In vivo observation of lipid droplets in coral calcifying cells: fat stores to fuel the reef-building process?

The calcifying cells of corals are responsible for skeleton formation, a process that is the basis for reef building. Their cell biology is therefore of primary interest, but current knowledge is limited because direct in vivo investigation of this cell type is challenging. Studies at the growing edge of laterally extending coral colonies allow for direct observations of actively calcifying cells in living samples. In the current study, we used this approach to study lipid droplets, which are considered storage organelles that fuel cells for physiological processes. Using the fluorescent lipid probe Nile Red and in vivo confocal microscopy, we observed lipid droplets in calcifying cells of the coral Stylophora pistillata and we suggest that they play a key role in coral calcification.


Introduction
Hermatypic scleractinian corals (i.e.reef-building corals), key architects of the three-dimensional reef ecosystem framework, are lipid-rich organisms (up to 47% dry tissue weight) (Kopp et al. 2015;Stimson 1987) compared to other cnidarians, including sea anemones and jellyfishes, which display 5-20% and 1-12% dry tissue weight in lipids, respectively (Jennison 1979;Raposo et al. 2022;Revel et al. 2016).Structural lipids (e.g.membrane phospholipids) lipid droplets in the calicodermis, which differ in number and size from the ones contained in the aboral gastrodermis.

Results and discussion
The calicodermis observed at the growing edge of S. pistillata microcolonies displays an abundance of spherical, refringent bodies (Fig. 1A).Staining with the fluorescent lipid probe Nile Red revealed that these bodies are lipid droplets contained within calcifying cells (Fig. 1B, C).The mean density of LDs in the calcifying cells is significantly higher than in the overlying gastrodermal cell layer (respectively 242.6 ± 51.5/1000 µm 2 versus 30.2 ± 8.2/1000 µm 2 , mean ± S.D., n = 4) (Fig. 1 and Table 1).Cnidarian gastrodermal cells, which line the gastrovascular system, are involved in heterotrophic digestion, and in the case of symbiotic corals, contain endosymbiotic dinoflagellates (Symbiodiniaceae, LaJeunesse et al. 2018) responsible for autotrophic nutrition.By contrast, calcifying cells are neither in direct contact with the coelenteric fluid nor do they contain symbiotic cells and are dependent on other epithelia for their nutrient supply.Thus, lipid droplets contained in the calicodermis likely originate from de novo, in situ synthesis inside the calcifying cells from soluble nutrients, e.g.glycerol and glucose, acquired from other cell layers (Colombo-Pallotta et al. 2010;Davy et al. 2012).Alternatively, lipids can be translocated from symbiotic cells to actively calcifying zones (Pearse and Muscatine 1971).Note that the growing edge of microcolonies contains few Symbiodiniaceae (Venn et al. 2011, Figure 1, 2024, in press), therefore lipids contained in this region probably originate from heterotrophy and/or translocation of nutrients from symbiont-rich zones.Such a transfer was observed in the branching coral Acropora cervicornis and in the massive coral Orbicella annularis, where organic compounds undergo directional transport from symbiont-rich zones to the branch tips or to the forward growing-edge, respectively, which are actively calcifying zones containing few symbionts (Taylor 1977).Interestingly, in A.cervicornis symbiont-free branch tips,  (Pearse and Muscatine 1971).Of note, coral calcifying cells contain a large amount of mitochondria (Barott et al. 2015;Tambutté et al. 2007), which is not surprising as mitochondria and LDs are functionally linked by lipid metabolism and are often closely associated (Olzmann and Carvalho 2019).
In addition to showing differences in LDs density, comparison of LDs between the gastrodermis and calicodermis shows significative differences in size, with LDs in calcifying cells displaying a relatively narrow distribution (mean diameter 752 ± 31 nm, Fig. 2A and Table 1), while gastrodermal LDs form a heterogeneous population containing large LDs (1043 ± 166 nm with some droplets up to 4 µm in diameter, Fig. 2B and Table 1) (mean ± S.D, p < 0.05).These large LDs (> 1200 nm) represent 85.4 ± 6.7% of the total LDs volume in the gastrodermis, while they account for only 25.2 ± 7.1% of the LDs volume in the calicodermis (mean ± S.D.).This suggests different functional roles for these two LDs pools.The small LDs observed in the calcifying cells, and to a lesser extent, in the gastrodermis present a high surface/volume ratio compared to the larger ones, allowing for increased lipid supply to cell metabolism and interaction with LD-associated proteins.By contrast, large LDs may allow lipid storage in a less easily mobilised reserve (Suzuki et al. 2011).
Taken altogether, these findings indicate that coral calcifying cells are rich in relatively small lipid droplets, suggesting an active metabolism.This is not surprising as the process of calcification is an energy-consuming process, requiring the synthesis of skeletal organic matrix (Tambutté et al. 2011) along with a transepithelial flux of ions through  the calicodermis to elevate and maintain the aragonite saturation state of the extracellular calcifying medium (ECM) necessary for crystal growth (Sevilgen et al. 2019;Sun et al. 2020;Venn et al. 2022).Indeed, it has been estimated that calcification represents up to 30% of coral energy budget (Allemand et al. 2011).Therefore, LDs in the calcifying cells potentially represent a metabolic energy store capable of sustaining the constant energy expenditure associated with skeletogenesis and may be essential to fuel the calcification process in periods of reduced nutrient availability.Moreover, the role of these LDs might be crucial in the context of ocean acidification (OA).For example, research by (Vidal-Dupiol et al. 2013) reports that genes involved in lipid catabolism in Pocillopora damicornis are upregulated under experimental OA conditions obtained through enrichment of seawater with CO 2 , suggesting an increase in lipid consumption at low pH.Furthermore, it was observed that in Acropora hyacintus, low ECM pH is correlated with skeletal dissolution in darkness under experimental OA conditions obtained through CO 2 bubbling, while the OA resilient S. pistillata, which regulates higher pH in the ECM, is capable of maintaining calcification in the same conditions, presumably at increased energetic cost (Venn et al. 2019).
A comparative study of LD abundance in the calicoblastic epithelium of coral species with different OA susceptibility may shed a new light on the role of these overlooked organelles, and on their potential contribution to resilience of corals facing environmental stress.
Compared with studies where lipid abundance and metabolism were investigated at the tissue or whole-organism scale through en masse methods, our approach allows for the first direct observation of lipid stores in the calcifying cells of living, actively calcifying coral samples.More broadly, observation at the growing edge of LDs dynamics, along with pH regulation, crystal growth, mitochondria dynamics or other energy-related processes, under various experimental conditions, may open perspectives on the energy budget and the resilience of hermatypic scleractinian corals facing environmental stress, while providing avenues to explore the cellular biology of calcification.

Coral samples maintenance
Experiments were conducted on Stylophora pistillata microcolonies laterally grown on glass coverslips in the long-term coral culture facility at the Centre Scientifique de Monaco.Aquaria were supplied with seawater from the Mediterranean Sea with an exchange rate of 50% h −1 , under an irradiance of 100 µmol photons m −2 s −1 of photosynthetically active radiation on a 12 h:12 h light/dark cycle and at 25 °C.
Corals were fed with live Artemia salina nauplii twice a week, and thrice a week with frozen rotifers.Microcolonies were prepared following the lateral skeleton preparative assay (Muscatine et al. 1997;Venn et al. 2011).Briefly, flat portions (0.5-1 cm 2 ) of S. pistillata mother colonies grown on glass supports were fixed with Devcon® epoxy resin on circular glass coverslips and left to grow across coverslip to a size of 1-2 cm 2 over 3 weeks in the culture conditions described above.Coverslips were cleaned twice a week with a razorblade to remove algae growing around the microcolony.

Reagents
Hoechst 33342 (20 mM) was obtained from ThermoFisher.Nile red and calcein were obtained from Sigma-Aldrich.Nile red stock solution was prepared in anhydrous DMSO at 30 mM.Calcein stock solution was prepared in filtered seawater at 30 mM, and pH was adjusted at 8 with NaOH.

Microcolony staining with fluorescent probes
Incubations of S. pistillata microcolonies with fluorescent probes were performed between 3:00 pm and 5:00 pm, i.e. after 7-9 h of exposure to photosynthetically active radiation but before corals daily feeding.Coral samples were incubated for 15 min in FSW containing 20 µM Hoechst 33342 and 20 µM Nile red (final DMSO concentration: 0.07%) to label nuclei and lipid droplets, respectively.Microcolonies were briefly rinsed with FSW before mounting in a microscope chamber (PeCon GmbH) containing 3 mL of FSW.Once the sample placed in the microscope and ready for imaging, FSW was replaced with FSW containing calcein at 100 µM and pH 8 to label extracellular spaces (i.e.ECM, mesoglea and coelenteron) and allow further delineation of the two aboral cell layers during image processing.Imaging was performed immediately after calcein addition.

Microscopy
Observations were carried out at the growing-edge of S. pistillata microcolonies with a confocal laser scanning Leica SP8 inverted microscope monitored by the LAS X software (version 3.5.6.21594).Imaging was conducted using a 63 × water immersion objective (numerical aperture: 1.2) and 4 × scanner zoom.The resolution was 1024 × 1024 pixels and the z-step size 0.15 µm, resulting in voxels of 45 nm (x) × 45 nm (y) × 150 nm (z).These voxel dimensions fulfill the Nyquist-Shannon sampling criterion, as determined using the online Nyquist Calculator (SVI) tool for a 488 nm excitation wavelength and 500 nm emission wavelength.High-speed stack acquisition was performed using resonant scanner (8.000 Hz) in bidirectional mode with a line averaging of 3. The pinhole size was 1 Airy Unit.The fluorescent signal from Nile red was acquired separately from Hoechst and calcein (sequential acquisition) to prevent dye crosstalk.Laser and detector settings are detailed in Table 2. Images were deconvoluted on-the-fly by the Las X-integrated Lightning tool using default parameters.In addition to fluorescence, a bright-field image was acquired using a transmitted light detector.

Image processing and analysis
Image processing and analysis were conducted with the Las X software (Leica) including a 3D-analysis module.Image stacks were cropped to separate the calicodermis and the gastrodermis, using the calcein-stained mesoglea and Hoechst-stained nuclei as visual controls.Threedimensional image analysis was conducted as follows: images were smoothed using a gaussian filter (radius 2.5) and thresholded using a Maximum Entropy algorithm (deviation − 750).Objects in the resultant binary image were further delineated using a morphological filter (Discard Details, radius 2) and the Split Objects tool (object size 0.5 µm and 1.6 µm for calicodermis and gastrodermis, respectively).LDs number, volume and minimum Feret diameter were obtained from 4 independent experiments.For LDs diameters, statistical significance was assessed with the Mann-Whitney U test using the IBM SPSS Statistics 26 software.

Fig. 1
Fig. 1 Lipid droplets observed with an inverted confocal microscope, at the growing edge of a S. pistillata microcolony grown on a coverslip.A-C Calicodermis; D-F aboral gastrodermis.A, D Transmitted light images.B, E confocal images displaying nuclei (cyan) and lipid droplets (magenta).C, F Overlay of transmitted light and confo-

Fig. 2
Fig. 2 Size distribution of lipid droplets observed at the growing edge of S. pistillata microcolonies grown on microscope coverslips.A Lipid droplets size distribution in the calicodermis; B lipid droplets

Table 1
Summary of LDs characteristics observed in the calicodermis and in the gastrodermis of 4 independent S. pistillata microcolonies