Extracellular carbonic anhydrase activity promotes a carbon concentration mechanism in metazoan calcifying cells

Significance

Biological calcifying systems are capable of generating a versatile construction material from metabolic CO₂. Here, the transformation of CO₂ to HCO₃⁻ is of critical importance, with very limited information about the underlying mechanisms. Carbonic anhydrases (CAs) are evolutionary ancient enzymes that catalyze the reversible hydration of CO₂ and have often been associated with calcifying tissues. Using the calcifying cells of the sea urchin larva as a model system, the present work uncovered a cellular carbon concentration mechanism that depends on the presence of an extracellular CA (Cara7). Molecular, cell physiological, and histological evidence support this conclusion and highlight the elegance of biomineralizing systems in using carbon resources in the most sustainable manner.

Abstract

Many calcifying organisms utilize metabolic CO₂ to generate CaCO₃ minerals to harden their shells and skeletons. Carbonic anhydrases are evolutionary ancient enzymes that have been proposed to play a key role in the calcification process, with the underlying mechanisms being little understood. Here, we used the calcifying primary mesenchyme cells (PMCs) of sea urchin larva to study the role of cytosolic (iCAs) and extracellular carbonic anhydrases (eCAs) in the cellular carbon concentration mechanism (CCM). Molecular analyses identified iCAs and eCAs in PMCs and highlight the prominent expression of a glycosylphosphatidylinositol-anchored membrane-bound CA (Cara7). Intracellular pH recordings in combination with CO₂ pulse experiments demonstrated iCA activity in PMCs. iCA activity measurements, together with pharmacological approaches, revealed an opposing contribution of iCAs and eCAs on the CCM. H⁺-selective electrodes were used to demonstrate eCA-catalyzed CO₂ hydration rates at the cell surface. Knockdown of Cara7 reduced extracellular CO₂ hydration rates accompanied by impaired formation of specific skeletal segments. Finally, reduced pH_i regulatory capacities during inhibition and knockdown of Cara7 underscore a role of this eCA in cellular HCO₃⁻ uptake. This work reveals the function of CAs in the cellular CCM of a marine calcifying animal. Extracellular hydration of metabolic CO₂ by Cara7 coupled to HCO₃⁻ uptake mechanisms mitigates the loss of carbon and reduces the cellular proton load during the mineralization process. The findings of this work provide insights into the cellular mechanisms of an ancient biological process that is capable of utilizing CO₂ to generate a versatile construction material.

Biomineralization is an evolutionary ancient process that is used by organisms from the five kingdoms to harden their tissues for protection and mechanical support (1). In the marine environment, the majority of organisms generate CaCO₃ minerals, including calcite and aragonite, to construct their skeletons and shells. The process of calcification has profound impacts on the geology of our planet and is an integral part of the global carbon cycle by generating large amounts of CaCO₃ bound in coral reefs, chalk mountains, and deep-sea sediments (2). To generate CaCO₃, marine organisms must acquire Ca²⁺ and dissolved inorganic carbon (DIC) (i.e., CO₂, HCO₃⁻, CO₃²⁻) that are concentrated at the site of mineral formation to promote precipitation (3–9). While calcium is usually absorbed from the environment, DIC can derive from the seawater as HCO₃⁻ or from the hydration of metabolic CO₂. Marine calcifiers—including corals (8), sea urchins (10), and foraminifera (11)—were demonstrated to utilize substantial amounts of metabolic CO₂ to produce their shells and skeletons. In the calcifying coral Stylophora pistillata ¹⁴C radioisotopic labeling methods demonstrated that 70 to 75% of the DIC, used for calcification derives from metabolic CO₂ (8, 11). Similarly, 63% of DIC incorporated into the skeleton of the sea urchin larva originates from respiratory CO₂, while the remaining 37% are absorbed from the seawater (10). These studies indicate that metabolic CO₂ is the primary carbon source for the calcification process in several key marine calcifiers.

Carbonic anhydrases (CAs) are metallo-enzymes that catalyze the reversible hydration of CO₂ to form H₂CO₃ (carbonic acid) that then dissociates into HCO₃⁻ (bicarbonate) or CO₃²⁻ (carbonate) (12–14). The direction of this chemical equilibrium is dictated by the pH, with alkaline conditions favoring the formation of HCO₃⁻ and CO₃²⁻, while acidic conditions shift the equilibrium toward CO₂. CAs occur ubiquitously in all organisms to mediate pH and CO₂ homeostasis on the cellular and organismic level (12). CAs are evolutionary ancient enzymes that comprise five different families—including α-CA, β-CA, γ-CA, δ-CA, ζ-CA—and are found in the five kingdoms of life. α-CAs are the most common CA family found in animals, plants, and prokaryotes (12, 15) and can occur as cytosolic form, membrane-bound, and secreted isozymes (16, 17). Membrane-bound CAs are characterized by the presence of a glycosylphosphatidylinositol (GPI)-anchor that attaches the enzyme to the plasma membrane (12). These extracellular and secreted forms were observed to be specifically associated with mineralizing tissues in corals (7, 18), sponges (16), mollusks (19), and sea urchins (20). Furthermore, pharmacological experiments using compounds (i.e., acetazolamide) that inhibit CAs demonstrated reductions in calcification rates in sponges (21), mollusks (19, 22), corals (23, 24), coccolithophores (25), and sea urchins (6, 26–28). While these studies point toward a central role of CAs in the calcification process, their exact mechanism of promoting cellular carbon concentration remains largely hypothetical.

In the sea urchin larva, the primary mesenchyme cells (PMCs) are responsible for the generation of the larval endoskeleton. PMCs form syncytial cables that enclose the calcitic spicules (29–31). The initial phase of CaCO₃ formation happens intracellularly by the precipitation of amorphous calcium carbonate (ACC) within vesicular compartments (32). The current model denotes that Ca²⁺ is endocytosed from the sea water-like fluids of the primary body cavity (33) into vesicles, in a process that resembles macropinocytosis (34). Following a substantial concentration of calcification substrates (i.e., Ca²⁺ and CO₃²⁻), probably by the removal of water and salts from the calcification vesicle (35), ACC is formed and exocytosed to build the calcitic endoskeleton (34, 36). While the route of Ca²⁺ uptake has been well described, the mechanisms of DIC transport remain less well understood. A previous study demonstrated that the PMC-specific Na⁺/HCO₃⁻ cotransporter, Sp-Slc4a10, is critically involved in the calcification process by absorbing HCO₃⁻ from the primary body cavity (5). Knockdown of Sp-Slc4a10 resulted in a substantial reduction in skeleton growth, associated with decreased capacities of HCO₃⁻ transport into the cytosol (5). While this transporter presents a route for the uptake of DIC from the environment, mechanistic knowledge for the conversion of metabolic CO₂ by CAs to promote cellular carbon concentration is currently lacking.

Using the calcifying PMCs of the sea urchin larva as a model, this work aimed at revealing the physiological role of CAs in the cellular carbon concentration mechanism (CCM). We used a suite of molecular tools—including in situ hybridization, expression analyses, single-cell transcriptomics, gene knockdown, and immunocytochemistry—to identify and localize calcification-relevant CAs in the sea urchin larva. We developed microfluorimetric and ion-selective microelectrode protocols for the in vivo measurement of intra- and extracellular CA (iCA and eCA, respectively) activity. These in vivo techniques, in combination with pharmacological inhibition of CAs and knockdown of Cara7 activity, revealed the differential contribution of iCAs and eCAs in the calcification process. The findings of this work led to a deep mechanistic understanding and a new working model for the role of CAs in the CCM of a marine calcifier.

Results and Discussion

Identification of iCAs and eCAs in Calcifying Cells.

Gene mining for CAs in the genomic data bases of the purple sea urchin Strongylocentrotus purpuratus confirmed the presence of three major CA isoforms expressed in the calcifying PMCs (37, 38), including Cara7 (LOC579101), Cara2 (LOC577690), and Cara14-like (LOC591885) (SI Appendix, Table S1). Our in situ hybridization result and single-cell transcriptomic analyses demonstrated that Cara14-like and Cara2 are also expressed in other tissues and cell types of ecto-, meso-, and endodermal origin (Fig. 1A and SI Appendix, Fig. S1 and Table S1). In accordance with earlier studies (37, 38), we confirm that Cara7 is exclusively and highly expressed in PMCs of actively calcifying regions, including the initial cluster of skeleton formation in the gastrula stage as well as PMCs of the scheitel and postoral rods in the pluteus larva (Fig. 1B and SI Appendix, Figs. S1 and S2). Cara7 is a downstream effector of key regulatory genes in the skeletogenic cell lineage, including VEGF (39). Within the first 72 h of development, Cara2 has its highest expression levels shortly before the initial formation of the skeleton (Fig. 1C), while Cara7 has its expression peak during the period of skeleton formation (Fig. 1D). This indicates that Cara7 is likely associated with the initiation of skeletogenesis.

Fig. 1.

Phylogenetic analysis and expression patterns of iCAs and eCAs in the sea urchin larva. Whole-mount in situ hybridization using antisense and sense probes against Cara2 (A) and Cara 7 (B) at 2, 3, and 4 dpf. Schematic illustration of the Cara2 and Cara7 expression pattern in the PMCs of the gastrula and pluteus larva (Bottom). Abbreviations: LV, lateral view; OV, oral view; VV, ventral view. Expression of iCA Cara2 (C) and eCA Cara7 (D) along the first 72 hpf [data obtained from Echinobase.org (64)]. (E) Phylogenetic tree of extracellular (blue lines) and intracellular (orange lines) α-CAs in metazoans. A complete list of species, genes, and accession numbers are provided in SI Appendix, Table S2. Red boxes indicate the two CAs (Cara2 and Cara7) from S. purpuratus investigated in this work and blue boxes highlight other experimentally confirmed eCAs and iCAs. The bootstrap values were labeled at branch nodes and the branch lengths information (scale bar) that is proportional to the amino acid divergence is shown.

In silico sequence analyses suggest that Cara2 is an iCA in PMCs, while Cara7 displays features of an eCA, including an N-terminal signal sequence and a C-terminal GPI anchor (SI Appendix, Fig. S3 and Table S1), and thereby corroborates earlier studies that identified this PMC-specific CA to be an extracellular isozyme (28, 38, 40). Cara7 also possesses the typical α-CA signature sequence, including a glycine-rich region and several histidines that represent the active center (H217) and zinc ion binding sites (i.e., H244, H246, and H269) (20). Our phylogenetic analysis of CAs highlights the deep separation of intra- and extracellular α-CAs in metazoans (Fig. 1E). The sea urchin Cara7 belongs to the clade of secreted or membrane-bound eCAs, while Cara2 clusters with iCAs from other animals, ranging from cnidarians over mollusks to arthropods and vertebrates.

iCAs and eCAs Have an Opposing Contribution to the Calcification Process.

In a next step, we used a pharmacological approach to address the contribution of iCA and eCA activity to the calcification process. Normally developing larvae were exposed to different concentrations of the CA inhibitor acetazolamide (AZM) and AZM coupled to a 6-kDa dextran molecule (Dex-AZM) during the time of skeleton formation (16 h postfertilization [hpf] to 64 hpf). Since AZM is a small uncharged molecule that can pass biological membranes within several seconds (41, 42), this compound targets both, iCAs and eCAs. To disentangle the role of iCAs vs. eCAs in the calcification process, we used Dex-AZM that only targets eCAs. The results of these experiments demonstrated no effect of AZM on skeleton formation up to a concentration of 100 µM (Fig. 2 A and B and SI Appendix, Fig. S4). However, using Dex-AZM we found a substantial reduction in spicule length from 247.9 ± 0.8 µm in DMSO treated larvae down to 101.5 ± 25.6 µm at a concentration of 100 µM. The reduction in skeleton length in the Dex-AZM–treated larvae is largely attributed to a shorter postoral rod that was often absent in the 100 µM Dex-AZM–treated larvae (Fig. 2 A and B and SI Appendix, Fig. S4C).

Fig. 2.

eCA activity is required for the calcification of the larval skeleton. (A) Representative images of larvae raised from 16 hpf to 64 hpf under the treatment of 0.1% DMSO, AZM, and Dex-AZM, respectively. Bracket indicates the postoral rod (PO). (B) Biometric analyses of the postoral rod and other skeletal segments (SI Appendix, Fig. S4) after treatment with different concentrations of AZM and Dex-AZM. Asterisks indicate significant differences compared to controls, with **P < 0.001 (n = 3, one-way ANOVA + post hoc test [Holm–Sidak]). Values are presented as mean ± SEM. (C) Determination of recalcification rates by measuring the growth rate of the dissolved skeleton under pharmacological inhibition of CA activity by four concentrations (0, 1, 10, 100 µM; red triangle) of AZM. For the controls (0 µM) only the vehicle DMSO was added. Values are presented as mean ± SEM; n = 6, *P < 0.05 (one-way ANOVA + post hoc test). (Right) Schematic model illustrating the site of AZM inhibition. (D) Recalcification rates during pharmacological inhibition of eCAs by Dex-AZM at four concentrations (0, 1, 10, 100 µM; blue triangle). Values are presented as mean ± SEM; n = 3, *P < 0.05; **P < 0.001 (one-way ANOVA + post hoc test [Holm–Sidak]). (Right) Schematic model illustrating the site of Dex-AZM inhibition. (E) Expression levels of Cara2 (iCA) and Cara7 (eCA) under recalcifying conditions along the period of 3 d. Expression levels were normalized to the internal control EF1a. mean ± SEM; n = 3 to 4. (F) Whole-mount in situ hybridization of Cara2 (Left) and Cara7 (Right) at 2 and 3 d of recalcification. Schematic illustration of Cara2 (Left) and Cara7 (Right) expression in PMCs under recalcifying conditions.

To assure no collateral effects of the inhibitors on developmental processes in the gastrula and blastula stage, we used a previously described recalcification assay where pluteus larvae recalcify their skeletons within 2 to 3 d (3). Recalcification rates were slightly reduced in the presence of the CA inhibitor AZM at a concentration of 1 µM (Fig. 2C), but these inhibitory effects were attenuated at higher doses (10 and 100 µM). However, in accordance with our inhibitor experiments performed on normally developing larvae, Dex-AZM substantially reduced recalcification rates in a dose-dependent manner from 1 to 100 µM (Fig. 2D), indicating that mainly eCA activity is required for the mineralization process. We explain the differential inhibitory effect of the membrane-permeable AZM and membrane-impermeable Dex-AZM by the following hypothesis. Low concentrations of cell-permeable AZM primarily inhibit eCAs, with little amounts of the inhibitor actually reaching the cytosol. However, higher concentrations of AZM also effectively inhibit iCAs, and thereby counteract the inhibition of eCAs. These pharmacological experiments suggest an opposing role of iCAs and eCAs in the calcifying PMCs of the sea urchin larva.

The importance of eCA activity in the calcification process is further underscored by a stimulation of Cara7 expression during recalcification of the larval skeleton (Fig. 2E). Increased expression levels of Cara7 during skeleton remineralization are due to a recruitment of nonmineralizing PMC along the recalcifying spicules (Fig. 2F). Such an activation of PMCs from the resting to calcification state has been previously demonstrated by increased mRNA and protein levels of other calcification-relevant ion transporters and channels in PMCs during skeleton recalcification (3, 43). In contrast, the major iCA in PMCs, Cara2, did not show a pronounced expression in remineralizing PMCs as observed for Cara7, underpinning a differential contribution of iCAs vs. eCAs to the calcification process.

The hypothesis for an opposing contribution of iCAs vs. eCAs in the calcification process is further corroborated by our iCA activity measurements. Intracellular pH recordings in combination with the CO₂-pulse method were used to measure the intracellular, CA catalyzed, hydration, and dehydration speed of CO₂ (Fig. 3). The intracellular pH (pH_i) recovery rate after washout of the out-of-equilibrium (OOE) solution (2.5% CO₂, pH 8.0) was decreased by AZM in a dose-dependent manner (IC₅₀ 7.5 µM), reflecting the iCA activity (Fig. 3 A and B). Similar IC₅₀ values using AZM were obtained for iCA activity measurements in muscle fibers (IC₅₀ 10 µM) (42) and corneal endothelial cells (IC₅₀ 7 µM) (44). Against our initial hypothesis that iCA activity is stimulated in recalcifying larvae to support the increased calcification rates, we found the opposite result. The recovery rate after the CO₂ pulse was substantially decreased in recalcifying PMCs compared to those of control larvae (Fig. 3 C and D). In addition, the share of the AZM-sensitive fraction of the CO₂ dehydration reaction was dramatically reduced, reflecting a reduction in iCA activity (Fig. 3D). Despite this seemingly counterintuitive observation, such a reduction in iCA activity during high mineralization rates may serve an important biological function. The extensive intracellular hydration of metabolic CO₂ to promote CaCO₃ formation generates large amounts of protons (i.e., 2 moles of H⁺ per mole of CaCO₃ precipitated) that may cause a substantial proton load to the calcifying cells. Thus, it can be suggested that reductions in cytosolic CA activity may help to decrease the cellular proton load during the mineralization process. Another benefit of reducing iCA activity lies in the fact that intracellular calcifiers draw on a cytosolic HCO₃⁻ pool (3, 45). Here the intracellular pH of ∼7.0, together with cytosolic CA, activity may shift the equilibrium toward the formation of CO₂, and thereby cause a loss of the otherwise membrane-impermeable HCO₃⁻ ion.

Fig. 3.

Intracellular pH recordings in combination with the CO₂-pulse method demonstrated iCA activity in PMCs that was decreased under recalcifying conditions. (A) pH_i was measured using the ratiometric pH-sensitive dye BCECF-AM. pH_i of PMCs was recorded during exposure to OOE solution (2.5% CO₂, pH 8.0) in the presence of 100 µM AZM (red line, n = 8, mean ± SEM) or DMSO (black line, n = 10, mean ± SEM). The rate in pH_i change during addition and removal of the OOE solution reflects the hydration and dehydration speed of CO₂ within the cell. The enlarged area depicts the area of interest that was used to analyze iCA activity during CO₂ removal. (B) The recovery rate from the 2.5% CO₂ pulse is inhibited by AZM in a dose-dependent manner with an IC₅₀ value of 7.5 µM reflecting the iCA catalyzed fraction of the dehydration reaction. Boxplots include single measurements (circles), mean values (crosses), 95th percentiles, and SD bars. The membrane-impermeable Dex-AZM had no effect on iCA activity (SI Appendix, Fig. S5). (C) iCA activity measurements performed with actively recalcifying larvae. Changes in pH_i during exposure to 2.5% CO₂ OOE solution were measured in the presence of DMSO (black line, n = 12, mean ± SEM) or 100 µM AZM (red line, n = 12, mean ± SEM). The enlarged area depicts the area of interest that was used to analyze iCA activity during CO₂ removal under recalcifying conditions. (D) Comparison of recovery rates (pH_i units per minute) of untreated (circles) and recalcifying larvae (diamonds) and the inhibitory effects of 100 µM AZM (red bars) compared to DMSO controls (black bars). Values are presented as mean ± SEM and statistical analyses were performed using Student’s t test with *P < 0.05; **P < 0.001; ***P < 0.0001.

Similar concepts for the opposing function of iCAs and eCAs to promote CCM were developed for photosynthetic marine algae (46–48). Depending on the species, diatoms can predominantly utilize carbon that derives from CO₂ (“CO₂-user”) or from HCO₃⁻ (“HCO₃⁻-user”). In HCO₃⁻-users the CO₂ leaking out of the cell is converted by eCA to HCO₃⁻ and subsequently taken up by HCO₃⁻ transporters. In contrast CO₂-users prevent extracellular conversion of CO₂ to HCO₃⁻ and disequilibrium in the boundary layer persists. Thus, a high pCO₂ in the boundary layer also increases the intracellular CO₂ partial pressure and thereby supports photosynthesis (47). Interestingly, these different modes of carbon concentration in marine algae are associated with the opposing presence of iCAs and eCAs. While CO₂-users are characterized by high iCA levels, HCO₃⁻-users usually have low iCA activities. In some HCO₃⁻-users like Synechococcus, iCA activity is even absent for minimizing CO₂ leakage (48). Further evidence for considering the calcifying PMCs of the sea urchin larva a HCO₃⁻-user type of CCM is seen in an up-regulation of the Na⁺/HCO₃⁻ cotransporter (Sp-Slc4a10) and Cara7 under low HCO₃⁻ conditions, while the iCA Cara2 is substantially down-regulated under reductions in environmental HCO₃⁻ (SI Appendix, Fig. S7). Based on the findings of the present work and the existing concepts for the CCM in photosynthetic organisms, we propose that in the calcifying cells of the sea urchin larva as well, iCAs and eCAs act in an opposing manner, with iCA activity being reduced during active mineralization to promote the cellular CCM.

The eCA Cara7 Promotes a CCM in Calcifying Cells.

In a next step, we used a multipronged approach to study the localization of the Cara7 protein and the effect of Cara7 genetic perturbation on skeletal development and function of the cellular CCM. Therefore, we developed a Cara7-specific antibody that demonstrated detection of an ∼50-kDa protein (Fig. 4A) that is in the predicted size range (49 kDa) for this protein. Preabsorption of the antibody with the immunization peptide largely reduced the signal demonstrating the specificity of the antibody to its Cara7-specific immunization peptide. To perturb the expression of Cara7 we used two different morpholinos (MO) targeting the 5′-untranslated region and another one designed close to the start codon of the Cara7 gene. Western blot analyses validated the Cara7 knockdown by demonstrating a 43% reduction in Cara7 protein abundance in Cara7 morphants compared to scramble MO-injected larvae (Fig. 4B).

Fig. 4.

Phenotype of Cara7 knockdown morphants and immunocytological localization of Cara7. (A) Western blot analysis of the Cara7 antibody using crude extracts of whole pluteus (3 dpf) larvae including a peptide compensation assay by preabsorption of the primary antibody with the immunization peptide. (B) Western blot analysis of Cara7 protein abundance in 3-dpf larvae injected with scramble or Cara7 MO at a concentration of 200 µM. The Cara7 protein abundance was normalized to total protein concentrations. Values are presented as mean ± SEM (n = 3 to 4). Student’s t tests *P < 0.05. (C) Representative phenotypes of Cara7 MO (300 µM) and scramble MO injected larvae at 4 dpf. The length of the postoral rod (indicated by arrowheads) is predominantly affected in Cara7 morphants. Relative postoral rod length as a function of different MO concentrations. MO injections were repeated three to five times and individual measurements (n = 12 to 44, gray dots) are presented including mean (red lines) ± SEM. (D) Positive immunoreactivity of the polyclonal Cara7-specific antibody in pluteus larvae raised for 4 d under control conditions. Oral view (OV) and anal view (AV) of pluteus larvae stained with the Cara7 antibody (green) and DAPI counter stain (blue). Negative control showed no autofluorescence (SI Appendix, Fig. S9). (E) High-magnification image of the positive immunoreactivity in PMCs of the postoral rod tips.

Knockdown of Cara7 led to a decreased development of the postoral rods in the pluteus larva (Fig. 4C). This relatively mild effect was confirmed by a second MO targeting the 5′-untranslated region of the Cara7 gene (SI Appendix, Fig. S6) and suggests that Cara7 is primarily involved in the formation of the postoral and anterolateral rods. This finding is corroborated by immunohistochemical analysis demonstrating that positive immunoreactivity of the Cara7 antibody is only found in the actively calcifying regions of the postoral and anterolateral rods (Fig. 4 D and E). Positive immunoreactivity was detected in membranes of PMC cell bodies, within vesicular structures, along the syncytial cables, and probably also the spicule itself. The latter is supported by proteomic analyses that demonstrated high concentration of Cara7 occluded in the larval skeleton (49). This observation suggests that besides the outer membrane facing the extracellular matrix, Cara7 must be additionally localized in the inner plasma membrane facing the spicule cavity. Cara7 potentially has another physiological function in the inner membrane or could be a result of endocytosis and vesicular trafficking from the outer to the inner membrane following the route of Ca²⁺ uptake described in PMCs (34). Here it remains an important future task to understand if Cara7 has other physiological functions in the inner membrane facing the spicule and within vesicles. The immunolocalization of Cara7 together with our single-cell transcriptomic analyses (SI Appendix, Fig. S8) demonstrated a colocalization of Cara7 with the Na⁺/HCO₃⁻ cotransporter (Sp_Slc4a10) (5) and the proton channel otop2l (43) in the same population of PMCs. Together, our results indicate that the Cara7 protein is localized in PMCs and spicule compartment at the edges of the postoral and anterolateral rods and its activity is essential for the elongation of these rods.

To address the function of Cara7, we specifically developed an in vivo method using pH microelectrodes to measure eCA activity at the surface of PMCs. The concept for this method is based on an earlier study that used CO₂ OOE solutions in combination with a stop-flow protocol to measure CA activity during lysis of red blood cells (50). We successfully modified this protocol to measure the CO₂ hydration kinetics during the relaxation of the OOE solution at the surface of PMCs (Fig. 5 A and B). Using this method, we determined the CO₂ hydration rate at the surface of PMCs that was reduced in the presence of AZM (Fig. 5 C–E) and in Cara7 knockdown larvae (Fig. 5 F–H). In addition, no differences in the hydration rate constant (KΔ[H⁺]) were measured between Cara7 morphants exposed to 100 µM AZM or 100 µM DMSO (SI Appendix, Fig. S10). These results demonstrate that Cara7 is responsible for eCA activity at the outer cell surface of PMCs associated with reductions in the formation of specific skeleton segments where this enzyme is highly expressed.

Fig. 5.

pH-selective microelectrode measurements demonstrated that Cara7 is responsible for extracellular CA activity at the surface of PMCs. (A) Brightfield image of PMCs attached to the larval skeleton at 2 dpf with the microelectrode positioned at the surface of one PMC. (B) Illustration of the principle used in the stop-flow method for the measurement of eCA activity using H⁺-selective microelectrodes. Upon stopping the flow of the OOE solution (2.5% CO₂/pH 7.8), surface pH decreased due to the relaxation of the solution toward the formation of HCO₃⁻ and H⁺. The speed of CO₂ hydration depends on the catalytic activity of CAs and was used to determine eCA activity. (C) Comparison of OOE relaxation kinetics in the bulk solution (background) and at the cell surface in the presence of 0.1% DMSO or 100 µM of AZM. (D) Presentation of the average (n = 18) hydration kinetics at the cell surface in the presence of DMSO or AZM after subtraction of the background CO₂ hydration curve. Here the increase of H⁺ at any time point is depicted, compared between control conditions (DMSO, black line) and CA inhibition (AZM, red line). (E) Dependence of the rate constant of the pH change on the presence of eCA activity. We obtained KΔ[H⁺] values from nonlinear least-squares curve fits like those presented in C and F, demonstrating increased AZM-sensitive CO₂ hydration at the cell surface of PMCs (n = 18). (F) Comparison of OOE relaxation kinetics in the bulk solution (background) and at the cell surface of Cara7 morphants and scramble MO-injected larvae. (G) Presentation of the average (n = 9 to 12) hydration kinetics at the cell surface in scramble MO-injected or Cara7 knockdown larvae after subtraction of the background CO₂ hydration curve. (H) Comparison of the rate constant of the pH change on the presence of eCA activity in scramble and Cara7 MO-injected larvae. Letters denote significant differences between treatments (one-way ANOVA + post hoc test [Holm–Sidak]).

Based on these observations, we propose that Cara7-mediated eCA activity in PMCs may serve as cellular CCM by generating HCO₃⁻ from metabolic CO₂ at the cell surface that is reabsorbed by HCO₃⁻ transport mechanisms. In previous studies, the Na⁺/HCO₃⁻ cotransporter (Sp-Slc4a10) has been demonstrated to control the transport of HCO₃⁻ into PMCs (3, 5). This essential role of HCO₃⁻ transport in the calcification process is underscored by our low HCO₃⁻ experiments that dramatically reduced skeleton growth by 75% in the absence of HCO₃⁻ (SI Appendix, Fig. S11). Such reductions in environmental HCO₃⁻ not only reduce a part of the carbon source used for calcification, but also impair the function of HCO₃⁻ transporters (51).

To test the concept for the involvement of Cara7 in cellular carbon concentration, we performed pH_i recordings in combination with the ammonia prepulse method to determine pH_i regulatory capacities of PMCs. Previous studies demonstrated that primary and secondary active pH_i regulatory capacities in PMCs largely reflect HCO₃⁻ uptake (5, 52), and to a lesser extent H⁺ export (3). Active pH_i regulatory capacities of PMC were decreased by ∼ 50% during pharmacological inhibition of eCAs (Fig. 6 A and B) or by Cara7 knockdown (Fig. 6 C and D). In these studies, pH_i regulatory capacities decreased by 65 to 70% during pharmacological inhibition of HCO₃⁻ transporters or by knockdown of the Na⁺/HCO₃⁻ cotransporter (Sp_Slc4a10) (5, 33). The present work further demonstrated that in the absence of external HCO₃⁻, addition of Dex-AZM further decreased pH_i regulatory capacities (Fig. 6 E and F), indicating that metabolic CO₂ must be an additional source for cellular HCO₃⁻ uptake.

Fig. 6.

Dependence of pH_i regulatory capacities on eCA activity and proposed CCM for PMCs. (A) pH_i regulatory capacities of PMCs in control larvae investigated by the ammonia prepulse method. Average traces of pH_i recordings with 0.1% DMSO (black line, n = 8) or 100 µM Dex-AZM (red line, n = 6) in the perfusion solution during the washout period. (B) pH_i regulatory capacities of control (DMSO, black) and Dex-AZM (red)–treated PMCs (mean ± SEM; n = 6 to 8; Student’s t test *P < 0.05). (C) Average traces of pH_i recordings in PMCs of Cara7 (red line) or scramble MO- (black line) injected larvae at 3 to 4 dpf. (D) Comparison of pH_i regulatory capacities in larvae injected with scramble MO (black) or with Cara7 MO (red) at a concentration of 200 µM. (mean ± SEM; n = 10 to 11; Student’s t test *P < 0.05). (E) Average traces of pH_i recordings with 0.1% DMSO (black line, n = 8) or 100 µM Dex-AZM (red line, n = 8) in the absence of HCO₃⁻ (0-Bic) in the ASW solution during the washout period. (F) pH_i regulatory capacities of control (DMSO, black) and Dex-AZM (red)–treated PMCs (mean ± SEM; n = 8; Student’s t test *P < 0.05) in the absence of HCO₃⁻ during the washout period. (G) Schematic model summarizing the CCM in PMCs of the sea urchin larva. PMCs form syncytial cables surrounding the calcitic spicule (Created with https://BioRender.com). Cross-section of an enlarged PMC cell including validated transporters (Na⁺/HCO₃⁻ cotransporter [NBC] and Otopetrin proton channel [Otop2l]) and CAs investigated in the present work. While a certain fraction of the metabolic CO₂ is hydrated by iCAs, the remaining CO₂ diffuses across the plasma membrane where alkaline (∼pH 8.0) extracellular conditions favor the formation of HCO₃⁻ catalyzed by Cara7. This HCO₃⁻ is reimported by the Na⁺/HCO₃⁻ cotransporter Sp-Slc4a10 and feeds back into the cellular carbon pool. DIC is imported into the calcification vesicle by so far unknown mechanisms. Intracellular hydration of CO₂ depends on iCA activity that is reduced during active mineralization to promote the CCM and to reduce the cellular proton load. The remaining protons generated by the intracellular formation of CaCO₃ can exit the cell through the proton channel Otop2l (43).

These results provide strong evidence that Cara7 contributes to the cellular CCM by generating HCO₃⁻ in the cellular boundary layer from metabolic CO₂, which is then reabsorbed by HCO₃⁻ transporters (summarized in Fig. 6G). Similar mechanisms for the concentration of carbon have been described in photosynthetic systems of microalgae (47, 53, 54). These photosynthetic systems mitigate the loss of carbon from the cell by recycling CO₂ in the boundary layer using an eCA and HCO₃⁻ transport systems (47). This interaction of CO₂ hydration by eCA and HCO₃⁻ transporters would be most efficient when these proteins are localized in close vicinity acting as a metabolon. Such an HCO₃⁻ transport metabolon has been described in red blood cells and cancer cells where HCO₃⁻ transporters physically bind to iCAs and eCAs to promote gas transport and intracellular acid-base homeostasis, respectively (55, 56). In the sea urchin larva, this extracellular conversion of CO₂ has several advantages with respect to carbon concentration and intracellular acid-base balance. On the one hand, this mechanism allows to recapture metabolic CO₂ that diffused out of the cell and feeds this otherwise “lost” carbon back into the cellular carbon pool. On the other hand, extracellular hydration of CO₂ and subsequent uptake of the generated HCO₃⁻ ion leaves 1 mole of protons per mole of hydrated CO₂ in the extracellular space and thereby substantially reduces the net intracellular proton load during calcification. A further benefit of this CCM lies in the fact that the extracellular pH (pH_e 8.1) is more alkaline than the intracellular conditions (pH_i 7.0), and thereby substantially favors the formation of HCO₃⁻ in concert with eCA activity. In this way, eCA activity in calcifying cells plays an important role in optimizing the CCM by mitigating the loss of carbon and by reducing the intracellular acid load during calcification.

Conclusions

Calcifying organisms evolved sophisticated mechanisms to generate Ca²⁺ minerals for the construction of protective and stabilizing shells and skeletons. Many of these calcifying systems were demonstrated to use metabolic CO₂ as a major carbon source (8, 10, 11), with CAs playing a central role in this process. The present work reveals the role of eCAs underlying a cellular CCM and acid-base balance in calcifying cells of a marine animal. Given the deep evolutionary origin of eCAs and their strong association with calcifying systems (7, 16, 18–20), the results of the present work have strong implications for understanding the role of CAs in other calcifying organisms.

Moreover, this work also contributes to a more differentiated view on how changes in oceanic carbonate chemistry affect calcifying systems in past and future oceans. On the one hand, reductions in seawater pH also reduce the pH of the primary body cavity (33), thereby making extracellular hydration of CO₂ by Cara7 less effective. Gene-expression analyses from the present work (SI Appendix, Fig. S7) and earlier studies (57) demonstrated a stimulation of the PMC-specific Cara7 and HCO₃⁻ transporter (Sp-Slc4a10) under hypercapnic conditions. These observations suggest a compensatory mechanism in place that, however, is associated with increased energetic demands that were reported for sea urchin larvae raised under elevated pCO₂ conditions (52, 58). On the other hand, calcifying systems that largely utilize CO₂ as a carbon source may benefit from increases in environmental pCO₂ similar to the situation in photosynthetic organisms. However, the reduction in carbonate saturation states along with increases in seawater [H⁺] may prevail the positive effects of ocean acidification on the calcification process. In fact, changes in proton gradients across the plasma membrane of calcifying cells were demonstrated to be the dominating factor leading to reductions in calcification rates of several marine animals and microalgae (43, 59, 60). In this way, a better mechanistic understanding underlying calcification in marine organisms may help to formulate unifying principles that explain and predict sensitivities of marine calcifiers to past and future oceanic conditions (61).

Finally, the present study unravels the mechanisms by which CAs promote cellular carbon concentration for the generation of a versatile construction material from CO₂. At the rising anthropogenic CO₂ emissions to fuel our life on this planet, this work highlights the elegance of biological systems in utilizing carbon resources in the most sustainable manner.

Materials and Methods

Larval Cultures.

Adult purple sea urchins (S. purpuratus) were collected from the coast of California (La Jolla, CA), and shipped to the Helmholtz Centre for Ocean Research Kiel (GEOMAR). The animals were maintained in a recirculating natural seawater system at 12 °C, a salinity of 32 and regular water changes were made three times a week. Animals were fed with Laminaria sp. three times a week. Spawning of males and females was induced by shaking and larval cultures were maintained at a concentration of 25,000 larvae/L at 15 °C with continuous aeration, as previously described (62).

Whole-Mount In Situ Hybridization.

Whole-mount in situ hybridization was essentially performed according to the protocol described by Walton et al. (63) with some minor modifications. Briefly, larvae were fixed with 4% PFA/filtered seawater (FSW) and dehydrated with MeOH and stored at −20 °C. Larvae were rehydrated in an ascending series of MeOH and diethyl pyrocarbonate (DEPC)-PBST, and fully washed with DEPC-PBST in the final step. Prehybridization was performed for at least 1 h at 60 °C. Primers for generating and cloning of the probes are listed in SI Appendix, Table S4. Digoxigenin-labeled probes (500 to 1,000 ng/mL) were prewarmed at 70 °C and hybridization was performed at 55 °C overnight. Samples were washed with hybridization buffer, wash buffer (50% formamide, 5× SSC, 0.1% Tween-20), and 2× SSC at 55 °C followed by another washing step with 2× SSC at 37 °C and by MABT (500 mM maleic acid, 750 mM NaCl, NaOH, pH 7.5, 0.1% Tween-20) at room temperature. Samples were washed with MABT for 2 h and transferred to a blocking solution (1× MAB, 2% Block reagent [Roche], 10% sheep serum [Jackson Immunoresearch], 0.1% Tween-20) for 2 h. Samples were incubated overnight with 1:2,000 sheep antidigoxigenin-AP Fab fragments (Roche) antibody at 4 °C followed by intensive washing with MABT at the next day. The color reaction was performed with BM Purple AP (Roche) according to the manufacturer’s instructions for about 3 to 28 h.

Recalcification Assay and Determination of Calcification Rates under Pharmacological Inhibition of CAs.

The recalcification assay was performed as previously described (3) and allows a specific examination of pharmacological effects on the calcification process independent from collateral effects on developmental processes. Briefly, larvae (3 days post-fertilzation, dpf) were exposed to 0.03 M MES-buffered FSW adjusted to pH 6.0 for 12 to 15 h to fully dissolve their calcitic spicules. To recalcify their skeletons, larvae were transferred back into FSW (pH 8.0). This transfer day is defined as day 0 (0D), while the following days are defined as 1D, 2D, and 3D throughout the text. This recalcification assay was used to test the effects of iCA and eCA inhibition on the calcification process. AZM (Sigma-Aldrich) and 6 kDa Dex-AZM (Ramidus) were used, which are specific inhibitors for CAs (18). While AZM has been demonstrated to be membrane-permeable (41, 42), Dex-AZM cannot pass biological membranes (45) and thereby only targets extracellular CAs. After dissolution of the skeleton by the low pH treatment, larvae were transferred to six-well plates filled with 10 mL FSW containing only 0.1% DMSO, and 1 µM, 10 µM, and 100 µM of AZM or DexAZM, respectively. Each well contained 200 larvae and three experimental replicates were performed for each inhibitor concentration with 24 wells in total (two inhibitors, four concentrations, and three replicates). From the start of the experiment (0D), 20 to 50 larvae were sampled every day along the period of 3 d (1D, 2D, 3D). Larvae were fixed with 4% PFA/FSW (pH 8.2) and the growth rate of the skeleton was determined by measuring the daily increase in body rod + postoral rod length. Pictures were taken by a Zeiss Observer A1 inverted microscope and images were analyzed using the Zeiss Zen 3.3 (blue edition) software. A second set of experiments exposed normally developing larvae from 16 hpf up to 64 hpf to AZM and Dex-AZM using the same concentrations as mentioned above. Three independent replicates (n = 3) were performed and from each replicate 20 to 70 larvae were used for biometric analyses. Biometric analyses of the skeleton were made by measuring the length of the total spicule (body rod + postoral rod) as well as only the length of the postoral rod.

Phylogenetic Analysis and Molecular Cloning.

The amino acid sequences for phylogenetic analysis were collected from EchinoBase (https://www.echinobase.org/entry/) (64), the Ensembl (https://useast.ensembl.org/index.html), Reefgenomics (http://reefgenomics.org) databases, and aligned and compared with the sequenced sequences from the sea urchin larvae via the online tool MAFFT (Multiple Alignment using Fast Fourier Transform) (https://www.ebi.ac.uk/Tools/msa/mafft/). The sequences were scanned for motives and functional groups with https://myhits.sib.swiss/cgi-bin/motif_scan, https://mendel.imp.ac.at/gpi/gpi_server.html, and http://gpcr.biocomp.unibo.it/predgpi/pred.htm For the phylogenetic tree Cara7 (XP_003726289.2) and Cara2 (XP_782997.3) sequences were blasted to the National Center for Biotechnology Information (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and Reefgenomics databases, and homologous sequences with E-values higher than 9e⁻⁴¹ (9e⁻⁷⁵ for iCA and 9e⁻⁴¹ for eCA) were excluded for the following analysis. Selected sequences were aligned via the MUSCLE program and further trimmed by the Gblocks Server (http://molevol.cmima.csic.es/castresana/Gblocks_server.html), and the 140-aa trimmed sequences were used for the phylogenetic analysis. The maximum-likelihood tree was generated by MegaX (65) with the best-fit model LG+G+I and 500 bootstrap replicates.

The primers for cloning and in situ probe synthesis (SI Appendix, Table S4) were designed with Primerblast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) and reblasted with (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The transcript sequences of the sea urchin larvae genes SPU_012518 (LOC579101) and SPU_008685 (LOC577690) were amplified via PCR. The amplified sequences were cloned into pGEM-T easy cloning vector (Promega) and cultured in New England Biolab 5α competent Escherichia coli cells. Plasmid extraction sequencing (GENEWIZ) was followed by the synthesis of the RNA probes for in situ hybridization using M₁₃-F and M₁₃-R primers and the DIG RNA labeling Mix (Roche).

qRT-PCR and Single-Cell Transcriptomic Analyses.

qRT-PCR was performed as previously described (5). Briefly, RNA from control and recalcifying larvae at 0D, 1D, 2D, and 3D from three experimental replicates was isolated by using the Direct-zol RNA MicroPrep kit (Zymo Research). RNA samples were reverse-transcribed by Super Script IV cDNA synthesis kit (Invitrogen) for quantitative RT-PCR. To measure the expression levels of the target genes, the 7500 Fast Real-Times PCR System (Applied Biosystems) was used and expression levels of target genes were normalized to the housekeeping gene EF1a that has been demonstrated to be stable along ontogeny and during skeleton recalcification (3, 43). qPCR primers used in this study are listed in SI Appendix, Table S4. Single-cell transcriptomic analyses were carried out as previously described (43).

MO Injection.

Microinjection was performed as previously described (5). Two gene-specific MO-substituted antisense oligonucleotides (5′-AAAATATATGCATTCATGTTGATCA-3′) complementary to the start codon region of the sea urchin Cara7 gene and a second MO (5′-CGACCTTCAGATATATTCTCACAAA-3′) complementary to the 5′-untranslated region were obtained from Gene Tools. In addition, a scramble MO 5′-CCTCTTACCTCAGTTACAATTTATA-3′ that has no biological target in the sea urchin was used to obtain the control groups. The MOs were dissolved in injection Buffer (20 mM Hepes, 120 mM KCl, 24% glycerol, pH 8.0) and were injected into the freshly fertilized egg (one-cell stage) using a microinjection system (Picospritzer III, Parker) mounted on an inverse microscope (Zeiss Observer D1) equipped with a cooling stage.

In Vivo Determination of iCA Activity and pH_i Regulatory Capacities in PMCs.

Measurements of pH_i of PMCs using the pH-sensitive dye BCECF-AM was essentially performed as previously described (5, 33, 43). Briefly, after removal of the ectoderm by the bag isolation method and incubation with 50 µM BCECF-AM for 30 to 45 min, larvae were immobilized on 1% protamine sulfate-coated coverslips that were attached to the bottom of a perfusion chamber. The fluorescence was monitored on an inverted microscope (Zeiss, Observer A1) equipped with an appropriate light source and the imaging software (Visitron). From each larva (3 to 4 dpf), six to eight PMCs in the rod tips were recorded and treated as one replicate (n = 1). A minimum of six larvae were measured for every inhibitor concentration in an alternating mode. To translate ratios to actual pH_i values we performed a Nigericin calibration, as previously described (33). Along the entire duration of the experiment, larvae were exposed to FSW pH 8.0 either containing 0.1% DMSO as control or the respective AZM concentration (1 µM, 10 µM, 100 µM, and 1 mM). The CO₂ pulse using CO₂ OOE solutions (for details, see ref. 3) was performed for 5 min before switching back to FSW containing the respective AZM concentration. All solutions were delivered to the perfusion chamber at a flow rate of 2 mL/min and at 15 °C by a Perfusor with 50 mL Perfusor syringes (B. Braun Melsungen AG). The recovery rate after removal of the OOE solution in the presence of AZM was used to characterize the catalytic activity of iCAs.

The ammonium prepulse method was used to quantify pH_i regulatory capacities of PMCs and performed as previously described (5). Briefly, during pH_i recordings using BCECF-AM PMCs were exposed to FSW containing 20 mM NH₃/NH₄⁺ followed by a washout using FSW containing either DMSO or 100 µM Dex-AZM (for inhibition of the eCA). These solutions were applied to the larvae at a flowrate of 6 to 8 mL/min through a water-jacket cooling system set to 15 °C. The recovery rate from the ammonia prepulse-induced acidosis was used to characterize the effects of eCA inhibition and Cara7 knockdown on pH_i regulatory capacities.

Cell Surface CA Activity Determinations Using Proton Selective Microelectrodes.

H⁺-selective microelectrodes were used for the in vivo measurement of eCA activity in PMCs. The construction of H⁺-selective microelectrodes has been previously described (5, 33, 62). Briefly, borosilicate glass capillaries were pulled on a DMZ-Universal puller (Zeiss Instruments) into micropipettes with tip diameters of 1 to 2 μm. These were vapor-silanized with dimethyl chlorosilane (Sigma-Aldrich) and front-loaded with a 200-μm column of liquid ion exchanger mixture (H⁺ ionophore III; Sigma-Aldrich). The micropipette was backfilled with an electrolyte solution (300 mM KCl, 50 mM NaPO₄, pH 7.0) to create an ion selective microelectrode. The microelectrode was connected to a FD223a dual channel differential electrometer (WPI) and the potential was measured against an Ag/AgCl electrode connected by a 3 M KCl agar bridge to the perfusion solution. Calibration was performed by placing the electrodes into different pH solutions and plotting the measured voltage against the respective pH. All electrodes used had a Nernstian slope > 52 mV for 1 pH unit.

To measure eCA activity, we performed a stopped-flow assay using OOE solutions (2.5% CO₂/pH 8.0) following a modified protocol for the in vitro measurement of CA activity (50). This OOE state is generated in the moment of mixing of a high CO₂ (5% CO₂ pH 6.2) and a Tris-buffered low CO₂ solution (0.04% CO₂ pH 8.0) (66). When the flow (mixing) is stopped, the solution will acidify due to the relaxation toward the formation of HCO₃⁻ and H⁺. This rate of change in [H⁺] representing the CO₂ hydration speed is detected by our pH-sensitive microelectrodes. After bag isolation (see above for methodology), 2-dpf larvae were immobilized on 1% protamine sulfate-coated cover slides and glued to a perfusion chamber using silicon grease. The chamber was mounted on an inverted microscope (Zeiss VisitronSystem Observer D1) and perfused at a rate of 5 mL min⁻¹ using an OOE solution that was generated by mixing artificial seawater (ASW; for ASW composition, see ref. 43) containing 2 mM Tris adjusted to pH 8.0 and ASW equilibrated with 5.0% CO₂. The solutions were maintained at 15 °C by a water-jacket cooling system. The pH microelectrode was placed at the surface of a PMC using a motorized micromanipulator (Patch Star, Scientifica) under visual control at a 40 × 10 magnification. When stable recordings of PMC surface pH were obtained, the flow was stopped, leading to a relaxation of the OOE solution within the perfusion chamber toward the formation of HCO₃⁻ and H⁺. The reaction followed a nonlinear saturation curve and the speed of this reaction depends on CA activity. We fitted the [H⁺] change as a function of time by the equation: [H⁺] (t) = A + B (1-e^{(−KΔ[H+])t}), where t is time, A is the [H⁺] value after full equilibration, B is the [H⁺] range, and K_Δ[H+] is the rate constant of the [H⁺] relaxation. We obtained A, B, and K_Δ[H+] using the curve-fitting analysis of Sigma Plot 13.0. [H⁺] recordings in combination with the stopped-flow method were performed in the extracellular matrix approximately 50 µM away from the PMC cell (background) or with the electrode touching the PMC surface. To inhibit CAs, AZM was added at a concentration of 200 µM to the Tris-buffered ASW solution, leading to a final concentration of 100 µM in the perfusion chamber. Since iCA activity has no effect on the detection of eCA activity by this method (that is restricted to detect eCA activity at the cell surface), we used AZM to pharmacologically demonstrate CA activity at the outer surface of PMCs. In addition, eCA activity measurements were carried out in scramble and Cara7 MO-injected larvae.

Immunofluorescence Staining and Western Blot Analysis of the Sea Urchin Cara7 Protein.

Immunofluorescence (IF) and Western blot analysis was performed as previously described (5). Briefly, for IF larvae were fixed with 4% PFA/FWS for 10 min, washed with PBS, and blocked with PBS containing 5% bovine serum albumin (BSA) for at least 1 h at room temperature. Afterward, the larvae were incubated overnight at 4 °C with the affinity purified polyclonal antibody (dilution 1:250) raised against the sea urchin Cara7-specific synthetic peptide (NH2-CGQNAVGNQHPQWSN-CONH2) (Pineda). After intensive washing with PBS, larvae were incubated with the secondary antibody, a fluorophore-conjugated Alexa fluor 488 anti-rabbit IgG (Invitrogen) at a dilution of 1:300, for 1 h at room temperature in the dark. After washing with PBS, larvae were mounted on glass slides and pictures were taken on a confocal microscope equipped with an Airyscan function (LSM800/P2, Zeiss) and analyzed with Zen 3.3 blue edition.

For Western blot, at least 200 to 300 larvae were collected, and extracted by direct mixing with 1:10 (w/vol) 1× Lämmli loading buffer. Samples were incubated at 95 °C for 5 min and proteins were fractionated by SDS/PAGE on a 10% polyacrylamide gel and blotted to a PVDF membrane (Bio-Rad), using a tank blotting system (Bio-Rad). Blots were blocked for at least 1 h at room temperature with PBS-Tween buffer containing 5% BSA. Afterward, the blots were incubated in a 1:10,000 dilution of the primary antibody at 4 °C overnight. After intensive washing with PBST buffer, the blots were incubated for 1 h at room temperature with the secondary antibody a peroxidase-conjugated goat anti-rabbit IgG (Dianova) at a 1:40,000 dilution. After washing with PBST buffer, the protein signals were detected with the ECL Select Western Blotting Detection Reagents (GE Healthcare) and photographed with a Gel Doc 2000 system equipped with a CCD camera (Bio-Rad) and evaluated with ImageLab. For the peptide compensation assay, the primary antibody was preincubated with the immunization peptide at a concentration 0.04 mg/mL overnight at 4 °C. Finally, Western blot analysis was used to validate our Cara7 knockdown. Approximately 300 larvae of Cara7 or scramble MO-injected larvae were collected and used for Western blot analysis as described above. The Cara7 protein abundance was normalized to total protein concentrations determined by subsequent Coomassie staining of the membrane. The blot was washed three times with H₂O and stained for ∼1 h with 0.1% Coomassie in 40% MeOH and 7% acetic acid. Afterward the membrane was destained with 50% MeOH and 7% acetic acid for ∼1 h, washed with water, air dried, and pictures were taken.

Statistical Analyses.

Our data were tested for variance homogeneity (Levene test) and were analyzed for significance using GraphPad Prism6. iCA measurements with the CO₂ pulse method, MO injection validation, and ammonia prepulse measurement were analyzed using the Student’s t test (two-sided) with P < 0.05 and P < 0.001. Recalcification experiments, eCA activity measurement, morphometric analyses of Cara7 knockdown, and qPCR results were analyzed for significance using one-way ANOVA, followed by post hoc tests.

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.