8
Molecular Ecology (2003) 12, 3477– 3484 doi: 10.1046/j.1365-294X.2003.01988.x © 2003 Blackwell Publishing Ltd Blackwell Publishing Ltd. Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef K. E. ULSTRUP * and M. J. H. VAN OPPEN , *Department of Phycology, Botanical Institute, University of Copenhagen, Copenhagen, Denmark; Australian Institute of Marine Science, PMB 3, Townsville MC, Qld 4810, Australia; School of Pharmacy and Molecular Sciences, James Cook University, Townsville, Australia Abstract Intra- and intercolony diversity and distribution of zooxanthellae in acroporid corals is largely uncharted. In this study, two molecular methods were applied to determine the dis- tribution of zooxanthellae in the branching corals Acropora tenuis and A. valida at several reef locations in the central section of the Great Barrier Reef. Sun-exposed and shaded parts of all colonies were examined. Single-stranded conformational polymorphism analysis showed that individual colonies of A. tenuis at two locations harbour two strains of Symbi- odinium belonging to clade C (C1 and C2), whereas conspecific colonies at two other reefs harboured a single zooxanthella strain. A. valida was found to simultaneously harbour strains belonging to two distinct phylogenetic clades (C and D) at all locations sampled. A novel method with improved sensitivity (quantitative polymerase chain reaction using Taqman™ fluorogenic probes) was used to map the relative abundance distribution of the two zooxanthella clades. At two of the five sampling locations both coral species were collected. At these two locations, composition of the zooxanthella communities showed the same pattern in both coral species, i.e. correlation with ambient light in Pioneer Bay and an absence thereof in Nelly Bay. The results show that the distribution of genetically distinct zooxanthellae is correlated with light regime and possibly temperature in some (but not all) colonies of A. tenuis and A. valida and at some reef locations, which we interpret as acclimation to local environmental conditions. Keywords: Acropora, niche partitioning, real-time quantitative PCR, Symbiodinium, symbiosis, zooxanthellae Received 14 May 2003; revision received 22 August 2003; accepted 22 August 2003 Introduction Obligate associations between scleractinian corals and endosymbiotic dinoflagellates (zooxanthellae) are ubi- quitous in tropical reef systems. The role of zooxanthellae in the symbiosis is generally considered vital for the health of shallow tropical reef systems (Muscatine 1990; Glynn 1996; Brown 1997). One of the biggest threats to the health of coral reefs today is the increasing frequency of bleaching of hermatypic corals (whitening of corals due to loss of either symbiotic algae or their pigments, or both). In severe cases corals do not recover and subsequently die (Brown 1997; Hoegh-Guldberg 1999). The severity of the bleaching re- sponse differs greatly between species of corals (Marshall & Baird 2000; Loya et al . 2001) and even across individual colonies (Ralph et al . 2002). It also varies spatially on local and regional scales (Glynn 2001). These spatial differences may be caused by genetically distinct zooxanthellae be- longing to the genus Symbiodinium responding differently to specific local environments and thus affecting the degree of bleaching (Baker 2001; Toller et al . 2001b). Symbiodinium consists of several extremely divergent groups of taxa and, based on ribosomal DNA (rDNA), seven distinct phylogenetic clades have been distinguished (A– G) to date Correspondence: M. J. H. Van Oppen. Fax: +61 7 4772 5852; E-mail: [email protected]

Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

Embed Size (px)

Citation preview

Page 1: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

Molecular Ecology (2003)

12

, 3477–3484 doi: 10.1046/j.1365-294X.2003.01988.x

© 2003 Blackwell Publishing Ltd

Blackwell Publishing Ltd.

Geographic and habitat partitioning of genetically distinct zooxanthellae (

Symbiodinium

) in

Acropora

corals on the Great Barrier Reef

K . E . ULSTRUP

*

and M. J . H. VAN OPPEN

,

*

Department of Phycology, Botanical Institute, University of Copenhagen, Copenhagen, Denmark;

Australian Institute of Marine Science, PMB 3, Townsville MC, Qld 4810, Australia;

School of Pharmacy and Molecular Sciences, James Cook University, Townsville, Australia

Abstract

Intra- and intercolony diversity and distribution of zooxanthellae in acroporid corals islargely uncharted. In this study, two molecular methods were applied to determine the dis-tribution of zooxanthellae in the branching corals

Acropora tenuis

and

A. valida

at severalreef locations in the central section of the Great Barrier Reef. Sun-exposed and shaded partsof all colonies were examined. Single-stranded conformational polymorphism analysisshowed that individual colonies of

A. tenuis

at two locations harbour two strains of

Symbi-odinium

belonging to clade C (C1 and C2), whereas conspecific colonies at two other reefsharboured a single zooxanthella strain.

A. valida

was found to simultaneously harbourstrains belonging to two distinct phylogenetic clades (C and D) at all locations sampled. Anovel method with improved sensitivity (quantitative polymerase chain reaction usingTaqman™ fluorogenic probes) was used to map the relative abundance distribution of thetwo zooxanthella clades. At two of the five sampling locations both coral species werecollected. At these two locations, composition of the zooxanthella communities showed thesame pattern in both coral species, i.e. correlation with ambient light in Pioneer Bay and anabsence thereof in Nelly Bay. The results show that the distribution of genetically distinctzooxanthellae is correlated with light regime and possibly temperature in some (but notall) colonies of

A. tenuis

and

A. valida

and at some reef locations, which we interpret asacclimation to local environmental conditions.

Keywords

:

Acropora

, niche partitioning, real-time quantitative PCR,

Symbiodinium

, symbiosis,zooxanthellae

Received 14 May 2003; revision received 22 August 2003; accepted 22 August 2003

Introduction

Obligate association

s

between scleractinian corals andendosymbiotic dinoflagellates (zooxanthellae) are ubi-quitous in tropical reef systems. The role of zooxanthellae inthe symbiosis is generally considered vital for the health ofshallow tropical reef systems (Muscatine 1990; Glynn 1996;Brown 1997). One of the biggest threats to the health ofcoral reefs today is the increasing frequency of bleaching ofhermatypic corals (whitening of corals due to loss of either

symbiotic algae or their pigments, or both). In severe casescorals do not recover and subsequently die (Brown 1997;Hoegh-Guldberg 1999). The severity of the bleaching re-sponse differs greatly between species of corals (Marshall &Baird 2000; Loya

et al

. 2001) and even across individualcolonies (Ralph

et al

. 2002). It also varies spatially on localand regional scales (Glynn 2001). These spatial differencesmay be caused by genetically distinct zooxanthellae be-longing to the genus

Symbiodinium

responding differentlyto specific local environments and thus affecting the degreeof bleaching (Baker 2001; Toller

et al

. 2001b).

Symbiodinium

consists of several extremely divergent groups of taxaand, based on ribosomal DNA (rDNA), seven distinctphylogenetic clades have been distinguished (A–G) to date

Correspondence: M. J. H. Van Oppen. Fax: +61 7 4772 5852;E-mail: [email protected]

Page 2: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

3478

K . E . U L S T R U P and M . J . H . V A N O P P E N

© 2003 Blackwell Publishing Ltd,

Molecular Ecology

, 12, 3477–3484

(LaJeunesse 2001; Rodriguez-Lanetty 2003). Five of these(A–D, F) are known to occur in corals.

Individual colonies can associate with several genetic-ally distinct symbionts simultaneously (Rowan & Knowlton1995; van Oppen

et al

. 2001). However, known examplesof mixed zooxanthella communities are limited (Rowan& Knowlton 1995; Toller

et al

. 2001a; van Oppen

et al

. 2001;Diekmann

et al

. 2002; LaJeunesse 2002). Although somestudies have found a lack of correlation between

Symbiod-inium

phylotypes and photo-physiology (Savage

et al

.2002), the distribution of zooxanthellae in some coral spe-cies correlates with light intensity and quality on differentspatial scales. For example, Rowan & Knowlton (1995) andRowan

et al

. (1997) found light-dependent distributions ofzooxanthella phylotypes within colonies of

Montastraea

spp., and van Oppen

et al

. (2001) reported distinct strainsbelonging to clade C (based on rDNA) in sun-exposed andshaded regions of

Acropora tenuis

. Light intensity andspectral composition are thus suggested to affect the com-position of the zooxanthella community within individualcolonies. Hence, corals appear to be able to acclimatize todifferent light regimes at intracolony spatial scales to limitthe extent of (light-induced) bleaching or even preventbleaching by hosting multiple genetically distinct zoo-xanthellae (Buddemeier & Fautin 1993; Baker 2001). Insuch cases, zooxanthellae are thought to be photo-adaptedrather than photo-acclimated (

sensu

Helmuth

et al

. 1997).Genetically distinct zooxanthellae have been found to

vary with regard to photo-physiological properties inculture (Chang

et al

. 1983; Iglesias-Prieto & Trench 1994;Santos

et al

. 2001). Similar studies of photo-physiologicalproperties of genetically distinct zooxanthellae

in hospite

are rare (Bhagooli & Hidaka 2003). In this study, we showthat algal symbionts from two acroporid coral speciesexhibit some niche partitioning that correlates with lightintensity and possibly temperature. The study is basedon examination of the distribution of zooxanthellae insun-exposed and shaded areas of

A. tenuis

and

A. valida

colonies collected over a geographical scale of 50–300 kmcovering several reefs that differ greatly in water clarity. In

A. valida

, we applied a novel method with improved sens-itivity [a real-time polymerase chain reaction (PCR) assay]to accurately estimate relative abundances of zooxanthel-lae belonging to different phylogenetic clades in distinctlight climates across coral colonies.

Methods and materials

Sampling

Ten colonies of

Acropora tenuis

were collected from each offour reefs and ten colonies of

A. valida

were collected fromthree reefs (Fig. 1). Sun-exposed and shaded branches ofeach colony were sampled and preserved in absolute

ethanol for genetic analysis. Black Reef (19

°

40

S, 149

°

20

E)and Chicken Reef (18

°

35

S, 147

°

45

E) (Fig. 1) are mid-shelfreefs with relatively high water clarity. Pioneer Bay(18

°

35

S, 146

°

20

E), West Point (19

°

16

S, 146

°

49

E) andNelly Bay (19

°

15

S, 146

°

50

E) (Fig. 1) are inshore, fringingreefs with relatively lower water clarity. The latter twosites are located around Magnetic Island, which is situatedclose to Townsville in north Queensland. The water qualityhere is particularly affected by several external physicalfactors, the most important of which are coastal develop-ment and by-passing of vessels to Townsville harbourcontributing to turbidity. Water clarity may vary furtherbecause of variable resuspension of sediments by currents,wind and wave action. In combination with depth, thesefactors are highly important for the light attenuation inthe water column. The distances between the reefs are

Fig. 1 Map of the five sampling sites, Pioneer Bay, Nelly Bay,West Point, Black Reef and Chicken Reef indicating whereAcropora tenuis and A. valida were collected. Black Reef andChicken Reef are mid-shelf reefs, whereas Pioneer Bay is aninshore and intermediate turbid fringing reef and Nelly Bay andWest Point are inshore and very turbid fringing reefs.

Page 3: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

N I C H E P A R T I T I O N I N G O F Z O O X A N T H E L L A E I N C O R A L S

3479

© 2003 Blackwell Publishing Ltd,

Molecular Ecology

, 12, 3477–3484

between 50 and 300 km (Fig. 1). Sampling depth at allreefs was

3 m at the astronomical high tide. Nelly Bayis generally somewhat more turbid than Pioneer Baybecause of its proximity to Townsville. The turbidity,however, may be subject to seasonal variation. All coralswere collected during the summer months (January–February 2002) except for

A. tenuis

colonies in Pioneer Bay,which were collected in September 2001 and

A. valida

colonies at West Point, which were collected in August2001.

Qualitative distribution of zooxanthellae in

Acropora tenuis

DNA extractions were carried out using the DNeasy™Tissue Kit (Qiagen) following the instructions provided inthe DNeasy™ protocol for DNA extraction of animaltissues. DNA was eluted twice from the spin column, oncewith 200

µ

L of elution buffer and a second time with150

µ

L. One microlitre of the second elution was used in a25-

µ

L PCR. Primers specific to zooxanthellae were usedin PCR to amplify rDNA internal transcribed spacer 1(ITS1). The ITS1 marker enables differentiation of zoo-xanthellae at subspecies (strain) level (Hunter

et al

. 1997; vanOppen

et al

. 2001; Rodriguez-Lanetty 2003). The forwardprimer, SymITSFP (5

-CTCAGCTCTGGACGTTGYGTTGG-3

), is located 105 bp upstream of ITS1 in the SSU rDNAgene, the reverse primer, SymITSRP (5

-TATCGCRCTTC-RCTGCGCCCT-3

), 11 bp downstream in the 5.8

S

rDNAregion (van Oppen

et al

. 2001). The total length of the frag-ment is

380 bp including the primer binding regions. ThePCR profile for amplification of the ITS1 region was 3 minat 94

°

C, 30 cycles of 30 s at 94

°

C, 30 s at 59

°

C and 30 s at72

°

C followed by 5 min at 72

°

C.

Single-stranded conformation polymorphism (SSCP)

To distinguish between the

Symbiodinium

C1 and C2 strain,DNA products were amplified using PCR as describedabove using a fluorescent Hex-labelled forward primer.The PCR product was mixed with formamide gel-loadingbuffer in a 1:3 v/v ratio. Samples were denatured at 95

°

Cfor 3 min and snap-cooled on ice for at least 3 min. Onemicrolitre of PCR product was pulse-loaded onto a 4%nondenaturing TBE-polyacrylamide gel (20 cm) for 25 s,followed by flushing of the well. Approximately 15 mLof solution was used for each gel (1.5 mL 10% acrylamide(37:1 bisacrylamide/acrylamide; SIGMA®); 12.925 mLddH

2

O; 1.2 mL 10

×

TBE; 0.375 mL 80% glycerol). Thirtymicrolitres of TEMED (SIGMA®) and 75

µ

L 10% ammoniumpersulphate (APS) were added to initiate polymerizationof the gel. Gels were run on the Gelscan 2000 (CorbettResearch) for 40 min (1200 V, 22

°

C) with 0.8

×

TBE-buffer(AMRESCO®).

Relative quantification of zooxanthellae in

Acropora valida

using real-time PCR

Real-time PCR detection on the ABI 7700 Prism SequenceDetection System (Applied Biosystems) was used to detectthe relative abundance of clade C vs. clade D in

A. valida

.Real-time PCR quantifies the amount of PCR productsynthesized as fluorescence. Primers are designed toamplify a short (

100 bp) PCR product (to ensure efficientPCR amplification) and a dual-labelled (a reporter and aquencher dye) probe anneals in between the forward andreverse primer. As long as the reporter and quencherdyes are in close proximity, the quencher suppresses thereporter’s fluorescent signal. The AmpliTaq Gold DNAPolymerase (Applied Biosystems) used in the PCR has 5

exonuclease activity and cleaves the probe during DNAstrand extension. This causes the reporter dye to comefree in solution and fluorescence can be detected uponexcitation by a light source. The amount of fluorescence isa direct measure of the amount of PCR product, as meas-ured in the exponential PCR phase. Quantitative real-timePCR is, therefore, more sensitive and allows much moreprecise quantification compared to conventional PCR.

The assay used in this study is based on nuclear rDNA.Two primer pairs and two dual-labelled probes are usedin the same reactions. One primer–probe set serves as acalibrator to correct for the amount of DNA added to thereaction as well as possible rDNA copy number differencesbetween different individuals and strains (Fig. 2). Thesecond primer–probe combination detects different zoo-xanthella clades specifically. The fluorescent dyes in thecalibrator and clade-specific reactions are VIC and FAM,respectively, the quencher dye is TAMRA. To design primersand probes, SSU sequences from zooxanthella clades A–Dwere downloaded from GenBank and aligned manually.Primers and probes were designed in

primer express

Version 1.5 (Applied Biosystems) to meet the ABI 7700real-time PCR amplification conditions and they weresubsequently blasted against the database to verify thatannealing with coral DNA would not occur. The calibratorprimers and probe anneal in the SSU rDNA gene (positions1109–1130, 1139–1160 and 1178–1199 in the

Symbiodinium

sequence from

Montipora verrucosa

GenBank Accessionno. AB016594, for the forward primer, the probe and the

Fig. 2 Schematic overview of the partial ribosomal cistronindicating the location of clade-specific primers (U-FP, C-RP andD-RP), the calibrator primers (cal-FP and cal-RP) and the twoprobes (cal-probe and ABCD probe).

Page 4: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

3480

K . E . U L S T R U P and M . J . H . V A N O P P E N

© 2003 Blackwell Publishing Ltd,

Molecular Ecology

, 12, 3477–3484

reverse primer, respectively). For the clade-specificamplifications, a ‘universal’ forward primer and probewere designed in combination with a clade-specific reverseprimer (Fig. 2, Table 1). The ‘universal’ FP aligns to the SSUrDNA gene at positions 1745–1769 and the A-B-C-D (i.e.‘universal’) probe anneals to position 1776 of the SSUgene to position 1 of the ITS1 region of the

Symbiodinium

sequence from

Montipora verrucosa

(GenBank Accessionno. AB016594). The C-specific reverse primer anneals topositions 29–48 of the

Symbiodinium

sequence from

Hippopushippopus

(GenBank Accession no. AF333518). The D-specificreverse primer anneals to positions 9–38 of the

Symbiodinium

sequence from

Montastraea annularis

(GenBank Accessionno. AF334660).

For analysis of the data a fluorescence threshold of 0.04was chosen, based on the variability of the baseline in testruns. The cycle-threshold (C

T

) cycle at which the fluores-cence of a sample exceeds the chosen threshold limit wasdetermined for the calibrator and clade-specific PCR ofeach sample. The C

T

for the calibrator is always lower thanor equal to that of the clade-specific reaction (i.e. more orequal number of copies present, respectively, comparedwith the clade-specific amplification product, in the caseof more than one or one zooxanthella clade present in thetissue). C

T

values decrease linearly with increasing inputtarget quantity and can be used as a quantitative measure-ment of the input target (Giulietti

et al

. 2001). The endo-genous control (calibrator) and target assays must haveidentical PCR efficiencies for the relative quantificationmethod used here (the ∆∆CT method, see below) to bevalid. This was tested by serially diluting a DNA sampleand demonstrating that the CT difference between thetarget and the endogenous control remained constant. Aconstant CT difference across a range of at least three logs(1000-fold) of initial template concentration verifies iden-tical PCR efficiency (Applied Biosystems SDS User Bulletin#2). The calibrator reaction was primer limited to avoidreaction components running out before the clade-specificreaction had reached the exponential phase. PCRs wererun in triplicate along with no template controls. The meanCT value was used to calculate relative abundance. The

calibrator CT value was generally ≈ 18–20, the clade-specificCT value tended to vary between 20 and 35, with a smallnumber of samples showing lower or higher CT values.The standard deviation of the CT value of each three repli-cates was in the order of 0.1–0.2.

The method used to quantify the relative abundancesof clade D vs. clade C is the comparative threshold (CT)method. The amount of clade D or clade C target within asample, normalized to the calibrator, is given by ∆CT = CT-value of the calibrator subtracted from CT of the target. Theclade D/C ratio is then calculated by calculating 2–∆∆CT,where ∆∆CT = ∆CT of clade D minus ∆CT of clade C. Thedifference represents the normalized abundance of thetarget gene in the unknown sample, relative to the norm-alized abundance of the calibrator samples (AppliedBiosystems SDS User Bulletin #2).

DNA extraction for A. valida was carried out as forA. tenuis using the Qiagen DNeasy™ Kit using standardprocedures. An aliquot of 2.5 µL of the second eluate fromthe Qiagen spin columns was used to amplify the rDNAregions in a 25-µL reaction (Table 1). The PCR reagentsfor one 25 µL real-time PCR were added in the follow-ing volumes: 12.5 µL of TaqMan® Universal PCR MasterMix, 2.5 µL (1.8 µm) of universal forward primer, 2.5 µL ofclade-specific reverse primer (1.8 µm), 1.25 µL of calibratorforward primer (1.8 µm), 1.25 µL of calibrator reverseprimer (1.8 µm), 1 µL of FAM probe (2.5 µm), 1 µL of VICcalibrator probe (2.5 µm), 0.5 µL ddH2O and 2.5 µL of DNAtemplate. The PCR profile for real-time detection of PCRruns for 2 min at 50 °C, 10 min at 95 °C (AmpliTaq Goldpreactivation), 40 cycles of 10 s at 95 °C (denaturation), and1 min at 60 °C (annealing and extension).

Statistical analysis

The occurrence of Symbiodinium genotypes C1 and C2from sun-exposed and shaded surfaces in A. tenuis wasidentified from SSCP gelscan images. The distribution ofzooxanthellae in different localities and intracolony lighthabitats was analysed statistically using χ2 tests.

The replicate calibrator amplifications of clade C andclade D of each A. valida sample were compared usinga paired t-test to test whether the clade-specific primersaffect the calibrator reaction and to control for pipettingerrors. One-way analysis of variance (anova) was usedto determine if significant differences were present inthe abundance of zooxanthellae belonging to clade C andclade D in shaded and sun-exposed surfaces of coloniesfrom each of the three individual sampling sites. More-over, the ratio of the amount of Symbiodinium D and C (i.e.D/C) in shaded and sun-exposed surfaces was comparedbetween sampling sites also using one-way anova. Thedata for these tests were log-transformed to meet theassumptions of normality and equal variance. The statistical

Table 1 DNA sequence of PCR primers and probes for quantitativeamplification of distinct Symbiodinium clades

Primer Sequence

Universal FP 5′-AAGGAGAAGTCGTAACAAGGTTTCC-3′C-specific RP 5′-AAGCATCCCTCACAGCCAAA-3′D-specific RP 5′-CACCGTAGTGGTTCACGTGTAATAG-3′Calibrator FP 5′-GTATGGTCGCAAGGCTGAAACT-3′Calibrator RP 5′-TTTCCCCGTGTTGAGTCAAATT-3′A B C D probe 5′-TGCGAATGATCCTTCCGCAGGTTC-3′Calibrator probe 5′-CCTGGTGGTGCCCTTCCGTCAA-3′

Page 5: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

N I C H E P A R T I T I O N I N G O F Z O O X A N T H E L L A E I N C O R A L S 3481

© 2003 Blackwell Publishing Ltd, Molecular Ecology, 12, 3477–3484

analyses were performed with spss 6.1 using a significancelevel of 0.05 to identify significant differences.

Results

Habitat partitioning of zooxanthella strains in Acropora tenuis

The sun-exposed parts of the 40 colonies sampled in eachof the four reef locations harboured a homogeneouspopulation of zooxanthellae with only one strain present.Thirty-eight of the 40 shaded coral samples also harboureda homogeneous population of zooxanthellae with a singlestrain present. In 2 samples of a total of 40 shaded coralfragments (both from Pioneer Bay), both zooxanthellastrains were found simultaneously (Table 2). The hetero-geneous zooxanthella population found in these twosamples may be the result of coral sampling in positions atthe boundary between sun-exposed and shaded lightclimates.

In Pioneer Bay, 6 of 10 and at Chicken Reef 9 of 10colonies were found to harbour only Symbiodinium C2 insun-exposed samples, whereas the remaining coloniesharboured Symbiodinium C1. In shaded regions, 7 of 10 col-onies in Pioneer Bay and 8 of 10 at Chicken Reef harbouredC1 zooxanthellae (Table 2). Hence, at these two locationsthe two zooxanthella strains are generally distributed inaccordance with the two distinct light microhabitats acrossindividual Acropora tenuis corals (χ2 test, d.f. = 1, P < 0.01).

Black Reef corals harboured a single zooxanthella strain(C2) independent of sun-exposed and shaded light cli-mates across the corals (Table 2). The environment at BlackReef is relatively pristine and similar to that of ChickenReef where two strains of zooxanthellae were found dis-tributed according to light climates. All but 1 of 10 coloniesfrom Nelly Bay also harboured a single homogeneouscommunity of zooxanthellae (C1) independent of sam-pling position on the corals. The one coral from Nelly Baythat harboured more than one strain did so in a light-dependent mode where C2 was harboured in sun-exposedsurfaces and C1 in shaded surfaces, similar to A. tenuiscolonies from Pioneer Bay and Chicken Reef (Table 2).

Although A. tenuis colonies from Black Reef and Nelly Bayharboured mainly a single strain, the zooxanthella strainswere different at the two localities. A. tenuis from the rela-tively pristine Black Reef harboured C2, whereas coralsfrom the ≈ 250 km removed turbid environment of NellyBay harboured mainly C1 (Table 2, Fig. 1).

Relative abundance of clade D vs. clade C zooxanthellae in Acropora valida

Real-time PCR in combination with SSCP analysis wasused to identify and estimate relative abundance ofdifferent clades present in the A. valida colonies at threelocations. Using clade A-, B-, C- and D-specific primers inreal-time PCR and the general ITS1 primers in SSCP, thecoral samples were found to harbour only zooxanthellaebelonging to clade C and D. This correlates with the reportof van Oppen et al. (2001) who found clade C and clade Din A. valida colonies from Pioneer Bay.

For the quantitative PCR assay, we first tested whetherthe efficiency of the PCR amplification was the same for thecalibrator and target amplifications, by plotting the ∆CTvalue for three serial dilutions against the log of the relativetemplate concentrations. The slope of the obtained linearregression line was −0.1117, which is quite acceptable forthe assay used here, as we are only interested in orders ofmagnitude differences in relative abundances of distinctzooxanthella clades. Next, we tested whether the efficiencyof the calibrator reaction was the same in the clade C andclade D reaction for each sample by comparing the trip-licate CT values of the calibrator amplification in the tworeactions using a paired t-test. In case the CT of the calib-rator reaction was significantly different in the C and D reac-tion, the sample was excluded from further analysis. Thismethod allowed us to include data from sun-exposed andshaded sides from 7 of 10 sampled colonies from WestPoint, 8 of 10 from Nelly Bay and all 10 from Pioneer Bay.The mean relative abundance of clade D vs. clade C (i.e. thenumber of clade D copies per one copy of clade C) fromsun-exposed and shaded light climates across individualcolonies is shown in Fig. 3. Statistical analysis of the meanrelative abundance of clade D vs. clade C shows that these

Col. # Light climate 1 2 3 4 5 6 7 8 9 10

PB* Sun-exposed C2 C2 C2 C2 C2 C1 C1 C1 C1 C2PB* Shaded C1 C1 C1 C1 C1 C1/C2 C1/C2 C1 C1 C2NB Sun-exposed C1 C1 C1 C1 C1 C1 C1 C1 C1 C2NB Shaded C1 C1 C1 C1 C1 C1 C1 C1 C1 C1BR Sun-exposed C2 C2 C2 C2 C2 C2 C2 C2 C2 C2BR Shaded C2 C2 C2 C2 C2 C2 C2 C2 C2 C2CR* Sun-exposed C2 C2 C2 C2 C2 C2 C2 C2 C2 C1CR* Shaded C1 C1 C1 C1 C1 C1 C1 C2 C2 C1

Table 2 Distribution of the two Symbio-dinium strains (C1 and C2 sensu van Oppenet al. 2001) within and among colonies ofAcropora tenuis from Pioneer Bay (PB),Nelly Bay (NB), Black Reef (BR) andChicken Reef (CR). Statistically significant*evidence (P < 0.01) exists for a lightdependent distribution of C1 and C2 inPioneer Bay and Chicken Reef

Page 6: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

3482 K . E . U L S T R U P and M . J . H . V A N O P P E N

© 2003 Blackwell Publishing Ltd, Molecular Ecology, 12, 3477–3484

are significantly different between all sampling locations,where both West Point and Nelly Bay have a significantlyhigher ratio of clade D vs. clade C compared with PioneerBay (P < 0.005). West Point has a lower ratio of clade D vs.clade C than Nelly Bay (P = 0.021). In Pioneer Bay the meanrelative abundance of the two clades is correlated withsun-exposed and shaded climates across colonies (P =0.046), in that Symbiodinium C is significantly less abundantin shaded areas of the colonies than in sun-exposed areasand Symbiodinium D is significantly more abundant inshaded areas than in sun-exposed areas. No distinction inD/C ratio was observed between light climates in theNelly Bay or West Point populations.

Discussion

Acropora tenuis colonies in both Pioneer Bay and ChickenReef, which are separated by > 200 km, show a distributionpattern in which C2 (sensu van Oppen et al. 2001) is foundin sun-exposed regions and C1 (sensu van Oppen et al.2001) in shaded regions of the corals (P < 0.01). This pre-ference in light climate of the two algal strains belonging tophylotype C suggests that these possess different photo-physiological characteristics that affect their abilities tocompete for light and hence increase the fitness of thecolonies. However, this apparent niche partitioning is notpresent at all locations sampled or within all colonies at agiven location. Colonies sampled from Black Reef harbourSymbiodinium C2 only and those in Nelly Bay harbourmainly Symbiodinium C1 (Table 2). These observationsindicate that the coral–algal relationship of A. tenuis is not

stable on a regional geographical scale in the centralsection of the Great Barrier Reef. The lack of strain C1 atBlack Reef is surprising as the light environment at thislocation is estimated to be intermediate to that of PioneerBay and Chicken Reef where both strains are present(Table 1). Also, A. tenuis colonies in Pioneer Bay, whichharbour both strains, and A. tenuis in Nelly Bay, whichassociate mostly with strain C1 are only ≈ 50 km apart(Fig. 1). This relatively short geographical distance suggeststhat the abundance of certain zooxanthella genotypes inhospite could be a consequence of local availability. It haspreviously been suggested that zooxanthellae may occurin a patchy spatial distribution across regional distances(Coffroth et al. 2001; van Oppen et al. 2001). Because BlackReef (where C1 is lacking in the sampled colonies) is thesouthern-most of the four locations (Fig. 1), other environ-mental factors such as seasonal sunlight and temperat-ure may also affect the presence of certain Symbiodiniumgenotypes. The lack of Symbiodinium C1 in A. tenuis atBlack Reef may thus be a consequence of latitude. Similarlatitudinal restrictions in the occurrence of certain zoo-xanthellae in corals have been reported previously. Forexample, Rodriguez-Lanetty et al. (2001) found that thecoral Plesiastrea versipora in the west Pacific harbours cladeB symbionts at high latitude and clade C symbionts at lowlatitude, and LaJeunesse & Trench (2000) observed that thesea anemone Anthopleura elegantissima harbours differentassemblages of zooxanthellae across a latitudinal gradientoff the west coast of North America.

At all three locations, A. valida harbours clade C andclade D simultaneously in all samples in both sun-exposedand shaded parts of the colonies (Fig. 3). The abundance ofclade D vs. clade C in A. valida colonies in Pioneer Bay isstrongly correlated with the ambient light regime acrosscolonies, which suggests that the two clades exhibit somedegree of niche partitioning depending on the light climateacross individual colonies and on the particular photo-synthetic characteristics of the algae involved. This partition-ing does not exist in Nelly Bay and West Point (MagneticIsland). A. tenuis was also sampled from Pioneer Bay andNelly Bay, and shows the same pattern of zooxanthellacommunity composition at these reefs, i.e. a light-relatedpattern in Pioneer Bay and a lack thereof in Nelly Bay. Thisindicates that there may be a lack of significant light dif-ferentiation across colonies at Magnetic Island probablydue to excessive turbid conditions.

The mean relative abundance of clade D, is highest inNelly Bay. At West Point, as well as at Nelly Bay, the meanrelative abundance ratio (D/C) is > 1 indicating that cladeD, is more abundant than clade C. In Pioneer Bay, the meanrelative abundance of Symbiodinium D vs. Symbiodinium Cis lowest and is < 1, indicating that clade C is more abund-ant than clade D. As Pioneer Bay is an intermediatelyturbid environment and Nelly Bay and West Point are very

Fig. 3 The mean relative abundance of Symbiodinium clade D vs.clade C (D/C ratio) in sun-exposed and shaded surfaces ofAcropora valida at West Point (n = 7), Nelly Bay (n = 8) and PioneerBay (n = 10). (�) Sun-exposed sides, (�) shaded sides of coralcolonies. The y-axis has a logarithmic scale. Vertical bars arestandard errors.

Page 7: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

N I C H E P A R T I T I O N I N G O F Z O O X A N T H E L L A E I N C O R A L S 3483

© 2003 Blackwell Publishing Ltd, Molecular Ecology, 12, 3477–3484

turbid, clade C could be characterized as ‘light-loving’ andclade D as ‘shade-loving’ genotypes of Symbiodinium inA. valida. However, other environmental factors, such astemperature or nutrition, may play a role as well. Glynnet al. (2001) have shown that far-eastern Pacific corals har-bouring symbionts of clade D had suffered less from the1998 coral bleaching event than those that associated withsymbionts of a different genotype, suggesting that zooxan-thellae of clade D increase the thermal tolerance of the coralhost. The water temperature around Magnetic Island(i.e. the Nelly Bay and West Point locations) is usually afew degrees higher than that around Orpheus Island (i.e.the Pioneer Bay location) and this may partly explain thehigher relative abundance of Symbiodinium D in theMagnetic Island corals.

The occurrence of niche partitioning of genetically dis-tinct zooxanthellae in coral colonies with distinct lighthabitats across colonies has been suggested to be an adaptivetrait of scleractinian corals to optimize photosynthesis ofzooxanthellae and thus metabolism of the holobiont (Baker2001; Goulet & Coffroth 2003). However, examples ofsymbiotic relationships between host and multiple endo-symbionts simultaneously are rare (Diekmann et al. 2002;LaJeunesse 2002), which may be a consequence of the rel-atively low sensitivity of the genetic methods commonlyused, as ‘background’ strains usually go undetected[strains below a relative abundance of 5–10% are usuallynot detected [(van Oppen personal observation; LaJeunesse,personal communication)]. Nevertheless, in this studyit has been shown that branching corals may have suchendosymbiotic populations at different genetic levelsof strains and clades, which may vary in distribution andabundance in relation to the ambient light regime inthe water column. The simultaneous presence of multiplezooxanthella strains as observed in the two Acropora spe-cies examined is not ubiquitous in the central section of theGreat Barrier Reef, which may be related to a lack of signi-ficant zonation in light climates across colonies (Goulet &Coffroth 2003) in some locations, a patchy distribution ofspecific zooxanthella genotypes over regional distances(Coffroth et al. 2001), and host–symbiont specificity.

Conclusion

This study suggests that in coral–algal symbioses partnerscan be acclimated in coexistence (i.e. at the level of theholobiont) to a specific light environment. The apparentpreference toward distinct optimal symbiont complexes indifferent light environments either within single coloniesor on different reefs with differing turbidity reducesthe likelihood of light-related stress and maximizes coralphotosynthesis, and hence the energy balance of the holo-biont. The results obtained here suggest that preferenceto light is likely to be genetically determined and that

symbionts belonging to clade C are more light-loving thanthose belonging to clade D, and within clade C, Sym-biodinium C2 is more light-loving than Symbiodinium C1.However, other environmental factors, such as temperatureor nutrients, may play an important role in the distributionof Symbiodinium strains as well.

Acknowledgement

We wish to acknowledge the many useful discussions with andhelp by all members of the Miller Laboratory, ComparativeGenomics Centre, Molecular Sciences Building, James Cook Uni-versity. Special thanks go to Dr David Miller. We also thank DavidRutter from Applied Biosystems for his help in the design ofprimers and probes for the quantitative PCR assay. This work wassupported by the Australian Research Council.

References

Baker AC (2001) Reef corals bleach to survive change. Nature, 411,765–766.

Bhagooli R, Hidaka M (2003) Comparison of stress susceptib-ility of in hospite and isolated zooxanthellae among five coralspecies. Journal of Experimental Marine Biology and Ecology, 291,181–197.

Brown BE (1997) Coral bleaching: causes and consequences. CoralReefs, 16, 129–138.

Buddemeier RW, Fautin DG (1993) Coral bleaching as an adaptivemechanism. Bioscience, 48, 320–326.

Chang SS, Prezelin BB, Trench RK (1983) Mechanisms of photoadap-tion in three strains of the symbiotic dinoflagellate Symbiodiniummicroadriaticum. Marine Biology, 76, 219–229.

Coffroth MA, Santos SR, Goulet TL (2001) Early ontogeneticexpression of specificity in a cnidarian–algal symbiosis. MarineEcology Progress Series, 222, 85–96.

Diekmann OE, Bak RPM, Tonk L et al. (2002) No habitat correla-tion of zooxanthellae in the coral genus Madracis on a Curaaoreef. Marine Ecology Progress Series, 227, 221–232.

Giulietti A, Overbergh L, Valckx D et al. (2001) An overview ofreal-time quantitative PCR. Applications to quantify cytokinegene expression. Methods, 25, 386–401.

Glynn PW (1996) Coral reef bleaching: fact, hypothesis and implica-tions. Global Change Biology, 2 (6), 495–509.

Glynn PW (2001) A collection of studies on the effects of the 1997–98 El Niño–Southern Oscillation events on corals and coral reefsin the eastern tropical Pacific — Preface. Bulletin of Marine Science,69 (1), 1–4.

Glynn PW, Mate JL, Baker AC et al. (2001) Coral bleaching andmortality in Panama and Equador during the 1997–98 ElNino–Southern Oscillation event: spatial/temporal patternsand comparisons with the 1982–83 event. Bulletin of MarineScience, 69 (1), 79–109.

Goulet TL, Coffroth MA (2003) Genetic composition of zooxan-thellae between and within colonies of the octocoral Plexaurakuna, based on small subunit rDNA and multilocus DNAfingerprinting. Marine Biology, 142, 233–239.

Helmuth BST, Timmerman BEH, Sebens KP (1997) Interplay ofhost morphology and symbiont microhabitat in coral aggrega-tions. Marine Biology, 130, 1–10.

Page 8: Geographic and habitat partitioning of genetically distinct zooxanthellae (Symbiodinium) in Acropora corals on the Great Barrier Reef

3484 K . E . U L S T R U P and M . J . H . V A N O P P E N

© 2003 Blackwell Publishing Ltd, Molecular Ecology, 12, 3477–3484

Hoegh-Guldberg O (1999) Climate change, coral bleaching andthe future of the world’s coral reefs. Marine and FreshwaterResearch, 50, 839–866.

Hunter CL, Morden CW, Smith CM (1997) The utility of ITSsequences in assessing relationships among zooxanthellae andcorals. Proceedings of the 8th International Coral Reef Symposium, 2,1599–1602.

Iglesias-Prieto R, Trench RK (1994) Acclimation and adaptation toirradiance in symbiotic dinoflagellates. I. Responses of thephotosynthetic unit to changes in photon flux density. MarineEcology Progress Series, 113 (1–2), 163–175.

LaJeunesse TC (2001) Investigating the biodiversity, ecology, andphylogeny of endosymbiotic dinoflagellates in the genus Sym-biodinium using the ITS region: in search of a ‘species’ levelmarker. Journal of Phycology, 377, 886–880.

LaJeunesse TC (2002) Diversity and community structure of sym-biotic dinoflagellates from Caribbean coral reefs. Marine Biology,141, 387–400.

LaJeunesse TC, Trench RK (2000) Biogeography of two species ofSymbiodinium (Freudenthal) inhabiting the intertidal sea anemoneAnthopleura elegantissima (Brandt). Biological Bulletin, 199, 126–134.

Loya Y, Sakai K, Yamazato K et al. (2001) Coral bleaching: thewinners and the losers. Ecology Letters, 4 (2), 122–131.

Marshall PA, Baird AH (2000) Bleaching of corals on the Great Bar-rier Reef: differential susceptibilities among taxa. Coral Reefs, 19,155–163.

Muscatine L (1990) The role of symbiotic algae in carbon andenergy flux in reef corals. In: Ecosystems of the World 25: CoralReefs (ed. Dubinsky Z), pp. 75–88. Elsevier, New York.

van Oppen MJH, Palstra FP, Piquet AM-T, Miller DJ (2001)Patterns of coral–dinoflagellate associations in Acropora: signi-ficance of local availability and physiology of Symbiodiniumstrains and host-symbiont selectivity. Proceedings of the RoyalSociety of London B., 268, 1759–1767.

Ralph PJ, Gademann R, Larkum AWD et al. (2002) Spatial hetero-geneity in active chlorophyll fluorescence and PSII activity ofcoral tissues. Marine Biology, 141, 639–646.

Rodriguez-Lanetty M (2003) Evolving lineages of Symbiodinium-like dinoflagellates based on ITS1 rDNA. Molecular Phylogeneticsand Evolution, 28, 152–168.

Rodriguez-Lanetty M, Loh W, Carter D, Hoegh-Guldberg O(2001) Latitudinal variability in symbiont specificity within thewidespread scleractinian coral Plesiastrea versipora. MarineBiology, 138, 1175–1181.

Rowan R, Knowlton N (1995) Intraspecific diversity and ecolo-gical zonation in coral–algal symbiosis. Proceedings of the NationalAcademy of Sciences of the USA, 92, 2850–2853.

Rowan R, Knowlton N, Baker A, Jara J (1997) Landscape ecologyof algal symbionts creates variation in episodes of coral bleach-ing. Nature, 388, 265–269.

Santos SR, Taylor DJ, Coffroth MA (2001) Genetic comparisons offreshly isolated versus cultured symbiotic dinoflagellates:implications for extrapolating to the intact symbiosis. Journal ofPhycology, 37 (5), 900–912.

Savage AM, Trapido-Rosenthal H, Douglas AE (2002) On thefunctional significance of molecular variation in Symbiodinium,the symbiotic algae of Cnidaria: photosynthetic response toirradiance. Marine Ecology Progress Series, 244, 27–37.

Toller WW, Rowan R, Knowlton N (2001a) Zooxanthellae of theMontastraea annularis species complex: patterns of distributionof four taxa of Symbiodinium on different reefs and acrossdepths. Biological Bulletin, 201, 348–359.

Toller WW, Rowan R, Knowlton N (2001b) Repopulation of zoo-xanthellae in the Caribbean corals Montastraea annularis andM. faveolata following experimental and disease-associatedbleaching. Biological Bulletin, 201, 360–373.

Karin Ulstrup conducted this research study as part of her MScdegree through the Department of Phycology, Botanical Institute,University of Copenhagen (Denmark). Her interest lies in the fieldof photo-physiology and genetics of the algal endosymbionts ofcorals. Madeleine van Oppen leads the ‘Risk and Recovery’research team at the Australian Institute of Marine Science(Townsville, Australia). Her research focuses on the adaptationand acclimatization of reef corals to increasing seawater tem-peratures, the use of genetics to aid in the design of MarineProtected Areas, the identification of marine stingers and theevolutionary genetics of reef corals.