Calcification responses of symbiotic and aposymbiotic corals to near-future levels of ocean acidification

Increasing the acidity of ocean waters will directly threaten calcifying marine organisms such as reef-building scleractinian corals, and the myriad of species that rely on corals for protection and sustenance. Ocean pH has already decreased by around 0.1 pH units since the beginning of the industrial revolution, and is expected to decrease by another 0.2–0.4 pH units by 2100. This study mimicked the preindustrial, present, and near-future levels of pCO2 using a precise control system ( ± 5 % pCO2), to assess the impact of ocean acidification on the calcification of recently settled primary polyps ofAcropora digitifera, both with and without symbionts, and adult fragments with symbionts. The increase in pCO2 of ∼ 100 μatm between the pre-industrial period and the present had more effect on the calcification rate of adultA. digitiferathan the anticipated future increases of several hundreds of micro-atmospheres of pCO2. The primary polyps with symbionts showed higher calcification rates than primary polyps without symbionts, suggesting that: (i) primary polyps housing symbionts are more tolerant to nearfuture ocean acidification than organisms without symbionts, and (ii) corals acquiring symbionts from the environment (i.e., broadcasting species) will be more vulnerable to ocean acidification than corals that maternally acquire symbionts.


Introduction
As humans continue to burn fossil fuels at an unprecedented rate, the concentration of CO 2 in the atmosphere is presently higher than it has been for the last 420 000 yr (Hoegh-Guldberg et al., 2007;IPCC, 2007).The oceans uptake a large proportion of that CO 2 , forcing them toward more acidic conditions (i.e., with high pCO 2 ), threatening the very foundation of calcifying marine organisms and coral reefs (Kleypas et al., 2006;Orr et al., 2005;Raven et al., 2005).Indeed, coral reefs support a wealth of calcifying organisms, of which scleractinian corals have been the most essential reef builder since the Triassic (Stanley and Fautin, 2001).
Since the pre-industrial period, we have witnessed a steady increase in pCO 2 concentrations from <300 µatm to present concentrations of 400 µatm.pCO 2 is predicted to reach 200 to 700 µatm above present levels by 2100.(IPCC, 2007).Such an increase in pCO 2 reduces both the pH and the concentration of carbonate ions in the water column, and increases the availability of bicarbonate ions (Kleypas et al., 1999).Several studies have found that coral calcification rates are directly related to the concentration of carbonate ions in the water column (Anthony et al., 2008;Gattuso et al., 1998;Kleypas et al., 2006;Marubini et al., 2008), whereas another study has shown a positive relationship between coral growth rates and the availability of bicarbonate ions (Jury et al., 2010).It has also been suggested that both the carbonate and bicarbonate ions affect coral calcification under acidified seawater conditions, but the extent of the effect differs in light and dark conditions (Comeau et al., 2013).Therefore, the response of coral growth and the state of the ocean's carbonate chemistry is under intensive investigation (Pandolfi et al., 2011).
The oceans are not homogeneous, and the temperature gradient from the tropics to the poles sets carbonate ion concentrations naturally higher in the tropics where coral reefs occur.Nevertheless, the decrease in carbonate ion concentrations from the pre-industrial period to the present has been greater in the tropics (∼ 29 µmol kg −1 ) than in the Southern Ocean (∼ 18 µmol kg −1 ) (Orr et al., 2005).Yet, symbiosis is prolific in the tropics, and the self-extending symbiosis theory tells us that organisms harboring symbionts should be more tolerant to environmental change than organisms without symbionts (i.e., aposymbiotic organisms) (Kitano, 2004;Kitano and Oda, 2006).These assertions lead to two pertinent questions: (i) Will calcifying coral species survive in high pCO 2 seawater?(ii) Are juvenile corals, without symbionts, more vulnerable to high pCO 2 seawater than juveniles and adult corals with symbionts?
Previous experiments that have mimicked the near-future pCO 2 conditions on coral reefs have adjusted the pH of seawater either by adding an acid or a base, or by bubbling CO 2 through the seawater in experimental chambers (Atkinson and Cuet, 2008).Adding an acid or a base results in seawater with different alkalinity, bicarbonate, and carbonate ion concentrations than when CO 2 is bubbled through seawater (Atkinson and Cuet, 2008).Thus, adding an acid or a base has not been used in recent ocean acidification studies.Although bubbling CO 2 through the seawater more closely reflects near-future conditions than adding acids, it is nevertheless difficult to achieve a stable pCO 2 environment, especially in flow-through systems (e.g., Leclercq et al., 2002;Suwa et al. 2010;Takahashi and Kurihara, 2013).To overcome these problems, our research group developed a system that produced stable pCO 2 concentrations in flow-through conditions (Fujita et al., 2011).
Using this system, we examined the effect of pCO 2adjusted seawater on the calcification rates of Acropora digitifera, one of the most common corals in the Pacific Ocean.Calcification was examined in five pCO 2 treatments: (i) pre-industrial pCO 2 , <300 µatm, (ii) present-day pCO 2 at 400 µatm, and at three near-future conditions of (iii) 600 µatm, (iv) 800 µatm, and (v) 1000 µatm.Within these treatments, we investigated the response of: (1) primary aposymbiotic coral polyps (i.e., without symbionts), (2) primary symbiotic polyps, and (3) adult symbiotic fragments.It was hypothesized that the calcification process of symbiotic corals was more tolerant to pCO 2 adjustments than aposymbiotic corals.

Experimental setup
To produce pCO 2 -adjusted seawater, we used a precise pCO 2 control system (Fujita et al., 2011).This system was used to generate five different pCO 2 levels, including one lower than the present level of atmospheric pCO 2 : (i) preindustrial, <300 µatm, (ii) present-day pCO 2 at 400 µatm, and at three near-future conditions of (iii) 600 µatm, (iv) 800 µatm, and (v) 1000 µatm.The pCO 2 -adjusted seawater was supplied to duplicate flow-through (150 mL min −1 ) aquaria systems (12 L).The seawater temperature was maintained at 27 • C, with a 12 : 12 h light : dark photoperiod (of 75 µmol m −2 s −1 ) under metal-halide lamps (Funnel2 150W, Kamihata, Japan) throughout all treatments.The aragonite saturation state of the seawater was estimated using the CO2SYS program (Lewis and Wallace, 1998) and the variables: temperature, pH, mean salinity, and total alkalinity were measured repeatedly during the experiments.The chemical and physical conditions of each pCO 2 treatment are summarized in Tables 1 and 2.

Primary polyp experiment
Several 20 cm A. digitifera colonies were collected from a fringing reef of Sesoko Island, Okinawa, Japan.Gametes from two colonies, which spawned on 29 May 2010, were combined in a flow-through aquarium, from which we derived several hundred planulae larvae.Primary polyps were prepared following the methods outlined in our previous report (Suwa et al., 2010) using 13 day-old planulae.To prepare the symbiotic primary polyps, primary polyps of A. digitifera were infected with the dinoflagellate Symbiodinium (clade A, Tanaka et al., 2013).These dinoflagellates were derived from the giant clam Tridacna crocea (a solution of 4 × 10 5 cells mL −1 ) and were used because the primary polyps could acquire algae from this bivalve more efficiently than from other hosts, including Acropora species (Hirose et al., 2008).Four days after inducing metamorphosis, primary polyps were exposed to the symbiont solution for one day.Three days after exposure to the symbiont solution, we confirmed symbiont infection using a dissecting microscope.On the final day of the experiment, many symbionts (which were identical to the symbionts in Tanaka et al., 2013) were observed in infected polyps.The primary polyps, both with and without symbionts, were subjected to four pCO 2 treatments: (i) pre-industrial, <300 µatm, (ii) present-day pCO 2 at 400 µatm, (iii) 800 µatm, and (iv) 1000 µatm.
Eight 6-well culture plates, containing the settled primary polyps, were placed into each of 8 aquariums (i.e., 4 plates for aposymbiotic primary polyps, and 4 plates for symbiotic primary polyps) for 10 days.A total of 20 polyps per treatment were used to evaluate skeletal growth of polyps.At the end of the experiment, soft tissues were removed from each polyp with a jet of high-pressure water from a dental-hygiene tool.The dry weight of each polyp skeleton was measured according to Inoue et al. (2011).The dry weight (µg) of the polyp skeleton, at the end of the experiment, was used to represent the amount of growth of each coral during the experiment.

Adult-coral-fragment experiment
Five >30 cm colonies of A. digitifera were collected in August 2009 from a shallow (2 m) fringing reef at Sesoko Island, Okinawa, Japan.The colonies, which were growing at least 10 m apart, were haphazardly selected.The A. digitifera colonies were kept in a flow-through aquarium for 3 weeks under natural light conditions at Sesoko Station, Tropical Biosphere Research Center, University of the Ryukyus (Okinawa, Japan).Fifty 2-3 cm fragments were cut from each parent colony and attached to plastic bolts with superglue.The fragments were kept in a flow-through aquarium for 2 weeks under natural light conditions until the coral tissues started to spread over the surfaces of the plastic bolts.Five of these fragments, from each parent colony, were maintained for 6 weeks in each of 10 aquaria, to which pCO 2 -adjusted seawater was supplied using the flow-through system (two aquaria per pCO 2 treatment).
The weight of each colony was measured as buoyant weight (Davies, 1989), which directly reflects skeletal weight (Anthony et al., 2008).The calcification rate was calculated as the percentage change in final weight relative to the initial weight, during the 6-week experiment (see also Fig. S1).During the adult fragment experiment, 29 fragments died (11.6 % of 250 fragments) and were excluded from the calcification analysis.To evaluate the photosynthetic fitness of zooxanthellae in the adult fragments, the symbionts' maximum photosynthetic quantum yields (F v /F m ) were measured after 6 weeks using a Diving-PAM Underwater Fluorometer (Walz, Germany) after at least 1 h of darkness.

Data analysis
In the primary polyp experiment, the dry weights of the primary polyp skeleton were analyzed using a two-factor crossed analysis of variance (ANOVA), in which pCO 2 (with four levels) and symbiosis (i.e., presence or absence of dinoflagellates) were incorporated into the model as fixedeffect factors.The subsequent pairwise comparisons among different pCO 2 levels were performed using Tukey's HSD (honestly significant difference) tests (α = 0.01).
In the adult-coral-fragment experiment, we used a generalized linear model to estimate the response of adultcoral calcification to pCO 2 (fixed-effect factor), aquarium (nested within pCO 2 ; fixed-effect factor), colony (fixed- To remove the variation of covariates, we calculated the adjusted mean final weights relative to the mean initial weight for each colony, assuming that their regression lines were heterogeneous among all the combinations of colony and pCO 2 .The adjusted final weight for each colony was independently analyzed using an ANOVA model with pCO 2 (fixed-effect factor) and aquarium (nested within pCO 2 ; fixed-effect factor) as the independent fixed factors.Statistically significant factors (α = 0.01) were subjected to pairwise comparisons (Tukey's HSD tests; α = 0.01) to specify significant combinations of treatment levels.The Fv/Fm values of adult fragments were analyzed using a one-way ANOVA model with pCO 2 as fixed-effect factors after an arcsine transformation.The subsequent pairwise comparisons among different pCO 2 levels were performed using Tukey's HSD tests (α = 0.01).

Results
With regard to the primary polyp experiment, the ANOVA results indicated that the pCO 2 × symbiosis interaction was statistically negligible (p > 0.05), and the main factors were all significant (p < 0.0001).The post hoc tests demonstrated that the skeletal weights at <300 and 400 µatm were significantly heavier than those at future-level treatments (i.e., 800 and 1000 µatm), regardless of whether polyps contain dinoflagellates or not (Fig. 1).When compared with the same pCO 2 level, the primary polyps with symbionts became heavier than those without dinoflagellates (Fig. 1).Because gametes from two colonies were added to each aquarium, genetic differences could not be incorporated into the model.However, it is unlikely that this changes our conclusion, because the error variance was small compared with the variance that was due to the main treatment effects in our data (see Table S1).
We evaluated the calcification rates of adult fragments of A. digitifera under five pCO 2 treatments.The ANOVA on the adult fragment weight, adjusted for initial size variation, indicated that a higher pCO 2 led to significantly slower growth rates in four out of the five colonies (Colony b-e; Fig. 2; Table S3).The results of the analysis also suggested that the potential environmental differences between two replicate aquaria were negligible in all five colonies (all p > 0.05).The subsequent Tukey's HSD tests indicated that the mean final weight of adult fragments, reared at <300 µatm, was significantly greater than those at the other pCO 2 conditions in all of the four colonies, showing significant pCO 2 effects (Fig. 2; Table S3).The maximum photosynthetic efficiencies of the adult fragments were above 0.6, and did not differ significantly among pCO 2 treatments (Fig. 2; Table S4).These observed values indicated that there was negligible or no light-induced damage caused by the lighting system used in the experiment.

Discussion
The differences in the skeletal weights between primary polyps with and without symbionts might reflect the difficulty that aposymbiotic corals may have in acquiring energy and resources, including organic matrix molecules, for calcification.Yet why would the primary polyps with symbionts be more responsive to pre-industrial treatments than aposymbiotic primary polyps?The increase in calcification in the pre-industrial pCO 2 treatment only occurred in corals that housed symbionts.Indeed, the adult colonies showed the same response as primary polyps with symbionts, clearly increasing calcification rates in low pCO 2 treatments.Moreover, the calcification rates of symbiotic adult A. digitifera fragments were higher in the pre-industrial seawater pCO 2 treatment than in the present-day pCO 2 treatment.
Higher calcification in the pre-industrial pCO 2 treatment was most likely attributed to a change in skeletal precipitation by the coral host, because there was no evidence of any dynamic photoinhibition (Enriquez et al., 2002) indicated as a decline in maximum photosynthetic quantum yield among the symbionts in the high-pCO 2 treatments (Fig. 2, Table S4).Still, there were no differences in calcification rates between present-day and near-future concentrations (Fig. 2).We note that this lack of difference in calcification between present-day and anticipated future pCO 2 treatments was not apparent for primary polyps (Fig. 1).These differences suggest a number of potential mechanisms that are not mutually exclusive.First, an increase in calcification in low pCO 2 environments was only apparent in the presence of symbionts.Therefore, such phenotypic plasticity in calcification potential was most likely attributed to the presence of the symbionts.Second, the adult colonies did not respond to higher pCO 2 environments, whereas the primary polyps with symbionts did show reduced calcification rates at high pCO 2 .Such results suggest a hierarchical response in tolerance to pCO 2 environments, from adult colonies with symbionts as the most tolerant, to symbiotic primary polyps showing some tolerance, to primary polyps without symbionts being the least tolerant to high pCO 2 treatments.
There is mounting evidence that symbiotic dinoflagellates facilitate calcification within corals through a positive feedback system between the host and the symbionts (Allemand et al., 2004;Muscatine, 1990;Yellowlees et al., 2008), although the detailed mechanisms have been under investigation.The glycerol and oxygen produced by the symbionts facilitate calcification through mitochondrial respiration and ATP (Adenosine triphosphate) production, which could be used for ion transport (Allemand et al., 2004;Colombo-Pallotta et al., 2010).CO 2 uptake by photosynthesis is also thought to stimulate calcification by changing the equilibrium of dissolved inorganic carbon (DIC) in coral tissue, although the mechanisms are unresolved (Allemand et al., 2004).Our results also indicate that the primary polyps with symbionts grew faster than aposymbiotic polyps (Fig. 1).Although the primary polyps with symbionts seemed to be more sensitive to acidified seawater than aposymbiotic polyps (Fig. 1), the faster growth induced by symbiosis could compensate for the decrease of calcification by acidified seawater.The reason why coral/algal symbiosis enhances coral calcification is not only attributed to algal photosynthesis but is also potentially related to the removal of substances inhibiting calcification, such as phosphates (Allemand et al., 2004).
Previous research indicates that acidified seawater increases the concentration of HCO − 3 , possibly followed by the activation of photosynthesis in coral symbionts (Jury et al., 2010;Marubini et al., 2008).In our experiments, however, there was no evidence that acidified seawater activates the photosynthesis of Acropora digitifera.The reason why the acidified seawater, with high pCO 2 concentration (1000 µatm), did not affect adult coral calcification and photosynthetic efficiency is unknown.We suspect that there were obvious advantages from symbiosis.For example, the removal of phosphates would facilitate calcification even in acidified seawater.Irrespective of the cellular mechanism involved, our results clearly showed that corals without symbionts were most vulnerable to pCO 2 increases, whereas corals that housed symbionts were more tolerant.
These results suggest that coral recruitment might be influenced by ocean acidification.Given that globally ∼ 80 % of the scleractinian corals are spawners that acquire symbionts from the "wild" after settlement (Baird et al., 2009), vulnerability of primary polyps to ocean acidification upon first settlement (in particular aposymbiotic polyps) could be at risk of decline in the near future.The same possibility was suggested by other recent studies (Albright et al., 2008;Cohen et al., 2009;Suwa et al., 2010;Albright and Langdon, 2011;Albright, 2011;de Putron et al., 2011;Dufault et al., 2012;Doropoulos et al., 2012;Dufault et al., 2013), although the comparative study between aposymbiotic and symbiotic primary polyps is only in its infancy (Inoue et al., 2012;Tanaka et al., 2013).This inference on recruitment may be particularly evident in the Indian and Pacific oceans, where most corals are spawners that horizontally transfer symbionts (Harrison and Wallace, 1990), acquiring them after settlement.By contrast, newly settled corals may do better in the Caribbean, where most corals are brooders and symbionts are maternally (i.e., vertically) acquired, and the planulae are symbiotic (Harrison and Wallace, 1990).
The degree of selective pressure by ocean acidification on newly settled polyps may therefore depend on how rapidly corals are able to support symbionts.Such selective filtering could lead to relative shifts in coral species abundance, changing reefs from those that primarily support spawners to reefs that primarily support brooders (that maternally acquire symbionts).Similar shifts in species composition have occurred in the Oligocene, when rapidly cooling oceans favored brooding corals over spawning corals in the Caribbean (Edinger and Risk, 1995).
In summary, the increase in pCO 2 of just 100 µatm between the pre-industrial period and the present had more effect on the calcification rate of adult A. digitifera than the anticipated future increases of several hundreds of microatmospheres of pCO 2 .Our results also suggest that ocean acidification has had adverse effects on reef corals since the www.biogeosciences.net/10/6807/2013/Biogeosciences, 10, 6807-6814, 2013 industrial revolution.Ocean acidification, therefore, may not be only a future problem but a direct and present threat to ocean ecosystems (Talmage and Gobler, 2010).However, we also need to consider that the seawater pH and pCO 2 in coral reefs can be variable over diel timescales (Suzuki et al., 1995;Ohde and van Woesik, 1999;Bates et al., 2001;Santos et al., 2011).Kitada et al. (2006) reported a relatively large pCO 2 diurnal variation of 680-290 µatm with seasonal variations in reef water in front of Sesoko Station.Thus, the natural pH and pCO 2 variation in coral reefs should be taken into account to accurately predict the effect of ocean acidification.
In conclusion, this study showed that the apparent sensitivity of primary polyps to near-future ocean acidification was a consequence of not housing symbionts, and those organisms harboring symbionts, at any life-history stage, are more tolerant to ocean acidification than organisms without symbionts.

Table 1 .
Summary of mean physical and chemical conditions in each pCO 2 treatment of the primary polyp experiment.Standard deviation is shown for pH T and pCO 2 .The mean salinity and total alkalinity measured repeatedly during the experiments were 34.6 and 2257 mol kg −1 , respectively.The value of arg was calculated using total alkalinity and pCO 2 .(i) pre-industrialpCO 2 , <300 µatm, (ii) present-day pCO 2 at 400 µatm, and at two near-future conditions of (iii) 800 µatm, and (iv) 1000 µatm.All values are shown as mean ±standard deviation.