Da

DRa

Ub

c

d

e

f

a

ARRAA

KSCPSH

1

emrad(tm

0h

Zoology 116 (2013) 262– 269

Contents lists available at ScienceDirect

Zoology

jou r n al hom epa ge: www . els ev ie r .com/ locate / zool

iversity in skeletal architecture influences biological heterogeneitynd Symbiodinium habitat in corals

enise M. Yosta, Li-Hsueh Wangb, Tung-Yung Fanb,c, Chii-Shiarng Chenb,d,e,aymond W. Leef, Emilia Sogina, Ruth D. Gatesa,∗

Hawaii Institute of Marine Biology, School of Ocean and Earth Science and Technology, University of Hawaii, 46-007 Lilipuna Road, Kaneohe, HI 96744,SANational Museum of Marine Biology and Aquarium, 2 Houwan Road, Checheng, Pingtung, Taiwan, ROCInstitute of Marine Biodiversity and Evolution, National Dong Hwa University, Pingtung, Taiwan, ROCInstitute of Marine Biotechnology, National Dong Hwa University, Pingtung, Taiwan, ROCDepartment of Marine Biotechnology and Resources, National Sun Yat-Sen University, Kaohsiung, Taiwan, ROCSchool of Biological Sciences, Washington State University, PO Box 644236, Pullman, WA 99164-4236, USA

r t i c l e i n f o

rticle history:eceived 10 December 2012eceived in revised form 24 June 2013ccepted 28 June 2013vailable online 3 August 2013

eywords:cleractinian coralsoral architectureerforate coralsymbiodinium habitatost–symbiont dynamics

a b s t r a c t

Scleractinian corals vary in response to rapid shifts in the marine environment and changes in reefcommunity structure post-disturbance reveal a clear relationship between coral performance and mor-phology. With exceptions, massive corals are thought to be more tolerant and branching corals morevulnerable to changing environmental conditions, notably thermal stress. The typical responses of mas-sive and branching coral taxa, respectively, are well documented; however, the biological and functionalcharacteristics that underpin this variation are not well understood. We address this gap by compar-ing multiple biological attributes that are correlated with skeletal architecture in two perforate (havingporous skeletal matrices with intercalating tissues) and two imperforate coral species (Montipora aequitu-berculata, Porites lobata, Pocillopora damicornis, and Seriatopora hystrix) representing three morphotypes.Our results reveal inherent biological heterogeneity among corals and the potential for perforate skele-tons to create complex, three-dimensional internal habitats that impact the dynamics of the symbiosis.Patterns of tissue thickness are correlated with the concentration of symbionts within narrow regionsof tissue in imperforate corals versus broad distribution throughout the larger tissue area in perforate

corals. Attributes of the perforate and environmentally tolerant P. lobata were notable, with tissues ∼5times thicker than in the sensitive, imperforate species P. damicornis and S. hystrix. Additionally, P. lobatahad the lowest baseline levels of superoxide and Symbiodinium that provisioned high levels of energy.Given our observations, we hypothesize that the complexity of the visually obscured internal environ-ment has an impact on host–symbiont dynamics and ultimately on survival, warranting further scientificinvestigation.

. Introduction

Scleractinian corals create reef habitats that provide criticalcosystem services worldwide. Corals have persisted over 500illion years, but have become increasingly threatened by the

apid changes in the marine environment linked to climate changend local human activities (Glynn, 1996). Corals have respondedramatically to environmental disturbances within recent decades

Berkelmans et al., 2004) resulting in large-scale global changes inhe community structure of reefs. These changes have prompted

any to question whether corals have the capacity to buffer,

∗ Corresponding author. Tel.: +1 808 236 7420; fax: +1 808 236 7443.E-mail address: rgates@hawaii.edu (R.D. Gates).

944-2006/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.zool.2013.06.001

© 2013 Elsevier GmbH. All rights reserved.

acclimatize and/or adapt to the dynamic environmental conditionspredicted to occur as a result of climate change and to survive intothe next century.

Observed ecological variation in the responses of corals andreef communities provides insight into which corals are likely topersist under challenging environmental conditions (Baker et al.,2004; van Woesik et al., 2011). Corals with massive morphologiesare among the most stress-tolerant corals, exhibiting much lowermortality rates following environmental disturbances (e.g., ther-mal stress) compared to branching and plating corals (Gates andEdmunds, 1999; McClanahan, 2004; Schloder and D’Croz, 2004).

Even within the same genus (e.g., Porites), massive coral mor-photypes appear to be less vulnerable to bleaching than theirbranching counterparts (McClanahan et al., 2001; but see also Guestet al. (2012)). Distinctive qualities of branching corals such as

logy 1

t(Etpcsfl(romethbct

a(twtiaisnAesmdr1st1o(aspaataS

utsi(mfhTwpcaipm

D.M. Yost et al. / Zoo

he combination of shallow tissue depths and limited resourcesLoya et al., 2001), along with high metabolic rates (Gates anddmunds, 1999), have been implicated as key factors that increasehe thermal sensitivity of these corals. Branching corals also dis-lay stronger responses to ocean acidification compared to massiveorals (Anthony et al., 2008). In addition to tissue depth, othertrategies for buffering environmental factors may include shuf-ing of endosymbiotic dinoflagellate (Symbiodinium) communitiessensu Baker, 2003) to optimize performance in response to envi-onmental change. Indeed, the functional integrity and persistencef the intimate associations between corals and Symbiodinium ulti-ately dictates whether corals survive in the face of changing

nvironmental conditions or not. That said, the biological charac-eristics that contribute to response variability among corals andow they compare among coral species is not well characterized,ut such analyses serve as important context for predicting howorals will respond to rising sea temperatures and ocean acidifica-ion.

Beneath the commonly known gross morphology of corals liesn interior calcium carbonate skeleton. Carbonate skeletal densitymicro-density) and porosity are key features of coral architec-ure and vary significantly between species, colonies, and evenithin a single colony (Bucher et al., 1998). Longitudinal sections

hrough perforate (porous) corals reveal skeletal matrices withntercalating tissues, whereas imperforate species typically have

‘veneer’ or surface covering of tissue that does not penetratento the skeleton as is the case with perforate corals. Perforatekeletal matrices are characteristic of many species in the domi-ant reef-building genera such as Acropora, Porites, Montipora andstreopora, but how skeletal porosity influences the biology andnvironmental range of corals is not well understood. There areeveral observations that suggest perforate architectural arrange-ents may have a positive impact on the survival and physiological

ynamics of the symbiosis. Deep tissues that penetrate perfo-ate skeletons are thought to enable the survival (Jokiel et al.,993) and rapid recovery (Krupp et al., 1993) of corals followingtressful events (low salinity, tissue damage), as well as facili-ate the calcification process (Buchsbaum-Pearse and Muscatine,971; Gladfelter, 1983), by promoting within-colony transportf Symbiodinium cells and potentially maximizing photosynthesisSantos et al., 2009). Additionally, corals with perforate skeletonsnd deep tissues appear more physiologically robust to thermaltress through reduced sunlight exposure in tissues and reducedhotodamage to Symbiodinium (Santos et al., 2009). Deep tissuesre known to co-occur with high levels of tissue-soluble proteinsnd high Symbiodinium densities in Porites lobata, which may givehese corals a competitive advantage over branching species suchs Pocillopora damicornis that have lower protein levels and fewerymbiodinium (Schloder and D’Croz, 2004).

In the present study, we address a fundamental gap in ournderstanding of coral biology by examining biological attributeshat are correlated with coral skeletal architecture. Four coralpecies (Montipora aequituberculata, Porites lobata, Pocillopora dam-cornis, and Seriatopora hystrix) representing three morphotypesfoliose, massive, and branching) were selected to provide infor-

ation on a wide range of the structural and biological complexityound in the scleractinians, as well as encompass corals known toave very different environmental thresholds (Loya et al., 2001).o compare the baseline biology of these four coral species,e evaluated their skeletal and tissue architecture and multiplehysiological traits, using a variety of analytical approaches. Thisomparison reveals very high levels of heterogeneity among corals

nd suggests that perforate skeletons may play an important rolen structuring internal architectures that create biologically com-lex Symbiodinium habitat. Our findings demonstrate that althoughassive macroarchitectures might suggest that the internal

16 (2013) 262– 269 263

architecture of P. lobata is simple, in fact it is not. As a result, inter-colating, deep tissues create a habitat within P. lobata that is uniqueamong the coral morphotypes investigated, a key biological featurethat in combination with other correlated attributes may explainsome of the ecological variation among corals.

2. Materials and methods

2.1. Experimental organisms

Qualitative and quantitative comparisons of the structural andbiological attributes of M. aequituberculata, P. lobata, P. damicor-nis, and S. hystrix (n = 32; 8 per species) were conducted using lightand confocal microscopy. Additionally, we measured total solubleprotein, total chlorophyll, Symbiodinium cell density, superoxidelevels and isotopic signatures of intact corals to investigate basestate physiology. Corals were selected (March 2011) from aquariumcollections maintained at the National Museum of Marine Biologyand Aquarium, Taiwan, allowing for high sample numbers (8 perspecies). The corals originated from the coastal reefs near Hobi-hou in southern Taiwan. Species-specific differences (e.g., betweenP. lobata and other species of Porites) were not explored in ourstudy due to the limited number of coral species maintained at theaquarium.

All corals were kept in the same large flow-through aquar-ium and experienced equivalent light and temperature regimes(nutrient levels were not measured). Prior to sampling, M. aequitu-berculata, P. lobata, P. damicornis and S. hystrix corals werefragmented (7.6 ± 0.8 cm2; average ± SEM) using a hammer andchisel, strung on monofilament line and hung in common gardensto acclimate for 1 week. All corals were kept (before and after selec-tion and sampling) under ambient conditions. Temperature datawere logged using HOBO temperature loggers (Onset ComputerCorp., Cape Cod, MA, USA) and averaged 23.4 ± 0.05 ◦C. Light condi-tions were recorded three times daily in the common garden usinga handheld probe and averaged 82.4 ± 9.9 �mol m−2 s−1.

2.2. Microscopy

Sub-fragments (approximately 1–2 cm2) of each coral werefixed in 4% paraformaldehyde for 1 h and then de-calcified in 10%HCl until the tissue tunics (intact biological tissues) were skeleton-free. The tissue tunics were stored in 1× PBS at 4 ◦C in the dark.Decalcified tissue tunics were bisected with a scalpel and visualizedusing confocal microscopy. Samples were scanned with excita-tions of violet (405 nm) and green (498 nm and 543 nm) light, andemissions were collected at 450 nm to visualize host tissues andat 600 nm to visualize autofluorescence of Symbiodinium. Mea-surements of tissue thickness were taken in triplicate at randomlocations on each tissue sample to determine a colony average.Coral tissue thickness was characterized by either (i) high biomass,thick tissues anastomosing through highly perforate skeletons ofperforate corals or (ii) low biomass, thin tissue ‘veneers’ of imper-forate corals. Thus, tissue thickness did not appear to be altered bywater absorption mechanisms (e.g., in the gastrovascular cavity)that may alter tissue thickness.

2.3. Physiological metrics

Following removal from tanks, corals were immediately air-brushed with filtered natural seawater (0.2 �m), and the totalvolume of the homogenate recorded. After airbrushing, all frag-

ments were dipped briefly in dilute bleach, left to dry, and theirsurface areas then measured using the paraffin wax technique(Stimson and Kinzie, 1991). Aliquots of the fresh homogenates wereimmediately frozen in liquid nitrogen and stored at −80 ◦C for later

2 logy 1

aiamWeCcufhisdc

2r

fowafiigtAtswtapu1tppnptfbSsdoaafccf1

2

b2ipwh

64 D.M. Yost et al. / Zoo

nalysis of protein and chlorophyll. Total protein was quantifiedn thawed homogenates using the BCA assay and bovine serumlbumin as a standard (Pierce, Rockford, IL, USA). Chlorophyll waseasured by passing 1 ml of thawed tissue homogenate across ahatman GF/F (glass fiber filter) and the filter was subsequently

xtracted in 90% acetone for 24 h at 4 ◦C (Parsons et al., 1984).hlorophyll in these extracts was measured spectrophotometri-ally and chlorophyll (a and c2) concentrations were calculatedsing the equation from Jeffrey and Humphrey (1975). Additionalresh homogenate aliquots were preserved in 3.6% paraformalde-yde to assess Symbiodinium cell density. Before counting, cells

n these aliquots were washed three times with filtered naturaleawater (0.2 �m) and resuspended in 1× PBS. Symbiodinium cellensities were determined using a Scepter handheld automatedell counter (EMD Millipore Corp., Billerica, MA, USA).

.4. Assay for superoxide ions in intact host tissues by NBTeduction

Nitroblue tetrazolium (NBT) is reduced by superoxide ions toorm its diformazan derivative (formazan). To determine baselinexidative loads in each coral, the reduction of NBT to formazanas measured in intact corals exposed to 1.2 × 10−4 M NBT (Nii

nd Muscatine, 1997) for 20 min. The NBT solution was made usingltered natural seawater (0.2 �m) and each coral was submerged

n a 100 ml volume of the solution in individual, aerated 200 mllass beakers. The beakers were floated in ambient temperatureanks in the light and aerated throughout the incubation period.fter the incubation, the NBT solution was poured off and each of

he corals rinsed by filling the beaker with fresh, filtered naturaleawater (0.2 �m). Each coral was then immediately airbrushedith filtered seawater (0.2 �m) to obtain a homogenate and the

otal volume recorded. Host and symbiont are difficult to separate;irbrushing, though not a perfect technique, is commonly used torovide representative samples of these fractions and permits these of different normalization indices (e.g., Szmant and Gassman,990; Yost and Mitchelmore, 2010). Each sample homogenate washen centrifuged at 1000 × g for 5 min. The supernatant containingrimarily host tissue was removed carefully so as not to disturb theellet that contained primarily the algae and formazan. The super-atants were frozen in liquid nitrogen and stored at −80 ◦C for laterrotein analysis. Symbiodinium pigments that can interfere withhe spectrophotometric measurement of the ethanol-insoluble NBTormazan (Nii and Muscatine, 1997) were extracted from the Sym-iodinium/formazan pellets using 1 ml of 95% EtOH for 1 h at 4 ◦C.amples were centrifuged at 5000 × g for 5 min and the EtOHupernatant removed. The pelleted formazan was dissolved in 1 mlimethylformamide (DMF) by sonication (3 min). The absorbancef the resulting solution was measured spectrophotometricallyt 550 nm. Spectrophotometric examination of the EtOH extractnd DMF solution verified the complete removal of algal pigmentsollowing EtOH extraction. The concentration of formazan was cal-ulated from the measured absorbance and the molar extinctionoefficient of the NBT formazan, with results expressed as nmol NBTormazan produced per milligram of protein (Nii and Muscatine,997).

.5. Inorganic carbon and nitrogen assimilation experiments

Intact corals (n = 5 per species) were exposed to 500 �M 13C-icarbonate and 100 �M 15NH3 for 30 min in individual, aerated00 ml glass beakers. Beakers were floated (two-thirds submerged)

n tanks under ambient conditions. At the end of the incubationeriod each coral was briefly placed in fresh, filtered natural sea-ater (0.2 �m) before being immediately airbrushed to obtain aomogenate (as detailed above) with deionized water (to minimize

16 (2013) 262– 269

the interference of chloride ions in the isotopic analysis). Dilutedhomogenates were subsampled and centrifuged at 12,000 × g for5 min at 4 ◦C to separate the coral host from Symbiodinium. The col-orless supernatant (coral host fraction) was pipetted into a freshmicrofuge tube and placed on ice. The remaining pellet (Symbio-dinium fraction) was resuspended in ice-cold artificial seawater.This washing procedure was repeated three times, and the pelletscontaining primarily Symbiodinium were resuspended in 10–20 mlartificial seawater. All samples were immediately frozen or driedfor later isotopic analysis.

2.6. Stable isotope analysis

Samples for the analysis of 15N/14N and 13C/12C ratios wereprepared by removing unincorporated 15NH3 and 13CO2, with1 N sodium hydroxide for 24 h, followed by 1 N phosphoric acid.Treated samples were dried and 0.5–1 mg subsamples analyzed bycontinuous flow isotope ratio mass spectrometry. Samples werecombusted and the resulting N2 and CO2 gases separated in an ele-mental analyzer (EuroVector S.p.A., Milan, Italy); these gases wereadmitted into a Micromass Isoprime mass spectrometer (GV Instru-ments, Manchester, UK) for determination of 15N/14N and 13C/12Cratios. Isotopic analyses are presented in atom % 13C or 15N calcu-lated as 15N/(14N + 15N) × 100 or 13C/(12C + 13C) × 100. Raw atom %values were corrected for background (control) values and used tocalculate atom % enrichment in the samples.

For each of the experiments, unenriched control corals (n = 5 perexperiment; no 13C-bicarbonate or 15NH3 addition) were sampledand analyzed. Atom % enrichment was calculated by subtractingaverage control atom % values from experimental coral host andSymbiodinium atom % values. Control atom % values ranged from1.08974 to 1.09145 atom % 13C and from 0.36928 to 0.36938 atom% 15N.

2.7. Statistical analysis

All data were checked for normality and homogeneity ofvariances before statistical analysis. Non-normal data were log-transformed prior to statistical analysis. Data were analyzed byanalysis of variance (ANOVA) or Student’s t-tests. Following sig-nificant results from ANOVA, pairwise differences were analyzedby the Tukey method. Pearson’s correlation coefficients were cal-culated to determine correlations between indices (pairwise for thefollowing endpoints: formazan, chlorophyll, Symbiodinium den-sity). All analyses were conducted using Minitab statistical software(Minitab Inc., State College, PA, USA), with ̨ = 0.05 for all tests.

3. Results

Distinct skeletal and tissue characteristics were apparent atmultiple spatial scales, underlining profound biological differencesamong morphotypes (Fig. 1). Branching corals have high levels ofexternal structural diversity due to their intricate skeletons thatallow for considerable habitat structure that scales with growth.Conversely, the foliose, and massive corals, have condensed three-dimensional architectures and comparatively reduced degrees ofmacro-morphological complexity (Fig. 1, column a). Under a lightmicroscope skeletal micro-architecture reveals a spectrum of polyparrangements: cup-like calices (branching P. damicornis and S. hys-trix), sub-surface polyps (massive P. lobata) and comparativelydispersed calices and elaborate coenosteum (foliose M. aequituber-

culata) (Fig. 1, column b). In column c of Fig. 1, simple schematicdiagrams of longitudinal sections through lobate, massive andbranching corals show different patterns of polyp, skeletal and tis-sue (with Symbiodinium) arrangements. Cross-sectional confocal

D.M. Yost et al. / Zoology 116 (2013) 262– 269 265

Fig. 1. Visual comparison of M. aequituberculata, P. lobata, P. damicornis and S. hystrix showing colony morphology, skeletal architecture and tissue tunics with Symbiodinium.Column (a): macroscopic photographs of external colony morphology appearing foliose or plate-like with papillae (bumps on skeleton surface) in M. aequituberculata,massive and simple in P. lobata, and branching and more externally complex (multiple branches) in P. damicornis and S. hystrix. Column (b): microscopic photographs oftissue-free skeletal micro-architectures showing the arrangement of corallites (skeleton produced by a single polyp) within the greater skeletal matrix. Column (c): simpleschematic diagrams of longitudinal sections through perforate (porous) and imperforate corals in the present study paralleling the confocal micrographs in column (d).Perforate morphology in the plate-like and massive corals reveals porous skeletal matrices with intercalating tissues and Symbiodinium distributed throughout, whereasthe imperforate branching species have a ‘veneer’ of tissue that does not penetrate into the skeleton and Symbiodinium populations that are near the surface and moreconcentrated in the polyps. Column (d): macroscopic confocal micrographs (longitudinal views) of decalcified tissue tunics showing perforate tissues with ‘holes’ whereskeleton was removed, or, alternatively, thin coenosarcs (tissue between corallites) that covered imperforate skeletons in branching species. Note different scale bars. Tissuedepth measurements are described in Section 2.2; P indicates an individual polyp; S. hystrix tissue tunic is circular due to horizontal sectioning through a branch. Column (e):confocal micrographs of tissue tunic surfaces showing polyp arrangements with the distribution of Symbiodinium that appear concentrated in polyps and near the surfacei d whito ith od

msacsSihba

TBti

n branching corals (host tissues appear as blue and green hues or gray in black anr black (black and white micrographs). Black and white micrographs are shown wistribution patterns.

icrographs of decalcified tissues show differences in both the tis-ue depths and the spatial distribution of Symbiodinium in tissuesmong coral species (Fig. 1, column d, similar to the schematics inolumn c). M. aequituberculata and P. lobata corals have perforatekeletons (porous with a three-dimensional mesh-like structure;antos et al., 2009), whereas the skeletons of the branching P. dam-cornis and S. hystrix corals are imperforate. The perforate corals

ave the greatest tissue depths, approximately three (M. aequitu-erculata) and five (P. lobata) times greater than branching corals,nd Symbiodinium are distributed throughout the intricate maze of

able 1iological attributes (mean ± SEM) of M. aequituberculata, P. lobata, P. damicornis and S. hystissue depths compared to perforate corals. Atom % 13C ratios are also higher in the brann the symbiosis. Other attributes show biological heterogeneity within and between imp

Attribute Montip

Tissue depth (�m) 608 ± 5Symbiodinium spatial pattern in tissue DispersTotal chlorophyll (pg cell−1) 3.44 ±

Symbiodinium cell diameter (�m) 9.00 ±

Atom % 13C ratio pellet:supernatant (primarily Symbiodinium:host) 0.30 ±

Atom % 15N ratio pellet:supernatant (primarily Symbiodinium:host) 0.20 ±

e micrographs (P. lobata, S. hystrix); Symbiodinium appear red (color micrographs)verlays of chlorophyll autofluorescence (red spheres) to emphasize Symbiodinium

coral tissue (Fig. 1, column d, and Table 1). Conversely, the branch-ing corals have shallow tissue layers and Symbiodinium are in closeproximity to one another, concentrated in the thin ‘veneer’ of coralpolyp and inter-calyx tissues atop their imperforate skeletons. Inaddition, the distribution of Symbiodinium is more concentratedin the polyps of the imperforate corals (Fig. 1, columns d ande).

The massive Porites corals exhibited comparatively highlevels of protein, high Symbiodinium densities and low Symbio-dinium chlorophyll concentrations (Fig. 2) when compared to the

rix corals. Imperforate, branching corals (P. damicornis and S. hystrix) have shallowerching corals, a notable difference with implications for carbon recycling dynamicserforate and perforate corals.

ora Porites Pocillopora Seriatopora

5 958 ± 22 183 ± 36 154 ± 27ed Dispersed Concentrated Concentrated

0.50 0.98 ± 0.12 3.18 ± 0.38 1.56 ± 0.110.08 8.15 ± 0.12 8.44 ± 0.06 7.66 ± 0.040.10 0.36 ± 0.07 1.65 ± 0.41 1.84 ± 0.320.01 0.43 ± 0.01 0.43 ± 0.01 0.31 ± 0.04

266 D.M. Yost et al. / Zoology 116 (2013) 262– 269

F (Pl),S oncens ant (s

bnS0(la(

FMsb

ig. 2. Comparison of physiological metrics in M. aequituberculata (Ma), P. lobataymbiodinium density (Symbiodinium/mg protein), (c) Symbiodinium chlorophyll cuperoxide radical levels; nmol formazan/mg protein). Letters indicate non-signific

ranching and foliose corals, which shared similar (i.e., not sig-ificantly different) physiological patterns among coral hosts andymbiodinium parameters. Total soluble protein levels ranged from.25 to 1.10 mg protein−1 cm−2 among all species investigated

Fig. 2). M. aequituberculata and S. hystrix corals contained simi-ar amounts of protein (0.39 ± 0.07; 0.50 ± 0.03 mg protein−1 cm−2)nd averaged 45% of the 1.10 ± 0.12 mg protein−1 cm−2 for P. lobataANOVA P < 0.05); the average protein content of the P. lobata

ig. 3. 13C and 15N atom % enrichment values (atom %/mg protein) in supernatant (prim. aequituberculata (Ma), P. lobata (Pl), P. damicornis (Pd), and S. hystrix (Sh) corals. Valu

ymbiosis (n = 20; 5 corals per species). Letters indicate non-significant (same letter) oetween species.

P. damicornis (Pd), and S. hystrix (Sh). (a) Soluble protein (mg protein/cm2), (b)trations (�g chlorophyll/mg protein), and (d) formazan concentration (indicates

ame letter) or significant (different letters) statistical differences between species.

homogenates was also significantly higher than that of P. dam-icornis by a margin of 88% (Fig. 2a). Though P. lobata hosted arelatively high number of Symbiodinium (Fig. 2b), the chlorophyllcontent of those algal cells was significantly lower than that of

M. aequituberculata, P. damicornis, and S. hystrix corals (Fig. 2c andTable 1), and was inversely correlated with Symbiodinium densities(Symbiodinium/mg protein; Fig. 2b) in a moderately negative rela-tionship (−0.316, P > 0.05). Among corals, the average diameter of

arily coral tissue; black bars) and pellet (primarily Symbiodinium; gray bars) fores indicate dynamics of carbon and nitrogen assimilation and cycling in the intactr significant (different letters) statistical differences for Symbiodinium:host ratios

logy 1

Se

mplcNvb(a((

5lcassifehll1

dwpti

%he(h5sbs

4

otiecbhi

stnapsovs

D.M. Yost et al. / Zoo

ymbiodinium (in �m ± SEM) was variable and significantly differ-nt between M. aequituberculata and S. hystrix (Table 1).

Baseline levels of superoxide ions in intact coral fragments, aseasured by the reduction of NBT to formazan (nmol formazan/mg

rotein), were significantly higher in M. aequituberculata than in P.obata, P. damicornis or S. hystrix, indicating that M. aequituberculataorals had the highest levels of superoxide ions during the 20 minBT exposure period (Fig. 2d). Additionally, these results showarying degrees of correlation (Pearson’s correlation coefficients)etween superoxide ion production and Symbiodinium densitiesSym) or chlorophyll concentrations (chl) within a species (forll correlations P > 0.05): M. aequituberculata: 0.377 (Sym), −0.066chl); P. lobata: 0.598 (Sym), 0.422 (chl); P. damicornis: −0.177Sym), −0.465 (chl); S. hystrix: 0.614 (Sym), 0.626 (chl) (Fig. 2b–d).

Following the 30 min isotopic exposure (100 �M 15N and00 �M 13C), both supernatant (primarily coral tissues) and pel-

eted fractions (primarily Symbiodinium) were enriched witharbon and nitrogen (Fig. 3). 13C enrichment in the supernatantsveraged 0.011 ± 0.001 atom % 13C/mg protein and there were noignificant differences in the levels of enrichment among coralpecies. In contrast, 13C enrichment patterns in the pellet fractionssolated from the different coral species were different. The pelletsrom S. hystrix and P. damicornis corals had the highest levels of 13Cnrichment; significant differences were evident in pellets from S.ystrix (0.023 ± 0.006) and M. aequituberculata (0.005 ± 0.002) or P.

obata (0.003 ± 0.001) (ANOVA P < 0.05). Additionally, pellets iso-ated from M. aequituberculata and P. lobata were significantly less3C enriched than their coral hosts (t-test P < 0.05; Table 1). Theseata show that the pellets from imperforate, branching speciesere significantly more enriched with 13C sodium bicarbonate (13Cellet to supernatant ratios were more than five times greater)han those in perforate, foliose or massive corals at the end of thencubation period.

15N enrichment patterns in pellets averaged 0.118 ± 0.010 atom15N/mg protein and were not significantly different among coralosts (Fig. 3). The supernatant fractions were, however, morenriched with 15N ammonium chloride compared to their pelletsANOVA P < 0.05; Table 1) and were significantly different amongosts. M. aequituberculata supernatants averaged approximately0% more atom % 15N/mg protein than those of the other coralpecies and were significantly more enriched (0.695 ± 0.105) thanoth P. lobata (0.381 ± 0.069) and P. damicornis (0.274 ± 0.069)upernatants.

. Discussion

The present study reveals inherent differences in the basic biol-gy of corals representing four important genera and emphasizeshe high levels of structural and biological heterogeneity that existn corals. The patterns that emerge show different combinations ofxternal and internal physical complexity, characteristics that areorrelated with fundamental biological attributes of the coral sym-iosis and the quality, quantity and complexity of Symbiodiniumabitat. Furthermore, our findings exemplify that skeletal porosity

s a trait related to known environmental thresholds in corals.Our investigation of two perforate and two imperforate coral

pecies revealed high variance in the mean values for most ofhe characteristics investigated. Within-group variation was pro-ounced in perforate species (e.g., protein and chlorophyll contentsnd superoxide radical levels) and variance was also high betweenerforate and non-perforate groups. These results suggest that

ampling more species per group would allow further explorationf the observed high levels of variation and how high- and low-ariability traits may pair with different skeletal features acrosspecies to influence response variability. In the present study, the

16 (2013) 262– 269 267

tradeoff of using a low number of species per group was balanced bya higher sampling number within individual species (n = 8), whichled to several possible conclusions notwithstanding high levels ofvariability.

Confocal micrographs of tissue cross-sections show the funda-mental differences in tissue architecture between the perforateand imperforate corals. The highly perforated skeletons of themassive P. lobata and foliose M. aequituberculata corals allow thecoral tissues to penetrate to much greater depths compared tothe imperforate, branching P. damicornis and S. hystrix corals. Thisinternal environment shapes the effective three-dimensional habi-tat for Symbiodinium, creating more dispersed, or comparativelyrestricted, spatial arrangements of Symbiodinium within the hostcoral in perforate versus imperforate corals, respectively. Thus,the relative partitioning of biological complexity is very differ-ent among the morphotypes investigated – branching imperforatecorals exhibit high levels of external structural complexity andspatial diversity, whereas massive and foliose perforate coralsexhibit much simpler external structure and much higher levelsof internal complexity. The internal environment that shapes theSymbiodinium habitat in perforate corals adds dimensionality tothe interaction between host and symbiont by providing a broaderscope for refuge from the external environment, a feature that isabsent from imperforate associations.

Massive Porites corals exhibit several unique features that dis-tinguish them from the other corals investigated. The comparisonof P. lobata and M. aequituberculata suggests that perforate skele-tons may be predictive of deeper tissues, but are not necessarilypredictive of other fundamental biological traits. Notably, P. lobatahad the highest levels of tissue soluble protein, and the lowestlevels of total chlorophyll compared to all other corals. Further-more, only P. lobata showed Symbiodinium densities that wereinversely correlated with chlorophyll concentrations in a mod-erately negative relationship, and levels of superoxide radicals(measured as formazan concentration) in P. lobata were lowestoverall. These results suggest that oxygen radical levels may belinked to Symbiodinium densities and/or the chlorophyll contentof Symbiodinium cells through positive or negative relationships.Both of these parameters merit concurrent consideration given thata coral’s complement of Symbiodinium is typically regarded as a pri-mary source of cellular reactive oxygen species for the coral host(Warner et al., 1999; Fitt et al., 2001; Weis, 2008). Many of theunique attributes described above appear to be correlated with P.lobata’s highly modified internal environment, an observation thatwarrants further investigation and suggests a broader relationshipbetween microenvironment and response variability among coraltaxa.

Specific features related to perforate skeletons might explainsome of the links between the physical and biological aspects ofthese corals. For example, Symbiodinium in perforate corals mayencounter lower light levels due to the light absorption propertiesof thicker coral tissues and higher Symbiodinium densities (Teranet al., 2010). Additionally, intercalating tissues create the opportu-nity for dispersal and migration of Symbiodinium within the tissuesto exploit the different microenvironments afforded by the com-plexity of the space (Santos et al., 2009). In combination with othertraits such as low permeability and high compression strength (Wuet al., 2009), massive Porites’ highly modified internal environmentprovides a greater scope for uniqueness as a result of its inherentlyhigh levels of biological variability.

In addition to our observations there are other unique featuresof P. lobata corals that are likely to contribute to this coral’s abil-

ity to survive changing environmental conditions. P. lobata is theonly species in the present study with a relatively fixed morphol-ogy, whereas the others are known to be phenotypically plastic,changing their external morphological composition in response

2 logy 1

tPSltCtn2g

tpthdosotpsddaiS(fsactSlt(p

gitM(aitalittsebibdaf

mcpvp

68 D.M. Yost et al. / Zoo

o environmental factors (Lesser et al., 1994; Kaandorp, 1999).. lobata’s complex internal environment may serve to diversifyymbiodinium habitats despite its simple external architecture. P.obata is also known to have a primarily autotrophic mode of nutri-ion, a low growth rate and high fidelity to Symbiodinium type15 symbionts (Stat et al., 2009 and references therein). Each ofhe above-mentioned features may in part determine this coral’sotable low bleaching susceptibility or long life span (Loya et al.,001; van Woesik et al., 2011), but it is likely complex and dynamicroupings of key features that ultimately inform such robustness.

The imperforate branching corals show similar patterns of struc-ural and biological organization that were not present among theerforate corals in our study. In addition to their imperforate skele-ons and highly complex external architectures, branching coralsave thin tissues, and similar chlorophyll levels (though Symbio-inium densities were significantly different) and baseline levelsf superoxide radicals. Furthermore, the imperforate corals exhibitimilar 13C enrichment patterns that differ significantly from thosef the perforate corals as a group (Symbiodinium more enrichedhan coral hosts). Corals can rely mainly on autotrophy, heterotro-hy or a combination of both (polytrophy) as energy acquisitiontrategies in nutrition (Holbreque and Ferrier-Pages, 2009). Theifferent atom % 13C ratios (pellet:supernatant (primarily Symbio-inium:host) fractions) between perforate and imperforate coralsre a key difference in this important aspect of symbiosis and mayndicate fast or slow recycling (respectively) of carbon betweenymbiodinium and the coral host. Additionally, the atom % 15N ratiospellet:supernatant) were similar for imperforate corals, but notor perforate corals, where M. aequituberculata supernatants wereignificantly more 15N enriched than their Symbiodinium fractionnd those of the other corals investigated. These results may indi-ate a greater capacity for nitrogen assimilation and/or recycling inhe M. aequituberculata corals or a potential mechanism to controlymbiodinium proliferation, which may be suppressed by nitrogenimitation (Falkowski et al., 1993). In future studies, mass balanceechniques may be used to explicitly determine which partnersymbiont or host) specifically contributes to isotope enrichmentatterns over time.

Although this factor is beyond the scope of this study, nitro-en limitation (and/or carbon limitation (Franklin et al., 2004)) maynfluence the oxidative state of Symbiodinium and could be a fac-or contributing to the high levels of superoxide radicals present in. aequituberculata. For example, a coral’s reliance on autotrophic

Symbiodinium) and/or heterotrophic sources of nutrition can bessessed using stable isotopes, although the interpretation of oursotopic results is complicated by biological processes innate tohe system, including respiration (McConnaughey et al., 1997)nd endosymbiotic photosynthesis (Swart et al., 2006). Nonethe-ess, the contrasting enrichment patterns between perforate andmperforate groups suggest different capacities for Symbiodiniumo provision carbon to their host (Stat et al., 2008) and indicatehat Symbiodinium performance may be driven in part by sub-urface habitat interactions that are not readily discernable from anxterior morphological perspective. It is also likely that the assem-lages of Symbiodinium genotypes differed among the coral taxa

nvestigated (Chen et al., 2005). The explicit determination of Sym-iodinium genotypes was not a focal point of this study, but knownifferences among types may explain some differences in the coralttributes investigated, warranting the inclusion of genotyping inuture comparisons.

Our comparative assessment of perforate and imperforate coralorphotypes demonstrates that perforate skeletons allow for and

reate fundamentally different arrangements of biological com-lexity in corals. The fixed external appearance of P. lobata coralseils an inwardly elaborate mosaic of skeleton and tissue that sup-orts a dense population of Symbiodinium that feeds this highly

16 (2013) 262– 269

autotrophic coral. In addition to the unique attributes we describe,the massive Porites are also reported to be slow growing andto exhibit high metabolic rates when compared to branchingspecies (reviewed in Buddemeier and Kinzie, 1976). Collectively,such biological attributes combine to promote enduring symbi-otic relationships that show increased resistance to environmentalchallenges. As corals do not respond uniformly to their environ-ment, analyses that consider the perforate/imperforate nature ofcoral skeletons in addition to other key aspects of coral biologysuch as morphology, feeding strategy and Symbiodinium speci-ficity are likely to improve the predictability of coral performancein response to dynamic environmental conditions. Additionally,species-specific comparisons will likely generate further insightsinto the ecological significance of morphological and physiologi-cal differences within genera. Predictions of coral response basedon taxonomy will lack the biological resolution of groups assem-bled according to functional or structural attributes. The use ofsuch groups and the assessment of their relative abundance onreefs will enhance community modeling frameworks and broadenthe conceptual context from which predictions of coral communityresponse to environmental change are made.

Acknowledgments

We thank all of those at the National Museum of Marine Biol-ogy and Aquarium, Taiwan, who assisted in this project. The studywas funded by the NSF OISE award #1042509 and represents HIMBcontribution number 1560 and SOEST contribution number 8975.

References

Anthony, K.R.N., Kline, D.I., Diaz-Pulido, G., Dove, S., Hoegh-Guldberg, O., 2008. Oceanacidification causes bleaching and productivity loss in coral reef builders. Proc.Natl. Acad. Sci. U. S. A. 45, 17442–17446.

Baker, A.C., 2003. Flexibility and specificity of coral–algal symbiosis: diversity, ecol-ogy, and biogeography of Symbiodinium. Annu. Rev. Ecol. Syst. 34, 661–689.

Baker, A.C., Starger, C.J., McClanahan, T.R., Glynn, P.W., 2004. Corals’ adaptiveresponse to climate change. Nature 430, 741.

Berkelmans, R., De’ath, G., Kininmonth, S., Skirving, W.J., 2004. A comparison ofthe 1998 and 2002 coral bleaching events of the Great Barrier Reef: spatialcorrelation, patterns and predictions. Coral Reefs 23, 74–83.

Bucher, D.J., Harriott, V.J., Roberts, L.G., 1998. Skeletal micro-density, porosity andbulk density of acroporid corals. J. Exp. Mar. Biol. Ecol. 228, 117–136.

Buchsbaum-Pearse, V., Muscatine, L., 1971. Role of symbiotic algae (zooxanthellae)in coral calcification. Biol. Bull. 141, 350–363.

Buddemeier, R.W., Kinzie III, R.A., 1976. Coral growth. Oceanogr. Mar. Biol. Annu.Rev. 14, 183–225.

Chen, A.C., Yang, Y.W., Wei, N.V., Tsai, W.S., Fang, L.S., 2005. Symbiont diversity inscleractinian corals from tropical reefs and subtropical non-reef communitiesin Taiwan. Coral Reefs 24, 11–22.

Falkowski, P.G., Dubinsky, Z., Muscatine, L., McCloskey, L., 1993. Population controlin symbiotic corals. Bioscience 43, 606–611.

Fitt, W.K., Brown, B.E., Warner, M.E., Dunne, R.P., 2001. Coral bleaching: interpreta-tion of thermal tolerance limits and thermal thresholds in tropical corals. CoralReefs 20, 51–65.

Franklin, D.J., Hoegh-Guldberg, O., Jones, R.J., Berges, J.A., 2004. Cell death and degen-eration in the symbiotic dinoflagellates of the coral Stylophora pistillata duringbleaching. Mar. Ecol. Prog. Ser. 272, 117–130.

Gates, R.D., Edmunds, P.J., 1999. The physiological mechanisms of acclimatizationin tropical reef corals. Am. Zool. 39, 30–43.

Gladfelter, E.H., 1983. Circulation of fluids in the gastrovascular system of the reefcoral Acropora cervicornis. Biol. Bull. 165, 619–636.

Glynn, P.W., 1996. Coral reef bleaching: facts, hypotheses and implications. Glob.Change Biol. 2, 495–509.

Guest, J.R., Baird, A.H., Jeffrey, A.M., Maynard, J.A., Muttaquin, E., Edwards, A.J., Camp-bell, S.J., Ywedall, K., Affendi, Y.A., Chou, L.M., 2012. Contrasting patterns of coralbleaching susceptibility in 2010 suggest an adaptive response to thermal stress.PLoS ONE 7, e33353.

Holbreque, F., Ferrier-Pages, C., 2009. Heterotrophy in tropical scleractinian corals.Biol. Rev. 84, 1–17.

Jeffrey, S.W., Humphrey, G.F., 1975. New spectrophotometric equations for deter-mining chlorophylls a, b, c1 and c2 in higher plants, algae, and naturalphytoplankton. Biochem. Physiol. Pflanzen 167, 191–194.

Jokiel, P.L., Hunter, C.L., Taguchi, S., Watari, L., 1993. Ecological impact of a fresh-water “reef kill” in Kaneohe Bay, Oahu, Hawaii. Coral Reefs 12, 177–184.

logy 1

K

K

L

L

M

M

M

N

P

S

S

the criteria for bone-tissue engineering scaffolds. J. Mater. Sci. Mater. Med. 20,

D.M. Yost et al. / Zoo

aandorp, J.A., 1999. Morphological analysis of growth forms of branching marinesessile organisms along environmental gradients. Mar. Biol. 134, 295–306.

rupp, D.L., Jokiel, P.L., Cartrand, T.S., 1993. In: Richmond, R.H. (Ed.), Asexual repro-duction by the solitary scleractinian coral Fungia scutaria on dead parent coralliain Kaneohe Bay, Oahu, Hawaiian Islands. Proceedings of the 7th Int. Coral ReefSymposium, vol. 1. University of Guam Press, UOG Station, Guam, pp. 527–534.

esser, M.P., Weis, V.M., Patterson, M.R., Jokiel, P.L., 1994. Effects of morphologyand water motion on carbon delivery and productivity in the reef coral, Pocillo-pora damicornis (Linnaeus): diffusion barriers, inorganic carbon limitation, andbiochemical plasticity. J. Exp. Mar. Biol. Ecol. 178, 153–179.

oya, Y., Sakai, K., Yamazato, K., Nakano, H., Sambali, H., van Woesik, R., 2001. Coralbleaching: the winners and the losers. Ecol. Lett. 4, 122–131.

cClanahan, T.R., 2004. The relationship between bleaching and mortality of com-mon corals. Mar. Biol. 144, 1239–1245.

cClanahan, T.R., Muthiga, N.A., Mangi, S., 2001. Coral and algal changes after the1998 coral bleaching: interaction with reef management and herbivores onKenyan reefs. Coral Reefs 19, 380–391.

cConnaughey, T.A., Burdett, J., Whelan, J.F., Paull, C.K., 1997. Carbon isotopes inbiological carbonates: respiration and photosynthesis. Geochim. Cosmochim.Acta 3, 611–622.

ii, C.M., Muscatine, L., 1997. Oxidative stress in the symbiotic sea anemone Aip-tasia pulchella (Carlgren, 1943): contribution of the animal to superoxide ionproduction at elevated temperature. Biol. Bull. 192, 444–456.

arsons, T.R., Maita, Y., Lalli, C.M., 1984. Plant pigments. In: Parsons, T.R., Maita, Y.,Lalli, C.M. (Eds.), A Manual of Chemical and Biological Methods for SeawaterAnalysis. Pergamon Press, Oxford, pp. 99–112.

antos, S.R., Toyoshima, J., Kinzie III, R.A., 2009. Spatial and temporal dynamics

of symbiotic dinoflagellates (Symbiodinium: Dinophyta) in the perforate coralMontipora capitata. J. Coral Reef Stud. 11, 139–147.

chloder, C., D’Croz, L., 2004. Responses of massive and branching coral species tothe combined effects of water temperature and nitrate enrichment. J. Exp. Mar.Biol. Ecol. 313, 255–268.

16 (2013) 262– 269 269

Stat, M., Morris, E., Gates, R.D., 2008. Functional diversity in coral–dinoflagellatesymbiosis. Proc. Natl. Acad. Sci. U. S. A. 105, 9256–9261.

Stat, M., Pochon, X., Cowie, R.O.M., Gates, R.D., 2009. Specificity in communi-ties of Symbiodinium in corals from Johnston Atoll. Mar. Ecol. Prog. Ser. 386,83–96.

Stimson, J., Kinzie III, R.A., 1991. The temporal pattern and rate of release ofzooxanthellae from the reef coral Pocillopora damicornis (Linnaeus) undernitrogen enrichment and control conditions. J. Exp. Mar. Biol. Ecol. 153,63–74.

Swart, P.K., Leder, J.J., Szmant, A., Dodge, R.E., 2006. The origin of variations in theisotopic record of Scleractinian corals: II. Carbon. Geochim. Cosmochim. Acta 60,2871–2886.

Szmant, A.M., Gassman, N.J., 1990. The effects of prolonged “bleaching” on the tissuebiomass and reproduction of the reef coral Montastrea annularis. Coral Reefs 8,217–224.

Teran, E., Mendez, E.R., Enriquez, S., Iglesias-Prieto, R., 2010. Multiple light scatteringand absorption in reef-building corals. Appl. Opt. 49, 5032–5042.

van Woesik, R., Sakai, K., Ganase, A., Loya, Y., 2011. Revisiting the winners and thelosers a decade after coral bleaching. Mar. Ecol. Prog. Ser. 434, 67–76.

Warner, M.E., Fitt, W.K., Schmidt, G.W., 1999. Damage to photosystem II in symbioticdinoflagellates: a determinant of coral bleaching. Proc. Natl. Acad. Sci. U. S. A.96, 8007–8012.

Weis, V.M., 2008. Cellular mechanisms of cnidarian bleaching: stress causes thecollapse of symbiosis. J. Exp. Biol. 211, 3059–3066.

Wu, Y.-C., Lee, T.-M., Chiu, K.-H., Shaw, S.-Y., Yang, C.-Y., 2009. A comparative studyof the physical and mechanical properties of three natural corals based on

1273–1280.Yost, D.M., Mitchelmore, C.L., 2010. Determination of total and particulate dimethyl-

sulfoniopropionate (DMSP) concentrations in four scleractinian coral species: acomparison of methods. J. Exp. Mar. Biol. Ecol. 395, 72–79.