closely related symbiodinium spp. differ in relative

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 284: 147–161, 2004 Published December 21

INTRODUCTION

The development and employment of moleculargenetic tools has initiated a new era in ecological, evo-lutionary and systematic discovery. This is especiallytrue for microscopic and morphologically cryptic orga-

nisms such as dinoflagellate endosymbionts (zooxan-thellae) that promote the survival and growth of stonycorals, soft corals, and related species that constructand/or thrive on tropical reefs (Rowan & Powers 1991,Rowan et al. 1997, LaJeunesse 2001, 2002, Baker 2003).Ribosomal gene sequence comparisons indicate that

© Inter-Research 2004 · www.int-res.com*Email: [email protected]

Closely related Symbiodinium spp. differ in relativedominance in coral reef host communities across

environmental, latitudinal and biogeographicgradients

T. C. LaJeunesse1, 2, 6,*, R. Bhagooli3, M. Hidaka3, L. deVantier4, T. Done4, G. W. Schmidt1, W. K. Fitt2, O. Hoegh-Guldberg5

1Department of Plant Biology, Plant Sciences Building, 2Institute of Ecology, University of Georgia, Athens, Georgia 30602, USA3Department of Chemistry, Biology and Marine Sciences, University of the Ryukyus, Nishihara, Okinawa 903-0213, Japan

4Australian Institute of Marine Science, Private Mail Box 3, Townsville, Queensland 4810, Australia5Centre for Marine Studies, University of Queensland, St. Lucia, Queensland 4072, Australia

6Present address: Department of Biology, Florida International University, University Park Campus, OE 245 11200SW 8th Street, Miami, Florida 33199, USA

ABSTRACT: The diversity and community structures of symbiotic dinoflagellates are described fromreef invertebrates in southern and central provinces of the Great Barrier Reef (GBR), Australia, andZamami Island, Okinawa, Japan. The symbiont assemblages from region to region were dominatedby Clade C Symbiodinium spp. and consisted of numerous host-specific and/or rare types (special-ists), and several types common to many hosts (generalists). Prevalence in the host community amongcertain host–generalist symbionts differed between inshore and offshore environments, across latitu-dinal (central versus southern GBR) gradients, and over wide geographic ranges (GBR versusOkinawa). One particular symbiont (C3h) from the GBR had a dramatic shift in dominance. Its pre-valence ranged from being extremely rare, or absent on high-latitude reefs to dominating thescleractinian diversity on a mid-latitude inshore reef. These changes occurred among coral faunawhose larvae must acquire symbionts from environmental sources (horizontal symbiont acquisition).Such differences did not occur among ‘vertical transmitters’ such as Porites spp., Montipora spp. andpocilloporids (corals that directly transmit symbionts to their offspring) or among those hosts display-ing ‘horizontal acquisition’, but that associate with specific symbionts. Most host-specialized typeswere found to be characteristic of a particular geographic region (i.e. Okinawa versus Central GBRversus Southern GBR). The mode of symbiont acquisition may play an important role in howsymbiont composition may shift in west Pacific host communities in response to climate change.There is no indication that recent episodes of mass bleaching have provoked changes in host–symbiont combinations from the central GBR.

KEY WORDS: Symbiodinium · Zooxanthellae biodiversity · Coral symbiosis · Phylogeography ·Community structure

Resale or republication not permitted without written consent of the publisher

Mar Ecol Prog Ser 284: 147–161, 2004

coral endosymbionts, in the genus Symbiodinium, arehighly diverse and evolutionarily old (Trench 1997,Rowan 1998). This revelation is not especially remark-able given the diverse environments, biological com-plexity and ages of the ecosystems where these sym-bioses are dominant. Nevertheless, our growingunderstanding of this diversity and its ecological sig-nificance has direct implications to addressing issuesregarding the impact of global climate change onnature conservation (Hoegh-Guldberg 1999).

Improved awareness of symbiont diversity, physio-logy and host-specificity contributes to hypotheses onhow corals respond to climate change (Buddemeier &Fautin 1993, Baker 2001). Differences in partner com-binations across latitudinal, longitudinal and environ-mental (e.g. irradiance) gradients have been reportedfor a number of host taxa (Baker & Rowan 1997, Rowanet al. 1997, LaJeunesse & Trench 2000, Rodriquez-Lanetty et al. 2001, Van Oppen et al. 2001, Burnett2002). Such patterns could be related to symbiontswith different sensitivities to thermal stress (Iglesias-Prieto et al. 1992, Iglesias-Prieto & Trench 1997,Jones et al. 1998, Warner et al. 1999), irradiance(Chang et al. 1983, Iglesias-Prieto et al. 2004), and/orhost-specificity (Colley & Trench 1983, LaJeunesse2001). These findings underlie the hypothesis thatpartner-switching creates new partnerships that arebetter adapted to changes in physical–environmentalconditions (Rowan & Powers 1991, Buddemeier &Fautin 1993, Baker 2001). Can the formation of newpartnerships (i.e. ‘switching’) take place rapidly overecological time scales or is this capacity limited byslower evolutionary processes that may not keep pacewith the current rate of environmental change (Hoegh-Guldberg et al. 2002)?

Further description of the variability between hostsand their symbionts over environmental, latitudinaland geographic ranges should help define the spatialand temporal limitations governing the extent to whichthese systems may respond and possibly adjust tofuture climate change. To this end, large-scale surveysof Symbiodinium spp. diversity and ecology have beeninitiated on stony corals and related host groups fromthe southern Great Barrier Reef (GBR), Hawaii, andCaribbean regions (Loh et al. 1998, Baker 1999, La-Jeunesse 2002, LaJeunesse et al. 2003, 2004).

The genus Symbiodinium is partitioned systemati-cally into a number of major phylogenetic divisions orclades. To date, most Pacific reef cnidarians are foundto harbor Symbiodinium spp. from Clade C, whereashost assemblages in the Caribbean associate com-monly with Symbiodinium spp. from Clades A and(especially) B (Baker & Rowan 1997, LaJeunesse et al.2003). The shared ecological dominance in theCaribbean of Clade B with Clade C Symbiodinium

spp. resulted possibly from environmental change inthis region during the Pliocene–Pleistocene transition(Jackson 1994, Collins et al. 1996, Budd 2000, Baker2003, LaJeunesse et al. 2003). Harsher (e.g. colder)physical–environmental conditions in the region dur-ing this time may have promoted the partial ecologicaldisplacement of Clade C by Clade B (Baker 2003,LaJeunesse et al. 2003).

Each Symbiodinium clade comprises a yet undeter-mined number of closely related ‘types’ or species thatcluster at the end of long diverging branches (Rowan1998, LaJeunesse 2002, Baker 2003). Molecular geneticidentification of ecologically different Symbiodiniumspp. within each major clade has been achieved bycomparing internal transcribed spacer region (ITS 1and 2) sequences of ribosomal RNA genes (Baillie et al.2000, LaJeunesse 2001, 2002, Van Oppen et al. 2001.These genetically distinct types possess unique envi-ronmental (depth zonation), ecological (host range)and geographic distributions. Some types have widegeographic distributions (Baillie et al. 2000, LaJeunesse2001) and are found in numerous host taxa (LaJeunesse2002). Most others display limited geographic rangesand/or associate specifically with a particular hostgenus or species. Because the host is a major axis ofresource (habitat) utilization, each type is defined asbeing a ‘generalist’ or ‘specialist’ depending on theirrelative capacity to associate with different host taxa(LaJeunesse et al. 2003).

The world’s largest contiguous reef system, the GreatBarrier Reef (GBR), off the east coast of Australia,offers an appropriate setting to explore host–symbiontdynamics because communities from different lati-tudes (a total range of nearly 15°), physical environ-ments (warm turbid inshore versus colder clear off-shore waters), and those with different bleachinghistories (including frequency and severity) can beexamined. Systematic surveys were first begun in themost southern region of the GBR, Heron Island of theCapricorn bunker group (LaJeunesse et al. 2003), afterinitial studies by Loh et al. (1998). The majority of hosttaxa from the southern GBR associate with one ofseveral generalist symbionts from the Clade C lineage.While these types were the most ecologically common,the greatest proportion of Symbiodinium spp. diversityconsisted of host-specific and/or rare types, special-ists, also from Clade C (LaJeunesse et al. 2003). Thelog normal distribution of symbiont abundance and/orprevalence (few highly prevalent generalists and manyrare specialists) is consistent with reefs surveyed fromthe Caribbean (LaJeunesse 2002).

Among other remaining questions, it is not known ifthe patterns described above hold for GBR reefs sys-tems closer to the equator, where many of the mid-shelf and inshore reefs suffered greater bleaching and

148

LaJeunesse et al.: West Pacific zooxanthellae communities

mortality than those in the southern GBR (Berkelmans& Oliver 1999, Skirving & Guinotte 2001). Moreover,there may be an inherent latitude-related pattern inhost–symbiont partnerships across coral communities,similar to that already documented for several host taxa(LaJeunesse & Trench 2000, Loh et al. 2001, Rodriquez-Lanetty et al. 2001). Finally, we sought to determineif the relatively low host-to-symbiont diversity in thesouthern GBR is consistent with lower-latitude GBRreefs. The coral diversity in the southern province ofthe GBR is substantially lower than in more northernareas, forming a high-latitude subset of the total GBRhost-species pool (Veron 1995, 2000).

We also conducted surveys at Zamami Island, partof the Kerama Islands, located 20 to 40 km west ofOkinawa, in the northern hemisphere, for purposes ofcomparing GBR Symbiodinium spp. diversity with thaton a reef from another region in the western Pacific.Unlike Okinawa to the immediate east, Zamami Islandwas not severely impacted by bleaching and representsa more ‘pristine’ reef analogous to Heron Island at thesame relative latitude in the opposite hemisphere.

MATERIALS AND METHODS

In May 2003, symbiotic invertebrates were collectedby SCUBA on reefs from the central Great BarrierReef. Collections were conducted in late March 2002 atZamami Island, of the Kerama Islands, approximately30 km west of the south end of Okinawa Island, Japan.

As with previous studies attempting to quantify thegeneral diversity of symbionts from a reef system (La-Jeunesse 2002, LaJeunesse et al. 2003, 2004), anemphasis was placed on sampling from a diverse rangeof hosts (over 75 genera, 154 species) consisting ofhard corals, soft corals, gorgonians, anemones, zoan-thids, corallimorphs, tridacnid clams and nudibranchs.Importantly, members of a host taxon found in a partic-ular reef environment and geographic region usuallypossess the same symbiont type (Baker 1999, La-Jeunesse 2002), with some exceptions (cf. Loh et al.2001) (i.e. 80 to 100% of individuals of Coral C sam-pled at Depth D on Reef R will have Symbiont S). Dif-ferences in host diversity between various reef habitatsand locations made sampling from a proportional num-ber of species difficult. For example, half of the hostgenera collected at Rib and Feather Reefs weresampled at the Curaçao Island fringing reef. The lowgeneric host diversity at this inshore location andlimited field time made collections at this reef unpro-portional to those of the mid-shelf reefs.

Collections from shallow (1 to 4 m) and deeper(>10 m) reef zones were made to obtain corals livingunder different irradiances. Host taxa distributed at

both depths were collected to identify possible ‘poly-morphisms’. Sampling colonies from deep and shal-low habitats increased the probability of identifyingcoral species that associate with more than 1 symbiont(Baker 1999). This work did not attempt to quantify thecomplete diversity of symbionts with which a particu-lar host associates, but rather represents a ‘snapshot’ ofthe symbionts across a wide array of hosts in a particu-lar community. A study of complete diversity wouldrequire exhaustive sampling of the host from everyenvironment in which it is found.

Before sampling, each host individual was photo-graphed (overall morphology and close-up) using aNikon Cool-Pix 5000 digital camera in an underwaterhousing (Subal, Netherlands) for later identification.The images obtained were compared with taxonomicreferences and identification guides (e.g. Veron 2000,Fabricius & Alderslade 2001), and most hosts wereidentified to the genus and species level.

Fragments representative of the host colony or indi-vidual were collected and processed to separate sym-bionts from host tissues, as previously described (La-Jeunesse et al. 2003). The resulting algal pellet waspreserved in 20% DMSO, 0.25 M EDTA in NaCl-saturated water (Seutin et al. 1991). Algal pellets fromZamami Island were preserved in CHAOS solution(4M guanidine thiocyanate, 0.5% sarkosyl, 2.5 mMTris-HCL (pH 8.0) and 0.1 M β-mercaptoethanol).

The Wizard DNA preparation protocol by Promega,modified by LaJeunesse et al. (2003), was used toextract nucleic acids. Approximately 30 mg of materialwas placed into 1.5 ml microcentrifuge tubes with250 µg 0.5 mm glass beads and 600 µl nuclei lysisbuffer (Promega) and bead-beaten for 2 min at 800 × gin a Biospec Mini-Beadbeater. The lysate was thenincubated with 0.1 mg ml–1 Proteinase K for 1 to 2 h at65°C, followed by incubation with 6 µg ml–1 RNase at37°C for 10 min. Protein precipitation buffer (250 µl)was then added and the extract incubated on ice for 10to 15 min. After centrifugation for 5 min at 24 000 × g,600 µl of supernatant was transferred to a second1.5 ml tube containing 700 µl isopropanol 100% and50 µl NaAc (3 M, pH 5.6). Following incubation on icefor 10 min, the precipitated DNA was centrifuged andthe pellet washed with 70% EtOH. The DNA was cen-trifuged again for 5 min, dried, and resuspended in70 µl H2O and 4 µl 10× Tris-EDTA.

While numerous molecular markers are now employedfor the study of Symbiodinium diversity (Baker 2003),the internal transcribed spacer region (ITS) methodappears to provide adequate resolution between eco-logically distinctive forms while being sufficientlyconserved to enable assessment of different types tobe compared among communities from distant geo-graphic areas. For each DNA extract, the ITS 2 region

149

Mar Ecol Prog Ser 284: 147–161, 2004

was amplified using primers ‘ITS 2 clamp’ and ‘ITSint-for 2’ (LaJeunesse & Trench 2000) with the touch-downthermal cycle given in LaJeunesse (2002). Productsfrom these PCR reactions were electrophoresed for9.5 h on denaturing gradient gels (45 to 80%) using aCBScientific system. Gels stained with Sybergreen(Molecular Probes) for 30 min were photographedusing a Kodak digital imaging system or with standardblack-and-white 677 polaroid film.

The PCR-DGGE fingerprint signatures from eachsample were compared with profiles from earlier datasets (LaJeunesse et al. 2003, 2004). The identificationof new symbiont types was verified by excising brightlystained bands from the denaturing gel. The DNA waseluted in 500 µl HOH for several days, re-amplifiedusing the same primer set without the guanine-cytosine-rich clamp extension in a standard PCR ther-mal cycle profile (annealing of 52°C for 40 cycles), andsequenced as previously described (LaJeunesse 2002).

To demonstrate the reproducibility of the PCR reac-tion by this method, 2 samples that showed a similarprofile involving 2 co-dominant ITS variants, but withdifferent relative band intensities, were repeatedlyamplified and compared. This verified whether cleardifferences in band intensity were the product of PCRartifact or reflected true differences in copy numberbetween variant sequences within a genome.

Maximum parsimony under the default settings ofPAUP* 4.0b10 (Swofford 1999) was employed for infer-ring a phylogeny. Sequences were edited and alignedmanually using Sequence Navigator Version 1.0 soft-ware (ABI, Division of Perkin Elmer). Clade C pos-sesses ITS 2 sequence types distinguished by a lownumber of base substitutions and/or insertion/dele-tions. With maximum parsimony (MP), informativesequence gaps as a 5th character state were included,delayed-transformation (DELTRAN) was chosen forcharacter state optimization, and no model of molecu-lar evolution was assumed. Sister lineages to Clade Crepresented by Fr1 (sensu Pochon et al. 2001) and F1in Clade F (Symbiodinium kawagutii; LaJeunesse2001) were used as outgroups (GenBank SequencesAJ291520 and AF333515). Phylogenetic analyses,neighbor-joining (NJ) and Bayesian inference of phy-logeny were also performed and the resulting treetopologies compared to MP. Baysian analyses wereimplemented using the software MrBayes Version3.0b4 (Huelsenbeck & Ronquiest 2001). We ran 500 000generations under the HKY85 models of sequence evo-lution, beginning with an unspecified tree topology,and no defined prior probabilities. The log probabilityreached stationarity between 50 000 and 75 000 gener-ations. This burn-in was not discarded and thereforethe posterior probabilities presented below are con-servative estimates. A bootstrap resampling was also

conducted for 500 replicates to assess relative branchsupport. Given that there are a small number of criticalbase substitutions (out of 321 aligned characters, 230were invariable and 32 were parsimony-informative),bootstrapping probably underestimates the actual sup-port for sub-clade radiations. A second bootstrappingwas performed with the resampling value doubled(642 characters). An unrooted phylogenetic analysis ofClade C sequences was also performed and topologywas compared with the rooted one.

Finally, symbiont community similarity from eachregion was assessed statistically using the Sørensencoefficient. It is preferred here to the Jaccard methodbecause it weights species, or types that are commonbetween regions over types found in only 1 area.

RESULTS

Symbiodinium spp. diversity

We identified 32 symbiont ‘types’ from biopsies of 74different host genera sampled at the 3 reefs surveyedfrom the central GBR (Table 1) and 20 types from 31host genera from Zamami island (Table 2). PCR-DGGEfingerprint profiles representing ‘new’ types from theCentral GBR and Okinawa are shown in Fig. 1a,brespectively. All belong to Clade C, the group mostprevalent among host taxa surveyed at all 3 regionallocations (Tables 1 & 2). New Clade C types character-ized were C1j (AY589732), C1k (AY589733), C31a(AY589746), C31b (AY589767), C40 (AY258485), C40a(AY589747), C40b (AY589748), C55 (AY589759), C56(AY589760), C57 (AY589761), C58 (AY589762), C59(AY589763), C60 (AY589764), C61 (AY589765), C62(AY589766), C64 (AY589768), C65/65a (AY589769/AY589770), C67 (AY686647), C68 (AY589772), C69/C69a (AY589773/AY686648), C70 (AY589774), andC71/C71a (AY589775/AY589776). A nexus alignmentof all Clade C ITS types from the west Pacific isavailable upon request from the corresponding author.

Due to the dominance of Clade C Symbiodiniumspp., the diversity and prevalence of symbionts fromother clades was low. Types A3 and A6 (AY686646)were identified from tridacnid clam tissues from thecentral GBR and Okinawa respectively. Clade B mem-bers were absent from our collections, except Type B1in the soft coral genus Nephthia from the southernGBR (LaJeunesse et al. 2003).

Only 3 ITS types in Clade D were identified (Fig. 2).Type D1a was found sporadically among faviids, mus-sids and oculinids from the central GBR, and agaricidsand fungiids from Zamami Island in typically shallowhabitats (1 to 4 m). Its fingerprint is known fromhosts from the western Indian Ocean, central Pacific

150


Ocean, and Caribbean Sea (LaJeun-esse 2002, LaJeunesse et al. 2005, A.Baker & T. LaJeunesse unpubl.). TypeD2 (AY686649) was found only amongsome Acropora spp., indicating that itmay be specialized to this particulargenus. Type D3 (AY686650) wasfound at only 1 location and in 1 hosttaxon (Clavularia sp. collected at 6 m).The rarity and probable host-specificnature of this type indicates thatClade D also consists of geographi-cally widespread host-generalists andmore localized, host-specific and/orrare forms.

PCR-DGGE analysis of Type C3h

The ITS 2 PCR-DGGE fingerprintprofiles of Type C3h from 19 host gen-era are illustrated in Fig. 3a. C3h wasnever identified in acroporids, how-ever. The genome of this particularSymbiodinium sp. contains 2 co-dominant ITS 2 sequences across theribosomal repeat array. Essentially it isintermediate between Types C3 andC21, a genome either caught in theprocess of concerted evolution fromthe ancestral sequence of C3 tothe more derived sequence of C21, orthe result of sexual recombination be-tween Types C3 and C21. The de-signation C3h was therefore given tothis fingerprint profile to distinguish itfrom Types C3 and C21. The upperC21/C3h band is consistently brighterthan the lower and suggests that therelative copy number between each

sequence is fixed in the genome of this organism. Therewas one exception; the C3 band from Goniastrea pecti-nata (Feather Reef, 15 m) was brighter than theC21/C3h band and the faint banding pattern foundhigher up in that lane differed from the other profiles.

To test whether this difference was a random artifactof the amplification process, 8 PCR replicate reactionswere conducted on the DNA extract from Goniastreapectinata that gave rise to this variant and on a repre-sentative that produced the normal profile. There wasno discernible fingerprint variation from reaction toreaction and the profiles from each example remainedconsistent with original analyses (Fig. 3b). The C3hprofile from G. pectinata is therefore different from theother C3h profiles. Because these similar fingerprint

151

Fig. 1. Symbiodinium spp. (a) Representative PCR-DGGE ITS 2 fingerprints (pro-files) of Lineage C types observed in hosts from the Central Great Barrier Reef.Identities, given as alphanumeric designations, and diagnostic band(s) to whicheach refers are compiled for those species newly reported: uppercase letters indi-cate lineage or clade, numbers represent ITS type, and lowercase letters denote arDNA paralog, when one is present within the genome and diagnostic of theentire fingerprint. (b) PCR-DGGE ITS 2 fingerprints of endosymbionts fromZamami Island, Okinawa Japan. Profiles from common symbionts C3, C1, and D1were pooled and run in the marker (MKR) lane. Examples of heteroduplexes areindicated; they are artifacts of the DGGE-PCR process present in fingerprints

of genomes with more than 1 dominant ITS 2 sequence

a

b

Fig. 2. Symbiodinium spp. PCR-DGGE ITS 2 fingerprints of Clade Didentified from the western Pacific.D1a was found sporadically amongfaviid, mussid and oculinid genera,usually related to shallow environ-ments but not completely dependenton depth. This fingerprint is found inhosts from the western Indian Ocean,central Pacific and Caribbean (La-Jeunesse 2002, A. Baker & T. La-Jeunesse unpubl.). Type D2 wasfound only in Acropora spp. from thecentral GBR and may represent a host-specialized symbiont. Type D3 wasfound at Curaçao Island in octocoralClavularia sp., but questions about itsprevalence and host-specificity remain

Cade D ‘types’

Mar Ecol Prog Ser 284: 147–161, 2004152

Tab

le 1

. Sym

bio

din

ium

sp

p. H

ost

spec

ies

and

sym

bio

nt

‘typ

e’ f

rom

sh

allo

w a

nd

dee

p r

eefs

of

the

cen

tral

GB

R. A

lph

anu

mer

ic id

enti

fier

s re

fer

to s

ymb

ion

t cl

ade

(up

per

case

lett

er),

IT

S t

ype

(nu

mb

er)

and

pre

sen

ce o

f a

char

acte

rist

ic c

o-d

omin

ant

par

alog

ue

in t

he

rib

osom

al a

rray

(lo

wer

case

lett

er).

Nu

mer

als

in p

aren

thes

es a

re n

um

ber

of

colo

nie

sin

wh

ich

a s

ymb

ion

t w

as f

oun

d.

‘Typ

es’

sep

arat

ed b

y so

lid

us

wer

e id

enti

fied

tog

eth

er i

n t

he

sam

e sa

mp

le.

v: v

erti

cal

tran

smit

ter;

h:

hor

izon

tal

tran

smit

ter

(Ric

hm

ond

&

Hu

nte

r 19

90, B

enay

ahu

199

7)

AN

TH

OZ

OA

Act

inia

ria

Het

erac

tis

mag

nif

ica

C67

C68

Sti

chod

acty

la g

igan

tea

(v?)

C69

aU

nk

now

n T

hal

assi

anth

idae

C1

Co

rall

imo

rph

aria

Dis

coso

ma

spp

.C

1 (3

)C

1C

1S

cler

acti

nia

Acr

opor

idae

Acr

opor

a ac

ule

us

(h)

D2

C3h

Acr

opor

a ce

real

is (

h)

C1,

C3k

C3i

C3

Acr

opor

a cy

ther

ea (

h)

C3

Acr

opor

a d

igit

ifer

a (h

)C

3kC

3i (

2)A

crop

ora

div

aric

ata

(h)

C3k

Acr

opor

a fl

orid

a (h

)C

3k, C

3iC

3k, C

3iA

crop

ora

form

osa

(h)

C3

Acr

opor

a g

emm

ifer

a (h

)C

3iA

crop

ora

gra

nd

is (

h)

C3

(2)

Acr

opor

a h

um

ilis

(h

)C

3k (

2), C

3iA

crop

ora

hya

cin

thu

s (h

)C

3kC

3A

crop

ora

lati

stel

la (

h)

C21

Acr

opor

a lo

rip

es (

h)

D2

C3

Acr

opor

a m

ille

por

a (h

)C

3kC

3kA

crop

ora

mon

ticu

losa

C3k

Acr

opor

a n

asu

ta (

h)

C3k

C3

C3k

C3k

C3i

Acr

opor

a n

obil

is (

h)

C1/

D2,

C3i

C3k

C3k

C3

Acr

opor

a p

alif

era

(v)

C3k

Acr

opor

a p

anic

ula

taD

2A

crop

ora

sarm

ento

sa (

h)

C3

Acr

opor

a se

cale

(h

)C

1A

crop

ora

ten

uis

(h

)C

1/D

2, C

1C

3C

1A

crop

ora

torr

esia

na

C3

Acr

opor

a va

lid

a (h

)C

3 (2

)A

crop

ora

yon

gei

(h

)C

3M

onti

por

a g

rise

a (v

)C

61 (

2)M

onti

por

a h

isp

ida

(v)

C26

aM

onti

por

a m

onas

teri

ata

(v)

C26

a, C

15f

Mon

tip

ora

turt

len

sis

(v)

C31

C26

aC

26a

Ag

aric

iid

aeG

ard

iner

oser

is p

lan

ula

taC

3ha

Lep

tast

rea

pu

rpu

rea

C1

C1b

Lep

tast

rea

pru

inos

aC

3hC

1L

epto

seri

s ya

bei

C3h

C1

Lep

tose

ris

exp

lan

ata

C21

Pac

hys

eris

ru

gos

a (h

)C

3h

Hos

t sp

ecie

sS

ymb

ion

t ‘t

ype’

Fea

ther

Rib

Cu

raça

ore

efre

efIs

lan

d1

–8

m10

–17

m1

–8

m10

–17

m1

–6

m

Pac

hys

eris

sp

ecio

sa (

h)

C3h

C21

Pav

ona

exp

lan

ula

taC

3h, C

21P

avon

a d

uer

den

iC

1D

end

rop

hyl

liid

aeT

urb

inar

ia f

ron

den

s (h

)C

40, C

3hC

40 (

2), C

3hT

urb

inar

ia r

enif

orm

is (

h)

C40

Tu

rbin

aria

ste

llu

lata

C40

, C1

Eu

ph

ylli

idae

Eu

ph

ylli

a d

ivis

a (h

)C

1P

lero

gyr

a si

nu

osa

C1

Fav

iid

aeC

aula

stre

a fu

rcat

a (h

)C

3C

3hC

yph

astr

ea m

icro

ph

thal

ma

(h)

C1

Cyp

has

trea

ser

aili

a (h

)C

1C

3C

1C

yph

astr

ea d

ecad

iaC

1D

iplo

astr

ea h

elio

por

aC

3E

chin

opor

a g

emm

acea

(h

)C

3C

3hC

3E

chin

opor

a h

irsu

tiss

ima

C40

C3h

Ech

inop

ora

hor

rid

a (h

)C

40a

Ech

inop

ora

lam

ello

sa (

h)

C40

C40

C3

C40

bE

chin

opor

a m

amm

ifor

mis

C3

Fav

ia p

alli

da

(h)

C1

Fav

ia s

pec

iosa

C1

Fav

ia s

tell

iger

aC

3F

avit

es a

bd

ita

C3

(2),

C3h

Gon

iast

rea

aust

rale

nsi

sC

1C

1C

1G

onia

stre

a ed

war

dsi

(p

ecti

nat

a)C

3G

onia

stre

a p

ecti

nat

a (h

)C

1C

3h (

2)C

3G

onia

stre

a re

tifo

rmis

(h

)C

3L

epto

ria

ph

ygra

(h

)C

3/D

1aM

onta

stre

a cu

rta

(h)

C3/

D1a

Mon

tast

rea

vale

nci

enes

si (

h)

C3

Ou

lop

hyl

lia

cris

pa

(h)

C3h

Pla

tyg

yra

dae

dal

ea (

h)

C3h

Pla

tyg

yra

lam

elli

na

(h)

C3

Pla

tyg

yra

pin

i (h

)C

3C

3, C

3hP

laty

gyr

a ry

uk

yuen

sis

C3h

Fu

ng

iid

aeF

un

gia

(cte

nac

tis)

ech

inat

aC

3hF

un

gia

fu

ng

ites

(h

)C

3h (

2)C

1F

un

gia

gra

nu

losa

C3h

C1

Fu

ng

ia h

orri

da

C3h

Fu

ng

ia p

aum

oten

sis

(h)

C3h

Hel

iofu

ng

ia a

ctin

ifor

mis

(h

)C

3hH

erp

olit

ha

web

eri

C1

Hos

t sp

ecie

sS

ymb

ion

t ‘t

ype’

Fea

ther

Rib

Cu

raça

ore

efre

efIs

lan

d1

–8

m10

–17

m1

–8

m10

–17

m1

–6

m

(Tab

le c

onti

nu

ed o

n n

ext

pag

e)

LaJeunesse et al.: West Pacific zooxanthellae communities 153

Lit

hop

hyl

lon

un

du

latu

mC

3hP

odab

acia

cru

stac

eaC

3hC

3hP

olyh

pyl

lia

talp

ina

(h)

C1

C3h

/C1

San

dal

olit

ha

rob

ust

a (h

)C

3hM

eru

lin

idae

Hyd

nop

hor

a ex

esa

C3h

(2)

, C3

C3h

Mer

uli

na

amp

liat

a (h

)C

1C

3C

3(2)

Mer

uli

na

scab

ricu

laC

1 (2

)M

uss

idae

Aca

nth

astr

ea e

chin

ata

(h)

D1a

Cyn

arin

a la

crym

alis

C3h

Lob

oph

ylli

a h

emp

rich

ii (

h)

C3h

C3

(2),

C3h

aC

3hS

coly

mia

au

stra

lis

C1

Sym

ph

ylli

a ag

aric

iaC

3S

ymp

hyl

lia

rad

ian

sC

40C

40S

ymp

hyl

lia

rect

aC

40O

culi

nid

aeG

alax

ea h

orre

sen

s (G

. acr

hel

ia)

D1a

Gal

axea

ast

reat

a (h

)C

1G

alax

ea f

asci

cula

ris

(h)

C1,

D1a

D1a

Pec

tin

iid

aeE

chin

oph

ylli

a as

per

a (h

)C

3ha

C3h

(2)

Ech

inop

hyl

lia

ech

inat

aC

3hE

chin

oph

ylli

a ec

hin

opor

oid

esC

1E

chin

oph

ylli

a or

ph

een

sis

(h)

C3h

Myc

ediu

m e

lep

han

totu

s (h

)C

3h (

2), C

3C

3h, C

1C

40O

xyp

ora

gla

bra

(h

)C

3hC

3ha

Oxy

por

a la

cera

(h

)C

3hC

3hP

ecti

nia

lac

tuca

(h

)C

3hP

ecti

nia

pae

onia

(h

)C

3hP

ecti

nia

sp

.C

3hP

ocil

lop

orid

aeP

ocil

lop

ora

dam

icor

nis

(v)

C1j

C1j

(2)

C1j

Poc

illo

por

a ey

dou

xi (

v)C

1cP

ocil

lop

ora

mea

nd

rin

a (v

)C

1cP

ocil

lop

ora

verr

uco

sa (

v)C

1cC

1cC

1cC

1cS

eria

top

ora

hys

trix

(v)

C3

C3

Sty

lop

hor

a p

isti

llat

a (v

)C

8C

8P

orit

idae

Alv

eop

ora

fen

estr

ata

C1

C31

b(2

)G

onio

por

a d

jib

outi

ensi

sC

1G

onio

por

a m

inor

C1

C1

Gon

iop

ora

ten

uid

ens

C1

Por

ites

an

nae

(v)

C15

(2)

C60

Por

ites

cyl

ind

rica

(v)

C15

C15

C15

C60

Por

ites

lic

hen

(v)

C15

C15

C15

C15

Por

ites

mas

sive

(v)

C15

C15

C15

C15

(2)

C15

Por

ites

nig

resc

ens

(v)

C15

Por

ites

ru

s (v

)C

15P

orit

es v

aug

han

i (v

)C

60C

15C

15

Hos

t sp

ecie

sS

ymb

ion

t ‘t

ype’

Fea

ther

Rib

Cu

raça

ore

efre

efIs

lan

d1

–8

m10

–17

m1

–8

m10

–17

m1

–6

m

Sid

eras

trei

dae

Cos

cin

arae

a co

lum

na

C1

C1

C1

(2)

Psa

mm

ocor

a co

nti

gu

aC

1P

sam

moc

ora

dig

itat

aC

1C

1P

sam

moc

ora

pro

fun

dac

ella

C1

Zo

oan

thid

eaP

alyt

hoa

sp

. (h

)C

1C

1, C

3Z

oan

thu

s sp

. (h

?)C

62 (

3)C

62C

62A

lcyo

nac

eaA

lcyo

nii

dae

Kly

xum

sp

.C

64L

obop

hyt

um

sp

p. (

h)

C1

C1

(2)

C1

Sar

cop

hyt

on s

pp

. (h

)C

1 (4

)C

65(2

)C

1S

inu

lari

a sp

. (h

)C

65 (

3)C

1cC

65 (

2)C

1 (2

)A

nth

oth

elid

aeIc

ilog

org

ia s

p.

C1

Bri

arei

dae

Bri

areu

m s

p.

C3

Cla

vula

rid

aeC

lavu

lari

a sp

. (h

?)D

3G

org

onid

aeH

ick

son

ella

exp

ansa

C65

Ru

mp

hel

la s

p.

C1

Ifal

uk

elli

dae

Plu

mig

org

ia s

p.

C1

Isid

idae

Isis

sp

.C

23T

ub

ipor

idae

Tu

bip

ora

mu

sica

C1b

Xen

iid

aeA

nth

elia

sp

. (v)

C64

Xen

ia s

p. (

v)C

64C

64C

64C

1kC

64H

eter

oxen

ia s

p. (

v)C

64C

64 (

2),C

15C

15e

Hel

iop

ora

cea

Hel

iop

ora

coer

ule

aC

1 (2

)

HY

DR

OZ

OA

Hyd

roid

aA

gla

oph

enia

sp

.C

15C

15M

ille

por

a ex

aesa

C1

C1

Mil

lep

ora

ten

ella

C15

e

MO

LL

US

CA

Tri

dac

nid

aeT

rid

acn

a d

eras

a (h

)C

1T

rid

acn

a g

igas

(h

)C

1, A

3T

rid

acn

a m

axim

a (h

)C

1 (3

)N

ud

ibra

nch

Pte

raeo

lid

ia i

anth

ina

(h)

C1

Hos

t sp

ecie

sS

ymb

ion

t ‘t

ype’

Fea

ther

Rib

Cu

raça

ore

efre

efIs

lan

d1

–8

m10

–17

m1

–8

m10

–17

m1

–6

m

Tab

le 1

(co

nti

nu

ed)

Mar Ecol Prog Ser 284: 147–161, 2004

profiles lack additional bands that could be sequencedto distinguish them on phylogenetic grounds, theywere both conservatively scored as the same type.Genetic structure occurs at all levels of a lineage downto the individual (Avise 2000). It is not surprising thatthe recent use of more variable markers indicates thatthese ITS lineages may be further subdivided (Santoset al. 2004).

Clade C phylogeny

The genetic relatedness of western Pacific Clade CSymbiodinium spp. is presented in Fig. 4. Internal

topology remained consistent under MP, NJ orBaysian inference of phylogeny methods, both rootedand unrooted. The exact point at which the outgroupbranch joins Clade C varied slightly, dependent onwhether Fr1 (Clade H; Pochon et al. 2005) and/orF1 (Clade F; LaJeunesse 2001) were used separatelyor together. The effect of long-branch attraction maysometimes supplant the outgroup connection withthe most ancestral sequence of this clade. However,a combination of ecological (host–generalist), biogeo-graphic (pandemic distribution), and phylogenetic(ancestral sequence to a radiation of numerous host-specific and/or rare types) evidence suggest TypesC3 or C1 are probably the most ancestral. The posi-

154

Table 2. Symbiodinium spp. Host species and symbiont ‘type collected’ at 1 to 10 m depth on reefs off Zamami Island, Okinawa, Japan. Further details as in Table 1

ANTHOZOA

ActiniariaUnkown anemone C70

ScleractiniaAcroporidae

Acropora abrolhosensis C1Acropora aspera C3 (2)Acropora cerealis C3 (2)Acropora copiosa C3Acropora digitifera (h) C3 (2)Acropora divaricata C3, C1 (2)Acropora donei C3 (3), C1Acropora exquisita C3Acropora florida (h) C3Acropora listeri C3Acropora microphthalma (h) C3Acropora nobilis (h) C1, C3Acropora palifera C1 (2)Acropora secale C3 (2)Acropora selago C3 (2)Acropora subglabra C3Acropora tenuis (h) C3 (2)Acropora valida (h)) C3, C3iAcropora verweyi C1Astreopora myriophthalma C1 (3)Montipora danae (v) C31Montipora efflorescens (v) C30Montipora mollis (v) C58 (2)Montipora venosa (v) C31a, C31Montipora verrucosa (v) C31Montipora spp. (v) C31, C31a, C31c, C1 (2)

AgariciidaePachyseris rugosa C1 (3)Pachyseris speciosa C27Pavona varians D1a

AstrocoeniidaePalauastrea ramosa C1 (3)

FaviidaeCaulastrea chalcidicum C1Cyphastrea japonica C1Cyphastrea sp. C21a (2)Echinopora lamellosa (h) C1, C3Echinopora pacificus C3Favia matthaii C3Favia stelligera C3Favites halicora C21aPlatygrya sp. (h) C55

FungiidaeFungia danai (h) C27Fungia scutaria (h) C1 (2)Fungia spp. (h) C27, C1Sandalolitha robusta (h) C1/D1a

MerulinidaeHydnophora exesa C21a (2), C27Hydnophora rigida C21a (2), C1

MussidaeLobophyllia robusta C21aSymphyllia sp. (h) C21aSymphyllia radians C21a (3)

OculinidaeGalaxea fascicularis (h) C21a

PectiniidaeEchonophyllia sp. C21aPectinia alcicornis C21aPectinia sp. C21a

PocilloporidaePocillopora damicornis (v) C1c (2)Pocillopora eydouxi (v) C1c (3)Seriatopora hystrix (v) C59Stylophora pistillata (v) C1 (4)

PoritidaeAlveopora sp. C27 (2)Porites cylindrica (v) C56a (2)Porites lichen (v) C56 (2)Porites lutea (v) C15 (3)Porites massive (v) C15 (3)Porites sp. (v) C56 (2)Porites rus (v) C15Porites silimaniana (v) C15

SiderastreidaeCoscinaraea sp. C1 (3)Coscinaraea exesa C1/C27

ZooanthideaZoanthus sp. (h) C1 (2)

AlcyonaceaAlcyoniidae

Sarcophyton sp. (h) C71a

HYDROZOAMillepora sp. C57 (6)

MOLLUSCATridacnidae

Tridacna sp. (h) A6

Host taxon Symbiont ‘type’ Host taxon Symbiont ‘type’

tion of Type C61 was unstable and certain branchconnections within the C21 sub-clade varied slightly.Several well-developed sub-cladal lineages contain-ing multiple types (viz. C15 and C21 and their off-shoots) correspond with host genera that transfertheir symbionts directly from generation to genera-tion (Fig. 4b). Some host-specialized and/or raresymbionts, especially those identified from species

of Porites, Montipora and Pocillopora, werefound at all locations, but many othersexhibited slight sequence differences thatdistinguished them regionally (Fig. 4b).

Similarity of symbiont communities anddifferences in relative dominance

Type C1 dominated most host communi-ties. Several other types were found at eachlocation, but the majority of diversity con-sisted of less common, geographicallyrestricted, rare and/or host-specific types.Reefs from distant geographic regions ex-hibited marked differences in their symbiontcommunities (Fig. 5). The Sørensen percent-age of dissimilarity estimates betweenZamami Island and Heron Island, ZamamiIsland and the central GBR, and the centralGBR and Heron Island were 68, 68, and 60%respectively. The 3 reefs surveyed from thecentral GBR region shared essentially thesame community composition of symbionttypes (dissimilarity estimates calculated be-tween Rib and Feather Reefs were 21%;Fig. 5). This valuation of community differ-ence is probably an overestimate because ofthe presence or absence of undersampledrare types and limited, uneven, host taxasampling at each reef. Because only half ofthe host genera collected a Rib and FeatherReefs were sampled at Curaçao Island, sta-tistical comparisons of their community com-positions were not made. Clearly, the sym-biont community at Curaçao is a subset ofthat found on offshore reefs.

Differences among the relative dominanceof some common symbiont types were ob-served from reef to reef (Fig. 5). C3h differedin its relative abundance between the mid-shelf reefs of Rib and Feather and the inshorereef at Curaçao Island (within 150 km of eachother). It was the most common type at therelatively turbid, inshore site at Curaçao,where it occurred in more than 50% of hostgenera surveyed. Often found in certain scle-

ractinian hosts sampled below 15 m on mid-shelf reefs(Fig. 3a), it occured in more genera on Feather Reef(approx. 30%) than on Rib Reef (approx. 8%). In con-trast, C3h was rare at the southern GBR location ofHeron Island, being found in 1 sample from Pavonamaldivensis.

As was the case for the outer GBR reefs, C1 was themost prevalent symbiont at the Zamami Island loca-


Fig. 3. Symbiodinium spp. (a) PCR-DGGE ITS2 fingerprints of Type C3h,collected from 19 different scleractinian genera from deep or low-lightreefs on the central GBR. Rare in the southern GBR, it achieved its highestprevalence in the host community surveyed from the turbid inshorelocation at Curaçao Island, central GBR. Relative intensities of each bandare similar from sample to sample and indicate that these sequences areintragenomic variants. It is postulated that ribosomal array contains 2 co-dominant ITS 2 sequences (C3 and C21) and represents an intermediatecondition of concerted evolution from C3 to C21. Alternatively, it couldrepresent a recombinant of these 2 distinctive species. (b) Repeatability ofthe PCR-DGGE method. The sample that gave rise to a distinctive ‘variant’whose C3 band was brighter (hence greater copy number) than theC3h/C21 band was amplified again along with a representative exhibitinga ‘normal’ profile. No discernible variation was detected from reaction toreaction, and fingerprint profiles remained consistent with original ana-lyses. Therefore, C3h profile from Goniastrea pectinata can be viewed asqualitatively different from the others but, because of limits in discerningthese 2 profiles through sequencing, they are both conservatively scored as

the same type

a

b

Mar Ecol Prog Ser 284: 147–161, 2004

tion. But here, Types C21a and C27 were also amongthe most common. The complex PCR-DGGE finger-print observed for Type C21a is thus far unique tothe Okinawa, NW Pacific, region. C27 was previouslyfound to occur in Pavona spp. from Hawaii (La-Jeunesse et al. 2004) and in 1 Pavona sp. specimenfrom Heron Island (LaJeunesse et al. 2003).

DISCUSSION

Variability in symbiont dominance is influenced bylife history of the host

External physical conditions may influence speci-ficity in symbioses that begin anew after each genera-

156

Fig. 4. Symbiodinium spp. Phylogenetic reconstruction of western Pacific Clade C based on ITS 2 and partial 5.8S sequences. Max-imum parsimony phylograms are both (a) rooted and (b) un-rooted; both have very similar topologies. (a) The point of connectionbetween Clade C and outgroup lineages differed depending on whether Fr1 or F1 were used separately or together. Dotted line in-dicates putative connection to Type C3 (bold vertical line) and is based on phylogeographic and ecological evidence for being an-cestral among Clade C types. Values indicated for each internal branch node are bootstrap estimates (first number), bootstrap withresampling doubled to compensate for high proportion of invariant characters (underlined), and Bayesian posterior probabilities(in parentheses); internal nodes that lack posterior probabilities are based on insertion/deletions not assessed by Bayesian methods.(b) Radiations of host-specific and/or rare types from a small number of widely distributed and/or host-generalist types are shownin this unrooted topology. Localization among types to specific geographic regions is indicated by symbols on branch termini:(h) southern GBR; (j) central GBR; (�) both GBR regions; d Zamami Island, Okinawa. Encircled types were found at all locations

a

b


Fig. 5. Symbiodinium spp. Comparison of diversity and community structures across latitudinal, environmental, and geographicgradients. Percentage of host genera surveyed is presented on y-axes and the different symbiont types in order of prevalenceamong host genera along the x-axes. Each community comprises a few types common to many host taxa and numerous host-specific and/or rare species that characterize each region surveyed. Type C1 is usually most common in each community. Exclud-ing Type C3, each region contains secondary generalists that are prevalent only in their respective regions (e.g. Types C21a andC27 in Zamami Island, C3h on the central GBR, and C21 from the southern GBR). Striped bars show relative prevalence ofType C3h at different locations over the GBR. Number of host genera surveyed at each site is given on upper-right of each graph,

with total numbers of symbiont types (displaying different ITS 2 PCR-DGGE fingerprints) in parentheses

Mar Ecol Prog Ser 284: 147–161, 2004

tion. A majority of western Pacific corals broadcast-spawn eggs and sperm that do not contain symbionts(Richmond & Hunter 1990). Therefore, coral popula-tions from this expansive region may be highly sus-ceptible to shifts in symbiont type from generation togeneration. We found that environment and latitudeaffected the relative dominance of certain Symbio-dinium spp. within hosts that acquire their symbiontsfrom environmental sources (Fig. 5, Tables 1 & 2).Because adult colonies often show long-term stabilitywith a particular symbiont type (Coffroth et al. 2001,Goulet & Coffroth 2003, Iglesias-Prieto et al. 2004,LaJeunesse et al. 2005), the present host–symbiontcommunity structure among broadcast-spawning coraltaxa probably reflects the environmental conditionsunder which these symbioses were initially estab-lished. Over several generations, change in environ-ment could cause significant shifts in the type of sym-biont dominating the host community; however, thisscenario assumes that the extent and rate of changedoes not exceed the physiological capability of eachpotential partner. Because there is the likelihood ofsignificant climatic disturbances in these communitiesover the coming decades (Hoegh-Guldberg 1999,Kleypas et al. 1999), continued monitoring will docu-ment the range of partner flexibility (Baker 2003) anddetermine if symbiont change can ultimately sustainthese reef corals.

Symbiont C1 was dominant on most reefs, but anumber of generalist taxa exhibited great regionalvariability in their relative dominance. Temperatureand light fluctuations associated with decreasing lati-tude may explain differences between the central GBRand southern GBR symbionts (Fig. 5). Within the cen-tral GBR region, higher temperatures and perhapsgreater turbidity at the inshore site of Curaçao Islandmay be key selective factors that explain the offshore-to-inshore gradient of C3h dominance. Also, little isknown about the environmental abundances of Sym-biodinium spp. If C3h is particularly abundant in theplanktonic or benthic assemblages nearer to shore, thismay in part explain differences in its host communitydominance (Baker & Rowan 1997).

Specialized and/or rare symbionts that displayedlimited ranges in geographic and host distributionaccounted for major differences in symbiont composi-tion from region to region. Many of these symbiontsassociate with hosts that pass on symbionts directly totheir offspring. For example, montiporid corals broad-cast eggs containing symbionts (Richmond & Hunter1990). They predictably associate with one of a numberof closely related types that are a part of the montiporidsub-clade radiating from C21 (Fig. 4b). The majority ofmembers from this sub-clade associate exclusivelywith Montipora spp. They have particular geographic

distributions, and are common in 1 region but rare orabsent in other places. Type C26a occurred rarely inMontipora spp. from the southern GBR, but was com-mon among these corals from the central GBR. It wasabsent from Zamami Island, yet was identified fromM. capitata in Hawaii living at depths below 20 m (La-Jeunesse et al. 2004) and, therefore, appears to have alarge geographic range. Another type, C31, was com-mon among montiporids from Zamami Island but rareon the central GBR. It associates with the commonbrown color morph of M. capitata (at depths above15 m) from Hawaii (LaJeunesse et al. 2004). Thesewidely distributed host-specialists contrast with othermontiporid symbionts surveyed from the NW Pacific,SW Pacific, and Hawaii that appear to have more lim-ited ranges. These distributions, involving widely dis-persed specialized types versus regionally endemictypes, are also observed for poritid and pocilloporidscorals, and among non-scleractinian groups such asthe alcyonarians.

Implications of mass coral bleaching

Many of the central GBR reefs experienced masscoral bleaching in 1998, 2002 and 2004, with mostsevere effects occurring on inshore reefs (Berkelmans& Oliver 1999, Marshall & Baird 2000). Such episodesof stress and reduction in symbiont population densitymay facilitate a shift in the symbiont type that becomesdominant upon recovery, either through uptake fromthe environment (Buddemeier & Fautin 1993, Baker2001) or proliferation of surviving cells remainingwithin the polyps (Baker 2001). For example, symbiontshifts in the population of ecologically dominant Pocillo-pora verrucosa were reported in the eastern Pacificfollowing the 1997 El Niño-related bleaching event(Glynn et al. 2001). It cannot be determined if the in-crease of Clade D (specifically Type D1) in proportionto Clade C (possibly involving different types withinthis clade) among individual coral colonies was due tonatural selection of thermally resistant combinations orwhether symbiont population shifts occurred via pro-liferation of a minor population of Clade D withincolonies as they recovered (Baker 2001).

Symbiodinium spp. from Clade D are often found toassociate with hosts from thermally variable and turbidenvironments and colonies or individuals recoveringfrom bleaching episodes (Baker 2001, Toller et al.2001, Van Oppen et al. 2001, Chen et al. 2003, Iglasias-Prieto et al. 2004). Various Clade D members areconsequently viewed as thermally tolerant, stress-re-sistant, opportunistic Symbiodinium spp. In thepresent study, Curaçao Island was also surveyed to de-termine if the inshore environment favored types from

158


Clade D over those from Clade C. Contrary to our ex-pectation that Clade D would dominate the host com-munity on the warmer inshore reefs (cf. Van Oppen etal. 2001), its prevalence was not different from that onreefs further offshore. Questions remain as to whatextent Clade D Symbiodinium spp. may allow coralpopulations to compensate under continued sea--surface warming (Baker 2001, Little et al. 2004). Com-parative physiological analyses will need to be em-ployed to learn what attributes or limitations accountfor the changes in the host community dominance ofthese Symbiodinium spp. These geographic and envi-ronmentally relatable shifts are ultimately of interest indescribing evolutionary processes between host andsymbiont lineages through time (Thompson 1994).

Heron Island and Zamami Island, both high-latitudereefs, have had limited exposures and experienceswith bleaching. While their symbiont community struc-tures are similar, the difference in composition ofClade C Symbiodinium spp. is clearly a product of theirgeographic separation. Thus for the present, the sever-ity or frequency of mass coral bleaching in the centralGBR does not seem to have resulted in a shift in sym-biont population structures that cannot be otherwiseexplained by geographic, latitudinal and normal envi-ronmental factors. These differences in Symbiodiniumspp. distribution probably took many generations tobecome established.

Symbiont diversity relative to host diversity

Many Pacific hosts associate with closely relatedSymbiodinium spp. in Clade C (Baker & Rowan 1997,Loh et al. 1998, Baker 1999, LaJeunesse et al. 2003,2004). Initially, host assemblages from southern GBRreefs were reported to have lower relative symbiont di-versity of ITS types than reefs from the Caribbean (La-Jeunesse et al. 2003). Subsequent work in Hawaii hasalso observed this inverse relationship between hostand symbiont diversity (LaJeunesse et al. 2004). Theenumeration of ITS types found on the central GBR waslow in relative proportion to the number of host generasurveyed. This trend offers further support of an inverserelationship between host and symbiont diversity.Genetic divergence and symbiont speciation throughhost-specialization is common for hosts in which sym-bionts are directly transferred from generation togeneration (Futuyma & Moreno 1988, Douglas 1998) aprocess that is exemplified by poritid, pocilloporid andmontiporid corals and their symbionts. The greatercomposition of hosts in the Caribbean (63% brood outof the 19 coral species investigated) and Hawaii (29%out of the 12 investigated) whose symbionts are verti-cally transferred in comparison to the west Pacific (GBR

6% out of 144, Okinawa 4% out of the 26 coral speciesinvestigated; Richmond & Hunter 1990), explains someof these inverse relations, but not all.

Symbiont-specificity and the presence of host-specificsymbionts were found for hosts that rely upon horizontalsymbiont transmission. Among others, species of Acrop-ora (with Types C3i, C3k and D2) (some Acropora [Iso-pora] brood their larvae) and Zoanthus (with Type C62)(Ryland 1997), possessed Symbiodinium spp. not identi-fied in other host taxa (cf. Van Oppen 2004). Regionswith a high diversity of hosts involved in acquiring sym-bionts from the environment may favor the maintenanceof highly prevalent generalist symbionts (Law 1985).Host rarity would present problems for passivelydispersed symbionts and newly settled aposymbioticlarvae in finding each other within a diverse hostcommunity over effectively vast spatial scales.

Emergent patterns of Symbiodinium spp. biogeography

Phylogeography, relating patterns of geographic dis-tribution with genetic relatedness, is a powerful tool inassessing historical and evolutionary processes (Avise2000). Our current understanding of Symbiodiniumdiversity and geographic distribution remains limited,but as more ITS data are gathered, patterns of dispersaland geographic isolation/connectivity are beginning toemerge. Widely distributed symbiont types tend to behost-generalists and are, phylogenetically, ancestral tosymbionts that are more specialized. Based on ecologi-cal, biogeographic, and phylogenetic grounds, we inter-pret Types C1 and C3 as the ancestral stock from whichnumerous host-specific, regionally endemic and/or raretypes have radiated. Divergent, host-specific and/or rareforms tend to have narrow geographic ranges and areprobably endemic. These patterns of geographic parti-tioning are reinforced by genetic surveys at the ‘popula-tion level’ that have identified clear geographic structurewithin specific ITS lineages (Santos et al. 2003, 2004).

Certain symbionts show exceptions to these basic pat-terns of geographic partitioning. Their host specificityand biogeography leads to hypotheses concerning theevolutionary processes that occur between host andsymbiont lineages. Type C27 is widely distributedthroughout the Pacific but has different host relations indifferent regions. It is rare in the southern GBR (LaJe-unesse et al. 2003), highly specific for the corals in thegenus Pavona from Hawaii (LaJeunesse et al. 2004), yetis a generalist among hosts from Zamami Island reefs.This example demonstrates the capacity for a symbiontto be specialized for 1 particular host in 1 region, butto display more generalized associations or greaterprevalence at other locations. The extent to whichhost–symbiont specificity is expressed in geographically

159

Mar Ecol Prog Ser 284: 147–161, 2004

separate populations is predicted by the geographicmosaic theory of coevolution (Thompson 1994).

Differences in symbiont communities between cen-tral and southern regions indicate that the GBR is bio-geographically partitioned. The similar symbiont com-positions found at Rib, Feather and Curaçao reefs areprobably possessed by other reefs from the centralGBR region. Except for the more common types (e.g.C1, C3, C21, C15), few others were identified in thesouthern GBR survey (LaJeunesse et al. 2003). On thenorthern GBR, Types C1, C40, C3h and C8 wereidentified in corals from Lizard Island (Baker 1999, A.Baker & T. LaJeunesse unpubl.). While these resultsindicate some similarity in symbiont assemblages be-tween northern and central regions, other Symbio-dinium spp. not identified from central or southernregions were also among those identified from LizardIsland (A. Baker & T. LaJeunesse unpubl.). Basedon the presence and/or absence of characteristic host-specific symbionts, the GBR may be divided into anumber of biogeographic provinces. Ayre & Hughes(2000) reported a lack of geographic partitioning, ascoral allozyme variation revealed moderate to highlevels of gene flow along the entire GBR. Trackingthe presence/absence of host-specific symbiont types,not influenced by environment across spatial scalesencompassed by the GBR is potentially a rapid methodfor determining genetic connectivity between andwithin major reef systems, information important toconservation-related decisions.

Acknowledgements. The authors would like to thank thecrew of the ‘Lady Basten’ for their hard work and support.S. Kininmonth produced the maps in Fig. 5 and waitedpatiently on the surface. A. Irikawa helped with coral identifi-cation from Okinawa. M. Wakeford assisted with diving andhelped with host identification on the GBR. D. Fautin advisedon anemone identifications and life-history traits. M. vanOppen and R. Berkelmans hosted our stay in Townsville. R.Berkelmans got us through the ‘Australian’ paper work. Thisresearch was made possible from funding provided by theAustralian Research Council (O.H.-G.: ARCL245G), IOC-UNESCO-World Bank Targeted Working Group on CoralBleaching and Related Ecological Factors, Australian Instituteof Marine Sciences, and the NSF (grant OCE-0137007 toW.K.F. and G.W.S.)

LITERATURE CITED

Avise JC (2000) Phylogeography: the history and formation ofspecies. Harvard University Press, Cambridge, MA

Ayre DJ, Hughes TP (2000) Genotypic diversity and gene flowin brooding and spawning corals along the Great BarrierReef, Australia. Evolution 54:1590–1605

Baillie BK, Belda-Baillie CA, Maruyama T (2000) Con-specificity and Indo-Pacific distribution of Symbiodiniumgenotypes (Dinophyceae) from giant clams. J Phycol 36:1153–1161

Baker AC (1999) The symbiosis ecology of reef-buildingcorals. PhD thesis, University of Miami, Miami

Baker AC (2001) Corals bleach to survive change. Nature 411:765–766

Baker AC (2003) Flexibility and specificity in coral–algalsymbiosis: diversity, ecology and biogeography of Sym-biodinium. Annu Rev Ecol Evol Syst 34:661–689

Baker AC, Rowan R (1997) Diversity of symbiotic dinoflagel-lates (zooxanthellae) in scleractinian corals of the Carib-bean and eastern Pacific. Proc 8th Int Coral Reef Symp2:1301–1306

Benayahu Y (1997) Developmental episodes in reef softcorals: ecological and cellular determinants. Proc 8th IntCoral Reef Sym 2:1213–1218

Berkelmans R, Oliver JK (1999) Large-scale bleaching ofcorals on the Great Barrier Reef. Coral Reefs 18:55–60

Budd A (2000) Diversity and extinction in the Cenozoichistory of Caribbean reefs. Coral Reefs 19:25–35

Buddemeier RW, Fautin DG (1993) Coral bleaching as anadaptive mechanism. BioScience 43:320–326

Burnett WJ (2002) Longitudinal variation in algal symbionts(zooxanthellae) from the Indian Ocean zoanthid Palythoacaesia. Mar Ecol Prog Ser 234:105–109

Chang SS, Prézelin BB, Trench RK (1983) Mechanisms ofphotoadaptation in three strains of the symbiotic dino-flagellate Symbiodinium microadriaticum. Mar Biol 76:219–29

Chen CA, Lam KK, Nakano Y, Tsai W (2003) A stable associ-ation of the stress-tolerant zooxanthellae, Symbiodiniumclade D, with the low-temperature tolerant coral, Oulas-trea crispata (Scleractinia: Faviidae) in subtropical non-reefal coral communites. Zool Stud 42:540–550

Coffroth MA, Santos SR, Goulet TL (2001) Early ontogeneticexpression of specificity in a cnidarian–algal symbiosis.Mar Ecol Prog Ser 222:85–96

Colley NJ, Trench RK (1983) Selectivity in phagocytosis andpersistence of symbiotic algae by the scyphistoma stage ofthe jellyfish Cassiopeia xamachana. Proc R Soc Lond B219:61–82

Collins LS, Budd A, Coates AG (1996) Earliest evolution asso-ciated with closure of the Tropical American Seaway. ProcNatl Acad Sci USA 93:6069–6072

Douglas AE (1998) Host benefit and the evolution of speciali-zation in symbiosis. Heredity 81:599–603

Fabricius K, Alderslade P (2001) Soft corals and sea fans: acomprehensive guide to the tropical shallow water generaof the central-west Pacific, the Indian Ocean and the RedSea. Australian Institute of Marine Science, Townsvilleand New Litho, Surrey Hills, Melbourne

Futuyma DJ, Moreno G (1988) The evolution of ecologicalspecialization. Annu Rev Ecol Syst 19:207–233

Glynn PW, Maté JL, Baker AC, Calderon MO (2001) Coralbleaching and mortality in Panama and Ecuador duringthe 1997–1998 El Niño–southern oscillation event: spatial/temporal patterns and comparisons with the 1982–1983event. Bull Mar Sci 69:79–109

Goulet TL, Coffroth MA (2003) Stability of an octocoral–algalsymbiosis over time and space. Mar Ecol Prog Ser 250:117–124

Hoegh-Guldberg O (1999) Climate change, coral bleachingand the future of the world’s coral reefs. Mar Freshw Res50:839–866

Hoegh-Guldberg O, Jones R, Ward S, Loh W (2002) Ecology—is coral bleaching really adaptive? Nature 415:601–602

Huelsenbeck JP, Ronquist F (2001) MrBays: Bayesian infer-ence of phylogenetic trees. Bioinformatics 17:754–755

Iglesias-Prieto R, Trench RK (1997) Photoadaptation, photo-

160


acclimation and niche diversification in invertebrate-dinoflagellate symbioses. Proc 8th Int Coral Reef Symp2:1319–1324

Iglesias-Prieto R, Matta JL, Robins WA, Trench RK (1992)Photosynthetic response to elevated temperature in thesymbiotic dinoflagellate Symbiodinium microadriaticumin culture. Proc Nat Acad Sci USA 89:10302–10305

Iglesias-Prieto R, Beltrán VH, LaJeunesse TC, Reyes-BonillaH, Thomé PE (2004) Different algal symbionts explain thevertical distribution of dominant reef corals in the easternPacific. Proc R Soc Lond B 271:1757–1763

Jackson JBC (1994) Constancy and change of life in the sea.Phil Trans R Soc Lond B 344:55–60

Jones R, Hoegh-Guldberg O, Larkum AWD, Schreiber U(1998) Temperature induced bleaching of corals beginswith impairment of dark metabolism in zooxanthellae.Plant Cell Environ 21:1219–1230

Kleypas JA, Buddemeier RW, Archer D, Gattuso JP, LangdonC, Opdyke BN (1999) Geochemical consequences ofincreased atmospheric carbon dioxide on coral reefs.Science 284:118–120

LaJeunesse TC (2001) Investigating the biodiversity, ecology,and phylogeny of endosymbiotic dinoflagellates in thegenus Symbiodinium using the internal transcribed spacerregion: in search of a ‘species’ level marker. J Phycol 37:866–880

LaJeunesse TC (2002) Diversity and community structure ofsymbiotic dinoflagellates from Caribbean coral reefs. MarBiol 141:387–400

LaJeunesse TC, Trench RK (2000) The biogeography of twospecies of Symbiodinium (Freudenthal) inhabiting theintertidal anemone, Anthopleura elegantissima (Brandt).Biol Bull (Woods Hole) 199:126–134

LaJeunesse TC, Loh WKW, van Woesik R, Hoegh-GuldbergO, Schmidt GW, Fitt WK (2003) Low symbiont diversity insouthern Great Barrier Reef corals relative to those of theCaribbean. Limnol Oceanogr 48:2046–2054

LaJeunesse TC, Thornhill DJ, Cox EF, Stanton FG, Fitt WK,Schmidt GW (2004) High diversity and host-specificityobserved among symbiotic dinoflagellates in reef coralcommunities from Hawaii. Coral Reefs 23:596–603

LaJeunesse TC, Lee S, Bush S, Bruno JF (2005) Persistence ofnon-Caribbean algal symbionts in Indo-Pacific mushroomcorals released to Jamaica 35 years ago. Coral Reefs (inpress)

Law R (1985) Evolution in a mutualistic environment. In:Boucher DH (ed) The biology of mutualism: ecology andevolution. Oxford University Press, New York, p 145–170

Little AF, Van Oppen JH, Willis BL (2004) Flexibility in algalendosymbioses shapes growth in reef corals. Science 304:1492–1494

Loh WK, Carter D, Hoegh-Guldberg O (1998) Diversity ofzooxanthellae from scleractinian corals of One Tree Island(The Great Barrier Reef). In: Greenwood JG, Hall NJ (eds)Proceedings of the Australian Coral Reef Society 75thAnniversary Conference, Heron Island. School of MarineSciences, University of Queensland, Brisbane, p 141–150

Loh WK, Loi T, Carter D, Hoegh-Guldberg O (2001) Geneticvariability of the symbiotic dinoflagellates from the wideranging coral species Seriatopora hystrix and Acroporalongicyathus in the Indo-West Pacific. Mar Ecol Prog Ser222:97–107

Marshall PA, Baird AH (2000) Bleaching of corals on theGreat Barrier Reef: differential susceptibilities amongtaxa. Coral Reefs 19:155–163

Pochon X, Pawlowski J, Zaninetti L, Rowan R (2001) Highgenetic diversity and relative specificity among Symbio-

dinium-like endosymbiotic dinoflagellates in soritid fora-miniferans. Mar Biol 139:1069–1078

Pochon X, LaJeunesse TC, Pawlowski J (2004) Biogeographicpartitioning and host specialization among foraminiferan di-noflagellate symbionts (Symbiodinium, Dinophyta). MarBiol 146:17–27

Richmond RH, Hunter CL (1990) Reproduction and recruitmentof corals: comparisons among the Caribbean, the tropicalPacific and the Red Sea. Mar Ecol Prog Ser 60:185–203

Rodriquez-Lanetty M, Loh W, Carter D, Hoegh-Guldberg O(2001) Latitudinal variability in symbiont specificity withinthe widespread scleractinian coral Plesiastrea versipora.Mar Biol 138:1175–1181

Rowan R (1998) Diversity and ecology of zooxanthellae on coralreefs. J Phycol 34:407–417

Rowan R, Powers DA (1991) A molecular genetic classificationof zooxanthellae and the evolution of animal–algal symbio-sis. Science 251:1348–1351

Rowan R, Knowlton N, Baker A, Jara J (1997) Landscape eco-logy of algal symbionts creates variation in episodes of coralbleaching. Nature 388:265–269

Ryland JS (1997) Reproduction in zoanthidea (Anthozoa: Hexa-corallia). Invertebr Reprod Dev 31:177–188

Santos SR, Gutierrez-Rodriquez C, Lasker HR, Coffroth MA(2003) Symbiodinium sp. associations in the gorgonianPseudopterogorgia elisabethae in the Bahamas: high levelsof genetic variability and population structure in symbioticdinoflagellates. Mar Biol 143:111–120

Santos SR, Shearer TL, Hannes AR, Coffroth MA (2004) Fine-scale diversity and specificity in the most prevalent lineage ofsymbiotic dinoflagellates (Symbiodinium, Dinophyceae) ofthe Caribbean. Mol Ecol 13:459–469

Seutin G, White BN, Boag PT (1991) Preservation of avianblood and tissue samples for DNA analyses. Can J Zool69:82–92

Skirving W, Guinotte J (2001) The sea surface temperature storyon the Great Barrier Reef during the coral bleaching event of1998. In: Wolanski E (ed) Oceanographic processes of coralreefs. Physical and biological links in the Great Barrier Reef.CRC Press, Boca Raton, p 301–313

Swofford DL (1999) PAUP*, phylogenetic analysis usingparsimony (*and other methods), Version 4.0b10. Sinauer,Sunderland, MA

Thompson JN (1994) The coevolutionary process. UniversityChicago Press, Chicago, IL

Toller WW, Rowan R, Knowlton N (2001) Zooxanthellae of theMontastraea annularis species complex: patterns of distrib-ution of four taxa of Symbiodinium on different reefs andacross depths. Biol Bull (Woods Hole) 201:348–359

Trench RK (1997) Diversity of symbiotic dinoflagellates and theevolution of microalgal-invertebrate symbioses. Proc 8th IntCoral Reef Symp 2:1275–1286

Van Oppen MJH (2004) Mode of zooxanthella transmissiondoes not affect zooxanthella diversity in acroporid corals.Mar Biol 144:1–7

Van Oppen MJH, Palstra FP, Piquet AMT, Miller DJ (2001) Pat-terns of coral–dinoflagellate associations in Acropora: signif-icance of local availability and physiology of Symbiodiniumstrains and host-symbiont selectivity. Proc R Soc Lond B268:1759–1767

Veron JEN (1995) Corals in space and time. University of NewSouth Wales Press, Sydney

Veron JEN (2000) Corals of the world, Vol 1–3. Australian Insti-tute of Marine Science, Townsville

Warner EM, Fitt WK, Schmidt GW (1999) Damage to photosys-tem II in symbiotic dinoflagellates: a determinant of coralbleaching. Proc Natl Acad Sci USA 96:8007–8012

161

Editorial responsibility: Otto Kinne (Editor), Oldendorf/Luhe, Germany

Submitted: March 24, 2004; Accepted: July 27, 2004Proofs received from author(s): December 16, 2004

closely related symbiodinium spp. differ in relative

Documents