embedding resin replaces water. The total intracellular concentration of calcium in coral cells is high in comparison to the few measurements of total intracellular calcium concentration available for animal cells. In some terrestrial invertebrate tissues calcium concentrations have been recorded at less than 5 mmol kg^–1 wet mass (reviewed by Gupta, 1993), although a concentration of 9 mmol kg^–1 wet mass has been reported for the cytoplasm of nematoblast cells in an anemone(Lubbock et al., 1981). Our values are considerably higher than the estimate of 2.8 mmol l^–1 derived from ⁴⁵Ca compartment analysis in Stylophora pistillata(Tambutté et al.,1996).

The gradient in total intracellular calcium concentration, increasing from the oral ectoderm in contact with seawater to the calicoblastic ectoderm in contact with the skeleton, suggests that a pool of intracellular calcium is accumulated for deposition in the skeleton. The high concentration of calcium in the mesogloea of the oral epithelium suggests that this compartment has some role in the transport of Ca²⁺ across the epithelium. Unfortunately, the dimensions of the mesogloea of the aboral epithelium were too small to permit measurement of calcium concentration.

In an x-ray microanalystical study of frozen-hydrated Galaxeapolyps, Clode and Marshall found that the calcium concentration in the extrathecal coelenteron (22 mmol kg^–1 wet mass) was significantly higher than in the mucus-containing seawater (16 mmol kg^–1 wet mass) (Clode and Marshall, 2002a). The latter was, in turn, significantly higher than in the bulk seawater (12 mmol kg^–1). These data suggested that active transport of calcium may occur across the oral epithelium. Our present data are consistent with this view. In freeze-substituted sections of Galaxea polyps, the calcium concentration in the extrathecal coelenteron was 24 mmol kg^–1embedded mass. In the mucus-containing external seawater layer, the calcium concentration was 21 mmol kg^–1 embedded mass. The latter was based on only three acceptable measurements and may not be a reliable estimate. However, the trend is similar to that observed in the analyses of frozen-hydrated polyps. Measurement by ion-selective electrodes(Marshall and Clode, 2003)indicates that the concentration of unbound calcium in this layer is 14.5 mmol l^–1, i.e. slightly lower than the concentration measured in frozen-hydrated polyps. The pattern of calcium distribution across the epithelia, mesogloea and extrathecal coelenteron is very similar to that seen in settled larvae of Pocillopora damicornis(Clode and Marshall, 2004).

These data indicate that Ca²⁺ is being transported across the oral epithelia against a concentration gradient. If transport occurs in the absence of an electrical gradient or against an electrochemical gradient then the process is likely to be an active one requiring energy. There are no in vivo measurements of transepithelial potentials in Galaxea or in any other corals as far as we are aware. However,transepithelial potentials of isolated preparations of oral epithelia of Lobophyllia hemprichii and Plerogyra sinuosa in Ussing chambers were less than 1 mV (gastrodermal side negative to the ectodermal side) (Wright and Marshall,1991). These preparations exhibited a net flux of Ca²⁺from the ectodermal side to the gastrodermal side. It seems possible,therefore, that Ca²⁺ is transported against a concentration gradient in the absence of a significant electrical gradient in intact polyps.

It is possible that a favourable transepithelial potential for Ca²⁺ entry could be generated in intact polyps by the release into the extrathecal coelenteron of OH^– from the photosynthesising zooxanthellae found in the oral and aboral gastrodermis. Certainly, pH in the coelenteron proper does increase when polyps are illuminated (A.T.M. and P.L.C., unpublished data). However, this mechanism seems unlikely because a high calcium concentration is maintained in the extrathecal coelenteron under dark conditions when photosynthesis is not occurring(Clode and Marshall,2002a).

Scleractinian corals transport prodigious quantities of Ca from seawater for incorporation into the CaCO₃ skeleton. Wright and Marshall measured net Ca²⁺ flux through isolated coral epithelia in Ussing chamber experiments at 1.1 μmol cm^–2 h^–1(Wright and Marshall, 1991). Making some simple assumptions about polyp dimensions, it can be calculated from data on skeletal ⁴⁵Ca incorporation(Marshall, 1996; Marshall and Clode, 2004) that Ca²⁺ flux is 4.8–9.6 μmol cm^–2h^–1. These values compare reasonably well with estimates of 1.7 μmol cm^–2 h^–1 Ca²⁺ flux in Acropora (Wilbur and Simkiss,1979). Using the calculated flux data and estimates of the volume of the extrathecal coelenteron from measurements made on slices of freeze-substituted Galaxea polyps, it can be shown that the Ca²⁺ content of the extrathecal coelenteron would be removed in 6–12 min in the light and in 22–45 min in the dark if no further Ca²⁺ entry occurred. These calculations are consistent with the observations of ⁴⁴Ca tracer entry in light and dark conditions. After incubation for 8 min in the light, ⁴⁴Ca had almost replaced ⁴⁰Ca in the external coelenteron whereas this was not the case in the dark. Very little exchange in the extrathecal coelenteron had occurred after 1 min.

After 8 min incubation in ⁴⁴Ca in the light, ⁴⁰Ca had almost completely been replaced by ⁴⁴Ca in the mucus-containing layer of external seawater, but the concentration of ⁴⁴Ca in the extrathecal coelenteron was higher, as shown in the ⁴⁴Ca/⁴⁰Ca ratio image and line scan. In the absence of a favourable transepithelial potential this could only be a consequence of some sort of active Ca²⁺ transport across the oral epithelium.

Analysis of samples incubated for 1 min and 8 min in ⁴⁴Ca by NanoSIMS showed clearly that Ca²⁺ rapidly entered the oral ectodermal cells and exchanged with approximately 30% of the total cell calcium. After 8 min incubation, the fraction of calcium exchanged in the ectodermal cells had risen only slightly to 33%. Thus, in these cells there is both a rapidly exchanging pool and a large slowly exchanging pool of Ca. This is consistent with the observations of Marshall and Wright, who observed a slowly exchanging Ca pool in the tissues by ⁴⁵Ca autoradiography(Marshall and Wright, 1998). In the mesogloea and the gastrodermal cells, the amount of Ca exchanged after 8 min was approximately 60 and 53%, respectively. The transport of Ca²⁺ appears to be principally transcellular in both the oral ectoderm and gastrodermis.

The uptake of ⁴⁵Ca by the skeleton has been shown to be inhibited by Ruthenium Red (Krishnaveni et al., 1989; Marshall,1996). This has been interpreted as evidence of the presence of a Ca-ATPase in coral tissues. In situ hybridization evidence indicates that this ATPase is principally located in the calicoblastic cells but is also present in the aboral and oral gastrodermal cells but not in the oral ectoderm(Zoccola et al., 2004). Thus,active transport of Ca²⁺ is thought to occur at the skeletal face of the calicoblastic cells, as proposed by McConnaughey(McConnaughey, 1994). Evidence of an active transport mechanism in the oral epithelium has been derived from Ussing chamber experiments (Wright and Marshall, 1991), while evidence based on x-ray microanalysis has been described by Clode and Marshall (Clode and Marshall, 2002a). However, Ca-ATPase may not be the conduit for active transport of Ca²⁺ across the oral epithelium because light-activated uptake of Ca²⁺ at the surface of the oral ectoderm in zooxanthellate corals is not inhibited by the Ca-ATPase inhibitor Ruthenium Red (Marshall and Clode,2003).

It seems probable that the oral mesogloea is involved in the transport of Ca²⁺ across the oral epithelium because total Ca in this compartment is high and ⁴⁰Ca is rapidly exchanged for ⁴⁴Ca; the mechanism, however, is obscure. In anemones, the mesogloea appears to be composed of collagen fibrils within an amorphous matrix that is composed of neutral protein–polysaccharide complexes(Gosline, 1971a; Gosline, 1971b; Koehl, 1973; Young, 1973). In Galaxea, the mesogloea is bounded by membrane-like structures(Clode and Marshall, 2002b)that are distinct from the adjacent cell membranes. These may be formed from laminins, as occur at the subepithelial boundaries of Hydra mesogloea(Sarras and Deutzmann, 2001). Although some charge shielding by inorganic cations may occur in the mesogloea of the anemone Metridium senile(Gosline, 1971a) to reduce electrostatic interactions between collagen and the matrix complexes, the number of charged sites is considered to be small. Our analysis of mesogloea in Galaxea indicates an increased Ca concentration compared with seawater. In the absence of extensive polyanionic charges, it seems unlikely that the increased Ca concentration is due to electrostatic interactions. It is interesting to note that Macklin, using autoradiography, found that calcium accumulated in high concentration in the mesogloea of Hydra and suggested that this accumulation resulted from active transport across the ectoderm (Macklin, 1967).

The data show that an intracellular concentration gradient for total calcium exists across the outer and inner epithelia of Galaxeapolyps. The gradient increases from the oral ectoderm to the calicoblastic ectoderm. Based on data from the present investigation and previous studies(Clode and Marshall, 2002a; Marshall and Clode, 2003; Clode and Marshall, 2004),there is also an increasing calcium gradient from bulk seawater to the mucus-containing external seawater layer, mesogloea and extrathecal coelenteron. The data also indicate that Ca²⁺ transport across the oral epithelium is transcellular and that entry into the extrathecal coelenteron is against a concentration gradient, possibly by some active transport process. Furthermore, the mesogloea is involved in this process. The movement of Ca²⁺ across the oral epithelium is initiated by light and is proportional to light intensity(Marshall and Clode, 2003). Thus, it is not surprising that the extrathecal coelenteron ⁴⁴Ca/⁴⁰Ca ratio measured in polyps incubated in the dark is lower than that measured in the light. This appears to be further evidence against paracellular Ca²⁺ transport since it would be expected that the ratio would be similar in light and dark conditions if Ca²⁺transport occurred by passive paracellular diffusion.

A possible explanation for the occurrence of active transcellular Ca²⁺ transport into the extrathecal coelenteron may be that the epithelium must be tight to prevent the dissipation of a proton gradient generated by the deposition of calcium carbonate at the skeletal surface. It is hypothesised that protons generated during the formation of calcium carbonate are exchanged for Ca²⁺via a Ca-ATPase in the calicoblastic epithelial cells (McConnaughy and Whelan, 1997). The protons are transported into the fluid-filled coelenteron where they keep the pH of the coelenteron seawater low and the concentration of CO₂ high for the photosynthetic needs of the symbiotic algae(Cohen and McConnaughy, 2003). Alternatively, the protons may neutralise OH^– produced by the photosynthesis of intracellular symbiotic algae present primarily in the oral gastrodermis (Allemand et al.,2004). In Galaxea, calcification occurs principally on the outside of the thecal walls of the corallite(Marshall and Wright, 1998). Covering these walls are the inner aboral epithelia and the outer oral epithelia separated by the extrathecal coelenteron. The latter is divided into longitudinal compartments that have restricted continuity with the internal coelenteron (see Fig. 6C).

Within each compartment, fluid circulates by ciliary action, frequently in countercurrents in adjacent compartments (A.T.M. and P.L.C., unpublished data). Thus, these compartments are semi-isolated and receive protons from calcium carbonate deposition and possibly hydroxyl ions from algal photosynthesis. If these compartments are functionally isolated from the inner coelenteron then it is perhaps not surprising that some form of active Ca²⁺ transport should occur across the oral epithelium. As shown by the NanoSIMS analysis, the passage of ⁴⁴Ca across the epithelium appears to be transcellular.

This research was supported by AINSE grants to A.T.M. and UWA Research and NANO TAP grants to P.L.C. The Cameca NanoSIMS50 facility at The University of WA was funded by a Major National Research Facility (MNRF) grant through the Nanostructural Analysis Network Organisation (NANO). Corals were collected under Great Barrier Reef Marine Park Authority permits to A.T.M.