Vectored dispersal of Symbiodinium by larvae of a
Caribbean gorgonian octocoral
HERMAN H. WIRSHING,* ,? KEVIN A. FELDHEIM? and ANDREW C. BAKER*
*Division of Marine Biology and Fisheries, Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600
Rickenbaker Causeway, Miami, FL 33149, USA, ?Pritzker Laboratory for Molecular Systematics and Evolution, Field Museum
of Natural History, 1400 S. Lake Shore Drive, Chicago, IL 60605, USA
Abstract
The ability of coral reefs to recover from natural and anthropogenic disturbance is
difficult to predict, in part due to uncertainty regarding the dispersal capabilities and
connectivity of their reef inhabitants. We developed microsatellite markers for the
broadcast spawning gorgonian octocoral Eunicea (Plexaura) flexuosa (four markers) and
its dinoflagellate symbiont, Symbiodinium B1 (five markers), and used them to assess
genetic connectivity, specificity and directionality of gene flow among sites in Florida,
Panama, Saba and the Dominican Republic. Bayesian analyses found that most
E. flexuosa from the Florida reef tract, Saba and the Dominican Republic were strongly
differentiated from many E. flexuosa in Panama, with the exception of five colonies
from Key West that clustered with colonies from Panama. In contrast, Symbiodinium
B1 was more highly structured. At least seven populations were detected that showed
patterns of isolation by distance. The symbionts in the five unusual Key West colonies
also clustered with symbionts from Panama, suggesting these colonies are the result of
long-distance dispersal. Migration rate tests indicated a weak signal of northward
immigration from the Panama population into the lower Florida Keys. As E. flexuosa
clonemates only rarely associated with the same Symbiodinium B1 genotype (and vice
versa), these data suggest a dynamic host?symbiont relationship in which E. flexuosa
is relatively well dispersed but likely acquires Symbiodinium B1 from highly struc-
tured natal areas prior to dispersal. Once vectored by host larvae, these symbionts may
then spread through the local population, and/or host colonies may acquire different
local symbiont genotypes over time.
Keywords: Caribbean, dispersal, octocoral, population genetics, specificity, Symbiodinium
Received 17 March 2013; revision received 17 May 2013; accepted 21 May 2013
Introduction
The temporal and spatial scales over which coral reef
ecosystems are connected are fundamental to under-
standing their evolutionary history (Hellberg 2007) and
resilience to natural and anthropogenic stressors (Jones
et al. 2009). The extent of dispersal of planktonic larval
stages, whether by large or small-scale oceanographical
features, and/or differential larval behaviour, largely
determines the connectivity of coral reef habitats
(Roberts 1997; Cowen et al. 2000, 2006). Because coral
reefs worldwide have experienced ecological degrada-
tion, both chronic and acute, over the last few decades
(Gardner et al. 2003; Hughes et al. 2003; Bellwood et al.
2004; Bruno & Selig 2007; De?ath et al. 2012), there is a
growing need for information regarding coral reef con-
nectivity that will help to better understand the natural
history of reef invertebrates (Hedgecock et al. 2007) and
maximize conservation efforts (Palumbi 2003; Van Op-
pen & Gates 2006; Jones et al. 2007).
Octocorals, together with a variety of other reef-
inhabiting invertebrates, form symbiotic relationships
Correspondence: Herman H. Wirshing, Fax: 202 633 0182;
E-mail: wirshingh@si.edu.
?Present address: Department of Invertebrate Zoology,
National Museum of Natural History, Smithsonian Institution,
Washington, DC 20560-0163, USA
? 2013 John Wiley & Sons Ltd
Molecular Ecology (2013) 22, 4413?4432 doi: 10.1111/mec.12405
with dinoflagellates in the genus Symbiodinium. This
mutualistic partnership benefits both members, and its
maintenance is critical to their survival (Baker 2003;
Coffroth & Santos 2005). Worldwide, octocorals are host
to Symbiodinium in clades A, B, C, D and G (sensu
Rowan & Powers 1991; Pochon & Gates 2010; Van
Oppen et al. 2005; Goulet et al. 2008), while in the Carib-
bean, most octocoral species (~88%) host only Symbiodi-
nium in clade B [genotyped as B1 using internal
transcribed spacer 2 (ITS-2), sensu LaJeunesse (2001);
and B184 using chloroplast large subunit ribosomal
DNA (cp23S-rDNA), sensu Santos et al. (2003a); referred
to here as ?Symbiodinium B1?] (Goulet & Coffroth 2004).
However, although most octocorals in the Caribbean
only host Symbiodinium B1, at least as the dominant
symbiont, ITS haplotype diversity within this subgroup
is disproportionately high compared with other Symbi-
odinium clades (Van Oppen et al. 2005), and microsatel-
lite analyses show substantial genotypical diversity
within Symbiodinium B1 among different octocoral hosts
(Santos et al. 2003b; Kirk et al. 2009; Andras et al. 2011).
Among some octocorals, specificity at the level of Sym-
biodinium clade is low during early ontogeny, and lar-
vae and newly settled polyps can acquire Symbiodinium
from at least three different clades (A, B and C;
Coffroth et al. 2001). Yet, over several months, cladal
diversity is winnowed to only clade B, and adult gorgo-
nian octocorals are always dominated by members of
this clade. However, adult colonies of the octocoral
Briarium are able to acquire different Symbioninium B1
types from the external environment after experimental
bleaching, suggesting some level of flexibility among
adult octocorals to acquire different Symbiodinium B1
types in response to environmental change (Lewis &
Coffroth 2004). Despite this, at the level of the individ-
ual, adult octocoral colonies appear to maintain an
acute genetic specificity to their symbionts at not only
the level of clade, but also at the level of genotype over
space and time (Goulet & Coffroth 2003a,b; Kirk et al.
2005; Hannes et al. 2009).
Gorgonian (or branching) octocorals are ecologically
important and abundant members of tropical reef envi-
ronments in the Caribbean/tropical western Atlantic
(e.g. Goldberg 1973; Lasker & Coffroth 1983; Sanchez
et al. 1997). Genetic studies from the region show differ-
ent patterns of population structure among gorgonian
octocorals, and the Symbiodinium they host. For exam-
ple, Pseudopterogorgia elisabethae exhibits significant
genetic structure in the Bahamas among sites separated
by distances up to 100 km (Gutierrez-Rodriguez &
Lasker 2004). However, the dominant symbiont of this
species, Symbiodinium B1, shows more extreme popula-
tion subdivision among similar sites in the region sepa-
rated by 10s of kilometres (Santos et al. 2003b). Similarly,
populations of another Caribbean/western Atlantic gor-
gonian octocoral, Gorgonia ventalina and its Symbiodinium
B1 symbionts are structured differently across their geo-
graphical range (Andras et al. 2011, 2013).
The Florida reef tract (FRT) is a ~260-km-long system
of coral reefs located at the northern latitudinal limit of
coral reef formation in the Caribbean (Spalding et al.
2001). Because much of the FRT has experienced wide-
spread ecological and habitat deterioration (Pandolfi
et al. 2005; Wilkinson & Souter 2008), larval replenish-
ment of degraded sites from intact sites upstream may
be critical for the persistence and sustainability of this
reef system. Studies of genetic connectivity along the
FRT, comprising taxonomically diverse invertebrates
from different phyla (e.g. Porifera, Cnidaria, Arthro-
poda and Echinodermata) with a variety of life history
characteristics, indicate a generally well-mixed system
with only modest population subdivision, at least using
markers available to date (e.g. Richards et al. 2007;
Vollmer & Palumbi 2007; Baums et al. 2010; Debiasse
et al. 2010; Hemond & Vollmer 2010). One exception to
this is Symbiodinium B1, which microsatellite analysis
has shown to be strongly subdivided over small spatial
scales (10s of km) along the FRT over horizontal (simi-
lar depths) and vertical (different depths) distances,
although intermittent gene flow and limited dispersal
are also likely to occur (Kirk et al. 2009). The gorgonian
octocoral Eunicea (Plexaura) flexuosa is a common species
in reef and hard-bottom environments throughout the
FRT and greater Caribbean/western Atlantic (Bayer
1961; Jaap 1984). Like most gorgonian octocorals of this
region, it only hosts Symbiodinium B1 as its dominant
dinoflagellate symbiont (Goulet & Coffroth 2004). As a
gonochoric broadcast spawner, E. flexuosa has the
potential for long-distance larval dispersal, with male
and female colonies releasing their gametes after the
summer full moons (Beiring & Lasker 2000). In this
study, microsatellite markers were developed for both
E. flexuosa and its dinoflagellate symbiont, Symbiodinium
B1, and were used for analyses of population structure,
spatial correlation, migration rates and genotypical
specificity among host and symbionts of the same colo-
nies to (i) quantify their genetic structure (and therefore
evaluate ecologically relevant dispersal and directional-
ity), (ii) assess their resilience capacity and (iii) measure
host/symbiont specificity among E. flexuosa and Symbi-
odinium B1 genotypes.
Materials and methods
Sample collection and microsatellite development
Samples of Eunicea flexuosa were collected from nine
sites along the Florida Keys and Biscayne Bay reef
? 2013 John Wiley & Sons Ltd
4414 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
tracts; three sites within the Bocas del Toro Province,
Panama; one site from Punta Cana, Dominican Repub-
lic; and Saba Bank (Fig. 1; Table 1). Only host colonies
>30 cm were collected, and sampling occurred over a
period of 3 years and 7 months. As growth rates in
E. flexuosa average ~1?2 cm/year (Beiring & Lasker
2000), all sampled colonies were likely >10 years old.
Individual colonies were sampled by clipping a ~3-cm
branch tip from one of the apical branches of the colony
and preserved in 95% ethanol or saline DMSO (Seutin
et al. 1991). To reduce the possibility of sampling the
same colony twice, a 15-m transect line was used on
sites with sufficient densities of the target species. On
sites with low colony densities, a random collecting
approach was used. If two divers were collecting on the
same site, they examined the area in opposite direc-
tions.
In order to increase the ratio of host (E. flexuosa) to
symbiont (Symbiodinium B1) DNA in total (host + sym-
biont) DNA extractions (Shearer et al. 2005), live
E. flexuosa branch clippings (~7 cm) were first bleached
using a photosynthetic inhibitor (DCMU; Jones 2004),
combined with increased temperature and irradiance.
Microsatellite isolation followed the approach of Glenn
& Schable (2005), which employs Dynabeads
 (Life
Technologies, Carlsbad, CA, USA) to sequester DNA
fragments with microsatellite regions with tri- or tetra-
nucleotide repeats. Sequences with microsatellite
regions were identified, and forward and reverse prim-
ers were designed using PRIMER3 (v. 0.4.0; Rozen &
Skaletsky 2000) and GENEIOUS PRO 4.8.5 (Drummond et al.
2008).
For screening of E. flexuosa- and Symbiodinium B1-
specific markers, tissue from the donor colony used for
microsatellite isolation was also harvested to isolate and
culture its symbionts. Denaturing gradient gel electro-
phoresis analysis (LaJeunesse 2001) of amplified ITS-2
region of rDNA genotyped the cultured symbiont as
clade B1 (see below for PCR protocol). Other Symbiodi-
nium types were also screened against the microsatellite
marker candidates from existing cultures (provided by
Scott Santos, Auburn University www.auburn.edu/
~santosr/index.php), including two from clade B (cul-
ture ID: PurPflex and 703) and one each from clades A
(ID: 719), C (ID: PtBr) and D (ID: 013).
Molecular analyses and marker validation
Total genomic DNA was extracted from each field-
collected sample using a modified organic protocol
Saba Bank
Punta Cana,
Dominican Republic
(A)
(B)
Caribbean Sea
Western Atlantic
500 km
N
Emerald Reef
Fowey Rocks
Long Reef
North Dry Rocks
Key Largo Patch
East Washerwoman Shoal
Delta Shoal
Western Sambo
Marker 32 50 km
Panama Mianland
Drago
Hospital 
Point
Crawl Cay
5 km
(A)
(B)
Florida Current
Caribbean Current
Fig. 1 Map of 14 sites across the Caribbean and western Atlantic where Eunicea flexuosa samples were collected. Shaded boxes are
enlarged for (A) the Florida reef tract and (B) Bocas del Toro, Panama sites. Arrows indicate the direction of the prevailing sea sur-
face currents discussed in the text.
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4415
(Baker et al. 1997). Host- and symbiont-specific primers
were fluorescently labelled on the forward primer. PCR
(10 lL volume) was conducted with 10 pmol of each
primer, 200 lM dNTPs, 2 mM MgCl2 and 1 lL of geno-
mic DNA using 0.6 U of GoTaq
 (Promega Inc.) poly-
merase and the manufacturer?s buffer. Thermal cycling
consisted of an initial denaturing step of 94 ?C for
3 min, followed by 35 rounds of 94 ?C for 1 min, 57 ?C
for 1 min and 74 ?C for 1 min, and a final extension of
72 ?C for 7 min. PCR products were analysed using an
ABI 3730 DNA Analyzer (Applied Biosystems) from the
Core Laboratories Center at Cornell University. Sizing
was achieved using an internal standard (Gene Scan
500-Liz; Applied Biosystems), and alleles were scored
using GENEMAPPER software 4.0 (Applied Biosystems).
For host-specific diploid microsatellite markers,
Hardy?Weinberg equilibrium (HWE) exact tests (Guo &
Thompson 1992) were implemented using ARLEQUIN
v.3.5.1.2 (Excoffier & Lischer 2010). For this test, we
used a Markov chain method, using the default param-
eter options, to determine the level of significance.
Additionally, chi-square tests for HWE, following the
methods of Hedrick (2000), were performed using GENE-
ALEX v.6.41 (Peakall & Smouse 2006). MICRO-CHECKER (Van
Oosterhout et al. 2004) was used to test for the presence
of null alleles using Bonferroni-adjusted 95% confidence
intervals. We tested for linkage disequilibrium for host-
and symbiont-specific markers using GENEPOP v.4.0
(Rousset 2008) with default parameters for the Markov
chain options.
Population structure, spatial analyses and specificity
Population structure was investigated using a Bayesian
clustering approach performed in STRUCTURE v.2.3.3
(Pritchard et al. 2000) using the Web-based Bioportal
server from the University of Oslo (www.bioportal.uio.
no). We employed ARLEQUIN v.3.5.1.2 (Excoffier &
Lischer 2010) to calculate pairwise FST values, using
Slatkin?s (1995) genetic distance with 1000 permutations
to determine significance, and analysis of molecular
variance (AMOVA; Excoffier et al. 1992) to examine
genetic partitioning among regions, sites within regions
and individuals within sites as defined in Table 1 (but
with the Florida Keys and Biscayne Bay pooled as one
region). For STRUCTURE, model options were set to allow
admixture, assumed allele frequencies to be correlated
(Falush et al. 2003), and did not prespecify populations
of origin. Because sampling location information set as
prior information can assist clustering for data sets with
few markers, few individuals or very weak structure
(Hubisz et al. 2009), the LOCPRIOR option was used.
The three sites from Biscayne Bay and each of the two
sites from the upper keys, middle keys and the lower
keys (Table 1) were pooled into separate ?locations?.
For E. flexuosa (?eflex? data sets), analyses were
initially performed using only those markers found to
be in HWE with duplicate genotypes removed. For
comparison, additional runs were performed that
included a marker found to deviate from HWE (Plfl7,
see Results). As gorgonian octocoral hosts may contain
more than one haploid Symbiodinium B1 genotype (Kirk
et al. 2009; Andras et al. 2011), analyses for Symbiodini-
um B1 were performed using two separate data sets.
One data set, ?singlehaps?, contained only those individ-
uals genotyped with a single allele for each marker.
This allowed analyses based on allele size. A second
data set, ?totalmatrix?, contained individuals with both
single alleles and those genotyped with multiple haplo-
types with at least one marker. This data set was scored
Table 1 Collection sites, their geographical regions and subregions, GPS position, total numbers collected and collection date
Region Subregion Site
GPS
n Date collectedLatitude Longitude
Biscayne Bay, Florida Offshore reefs Emerald Reef N25 40.450 W080 05.920 54 July 2010
Fowey Rocks N25 35.417 W080 05.800 65 August 2009
Long Reef N25 24.734 W080 07.642 76 November 2009
Florida Keys Upper Keys North Dry Rocks N25 08.180 W080 17.359 52 May 2010
Key Largo Patch N25 05.580 W080 19.380 82 May 2010
Middle Keys East Washerwoman Shoal N24 40.000 W081 04.437 70 August 2010
Delta Shoal N24 38.048 W081 05.342 77 August 2010
Lower Keys Western Sambo N24 28.709 W081 43.812 54 August 2010
Marker 32 N24 28.450 W081 44.561 42 August 2010
Panama Bocas del Toro Drago N09 24.833 W082 19.997 9 August 2007
Hospital Point N09 19.798 W082 13.264 25 August 2007
Crawl Cay N09 15.696 W082 07.389 21 August 2007
Saba Bank Conch Valley N17 20.000 W063 14.844 12 October 2008
Dominican Republic Punta Cana N18 31.953 W068 21.074 22 March 2011
? 2013 John Wiley & Sons Ltd
4416 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
as a presence/absence matrix (1s and 0s) of all possible
alleles found at each locus for each individual, similar
to analyses of amplified fragment length polymor-
phisms. STRUCTURE options for the ?totalmatrix? data set
were the same as the ?singlehaps? data set, but included
the RECESSIVEALLELES=1 option, and ?0s? set as the
recessive allele (Pritchard 2010). Clonal (duplicate) hapl-
otypes were not removed from either Symbiodinium B1
data set, as their inclusion does not alter STRUCTURE anal-
yses, and more accurately represents Symbiodinium
diversity among individual colonies (Andras et al.
2011). For both E. flexuosa and Symbiodinium B1, Markov
chain runs consisted of an initial ?burn-in? (values dis-
carded) of 2 9 106 steps followed by a final 2 9 106
iterations. Three independent runs were performed for
1?12 Ks. The number of population clusters (K) was
chosen using the ?K method (Evanno et al. 2005), as
implemented in STRUCTURE HARVESTER v0.6.8 (Earl &
vonHoldt 2011), and a ?standard? approach (Pritchard
2010) in which the most likely Ks are chosen among the
least negative plots of the estimated log probability of
the data vs. each of the 12 Ks. Among the least negative
plots, the most likely K is then chosen as the one that
most robustly recovers structure in the data and is bio-
logically reasonable. CLUMPP v.1.1.2 (Jakobsson & Rosen-
berg 2007) was used to fuse the results of the three
independent runs for the chosen K, and DISTRUCT v1.1
(Rosenberg 2004) was used to visualize the results as
the probability of each individual?s membership to K
populations.
Correlations between geographical and genetic dis-
tance (i.e. isolation by distance) and between genetic
distances of the host and symbiont were tested using
Mantel tests (Mantel 1967) performed with GENALEX
v.6.41 (Peakall & Smouse 2006). BAYESASS v1.3 (Wilson &
Rannala 2003a) was used to determine directionality of
gene flow among the sampling locations by estimating
recent migration rates of individuals (over the last few
generations). Default settings were used for number of
iterations (3 000 000), ?burn-in? (999 999) and sampling
frequency (2000).
To examine host/symbiont genotypical specificity,
genetic clones for E. flexuosa and Symbiodinium B1 were
identified and assigned a genotype label using GENEALEX
v.6.41 (Peakall & Smouse 2006). Eunicea flexuosa clones
were compared with the genotypes of their Symbiodini-
um B1 partners, and vice versa. Probability of identity
(PI; the average probability of randomly drawing two
individuals from a population that, by chance, have
identical genotypes) analyses were performed in GENALEX
v.6.41 to evaluate the ability of the E. flexuosa markers to
reliably identify clones. Fisher?s exact tests (Fisher 1944)
were used to test for nonrandom associations among
Symbiodinium B1 and E. flexuosa genotypes using 2 9 4
contingency tables calculated with In-Silico Online
(Joosse 2011). Structure outputs were used to identify
the most genetically dissimilar host and symbiont geno-
types.
Results
Microsatellite markers for E. flexuosa and
Symbiodinium B1
Fifty-two sequenced clones were obtained with micro-
satellite regions containing either tri- or tetra-nucleotide
repeats. After screening for Symbiodinium (see Materials
and Methods), four were E. flexuosa specific (~8%)
(Table 2); 17 (~33%) were specific to Symbiodinium B1; 18
(~35%) robustly PCR-amplified the host Symbiodinium
B1 culture plus at least one of the other Symbiodinium
cultures; and the remaining clones performed poorly
with PCR. Of the Symbiodinium B1-specific candidates
that were variable and amplified well across all sampled
locations (data not shown), five were used for analyses
(Table 3).
For E. flexuosa, exact tests for heterozygote deficien-
cies (Table S1, Supporting Information) and chi-square
tests (Table S2, Supporting Information) performed for
each marker and sampling location (which represent
arbitrarily chosen populations) yielded departures from
HWE at some sites. Only Plfl7 showed consistent and
highly significant departures from HWE (P < 0.001) for
all sites using both tests. Positive FIS values were nota-
bly greater (0.40?0.79) with Plfl7 compared with the
other three markers. Tests for null alleles were signifi-
cant at all sampling sites for Plfl7, indicating missed
alleles likely influenced the homozygote excess found
with this marker. In addition, Marker 32 and Panama
were significant for null alleles for all of the markers
(except Plfl67 at Marker 32). However, the bias towards
null alleles at these sites may have been influenced by
individuals whose genotypes were strongly assigned to
separate clusters (see Fig. 2A). Tests of linkage disequi-
librium for each E. flexuosa marker revealed no signifi-
cant deviations from linkage equilibrium for all sites
except Panama (Table S3, Supporting Information). At
this site, three of the six marker pairs showed signifi-
cant departures from linkage equilibrium.
For Symbiodinium B1, 53% (n = 326) contained single
alleles, and 47% (n = 284) contained at least one multi-
ple allele at any locus. Tests of linkage disequilibrium
(?singlehaps? data set) found significant deviations from
linkage equilibrium for all, or most, marker pairs at six
of the 12 sites (Table S4, Supporting Information).
However, when individuals within each sampling site
were pooled into their population clusters as deter-
mined by STRUCTURE (see below), only one site (Panama)
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4417
T
ab
le
2
L
oc
u
s
n
am
e,
p
ri
m
er
se
qu
en
ce
an
d
m
ar
ke
r
ch
ar
ac
te
ri
st
ic
s
of
fo
u
r
m
ic
ro
sa
te
lli
te
lo
ci
in
E
un
ic
ea
fle
xu
os
a
L
oc
u
s
N
am
e
P
ri
m
er
Se
qu
en
ce
R
ep
ea
t
m
ot
if
T
a
(?
C
)
Si
ze
ra
n
ge
(b
p
)
N
u
m
be
r
of
al
le
le
s
H
O
H
E
P
lfl
19
F-
FA
M
_C
A
A
C
A
T
C
G
T
C
A
C
C
A
G
T
C
A
C
C
(T
C
A
) 2
C
C
A
(T
C
A
) 3
C
C
A
A
(T
C
A
) 4
A
T
A
A
(T
C
A
) 3
A
T
C
A
(T
C
A
) 2
A
T
A
A
(T
C
A
) 2
A
T
A
A
(T
C
A
) 2
A
(T
C
A
) 3
57
13
6?
29
8
25
0.
46
(S
E
0.
01
7)
0.
50
(S
E
0.
03
3)
R
-
T
G
G
A
T
T
G
T
G
G
T
T
G
G
A
C
A
G
T
G
P
lfl
67
F-
V
IC
_A
T
T
T
A
A
C
G
T
A
A
T
T
C
A
G
C
C
T
C
T
G
G
(T
A
T
C
) 6
57
17
9?
25
1
18
0.
50
(S
E
0.
01
5)
0.
60
(S
E
0.
02
5)
R
-
C
C
A
C
A
A
A
T
C
A
T
T
T
A
G
T
C
A
T
A
T
T
G
C
P
lfl
19
9
F-
N
E
D
_G
C
G
T
T
T
C
G
T
T
C
A
G
G
C
T
T
T
A
G
(T
T
G
) 5
57
13
4?
29
6
13
0.
22
(S
E
0.
05
5)
0.
29
(S
E
0.
05
8)
R
-
T
G
C
A
G
C
A
T
G
G
T
C
A
A
G
A
T
A
C
C
P
lfl
7
F-
P
E
T
_T
A
G
T
G
G
G
A
A
T
G
C
A
C
A
T
C
T
C
G
(A
A
C
) 6
A
A
G
(A
C
C
) 2
A
A
G
(A
A
C
) 2
A
A
G
(A
A
C
) 2
57
13
9?
22
6
19
0.
32
(S
E
0.
02
5)
0.
80
(S
E
0.
01
2)
R
-
G
C
T
T
C
C
G
A
G
A
T
A
G
T
T
T
G
T
A
G
G
G
T
ab
le
3
L
oc
u
s
na
m
e,
p
ri
m
er
se
qu
en
ce
an
d
m
ar
ke
r
ch
ar
ac
te
ri
st
ic
s
of
fi
ve
Sy
m
bi
od
in
iu
m
ty
p
e
B
1
m
ic
ro
sa
te
ll
it
e
lo
ci
(h
os
t?
E
un
ic
ea
fle
xu
os
a)
L
oc
u
s
na
m
e
P
ri
m
er
se
qu
en
ce
R
ep
ea
t
m
ot
if
T
a
(o
C
)
Si
ze
ra
n
ge
(b
p
)
N
u
m
be
r
of
al
le
le
s
(s
in
gl
e
h
ap
lo
ty
p
es
)
N
u
m
be
r
of
al
le
le
s
(m
u
lt
ip
le
h
ap
lo
ty
p
es
)
P
lfl
sy
m
17
F-
P
E
T
_A
G
G
C
T
G
C
A
G
A
C
A
C
A
A
A
T
G
C
(A
A
C
) 2
4
57
15
0?
29
1
33
36
R
-
T
T
T
G
T
C
T
C
A
A
T
G
G
C
A
T
C
A
G
C
P
lfl
sy
m
21
F-
FA
M
_A
A
T
C
A
T
T
T
T
G
G
A
A
G
G
C
G
A
T
G
(A
A
C
) 2
0(
A
A
G
) 5
57
14
7?
25
8
22
25
R
-
T
G
A
G
T
G
G
G
A
A
T
G
A
A
C
T
T
G
T
G
A
P
lfl
sy
m
19
6
F-
V
IC
_C
T
T
G
A
T
C
G
C
A
T
G
T
G
C
A
T
C
T
C
(T
T
G
) 2
3
57
14
5?
21
4
20
23
R
-
C
C
G
G
A
T
T
C
G
T
G
T
T
T
C
A
A
G
A
T
P
lfl
sy
m
21
1
F-
N
E
D
_G
C
G
G
A
T
A
T
G
G
T
T
T
C
T
T
G
G
A
G
(A
T
C
) 1
5A
C
C
(A
T
C
) 1
2
57
17
7?
30
0
27
31
R
-
C
C
C
C
C
T
T
T
T
G
A
A
A
G
T
G
A
A
C
A
P
lfl
sy
m
71
/
72
F-
P
E
T
_G
A
C
C
T
T
G
C
C
A
A
T
T
C
A
T
G
T
C
C
(T
A
C
A
) 9
T
T
C
A
(T
A
C
A
) 2
T
T
C
A
(T
A
C
A
) 5
57
12
5?
23
7
14
21
R
-
G
A
C
A
T
G
A
C
A
T
G
A
C
A
T
G
A
A
A
T
G
C
T
a,
an
n
ea
li
n
g
te
m
p
er
at
u
re
.
? 2013 John Wiley & Sons Ltd
4418 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
demonstrated significant departures from linkage equi-
librium, with seven of the 10 primer pairs (Table S5,
Supporting Information).
Population structure and spatial analyses
For E. flexuosa, duplicate genotypes (i.e. clones) were
found within and among collection sites and were
removed from each sampling location. As a result, a par-
ticular genotype could only be found once per sampling
location, but could also occur once in other sampling
locations. For Bayesian analyses in STRUCTURE of the
?eflex? data sets, ?K plots and a ?standard? approach (see
Materials and Methods) for determining the number of
K population clusters indicated K to be 2 for data sets
run with markers in HWE and with an additional mar-
ker (Plfl7; ?K method only) found to deviate from HWE.
Population assignments were the same for both data sets
at this value of K (Fig. 2A). All of the Florida Keys/Bis-
cayne Bay sites, in addition to Punta Cana, Dominican
Republic, were strongly assigned to the same population
cluster (colour = blue). Panama contained individuals
with high membership probabilities (>0.90) to a second
yellow population cluster, in addition to the blue popu-
lation cluster. Five individuals from Marker 32 (lower
keys) were also assigned to the yellow population clus-
ter, but with lower membership probabilities (0.40?0.83).
Saba individuals showed strong membership to the blue
cluster (>0.90) except for one individual with a robust
assignment to the yellow cluster (0.89).
Eunicea flexuosa estimates of pairwise FST for all loca-
tions ranged from 0.000 to 0.149 (overall = 0.020; Table
S6, Supporting Information) and revealed a similar pat-
tern of population differentiation to that of STRUCTURE.
For example, paired FRT sites showed no significant FST
values with the exception of Marker 32 (lower keys)
with three other sites (Fowey Rocks, FST 0.048; Delta
Shoal, FST 0.033; Western Sambo, FST 0.072; Ps < 0.05),
and Panama FSTs differed significantly from all other
sites (FSTs 0.056?0.119, P < 0.05) except Marker 32 (FST
0.003, P > 0.05). However, two FRT sites (Fowey Rocks
and Western Sambo) showed significant population
structure with the Dominican Republic (FSTs 0.049,
0.045; P < 0.05). Similarly, AMOVAs indicated a significant
partitioning among regions (e.g. FRT and Panama;
P < 0.000) and individuals within sites (P = 0.001), but
not among sites within regions (e.g. sites within the
FRT and Panama; P = 0.093; Table 4).
For Symbiodinium B1, the ?singlehaps? and ?totalma-
trix? data sets revealed the value of K populations to be
2 and 3, respectively, for the ?K method and 7 for the
?standard? approach with both data sets. At K = 2, Sym-
biodinium B1 group into two principal clusters com-
prised of Panama differentiated from most of the FRT,
Dominican Republic, and Saba (Fig. S1, Supporting
Information). However, at each increasing value of K
from 2 to 7, both data sets show individuals with strong
membership probabilities (>0.90) to new, and more sub-
tle, population clusters up to K = 7. At K = 8, admix-
ture begins to confound the results. The difference
Florida Reef Tract
Lower
Keys
Middle Keys Upper Keys Biscayne BayPanama DR SB
M32 WS DS EWS KLP NDR LR FR ER PC CVDG HP CC
Florida Reef Tract
Lower Keys Middle Keys Upper Keys Biscayne BayPanama DR SB
M32 WS DS EWS KLP NDR LR FR ER PC CVDG HP CC
(A)
(B)
(C)
Florida Reef Tract
Lower Keys Middle Keys Upper Keys Biscayne Bay Panama DR SB
M32 WS DS EWS KLP NDR LR FR ER PC CVDG HP CC
Fig. 2 Graphical depiction of population structure for (A) Eunicea flexuosa (K = 2), (B) Symbiodinium B1 for the ?singlehaps? (K = 7)
and (C) ?totalmatrix? (K = 7) data sets as inferred by Bayesian clustering analyses. Each vertical line represents one individual, with
that individual?s assignment fraction to each of K population clusters.
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4419
between the values of K for the ?K and the ?standard?
approach may be the result of population structuring in
Symbiodinium B1 that is not effectively hierarchical. The
?K method largely captures the highest hierarchical
level among populations (Evanno et al. 2005) and may
therefore underestimate population structure in groups
that do not conform to this type of subdivision
(Kalinowski 2011), including octocorals (Aurelle et al.
2011; Andras et al. 2013). Therefore, despite the DK
method identifying K = 2 or 3 as the most likely num-
ber of populations, we believe that K = 7, as predicted
by the ?standard? approach, more robustly captures the
complexity of the genetic signal of Symbiodinium B1 for
both ?singlehaps? and ?totalmatrix? data sets (Fig. 2B, C).
For the ?singlehaps? data set (Fig. 2B), each sampling
location (i.e. reef) along the FRT contained Symbiodinium
B1 strongly assigned (>0.90) to up to four population
clusters. Saba and the Dominican Republic showed
moderate signs of admixture, indicating a weaker sig-
nal, but were principally assigned to the yellow cluster
(>0.50). No population subdivision was detected among
Symbiodinium B1 from three sites at Bocas del Toro,
Panama (red cluster). A separate analysis consisting of
only the Bocas del Toro sites similarly did not detect
any subdivision (data not shown). Population assign-
ments based on the ?totalmatrix? data set (Fig. 2C) were
virtually identical to those of the ?singlehaps? data set
(Fig. 2B). An exception was the assignment of five indi-
viduals from the lower keys site, Marker 32, to the red
cluster (>0.50). Both host and symbiont of these five col-
onies were assigned to population clusters that were
primarily assigned to samples from Panama (yellow
and red, respectively; Fig. 3). Symbionts of three other
individuals spanning the FRT and two individuals from
Saba were also assigned to the red cluster.
Symbiodinium B1 pairwise FST values were largely
significant (Table S6, Supporting Information). With the
?singlehaps? data set, FST values ranged from 0.000 to
0.431 (overall = 0.129) and showed high population
subdivision with 77.3% of paired sites exhibiting signifi-
cant differentiation. The ?totalmatrix? data set FST values
ranged from 0.010 to 0.298 (overall = 0.109), and 93.9%
of paired sites were significant. AMOVA analyses of both
Symbiodinium B1 data sets also indicated acute genetic
partitioning with significant differences found among
regions (P < 0.000), sites within regions (P < 0.000) and
individuals within sites (P = 0.025, ?singlehaps? only;
Table 4).
The robustness of population assignments for both
E. flexuosa and Symbiodinium B1 data sets was visualized
by pooling only those individuals with assignment
probabilities >0.90 (with a few exceptions) at each sam-
pling location and mapping them over their geographi-
cal location (Figs 4 and 5; Tables S7 and S8, Supporting
Information). Mantel tests for pairwise comparisons
among all of the sampling locations in the data set
(FRT, Panama, Saba and Dominican Republic) were not
significant for E. flexuosa, although the trend was posi-
tive (Rxy = 0.272, P = 0.150). Similarly, E. flexuosa com-
parisons among sites exclusive to the FRT were also
nonsignificant (Rxy = 
0.103, P = 0.250). In contrast, for
Symbiodinium B1, Mantel tests among all of the
sampling locations revealed a significant trend of
increasing genetic distance with geographical distance
(Rxy = 0.691, P-value = 0.010). Sites within the FRT also
displayed a significant positive trend (Rxy = 0.459,
P-value = 0.010). In addition, Mantel test comparing
pairwise genetic distances of E. flexuousa with
Symbiondinium B1 by reef site across the FRT showed
no significant correlation (Rxy = 
0.222, P-value = 0.160)
Table 4 AMOVA results for Eunicea flexuosa (?eflex? data set with marker Plfl7 and duplicates removed) and Symbiodinium B1 (?single-
haps? and ?totalmatrix? data sets)
Source of variation d.f. Sum of squares % Of variation Fixation indices P value
?eflex?
Among regions 3 4.54 5.41 0.061 0.000*
Among sites within regions 10 4.02 0.73 0.008 0.093
Among individuals within sites 782 209.91 93.85 0.054 0.001*
?singlehaps?
Among regions 3 7.95 8.56 0.099 0.000*
Among sites within regions 8 12.55 9.01 0.176 0.000*
Among individuals within sites 305 117.8 82.42 0.086 0.025*
?totalmatrix?
Among regions 3 32.18 2.15 0.104 0.000*
Among sites within regions 8 126.1 10.15 0.123 0.000*
Among individuals within sites 599 1209.44 87.7 0.022 0.166
Regions and sites are as defined in Table 1 except with all Florida sites pooled as one region.
*Significance (P < 0.05).
? 2013 John Wiley & Sons Ltd
4420 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
and suggests no relationship among the genetic struc-
ture of E. flexuosa and Symbiodinium B1.
Migration rates
For E. flexuosa, outputs from BAYESASS were used to
determine directionality of recent gene flow among
three principle regions, namely Panama (SW Caribbean),
eastern Florida (western Atlantic) and Saba/Dominican
Republic (NE Caribbean; Table 5). Of the nine sampling
locations across the FRT, seven contained ~67% nonim-
migrants (i.e. self-seeding). This is the minimum amount
allowed by BAYESASS and suggests that there is not
enough genetic differentiation among these sites (with
these markers) to robustly assess migration (i.e. the sites
are genetically well mixed; Wilson & Rannala 2003b).
However, Emerald Reef and North Dry Rocks were 99%
and 95% self-seeding (CIs straddled the upper null CI
limit), respectively, which suggests sufficient resolution
to detect migration, although less reliably for North Dry
Rocks than Emerald Reef. These two locations were the
most likely source of ~8?29% of immigrants to 8 reefs
within the FRT (six reefs with CIs outside the null, and
two reefs with CIs that straddled the upper null CI
limit). The overall directionality within the FRT was
from north to south. Unexpectedly, FRT sites were also
the source of 6?15% of immigrants to sites in Panama,
Saba and Dominican Republic, although their CIs indi-
cated low reliability. This may represent an artefact of
limited sampling from other candidate sites in the
region and/or insufficient marker resolution. Con-
versely, Saba and the Dominican Republic were not a
significant source of immigrants to either Panama or the
FRT (?1%). Immigrants from Panama into all other loca-
tions were <1%, with the exception of Marker 32 (3%)
and Saba (2%), but contained unreliable CIs. Although
the signal was weak, the trend of migration from Pan-
ama into the FRT via Marker 32 (3%) was greater than
an inverse migration of Marker 32 to Panama (0.0%).
Host and symbiont genotype specificity
For a conservative approximation of specificity, only
Symbiodinium B1 samples with single alleles at all five
microsatellite loci were compared with host genotypes
that amplified at all 4 host-specific markers. For E. flexu-
osa, 1?5 clonal host genotypes were found (each com-
prising 2?6 individuals) within a given sampling
location (i.e. reef), and they rarely associated with the
same symbiont genotype (Table 6A). Conversely for
Symbiodinium B1, comparisons of symbiont clonal geno-
types to host genotypes mostly contained cases of the
same symbiont genotype being found in different host
genotypes (Table 6B). Tests for PI for E. flexuosa from
each sampling location revealed a low probability of
two individuals having the same genotype by chance
using this suite of markers (0.000?0.010). Fowey Rocks
contained the highest PI (0.010), and the expected num-
ber of individuals at that site to contain the same geno-
type is 0.432, or less than one-half of an individual. All
other sites contained lower PIs and therefore lower
expected numbers of individuals to contain the same
genotype by chance. Consequently, the identical geno-
types found at each location are not likely to be the
result of inadequate marker resolution, but more likely
represent clonal individuals.
Fisher?s exact tests for nonrandom associations
between host and symbiont genotypes revealed a
random association among three of the four sites exam-
ined along the FRT (Emerald Reef, P = 0.087; Key Largo
Patch, P = 0.178; Delta Shoal, P = 0.092). However,
Marker 32 demonstrated a significant nonrandom
Eunicea flexuosa Symbiodinium B1
M32-13 M32-23 M32-28 M32-31 M32-41 M32-13 M32-23 M32-28 M32-31 M32-41
Fig. 3 Assignment probabilities of Euni-
cea flexuosa (from Fig. 2A) and Symbiodi-
nium B1 (from Fig. 2C) of individuals
from Marker 32 (M32) in the lower Keys.
Both host and symbiont of five colonies
were assigned to population clusters
(yellow and red, respectively) that were
primarily assigned to samples from Pan-
ama. Sample identification numbers are
located beneath each individual.
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4421
pattern of host/symbiont genotypes (P < 0.000; i.e. it is
unlikely that the association of the five colonies from
Marker 32 to ?Panamanian? clusters for both host and
symbiont occurred by chance).
Discussion
Hardy?Weinberg and linkage equilibrium tests
It is not uncommon for both scleractinian corals and
gorgonian octocorals to display significant departures
from HWE, particularly heterozygote deficits, with
microsatellite markers (e.g. Gutierrez-Rodriguez &
Lasker 2004; Underwood et al. 2007; Baums et al. 2010;
Ledoux et al. 2010; Starger et al. 2010; Mokhtar-Jama??
et al. 2011). These deficiencies are generally attributed
to biological characteristics, such as overlapping genera-
tions and nonrandom mating due to inbreeding. Three
of the four microsatellite loci developed for E. flexuosa
were in HWE (Plfl19, Plfl199 and Plfl67) for most sam-
pling locations using exact and chi-square tests. The
deficiency found at some of the sites may be attributed
(A)
(B)
(C)
Emerald Reef
Reef
Dry Rocks
Largo Patch
Washerwoman Shoal
Shoal
Sambo
Point
Cay
Cana
Bank
Rocks
Fig. 4 The fraction of Eunicea flexuosa colonies with >0.90 assignment to either blue or yellow population clusters from the (A) Florida
reef tract, (B) three sampling locations within the Bocas del Toro Province, Panama, and (C) Punta Cana, Dominican Republic and
Saba Bank, mapped over their geographical location. A cross indicates individuals for that location with <0.90 assignment probabili-
ties (see Table S8, Supporting Information for data).
? 2013 John Wiley & Sons Ltd
4422 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
to the inability to capture the true allelic diversity of
the likely large FRT-wide population of E. flexuosa at a
single reef. Locus Plfl7 in E. flexuosa showed extreme
and consistent departures from HWE in all sampling
locations and displayed the highest difference between
observed/expected heterozygote frequencies compared
with the other loci. However, it did not significantly alter
the estimation of K populations using Bayesian analyses.
Although Plfl7 may not conform to HWE expectations
among the collection sites of this study, sampling from
other locations across the Caribbean/western Atlantic
may reveal this marker to be more useful.
For Symbiodinium B1, pairwise tests for linkage dis-
equilibrium based on sampling location revealed signifi-
cant levels of linkage disequilibrium within sites and
among locus pairs. High linkage disequilibrium is likely
a consequence of asexually reproducing organisms that
do not sexually recombine (Istock et al. 1992), and it is
Saba Bank
Punta Cana
N
Emerald Reef
Fowey Rocks
Long Reef
North Dry Rocks
Key Largo Patch
East Washerwoman Shoal
Delta Shoal
Western Sambo
Marker 32
500 km
50 km
5 km
(A)
(B)
(C)
Drago
Hospital Point
Crawl Cay
Fig. 5 The fraction of Symbiodinium B1 with >0.90 assignment to their respective population cluster with the ?singlehaps? and ?to-
talmatrix? data sets from (A) the Florida reef tract [?singlehaps? (right/bottom, ?totalmatrix? (left/top)], (B) three sampling locations
within the Bocas del Toro Province, Panama, and (C) Punta Cana, Dominican Republic and Saba Bank [?singlehaps? (bottom), ?totalm-
atrix? (top)], mapped over their geographical location. A cross indicates individuals for that location with <0.90 assignment probabili-
ties (see Table S7, Supporting Information for data).
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4423
T
ab
le
5
E
un
ic
ea
fle
xu
os
a
im
m
ig
ra
ti
on
m
at
ri
x.
P
er
ce
nt
ag
e
of
im
m
ig
ra
n
ts
is
li
st
ed
fr
om
th
e
so
u
rc
e
lo
ca
ti
on
s
(?
Fr
om
?
co
lu
m
n
)
to
th
ei
r
en
d
lo
ca
ti
on
(?
In
to
?
ro
w
).
B
ol
d
/
bo
xe
d
va
lu
es
in
d
ia
go
n
al
re
p
re
se
n
t
n
on
m
ig
ra
n
ts
(s
el
f-
re
cr
u
it
m
en
t)
fo
r
ea
ch
lo
ca
ti
on
.
D
ar
k
sh
ad
ed
ce
ll
s
co
nt
ai
n
im
m
ig
ra
ti
on
ra
te
s
>
5%
.
A
m
on
g
th
e
bo
ld
/
bo
xe
d
an
d
d
ar
k
sh
ad
ed
ce
ll
s,
u
n
d
er
-
lin
ed
va
lu
es
co
nt
ai
n
co
n
fi
d
en
ce
in
te
rv
al
s
(C
I;
li
st
ed
in
p
ar
en
th
es
es
)
ou
ts
id
e
of
th
e
u
p
p
er
bo
u
nd
of
th
e
?n
u
ll?
C
I?
n
on
m
ig
ra
n
ts
(0
.6
75
?
0.
99
2)
,
an
d
m
ig
ra
n
ts
(0
?
0.
11
0)
.
T
h
os
e
n
ot
u
n
d
er
li
n
ed
co
nt
ai
n
ed
a
C
I
th
at
st
ra
d
d
le
d
th
e
u
p
p
er
?n
u
ll?
C
I
li
m
it
.
T
w
o
li
gh
t-
sh
ad
ed
ce
ll
s
sh
ow
th
e
bi
as
ed
d
ir
ec
ti
on
al
it
y
of
im
m
ig
ra
n
ts
to
P
an
am
a
fr
om
M
ar
ke
r
23
(a
n
d
n
ot
vi
ce
ve
rs
a)
,
al
th
ou
gh
th
ei
r
C
Is
in
d
ic
at
e
lo
w
re
li
ab
il
it
y.
In
to
 
Fr
om
 
E
m
er
al
d
 
R
ee
f
Fo
w
ey
 R
oc
ks
L
on
g 
R
ee
f
N
or
th
 D
ry
 
R
oc
ks
K
ey
 L
ar
go
 
P
at
ch
E
as
t 
W
as
he
rw
om
an
 
Sh
oa
l
D
el
ta
 S
ho
al
W
es
te
rn
 
Sa
m
bo
M
ar
ke
r 
32
P
an
am
a
Sa
ba
D
om
in
ic
an
 
R
ep
ub
li
c
0.
99
0.
21
0.
29
0.
06
0.
10
(0
.9
7-
0.
99
)
(0
.1
2-
0.
31
)
(0
.2
2-
0.
33
)
(0
.0
0-
0.
16
)
(0
.0
0-
0.
23
)
Fo
w
ey
 R
oc
ks
0.
00
0.
68
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
01
0.
00
0.
01
0.
01
L
on
g 
R
ee
f
0.
00
0.
00
0.
68
0.
00
0.
00
0.
00
0.
00
0.
00
0.
01
0.
00
0.
01
0.
01
0.
28
0.
28
0.
95
0.
08
0.
29
0.
28
0.
19
0.
15
0.
11
0.
13
(0
.2
2-
0.
32
)
(0
.2
2-
0.
32
)
(0
.8
4-
0.
99
)
(0
.0
0-
0.
17
)
(0
.2
2-
0.
32
)
(0
.2
0-
0.
32
)
(0
.0
9-
0.
27
)
(0
.0
9-
0.
21
)
(0
.0
2-
0.
24
)
(0
.0
2-
0.
27
)
K
ey
 L
ar
go
 
P
at
ch
0.
00
0.
00
0.
00
0.
00
0.
68
0.
00
0.
00
0.
00
0.
03
0.
01
0.
02
0.
01
E
as
t 
W
as
he
rw
om
an
 
Sh
oa
l
0.
00
0.
00
0.
00
0.
00
0.
00
0.
68
0.
00
0.
00
0.
01
0.
00
0.
01
0.
01
D
el
ta
 S
ho
al
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
67
0.
00
0.
01
0.
00
0.
01
0.
01
W
es
te
rn
 S
am
bo
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
68
0.
01
0.
00
0.
01
0.
01
0.
00
(0
.0
0-
0.
03
)
0.
03
(0
.0
0-
0.
10
)
Sa
ba
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
01
0.
00
0.
70
0.
01
D
om
in
ic
an
 
R
ep
ub
li
c
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
01
0.
00
0.
01
0.
69
U
p
p
er
 
K
ey
s
N
or
th
 D
ry
 
R
oc
ks
0.
00
0.
02
B
is
ca
yn
e
U
p
p
er
 K
ey
s
M
id
d
le
 K
ey
s
L
ow
er
 K
ey
s
B
is
ca
yn
e
E
m
er
al
d
 R
ee
f
0.
01
0.
01
0.
02
0.
00
0.
01
0.
01
0.
03
0.
01
M
id
d
le
 
K
ey
s
L
ow
er
 
K
ey
s
M
ar
ke
r 
32
0.
00
0.
00
0.
01
0.
01
P
an
am
a
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
00
0.
69
0.
00
0.
00
0.
80
0.
02
0.
01
? 2013 John Wiley & Sons Ltd
4424 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
Table 6 Comparison of clonal host genotypes (4 markers) to their symbiont genotypes (5 markers) (A), and replicate symbiont geno-
types to their host genotypes (B). Genotypes are arranged by location (i.e. reef)
(A)
Sampling location Sample ID
Host genotype
label
Symbiont genotype
label
% Host colonies with same genotype
that have the same Symbiodinium genotype
Emerald Reef (No duplicate host genotypes)
Fowey Rocks FR_07 O R 100% (n = 2)
Fowey Rocks FR_50 O R
Fowey Rocks FR_02 OO 178 33% (n = 6)
Fowey Rocks FR_22 OO 41
Fowey Rocks FR_40 OO 151
Fowey Rocks FR_30 OO 145
Fowey Rocks FR_12 OO F
Fowey Rocks FR_31 OO F
Fowey Rocks FR_03 P T 0% (n = 2)
Fowey Rocks FR_36 P F
Long Reef (No duplicate host genotypes)
North Dry Rocks NDR_06 GG M 0% (n = 3)
North Dry Rocks NDR_30 GG 50
North Dry Rocks NDR_34 GG 117
North Dry Rocks NDR_31 JJ 173 0% (n = 2)
North Dry Rocks NDR_38 JJ 172
North Dry Rocks NDR_15 OO 127 0% (n = 2)
North Dry Rocks NDR_40 OO 142
Key Largo Patch KLP_28 FFF D 100% (n = 2)
Key Largo Patch KLP_64 FFF D
East Washerwoman Shoal EWS_18 JJ 80 0% (n = 2)
East Washerwoman Shoal EWS_65 JJ 132
East Washerwoman Shoal EWS_31 LL 12 0% (n = 2)
East Washerwoman Shoal EWS_40 LL 8
East Washerwoman Shoal EWS_26 NN E 0% (n = 2)
East Washerwoman Shoal EWS_50 NN H
East Washerwoman Shoal EWS_01 O E 67% (n = 3)
East Washerwoman Shoal EWS_09 O B
East Washerwoman Shoal EWS_22 O B
East Washerwoman Shoal EWS_24 OO C 0% (n = 3)
East Washerwoman Shoal EWS_35 OO K
East Washerwoman Shoal EWS_63 OO 66
Delta Shoal DS_55 III 135 0% (n = 2)
Delta Shoal DS_65 III 105
Delta Shoal DS_59 JJJ 64 0% (n = 2)
Delta Shoal DS_66 JJJ 107
Western Sambo (No duplicate genotypes)
Marker 32 (No duplicate genotypes)
Panama PA_40 SSS 7 0% (n = 2)
Panama PA_57 SSS 4
Saba (No duplicate genotypes)
Dominican Republic DR_16 III 130 0% (n = 2)
Dominican Republic DR_20 III 100
Dominican Republic DR_12 UUU G 100% (n = 2)
Dominican Republic DR_21 UUU G
(B)
Sampling location Sample ID
Symbiont genotype
label
Host genotype
label
% Symbiont with the same genotype
found with the same host genotype
Emerald Reef (No duplicate symbiont genotypes)
Fowey Rocks FR_18 F 74 25% (n = 8)
Fowey Rocks FR_58 F 89
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4425
possible that the linkage disequilibrium observed here
is a result of the asexual mode of reproduction in
Symbiodinium when in symbiosis. However, when Sym-
biodinium samples were pooled according to assigned
population clusters as predicted by STRUCTURE, the
amount of significant linkage disequilibrium among
locus pairs dropped dramatically. This suggests that
haploid Symbiodinium B1 undergo considerable asexual
reproduction, but may also experience recombination
(LaJeunesse 2001; Santos et al. 2003b), possibly by
Table 6 Continued
(B)
Sampling location Sample ID
Symbiont genotype
label
Host genotype
label
% Symbiont with the same genotype
found with the same host genotype
Fowey Rocks FR_39 F 105
Fowey Rocks FR_49 F 257
Fowey Rocks FR_48 F J
Fowey Rocks FR_12 F OO
Fowey Rocks FR_31 F OO
Fowey Rocks FR_36 F P
Fowey Rocks FR_07 R O 100% (n = 2)
Fowey Rocks FR_50 R O
Fowey Rocks FR_03 T P 0% (n = 2)
Fowey Rocks FR_41 T QQ
Long Reef LR_50 I 94 0% (n = 2)
Long Reef LR_60 I 309
North Dry Rocks NDR_06 M GG 0% (n = 2)
North Dry Rocks NDR_08 M 55
Key Largo Patch KLP_28 D FFF 100% (n = 2)
Key Largo Patch KLP_64 D FFF
Key Largo Patch KLP_24 F 101 0% (n = 2)
Key Largo Patch KLP_35 F 174
Key Largo Patch KLP_23 P 314 0% (n = 2)
Key Largo Patch KLP_62 P 311
Key Largo Patch KLP_27 U 126 0% (n = 2)
Key Largo Patch KLP_47 U 125
East Washerwoman Shoal EWS_20 B 33 50% (n = 4)
East Washerwoman Shoal EWS_39 B 291
East Washerwoman Shoal EWS_09 B O
East Washerwoman Shoal EWS_22 B O
East Washerwoman Shoal EWS_01 E O 0% (n = 2)
East Washerwoman Shoal EWS_26 E NN
East Washerwoman Shoal EWS_05 H 217 0% (n = 2)
East Washerwoman Shoal EWS_50 H NN
East Washerwoman Shoal EWS_33 K 48 0% (n = 2)
East Washerwoman Shoal EWS_35 K OO
Delta Shoal DS_48 F 269 0% (n = 2)
Delta Shoal DS_74 F 22
Delta Shoal DS_36 J 315 0% (n = 2)
Delta Shoal DS_53 J 171
Western Sambo (No duplicate symbiont genotypes)
Marker 32 (No duplicate symbiont genotypes)
Panama PA_10 A 47 0% (n = 2)
Panama PA_11 A HH
Saba (No duplicate symbiont genotypes)
Dominican Republic DR_12 G UUU 100% (n = 2)
Dominican Republic DR_21 G UUU
Dominican Republic DR_08 O UU 0% (n = 2)
Dominican Republic DR_09 O 20
Individuals in bold highlight clones with matching host and symbiont genotypes. Genotype labels assigned by GENEALEX v.6.41 (Peakall
& Smouse 2006).
? 2013 John Wiley & Sons Ltd
4426 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
sexual reproduction in a free-living stage (Andras et al.
2011) within these population clusters.
Population structure and dispersal
All analyses indicated a single population of E. flexuosa
among nine sampling locations covering the north-
eastern and south-western boundaries of the FRT
(~225 km). Only the south-westernmost location,
Marker 32 (Key West), contained a few individuals with
assignment profiles to a second population using clus-
tering analysis. The population homogeneity observed
for E. flexuosa across the FRT is consistent with other
studies that examined the connectivity of scleractinian
corals (e.g. Baums et al. 2010; Hemond & Vollmer 2010)
and other invertebrates (e.g. Richards et al. 2007; but see
Debiasse et al. 2010). However, additional markers may
reveal further population subdivision among E. flexuosa
along the FRT not detected with this suite of markers.
Nonetheless, gene flow and therefore larval dispersal in
E. flexuosa in the FRT appears to be high and suggests a
strong potential for resilience in this species following
disturbance events.
Symbiodinium B1 on the FRT showed strong popula-
tion structure among sites and at individual collection
sites (i.e. reefs) over distances on the scale of metres.
Each site contained individuals assigned to 2?4 differ-
ent population clusters with high (>0.90) assignment
probabilities. This acute population structuring suggests
some level of dispersal among Symbiodinium B1 geno-
types to different reef locations of the FRT, but with
minimal sexual recombination. However, the large
number of samples with multiple symbiont genotypes
(47%), and the robust levels of admixture found among
many samples using Bayesian analyses, also suggests
some level of mixing among Symbiodinium B1, likely in
the form of mixed symbiont communities. As gorgo-
nian octocorals have not been recorded to routinely
shuffle their symbiont communities, especially in the
absence of any stress-induced bleaching (Kirk et al.
2005; Goulet 2006; Baker & Romanski 2007; Hannes
et al. 2009), it is unlikely that the observed genetic
pattern of Symbiodinium was caused by seasonal or syn-
chronous changes in symbiont communities during the
3.6 years over which sampling occurred. The opera-
tional timescales of this data set are likely much longer
(>10 years).
Unexpectedly, E. flexuosa from the FRT was not dif-
ferentiated from either Saba or the Dominican Republic
(distances > 1000 km) with Bayesian analyses. How-
ever, the Dominican Republic did show significant
differentiation from three sites of the FRT with FST anal-
yses. This mixed signal contrasts with the scleractinian
coral, Acropora cervicornis (also a broadcast spawner),
which was found to be genetically well connected along
the FRT, but which showed clear population subdivi-
sion at sites from St. Thomas and Honduras (Baums
et al. 2010), which are as distant from the FRT as Saba
and the Dominican Republic. The lack of a strong popu-
lation signal in E. flexuosa from these three regions
could be the result of insufficient resolution based on
the number of loci used for analyses (3 or 4 in this
study compared with 7 and 8 in Baums et al. 2010).
However, for broadcast spawning corals, it is possible
that gene flow may be maintained over long geographical
distances (Nunes et al. 2009). In contrast, Symbiodinium
B1 from the FRT were differentiated from those in Saba
and Dominican Republic using Bayesian analyses, with
the latter two sites sharing a similar, albeit admixed,
population signal. A single-population cluster at Saba
and the Dominican Republic, although only weakly
supported, is plausible because of their relatively close
proximity to one another.
Migration rate analyses in BAYESASS indicated a pri-
marily southerly dispersal of E. flexuosa along the FRT,
in contrast to the northerly flow followed by the Florida
Current (Fig. 1). Although tests indicated a well-mixed
system along the FRT (nonmigrant values of ~68%)
among most sampled sites, this suite of markers con-
tained sufficient resolution to assess migration from
two sites (Emerald Reef and North Dry Rocks), both of
which signified a predominately southerly flow of
migrants along the FRT. Conversely, none of the middle
keys (East Washerwoman Shoal and Delta Shoal) or
lower keys (Western Sambo and Marker 32) sites
showed migration rates >1% to any of the upper keys
or Biscayne Bay sites, although many immigrant values
were at the maximum allowed by BAYESASS (~30%),
suggesting a weak signal. Richards et al. (2007) and
Debiasse et al. (2010) found a similar southerly migra-
tion pattern among amphipods and a reef sponge,
respectively, along the FRT and attributed the pattern
to inshore counter currents running north to south,
west of the Florida Current (Lee & Williams 1999;
Yeung & Lee 2002). The E. flexuosa samples in this
study were collected primarily from inshore patch reefs,
which could be influenced by inshore counter currents.
Furthermore, smaller-scale (both spatial and temporal)
oceanographical features, such as mesoscale eddies,
may potentially affect larval dispersal across the FRT,
counter to generalized current patterns (D?Alessandro
et al. 2007; Parks et al. 2009). Migration rate tests also
suggest the immigration of E. flexuosa to sites outside
the FRT (Saba, Dominican Republic and Panama) from
Emerald Reef and North Dry Rocks, although the CIs
for these sites indicate a tenuous signal. The pattern of
gene flow (and low migration rate reliability) observed
among these sites is likely an artefact of inadequate
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4427
sampling. Additional sampling between these sites,
together with the addition of other markers, may reveal
different patterns of dispersal.
A second-population cluster of E. flexuosa was
assigned predominantly to individuals sampled from
three sites at Bocas del Toro, Panama, in addition to
five individuals from a site off Key West (Marker 32).
These individuals were found sympatrically (on the
same reef site) among other individuals that were
assigned to a separate population, sometimes separated
by distances of only a few metres. This extreme within-
reef population subdivision could be the result of fine-
scale niche partitioning and adaptation to microreef
environments (Finke & Snyder 2008), or cryptic specia-
tion (Bickford et al. 2007). Eunicea flexuosa is highly plas-
tic and can vary morphologically in different areas of a
reef (Kim et al. 2004; Prada et al. 2008). Whether the
genetic subdivision of E. flexuosa within these reefs
corresponds to morphological variation needs to be
examined further.
Migration rate tests indicated very weak directional-
ity of immigrants from Panama to the FRT (Marker 32,
3%), but not vice versa (Marker 32 to Panama, 0%).
Although tenuous, this result agrees with the major
oceanographical currents of the western Caribbean
(Fig. 1; Roberts 1997) and suggests that the individuals
from Marker 32 (Key West) that were assigned to the
second (yellow) population are immigrants from
Panama (excluding ghost populations not sampled).
However, the dispersal of individuals from this second
population, for example, from Marker 32 to more north-
ern regions of the FRT was not detected using assign-
ment tests. This suggests that successful recruitment
into the FRT from Panama may not only be limited, but
migrants from the second population are not dispersing
further into the FRT. Instead, these individuals may be
mating with local individuals, not of the same popula-
tion, as indicated by the admixture of the blue and yel-
low populations among some individuals of Marker 32.
Vectored dispersal of Symbiodinium B1 by E. flexuosa
larvae
The high level of population subdivision among Symbi-
odinium B1 along the FRT contrasts markedly with that
of its host, E. flexuosa, which formed a single-population
cluster. The discrepancy between E. flexuosa and its
symbiont populations is in agreement with the life his-
tory of E. flexuosa (i.e. as a broadcast spawner, its larvae
obtain symbionts from the external environment) and
suggests that adult colonies acquire (and maintain)
Symbiodinium B1 found in local environmental pools
after settlement. Analyses of E. flexuosa and Symbiodini-
um B1 indicate high dispersal of the host across the FRT
and Caribbean, but not in the symbiont. Therefore, if
symbiont acquisition occurs at, or close to, settlement,
E. flexuosa larvae that disperse long distances will likely
be exposed to genetically different symbionts compared
with those found in their natal symbiont habitat. How-
ever, as gorgonian octocoral larvae typically obtain their
symbionts quickly (during the first few days) and are
nonspecific during early symbiont uptake (Coffroth
et al. 2001), E. flexuosa larvae may initially incorporate
Symbiodinium B1 from their natal site and only later
acquire local symbionts after settlement. Another possi-
bility is that E. flexuosa can acquire natal symbionts,
transport them to new settlement sites and retain them
over time. For example, five samples from the south-
westernmost site in the lower keys (Marker 32) were
assigned to the red Symbiodinium B1 population cluster,
to which mostly individuals from Panama were
assigned (Fig. 3). Moreover, the hosts in which these
five Symbiodinium samples were found were assigned to
the yellow E. flexuosa population, which was also prin-
cipally assigned to individuals from Panama. This pat-
tern of association among host/symbiont genotypes in
Marker 32 was nonrandom (Fisher?s exact test) and is
not likely to have occurred by chance. Therefore, as the
direction of E. flexuosa immigrants is likely to move
from Panama to the lower keys (Marker 32), and not
vice versa, it is hypothesized that the larvae of these
five E. flexuosa colonies originated in Panama (disre-
garding unsampled locations) and acquired natal sym-
bionts there. They were then subsequently transported
to the lower keys (perhaps via intermediate populations
that were not sampled), where they eventually settled
and, in this case, retained their natal symbionts. Simi-
larly, one individual from Saba was also assigned to
both the yellow host and red symbiont ?Panamanian?
populations, unlike the other samples of that site, and
likely exhibited a similar pattern of symbiont acquisi-
tion, dispersal and retention.
As previously discussed, it does not appear that the
red ?Panamanian? symbiont genotypes are specific to
the yellow ?Panamanian? host genotypes. Excluding the
five individuals from Marker 32, the red ?Panamanian?
symbiont population cluster was also strongly assigned
to at least three other individuals (and weakly to
another four) in the FRT whose hosts were assigned to
the blue cluster (Delta Shoal, Key Largo Patch, and
Emerald Reef). This suggests that Symbiodinium B1 from
Panama, after arriving in the FRT with E. flexuosa lar-
vae, can then infect local E. flexuosa hosts in the FRT
and mix with other (local) Symbiodinium B1 genotypes
within these hosts. This may explain why the red ?Pana-
manian? symbiont genotype was strongly admixed with
local (FRT) symbiont genotypes in four of the FRT
hosts.
? 2013 John Wiley & Sons Ltd
4428 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
Symbiodinium from three sites from Bocas del Toro,
Panama, were principally assigned to a single (red)-
population cluster, with minimal admixture. In contrast
to most of the FRT, E. flexuosa within these sites were
assigned to separate population clusters (blue and
yellow). This suggests that two genetically distinct
groups of E. flexuosa occur sympatrically at these sites,
but are nonspecific to the Symbiodinium B1 genotypes
with which they associate. If E. flexuosa were strictly
specific to a particular Symbiodinium B1 genotype (San-
tos et al. 2004), then individuals from the two E. flexu-
osa populations would be expected to host different
populations of Symbiodinium B1. This was not the case
and is contrary to what has been suggested regarding
host/symbiont specificity in gorgonian octocorals to
date (Coffroth et al. 2001; Goulet & Coffroth 2003a,b;
Santos et al. 2004). In addition, Fisher?s exact tests com-
paring groups of the most genetically dissimilar
E. flexuosa genotypes along the FRT with the popula-
tion cluster of Symbiodinium B1 genotypes with which
they associate showed no significant nonrandom rela-
tionship. This suggests that E. flexuosa genotypes of the
FRT are not specific to a particular Symbiodinium B1
genotype, nor do they continually maintain them over
time.
A comparison of clonal host genotypes to their sym-
biont genotypes revealed very few instances of geno-
typical matches, and PI tests suggest that it is unlikely
the clonal genotypes found are a result of insufficient
marker resolution. This host/symbiont flexibility con-
trasts with other studies that have shown strict (100%)
specificity of clonemates of a particular host genotype
to a single symbiont genotype (Goulet & Coffroth
2003a,b). Conversely, Andras et al. (2013) did not find
duplicate genotypes of the gorgonian octocoral G. vent-
alina using 10 microsatellite loci. However, G. ventalina
is not known to regularly fragment asexually and is
phylogenetically unrelated to E. flexuosa at the family
level. The extent to which E. flexuosa routinely frag-
ments is unknown, although fragmentation among a
more closely related gorgonian octocoral, Plexaura
kuna, can be extensive (Lasker 1984). Replicate Symbi-
odinium B1 genotypes were not found to be restricted
to a particular E. flexuosa genotype. Instead, in most
cases, the same Symbiodinium B1 genotype was found
in different E. flexuosa genotypes. This suggests that
Symbiodinium B1 genotypes, reproducing asexually
within hosts, may not be restricted to that individual,
but may be transferred and shared among different
E. flexuosa individuals. Alternatively, many Symbiodini-
um B1 with the same genotype may each initially
associate with a different E. flexuosa genotype and
remain in those specific symbioses. Future work
using temporal data (e.g. Goulet & Coffroth 2003b)
may be necessary to accurately distinguish these two
possibilities.
It is likely that the relatively complex population
structure observed among Symbiodinium B1 in E. flexu-
osa is the result of a passive process of stochastic
dispersal and host acquisition, although a selection-
mediated process of local adaptation cannot be ruled
out. To understand these processes more clearly, a thor-
ough understanding of the relationship between the
genetic diversity of Symbiodinium B1 and ecological/
physiological adaptation is needed. Nevertheless,
environmental pools of Symbiodinium B1 appear to be
genetically diverse, and both host and symbiont likely
exhibit short- and long-range dispersal over ecological
time periods. This could, in turn, contribute to the
spread of locally viable genetic combinations of host
and symbiont in response to environmental change.
Acknowledgements
Special thanks to S. Gray, L. Krimsky, S. Manley, J. Sanchez and
P. Etnoyer for their assistance in the field and boat captaining. K.
Darois, K. Lynch and H. Tieslink contributed invaluable support
in the laboratory. D. Crawford and A. Wilson provided
indispensible advice and suggestions on early drafts of the
manuscript. Microsatellite enrichment was carried out in the
Pritzker Laboratory for Molecular Systematics and Evolution
operated with support from the Pritzker Foundation. This work
was partially funded by Knight and Rowlands Fellowships at
the University of Miami?s Rosenstiel School.
References
Andras JP, Kirk NL, Drew Harvell CD (2011) Range-wide pop-
ulation genetic structure of Symbiodinium associated with the
Caribbean Sea fan coral, Gorgonia ventalina. Molecular Ecology,
20, 2525?2542.
Andras JP, Rypien KL, Harvell CD (2013) Range-wide popula-
tion genetic structure of the Caribbean sea fan coral, Gorgonia
ventalina. Molecular Ecology, 22, 56?73.
Aurelle D, Ledoux JB, Rocher C, Borsa P, Chenuil A, Feral JP
(2011) Phylogeography of the red coral (Corallium rubrum):
inferences on the evolutionary history of a temperate gorgo-
nian. Genetica, 139, 855?869.
Baker AC (2003) Flexibility and specificity in coral-algal symbi-
osis: diversity, ecology and biogeography of Symbiodinium.
Annual Review of Ecology and Systematics, 34, 661?689.
Baker AC, Romanski AM (2007) Multiple symbiotic partner-
ships are common in scleractinian corals, but not in octoco-
rals: comment on Goulet (2006). Marine Ecology Progress
Series, 335, 237?242.
Baker AC, Rowan R, Knowlton N (1997) Symbiosis ecology of
two Caribbean acroporid corals. Proceedings of the 8th Interna-
tional Coral Reef Symposium, Panama, 2, 1295?1300.
Baums IB, Johnson ME, Devlin-Durante MK, Miller MW (2010)
Host population genetic structure and zooxanthellae diver-
sity of two reef-building coral species along the Florida Reef
Tract and wider Caribbean. Coral Reefs, 29, 835.
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4429
Bayer FM (1961) The Shallow-Water Octocorallia of the West
Indian Region: A Manual for Marine Biologists. Martinus Nijho,
The Hague.
Beiring EA, Lasker HR (2000) Egg production by colonies of a
gorgonian coral. Marine Ecology Progress Series, 196, 169?177.
Bellwood D, Hughes T, Folke C, Nystr?om M (2004) Confront-
ing the coral reef crisis. Nature, 429, 827?833.
Bickford D, Lohman DJ, Sodhi NS et al. (2007) Cryptic species
as a window on diversity and conservation. Trends in Ecology
& Evolution, 22, 148?155.
Bruno JF, Selig ER (2007) Regional decline of coral cover in the
Indo-Pacific: timing, extent, and subregional comparisons.
PLoS ONE, 2, e711.
Coffroth M, Santos S (2005) Genetic diversity of symbiotic
dinoflagellates in the genus Symbiodinium. Protist, 156, 19?34.
Coffroth MA, Santos SR, Goulet TL (2001) Early ontogenetic
expression of specificity in a cnidarian-algal symbiosis. Mar-
ine Ecology Progress Series, 222, 85?96.
Cowen RK, Lwiza KMM, Sponagule S, Paris CB, Olson DB
(2000) Connectivity of marine populations: open or closed?
Science, 287, 857?859.
Cowen RK, Paris CB, Srinivasan A (2006) Scaling of connectiv-
ity in marine populations. Science, 311, 522?527.
D?Alessandro E, Sponaugle S, Lee T (2007) Patterns and
processes of larval fish supply to the coral reefs of the
upper Florida Keys. Marine Ecology Progress Series, 331,
85?100.
De?ath G, Fabricius KE, Sweatman H, Puotinen M (2012) The
27?year decline of coral cover on the Great Barrier Reef and
its causes. Proceedings of the National Academy of Sciences, 109,
17995?17999.
Debiasse MB, Richards VP, Shivji MS (2010) Genetic assess-
ment of connectivity in the common reef sponge, Callyspon-
gia vaginalis (Demospongiae: Haplosclerida) reveals high
population structure along the Florida reef tract. Coral Reefs,
29, 47?55.
Drummond AJ, Ashton B, Cheung M et al. (2008) Geneious
v4.8.5. Available from http://www.geneious.com/
Earl DA, vonHoldt BM (2011) STRUCTURE HARVESTER: a
website and program for visualizing STRUCTURE output
and implementing the Evanno method. Conservation Genetics
Resources, 4, 359?361.
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software STRUCTURE: a
simulation study. Molecular Ecology, 14, 2611?2620.
Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new
series of programs to perform population genetics analyses
under Linux and Windows. Molecular Ecology Resources, 10,
564?567.
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecu-
lar variance inferred from metric distances among DNA
haplotypes, application to human mitochondrial DNA
restriction data. Genetics, 131, 479?491.
Falush D, Stephens M, Pritchard JK (2003) Inference of popula-
tion structure: extensions to linked loci and correlated allele
frequencies. Genetics, 164, 1567?1587.
Finke DL, Snyder WE (2008) Niche partitioning increases
resource exploitation by diverse communities. Science, 321,
1488?1490.
Fisher RA (1944) Statistical Methods for Research Workers. Oliver
and Boyd, Edinburgh, section 21.02.
Gardner TA, Cote IM, Gill JA, Grant A, Watkinson AR (2003)
Long-term region-wide declines in Caribbean corals. Science,
301, 958?960.
Glenn TC, Schable NA (2005) Isolating microsatellite DNA loci.
Methods in Enzymology, 395, 202?222.
Goldberg W (1973) Ecological aspects of salinity and tempera-
ture tolerances of some reef-dwelling gorgonians from Flor-
ida. Caribbean Journal of Science, 13, 465?488.
Goulet TL (2006) Most corals may not change their symbionts.
Marine Ecology Progress Series, 321, 1?7.
Goulet TL, Coffroth MA (2003a) Genetic composition of zoo-
xanthellae between and within colonies of the octocoral
Plexaura kuna, based on small subunit rDNA and multilocus
DNA fingerprinting. Marine Biology, 142, 233?239.
Goulet TL, Coffroth MA (2003b) Stability of an octocoral-algal
symbiosis over time and space. Marine Ecology Progress Series,
250, 117?124.
Goulet TL, Coffroth MA (2004) The genetic identity of dinofla-
gellate symbionts in Caribbean octocorals. Coral Reefs, 23,
465?472.
Goulet TL, Simmons C, Goulet D (2008) Worldwide biogeogra-
phy of Symbiodinium in tropical octocorals. Marine Ecology
Progress Series, 355, 45?58.
Guo SW, Thompson EA (1992) Performing the exact test of
Hardy?Weinberg proportion for multiple alleles. Biometrics,
48, 361?372.
Gutierrez-Rodriguez C, Lasker HR (2004) Microsatellite varia-
tion reveals high levels of genetic variability and popula-
tion structure in the gorgonian coral Pseudopterogorgia
elisabethae across the Bahamas. Molecular Ecology, 13, 2211?
2221.
Hannes AR, Barbeitos M, Coffroth MA (2009) Stability of sym-
biotic dinoflagellate type in the octocoral Briareum asbestinum.
Marine Ecology Progress Series, 391, 65?72.
Hedgecock D, Barber PH, Edmands S (2007) Genetic
approaches to measuring connectivity. Oceanography, 20,
70?79.
Hedrick PW (2000) Genetics of Populations, 2nd edn. Jones and
Bartlett, Boston, Massachusetts, USA.
Hellberg ME (2007) Footprints on water: the genetic wake of
dispersal among reefs. Coral Reefs, 26, 463?473.
Hemond EM, Vollmer SV (2010) Genetic diversity and connec-
tivity in the threatened staghorn coral (Acropora cervicornis)
in Florida. PLoS ONE, 5, e8652.
Hubisz M, Falush D, Stephens M, Pritchard J (2009) Inferring
weak population structure with the assistance of sample
group information. Molecular Ecology Resources, 9, 1322?
1332.
Hughes TP, Baird AH, Bellwood DR et al. (2003) Climate
change, human impacts, and the resilience of coral reefs. Sci-
ence, 301, 929?933.
Istock CA, Duncan KE, Ferguson N, Zhou X (1992) Sexuality in
a natural population of bacteria-Bacillus subtilis challenges
the clonal paradigm. Molecular Ecology, 1, 95?103.
Jaap WC (1984) The Ecology of the South Florida Coral Reefs: A
Community Profile. US Department of the Interior, Fish and
Wildlife Service, Minerals Management Service, Washington,
DC, 138 pp.
Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster match-
ing and permutation program for dealing with label switch-
? 2013 John Wiley & Sons Ltd
4430 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER
ing and multimodality in analysis of population structure.
Bioinformatics, 23, 1801.
Jones RJ (2004) Testing the ?photoinhibition? model of coral
bleaching using chemical inhibitors. Marine Ecology Progress
Series, 284, 133?145.
Jones GP, Srinivasan M, Almany GR (2007) Population connec-
tivity and conservation of marine biodiversity. Oceanography,
20, 100?111.
Jones GP, Almany GR, Russ GR et al. (2009) Larval retention
and connectivity among populations of corals and reef
fishes: history, advances and challenges. Coral Reefs, 28,
307?325.
Joosse SA (2011) Fisher?s exact test. Available from http://
in-silico.net/statistics/fisher_exact_test.
Kalinowski ST (2011) The computer program STRUCTURE
does not reliably identify the main genetic clusters with spe-
cies: simulations and implications for human populations
structure. Heredity, 106, 625?632.
Kim E, Lasker H, Coffroth M, Kim K (2004) Morphological and
genetic variation across reef habitats in a broadcast-spawn-
ing octocoral. Hydrobiologia, 530, 423?432.
Kirk NL, Ward JR, Coffroth MA (2005) Stable Symbiodinium
composition in the sea fan Gorgonia ventalina during temper-
ature and disease stress. Biological Bulletin, 209, 227?234.
Kirk NL, Andras JP, Harvell CD, Santos SR, Coffroth MA (2009)
Population structure of Symbiodinium sp. associated with the
common sea fan, Gorgonia ventalina, in the Florida Keys across
distance, depth, and time. Marine Biology, 156, 1609?1623.
LaJeunesse TC (2001) Investigating the biodiversity, ecology
and phylogeny of endosymbiotic dinoflagellates in the genus
Symbiodinium using the ITS region: In search of a ?species?
level marker. Journal of Phycology, 37, 866?880.
Lasker H (1984) Asexual reproduction, fragmentation, and
skeletal morphology of a plexaurid gorgonian. Marine Ecol-
ogy Progress Series, 19, 261?268.
Lasker H, Coffroth M (1983) Octocoral distributions at Carrie
Bow Cay, Belize. Marine Ecology Progress Series, 13, 21?28.
Ledoux J-B, Mokhtar-Jamai K, Roby C et al. (2010) Genetic sur-
vey of shallow populations of the Mediterranean red coral
[Corallium rubrum (Linnaeus, 1758)]: new insights into evolu-
tionary processes shaping nuclear diversity and implications
for conservation. Molecular Ecology, 19, 675?690.
Lee T, Williams E (1999) Mean distribution and seasonal vari-
ability of coastal currents and temperature in the Florida
Keys with implications for larval recruitment. Bulletin of
Marine Science, 64, 35?56.
Lewis CL, Coffroth MA (2004) The acquisition of exogenous
algal symbionts by an octocoral after bleaching. Science, 304,
1490?1492.
Mantel N (1967) The detection of disease clustering and a gen-
eralized regression approach. Cancer Research, 27, 209?220.
Mokhtar-Jama?? K, Pascual M, Ledoux J-B et al. (2011) From glo-
bal to local genetic structuring in the red gorgonian Paramu-
ricea clavata: the interplay between oceanographic conditions
and limited larval dispersal. Molecular Ecology, 20, 3291?3305.
Nunes F, Norris R, Knowlton N (2009) Implications of isolation
and low genetic diversity in peripheral populations of an
amphi-Atlantic coral. Molecular Ecology, 18, 4283?4297.
Palumbi SR (2003) Population genetics, demographic connec-
tivity, and the design of marine reserves. Ecological Applica-
tions, 13, S146?S158.
Pandolfi J, Jackson J, Baron N et al. (2005) Are US coral reefs
on the slippery slope to slime? Science, 307, 1725?1726.
Parks AB, Shay LK, Johns WE, Martinez-Pedraja J, Gurgel K-W
(2009) HF radar observations of small-scale surface current
variability in the Straits of Florida. Journal of Geophysical
Research Oceans, 114, C08002.
Peakall R, Smouse P (2006) GENALEX 6: genetic analysis in
Excel. Population genetic software for teaching and research.
Molecular Ecology Notes, 6, 288?295.
Pochon X, Gates RD (2010) A new Symbiodinium clade (Dino-
phyceae) from soritid foraminifera in Hawai?i. Molecular Phy-
logenetics and Evolution, 56, 492?497.
Prada C, Schizas N, Yoshioka P (2008) Phenotypic plasticity or
speciation? A case from a clonal marine organism. BMC
Evolutionary Biology, 8, 47.
Pritchard JK (2010) Documentation for structure software: Ver-
sion 2.3.
Pritchard JK, Stephens M, Donnelly P (2000) Inference of popu-
lation structure using multilocus genotype data. Genetics,
155, 945?959.
Richards VP, Thomas JD, Stanhope MJ, Shivji MS (2007)
Genetic connectivity in the Florida reef system: comparative
phylogeography of commensal invertebrates with contrasting
reproductive strategies. Molecular Ecology, 16, 139?157.
Roberts CM (1997) Connectivity and management of Caribbean
coral reefs. Science, 278, 1454.
Rosenberg NA (2004) DISTRUCT: a program for the graphical
display of population structure. Molecular Ecology Notes, 4,
137?138.
Rousset F (2008) genepop?007: a complete re-implementation of
the genepop software for Windows and Linux. Molecular
Ecology Resources, 8, 103?106.
Rowan R, Powers DA (1991) A molecular genetic classification
of zooxanthellae and the evolution of animal-algal symbiosis.
Science, 251, 1348?1351.
Rozen S, Skaletsky H (2000) Primer3 on the WWW for general
users and for biologist programmers. Methods in Molecular
Biology, 132, 365?386.
Sanchez J, Diaz J, Zea S (1997) Gorgonian communities in two
contrasting environments on oceanic atolls of the southwest-
ern Caribbean. Bulletin of Marine Science, 61, 453?465.
Santos SR, Gutierrez-Rodr?guez C, Coffroth MA (2003a)
Phylogenetic identification of symbiotic dinoflagellates via
lengthheteroplasmy in domain V of chloroplast large subunit
(cp23S)-ribosomal DNA sequences. Marine Biotechnology, 5,
130?140.
Santos S, Gutierrez-Rodriguez C, Lasker H, Coffroth M (2003b)
Symbiodinium sp. associations in the gorgonian Pseudoptero-
gorgia elisabethae in the Bahamas: high levels of genetic vari-
ability and population structure in symbiotic dinoflagellates.
Marine Biology, 143, 111?120.
Santos S, Shearer T, Hannes A, Coffroth M (2004) Fine-scale
diversity and specificity in the most prevalent lineage of
symbiotic dinoflagellates (Symbiodinium, Dinophyceae) of the
Caribbean. Molecular Ecology, 13, 459?469.
Seutin G, White BN, Boag PT (1991) Preservation of avian
blood and tissue samples for DNA analyses. Canadian Journal
of Zoology, 69, 82?90.
Shearer TL, Gutierrez-Rodriguez C, Coffroth MA (2005) Gener-
ating molecular markers from zooxanthellate cnidarians.
Coral Reefs, 24, 57?66.
? 2013 John Wiley & Sons Ltd
POPULATION GENETICS OF A REEF CORAL SYMBIOSIS 4431
Slatkin M (1995) A measure of population subdivision based
on microsatellite allele frequencies. Genetics, 139, 457?462.
Spalding MD, Ravilious C, Green EP (2001) World Atlas of Coral
Reefs. University of California Press, Berkeley, Los Angeles,
London.
Starger CJ, Barber PH, Ambariyanto Baker AC (2010) The
recovery of coral genetic diversity in the Sunda Strait
following the 1883 eruption of Krakatau. Coral Reefs, 29,
547?565.
Underwood JN, Smith LD, Van Oppen MJH, Gilmour JP (2007)
Multiple scales of genetic connectivity in a brooding coral on
isolated reefs following catastrophic bleaching. Molecular
Ecology, 16, 771?784.
Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P
(2004) MICRO-CHECKER: software for identifying and cor-
recting genotyping errors in microsatellite data. Molecular
Ecology Notes, 4, 535?538.
Van Oppen MJH, Gates RD (2006) Conservation genetics and
the resilience of reef-building corals. Molecular Ecology, 15,
3863?3883.
Van Oppen MJH, Mieog JC, Sanchez CA, Fabricius KE (2005)
Diversity of algal endosymbionts (zooxanthellae) in octoco-
rals: the roles of geography and host relationships. Molecular
Ecology, 14, 2403?2417.
Vollmer SV, Palumbi SR (2007) Restricted gene flow in the Carib-
bean staghorn coral Acropora cervicomis: implications for the
recovery of endangered reefs. Journal of Heredity, 98, 40?50.
Wilkinson C, Souter D (2008) Status of Caribbean Coral Reefs
After Bleaching and Hurricanes in 2005. Global Coral Reef
Monitoring Network, and Reef and Rainforest Research Cen-
tre, Townsville, 152 p.
Wilson G, Rannala B (2003a) Bayesian inference of recent
migration rates using multilocus genotypes. Genetics, 163,
1177?1191.
Wilson GA, Rannala B (2003b) Documentation for BayesAss 1.3.
Yeung C, Lee T (2002) Larval transport and retention of the
spiny lobster, Panulirus argus, in the coastal zone of the Flor-
ida Keys, USA. Fisheries Oceanography, 11, 286?309.
H.H.W. is interested in the population genetics, evolu-
tion and systematics of reef corals, particularly octocor-
als. This research was part of his Ph.D. dissertation in
A.C.B.?s lab at the University of Miami?s Rosenstiel
School. A.C.B.?s research focuses on the biology, ecol-
ogy and conservation of coral reefs, with a focus on the
impacts of climate change on these ecosystems. K.A.F is
manager of the Pritzker Laboratory for Molecular Sys-
tematics and Evolution at The Field Museum, Chicago.
His research focuses on the mating systems and popula-
tion biology of sharks using microsatellites. K.A.F.
helped with the development of the microsatellite
markers used in this study and contributed to writing
the manuscript.
Data accessibility
Microsatellite genotype data sets for Eunicea flexuosa
?eflex? and Symbiodinium B1 ?singlehaps? and ?totalma-
trix?: DRYAD, doi: 10.5061/dryad.82ms3.
Supporting information
Additional supporting information may be found in the online
version of this article.
Fig. S1 Estimation of K populations for the ?singlehaps? (A)
and ?totalmatrix? (B) datasets.
Table S1 Exact tests (Eunicea flexuosa) for Hardy?Weinberg
equilibrium for each locus and sampling location with FIS esti-
mates using Weir & Cockerham (1984) [W&C] and Robertson
& Hill (1984) [R&H] methods, observed and expected heterozy-
gosity, P-value, and significance.
Table S2 Chi-square tests (Eunicea flexuosa) for Hardy?Wein-
berg equilibrium (H1 = significant deviations from HWE fre-
quences) for each locus and sampling location, degrees of
freedom, chi-square value, P-value, and significance.
Table S3 Tests for linkage disequilibrium (Eunicea flexuosa) for
each primer pair in each sampling location, standard error,
P-value, and significance.
Table S4 Tests for linkage disequilibrium for each Symbiodini-
um primer pair in each sampling location, P-value, and signifi-
cance for the ?singlehaps? dataset.
Table S5 Linkage disequilibrium tests for individuals grouped
by within-site population clusters as determined by STRUCTURE
for each primer pair, P-value, and significance for the ?single-
haps? dataset.
Table S6 Pairwise FST values among all sampled locations
(reefs) for Eunicea flexuosa (?eflex? dataset without marker Plfl7
and duplicates) and Symbiodinium B1 (?singlehaps? and ?totalm-
atrix? datasets).
Table S7 Numbers of host individuals with Symbiodinium geno-
types with >90% assignment to a particular population includ-
ing the total number of individuals collected at each sampling
location, total number of individuals with >90% assignment per
location, the percent of individuals with >90% assignment for
each location, and individuals with >90% assignment and their
predicted population.
Table S8 Numbers of Eunicea flexuosa colonies with >90%
assignment to a particular population including the total num-
ber of individuals collected at each sampling location, the total
number of individuals with >90% assignment per location, the
percent of individuals with >90% assignment for each location,
and individuals with >90% assignment and their predicted pop-
ulation (data set excludes marker Plfl7).
? 2013 John Wiley & Sons Ltd
4432 H. H. WIRSHING, K. A. FELDHEIM and A. C. BAKER