Similarities in biomass and energy reserves among coral colonies from contrasting reef environments

Coral reefs are declining worldwide, yet some coral populations are better adapted to withstand reductions in pH and the rising frequency of marine heatwaves. The nearshore reef habitats of Palau, Micronesia are a proxy for a future of warmer, more acidic oceans. Coral populations in these habitats can resist, and recover from, episodes of thermal stress better than offshore conspecifics. To explore the physiological basis of this tolerance, we compared tissue biomass (ash-free dry weight cm−2), energy reserves (i.e., protein, total lipid, carbohydrate content), and several important lipid classes in six coral species living in both offshore and nearshore environments. In contrast to expectations, a trend emerged of many nearshore colonies exhibiting lower biomass and energy reserves than colonies from offshore sites, which may be explained by the increased metabolic demand of living in a warmer, acidic, environment. Despite hosting different dinoflagellate symbiont species and having access to contrasting prey abundances, total lipid and lipid class compositions were similar in colonies from each habitat. Ultimately, while the regulation of colony biomass and energy reserves may be influenced by factors, including the identity of the resident symbiont, kind of food consumed, and host genetic attributes, these independent processes converged to a similar homeostatic set point under different environmental conditions.


Methods
Sample collection and preparation. Coral fragments were collected in March 2017 from six coral species (Coelastrea aspera, Cyphastrea chalcidicum, Favites abdita, Pachyseries rugosa, Porites cylindrica, and Porites rus) at both a nearshore (Ngermid Bay, also known as Nikko Bay, 7° 19.470′ N, 134° 29.634′ E, Fig. 1) and offshore reef (Rebotel Reef, 7° 14.930′ N, 134° 14.149′ E, Fig. 1). Coral colonies (n = [3][4][5][6][7][8][9][10][11][12][13][14] were sampled from each species at a depth of 5-10 m (offshore) or 2-5 m (nearshore), at least 10 m apart (see Table 2 for specific sample sizes). Differences in collection depth were due to the natural distribution of these species at each location and to ensure all colonies were collected from similar light conditions (maximal light 800-1000 μmol quanta m −2 s −1 ). All colonies sampled were similar sizes, representative of typical sizes for each species, and fragments were taken from the top and center of the colonies. Coral colonies showed no visual evidence of stress, and no thermal anomalies or bleaching events were reported prior to sample collection. All coral colonies were transported back to the Palau International Coral Research Center (PICRC) in seawater filled coolers. Coral samples were placed into individual Whirlpaks ® and immediately frozen (− 40 °C) at PICRC and kept frozen while transported to the University of Alabama at Birmingham (UAB) in the United States of America where they were stored at -80 °C until processing. While frozen, coral fragments were cut into ~ 4 cm 2 pieces via a Torque Master Tile Saw (QEP) with a diamond blade. All excess skeleton, boring sponges, and epibionts were removed. To determine coral surface area 35 , each fragment was 3D scanned using a Capture Mini 3D scanner (Geomagic ® Controlx64™ software, 3DSystems). Coral fragments were then lyophilized for 36 h (Labconco Freeze Dry System) and weighed for total dry mass. Coral fragments were individually pulverized (SPEX Sample Prep ball mill) into a fine, homogenized powder, encompassing the coral holobiont (animal host, endosymbiotic dinoflagellate communities, and microbiome), and partitioned for sample analysis (tissue biomass and energy reserves i.e., total lipids, soluble protein, and carbohydrates). All laboratory sample preparation and analysis took place at UAB. Tissue biomass. Dry powdered fragments were weighed (~ 0.5-3.3 g) and combusted in a muffle furnace for 12 h at 500 °C to determine the total organic content. Ash-free dry weight (AFDW) was calculated as the dif- www.nature.com/scientificreports/ ference between dry weight and ash weight following combustion. The proportion of AFDW to total dry weight was used to calculate the total AFDW and represents the total coral tissue biomass of the entire fragment per surface area. Sampling sizes varied for the energy reserves measured for some species according to sample availability, offshore (n = 3-14) and nearshore (n = 6-8). See Table 2 for habitat and species-specific sample details.
Lipid analyses. Total lipids were quantified from ~ 0.6 g of lyophilized coral fragment powder using a modified Folch method 36 . Triplicate, independent lipid extractions were conducted for each sample. Briefly, a solvent system (chloroform, methanol, and 0.88% NaCl at 8:4:3 ratios) was used to extract total lipids, at a 20:1 ratio, by solvent volume to dry sample weight. The lower chloroform phase was placed in pre-weighed, glass tubes and dried under a steady N 2 gas stream. Total lipid concentration was gravimetrically determined and converted to Joules (J) AFDW g −126 , then 100% chloroform was added to achieve a concentration of 10 mg mL −1 .
Carbohydrate and protein analyses. Soluble proteins were quantified from ~ 0.5 g of lyophilized, crushed coral sample. Following the protocol of McLachlan et al. 41 , samples were placed in 15 mL tubes and 1 mL of diluted 1 × solution of radioimmunoprecipitation (RIPA, Sigma-Aldrich) was added to the samples. Samples were placed in the freezer and three freeze-thaw cycles were used to lyse cells and solubilize proteins. Samples were centrifugated for 20 min at 4122g relative centrifugal force (RCF) at 4 °C then 1 mL of the supernatant containing the solubilized protein was removed and placed in a 2 mL tube. A modified Bradford assay was then used to quantify soluble protein using bovine serum albumin (BSA) as a standard. All samples were run in triplicate on a microplate reader (EPOCH 2, Agilent) measuring absorbance at 465 and 595 nm. Soluble proteins were converted to J AFDW g −126 .
Carbohydrates were quantified from ~ 0.3 g of lyophilized, crushed coral sample placed in a 2 mL tube with 1 mL of Milli-Q water and sonicated (Sonicator model CL-18, Fischer Scientific) at 35% for 2 min. Samples were then centrifugated for 10 min at 1000g and 1 mL of the supernatant containing total carbohydrate was removed. A modified DuBois method, as described in Masuko et al. 42 , was used to quantify carbohydrates with glucose used to create a standard curve. All samples were run in triplicate on a microplate reader (EPOCH 2, Agilent) measuring absorbance at 485 nm and 750 nm for total carbohydrate determination, and then converted to J AFDW g −126,42 .
Statistical analyses. Data analyses were performed in R (version 1.2.5033) using the car library. All data were tested for normality and homogeneity of variance using the Shapiro-Wilks and Bartlett's tests. If either test was significant (p < 0.05), data were log or square root transformed to achieve normality and homoscedasticity. Despite transformations, some data remained non-normal and heteroscedastic, so non-parametric analyses were conducted. A two-way ANOVA was used to establish species and community trends for AFDW, soluble protein, WAX, and phospholipids, while the non-parametric Kruskal-Wallis test was used to assess total lipid, carbohydrates, TAG, and ST, using a significance level (α) no greater than 0.05. When appropriate, an ANOVA was followed by Tukey Honest Significant Difference (HSD). Intraspecific, across-site comparisons were assessed using Welch's t-tests or non-parametric Mann-Whitney-Wilcoxon tests. All biomass measurements (AFDW, soluble protein, total lipid, and carbohydrate) and lipid class composition (WAX, TAG, ST, and phospholipids) were used in two separate principal component analyses (PCA) to determine the impact of the aforementioned variables on the distribution of samples within two axes that best describe the data. A permutational multivariate analysis of variance (PERMANOVA) was used to determine the significance of species and site, as well as their interaction on sample distribution within each PCA, and a distance-based redundancy analysis (db-RDA) was used and included all biomass measurements and lipid class composition as predictors of the sample distribution.

Results
Interspecific within site comparisons. Ash-free dry weight (AFDW) of coral species was significantly different within reef site (F 5,68 = 19.10, p = 0.01, Fig. 2a, Table 1), with differences driven by the overall lower biomass of P. rus and P. cylindrica. F. abdita and C. chalcidicum had significantly higher AFDW than P. cylindrica offshore (p < 0.001, for both species, S1). Nearshore C. chalcidicum had significantly higher AFDW than P. rus (p < 0.001, S1). Total carbohydrate content was significantly different across species (p < 0.001, Table 1), with F. abdita influencing these differences, because of higher carbohydrate content offshore (p = 0.012, Fig. 2d, Table 2). Protein was significantly different among coral species as well (p = 0.006, Fig. 2c, Table 1), and likely driven by higher protein in C. aspera and P. rugosa nearshore, while C. chalcidicum and P. rus tended to have higher pro-  Fig. 3d, Table 1) significantly differed by coral species but WAX did not (Fig. 3a, Table 1). AFDW of C. aspera significantly differed from all other coral species (Tukey HSD, p < 0.05, S1).

Physiological traits and lipid class composition.
Overall coral physiological traits were significantly impacted by site (PERMANOVA, F 1,47 = 8.86, p = 0.002), species (PERMANOVA, F 5,47 = 7.47, p < 0.001) and the interaction of the two (PERMANOVA, F 5,47 = 4.88, p < 0.001), with species differences primarily driving sample distribution (Fig. 4, Table 3). Total lipids (db-RDA, p = 0.001) and carbohydrates (db-RDA, p = 0.001) further  Table 1, with post-hoc analyses listed in S1. Multivariate analyses results, comparing the impact of site and species and each variable has on sample distribution in the above PCA, can be found in Tables 3 and 4, respectively. www.nature.com/scientificreports/ contributed to sample distribution and AFDW also was a significant predictor in determining sample distribution (db-RDA, p = 0.001, Table 4). P. cylindrica and P. rus have higher carbohydrate concentrations, thus driving the separation of the Porites species from all other species investigated in this study (Figs. 2d, 4). Lipid class composition was only significantly influenced by species (PERMANOVA, F 5,67 = 2.03, p = 0.05), contributing more to the overall sample distribution (Fig. 5, Table 3). WAX and phospholipids were significant predictors of sample distribution (db-RDA, p = 0.001, Fig. 3a,d), followed by ST (db-RDA, p = 0.006, Fig. 3c), with phospholipids best describing the data distribution (Table 4).

Discussion
Underlying physiological traits, and acclimatization potential, among reef-building corals are influenced by multiple factors working in synergy, including biomass, energy reserves, and tolerance of symbiotic dinoflagellates 16,[18][19][20] . We established baseline population characteristics and compared the biochemical components of coral tissue, known to increase overall holobiont thermal resilience, across contrasting environments 17,[43][44][45] . Our data indicate the relative importance of biomass, energy reserves, and resident symbiotic dinoflagellate identity in maintaining the stability of these mutualisms under contrasting environmental conditions that enable the persistent health and well-being of the holobiont 12,32 . Colony tissue composition and available energy stores affect coral health and survivorship. The unique attributes of different coral species (e.g., polyp size, colony morphology, growth rates, diet, etc.) influence differences in tissue thicknesses and biochemical (i.e., energy reserves) composition 19,28,46 . Species differences in AFDW, soluble proteins, carbohydrates, total lipids, and lipid class composition were similar at both sites (Tables 1, 3), and species with the greatest biomass in one habitat also tended to have the highest biomass in the other habitat   Table 1, with post-hoc analyses listed in S1. Multivariate analyses results, comparing the impact of site and species and each variable has on sample distribution in the above PCA, can be found in Tables 3 and 4, respectively. www.nature.com/scientificreports/ (Fig. 2a). As expected, biomass (AFDW cm −2 ) was greatest among the four non-branching and slower growing coral species Coelastrea aspera, Cyphastrea chalcidicum, Favites abdita, and Pachyseries rugosa 46,47 . The greater biomass of these species likely influences their tolerance to marine heatwaves 28,46 , as they can supplement energetic demand by catabolizing energy reserves during thermal stress and bleaching, when the flow of nutrients from their symbionts is interrupted 17,22,43,48 . Greater coral tissue biomass may also enhance the capacity of a colony to modulate radiant flux to the symbiotic dinoflagellates, thus, providing additional photoprotection 49 . The two representatives from the genus Porites, P. cylindrica and P. rus, clustered separately from the other four species because of their biochemical composition (Fig. 4). This was primarily driven by both species having less total biomass and a greater proportion of carbohydrates than the other coral species (Fig. 2a,d). These high carbohydrate concentrations may be explained by Porites spp. ability to produce large amounts of mucus from dense arrays of mucocytes, comprising the ectoderm, like many other poritids 50 .
Although total biomass (AFDW cm −2 ) is informative for coral physiology and health comparisons, direct analyses of tissue biochemical composition can reveal specific components influencing physiology and help to better understand a colony's metabolism. Energetic reserves are macromolecules (i.e., proteins, lipids and carbohydrates), and they should be converted to energetic equivalents using the enthalpies of combustion to determine energy available to the organism 23,26 . Macromolecule concentrations are dictated by several ecological and physiological processes, which are highly specific to species, geography, and prevailing environmental conditions. Total lipids, and lipid classes, were unexpectedly similar among colonies from offshore and nearshore environments (Table 1, Figs. 2, 3). Lipids are abundant macromolecules in coral tissues making up 10-40% of the total biomass 24,51 . Moreover, they are energy-rich compounds (specific enthalpy of combustion − 39.5 kJ g −1 ), making them important catabolic energy sources 24,26,48,51,52 . Similar to tissue biomass, there were species differences in total lipid (J AFDW g −1 ) in both environments (Table 1). For example, there was a ~ sixfold difference in total lipids among nearshore corals, with F. abdita and P. rugosa having the greatest concentration and C. aspera the lowest concentration (Fig. 2b). Offshore corals' total lipid (J AFDW g −1 ) had less variation with F. abdita, P. rugosa, P. cylindrica, and P. rus having approximately double the amount of total lipids compared to C. aspera and C. chalcidicum (Fig. 2b). The high variation in total lipid (J AFDW g −1 ) among nearshore and offshore species likely reflects their differences in tissue biomass, lipid storage, and catabolic rates based on each species' metabolic demand. While lipid content provides a good proxy for energetic capacity, quantifying lipid classes important for energy storage and cellular structure offers additional insight and more accurate estimates of available energy reserves.
Lipid classes were different among species (Table 2). For example, nearshore P. rus had the greatest amounts of the storage lipid, triacylglycerols (mg of AFDW g −1 ; Fig. 3b). These lipids are mobilized during thermal stress providing metabolic energy, thereby mitigating some of the lost nutrition during coral bleaching 52,53 . We also noted species differences in the phospholipids and sterols (Table 1). Both lipid classes serve structural functions and were highly variable between species in both environments (Fig. 3c,d).
Proteins constituted a large proportion of macromolecules in coral tissues and energy provided from protein catabolism (specific enthalpy of combustion − 23.9 kJ g −1 ) is potentially substantial 26,54 . Coral soluble protein (J AFDW g −1 ) was different between species irrespective of their habitat origin (Table 1, Fig. 2c). Colonies of the plating species P. rugosa had the greatest protein concentrations in both environments (Table 1; Fig. 2c). While protein synthesis and maintaining protein reserves are important for growth and calcification, especially when conditions are physiologically demanding, the higher than usual protein content in both environments may simply be a species specific attribute of P. rugosa 17,[54][55][56] .
Carbohydrates, provide the least amount of energy (specific enthalpy of combustion − 17.5 kJ g −1 ) to the coral host 26,27 , and comprise the smallest proportion of overall coral energy reserves (Fig. 2d). We found carbohydrate concentrations (J AFDW g −1 ) were highly variable between species with members of the genus Porites, P. cylindrica and P. rus, having the greatest carbohydrate concentrations in colonies from both habitats ( Table 1, Fig. 2d). These molecules are synthesized quickly and catabolized more rapidly by the host than both lipids and proteins 45 . Since carbohydrates are important energy sources, the rapid catabolism of high concentrations in Porites could aid colonies of this genus when acclimating to thermal stress 11,32 .
While physiological differences between coral species are clear, differences between colonies of the same species are also critical for anticipating population level responses to thermal stress events 57 . Biomass composition is an important indicator of colony health and colony biomass is influenced by spatial and temporal environmental differences 13,20,43,47 . Despite marked differences in temperature, pH, symbiont association, and feeding ecologies, biomass and energy reserves were often similar between conspecific colonies from nearshore and offshore populations (Figs. 2, 3). Counter to our expectations, two offshore species (C. aspera and P. rus) had significantly greater relative biomass (AFDW cm −2 ) than their nearshore conspecifics (Table 1, Fig. 2a). We expected higher biomass and energy reserves in all nearshore corals because of their high temperature tolerance and resistance to bleaching 12,32 . Furthermore, C. aspera energetic reserves also differed between colonies from each habitat (Figs. 2, 3). Colonies from offshore populations had greater total lipid (J AFDW g −1 ), but soluble protein between populations remained the same (Fig. 2b,c). This common coral is regarded as one of the most environmentally tolerant species throughout the Indo-Pacific 13,28,58 . These findings highlight the broad physiological range of C. aspera in energy storage, possibly explaining this species' prevalence across many environments throughout the Indo-Pacific.
Remarkably few differences across conspecifics were found in lipid classes between nearshore and offshore populations (mg AFDW g −1 ; Table 1, Fig. 4). These lipid findings were unexpected, as lipid reserves and specific lipid classes are important in coral physiology, especially during thermal stress 44,52,54 . Unlike our conspecific comparisons for total biomass, total lipids, and soluble protein, the carbohydrates in C. aspera, F. abdita, and P. rugosa colonies from nearshore populations were significantly lower (J AFDW g −1 ) than in offshore populations (Table 1, Fig. 2d). High carbohydrate catabolism may be a consequence for colonies living in warmer more www.nature.com/scientificreports/ acidic environments. Catabolizing carbohydrate reserves is associated with the high bleaching threshold of Stylophora pistillata from the Red Sea 45 and differences in carbohydrates could be driven by increased reliance on autotrophically derived carbon in corals from offshore habitats 34 . Moreover, the lower carbohydrate content of nearshore colonies does not explain local acclimatization of these populations to higher mean temperatures 12,32 . Recent isotopic analyses of nearshore corals in Palau revealed increased reliance on zooplankton consumption compared to offshore coral colonies 34 . Typically, increased heterotrophy leads to corals with higher total biomass, including lipids and proteins 44,[59][60][61] . Despite the abundant zooplankton at these nearshore habitats [62][63][64] and coral host δ 15 N providing evidence that nearshore corals consume more zooplankton than offshore conspecifics 34 , nearshore corals typically had lower biomass (AFDW cm −2 ), soluble protein (J AFDW g −1 ), and carbohydrates (J AFDW g −1 ). This reveals a discrepancy between previously observed feeding ecologies and the current quantification of biomass, as well as energy reserves, of nearshore corals.
Corals in Palau's nearshore habitats have comparable skeletal extension, density, and calcification rates to offshore corals living in cooler and less acidic reefs 30 . Warmer temperatures (1-2 °C) and lower pH (~ 0.3 pH units) of nearshore habitats (Fig. 1) likely raise host metabolic demand through increased respiration (Q 10 effect) and proton pumping 65,66 . Consequently, nearshore corals may be catabolizing energy reserves to offset increased metabolic demand, which could explain nearshore coral colonies having similar or lower energy reserves compared to offshore coral colonies. High pCO 2 (lower pH) exposure resulted in a significant loss in biomass and energy reserves in Pocillopora acuta 67 , which could explain the overall trend of lower biomass within nearshore coral colonies. Physiological stress from ocean warming, and acidification, has likely altered metabolic demands of nearshore Palauan coral colonies to maintain calcification and cellular function, resulting in these coral populations catabolizing energy reserves to meet increased energy demands 65,67 but see Drenkard et al. 68 .
Colonies from nearshore coral communities in Palau harbor symbiotic dinoflagellates adapted to warm environments 12,32 . Specifically, Durusdinium trenchii is the dominant endosymbiont in many nearshore colonies, while nearshore and offshore Porites spp. colonies harbor genetically distinct Cladocopium C15, which differ in physiology and temperature tolerance 32 . Hence, similar total energetic reserves between nearshore and offshore coral conspecifics, suggest the symbiont species may have a larger effect on overall physiological acclimatization.
Host genetics strongly influences physiology and may provide another explanation for nearshore colonies thriving in these habitats [69][70][71] . Coral larvae in the nearshore habitats may experience higher post settlement selection than offshore corals. Furthermore, previous attempts to transplant offshore colonies to nearshore habitats had limited success (unpublished data), suggesting offshore colonies may not have a sufficient combination of genetic attributes for life in these more restrictive habitats. Recent analyses of Porites cf. lobata from Palau 72 , and other environments, suggest distinct populations are found in abnormally warm and variable environments 56,73-75 , hence, an individual's genetic make-up may also supersede the importance of energetic reserves to facilitate coral survival and persistence within nearshore habitats of the Rock Islands 76 .
In conclusion, coral tissue biomass, and associated energy reserves, regulation are influenced by several independent factors but overall these processes appear to converge upon a similar homeostatic set point, irrespective of environmental conditions. These findings provide insight into biochemical and energetic states of diverse coral species, while also contributing to our understanding of factors critical in the acclimatization of coral-dinoflagellate mutualisms to warm and acidic environments. The similarity in several biomass proxies, and energetic macromolecules, suggest differences in symbiont associations, kind of food consumed, as well as host genetics, may further enable nearshore populations to persist and thrive 77,78 . Ultimately, the recognition of biological processes promoting coral growth and persistence under different environmental conditions improves forecasting the future distribution and composition of reef coral communities.

Data availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.