Characterization and role of carbonic anhydrase in the calcification process of the azooxanthellate coral Tubastrea aurea
- First Online:
- Cite this article as:
- Tambutté, S., Tambutté, E., Zoccola, D. et al. Mar Biol (2007) 151: 71. doi:10.1007/s00227-006-0452-8
- 348 Views
In zooxanthellate corals, the photosynthetic fixation of carbon dioxide and the precipitation of CaCO3 are intimately linked both spatially and temporally making it difficult to study carbon transport mechanisms involved in each pathway. When studying Tubastrea aurea, a coral devoid of zooxanthellae, we can focus on carbon transport mechanisms involved only in the calcification process. We performed this study to characterize T. aurea carbonic anhydrase and to determine its role in the calcification process. We have shown that inhibition of tissular carbonic anhydrase activity affects the calcification rate. We have measured the activity of this enzyme both in the tissues and in the organix matrix extracted from the skeleton. Our results indicate that organic matrix proteins, which are synthesized by the calcifying tissues, are not only structural proteins, but they also play a crucial catalytic role by eliminating the kinetic barrier to interconversion of inorganic carbon at the calcification site. By immunochemistry we have demonstrated the presence of a protein both in the tissues and in the organic matrix, which shares common features with prokaryotic carbonic anhydrases.
KeywordsCarbonic anhydraseCarbonCalcificationCoralBiomineralizationOrganic matrix
Bovine serum albumin
Dissolved inorganic carbon
Phosphate buffered saline
Protease inhibitor cocktail
Soluble organic matrix
Sodium dodecyl sulphate
Tris buffered saline
Dissolved inorganic carbon
Scleractinians (stony corals) are coelenterates that form aragonitic calcium carbonate (CaCO3) skeletons. They are classically functionally divided into two groups: the hermatypic (reef-building) and the ahermatypic (non-reef-building) corals. The vast majority of the hermatypic corals are found in shallow, tropical oceans and characteristically contain within their tissues large populations of symbiotic dinoflagellates called zooxanthellae. In these zooxanthellate scleractinians, the photosynthetic fixation of carbon dioxide (CO2) and precipitation of CaCO3 are intimately linked both at spatial (cell to ecosystem) and temporal (day–night) scales rendering it difficult to study the carbon transport mechanisms involved in each pathway. On the other hand, the vast majority of ahermatypic corals are devoid of zooxanthellae. Thus when studying these corals, it is possible to focus on carbon transport mechanisms involved only in calcification processes.
In corals, skeleton formation is a process of “extracellular biologically-controlled biomineralization” and as such involves a mineral fraction and an organic matrix. The means by which corals may influence CaCO3 precipitation include (1) control of the levels of inhibitors, promoters, and regulators of calcification by the means of a set of macromolecules (called the organic matrix) surrounding the crystal or included within the mineral and (2) availability of substrates. Since coral skeleton formation results from the delivery of calcium and inorganic carbon to the site of calcification, these two substrates are crucial to study. Recently, most of the research involving coral calcification has focused either upon the structure and composition of organic matrices of skeletons (Gautret et al. 1997, 2000; Cuif et al. 1999, 2003; Dauphin 2001) or on the uptake and mechanisms of deposition of calcium ions (Wright and Marshall 1991; Tambutté et al. 1995, 1996). However, for invertebrate mineralization, carbonate ions are as important as calcium ions. Pearse (1970) established that skeletal carbonate can originate from two different carbon sources: soluble carbonates from sea water or CO2 produced by animal metabolism. Furla et al. (2000) demonstrated that in the zooxanthellate coral Stylophora pistillata, the major source of DIC for coral calcification is metabolic CO2 and not inorganic carbon originating from seawater. Similar results were obtained in the non-zooxanthellate octocoral Leptogorgia virgulata and Corallium rubrum respectively by Lucas and Knapp (1997) and Allemand and Grillo (1992). In addition, these last authors have shown that DIC supply is rate-limiting for calcification. However, these conclusions can not be generalized to all corals since Adkins et al. (2003) demonstrated that there is little or no metabolic CO2 in the skeleton of the deep-sea non-zooxanthellate coral, Desmophyllum cristagalli.
Carbonic anhydrases are ubiquitous enzymes known to act as catalysts for the interconversion between CO2 and HCO3−. Since the limiting step in the conversion from CO2 to carbonate ion is the hydration step, CA can play an important role when calcification is carbon limited. In avians, CA facilitates eggshell formation (Nys and de Laage 1984) and in fishes, CA is supposed to play an important role in otolith formation (Payan et al. 1997; Tohse and Mugiya 2001; Tohse et al. 2004). In the case of invertebrates, this enzyme has been found to play a role in the calcification of calcareous sponges (Jones and Ledger 1986), scleractinian corals (Goreau 1959; Isa and Yamazato 1984; Marshall 1996, Furla et al. 2000; Al-Horani et al. 2003), octocorallians (Kingsley and Watabe 1987; Allemand and Grillo 1992; Lucas and Knapp 1996, 1997; Rahman et al. 2005, 2006), molluscs and echinoderms (Miyamoto et al. 1996; Mitsunaga et al. 1986). Carbonic anhydrase has been described in many tissues but its presence in extracellular calcified structures suggests that this enzyme could also play an important role during the precipitation step of the mineral. The aim of this study was to characterize T. aurea carbonic anhydrase and determine its role in the calcification process.
Materials and methods
Parent coral colonies of Tubastrea aurea (Cœlenterata:Anthozoa:Scleractinia), indigenous to the Indo-Pacific, and sea anemones Aiptasia pulchella were maintained at the Centre Scientifique de Monaco in the following conditions: semi-open circuit, Mediterranean sea water heated to 26 ± 0.2°C, and fed with Artemia nauplii twice a week. T. aurea was maintained in low light conditions at a constant irradiance of 15 μmol photons m−2 s−1 on a 12 h/12 h light/dark cycle, and A. pulchella was maintained in the same conditions but with a constant irradiance of 250 μmol photons m−2 s−1. For calcification rate experiments, coral colonies were cut with a bone cutter in order to obtain fragments of three to four polyps called nubbins. The sectioned skeleton was coated with epoxy resin so that only the tissues were in contact with seawater. Nubbins were used for experiments after a period of 4–5 weeks, when new tissues entirely covered the junction between the resin and the skeleton. Cleaning was performed daily.
Preparation of tissues for carbonic anhydrase activity assay
Six to seven polyps were cut from parent colonies, put on ice and homogenized with a mortar in about 3 ml cold veronal buffer prepared according to Weis et al. (1989): 25 mM veronal containing 5 mM ethylenediaminetetraacetate (EDTA), 5 mM dithiothreitol (DTT), 10 mM MgSO4 with pH adjusted to 8.2. Protease inhibitor cocktail (PIC, SIGMA) 0.1% was then added. The mixture was sonicated for 1 min and then centrifuged at 765g for 20 min at 4°C. The supernatant was centrifuged again in the same conditions and then aliquoted in Eppendorf tubes for storage at −80°C.
Preparation of organic matrix for carbonic anhydrase activity assay
Polyps were cleaned by removing soft tissues with 2N NaOH for 2 h at 70°C. The skeletons were rinsed with ultrapure water, dried at 60°C overnight and ground to a fine powder with a mortar. The powder was demineralized for one night at 4°C in 0.5 M EDTA containing 0.1% PIC and 5 mM phenanthrolin. After complete dissolution of aragonite, centrifugation (10 min, 10,000g, 4°C) allowed soluble and insoluble matrices to be separated. To desalt soluble components, the supernatant containing the soluble organic matrix (SOM) was filtered and concentrated using Centricon® (Amicon, cut-off 5 kDa) according to the manufacturer’s instructions. The retentate was aliquoted and stored at −80°C.
In vitro assay for carbonic anhydrase activity
The in vitro assay for CA activity is described in detail by Weis et al. (1989). Briefly CA activity in crude homogenates was measured by the decrease of pH resulting from the hydration of CO2 to HCO3− and H+ after the addition of substrate. All experiments were performed at 4°C. CO2−saturated distilled H2O served as a substrate and was prepared by passing gaseous CO2 through an airstone for 30 min (pH 3.5). To run the assay, 5 ml veronal buffer (pH 8.2) were transferred to a small beaker and 1 ml of homogenate diluted in veronal buffer was added to obtain quantities of proteins ranging from 0 to 12 mg for tissues and 0–16 μg for organic matrix. The mixture was constantly stirred with a magnetically driven stirring bar. Four ml of substrate was then added rapidly and the decrease in pH was recorded by a Ag/AgCl pH probe immersed in the mixture and connected to a Metler DL70 pH meter fitted with a chart recorder. As a control for the non-catalyzed reaction, the same experiment was performed without homogenate.
Carbonic anhydrase activity was calculated as (t0 − t)/t, where t0 is the time needed for the non-catalyzed reaction and t is the time for the catalyzed reaction to obtain a pH decrease from 8 to 7.5. Units of enzyme activity (EU) were normalized to the weight of soluble proteins.
Inhibition of carbonic anhydrase activity
To test the effect of the inhibitor ethoxyzolamide on CA activity, the assay was performed as described above, but 0–10 μM ethoxyzolamide was added to the veronal buffer before addition of tissue. Results are expressed as percent inhibition calculated from 100 − [(CA activity in presence of inhibitor/CA activity in absence of inhibitor) × 100)]. IC50 represents the concentration of inhibitor, which inhibits half of the enzyme activity measured in the absence of inhibitor.
Effect of inhibition of CA activity on calcification rate
Measurement of calcification rate was made according to the method of Tambutté et al. (1995) adapted for higher volumes. Measurements were made at equivalent times of day in order to avoid possible variation caused by endogenous circadian rhythms (Buddemeier and Kinzie 1976). Nubbins grown on epoxy resin were incubated for 2 h 15 min in 60 ml beakers containing approximatively 800 kBq of 45Ca (as 45CaCl2, NEZ013, Perkin Elmer) dissolved in filtered seawater. For inhibition experiments, 10 μM ethoxyzolamide was added in the incubation medium. Water motion was provided during each incubation by small stirring bars in order to reduce as much as possible diffusion limitation by boundary layers. Exposure to air was limited to less than 5 s during transfer to the incubation beakers and incubations were made under low light conditions.
At the end of the labelling period, each nubbin was immersed for 20 s in a beaker containing 1 l FSW. Labelled nubbins were then incubated in a beaker containing 150 ml FSW for 180 min to monitor 45Ca efflux into the rinse medium. Water motion was provided in the efflux medium by stirring bars. Upon completion of the efflux, nubbins were dissolved completely over a period of 20 min in approximately 5 ml of 1 N NaOH at 90°C. Each skeleton was then rinsed six times in 5 ml of distilled water, dried and dissolved in 10 ml of 6 N HCl overnight (“HCl-soluble fraction”).
Radioactive samples were added to 4 ml Ultima Gold (Packard) and emissions were measured using a liquid scintillation analyzer (Tricarb, 2100TR, Packard).
For the efflux, results are expressed as dpm mg−1 protein. For the calcification rate, results are expressed as nmol Ca2+ h−1 mg−1 protein in tissues or nmol Ca2+ h 1 g−1 skeleton and represent means ± S.D. for at least three measurements. Calculation of the half-time for calcium washout (T1/2) and its corresponding rate constant as well as calculation of the size of the coelenteric pool were made according to Tambutté et al. (1995). In the presence of ethoxyzolamide, results are expressed as percentage inhibition of calcification rate.
A control experiment was performed with killed nubbins (paraformaldehyde) in order to determine non specific binding of 45Ca.
Extraction of proteins for immunoblotting
Proteins from tissues: Six polyps of T. aurea were cut from a colony with a bone cutter and rinsed in FSW. They were then homogenized with a mortar maintained on ice, in about 8 ml of extraction buffer (50 mM Tris, 100 mM NaCl, 5 mM EDTA, 1% Triton X100, 0.1% PIC, and 5 mM phenanthrolin). The mixture was twice centrifuged at 765g for 10 min at 4°C to eliminate skeletal debris. The supernatant was maintained on ice for 20 min while vortexing every 5 min, the time necessary for the buffer to extract proteins. The supernatant was aliquoted in Eppendorf tubes for storage at −80°C before experiments. Proteins were also extracted from the sea anemone A. pulchella with the same protocol except that they were directly homogenized in the extraction buffer and sonicated.
Proteins from organic matrix: The protocol was the same as the preparation of organic matrix for the carbonic anhydrase activity assay (see paragraph above).
The activity of the antibody was examined by a dot-blot assay on proteins extracted either from whole tissues (40 μg of proteins) or organic matrix (12 μg of proteins). Experiments were performed at room temperature.
Briefly, the samples were deposited on nitrocellulose membranes which were saturated with 1% BSA for 1 h in TBS (140 mM NaCl, 5 mM Tris, pH 7.4) and labelled for 1 h with primary antibodies. The primary antibodies were either (1) anti phycoerythrin (AbCam), 1:20,000 dilution or (2) rabbit anti-human erythrocyte carbonic anhydrase II antibody (Rockland immunochemicals), 1:10,000 dilution or (3) rabbit anti N-terminal β-carbonic anhydrase from Synecchococcus sp (generous gift from Mak Saito and François Morel), 1:25,000 dilution, in TBS-BSA 1%). Membranes were then rinsed and incubated for 1 h with secondary antibodies (horseradish peroxidase-linked anti-rabbit IgG, Sigma, 1:2,000 dilution in TBS-BSA 1%). Immunoreactive dots were then revealed with ECL kit (GE Healthcare). Controls were made with the preimmune serum as the primary antibody.
Electrophoresis, protein transfer and Western blot
Proteins extracted either from whole tissues (100 μg of proteins) or just the organic matrix (20 μg of proteins) were homogenized in Laemmli sample buffer (Laemmli 1970). Samples were resolved in SDS–PAGE (12% acrylamide for resolving gel, 4% acrylamide for stacking gel) using a Mini Protean II apparatus (BIORAD). Proteins were then electrophoretically transfered from unstained gels onto PVDF membranes using a transfer apparatus (Mini Transblot Cell, BIORAD). After transfer, membranes were saturated with 5% skimmed milk in TBS containing 0.1% Tween and labelled for 1 h with primary antibodies either (1) anti phycoerythrin 1:20,000 dilution, or (2) anti-human erythrocyte carbonic anhydrase II antibody, 1:10,000 dilution, or (3) rabbit anti-β carbonic anhydrase from Synecchococcus sp., 1:10,000 dilution, in TBS containing 1% skimmed milk and 0.1% Tween® 20). Membranes were then rinsed and incubated for 1 h with secondary antibodies (horseradish peroxidase-linked anti-rabbit IgG, Sigma, 1:2,500 dilution in TBS containing 1% skimmed milk and 0.1% Tween® 20). Immunoreactive dots were then revealed with an ECL kit (GE Healthcare). Controls were made with the preimmune serum as the primary antibody.
Preparation of samples for immunolocalization
Demineralized samples: One polyp was fixed in 3% paraformaldehyde in S22 buffer (NaCl 450 mM, KCl 10 mM, MgCl2 58 mM, CaCl2 10 mM, Hepes 100 mM, pH 7.8) at 4°C overnight and then decalcified using 0.5 M EDTA in Ca-free S22 with 3% PAF at 4°C. It was then dehydrated in an ethanol series, cleared with xylene and embedded in Paraplast. Cross sections (7 μm thick) were cut and mounted on silane-coated glass slides.
Mineralized samples: One polyp including skeleton was fixed in 3% paraformaldehyde in S22 buffer (NaCl 450 mM, KCl 10 mM, MgCl2 58 mM, CaCl2 10 mM, Hepes 100 mM, pH 7.8) at 4°C overnight. It was then dehydrated in ethanol and embedded in LR White resin. Sections were cut with a low speed saw (Buehler, Isomet) in thick slices (about 1 mm), etched with EDTA 1% for 1 h to expose antigenic epitopes and rinsed in ultrapure water.
Immunolocalization of carbonic anhydrase
Deparaffinized sections of tissues or samples of skeleton prepared as described above were saturated with 5% BSA in 0.05 M PBS, pH 7.4, containing 0.2%, teleostean gelatin, 0.2% Triton X100. The samples were then incubated with primary antibodies from rabbit anti-β carbonic anhydrase 1:1,000 dilution in BSA-saturated PBS solution (PBS 0.05 M, pH 7.4, containing 0.2%, teleostean gelatin, 0.2% Triton X100, 5% BSA), 1 h at RT and overnight at 4°C in moist chamber. After rinsing in BSA-saturated PBS solution, they were incubated with biotinylated anti-rabbit antibodies (GE Healthcare 1:250 dilution, 1 h at RT) as secondary antibodies. They were finally incubated for 20 min with streptavidin-Alexa Fluor 568 (Molecular probes, 1:50 dilution) and DAPI (2 μg ml−1, SIGMA). Controls were routinely performed with the rabbit preimmune serum as the primary antibody. Samples were embedded in Pro-Long antifade medium (Molecular probes) and observed with a confocal laser scanning microscope (Leica, TCS4D) at 568 nm excitation, 600 nm emission.
Cross sections of demineralized samples or thick slices of mineralized samples (see paragraphs above for preparation) were stained with hemalun, eosin, and acetified anilin blue solutions.
Media and chemicals
Unless otherwise specified, all chemicals were obtained from Sigma or Biorad and were of analytical grade.
FSW was obtained by filtering seawater on 0.22 μm Millipore membranes.
The carbonic anhydrase inhibitor ethoxyzolamide was dissolved in DMSO to a concentration of 60 mM and buffered with 1 M Tris to pH 8.2.
Protein concentration was measured in a microplate using the BCA Protein Assay Kit (Uptima® UP40840A). BSA was used as a standard.
Statistical analysis of the data
The effect of the carbonic anhydrase inhibitor, ethoxyzolamide, on calcification rate was tested using a t test (software Jump 5.1, SAS Institute, Cary, USA). Results are considered statistically significant when P < 0.05.
The approach we have taken towards the long-range goal of understanding the mechanisms of biomineralization in corals has been to characterize Tubastrea aurea carbonic anhydrase and then to determine its role in the calcification process. We report results of the measurement of enzyme activity in tissues and organic matrix and the effect of inhibition of carbonic anhydrase on the calcification rate. We have revealed by Western blotting and immunohistochemistry the presence of a protein both in the tissues and in the organic matrix, which reacts with an antiserum against prokaryotic carbonic anhydrases.
Carbonic anhydrase activity
Effect of inhibition of carbonic anhydrase activity on the calcification rate
Before using the antibodies for immunolocalization, we tested their reactivity by immunochemistry. Dot blots were performed with the proteins/enzymes in their non denaturated form whereas Western blots were performed with the proteins/enzymes in their denaturated form (due to the presence of SDS and β-mercapto-ethanol during electrophoresis).
In the present work, we investigated carbonic anhydrase from the tissues and the skeletal fraction of T. aurea by (1) measuring the enzyme activity, (2) determining its involvement in biomineralization, (3) determining some of its biochemical properties.
Measurement of carbonic anhydrase activity
While it appears that cellular carbonic anhydrase plays a key role in the availability of carbon involved in different physiological processes, there are few studies dealing with the presence and the role of extracellular CA in the calcium carbonate deposition process. In calcified skeletal structures of invertebrates, where calcium carbonate is the major component, a set of molecules grouped under the term of “organic matrix” are always present (Weiner 1984; Wheeler and Sikes 1984; Constantz and Weiner 1988; Falini et al. 1996). The roles of these molecules in the calcification process are numerous: initiation/inhibition of crystal growth, crystal morphology, and calcium binding (for review see Wheeler et Sikes 1984; Watanabe et al. 2003). Miyamoto et al. (1996) were the first to discover a carbonic anhydrase domain within nacrein, a soluble organic matrix protein of the nacreous layer in the mollusc Pinctada fucata. Since then, a cDNA that encodes a shell matrix protein composed of carbonic anhydrase-like domains has been cloned in the oyster Pinctada maxima by Kono et al. (2000). Watanabe et al. (2003) have also found an internal sequence in T. aurea that exhibits similarity to a part of the carbonic anhydrase sequences. CA activity was also recently found in various biominerals (Borelli et al. 2003; Rahman 2005, 2006) suggesting a widespread distribution of CA in calcium carbonate biominerals. In the present work we measured carbonic anhydrase activity in the organic matrix of T. aurea. The value of this activity is higher for organic matrix than for tissues, which can be due either to (1) a truly higher activity in the organic matrix, (2) a different CA between tissues and the organic matrix, (3) an artefact with standardization methods related to protein assays. Indeed, invertebrate organic matrix proteins have special biochemical features which render their characterization difficult, for example when performing staining after electrophoresis (Gotliv et al. 2003). A similar reason could explain the problems encountered when using protein assays on the organic matrix that probably underestimate the protein content (unpublished results). Nevertheless the important point to consider is that a carbonic anhydrase is present in the organic matrix and that this enzyme possesses an activity. Carbonic anhydrase activity in the organic matrix is inhibited by ethoxylamide and the inhibition constant IC50 in this matrix is 200 times higher than for tissues. It is important to note the high resistance of the organic matrix-linked CA, because its activity is preserved after the long demineralization and purification steps necessary to obtain organic matrix itself.
Determination of CA involvement in biomineralization
We looked at the role of CA in biomineralization by using a pharmacological approach. We first measured the calcification rate in control conditions and then tested the effect of CA inhibitors. We determined that the value of calcium deposition in the skeleton is similar (i.e. 0.43 ± 0.09 μmol Ca2+ h−1 g−1 skeleton) to the value obtained by Marshall (1996) on the same species (i.e. 0.48 ± 0.03 μmol Ca2+ h−1 g−1 skeleton). The volume of the coelenteric cavity (i.e. 12.71 ± 3.6 μl mg−1 protein) is comparable to the value obtained for the scleractinian coral S. pistillata (i.e. 7.3 ± 1.2 μl mg−1 protein, Tambutté et al. 1995). Our results on the inhibition of calcium deposition into the skeleton in the presence of ethoxyzolamide indicate that CA is involved in the calcification process. This type of inhibition of the calcification rate was also observed in hermatypic corals (Goreau 1959; Tambutté et al. 1996; Furla et al. 2000), sea urchin spines (Heatfield 1970), barnacle shells (Yule et al. 1982), molluscans (Wilbur and Jodrey 1955), crustaceans (Roer 1980) and the red coral Corallium rubrum (Allemand et al. 1992). Using the same kind of approach with 45Ca, Kingsley and Watabe (1987) obtained the opposite results in the gorgonian Leptogorgia virgulata with an increase of Ca uptake in the spicules and the axes in the presence of carbonic anhydrase inhibitors. Nevertheless, while the mechanism of CA action seems to differ depending on the species, in all cases, the results show that CA is involved in the calcification process
- 1.If seawater HCO3− is the source of DIC, CA may catalyze its conversion to CO2 to buffer the acidity produced by the conversion of HCO3− into CO32−, as already suggested by Sikes et al. (1980) for coccolithophorids, following the equations:
- 2.If intracellular CO2 is the form of DIC used at the site of calcification, then the extracellular organic matrix-linked CA may help in converting this CO2, which diffuses from the tissue to the skeletogenic fluid due to the probable high pH of this calcifying region (Furla et al. 2000; Al-Horani et al. 2003), into HCO3−, following the equations:
Determination of some CA properties
Since we have shown that CA is present in both the tissues and the organic matrix and we demonstrated that this enzyme plays a role in the calcification process, we tried to constrain some of its biochemical properties and its localization both in the tissues and the organic matrix. Five groups of cellular CA are described in the literature (Hewett-Emmett and Tashian 1996; Cox et al. 2000; Lane et al. 2000, 2005) (1) α-CA, mostly found in eukaryotes, (2) β-CA, characteristic of eubacteria and plants (3) γ-CA, characteristic of archaebacteria (4) δ-CA and ɛ-CA described in the marine diatom Thalassiosira weisflogii. The enzymes of the type alpha, beta, and gamma use zinc as cofactor whereas delta-CA can switch between zinc, cadmium and cobalt, and epsilon-CA uses cadmium. Weis and Reynolds (1999) have shown, by Western blotting with an antibody raised against human CA, the presence of an α-carbonic anhydrase in the tissues of a sea anemone. We found the same result in the present study using the sea anemone A. pulchella. Watanabe et al (2003) found in T. aurea an internal sequence of a matrix protein that exhibits sequence similarity with α-CA sequences. However, in the present study, in T. aurea, we could not detect CA using an immunological approach with an anti α-CA antibody. Thus, if an α-CA is present in T. aurea, it possesses epitopes in the catalytic site that are not recognized by the antibody against human α-CA. This result could be due to a difference in structure between mammal α-CA and coral α-CA. On the other hand, Western blotting with an antibody raised against a carbonic anhydrase from Synecchococcus sp. shows that there is a protein that immunoreacts with this antibody raised against a prokaryotic CA in both the tissues and in the organic matrix. Since no labelling was obtained with the antibody against phycoerythrin, a cyanobacterial marker, we can suggest that the labelling observed with the antibody against the prokaryotic CA is specific and not due to the presence of cyanobacteria in coral tissues. The presence of a protein, which shares properties with enzymes found in prokaryotes appears surprising. However, prokaryotic-like proteins have already been described in Cnidarians. For example, Richier et al. (2003) found an extra-mitochondrial, monomeric Mn-superoxide dismutase and a Fe-superoxide dismutase, both enzymes characteristic of prokaryotes, within the tissues of a sea anemone. More recently, by studying 26,845 ESTs from a coral and a sea anemone, Technau et al. (2005) found that about 1.3–2.7% of cnidarian proteins only matched with non-metazoan sequences (i.e. fungi, prokaryotes, plants, and protists), many matching only with bacterial sequences. Among these bacterial sequences, they identified the bacterial universal stress protein (UspA). To explain this result, these authors suggested either a conservation of ancient genes within the genomes of basal metazoans or lateral gene transfer. It is noteworthy that such observations are not only limited to cnidarians since in tunicates it has also recently been suggested that enzymes involved in cellulose biosynthesis are likely acquired by horizontal transfer from bacteria (Sasakura et al. 2005).
The protein characterized in our Western-blots by its immunoreactivity with the prokaryotic CA antibody has an apparent molecular mass in the tissues of 35 kDa, which is smaller than the one found in the organic matrix (37 kDa). This difference can be accounted for by oligosaccharide chains (Waheed et al. 1992; Wilson et al. 2000) since, in corals, matrices are highly glycosylated (Dauphin 2001). When chemical fluorochrome staining or immunolocalization was performed with this antibody on the skeleton, labelling displayed a pattern typical of the organic matrix (Gautret et al. 2000; Puverel et al. 2005). The existence of a protein positively labelled by the same antibody both in the tissues and in the organic matrix suggests that high homologies exist between these two proteins, the latter could be the secreted form of the former. In this case, however, since we have determined that carbonic anhydrases react differently to inhibitors, it is probable that these two enzymes are different isoforms. Another possibility is that the pharmacologic differences between these two isoforms only result from a modification, before secretion of the protein, by glycosylation. Further characterization is needed to solve this point. Furthermore, the differences in molecular weight observed in Western-blot confirms that, as in zooxanthellate corals (Puverel et al. 2005), organic matrix proteins of azooxanthellate corals do not result from the trapping of the whole soft tissues as suggested by Constantz (1986). By immunohistochemistry, we showed that only the tissues facing the skeleton (i.e. the calcifying tissues) are labeled with the antibody raised against prokaryotic CA, suggesting that this protein is involved in the biomineralization process.
Our work was performed on the azooxanthellate coral T.aurea, allowing us to study the transport of carbon used for calcification while eliminating the complications due to photosynthesis. We demonstrated the presence of an active carbonic anhydrase both in the tissues and in the organic matrix with a direct role of this enzyme in the calcification process. We suggest that this activity may be due a protein, which shares common features with prokaryotic CA. This protein shows similar features in the tissues and in the organic matrix suggesting that the calcifying tissues could be responsible for the secretion of this protein. This result does not exclude the possibility that other types of CA are responsible for the activity observed in the tissue and the organic matrix. Our results also demonstrate that in corals, organic matrix proteins are not only structural proteins but also catalytic proteins and provide a crucial enzyme to eliminate the kinetic barrier in the conversion of inorganic carbon. This new understanding of the chemistry in the calcifying region is essential to account for the mechanisms underlying the carbon and oxygen isotope fractionations seen in skeletal carbonates (Adkins et al. 2003).
We thank Prof. François Morel from Princeton University and Mak Saïto from the Woods Hole Oceanographic Institution for providing the antibody, anti-β-carbonic anhydrase from Synecchococcus sp. This study was conducted as part of the Centre Scientifique de Monaco 2000–2004 research program. It was supported by the Government of the Principality of Monaco and by the California Institute of Technology, USA.