Introduction

The oceans absorb approximately 30% of the carbon dioxide (CO2) released by anthropogenic activities such as fossil fuel burning and land use change (Gruber et al. 2019). Absorption of CO2 creates carbonic acid (H2CO3) causing a decrease in the pH of seawater and numerous associated chemical changes collectively referred to as ocean acidification (OA) (Doney et al. 2009). Most notable among these changes are decreased carbonate ion (CO 23 ) concentrations and increased bicarbonate (HCO3−) concentrations, which are consequences of re-equilibration of the carbonate buffer system after CO2 absorption. OA affects marine organisms, populations, and ecosystems in diverse but often subtle ways (Hofmann et al. 2010; Kroeker et al. 2013). Photosynthetic organisms such as algae and seagrasses often benefit from increased CO2 concentrations that either directly or indirectly stimulate photosynthesis (Koch et al. 2013; Mackey et al. 2015). The relative benefit of OA for photosynthetic performance accrues differentially to community members leading to shifts in the relative species abundance (Tortell et al. 2008).

While OA may benefit certain photoautotrophs, it is generally detrimental to the many marine organisms that make calcium carbonate skeletons or shells. OA reduces the rate of calcification in many corals, mollusks, and other marine calcifiers (Riebesell et al. 2000; Ries et al. 2009; Kroeker et al. 2010). However, the mechanism by which OA inhibits calcification has not been definitively identified and likely varies between taxa. OA causes a reduction in carbonate concentrations, the form of inorganic carbon ultimately required to make calcium carbonate minerals, but many calcifiers can make use of bicarbonate as well as carbonate by carefully controlling fluid chemistry in the spaces where calcium carbonate is deposited (Tambutte et al. 2011; Drake et al. 2020). This ability to use more than one inorganic carbon source implies that the physiological consequences of OA that lead to reduced calcification rates are more complex than simple lack of substrate (carbonate) availability (Comeau et al. 2013). This complexity is most evident in the few unusual species that calcify more rapidly with OA, perhaps taking advantage of the increased bicarbonate concentration (Ries et al. 2009).

Scleractinian corals have received substantial attention regarding OA given their ecological importance as reef builders and the current pressures from thermal bleaching, diseases, overfishing, and eutrophication (Edmunds et al. 2016; Zaneveld et al. 2016; Hughes et al. 2018; Muller et al. 2020). Calcification in corals is a highly regulated process, and the mechanism by which OA impacts corals is not yet fully understood. Corals deposit calcium carbonate from the calcifying fluid, a thin layer of fluid generated between the coral’s innermost tissue layer and existing skeleton (Cohen and McConnaughey 2003; Tambutte et al. 2011). Active proton export from the calcifying fluid increases its pH (typically pH 8–9) converting inorganic carbon into carbonate, thereby raising the calcium carbonate saturation state to maintain calcification (Sevilgen et al. 2019). Inorganic carbon is delivered to the calcifying fluid to support calcification, and this inorganic carbon is ultimately derived from a mix of respiratory CO2 and inorganic carbon imported from seawater (Furla et al. 2000). Delivery of inorganic carbon to the calcifying fluid is facilitated by the enzyme carbonic anhydrase, which accelerates the interconversion of CO2 and bicarbonate (Bertucci et al. 2013). Carbonic anhydrase is involved in CO2 uptake by corals (Tansik et al. 2015), routing of inorganic carbon through coral tissues layers (Bertucci et al. 2013), and calcification (Moya et al. 2008).

Given the complexity of calcification, it has been challenging to identify the critical processes negatively affected by OA. One key piece of evidence is that OA reduces the pH of the calcifying fluid in scleractinian corals, which lowers the aragonite saturation state and slows calcification rates in some species (e.g., Stylophora pistillata, Acropora yongei), while others are more resilient (e.g., Pocillopora damicornis) (Venn et al. 2013; Comeau et al. 2018). How OA decreases calcifying fluid pH is an open question, but one hypothesis is that increased CO2 concentrations in the water drive increased CO2 fluxes to the calcifying fluid, acidifying the calcifying fluid in a manner analogous to that by which absorption of anthropogenic CO2 acidifies seawater (Cai et al. 2016).

The relative impact of OA on calcification of scleractinian corals seems to be modulated by species, growth characteristics, and environmental factors (Chan and Connolly 2013; Comeau et al. 2014). The coral host could theoretically benefit from the increased photosynthetic potential that elevated seawater CO2 concentrations offer to their symbiotic algae. Generation of additional photosynthetic energy could be used to compensate for the otherwise detrimental effect of OA on calcification. Though there have been sporadic reports of stimulation of photosynthesis by OA (Langdon and Atkinson 2005), the symbiotic algae of most coral species are impacted neither positively nor negatively (Anthony et al. 2008; Hoadley et al. 2015).

Although tropical coral growth is commonly slowed by near future (~ year 2100) levels of OA, many subtropical and temperate corals, including Oculina arbuscula, are more resilient to this stress (Ries et al. 2010; Rodolfo-Metalpa et al. 2010; Varnerin et al. 2020). Oculina arbuscula is a common branching scleractinian coral found in subtropical to temperate hard bottom habitats along the southeastern coast of the USA (Gleason et al. 2018). It is facultatively symbiotic with dinoflagellates from the family Symbiodiniaceae, most commonly associating with Brevolium psygmophilum (Thornhill et al. 2008; Leydet and Hellberg 2016; LaJeunesse et al. 2018), with colonies from well-lit environments generally hosting abundant symbionts and colonies from darker habitats having fewer to no symbionts and instead relying on heterotrophy. Unlike tropical corals, O. arbuscula in shallow water habitats (< 30 m depth) of the southeastern USA are often exposed to highly seasonal environments where temperatures and CO2 concentrations vary greatly throughout the year (Stachowicz and Hay 1999; Xue et al. 2016). Perhaps in part due to the natural variability in CO2, calcification rates of adults and juveniles of this species are resilient to OA (Ries et al. 2010; Varnerin et al. 2020). Juveniles were only modestly affected by CO2 concentrations similar to predicted year 2100 concentrations (1300 ppm; Varnerin et al. 2020), conditions that substantially inhibit calcification in many tropical corals (Comeau et al. 2014; Langdon et al. 2018). Adults were similarly resilient until CO2 concentrations reached ~ 3000 ppm when calcification was dramatically reduced (Ries et al. 2010).

While it is clear that growth and calcification of O. arbuscula are relatively unaffected by OA, the mechanisms behind this resilience, including any compensating physiological adjustments, have not been examined. Here, we sought to assess the effects of OA on components of the inorganic carbon delivery system, carbonic anhydrases, and photosynthetic physiology. We hypothesized that adjustments to the inorganic carbon delivery processes induced by OA could mitigate its detrimental effects or allow reallocation of resources to other processes that could maintain calcification, such as proton export from the calcifying fluid (McCulloch et al. 2012). As an alternate potential response, we hypothesized that the photosynthetic system may acclimate to take advantage of increased CO2 availability, enhancing photosynthetic energy generation that could be used to mitigate detrimental effects of OA on the calcification process.

Materials and methods

Coral collection and experimental design

A total of 22, 6–10-cm-long fragments of O. arbuscula were collected on May 23, 2018, at approx. 20 m depth at J–Y Reef (31° 36.056, 80° 47.431), a relict scallop shell reef located off the coast of Georgia, USA. At the time of collection, daily average water temperature was 24.6 ± 0.3 C and seawater pCO2 was 524 ± 6 ppm at the nearby Gray’s Reef Buoy (https://www.ndbc.noaa.gov/station_page.php?station=41008). Care was taken to sample from coral colonies that were consistently brown in color with no visible signs of bleaching and that were separated by ≥ 1 m so that they were more likely to be genetically distinct. Coral fragments were removed from larger colonies using a chisel and rock hammer. Oculina arbuscula fragments were transported in insulated coolers inside temperature-controlled vehicles to the laboratory at Georgia Southern University (Statesboro, GA) in seawater collected from the reef. The coral fragments were left in the cooler with the lid open and with aeration for the next 24 h to allow the sea water to cool down to the room temperature of 23 °C. Following their initial 24-h temperature acclimation, each colony fragment was mounted to a 4.7 × 4.7 cm ceramic tile using underwater epoxy (Z-Spar A-788 Splash Zone Compound). Fragments and their attached ceramic tiles were held at 23 °C in 38-L aquaria containing artificial seawater (Instant Ocean) until they were transported to the University of Georgia (Athens, GA) 9 days later where OA experiments were conducted. At the University of Georgia, corals were again allowed to acclimate to new tank conditions (detailed below) for 1 week prior to beginning CO2 manipulation.

The OA experiment consisted of two treatments, ambient and elevated CO2, with three replicate tanks per treatment. To allocate the 22 coral fragments among treatments, the corals were paired by mass (30–95 g) and one member of each pair was assigned to each treatment to ensure the size of corals was similar between treatments. The corals were then randomly assigned to tanks within treatments. Tanks (38 L) were filled with artificial seawater (Red Sea ‘Red Sea Salt Mix’) made up in deionized water at a salinity of 35 psu. Prior to addition to the tanks, elevated alkalinity concentrations in the artificial seawater were reduced to 2.4 mM using HCl. Temperature was maintained between 25 and 27 °C. Aquarium water was continuously filtered through physical and charcoal filters. Ambient CO2 treatments were bubbled with laboratory air (~ 420 ppm CO2) using aquarium air pumps through a fritted air stone to increase gas exchange rates. Elevated CO2 treatments were bubbled with high CO2 air (1000 ppm CO2) through a fritted air stone. High CO2 air was obtained by mixing laboratory air with pure CO2 using mass flow controllers (Aalborg), and the CO2 content was verified using an infrared gas analyzer (LiCor LI-840A). The seawater inorganic carbon system in each tank was monitored routinely by measuring pH (3 times per week) and alkalinity (at minimum weekly). pH was measured on the total hydrogen ion scale using the thymol blue spectrophotometric method (Zhang and Byrne 1996). Alkalinity was measured by an open-cell Gran titration (Dickson and Goyet 1994) using an automated titrator (Apollo SciTech ALK1+) that was calibrated with certified reference material (Reference material for oceanic CO2 measurements, Scripps Institution of Oceanography, Batch 132). CO2 levels in the tanks generally matched target conditions (Table 1, Supplementary Figure 1) with ambient CO2 concentrations approximating the elevated summertime pCO2 conditions at Gray’s Reef (~ 400–500 ppm; Xue et al. 2016) and high CO2 treatments of approximately 840 ppm, consistent with business as usual scenarios for year 2100 (Riebesell et al. 2011; IPCC 2014). Alkalinity in the tanks (2.3–2.36 mM; Table 1; Supplementary Figure 1) was also consistent with spring to summertime concentrations at Gray’s Reef (2.3–2.4 mM; Xue et al. 2016).

Table 1 Tank inorganic carbon chemistry averaged over the course of the experiment
Full size table

The aquariums were illuminated with LED lights at 40–60 μmol photons m−2 s−1 for 12 h per day, similar to conditions in Georgia hard bottom environments (40–155 μmol photons m−2 s−1; Matterson 2012). Corals were fed with freshly hatched Artemia sp. nauplii ad libitum three times per week for 2 h. Feeding was conducted in an isolated tray within the aquarium, and any uneaten nauplii were removed from the aquarium after feeding. Partial water changes (~ 25%) and tank cleanings were conducted once a week.

Coral and algal symbiont biometrics

Coral calcification rates were measured using the change in buoyant weight over the course of the 7-week experiment (Jokiel et al. 1978). At the beginning and end of the experiment, each coral fragment was suspended from a monofilament line attached to its base on the underhook of an analytical balance. The coral fragments were immersed in aquarium water during the measurement, and weights were recorded to the nearest milligram. At the end of the experiment, coral tissue was removed using either a water-pik or airbrush filled with artificial seawater depending on whether carbonic anhydrase activities were measured on the fragment. Airbrushing produces a more concentrated suspension needed for carbonic anhydrase measurements (described below). The resulting coral–algal slurry was homogenized, and algae were removed from the homogenate via centrifugation at 2350 g for 5 min. Fragment surface area was measured in duplicate from the weight of aluminum foil wrapped around the fragment (Marsh 1970). A calibration curve was generated by weighting foil pieces of known size to convert foil weight to coral surface area.

Algal symbiont cell density and pigment content were determined from the centrifuged coral homogenate. The resulting algal pellet was resuspended in artificial seawater, and a fraction was preserved with formalin (1%) for cell density estimation. Algal cells in the preserved fraction were enumerated (three times) using a hemacytometer, and cell densities were normalized to coral surface area. A systematic downward bias in algal cell densities was observed in corals from which tissue had been removed using an airbrush as opposed to a water-pik. The magnitude of this bias was estimated by dividing the average cell densities in all water-piked corals by the average cell densities in all airbrushed corals. This bias factor (2.2 ± 1.3) was used to scale up algal cell densities in airbrushed corals. The source of this bias is suspected to be clumping of algae in the airbrushed coral samples, clumps that were then not included in the hemacytometer cell. We verified that effectively all algae were pelleted in the centrifugation procedure and there were no biases in other parameters. Most notably there was no evident bias in algal pigment concentrations per unit coral surface area in airbrushed versus water-piked samples, which would indicate loss of algae at some point or incomplete tissue removal in the airbrushed samples.

Algal pigment concentrations were determined via Gaussian curve fitting to UV/Vis absorption spectra of pigment extracts. A portion of the coral–algal homogenate obtained via airbrushing or water-piking was centrifuged (2350 g, 5 min), and the algal pellet was stored at − 20 °C for less than 1 week. To extract pigments, a 90% acetone–10% water solution was added to the algal pellet and sonicated with a probe sonicator for 1 min to disrupt algal aggregates. The solution was allowed to extract for 48 h at − 20 °C. The extract was clarified by brief centrifugation, and then the UV/Vis absorption spectrum was measured from 350 to 750 nm. The concentrations of selected pigments were determined by nonlinear least squares fitting of Gaussian components to the absorption spectrum (400–700 nm) using the method described in Thrane et al. (2015). Pigments known to be found in dinoflagellates (chlorophyll a, chlorophyll c2, peridinin, beta-carotene, dinoxanthin, diatoxanthin, and diadinoxanthin) were included in the fitting procedure using reference spectra supplied by Thrane et al. (2015). The dominant pigments were determined to be chlorophyll a, chlorophyll c2, and peridinin whose concentrations are reported.

Carbonic anhydrase activity

Internal and external carbonic anhydrase (iCA and eCA, respectively) activities were measured in corals, and iCA activity was measured in algal symbionts using 18O exchange methods. eCA activity includes any CA exposed to the environment on the outer surface of polyps or coenosarc tissue. Coral iCA activity includes any CA within the tissues, mesoglea, or calcifying fluid, and symbiont iCA activity includes any CA within the algal symbionts regardless of localization (cytoplasm, chloroplast, etc.). Due to time constraints, only a subset of the coral fragments (coral eCA: 8 per treatment, coral iCA: 6 per treatment, symbiont iCA: 6 per treatment) were assayed for CA activity. CA activities were determined based on the rate of removal of an 18O-label from DIC, monitored via membrane inlet mass spectrometry (MIMS). As CO2 is hydrated and bicarbonate-dehydrated, the 18O-label is gradually exchanged for 16O from water; CA accelerates the hydration/dehydration reaction and so accelerates removal of the 18O-label (Silverman 1982).

Coral eCA activity was measured on intact coral fragments as detailed in Tansik et al. (2015). The measurements were conducted in a water-jacketed (26 °C), stirred chamber filled with pH-buffered, artificial seawater (Red Sea Salt Mix, pH = 8.0, 10 mM Bicine buffer) from which water was continuously removed at a rate of ~ 500 μL min−1 using a peristaltic pump. The water was passed through a flow-through MIMS inlet, which consisted of a 25 mm length of gas-permeable silicone tubing under high vacuum, to extract gases including CO2. The extracted gas passed through a dry ice/ethanol water trap (− 75 °C) and into a mass spectrometer (Pfeiffer PrismaPlus QMS220), which continuously monitored isotopologues of CO2. To initiate the experiment, 18O–13C-labeled DIC (100 μM) was added to the chamber and the background rate of 18O exchange was monitored for ~ 20 min. A coral fragment was then added to the chamber, and any acceleration of 18O removal was monitored for ~ 25 min at which point an eCA inhibitor, acetazolamide (50 μM), was added to the chamber and 18O exchange was monitored for a further ~ 25 min. In the absence of the eCA inhibitor, accelerated 18O exchange is due to both eCA and iCA, whereas after the application of the inhibitor accelerated 18O exchange is only due to iCA. The 18O exchange data were analyzed to determine eCA activity by fitting a numerical model to the data as detailed in Tansik et al. (2015). The first-order rate constant for CO2 hydration (units of cm3 s−1), determined from the model–data fit, was normalized by coral surface area to provide an estimate of eCA activity per unit surface area (units of cm s−1).

Coral iCA activity was measured on homogenized coral tissue subtracting the contribution of coral eCA as detailed in Hopkinson et al. (2015). To obtain a concentrated coral homogenate necessary for this assay, coral tissue was removed using an airbrush filled with assay buffer [DIC-free artificial seawater (Price et al. 1988) buffered to pH 8.0 using 20 mM TRIS] to which a protease inhibitor cocktail (Roche cOmplete protease inhibitor, 3 mini tablets per 50 mL) was added. The algal–coral slurry was homogenized and then centrifuged (2350 g, 5 min) to pellet symbiotic algae to produce a purified coral tissue homogenate. The CA activity of this purified homogenate was measured using an 18O removal technique. The mass spectrometer described above was interfaced to a small, water-jacketed (26 °C) chamber that holds ~ 1 mL of solution and has a Teflon membrane on the bottom through which gases can diffuse into the mass spectrometer. This chamber was filled with 800 μL assay buffer, and 2 mM 18O–13C-labeled DIC was added to the chamber. After monitoring background 18O-exchange for ~ 10 min, 200 μL of the purified coral homogenate was added to the chamber and accelerated 18O-exchange was monitored for ~ 10 min at which point a CA inhibitor (50 μM acetazolamide) was added. Coral iCA activity was determined by fitting a numerical model to the 18O-exchange data as detailed in Hopkinson et al. (2015). The first-order rate constant for CO2 hydration (units of cm3 s−1), determined from the model–data fit, was normalized by coral surface area to provide an estimate of iCA activity per unit surface area (units of cm s−1).

Algal symbiont iCA activity was measured on algae pelleted from the airbrushed corals and resuspended in assay buffer following Hopkinson et al. (2011). The procedure for algal iCA activity is similar to that described for the coral iCA except that 200 μL of algal suspension was added to the MIMS chamber instead of coral homogenate and the assay buffer contained 50 μM acetazolamide to inhibit eCA or any residual coral tissue CA activity. Algal symbiont iCA activity was determined by fitting a numerical model to the 18O-exchange data as detailed in Hopkinson et al. (2011). The first-order rate constant for CO2 hydration (units of cm3 s−1), determined from the model–data fit, was normalized by algal cell volume resulting in units of s−1 and providing an estimate of the CA activity, assuming that it is homogenously distributed throughout the cell (Hopkinson et al. 2011). The MATLAB code used to determine coral iCA and eCA and algal iCA is available at: https://github.com/bmhopkinson/18O_processing.

Photosynthetic rates and physiology

Photosynthetic rates as a function of irradiance were determined on coral holobionts (coral + algal symbionts) by measuring O2 production rates in metabolic chambers. Coral fragments were placed on a fritted base in a water-jacketed (26 °C) acrylic chamber containing artificial seawater whose pH was adjusted to match experimental conditions (ambient: 7.96 ± 0.02; high CO2: 7.75 ± 0.02; Table 1). The chamber was stirred using a magnetic stir bar at 500 rpm, and oxygen was continuously monitored in the chamber using a calibrated Firesting oxygen optode. Irradiance was controlled by a LED light source (Cool-White 5000 K Cree) calibrated for the chamber using a Walz PAR meter. Irradiance was increased stepwise (0, 50, 100, 200, 300, 450, 650, and 900 μmol photons m−2 s−1) holding light levels constant for 8 min. P versus E data were fit to a saturating exponential function to extract the maximal gross photosynthetic rate (Pmax), the light saturation parameter (Ek), and the dark respiration rate (R), which were subsequently normalized to specimen surface areas calculated as described above.

To determine the role of eCA in photosynthesis, experiments were conducted in which an eCA inhibitor (acetazolamide) was added to corals while they were photosynthesizing. To conduct these experiments, a coral fragment was placed in the same apparatus used to conduct photosynthesis versus irradiance experiments. Initially, oxygen depletion was monitored in the dark for 10 min to constrain the respiration rate, at which point the LED lights were turned on to 600 μmol photons m−2 s−1 and the photosynthetic rate was monitored from oxygen production. After 10 min, 50 μM acetazolamide was added to the chamber and oxygen production was monitored for a further 10 min. Coral fragments were rinsed with seawater and returned to the tank and allowed to rest for at least 2 days before any other physiological experiments were performed.

Photosystem II (PSII) physiological parameters were determined using chlorophyll fluorescence saturation and relaxation kinetics. Chlorophyll fluorescence kinetics were measured at three locations on each coral fragment using a fiber-optic probe interfaced to a Satlantic FIRe chlorophyll fluorometer. During the single-turnover saturation phase, light (blue LED) was continuously applied for 200 μs, after which relaxation was monitored over 60 intervals with a base interval of 60 μs. During the multi-turnover saturation phase, light was applied continuously for 200 ms, after which relaxation was monitored over 40 intervals with a base interval of 100 μs. The single-turnover phase data were fit to the Kolber et al. (1998) model of chlorophyll fluorescence using a custom MATLAB script to calculate the initial fluorescence (Fo), the maximal fluorescence during the single-turnover phase (Fm), and the turnover time of QA (τQA). Variable fluorescence (Fv) was calculated as the difference between Fm and Fo. A single exponential decay was fit to the multi-turnover relaxation phase data to determine the turnover time of the plastoquinone pool (τPQ).

Timing of measurements

After collection and acclimation, the experiment was deemed to have officially started when CO2 conditions were manipulated, which was coincident with initial measurement of buoyant weights. After 7 full weeks of exposure to manipulated CO2 conditions, measurements began with a final buoyant weight, then proceeded to nondestructive physiological measurements, and finally ended with destructive measurements over the course of 2 weeks. Of the nondestructive measurements, chlorophyll fluorescence kinetics were measured first in a single day, since this measurement is quick, and then followed by measurements of photosynthesis versus irradiance, coral eCA activity, and the effect of acetazolamide on photosynthesis. Other than chlorophyll fluorescence kinetics, these measurements were not conducted in a consistent order to facilitate throughput. However, care was taken to ensure coral fragments were allowed to rest for 2 days after exposure to acetazolamide. Finally, samples for destructive measurements were taken on each fragment simultaneously. These destructive measurements were: algal pigment content, algal symbiont density, and when taken coral iCA activity and algal symbiont iCA activity.

Statistical analysis

For each variable, a nested ANOVA was conducted to assess the effect of the treatment (CO2), accounting for any potential effect of individual tanks, which were nested within treatment. Because the experimental design was unbalanced (the number of corals in each tank varied), the analysis was conducted as a linear mixed effects model in R using the lme function from the nlme package. CO2 treatment was specified as a fixed effect, and tank was specified as a random effect. The significance of the treatment effect for each variable is reported. The code and data are included in the Supplemental Materials.

Results

The coral fragments appeared to maintain good health over the course of the experiment as evidenced by extension of polyps in the day, active feeding on Artemia sp., and lack of mortality or tissue loss. One coral in a high CO2 treatment tank exhibited partial bleaching that began roughly half-way through the experiment and continued for the remainder of the experiment. This coral was removed from subsequent analyses. Over the course of the 7-week experiment, corals in both the ambient and high CO2 treatments increased skeletal weight by ~ 3% at an average rate of 0.52 ± 0.10 mg cm−2 d−1 and there was no significant difference in growth between the treatments (Fig. 1; Table 2). Algal symbiont density was approximately 2 × 106 cells cm−2 and was not significantly different between treatments (Fig. 2a; Table 2). It should be noted that the algal symbiont densities from coral fragments whose tissue was removed via airbrushing appeared to be underestimates, likely due to clumping, and consequently were corrected as detailed in the Materials and methods section. The dominant algal pigments detected were chlorophyll a and peridinin, present at ~ 1.25 pg cell−1, and chlorophyll c2 present at ~ 0.4 pg cell−1 (Fig. 2b). Lower concentrations of beta-carotene and dinoxanthin were also detected. Average per cell pigment contents were not significantly affected by CO2 treatment (Fig. 2b; Table 2).

Fig. 1
figure 1

Skeletal growth at ambient (490 ppm) and high (840 ppm) CO2 treatments. Skeletal growth rates were calculated from differences in buoyant weight from the beginning and end of the experiment. The lower and upper bounds of the boxes indicate the 25th and 75th percentiles of the data, and the extent of the whiskers indicates the 10th and 90th percentiles. The circles overlying the box plot are the individual data points (n per treatment = 9–11)

Full size image
Table 2 Results of nested one-way ANOVAs to assess differences in coral and algal biometrics between ambient and high CO2 treatments (n per treatment = 9–11)
Full size table
Fig. 2
figure 2

a Algal symbiont cell density and b algal pigment content for O. arbuscula in ambient (490 ppm) and high (840 ppm) CO2 treatments. Algal symbiont densities were determined by direct microscopic count at the end of the experiment. Pigment content was determined by deconvolution of UV/Vis spectra of organic extracts obtained at the end of the experiment. The lower and upper bounds of the boxes indicate the 25th and 75th percentiles of the data, and the extent of the whiskers indicates the 10th and 90th percentiles. The circles overlying the box plot are the individual data points (n per treatment = 9–11)

Full size image

The enzyme CA is involved in inorganic carbon processing for photosynthesis and calcification and is consequently one of the points at which OA may affect corals. Coral eCA and iCA activities were assessed as was algal iCA activity. Coral eCA activity was ~ 1 cm s−1 in most colonies and was not significantly different between CO2 treatments (Fig. 3c; Table 3). The 18O-exchange model used to extract eCA activities generally fit the data well, indicating that the rates were accurately determined, and there was a clear effect of the eCA inhibitor on 18O-exchange demonstrating that eCA was present (Fig. 3a). Coral iCA activity was 10–15 cm s−1 and was not significantly different between treatments (Fig. 3d; Table 3). The model used to analyze the data fit well in most cases (Fig. 3b), but in one case biphasic depletion of 18O was observed. Biphasic depletion behavior is indicative of intact cells that contain CA in the homogenate, possibly due to incomplete breakage of coral cells or incomplete removal of algal symbionts. This sample was removed from the analysis. Algal symbiont iCA activities were very high, between 1000 and 2500 s−1 (Fig. 3e), compared with activities of 20–500 s−1 in diatoms (Hopkinson et al. 2011, 2013). There was no significant difference in algal symbiont iCA activity between treatments (Table 3).

Fig. 3
figure 3

Carbonic anhydrase activity for O. arbuscula in ambient (490 ppm) and high (840 ppm) CO2 treatments. a Sample results from a coral eCA assay illustrating acceleration of 18O removal after addition of coral fragment (t = 0 s, gray line) and partial slowing of 18O removal upon addition of the eCA inhibitor AZ (t = 1800 s, gray line). b Sample results of a coral iCA assay showing acceleration of 18O removal after addition of coral tissue homogenate (t = 0 s, gray line). In a and b, the colored open circles are the isotopologue measurements (blue = 13C16O16O, green = 13C18O16O, red = 13C18O18O) and the solid lines are the model fit to the isotopologue data used to determine CA activities. c Coral eCA (n per treatment = 5–6), d coral iCA (n per treatment = 8) activities, e algal symbiont iCA activities (n per treatment = 5–6). In c, d, and e, the lower and upper bounds of the boxes indicate the 25th and 75th percentiles of the data and the extent of the whiskers indicates the 10th and 90th percentiles. The circles overlying the box plot are the individual data points

Full size image
Table 3 Results of nested one-way ANOVAs to assess differences in carbonic anhydrase activities between ambient and high CO2 treatments [n per treatment = 5–6 (coral iCA, algal iCA), 8 (coral eCA)]
Full size table

OA can alter photosynthesis, and so photosynthetic physiology was characterized through photosynthesis versus irradiance experiments, chlorophyll fluorescence kinetics, and the effect of eCA inhibition on photosynthesis. Photosynthesis versus irradiance experiments showed that maximal photosynthetic rates were ~ 4.5 × 10−4 μmol cm−2 s−1, respiration rates averaged ~ 1 × 10−4 μmol cm−2 s−1, and Ek was ~ 225 μmol photons m−2 s−1 (Fig. 4). None of the photosynthesis versus irradiance parameters were significantly different between CO2 treatments (Table 4). PSII characteristics, as assessed using chlorophyll fluorescence kinetics, were not affected by CO2 treatments. Fv/Fm, a proxy for the maximal photosynthetic efficiency of PSII, was ~ 0.4 in both treatments (Fig. 5a). The turnover time of the first electron acceptor in PSII (τQA) was ~ 1500 μs (Fig. 5b), and the turnover time of the plastoquinone pool (τPQ) was ~ 20 ms (Fig. 5c). Application of an inhibitor of eCA (acetazolamide) inhibited photosynthesis in nearly all coral fragments reducing photosynthetic rates by 39 ± 16% and by 25 ± 12% at high CO2, but the differences are not statistically significant (Table 4).

Fig. 4
figure 4

Photosynthetic parameters determined from photosynthesis versus irradiance experiments on O. arbuscula fragments in ambient (490 ppm) and high (840 ppm) CO2 treatments. a The maximal photosynthetic rate at saturating irradiance (Pmax) and the dark respiration rate (R), and b plots the light saturation parameter (Ek). The lower and upper bounds of the boxes indicate the 25th and 75th percentiles of the data, and the extent of the whiskers indicates the 10th and 90th percentiles. The circles overlying the box plot are the individual data points (n per treatment = 8)

Full size image
Table 4 Results of nested one-way ANOVAs to assess differences in photophysiological parameters between ambient and high CO2 treatments
Full size table
Fig. 5
figure 5

Photosynthetic characteristics determined from chlorophyll fluorescence kinetics for O. arbuscula in ambient (490 ppm) and high (840 ppm) CO2 treatments. The characteristics measured were a the maximal photosynthetic efficiency of PSII (Fv/Fm), b the turnover time of QA (τQa), and c the turnover time of the plastoquinone pool (τPQ). The lower and upper bounds of the boxes indicate the 25th and 75th percentiles of the data, and the extent of the whiskers indicates the 10th and 90th percentiles. The circles overlying the box plot are the individual data points (n per treatment = 10–11)

Full size image

Discussion

This study sought to examine the physiological mechanisms that enable resilience to OA in the temperate coral O. arbuscula. Elevated CO2 concentrations had no statistically significant effects on photosynthetic physiology or carbonic anhydrase activity, a key component of inorganic carbon-processing machinery. These findings suggest that other characteristics such as the ability to maintain calcifying fluid pH or the ability to acquire energy from heterotrophy may be more important for resilience to OA.

The temperate scleractinian coral O. arbuscula is an important habitat structuring species in hard bottom communities on the continental shelf of the southeastern USA (Gleason et al. 2018). These environments are characterized by dramatic seasonal variation in temperature, light, and CO2 (Xue et al. 2016) requiring broad tolerance in resident organisms, especially sessile species such as O. arbuscula. Oculina arbuscula is capable of surviving a wide range of temperatures (10–30 °C), light intensities (1–100% of surface irradiance), and CO2 conditions (> 2500 ppm) (Miller 1995; Thornhill et al. 2008; Ries et al. 2010). Although it is clear that this species is capable of tolerating widely varying environmental conditions, the underlying physiological, cellular, and genetic bases of environmental tolerance are generally unknown. The ability of scleractinian corals to withstand high CO2 concentrations has received increasing focus recently with the recognition that OA depresses calcification rates in diverse calcifying taxa including scleractinian corals, bivalves, and coccolithophores (Ries et al. 2009; Kroeker et al. 2010). In this study, physiological traits that may enable O. arbuscula’s demonstrated resilience to OA (Ries et al. 2010, Varnerin et al. 2020) were examined. In particular, this study focused on CA activity, which is involved in inorganic carbon supply for calcification and photosynthesis, and photosynthetic performance. Inorganic carbon supply and photosynthesis may be enhanced by increased CO2 and potentially offset otherwise detrimental effects of OA on calcification.

In our experiments, raising CO2 concentrations from ~ 500 to ~ 850 ppm had minimal effects on O. arbuscula at either the organismal or physiological level. Elevated CO2 had no effect on organismal-level metrics including skeletal growth (calcification) or algal symbiont densities. The lack of effect of CO2 on O. arbuscula growth is consistent with several previous studies (Ries et al. 2010; Varnerin et al. 2020), and the principle aim of this work was to assess potential mechanisms underlying this resilience to high CO2. One hypothesis was that high CO2 could stimulate photosynthetic output by O. arbuscula’s symbiotic algae providing additional resources that could be used to mitigate the detrimental effect of OA on calcification. However, there was no evidence for this hypothesis as elevated CO2 did not affect photosynthesis versus irradiance characteristics (Pmax, Ek, R), PSII efficiency (Fv/Fm), photosynthetic electron transport properties (τQA, τPQ), or algal symbiont densities. Consequently, there is no evidence that increased energetic supplies from symbiotic autotrophs enabled resilience of the coral host.

In contrast to tropical corals, temperate corals likely obtain a higher proportion of their total energy budget from heterotrophy and so photosynthetic carbon supplies may be less critical for enabling OA resilience. Oculina arbuscula was fed generously three times a week in these experiments, and light levels were low (~ 50 μmol photons m−2 s−1) to match in situ conditions (Matterson 2012). Although we did not construct an energy budget, these considerations suggest that heterotrophy may have been a significant energy source. In tropical corals, heterotrophy helps corals withstand stressors such as OA (Edmunds 2011; Towle et al. 2015) and bleaching (Grottoli et al. 2006). Heterotrophy has been shown to mitigate the effects of temperature stress on O. arbuscula specifically (Aichelman et al. 2016) and could reasonably be expected to help this species withstand OA as well. Increased heterotrophy is linked to greater energy availability and reserves through lipid accumulation under both temperature and CO2 stress (Grottoli et al. 2004, 2006; Towle et al. 2015).

A second possible source of resilience to high CO2 lies in the inorganic carbon-processing system that delivers inorganic carbon for calcification and photosynthesis. Corals import CO2 and bicarbonate from the external environment and recycle-respired CO2 to support photosynthesis and calcification (Tambutte et al. 2011; Tansik et al. 2015). These inorganic carbon sources are transported through tissues layers by both active and diffusive mechanisms, the details of which have yet to be fully resolved. CAs serve critical roles in carbon routing by interconverting CO2 and bicarbonate (Moya et al. 2008; Bertucci et al. 2011). To gain insight into potential shifts in inorganic processing at high CO2, O. arbuscula’s eCA and iCA activities were measured as a proxy for inorganic processing capacity (Bertucci et al. 2013; Hopkinson et al. 2015). However, no significant differences in either of these parameters were detected, suggesting that no major rearrangements in carbon processing occurred, with the caveat that these measurements represent bulk CA activities and do not resolve the activities of specific isoforms with differing functional roles. Maintenance of high CA activity at elevated CO2 could be a sign of OA resilience since CA gene expression is often, though not always, down-regulated at high CO2 in tropical corals suggesting a reduced capacity for carbon processing when subject to OA (Moya et al. 2012; Kaniewska et al. 2015). Previous physiological and transcriptomic work on tropical corals has indicated that the ability to maintain a high pH in the calcifying fluid at high CO2 confers resilience to OA (McCulloch et al. 2012; Venn et al. 2013; Davies et al. 2016; Strader et al. 2020), and there is some evidence for this in temperate corals as well (Trotter et al. 2011). Characteristics of the calcifying fluid were not assessed in this study, and so it may be that maintaining a high calcifying fluid pH (McCulloch et al. 2012), likely through enhanced proton export, or additional calcium import (DeCarlo et al. 2018), is critical to OA resilience in temperate corals as well.

There was substantial variance in many traits (algal symbiont density, CA activity) within treatments that limited our ability to detect any potentially moderate effects of OA on coral physiology. This individual variation may have both environmental and genetic components. Though the specimens had been in aquaria with controlled environmental conditions for approximately 10 weeks at the time of harvest, acclimation in corals to new conditions can take months (Gates and Edmunds 1999; Mayfield et al. 2018). Consequently, some of the individual variation could reflect residual effects of differences in environmental conditions in which parent colonies grew. Genetic differences can also lead to variability in physiological responses among individuals exposed to similar conditions, as has been well documented in coral response to temperature change (Dixon et al. 2015; Fuller et al. 2020). This genotype by environment interaction can be partially mitigated by ensuring that replicates of a single genotype are distributed among treatments. However, permit restrictions on the size and number of O. arbuscula colonies that could be collected precluded taking coral samples that were large enough to be subdivided and dispersed in both OA treatments.

While no major changes in CA activity or photosynthetic properties were detected in O. arbuscula, another possibility is that O. arbuscula’s physiology is exceptional in some way, even at ambient CO2, preventing harmful effects of OA. Comparing O. arbuscula’s photosynthetic physiology with those of tropical corals, as compiled in Sawall and Hochberg (2018), revealed very similar characteristics between the two groups. Maximal photosynthetic rates of O. arbuscula (~ 4.5 × 10−4 μmol O2 cm−2 s−1) are nearly identical to average tropical rates (4.6 ± 2.3 × 10−4 μmol O2 cm−2 s−1, average ± standard deviation). The light saturation parameter (Ek) varies more widely in tropical corals since it is highly dependent on environmental light conditions, but O. arbuscula’s Ek under our experimental conditions (229 ± 62 μmol photons m−2 s−1) is well within the typical range of tropical corals (186 ± 100 μmol photons m−2 s−1, average ± standard deviation). This comparison suggests that O. arbuscula’s photosynthetic potential is similar to that of tropical corals, which are able to support most of their energetic needs through photosynthesis (Muscatine et al. 1981; Edmunds and Davies 1986). The combination of substantial photosynthetic and heterotrophic energy sources in this experiment may have provided sufficient energy to counteract negative effects of OA, as noted above. Comparisons of O. arbuscula’s CA activities with the few previous quantitative CA activity measurements also indicated that these traits are similar in tropical coral and O. arbuscula. eCA activities in three tropical corals, Orbicella faveolata, Porites astreoides, and Siderastrea radians, ranged between maximum values of ~ 6 cm s−1 in O. faveolata down to ~ 0.1 in P. astreoides (Tansik et al. 2015), and O. arbuscula averaged ~ 1 cm s−1. iCA activities in the same three tropical corals were in the range of 4–12 cm s−1 (Hopkinson et al. 2015), and O. arbuscula averaged ~ 14 cm s−1, on the high end but not dramatically different from tropical corals. Notably, all of these tropical corals have impaired growth under moderate OA conditions (Okazaki et al. 2013; Langdon et al. 2018; Bove et al. 2019), suggesting CA activity, as a proxy for inorganic processing capacity, is not a good indicator of OA resilience.

Like many other temperate (Rodolfo-Metalpa et al. 2010) and slow-growing tropical corals (Comeau et al. 2014), O. arbuscula’s growth and calcification rates are resilient to OA, both in adult (Ries et al. 2010) and in juvenile (Varnerin et al. 2020) life stages. Here, the potential roles of changes in photosynthetic physiology and inorganic carbon processing in conferring resilience were assessed. There was no evidence that increased photosynthetic production at higher CO2 provided additional energy to counteract OA, perhaps because the coral host regulates CO2 supply to their algal symbionts (Tansik et al. 2017). Similarly, carbonic anhydrase activity of the coral was unaltered by elevated CO2 concentrations though this does suggest that inorganic carbon flows to photosynthesis and calcification are maintained at high CO2 in contrast to several tropical corals that down-regulate CA under OA conditions (Moya et al. 2012; Kaniewska et al. 2015). Explanations for why temperate and slow-growing corals are resilient to OA may lie in how the chemistry of the calcifying fluid is regulated (McCulloch et al. 2012; Strader et al. 2020), increased reliance on heterotrophy, or simply reflect the fact that slower growing corals invest less energy in skeletal formation (Comeau et al. 2014).