Paracellular transport to the coral calcifying medium: effects of environmental parameters

The ocean's most biodiverse ecosystems, coral reefs, rely on a structural framework formed largely from the skeletons of scleractinian corals. Aspects of global climate change that directly impinge on the ability of corals to form their skeletons by calcification are therefore considered as major threats to these ecosystems (Hoegh-Guldberg et al., 2007; Kornder et al., 2018; Wild et al., 2011). Ocean acidification, which involves decreases in seawater pH and associated changes in carbonate chemistry caused by seawater uptake of anthropogenic carbon dioxide, impairs coral calcification rates and causes changes in skeletal structure (Chan and Connolly, 2013; Fantazzini et al., 2015; Tambutté et al., 2015). Temperature increases associated with climate change also have deleterious effects on coral calcification rates (Cantin et al., 2010; Carricart-Ganivet et al., 2012; Lough and Cantin, 2014). In light of the ecological importance of corals and their environmental sensitivity, recent years have seen a surge of interest in attaining a better mechanistic understanding of coral calcification (Weis and Allemand, 2009).

Furthering understanding of the coral calcification mechanism relies on an improved knowledge of the underlying physiology set in context with coral histology. Like all cnidarians, scleractinian corals are diploblastic, composed of two ectoderm and two endodermal cell layers. The endoderm cell layers line the internal coelenteron lumen and can contain dinoflagellate algae (family Symbiodiniaceae) in photosymbiotic coral species (LaJeunesse et al., 2018; Venn et al., 2008). The oral ectoderm faces the external seawater and the aboral ectoderm (calicoblastic epithelium) faces the skeleton. An extracellular calcifying medium (ECM) is situated between the calicoblastic cells and the growing front of the skeleton (Tambutté et al., 2011). The coral skeleton itself is composed of calcium carbonate arranged in an intricate architecture at both micro- and macroscales (Allemand et al., 2011; Cuif and Dauphin, 2005; Drake et al., 2020). Recent studies suggest that calicoblastic cells produce amorphous calcium carbonate intracellularly, which is introduced into the ECM (Drake et al., 2020; Mass et al., 2017; Neder et al., 2019). Research also shows that calicoblastic cells secrete an organic matrix that influences calcium carbonate formation (Goffredo et al., 2011; Mass et al., 2013; Tambutté et al., 2006; Von Euw et al., 2017). Importantly, the calicoblastic cells also supply ions to the ECM and regulate its composition by transcellular ion transport processes (Tambutté et al., 2011; Tresguerres et al., 2017). This involves a biologically controlled elevation of pH and calcium ion concentration ([Ca²⁺]) in the ECM (Decarlo et al., 2018; Sevilgen et al., 2019; Venn et al., 2011). Dissolved inorganic carbon (DIC) concentration ([DIC]) is also elevated in the ECM, with DIC being derived principally from metabolic sources, but also from seawater (Allison et al., 2014; Furla et al., 2000; McCulloch et al., 2017; Sevilgen et al., 2019). The result of the elevation of pH, [Ca²⁺] and [DIC] in the ECM is to increase the saturation state of aragonite (the dominant mineral form of calcium carbonate in coral skeletons), which promotes the calcification reaction (DeCarlo et al., 2017; Sevilgen et al., 2019). In symbiotic scleractinians, the photosynthetic activity of dinoflagellate symbionts in the endoderm cell layer increases coral calcification rates in a process called light-enhanced calcification (LEC) (Goreau, 1959). Although the mechanism underlying LEC has not yet been fully determined, it has been proposed to involve the influence of coral symbionts on certain aspects of biological control of calcification mentioned above (Gattuso et al., 1999; Tambutté et al., 2011). Furthermore, previous work indicates that mechanisms of biological control of calcification could shape how corals respond to ocean acidification (Comeau et al., 2017; McCulloch et al., 2012; Raybaud et al., 2017; Tambutté et al., 2015).

In parallel to transcellular ion transport mechanisms, paracellular transport is also potentially important to coral calcification (Gagnon et al., 2012; Ganot et al., 2015; Tambutté et al., 2012). Generally speaking, paracellular transport involves the passive movement of ions and molecules along their electrochemical gradient in between cells, contrary to transcellular transport, which involves primary or secondary active transport across cell membranes (Anderson and Van Itallie, 2009; Madara, 1998). Intercellular junctions can serve to control paracellular transport through their permeability and selectivity to particular ions and molecules. This function is carried out by occluding junctions, of which there are two types: tight junctions and their invertebrate analogue, septate junctions (Banerjee et al., 2006; Magie and Martindale, 2008). Other types of junctions serve to provide mechanical binding between cells, ensuring structural integrity of the epithelium, including desmosomes and adherens junctions (Garcia et al., 2018). In corals, there are only a limited number of investigations into intercellular junctions, but histological work has identified the presence of septate junctions (Clode and Marshall, 2002; Johnston, 1980; Tambutté et al., 2007), and their molecular structure has been characterised in the model coral Stylophora pistillata Esper 1797 (Ganot et al., 2015).

For some time, contradictory views have been held over the leakiness versus tightness of paracellular permeability in corals. The degree to which ions and molecules can pass from seawater to the ECM via the paracellular pathway could have important implications for the composition of the ECM and its biological control by the calicoblastic cells. Many models of coral calcification invoke the direct transport of seawater to the site of calcification, implying the existence of a leaky, open paracellular pathway between the ECM and seawater (Adkins et al., 2003; Cohen and McConnaughey, 2003; Saenger and Erez, 2016). This view is supported by studies measuring the uptake of metal/calcium isotopes from seawater into the coral skeleton and by the passage of high-molecular-mass fluorescent dyes via the paracellular pathway (Erez and Braun, 2007; Gagnon et al., 2012; Tambutté et al., 2012). In contrast, other models argue for a tight paracellular pathway, with restricted exchange of the ECM with seawater imposed by the intercellular junctions (Hohn and Merico, 2015; Tambutté et al., 2012; Taubner et al., 2017). This view is supported by relatively high levels of transepithelial resistance of the coral tissues in electrophysiology experiments (Tambutté et al., 2012; Taubner et al., 2017).

Research on vertebrate and invertebrate model organisms, and cell cultures has shown that paracellular permeability is not necessarily fixed, but can be variable, changing between states of greater leakiness or tightness (Anderson and Van Itallie, 2009; Beyenbach and Piermarini, 2011). Studies have shown that changes in paracellular permeability can occur in response to physiological or environmental stress, including work on invertebrate septate junctions, which vary in their permeability due to temperature stress (Livingston et al., 2020; MacMillan et al., 2017). This raises the possibility that paracellular permeability in corals could also be shaped by temperature and other environmental stressors such as ocean acidification. Until now, this possibility has received relatively little attention in corals, although mathematical models have considered effects on ocean acidification on paracellular permeability and their consequences on coral calcification (Hohn and Merico, 2015). The consequences of changes in paracellular permeability driven by environmental stress could involve an increase in ion exchange between the surrounding seawater and the ECM, potentially making biological control of ECM chemistry more difficult, leading to decreases in calcification rate (Galli and Solidoro, 2018; Ganot et al., 2015; Hohn and Merico, 2015). Conversely, if decreases in permeability were to occur, they may effectively isolate the ECM, enhancing biological control, providing more favourable conditions for calcification. In the case of ocean acidification, previous investigations with S. pistillata indicate that acidification causes decreases in pH in the ECM, leading to decreased ECM saturation states and calcification rates (McCulloch et al., 2012; Venn et al., 2013, 2019). However, it is not known whether this is accompanied by an impact of ocean acidification on paracellular permeability. In the case of temperature stress, previous studies show that calcification rates decline outside an optimum temperature range (Bernardet et al., 2019; Castillo et al., 2014; Jokiel and Coles, 1977). Little is known about the mechanistic basis underlying thermally driven declines in calcification rate, and the impacts of paracellular transport have not been investigated. It is also not known, from in vivo studies, whether temperature causes differences in pH in the ECM.

In the current investigation, we hypothesised that paracellular permeability between seawater and the ECM is not fixed, but increases (becomes leakier) under exposure to environmental stress, which is known to decrease calcification rates. To investigate this hypothesis, we developed an in vivo method to measure relative changes in paracellular permeability in corals. Our technique was based on an approach frequently used in vertebrate and invertebrate model systems that involves measuring the transepithelial flux of cell-impermeable dye tracers, such as calcein, lucifer yellow and sodium fluorescein (Caswell et al., 2013; Hanani, 2012; Helms et al., 2016; Inokuchi et al., 2009; Livingston et al., 2020; MacMillan et al., 2017; Molenda et al., 2014; Yang et al., 2017, 2018). As fluorescent dyes have previously been used in corals to image both intercellular spaces and the ECM (Ohno et al., 2017a,b; Tambutté et al., 2012; Venn et al., 2011, 2013, 2019), there was good potential to use this approach to investigate paracellular permeability. In our study, we used the cell-impermeable fluorescent dye calcein, a non-toxic compound that has been widely used in studies on coral calcification (Holcomb et al., 2013). Using confocal microscopy, we worked with the common reef-building, Indo-Pacific symbiotic coral S. pistillata (order Scleractinia, family Pocilloporidae) using time-lapse imaging to measure the rate at which calcein entered the ECM from the surrounding seawater. We derived a half-time of calcein influx from these data as a measurement of paracellular permeability to the ECM. Initially, we tested whether variation in paracellular permeability could be detected by measuring calcein influx rates in the coral ECM by exposing corals to hyperosmotic shock, a known modulator of intercellular junction permeability (Dan-Sohkawa et al., 1995). Next, we used the half-time of calcein influx approach to investigate whether seawater acidification caused relative changes in paracellular permeability. For our experimental treatment, we targeted a level of seawater acidification (pH 7.2) that is already known to decrease pH in the ECM and calcification rates in our model species S. pistillata (Tambutté et al., 2015; Venn et al., 2013). Finally, we investigated the effect of decreased and elevated seawater temperature on paracellular permeability using the half-time of calcein influx approach. Again, we used experimental treatments, 20°C and 31°C, that have previously been shown to decrease calcification rates in this species (as shown on its calcification thermal performance curve) (Bernardet et al., 2019). Whereas pH in the ECM and the calcifying cells has been characterised previously under seawater acidification in S. pistillata, these parameters have not been analysed under temperature stress. As such, we also measured pH in the ECM and the calcifying cells in corals exposed to 20°C and 31°C.

Stylophora pistillata was grown in the long-term coral culture facilities at the Centre Scientifique de Monaco. Aquaria were supplied with flowing seawater from the Mediterranean Sea, filtered to 5 µm, with a flow through exchange rate of 2% h⁻¹, a salinity of 38, pH 8, under an irradiance of 200 µmol photons m⁻² s⁻¹ of photosynthetically active radiation (400–700 nm) on a 12 h:12 h light:dark cycle. A temperature monitoring system (Enoleo, Monaco) maintained the temperature at 25±0.3°C (mean±s.d.). Corals were fed daily with frozen rotifers and twice weekly with live Artemia salina nauplii.

Stylophora pistillata used in the investigation was prepared from larger colonies (Fig. S1A,B) that had been propagated by fragmentation from a single colony collected from the Red Sea over 30 years ago. Microcolonies were made by the lateral skeleton preparative (LSP) assay on glass coverslips (Fig. 1A, inset; Fig. S1C–F) (Muscatine et al., 1997; Raz-Bahat et al., 2006; Venn et al., 2011). Briefly, branch tips of larger colonies were placed on glass slides, where they grew out laterally covering the slide as a thin sheet. Pieces of sheets were sectioned from the colonies with a razor blade and fixed with resin (Devcon™) to circular glass coverslips. These pieces (termed microcolonies here) were then left to grow out across glass coverslips under the conditions described above to a size of 1 cm² over 3 weeks. Separate microcolonies were used for all experiments.

Stock solutions of calcein (Sigma-Aldrich) were prepared at 2 g l⁻¹ as in Tambutté et al. (2012). From this stock, two types of solution were prepared in 0.45 µm filtered seawater: a spike solution of 300 µmol l⁻¹ calcein [designed to be added to the perfusion chamber (see below)] and a perfusion solution of 100 µmol l⁻¹ calcein (designed to provide a constant concentration of calcein under flowing conditions during the experiment).

Coral microcolonies were fitted into a perfusion chamber (PeCon, Germany) [which was then mounted on the confocal microscope (details below)] containing 2 ml seawater and perfused with seawater at a rate of 1.5 ml min^–1 in darkness for a 20 min period. At time zero of each time lapse, seawater perfusion was halted and 1 ml of the calcein spike solution (300 µmol l⁻¹) was mixed by pipette with the seawater in the perfusion chamber to bring the final volume to 3 ml and the final calcein concentration to 100 µmol l⁻¹. At the same time, the perfusion of 100 µmol l⁻¹ calcein in seawater was started and continued for the duration of the time-lapse imaging period at a rate of 1.5 ml min^–1. All time lapses in the investigation were made in darkness and the temperature kept at 25°C except in temperature treatments (described below).

Perfusion chambers with corals were mounted on an inverted confocal laser scanning microscope (Leica, Germany) for analysis. Hyperosmotic and seawater acidification experiments (described below) were carried out with a Leica SP5 model; temperature experiments (described below) were conducted with a Leica SP8. Imaging was conducted using a 40× oil immersion lens with laser excitation at 488 nm (2% intensity), emission capture at 520±10 nm, Gain 800, with the pinhole set at 1 Airy unit. Using these settings, sequences of five to six optical sections in the vertical (z) axis (z-stacks), slice depth 1 µm, were captured in the ECM and calcifying cell layer at each step of the time lapse. The initial z-stack (time zero) was taken before calcein addition (to check for background fluorescence), the next at 1 min after calcein addition and every 3 min thereafter for a maximum duration of 65 min. Wide-field images were also captured simultaneously. Interference by green fluorescent proteins (GFPs) was checked by measuring background fluorescence in samples not stained with calcein using identical parameters. GFPs are sometimes visible in the endoderm layer of S. pistillata, but were not detected in the calcifying cell layer with the instrument settings used here.

The half-time of calcein influx was determined by measuring the rate of increase of calcein fluorescence in the ECM, from the time course of confocal z-stacks using LAS X software (v. 3.5.2; Leica). Mean ECM calcein fluorescence was obtained from three to five digital regions of interest (ROIs) created in z-stacks of the ECM at each time point. ROIs did not overlap with the brightly stained crystals that were saturated in fluorescence. Where necessary, images could be viewed using the glow look-up table (LUT) that clearly differentiated the crystals as a separate colour to the ECM (Fig. S2). The half-time of calcein influx was determined using an approach adapted from Tambutté et al. (1996). Calcein fluorescence (F) data were plotted against time and fitted with an exponential curve. Fluorescence at equilibrium (F_eq) was calculated by taking the mean fluorescence value of the first eight time points at the plateau (where calcein fluorescence was stable). A linear plot of ln(F_eq–F) against time was produced in which the slope is the rate constant (k). Half-time of calcein influx was then calculated from k using the equation ⁠.

Hyperosmotic shock experiments were conducted to test whether measurements of calcein influx rate (see below) could be used to detect changes in paracellular permeability. For these experiments, seawater solutions with and without calcein were prepared by adding glycine to a final concentration of 200 mmol l⁻¹, pH 8.0 (Tambutté et al., 2012). At the start of experiments, microcolonies were initially perfused with seawater containing 200 mmol l⁻¹ glycine (no calcein) for 15 min in darkness. At time zero, addition of the calcein spike and perfusion solutions was carried as above, but with each solution also containing 200 mmol l⁻¹ glycine. The same procedure was used for controls in the absence of glycine. The calcein influx time-lapse imaging procedure was carried out as described above in darkness. Measurements were carried out in five independent samples in each condition (N=5).

For exposure to seawater acidification, microcolonies were transferred to 30 l treatment aquaria for 1 week at pH 7.2 or pH 8.1, at 25±1°C. The seawater acidification aquarium setup has been described in several previous publications (e.g. Tambutté et al., 2015; Venn et al., 2013, 2019), and the seawater carbonate chemistry parameters for the current study are reported in Table S1 and are the same as those reported in Venn et al. (2019). Briefly, seawater was supplied at a renewal rate of 60% h^–1, and the pH 7.2 treatment aquaria were bubbled with CO₂ to manipulate carbonate chemistry and maintain pH at 7.2, whereas no CO₂ bubbling was carried out in the pH 8.1 treatment. Corals were fed daily with frozen rotifers and twice weekly with live Artemia salina nauplii. pH and temperature were monitored continuously at 1 min intervals with a custom-made monitoring system (Enoleo, Monaco), which controlled CO₂ bubbling rates and heating elements. Additionally, spectrophotometric pH measurements and total alkalinity measurements were taken weekly as described in Venn et al. (2019). Five independent microcolonies were exposed at each pH (N=5). Each pH treatment (pH 8.1 and pH 7.2) was represented by two aquaria with the microcolonies split between them (two in one aquarium, three in the other). After 1 week at a seawater pH of 7.2 or 8.1, microcolonies were analysed with the calcein influx time-lapse imaging procedure described above in darkness, with calcein-seawater solutions at pH 7.2 or pH 8.1 at 25°C, prepared with water drawn directly from the pH treatment aquaria. Five independent microcolonies were analysed at each pH (N=5). There was no observed direct effect of pH on calcein fluorescence in seawater, as its fluorescence is insensitive to pH changes between pH 6.5 and pH 9 (Bégu et al., 2002). We confirmed this by measuring calcein fluorescence in seawater at pH 7.2 and pH 8.

Microcolonies were transferred to treatment aquaria at 25±0.5°C (mean±s.d.), 20±0.5°C or 31±0.5°C at pH 8 for 1 week. The temperature treatment aquarium setup has been described recently (Bernardet et al., 2019). Flow-through seawater was supplied at 60% h^–1 renewal rate, and temperature and pH were monitored continuously at 1 min intervals with a custom-made monitoring system (Enoleo, Monaco). Each temperature level was represented by two 30 l aquaria and independent replicate microcolonies were split as evenly as possible between the two aquaria at each temperature (e.g. two in one aquarium, three in the other if N=5). Replicate numbers at each temperature were as follows: N=6 at 25°C, N=5 at 20°C, N=7 at 31°C. After 1 week, imaging was carried out according to the calcein influx time-lapse imaging procedure described above in darkness, with calcein-seawater solutions at 25, 20 or 31°C.

Microcolonies were transferred to the same treatment aquaria described above for 1 week with a temperature of 25±0.5°C (control), 20±0.5°C or 31±0.5°C at a seawater pH of 8. After 1 week, measurements of pH in the ECM (pH_ECM) and intracellular pH (pH_i) of calcifying cells were made in separate microcolonies according to methods published previously (Venn et al., 2011, 2019). Briefly, measurements were performed using inverted confocal microscopy (Leica SP5) and the ratiometric pH-sensitive dye SNARF-1 (Invitrogen). Using the same perfusion apparatus and flow rates used for permeability experiments, we used 45 µmol l⁻¹ cell-impermeable SNARF-1 for pH_ECM analysis and 10 µmol l⁻¹ cell-permeable SNARF-1 AM for pH_i analysis. All measurements were carried out at 40× magnification by excitation at 543 nm at 30% laser intensity, and fluorescence captured at emission wavelengths of 585±10 nm and 640±10 nm, Gain 1000, with the pinhole set at 1 Airy unit. For each measurement, 10 to 12 optical sections were captured in a z-stack with an optical slice thickness of 1 µm. Separate calibrations of SNARF-1 were made at each temperature for both pH_ECM and pH_i according to Venn et al. (2011). pH_ECM was analysed in dark conditions in four to five samples [20°C (N=4), 25°C (N=5) and 31°C (N=5)]. Calicoblastic cell pH_i was analysed separately in darkness in four to five colonies [20°C (N=4), 25°C (N=4) and 31°C (N=5)]. Independent microcolonies were used and replicates were divided between the two aquaria at each temperature level as described above.

Linear and exponential fits were performed using the software R version 1.2.1335. Statistical analysis was performed using IBM SPSS software v. 24. Data plots and Shapiro–Wilk tests were used to check for data normality; Levene's tests were used to check for homogeneity of variance. Student's t-tests were used to compare data in the hyperosmotic and acidification experiments. A Kruskal–Wallis test and Mann–Whitney U post hoc analysis were used to compare between temperature treatments. One-way ANOVA was used to analyse pH_ECM. Data are reported in the text as means±s.d. In figures, data from each treatment are presented as box plots.

Inverted confocal imaging allowed visualisation of the ECM, calicoblastic cells and newly formed isolated calcium carbonate crystals at the distal growing edge of corals prepared by the LSP assay (consistent with our previous work, e.g. Venn et al., 2011). Addition of calcein to the seawater medium surrounding the corals resulted in staining of the calcium carbonate crystals and the ECM, with calcein fluorescence being strongest in the crystals, making them easy to distinguish against the ECM, which exhibited a lower fluorescence (Fig. 1A,B). By contrast, calcein was excluded from the cytosol of the calicoblastic cells, which appeared black. When corals were maintained on the microscope for 1–5 h, crystals could be seen to be actively growing in contact with the ECM and the calicoblastic cells (Fig. 1C–F).

To analyse the dynamics of movement of calcein from the surrounding seawater to the ECM via the paracellular pathway, we carried out time-lapse confocal imaging of calcein fluorescence in ROIs in the ECM (Fig. 2). Following its addition to the perfusion chamber at time zero, calcein fluorescence exhibited saturable kinetics in the ECM, appearing in the first minute, then increasing over time until reaching a plateau after 25–40 min (Fig. 2A–E). Semi-logarithmic treatment of these data (Fig. 2E, inset) allowed us to determine the half-time of the influx of calcein to the ECM of each colony to serve as an indicator of renewal rate of calcein in the ECM by the paracellular pathway.

When measuring calcein influx in ROIs at different positions in the ECM, we found that the half-time of calcein influx could vary spatially. This spatial variation was not related to the distance of the ROI to the colony edge within the field of view captured by our images (Fig. 3). To take this spatial variation into account, we determined the half-time of calcein influx in multiple ROIs in the ECM for each coral sample and used the mean value as the measure of permeability for that sample. Measurement of five colonies of S. pistillata yielded a mean half-time of calcein influx into the ECM of 6.37±0.73 min (mean±s.d.).

Previous work on invertebrates and vertebrates, including S. pistillata, have found that exposure to hyperosmotic conditions affects intercellular junction permeability, increasing leakiness of the paracellular pathway (Dan-Sohkawa et al., 1995; Inokuchi et al., 2009; Nilsson et al., 2007; Tambutté et al., 2012). Here, we tested whether hyperosmotic effects on paracellular permeability could be detected by measuring calcein influx rates in the coral ECM. Colonies perfused with a hyperosmotic seawater solution exhibited decreased half-times of calcein influx of 3.80±0.79 min relative to values of 6.87±0.9 min in seawater controls, indicating an increase in the permeability of the paracellular pathway to the ECM (Fig. 4) [t₈=−5.676; P<0.001].

To investigate whether seawater acidification affects paracellular permeability to the ECM, we exposed colonies to seawater at pH 8.1 or at pH 7.2 for 1 week (see Table S1 for carbonate chemistry). Half-times of calcein influx were significantly decreased in corals exposed to pH 7.2 (5.20±0.86 min) relative to pH 8.1 (6.94±0.95 min), indicating an increase in paracellular permeability under these conditions [t₈=3.032; P<0.05] (Fig. 5).

Exposure of coral colonies to elevated and decreased temperature for 1 week elicited different responses (Fig. 6). At 31°C, half-times of calcein influx [mean 7.87±2.19 (mean±s.d.)] were not significantly different from those at 25°C (6.88±1.11 min), whereas at 20°C, half-times were significantly higher than at 25°C (9.31±0.91 min) (Kruskal–Wallis: =6.513; P<0.05, Mann–Whitney U post hoc analysis). These results suggest that although paracellular permeability was unaffected by elevated temperature, it significantly decreased at low temperature.

In temperature treatments, pH_ECM remained above the surrounding seawater (pH 8) at similar levels in the three temperatures: pH 8.2±0.05 at 20°C, pH 8.2±0.1 at 25°C and pH 8.1±0.1 at 31°C (Fig. 7A) [one way ANOVA: F_2,11=1.242; P>0.05]. Similarly, no effect of temperature on pH_i was observed in the calcifying cells (pH 7.3±0.05 at 20°C, pH 7.3±0.04 at 25°C and pH 7.3±0.05 at 31°C) (Fig. 7B) [one way ANOVA: F_2,10=0.820; P>0.05].

The paracellular pathway is a key consideration in models of ion transport for calcification and could also be important in understanding the sensitivity of coral calcification to environmental change (Hohn and Merico, 2015). Currently, however, very little functional information is available on the properties of this pathway and its permeability in corals and marine calcifying organisms in general. Here, we developed an approach to characterise rates of influx of the dye calcein from seawater to the ECM and used this technique to determine whether environmental parameters affect paracellular permeability.

Fluorescent dyes are widely used in studies on the paracellular pathway to assess relative changes in its permeability (e.g. MacMillan et al., 2017). In the current investigation, the dye calcein was selected because it is non-toxic to corals and several other marine calcifiers (Fox et al., 2018; Holcomb et al., 2013), and it is brightly fluorescent even at relatively low laser-excitation intensities, thus making it highly amenable to in vivo imaging. Furthermore, our previous research has imaged this dye in intercellular spaces in S. pistillata, demonstrating that movement of calcein from seawater to the ECM occurs via the paracellular pathway (Tambutté et al., 2012). Although calcein can be incorporated in vesicles in S. pistillata (Neder et al., 2019; Ohno et al., 2017b), recent research shows that vesicles incorporating fluorescent dyes move in an apical-to-basal direction in each individual cell layer, and do not traverse the coral tissues from seawater to the ECM (Ganot et al., 2020). As such, vesicle transport is therefore unlikely to move calcein from seawater to the ECM.

In a first step, we tested whether we could use measurements of the half-time of calcein influx to detect changes in paracellular permeability. For this, we exposed corals to a well-known modulator of paracellular permeability: a shift to hyperosmotic conditions. Previously, a variety of techniques have been used to observe this effect in vertebrates and invertebrates including dyes, metal tracers and electrophysiology (Dan-Sohkawa et al., 1995; Högman et al., 2002; Inokuchi et al., 2009; Nilsson et al., 2007). All these studies show that hyperosmotic treatment leads to a shift towards a ‘leakier’, less restrictive, paracellular permeability, and this has also been shown in corals by measuring transepithelial electrical resistance (Tambutté et al., 2012). The mechanism underlying hyperosmotic shock-induced changes in paracellular permeability involves changes in the tertiary structure of proteins in intercellular junctions and cell-to-cell contacts (Dan-Sohkawa et al., 1995; Nilsson et al., 2007).

In the context of the current study, our experiments with hyperosmotic conditions allowed us to check that measurements of the half-time of calcein influx to the ECM were sensitive to changes in paracellular permeability. In hyperosmotic treatments, we observed that the half-time of calcein influx to the ECM decreased, consistent with a shift from a more restrictive permeability under normal osmotic conditions to leakier paracellular permeability under hyperosmotic conditions. This finding demonstrates that the half-time of calcein influx can be used as an indicator of change of paracellular permeability to the coral ECM.

To date, relatively little consideration has been given to the impacts of environmental stress on paracellular permeability in corals. Here, we investigated the effects of decreases in pH (seawater acidification) and temperature stress, which have both been shown to influence paracellular permeability in other organisms (Ferreira and Hill, 1982; MacMillan et al., 2017). We summarise our findings in Fig. 8 in context with responses of pH_ECM, pH_i of the calicoblastic cells and calcification in the current study and in the literature. In the current study, exposure to a seawater acidification of pH 7.2 resulted in an increase in paracellular permeability, and it has been shown previously that the same seawater acidification treatment of pH 7.2 depresses pH_ECM, pH_i and calcification rates in S. pistillata (Tambutté et al., 2015; Venn et al., 2013, 2019). By contrast, during temperature stress, paracellular permeability decreased or remained unchanged. Under these conditions, there is no impact of temperature on either pH_ECM or pH_i, but calcification rates declined at both decreased and elevated temperatures (Bernardet et al., 2019). In the following paragraphs, we consider these findings in greater detail.

Turning first to seawater acidification, our choice of treatment (pH 7.2) was based on our previous investigations of S. pistillata, in which significant declines in calcification rate, a shift in skeletal architecture, and declines in both pH_ECM and pH_i were found to occur at a seawater pH of 7.2 in this species (Tambutté et al., 2015; Venn et al., 2013, 2019). We thus selected this seawater acidification treatment for the current investigation. Following a 1-week exposure to pH 7.2 seawater, we found significant decreases in the half-time of calcein influx relative to pH 8.1 treatments, pointing to an increase in paracellular permeability (leakiness). The mechanism underlying this acidification-induced change in permeability in corals is unknown, but work on other organisms shows that decreases in pH cause changes in permeability through their effects on intercellular junction structure. For example, in the human stomach, acid-induced changes in paracellular permeability occur through modifications of the gastric epithelial junctions (Marcus et al., 2013). Investigations on frog skin have shown how exposure to low pH can increase paracellular permeability through opening of tight junctions and widening of intercellular spaces (Ferreira and Hill, 1982). A further possibility is that changes in paracellular permeability are linked to depressions in intracellular pH (pH_i) of the calicoblastic cells that have been recorded previously at a seawater pH of 7.2 for this species (Venn et al., 2013, 2019). Alteration of pH_i affects a myriad of cellular processes, including regulation of cell volume and the cytoskeleton (Casey et al., 2010). Changes in pH_i could have affected cellular morphology at the level of cell-to-cell contact and/or the conformation of junction proteins and their anchoring with the cytoskeleton, possibility increasing paracellular permeability. Further research is clearly required to investigate these possibilities and could involve ultrastructural observations of calicoblastic cells (Tambutté et al., 2007).

Although the mechanism by which pH affects coral paracellular permeability is not yet known, its physiological consequences could be relevant to understanding how low seawater pH can affect ECM chemistry. It is known that exposure of S. pistillata to pH 7.2 seawater causes decreases in pH_ECM, but these decreases in pH_ECM are mitigated by mechanisms of acid-base regulation that maintain a greater proton gradient between the ECM and the surrounding seawater, presumably at greater energetic cost (Ries, 2011; Venn et al., 2013). If the changes in paracellular permeability identified here extend to the paracellular flux of protons and other ions, then the challenge to acid-base regulation of the ECM would be anticipated to be even greater than previously thought, because increased paracellular leakiness could enhance H⁺ entry and loss of CO₃²⁻ from the ECM along their respective concentration gradients, which could potentially decrease pH_ECM and ECM buffering capacity. This hypothetical mechanism linking paracellular permeability to calcification has been proposed in previous studies (Galli and Solidoro, 2018; Ganot et al., 2015; Hohn and Merico, 2015), and although it warrants further investigation, our results suggest that paracellular permeability is an important parameter for calcification under seawater acidification because it can affect ECM chemistry.

Turning next to temperature stress, it is known from work on septate junctions in Drosophila that exposure to low temperatures can increase paracellular permeability, leading to a loss of gut epithelia barrier function (MacMillan et al., 2017). In corals, little is known about the effect of temperature on paracellular transport, and we investigated whether temperature variation could increase paracellular permeability because it could provide mechanistic insight into why high and low temperatures lead to decreases in coral calcification rate (Bernardet et al., 2019). We hypothesised that although temperature does not directly affect the pH gradient with the surrounding seawater (as in the case of seawater acidification), temperature-induced increases in paracellular permeability could lead to increased proton flux between the ECM and seawater, leading to a decrease in pH_ECM, which could, in turn, be detrimental to calcification rates. However, our measurements of paracellular permeability, pH_ECM and pH_i, do not support this hypothesis, and our observations are not easy to relate to how calcification responds in S. pistillata. When we measured both pH_ECM and pH_i at high and low temperatures, we found no change in either parameter with respect to 25°C. Furthermore, despite temperature changes, pH_ECM remained upregulated at similar levels observed in previous measurements at a seawater pH of 8 (Venn et al., 2019). For calcification, decreases are known to occur in S. pistillata at both high and low ends of its thermal performance window (Bernardet et al., 2019), but temperature change appeared to only affect paracellular permeability at low temperatures. Recent work on gene expression in S. pistillata suggests that declines in calcification rates at each end of the thermal tolerance window are associated with disruption to transcellular transport and interconversion of DIC potentially used for calcification, whereas proton transporters are not affected (Bernardet et al., 2019). Furthermore, recent geochemical analysis of seasonal effects on ECM chemistry also indicate that pH_ECM remains stable despite seawater temperature change, although DIC levels in the ECM are altered (McCulloch et al., 2017). These previous studies are consistent with the lack of change in pH_ECM observed in the current study, and, together with our findings, they are a starting point from which more mechanistic research into the thermal sensitivity of coral calcification must be undertaken.

The relatively bulky nature of calcein [ionic radius of 6.5 Å and molar mass of 622.55 g mol⁻¹ (Zarnitsyn et al., 2008)] makes it tempting to assume that changes in the paracellular influx of calcein to the ECM represent changes in the influx of bulk seawater or some of its constituent ions. Indeed, the incorporation of calcein in the coral skeleton is considered by some authors to be a marker of direct bulk seawater transport to the site of calcification (Braun and Erez, 2004; Erez and Braun, 2007; Gagnon et al., 2012; Saenger and Erez, 2016). However, it should be kept in mind that intercellular junctions define selectivity to the passage of solutes with different properties (permselectivity) (Anderson and Van Itallie, 2009). Indeed, in vertebrates, intercellular tight junctions exhibit different permselectivity for substances that vary in size and charge (Schneeberger and Lynch, 2004). Invertebrate septate junctions are less well characterised, but selective permeability is known to occur, including in anemones and corals, although the limited number of studies means the permeability properties need further investigation (Bénazet-Tambutté et al., 1996; Taubner et al., 2017). Better-studied invertebrate models such as the mosquito Malpighian tubules even show that permselectivity can be dynamically regulated (Beyenbach and Piermarini, 2011). Furthermore, the paracellular passage of macromolecules, such as calcein, can be regulated by tricellular junctions without affecting ion permeability (Krug et al., 2009). All these considerations mean that we urge caution about assuming that the half-life of calcein influx can represent the paracellular passage of all ions and molecules in seawater. Additionally, it should be remembered that although calcein fluorescence reacts with Ca²⁺ and has sometimes been interpreted to be a tracer of seawater Ca²⁺, calcein also reacts with other metal ions including Mg²⁺, and can be quenched strongly by Cu²⁺, Fe³⁺ and Mn²⁺ (Markuzewski, 1976). As such, its fluorescence cannot be reliably used to quantify paracellular flux of Ca²⁺.

It is not known whether the observations made here for S. pistillata apply to other coral species. Previous studies have shown that coral species vary in their ability to regulate pH in the ECM against acidification (McCulloch et al., 2012). For example, seawater acidification decreases pH_ECM much more readily in Pocillopora damicornis and Acropora hyacinthus than in S. pistillata (Venn et al., 2019). The possible links between pH and ion regulation of the ECM and paracellular permeability require further research, but it would be interesting to investigate whether interspecies differences in the capacity of pH_ECM regulation against ocean acidification are shaped by differences in paracellular permeability (Galli and Solidoro, 2018). Future research could exploit the techniques developed in the current study to investigate this idea. In addition to investigating interspecies differences in paracellular permeability, it would also be useful to explore intraspecific variation in this trait. A caveat of the current study is that it was carried out on individuals propagated from a single genotype; therefore, future studies should consider incorporating several genotypes.

Future research on coral paracellular permeability should also incorporate light into the experimental design. The effect of light was not considered here, and analysis of calcein influx was carried out in darkness. Symbiotic dinoflagellate photosynthetic activity boosts coral calcification rates (light-enhanced calcification, see Introduction) and is an important factor for determining how coral calcification responds to both ocean acidification and temperature stress (e.g. van der Zande et al., 2020). Light exposure and photosynthetic activity can enable corals to maintain higher levels of pH_ECM under conditions of seawater acidification, which warrants investigation into how paracellular permeability responds to light under acidification (Venn et al., 2019). Similarly, decreases in calcification rate in S. pistillata that occur at elevated temperature in darkness are mitigated by photosynthetic activity during light exposure, which also justifies investigation into how paracellular permeability responds to light exposure in these conditions (Bernardet et al., 2019). Furthermore, exposure to elevated temperatures impairs photosynthesis and leads to dysfunction of the coral-algal symbiosis (coral bleaching), which can be exacerbated by ocean acidification (Anthony et al., 2008; Oakley and Davy, 2018). Given the importance of the coral-algal symbiosis to coral metabolism and calcification, future mechanistic experiments on the links between paracellular permeability and calcification should take the role of the dinoflagellate symbionts into account.

In conclusion, the current study used time-lapse imaging of calcein influx into the ECM to show that the environmental parameters of seawater pH and temperature can alter paracellular permeability in S. pistillata. The responses of paracellular permeability to increases and decreases in temperature were distinct to the response to decreased seawater pH. While seawater acidification increased paracellular permeability, exposure to low temperature caused decreased permeability, whereas elevated temperature resulted in no change. Much further work is required to better understand the mechanisms underlying these changes, and to better characterise paracellular permeability and selectivity in corals in general. The technique established in the current study is amenable to the use of other fluorescent dye markers with different properties to calcein, which could be used to better characterise selectivity of paracellular transport and, ultimately, more accurately define renewal rates of the ECM. Importantly, future studies are required to better interpret the laboratory-based findings of the current study in terms of their environmental relevance, with the ultimate goal of increasing understanding of the vulnerability and resilience of corals in a rapidly changing ocean.

We thank Natacha Segonds, Nathalie Techer, Dominique Desgré and Eric Elia for technical assistance and coral culture.