Microbial Ecology

https://doi.org/10.1007/s00248-023-02299-3

RESEARCH

Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont

Assemblages of Corals and Hydrocorals During El Niño 2015/2016
Amana Guedes Garrido1,2,3 · Laís Feitosa Machado4 · Cristiano Macedo Pereira3,5 · Douglas Pinto Abrantes5 ·
Emiliano Nicolas Calderon3,6 · Carla Zilberberg3,6

Received: 26 April 2023 / Accepted: 30 August 2023

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023

Abstract
Reef corals have been threatened by climate change, with more frequent and intense bleaching events leading to extensive
coral mortality and loss of coral cover worldwide. In the face of this, the corals’ photosymbiont assemblages have received
special attention as a key to better understand the bleaching process and its recovery. To assess the effects of thermal anoma-
lies, the coral Mussismilia harttii and the hydrocoral Millepora alcicornis were monitored through the El Niño 2015/2016 at
a Southwestern Atlantic (SWA) coral reef. A severe bleaching event (57% of colonies bleached) was documented, triggered
by a < 3 °C-week heatwave, but no mortality was detected. The hydrocoral was more susceptible than the scleractinian,
displaying bleaching symptoms earlier and experiencing a longer and more intense bleaching event. The composition of
photosymbionts in the M. alcicornis population was affected only at the rare biosphere level (< 5% relative abundance), with
the emergence of new symbionts after bleaching. Conversely, a temporary dysbiosis was observed in the M. harttii population,
with one of the dominant symbiodiniaceans decreasing in relative abundance at the peak of the bleaching, which negatively
affected the total β-diversity. After colonies’ complete recovery, symbiodiniaceans’ dominances returned to normal levels
in both hosts. These results highlight critical differences in how the two coral species cope with bleaching and contribute to
the understanding of the role of photosymbionts throughout the bleaching-recovery process.

Keywords Zooxanthellae · Symbiodiniaceae · Coral bleaching · South Atlantic reefs · Thermal anomalies · Global climate
change

Introduction maximum monthly average temperature can lead to coral

bleaching [2]⁠, a phenomenon in which the symbiosis between
In the past decades, reef corals have been constantly threat- reef corals and dinoflagellates from the family Symbiodiniaceae
ened by climate change with more frequent and intense thermal is broken [3]⁠. Corals can deal with short and isolated bleaching
anomalies [1]⁠. Thermal stresses of just 1 °C above a region’s

1
Programa de Pós‑Graduação em Biodiversidade e Biologia
Evolutiva, Instituto de Biologia, Universidade Federal
* Amana Guedes Garrido do Rio de Janeiro (UFRJ), Rio de Janeiro, RJ, Brazil
[email protected] 2
Centro de Biologia Marinha, Universidade de São Paulo,
Laís Feitosa Machado São Sebastião, SP, Brazil
[email protected] 3
Instituto Coral Vivo, Santa Cruz Cabrália, BA, Brazil
Cristiano Macedo Pereira 4
Universidade Federal do Vale do São Francisco (UNIVASF),
[email protected]
Senhor do Bonfim, BA, Brazil
Douglas Pinto Abrantes 5
Programa de Pós‑Graduação em Zoologia, Museu
[email protected]
Nacional, Universidade Federal do Rio de Janeiro (UFRJ),
Emiliano Nicolas Calderon Rio de Janeiro, RJ, Brazil
[email protected] 6
Instituto de Biodiversidade e Sustentabilidade – NUPEM,
Carla Zilberberg Universidade Federal do Rio de Janeiro (UFRJ), Macaé, RJ,
[email protected] Brazil

13
Vol.:(0123456789)
 A. G. Garrido et al.

occurrences, but long and intense events have led to extensive depth range of coral species, the dominance of more resilient
coral mortality and loss of coral cover worldwide [1, 2, 4]⁠. massive corals, and the generalism of symbiosis patterns,
Global climate change impacts (e.g., increasing sea sur- including the dominance of generalist host species and sym-
face temperature) have been intensified by pulse heat stress biont lineages [21]⁠. In addition, many studies have revealed
events, such as El Niño, the warm phase of the El-Niño a high diversity of symbiodiniaceans associated with SWA
Southern Oscillation (ENSO) [1]⁠. In particular, the ENSO corals, following Caribbean patterns [22–25]⁠. However, the
2015/2016 summed with local thermal anomalies triggered dynamics of these diverse photosymbiont assemblages during
the third global-scale coral bleaching event between 2015 bleaching events are still poorly understood.
and 2017 [1, 5]⁠. In the face of this alarming scenario, the To assess the effect of thermal anomalies predicted by
diversity and dynamics of corals’ photosymbiotic assem- the NOAA Coral Reef Watch Program (CRW), during the
blages have received special attention as a key to better ENSO 2015/2016, over the composition of symbiodiniacean
understand the process of coral bleaching and its recovery assemblages, we monitored for 6 months the bleaching sta-
[6, 7]. Consequently, knowledge of symbiodiniaceans’ diver- tus and the profile of Symbiodiniaceae ITS2 rDNA phylo-
sity has increased drastically over the past few years, with types, assessed through next-generation sequencing, in the
new Symbiodiniaceae genera now being recognized [8]⁠ , populations of the hydrocoral Millepora alcicornis and the
with each of them encompassing a high diversity of physio- scleractinian coral Mussismilia harttii at three reef sites in a
logically distinguished lineages or species [7, 9]⁠. Since some model coral reef at the SWA.
corals can harbor multiple genera of Symbiodiniaceae, with
distinct levels of thermotolerance [10]⁠, and this assemblage
could be highly dynamic over time and space [11, 12]⁠, these Methods
photosymbionts can display a relevant role in the holobiont
thermal adaptation [7]⁠. Study Area
According to the Adaptive Bleaching Hypothesis [13,
14]⁠, symbiosis breakdown could allow the coral’s tissue to The study was carried out at Recife de Fora Marine Pro-
be colonized by a different set of symbiodiniaceans [15]⁠. tected Area (no-take zone), Porto Seguro, Brazil (Fig. 1, 16
During bleaching recovery, symbiont assemblage could 23.931′S, 038 59.193′W), a reef of 17.5 k­ m2. North of the
re-establish itself in a different composition better adapted Abrolhos Reef Complex, the South Atlantic center for coral
to the new thermal condition [15]⁠. The reorganization of diversity, Recife de Fora is a biogenic reef that harbors almost
the photosymbiont assemblage altering dominant lineages, all the coral species occurring in this province [16]⁠ and is a
could be a mechanism by which corals might adapt to cli- shallow (< 15 m depth; [26]) and easily accessible reef, 3.5
mate change [9]⁠. km offshore the Porto Seguro Bay. Because of these charac-
Brazilian coral reefs are the only reefs in the South Atlan- teristics, Recife de Fora has been used as a model coral reef
tic Ocean whose generally turbid waters harbor a coral fauna for ecological studies in Brazilian reef systems [26–30]⁠.
with a high degree of endemism [16]⁠, including the four
massive reef-building scleractinian species of the genus Monitoring Bleaching and Symbiotic Assemblages
Mussismilia Ortmann, 1890, as well as amphi-Atlantic spe-
cies, such as the branched hydrocoral Millepora alcicornis Three sites, Mourão (16 25′02.6″ S, 038 58′55.8″ W),
Linnaeus, 1758. Those species are of special interest due Funil (16 24′39.0″ S, 038 59′06.1″ W), and Taquaruçu (16
to their ecological relevance as reef-builders and three- 24′25.3″ S, 038 58′41.0″ W) were monitored over 6 months
dimensional complexity promoters in Southwestern Atlantic from December ­15th, 2015 to June ­15th, 2016 by biweekly
(SWA) reefs and their conservation status. This is because expeditions (totaling 12 sampling days). Mourão is a region
the scleractinian coral Mussismilia harttii (Verrill, 1868) is located west of the reef plateau, characterized by an exten-
an endangered species facing losses of coral cover and popu- sive coral cover found at depths ranging from 2 to 6 m,
lation decline in some coastal reefs [17]⁠, while the common where corals of the genera Mussismilia and Millepora are
M. alcicornis is the only branching species in the region and common. The Funil region, situated south of the reef pla-
the main generator of structural complexity, but has been teau, consists of multiple reef patches with depths and fauna
very susceptible to coral bleaching [18]⁠. similar to Mourão’s. Taquaruçu, on the other hand, is a tidal
Bleaching has been recorded for SWA reefs since the pool located on the eastern side of the reef. It features a
1990s, and such events are becoming more frequent and wide- sandy bottom and is characterized by the shallowest depths,
spread, being strongly correlated with El Niño years [17–20]⁠. ranging from 2 to 4 m. In this area, coral species such as
Nevertheless, SWA reefs have been highlighted as less suscep- Mussismilia harttii and Millepora alcicornis can be easily
tible to mortality from bleaching than the Caribbean and Indo- observed along the pool’s walls. The sites were monitored
Pacific reefs due to characteristics, such as water turbidity, the for bleaching through 20 × 1 m transects placed randomly at

13
Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont Assemblages of Corals and…

Fig. 1  The three monitored sites at Recife de Fora, Porto Seguro, Bahia, Brazil: Taquaruçu (T), Mourão (M), and Funil (F). The dashed square
indicates Recife de Fora Marine Protected Area

approximately 2 m depth during the low tide at each site. All (CRW; https://​coral​reefw​atch.​noaa.​gov/​produ​ct/​5km/​metho​
M. harttii and M. alcicornis coral colonies inside the tran- dology.p​ hp#d​ hw) as the accumulation of temperature anom-
sects were recorded and classified according to their color alies exceeding the MMM over the last 84 days (12 weeks)
following the Coral Health Chart color scale (https://​coral​ and converted to degree heating weeks (DHW) by divid-
watch.​org/; [31]⁠). A fragment of ~ 0.5 ­cm2 was sampled ing it by 7 days [32, 33]. We, therefore, used the long-term
from four randomly selected individuals of each species per regional MMM of 26.87 °C for our analyses.
site with a hammer and chisel. Tissue samples were stored
in liquid nitrogen and afterward transferred to CHAOS solu- Sample DNA Extraction, Processing, and Sequencing
tion (4 M guanidine thiocyanate, 0.1% N-lauroylsarcosine
sodium, 10 mM Tris pH 8.0, 0.1 M 2-mercaptoethanol) and Total DNA was extracted by phenol:chloroform:isoamyl
maintained at room temperature until analyzed. alcohol protocol [34]. Extracted samples were purified with
ProNex Size-Selective Purification System (Promega®) for
Thermal Stress Assessment DNA fragments of 600 bp, according to the manufacturer’s
instructions. PCR amplified the region of the internal tran-
Water temperatures at each monitored site were measured scribed spacer 2 (ITS2) rDNA with the specific primers
by in situ temperature data loggers (HOBO Pendant model SYM_VAR_5.8S2 (5′ GAAT ​ TGC​ AGA​ ACT​ CCG​ TGA​ ACC
UA-002-64, onset®), continuously recording every 15 min 3′) and SYM_VAR_REV (5′ CGGG ​ TTC​ WCT​ TGT​ YTG​ AC​
from December ­20th, 2015, to June ­20th, 2016. The long- TTC​ATG​C 3′; [35]⁠) linked to Illumina adapters (forward:
term maximum of the monthly mean sea surface temperature 5′ TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA
(MMM) from NOAA’s 5 km virtual station, Recife de Fora, CAG 3′; reverse: 5′ GTC TCG TGG GCT CGG AGA TGT
Brazil, was assessed and combined with in situ temperature GTA TAA GAG ACA G 3′; following the 16S Metagen-
data to calculate the cumulative heat stress as the sum of the omic Sequencing Library Preparation Illumina protocol).
positive anomalies ≥ 1 °C of the daily mean temperature The target amplicon was approximately 400 bp. The PCR
exceeding the MMM. Degree heating days (DHD) were cal- reaction contained 1 μl of DNA (~ 5 ng/μl), 0.2 μM of each
culated according to the NOAA Coral Reef Watch Program primer, 0.025 U/μl of Pfu DNA polymerase (Promega®),

13
 A. G. Garrido et al.

1X Pfu reaction buffer with ­MgSO4, 200 μM dNTP mix of similarity with the reference sequence were accepted to
(Promega®) and molecular grade water up to a total reaction avoid incorrect identifications. When a contig was classified
volume of 30 μl. PCR cycles were as follows: 98 °C for 2 in more than one OTU with the same percentage of simi-
min, 35 cycles of 98 °C for 30 s, 56 °C for 30 s, and 72 °C larity, all corresponding OTUs were attributed to the con-
for 60 s, and a final extension at 72 °C for 5 min. tig (e.g., C1/C3), and when similarity was lower than 98%,
Triplicate PCRs for each sample were pooled, and PCR a prefix TL- (from “too-low similarity”) was added to the
clean-up was performed with the ProNex System to select OTU (e.g., TL-B23/B42). When similarity was lower than
350 bp amplicons. Illumina indexing primers (Nextera® XT 95% with any reference sequence, the contig was named
Index Kit—96 indexes, 384 samples) were added to 5 μl of “unclassified.” Differences in library sizes across samples
mixed purified PCR product, and a new PCR of 50 μl was were adjusted by rarefaction using the smallest sample (1024
run to incorporate unique barcodes for each sample. This reads) as the threshold. The rarefied samples were used to
PCR reaction contained 5μl of cleaned PCR product, 5 μl compare Symbiodiniaceae assemblage composition between
of Illumina Indexed Primer 1 (i5), 5 μl of Illumina Indexed monitored sites and host species during the collection period.
Primer 2 (i7), 1 μl of HotStarTaq (Qiagen®), 10 μl of 1X
HotStarTaq Buffer with dNTP, and 24 μl PCR grade water Statistical Analyses
for a total reaction volume of 50 μl. PCR cycles were as fol-
lows: 95 °C for 3 min, 8 cycles of 94 °C for 30 s, 54 °C for All analyses were performed in the R statistical environment
30 s, and 72 °C for 30 s, and lastly, 72 °C for 5 min. PCR (version 1.4.1717; [36]), and the significance level adopted
clean-up was completed one last time with the ProNex Sys- was 95% (ɑ = 0.05). Permutational multivariate analysis
tem, and the resulting products were quantified by fluorom- of variance (PERMANOVA) with the function adonis from
etry with Qubit dsDNA HS Assay Kit by Qubit Fluoromet- “vegan” compared the Bray-Curtis Index dissimilarity of
ric Quantification (ThermoFisher®). Indexed and purified relative abundance of Symbiodiniaceae assemblage com-
samples were pooled in equal molar amounts (4 nM). The position of each host species population for the effect of
two obtained amplicon libraries, with 96 samples each, were sites and sampling period. For significant results, pairwise
sequenced on the Illumina MiSeq platform using a MiSeq PERMANOVAs were performed with the function pairwise
Reagent Kit v2 (500-cycles) paired-end run with 25% PhiX adonis from “vegan,” both with 999 permutations. Indica-
at the Integrated Functional Genomics Unit at Instituto de tor species analysis (ISA) was performed with the func-
Biodiversidade e Sustentabilidade - NUPEM/UFRJ, Brazil. tion multipatt from the package “indicspecies” to identify
symbiodiniacean lineages found more often in one group
Bioinformatic Data Processing than another, considering the groups indicated by PER-
MANOVA. For β-diversity metrics, differences in symbio-
Illumina’s real-time analysis was run during sequencing diniaceans assemblages from each host species were com-
using the default settings to remove clusters with the least pared through the relative abundance of Symbiodiniaceae
reliable data. Demultiplexed fastq files were generated with ITS2 phylotypes and visualized in a non-metric multidimen-
Illumina’s BaseSpaceFS (version 1.5.964), and reads were sional scaling (nMDS) plot based on Bray-Curtis Index dis-
processed in FastQC online platform. Samples with less similarity with the function metaMDS from “vegan.”
than 200 reads (N = 4) were removed from the dataset (N
= 182 samples remained, including at least three random
colonies of each species per location and collection time; Results
Table S1). All remaining sample reads were processed with
the Geneious software, where data were cleaned of prim- All three reef sites at Recife de Fora were hit by the marine
ers, adapters, and indexes with the BBDuk plugin and were heatwave and suffered bleaching during the summer of
filtered by quality (> 30) and length (> 200 bp). Chimeric 2015/2016 (as seen by [30]⁠ ; Fig. 2; Fig. S1). Although
reads were removed using UCHIME v4.2.40, and the remain- bleaching was severe, no mortality was recorded for either
ing reads were clustered into contigs with 98% similarity Mussismilia harttii or Millepora alcicornis. The sea surface
using de novo assembler for each sample separately. Contigs temperature (SST) at two of the monitored sites exceeded
encompassing more than 10 reads per sample were classi- the long-term maximum of the monthly mean SST (MMM)
fied into operational taxonomic units (OTUs) according to in late December and early February and stayed above the
a Symbiodiniaceae ITS2 rDNA reference database contain- MMM for 30 and 33 consecutive days at Funil and Mourão,
ing sequences downloaded from the GenBank (NCBI) and respectively, between March and April (Fig. 2a). Although
SymPortal (symportal.org). To remove non-symbiodini- SST at Taquaruçu never rose over the regional MMM, this
aceans’ sequences, a minimum sequence cover of 75% was site recorded the most intense bleaching status for both mon-
applied to classify the contigs, and classifications with > 98% itored species (100% of M. alcicornis, N = 4–9 between

13
Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont Assemblages of Corals and…

Fig. 2  a Daily mean water temperature (°C; solid lines) and accumu- ature. The dark continuous line is the bleaching threshold (long-term
lated heat stress (degree heating weeks, DHW; dashed lines) during maximum of the monthly mean sea surface temperature (MMM) + 1
the 6-month monitoring period (December 2015 to June 2016) meas- °C). b Relative occurrence of Millepora alcicornis and Mussismilia
ured in three sites at Recife de Fora Marine Protected Area (Porto harttii colonies in three sites according to color index based on Coral
Seguro, Bahia, Brazil) and from NOAA’s 5 km virtual station, Recife Health Chart along the monitoring period. Sample size indicated for
de Fora. Shadows indicate the daily maximum and minimum temper- each bar

March and May, and 66.67% of M. harttii colonies, N = 12 colonies in early June (N = 64) and to zero in final June (N
in final-April; Fig. 2b). = 52; Fig. 2b). We have considered “bleached” all colonies
Although the degree heating weeks (DHW) did not reach with a color index of 1 or 2 (e.g., Fig. S1), since it was a
the bleaching threshold (4 °C-weeks DHW, according to two-point reduction on the most common color index of the
Coral Reef Watch Program - CRW), bleaching occurrences coral population (4 for both species; [31]).
were recorded between March and May at all monitored After filters and data cleaning, 709,341 sequences were
sites, with a peak in final April to May, with 57–55% of the obtained from the two sequencings, library size varied between
colonies bleached (N = 63 and 54, for final April and May, 1394 and 13,939 sequences per sample (Table S1). Symbiod-
respectively). Millepora alcicornis was more susceptible to iniaceae ITS2 rDNA phylotype profiles were distinct between
bleaching, displaying bleaching symptoms about 1 month host species (Fig. 3) and responded differently to environmen-
before M. harttii. Additionally, M. alcicornis experienced a tal changes. Millepora alcicornis was dominated by Breviolum
longer and more intense event, for 3 months (March–May), phylotype B23, and no change in its dominance was observed
with 75% of colonies bleached at the peak of the event (N throughout the bleaching event (Fig. 3). Conversely, M. hart-
= 28 colonies in final-April), whereas M. harttii colonies tii was dominated by Cladocopium phylotype C3 at Funil and
experienced bleaching for 2 months (April–May) with 47% Mourão, while at Taquaruçu colonies were co-dominated by
of them bleached (N = 32 colonies in final-April). In June, the same C3 and Symbiodinium linucheae (ITS2 phylotype
recovery was observed with a drop to 0.08% of bleached A4) (Trench & Thinh) LaJeunesse 2018 (Fig. 3). However, in

13
 A. G. Garrido et al.

Fig. 3  Relative abundance of Symbiodiniaceae ITS2 rDNA phylo- 6-month monitoring period (December 2015 to June 2016) at Recife
types representatives of complete and rare biospheres (relative abun- de Fora Marine Protected Area (Porto Seguro, Bahia, Brazil). Sam-
dance < 5%) associated with the hydrocoral Millepora alcicornis pling periods with bleaching incidence are in red
and the coral Mussismilia harttii at the three sampled sites along the

M. harttii there was a change in dominance during the bleach- difference in symbiodiniacean phylotypes’ relative abun-
ing period, where C3 was replaced by A4 at Mourão and A4 dance between collection periods for this host (Table 1).
by C3 at Taquaruçu. After recovery, the symbiont assemblage The rare biospheres of photosymbionts (< 5% relative
resumed its original composition (Fig. 3). abundance) did not suffer marked changes, except for the
In general, β-diversity analysis (non-metric multidi- presence of B43 and the increase of B18b-related phylo-
mensional scaling, nMDS, and Table S2) showed that types in M. alcicornis during and after bleaching (Fig. 3),
photosymbiont assemblages of M. alcicornis were distinct but the relative abundance of M. alcicornis’ photosymbi-
between pre- and post-bleaching, with post-bleaching onts changed significantly throughout the monitored period
assemblages forming a more heterogeneous group (Fig. 4). (PERMANOVA: ­F7 = 2.4861, p < 0.01; Table 1). Indicator
Conversely, M. harttii colonies collected before bleach- species analysis showed that three OTUs were significantly
ing were more divergent from one another than colonies associated with both bleaching and recovery periods in M.
collected during the bleaching event (Fig. 4). After recov- alcicornis: TL-B43/B23 (stat = 0.452, p = 0.0026), TL-B40
ery, colonies returned to their original assemblages, and (stat = 0.451, p = 0.0022) and TL-B18b (stat = 0.559, p <
β-diversity was recovered, but there was no significant 0.0001), being present only from March onwards.

13
Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont Assemblages of Corals and…

Fig. 4  Non-metric multidi-

mensional scaling (nMDS)
biplot based on Bray-Curtis
index dissimilarity showing the
configuration of the relative
abundances of Symbiodiniaceae
ITS2 rDNA phylotypes within
different coral host species
along the sampling months
grouped by bleaching incidence
(ellipses)

Table 1  PERMANOVA based on Bray-Curtis index dissimilarity of TL-B30/B23 at Mourão (stat = 0.269, p = 0.0242), Bre-
the Symbiodiniaceae composition present in Millepora alcicornis and violum TL-B40 at Taquaruçu (stat = 0.286, p = 0.0167),
Mussismilia harttii from three sites at Recife de Fora, Brazil, over
Breviolum TL-B43/B23 grouped Funil+Taquaruçu (stat =
6-month monitoring (8 sampling periods). Pairwise PERMANOVA
only for the “collection site” factor 0.265, p = 0.0283), and Breviolum TL-B18b/B23 and TL-
B18b grouped Mourão+Taquaruçu (stat = 0.376/0.281, p
Factor Df Sum Sq F R2 p-value
= 0.0009/0.0233, respectively). Regarding the photosym-
Millepora alcicornis bionts of M. harttii, four were significantly associated with
Sampling period 7 0.0047 2.4861 0.16353 0.001 reef sites: Symbiodinium A4, Symbiodinium A1dh, and
Collection site 2 0.0028 5.1621 0.09701 0.001 Breviolum B23 at Taquaruçu (stat = 0.578/0.574/0.422,
Period:Site 13 0.0034 0.9763 0.11926 0.516 respectively, p < 0.0001), and Cladocopium C3 grouping
Funil vs Mourão 1 0.0018 8.2160 0.1279 0.003 Funil+Mourão (stat = 0.507, p < 0.0001).
Funil vs Taquaruçu 1 0.0018 5.5412 0.0845 0.003
Mourão vs Taquaruçu 1 0.0004 1.1018 0.0193 0.969
Mussismilia harttii Discussion
Sampling period 7 0.4795 0.7978 0.0433 0.617
Collection site 2 2.9135 16.9661 0.2633 0.001 The coral bleaching event during the summer of 2015/2016
Period:Site 14 1.8336 1.5253 0.1657 0.089 was severe at Recife de Fora (> 50% of the colonies bleached;
Funil vs Mourão 1 0.0551 1.1939 0.0192 0.999 [37]⁠), as other reefs of south Bahia [18, 38] even with rela-
Funil vs Taquaruçu 1 2.3137 18.8833 0.2456 0.003 tively low thermal stress estimations (< 4 °C-weeks degree
Mourão vs Taquaruçu 1 2.2420 21.5365 0.2674 0.003 heating weeks; DHW), but no coral mortality was recorded.
Bleaching occurrences without DHW values ≤ 4 °C-weeks
p-values in bold are significative (< 0.05)
have already been recorded, being attributed to the accumu-
lation of thermal anomalies lower than 1 °C persisting over
The effect of the collection site was significant for sym-
months [33, 39]⁠, or to the synergic effect of thermal and lumi-
biodiniaceans’ relative abundance from both host species
nous stress in shallow reefs [40, 41], or even that NOAA’s
(PERMANOVA: M. alcicornis - F ­ 7 = 5.1621, p < 0.01;
long-term MMM is not adequate, as a threshold, to estimate
M. harttii - ­F7 = 16.966, p < 0.01; Table 1). Pairwise PER-
thermal stress on corals at a local scale. However, this is an
MANOVA showed that only Mourão was not different from
indication that CRW general bleaching threshold may not be
Taquaruçu for M. alcicornis symbionts, and only Funil was
suitable as a bleaching predictor in Southwestern Atlantic
not different from Mourão for M. harttii (Table 1). Indica-
(SWA) reefs, as reported for other regions [39]⁠, and consistent
tor species analysis showed that 7 OTUs associated with M.
long-term bleaching and temperature monitoring coupled with
alcicornis were significantly associated with reef sites: Bre-
analyses of historical coral bleaching are necessary to improve
violum TL-B23 and TL-B19/B18b/B18a/B18 at Funil (stat
bleaching threshold prediction for SWA corals.
= 0.524/0.335, p = 0.0001/0.0096, respectively), Breviolum

13
 A. G. Garrido et al.

Although no thermal anomalies were recorded at period have never been observed. They may have helped the
Taquaruçu, the higher bleaching occurrence at this site could holobiont to deal with the thermal stressful period.
have resulted from the high levels of irradiance experienced Changes in photosymbionts dominance due to bleaching
by corals at this shallow tide pool [40]⁠. Unlike the other two events could represent a way for coral adaptation to climate
studied sites that receive warm water washed over the reef change [7]⁠, and those changes have already been observed
during tidal rises, this one faces the open sea and receives [9, 46]. However⁠ , they must be long-lasting and reach
cooler water during high tides, decreasing the daily mean most of the coral population. However, without permanent
SST locally. So, even though the stress faced by the cor- changes in dominance, we found no evidence of adaptation
als was higher at this site, the DHW calculation based on after bleaching for M. harttii, at least regarding this domi-
NOAA 5 km virtual station long-term MMM was not repre- nant portion of the photosymbiotic assemblage. In this case,
sentative of this stress in this locality. the changes in photosymbiotic assemblage observed in M.
As expected, Mussismilia spp. were less susceptible to harttii during bleaching seem more related to a dysbiosis
bleaching due to morphological characteristics and high het- process. The incidence of coral bleaching by itself has been
erotrophy capability [20, 21]⁠. On the contrary, milleporids defined as a dysbiosis simply due to the breakdown of the
are historically classified as “losers” been highly susceptible relationship of the coral with the most common photosymbi-
to bleaching and mortality [18, 42, 43], although its het- onts [47, 48]. Moreover, depending on the path taken by the
erotrophic feeding can also supply energetic demands dur- holobiont, distinct dysbiosis profiles could be determinants
ing bleaching [44, 45],⁠ and no mortality was observed here. for coral death or survival after bleaching. In this sense, the
According to Marangoni et al. [30], who assessed the oxi- temporary dominance changes in symbiont relative abun-
dative stress of the samples analyzed herein, M. alcicornis dances in the population of M. harttii during the bleaching
had shown an oxidative stress condition before and during peak may be a result of the metabolic imbalance inherent to
bleaching, when total antioxidant capacity (TAC) and lipid bleaching [47], or a strategy to maintain essential functions
damage increased. On the other hand, M. harttii naturally of the holobiont during stress [48].
operated with maximal capacity against reactive oxygen The presence of Breviolum phylotype B23 in M. alci-
species formation, with no distinction between healthy and cornis was recently described as dominant for this reef
bleached colonies, but TAC decreased in March, before the region [25]⁠, but the rare biosphere of photosymbionts of
peak of bleaching, which might indicate an imbalance in the this hydrocoral had never been assessed until now. Contrast-
oxidative status before the occurrence of bleaching. Also, ing with M. harttii, post-bleaching samples of M. alcicornis
oxidative stress biomarkers indicated a recovery, resuming were more heterogeneous and different from pre-bleaching
healthy levels by June 2016 [30]⁠. In addition, M. harttii was samples. As we did not observe changes in dominance in the
able to increase heterotrophic activity 15 to 30 days before composition of assemblages before, during or after bleach-
bleaching symptoms appeared, while a lower heterotrophic ing, this result indicates that changes in the rare biosphere
activity was observed for M. alcicornis [29]⁠, which may was mainly responsible for these observed differences over
have also influenced the difference in bleaching intensity time. The emergence of new symbionts through switching
between these two species. or shuffling after bleaching has already been reported for
The dominant symbiodiniaceans found in M. harttii at branched corals [49]⁠ and has been attributed to coral accli-
Recife de Fora (Cladocopium phylotype C3 and Symbiod- mation/adaptation to the new conditions of the environment.
inium linucheae, ITS2 phylotype A4) have already been In this case, the dysbiosis process may confer certain advan-
described in association with other Mussismilia spp. [22, 23] tages to the holobiont (adaptive dysbiosis; [48]). Background
and recently with M. harttii (Maia et al. in prep). Interestingly, photosymbiotic lineages can play a relevant role in host fit-
Symbiodinium spp. has already been identified in Mussismilia ness and resilience [50]⁠ since coral tissue repopulation can
spp. from well-lit environments [22, 23]⁠, such as Taquaruçu start by multiplying remaining symbionts [51]. However⁠,
shallow tide pool. In contrast, Cladocopium spp. were fre- its influence on holobiont physiology and its contribution
quently found in more turbid habitats [23]⁠, such as Funil and to host recovery is poorly understood [50]⁠.
Mourão, the washed patch reefs on the inner and southern The disruption of the homeostasis between coral and
edges of Recife de Fora. Nevertheless, it is difficult to attribute symbionts by thermal stress can also lead to an outbreak of
putative function or assume physiological characteristics to opportunistic symbiont lineages. Symbiodiniaceans from the
OTUs identified as “too-low similarity” (TL-), particularly genus Durusdinium have frequently been defined as oppor-
for the genus Breviolum, which presents a high diversity of tunistic or thermotolerant [10] due to their wide distribution,
physiologies [10]⁠ and for which several new lineages have physiologic characteristics, and documented replacement of
been discovered in the Southwestern Atlantic (SWA; [25]⁠). healthy coral’s symbionts after induced thermal stress [9,
However, the temporal changes in the relative abundance of 46] and in Caribbean coral populations after bleaching [52].
Symbiodinium A4 in M. harttii colonies during the bleaching The dominance effects of opportunistic symbiodiniaceans in

13
Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont Assemblages of Corals and…

coral fitness is still not well known, but a few studies have Data Availability The molecular data generated and analyzed during
shown that they have reduced coral growth rates [53]. To the current study are available in the NCBI Sequence Read Archive
database under the BioProject ID PRJNA943220.
date, Durusdinium spp. has never been found in the SWA.
Some Symbiodinium spp. have also been considered oppor- Declarations
tunistic for some Caribbean corals [54]. However, since S.
linucheae (phylotype A4) is commonly found in healthy cor- Ethics Approval This study was performed under the sampling per-
als in the SWA (this study and [22–25]), we believe that its mission (# 47714-1) of the Brazilian Institute of Environment and
Renewable Natural Resources (IBAMA) / Chico Mendes Institute for
increased dominance in M. harttii tissue at Taquaruçu may Biodiversity Conservation (ICMBio), under the Instruction Normative
be due to better fitness of this phylotype during the bleaching n° 03/2014 of System Authorization and Information on Biodiversity
period when compared to the co-dominant Cladocopium C3. (SISBIO).
The present study documented the distinct responses of
Competing Interests The authors declare no competing interests.
a coral and a hydrocoral facing a severe bleaching event in
a model reef in the Southwestern Atlantic during the ENSO
2015/2016, with no mortality records. The factors contrib-
uting to holobionts’ survival are still poorly understood.
However, they could be attributed to the host mechanisms of References
oxidative stress mitigation [29, 30]⁠ and feed mode balancing
[20]⁠, but also to adjustments in photosymbiont assemblages 1. Hughes TP, Anderson KD, Connolly SR et al (2018) Spatial and
temporal patterns of mass bleaching of corals in the Anthropo-
composition during and after bleaching or all this together. cene. Science 359:80–83. https://​doi.​org/​10.​1126/​scien​ce.​aan80​
Even subtle changes in photosymbiont composition, such as 48
the documented dysbiosis, may contribute to the acclimati- 2. Brown BE (1997) Coral bleaching: causes and consequences.
zation of the holobiont [50]⁠. Nevertheless, assessing sym- Coral Reefs 16:S129–S138
3. Glynn PW (1996) Coral reef bleaching: facts, hypotheses and
bionts’ taxonomic and physiologic diversity is essential to implications. Glob Chang Biol 2:495–509. https://​doi.​org/​10.​
better comprehend the role of symbiodiniaceans in host resil- 1111/j.​1365-​2486.​1996.​tb000​63.x
ience in the face of climate change. These findings follow the 4. Hoegh-Guldberg O, Mumby PJ, Hooten AJ et al (2007) Coral
observed resilience of Brazilian coral reefs to bleaching [21]⁠, reefs under rapid climate change and ocean acidification. Sci-
ence 318:1737–1742
since both species were able to deal with heat stress without 5. Claar DC, Szostek L, McDevitt-Irwin JM et al (2018) Global
mortality. However, the recently documented vulnerability patterns and impacts of El Niño events on coral reefs: a meta-
of those species, mainly M. alcicornis, to more severe heat analysis. PLoS One 13:e0190957. https://​doi.​org/​10.​1371/​journ​
anomalies [18]⁠ reinforces the urgent need for monitoring al.​pone.​01909​57
6. Coffroth MA, Santos SR (2005) Genetic diversity of symbiotic
programs and research focused on SWA corals to determine dinoflagellates in the genus Symbiodinium. Protist 156:19–34.
priorities for coral reef conservation and management. https://​doi.​org/​10.​1016/j.​protis.​2005.​02.​004
7. Oliver TA, Palumbi SR (2011) Many corals host thermally
Supplementary Information The online version contains supplemen- resistant symbionts in high-temperature habitat. Coral Reefs
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00248-0​ 23-0​ 2299-3. 30:241–250. https://​doi.​org/​10.​1007/​s00338-​010-​0696-0
8. LaJeunesse TC, Parkinson JE, Gabrielson PW et al (2018) Sys-
Acknowledgements We would like to thank Dr. Rodrigo Nunes and tematic revision of symbiodiniaceae highlights the antiquity and
MSc. Bruno Rodrigues for providing the technical support for sequenc- diversity of coral endosymbionts. Curr Biol 28:2570–2580.e6.
ing and the infrastructure of the Integrated Functional Genomics Unit, https://​doi.​org/​10.​1016/j.​cub.​2018.​07.​008
and Dr. Raquel Peixoto for providing the storage of refrigerated sam- 9. Rowan R (2004) Thermal adaptation in reef coral symbionts.
ples. The Coral Vivo Project and its sponsors Petrobras, through the Nature 430:742
Petrobras Environmental Program, and Arraial d’Ajuda Eco Parque are 10. Swain TD, Chandler J, Backman V, Marcelino L (2017) Con-
acknowledged for supporting field research and logistics. sensus thermotolerance ranking for 110 Symbiodinium phylo-
types: an exemplar utilization of a novel iterative partial-rank
Author Contribution All authors contributed to the study’s conception aggregation tool with broad application potential. Funct Ecol
and design. Amana G. Garrido, Laís F. Machado, Cristiano M. Pereira, 31:172–183. https://​doi.​org/​10.​1111/​1365-​2435.​12694
and Douglas P. Abrantes performed material preparation, data collec- 11. Rouzé H, Lecellier G, Pochon X et al (2019) Unique quantitative
tion, and analysis. Carla Zilberberg and Emiliano N. Calderon led the Symbiodiniaceae signature of coral colonies revealed through
funding acquisition. Amana G Garrido wrote the manuscript’s first spatio-temporal survey in Moorea. Sci Rep 9:7921. https://​doi.​
draft, and all authors commented on previous versions. All authors org/​10.​1038/​s41598-​019-​44017-5
read and approved the final manuscript. 12. de Souza MR, Caruso C, Ruiz-Jones L et al (2022) Commu-
nity composition of coral-associated Symbiodiniaceae differs
Funding This work was supported by the Coral Vivo Project and its across fine-scale environmental gradients in Kãne’ohe Bay. R
sponsors Petrobras through the Petrobras Environmental Program Soc Open Sci 9:212042. https://​doi.​org/​10.​1098/​rsos.​212042
and Arraial d’Ajuda Eco Parque. A G Garrido was a Ph.D. fellow 13. Buddemeier RW, Fautin DG (1993) Bleaching as an adaptive
from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior mechanism. Bioscience 43:320–326
(CAPES) and the Fundação de Amparo à Pesquisa do Estado do Rio
de Janeiro (FAPERJ - Programa Bolsa Nota 10 # 2/2021).

13
 A. G. Garrido et al.

14. Ware JR, Fautin DG, Buddemeier RW (1996) Patterns of coral 30. de Barros MLF, Dalmolin C, Marques JA et al (2019) Oxidative
bleaching: modeling the adaptive bleaching hypothesis. Ecol stress biomarkers as potential tools in reef degradation monitor-
Model 84:199–214. https://​doi.​org/​10.​1016/​0304-​3800(94)​00132-4 ing: a study case in a South Atlantic reef under influence of the
15. Kinzie III RA, Takayama M, Santos SR, Coffroth MA (2001) 2015–2016 El Niño/Southern Oscillation (ENSO). Ecol Indic
The adaptive bleaching hypothesis: experimental tests of critical 106:105533. https://​doi.​org/​10.​1016/j.​ecoli​nd.​2019.​105533
assumptions. Biol Bull 200:51–58 31. Siebeck UE, Marshall NJ, Klüter A, Hoegh-Guldberg O (2006)
16. Castro CB, Pires DO (2001) Brazilian coral reefs: what we Monitoring coral bleaching using a colour reference card. Coral
already know and what is still missing. Bull Mar Sci 69:357–371 Reefs 25:453–460. https://​doi.​org/​10.​1007/​s00338-​006-​0123-8
17. de Oliveira SM, Teixeira CEP, Ferreira SMC et al (2019) Ther- 32. Liu G, Strong AE, Skirving W (2003) Remote sensing of sea sur-
mal stress and tropical reefs: mass coral bleaching in a stable face temperatures during 2002 barrier reef coral bleaching. EOS
temperature environment? Mar Biodivers. https://​doi.​org/​10.​ Trans Am Geophys Union 84:137–141. https://​doi.​org/​10.​1029/​
1007/​s12526-​019-​00994-4 2003E​O1500​01
18. Ferreira LCL, Grillo AC, Filho FPMR et al (2021) Different 33. Le Nohaïc M, Ross CL, Cornwall CE et al (2017) Marine heat-
responses of massive and branching corals to a major heatwave wave causes unprecedented regional mass bleaching of thermally
at the largest and richest reef complex in South Atlantic. Mar resistant corals in northwestern Australia. Sci Rep 7:14999.
Biol 168:54. https://​doi.​org/​10.​1007/​s00227-​021-​03863-6 https://​doi.​org/​10.​1038/​s41598-​017-​14794-y
19. Lisboa DS, Kikuchi RKP, Leão ZMAN (2018) El Niño, sea 34. Fukami H, Budd AF, Levitan DR et al (2004) Geographic differ-
surface temperature anomaly and coral bleaching in the South ences in species boundaries among members of the Montastraea
Atlantic: a chain of events modeled with a Bayesian approach. annularis complex based on molecular and morphological mark-
J Geophys Res Ocean 123:2554–2569. https://​doi.​org/​10.​1002/​ ers. Evolution (N Y) 58:324–337. https://​doi.​org/​10.​1111/j.​0014-​
2017J​C0128​24 3820.​2004.​tb016​48.x
20. Banha TNS, Capel KCC, Kitahara MV et al (2020) Low coral 35. Hume BCC, Ziegler M, Poulain J et al (2018) An improved primer
mortality during the most intense bleaching event ever recorded set and amplification protocol with increased specificity and sen-
in subtropical Southwestern Atlantic reefs. Coral Reefs 39:515– sitivity targeting the Symbiodinium ITS2 region. PeerJ 6:e4816.
521. https://​doi.​org/​10.​1007/​s00338-​019-​01856-y https://​doi.​org/​10.​7717/​peerj.​4816
21. Mies M, Francini-Filho RB, Zilberberg C et al (2020) South 36. Team RC (2018) R: a language and environment for statistical
Atlantic coral reefs are major global warming refugia and less computing. R Foundation for Statistical Computing, Vienna, Aus-
susceptible to bleaching. Front Mar Sci 7:514. https://​doi.​org/​ tria. Available online at https://​www.R-​proje​ct.​org/
10.​3389/​FMARS.​2020.​00514 37. Donner SD, Rickbeil GJM, Heron SF (2017) A new, high-
22. Silva-Lima AW, Walter JM, Garcia GD et al (2015) Multiple resolution global mass coral bleaching database. PLoS One
Symbiodinium strains are hosted by the brazilian endemic corals 12:e0175490. https://​doi.​org/​10.​1371/​journ​al.​pone.​01754​90
Mussismilia spp. Microb Ecol 70:301–310. https://​doi.​org/​10.​ 38. de Oliveira Soares M (2018) Climate change and regional human
1007/​s00248-​015-​0573-z pressures as challenges for management in oceanic islands, South
23. Picciani N, de Lossio e Seiblitz IG, de Paiva PC et al (2016) Atlantic. Mar Pollut Bull 131:347–355. https://​doi.​org/​10.​1016/j.​
Geographic patterns of Symbiodinium diversity associated with marpo​lbul.​2018.​04.​008
the coral Mussismilia hispida (Cnidaria, Scleractinia) correlate 39. Liu B, Guan L, Chen H (2021) Detecting 2020 Coral Bleaching
with major reef regions in the Southwestern Atlantic Ocean. Event in the Northwest Hainan Island Using CoralTemp SST and
Mar Biol 163:236. https://​doi.​org/​10.​1007/​s00227-​016-​3010-z Sentinel-2B MSI Imagery. Remote Sens 13:4948. https://​doi.​org/​
24. Teschima MM, Garrido A, Paris A et al (2019) Biogeography 10.​3390/​rs132​34948
of the endosymbiotic dinoflagellates (Symbiodiniaceae) com- 40. Brown B, Dunne R, Goodson M, Douglas A (2002) Experience
munity associated with the brooding coral Favia gravida in the shapes the susceptibility of a reef coral to bleaching. Coral Reefs
Atlantic Ocean. PLoS One 14:1–18. https://​doi.​org/​10.​1371/​ 21:119–126. https://​doi.​org/​10.​1007/​s00338-​002-​0215-z
journ​al.​pone.​02135​19 41. Abrego D, Ulstrup KE, Willis BL, Van Oppen MJH (2008) Spe-
25. Garrido AG, de Assis Leite DC, Machado LF et al (2022) cies-specific interactions between algal endosymbionts and coral
Highly diverse and geographically differentiated Symbiodini- hosts define their bleaching response to heat and light stress. Proc
aceae communities associated with the hydrocoral Millepora R Soc B Biol Sci 275:2273–2282. https://​doi.​org/​10.​1098/​rspb.​
alcicornis in the Atlantic Ocean. Coral Reefs. https://​doi.​org/​ 2008.​0180
10.​1007/​s00338-​022-​02266-3 42. Cook CB, Logan A, Ward J et al (1990) Elevated temperatures and
26. Tedesco EC, Calderon EN, Seoane JCS et al (2018) Coral reef bleaching on a high latitude coral reef: the 1988 Bermuda event.
benthic assemblages of a Marine Protected Area in Eastern Coral Reefs 9:45–49. https://​doi.​org/​10.​1007/​BF006​86721
Brazil: effect of reef habitats on the spatial pattern of species. 43. Banaszak AT, Ayala-Schiaffino BN, Rodríguez-Román A et al
J Nat Hist 52:2723–2743. https://​doi.​org/​10.​1080/​00222​933.​ (2003) Response of Millepora alcicornis (Milleporina: Mille-
2018.​15523​32 poridae) to two bleaching events at Puerto Morelos reef, Mexican
27. Santos HF, Carmo FL, Duarte G et al (2014) Climate change Caribbean. Rev Biol Trop 51:57–66. https://​doi.​org/​10.​15517/​rbt.​
affects key nitrogen-fixing bacterial populations on coral reefs. v51i4.​26429
ISME J 8:2272–2279. https://​doi.​org/​10.​1038/​ismej.​2014.​70 44. Lewis JB (1992) Heterotrophy in corals: zooplankton predation
28. Leite DCA, Leão P, Garrido AG et al (2017) Broadcast spawning by the hydrocoral Millepora complanata. Mar Ecol Prog Ser
coral Mussismilia hispida can vertically transfer its associated 90:251–256. https://​doi.​org/​10.​3354/​meps0​90251
bacterial core. Front Microbiol 8:176. https://​doi.​org/​10.​3389/​ 45. Olguín-López N, Hérnandez-Elizárraga VH, Hernández-Mate-
fmicb.​2017.​00176 huala R et al (2019) Impact of El Niño-Southern Oscillation 2015-
29. de Barros MLF, Mies M, Güth AZ et al (2019) Peroxynitrite 2016 on the soluble proteomic profile and cytolytic activity of
generation and increased heterotrophic capacity are linked to the Millepora alcicornis (“fire coral”) from the Mexican Caribbean.
disruption of the coral–dinoflagellate symbiosis in a Scleractinian PeerJ 2019:1–26. https://​doi.​org/​10.​7717/​peerj.​6593
and hydrocoral species. Microorganisms 7:426. https://d​ oi.o​ rg/1​ 0.​ 46. Baker AC (2003) Flexibility and specificity in coral-algal symbio-
3390/​micro​organ​isms7​100426 sis: diversity, ecology, and biogeography of Symbiodinium. Annu

13
Marine Heatwave Caused Differentiated Dysbiosis in Photosymbiont Assemblages of Corals and…

Rev Ecol Evol Syst 34:661–689. https://​doi.​org/​10.​1146/​annur​ev.​ 52. LaJeunesse TC, Smith RT, Finney J, Oxenford H (2009) Outbreak
ecols​ys.​34.​011802.​132417 and persistence of opportunistic symbiotic dinoflagellates during
47. Bosch TCG, Miller DJ (2016) Bleaching as an obvious dysbiosis the 2005 Caribbean mass coral ‘bleaching’ event. Proc R Soc B
in corals. The Holobiont Imperative. Springer, Vienna. https://d​ oi.​ 276:4139–4148. https://​doi.​org/​10.​1098/​rspb.​2009.​1405
org/​10.​1007/​978-3-​7091-​1896-2_9 53. Jones A, Berkelmans R (2010) Potential costs of acclimatization
48. Boilard A, Dubé CE, Gruet C et al (2020) Defining coral bleach- to a warmer climate: growth of a reef coral with heat tolerant vs.
ing as a microbial dysbiosis within the coral holobiont. Microor- sensitive symbiont types. PLoS One 5:e10437. https://​doi.​org/​10.​
ganisms 8:1682. https://d​ oi.o​ rg/​10.3​ 390/m
​ icro​organi​ sms81​ 11682 1371/​journ​al.​pone.​00104​37
49. Jones A, Berkelmans R, van Oppen M et al (2008) A community 54. Toller WW, Rowan R, Knowlton N (2010) Repopulation of zoox-
change in the algal endosymbionts of a scleractinian coral follow- anthellae in the Caribbean corals Montastraea annularis and M.
ing a natural bleaching event: field evidence of acclimatization. faveolata following experimental and disease-associated bleach-
Proc R Soc B Biol Sci 275:1359–1365. https://​doi.​org/​10.​1098/​ ing. Biol Bull 201:360–373. https://​doi.​org/​10.​2307/​15436​14
rspb.​2008.​0069
50. Ziegler M, Eguíluz VM, Duarte CM, Voolstra CR (2018) Rare sym- Springer Nature or its licensor (e.g. a society or other partner) holds
bionts may contribute to the resilience of coral–algal assemblages. exclusive rights to this article under a publishing agreement with the
ISME J 12:161–172. https://​doi.​org/​10.​1038/​ismej.​2017.​151 author(s) or other rightsholder(s); author self-archiving of the accepted
51. Boulotte NM, Dalton SJ, Carroll AG et al (2016) Exploring the manuscript version of this article is solely governed by the terms of
Symbiodinium rare biosphere provides evidence for symbiont such publishing agreement and applicable law.
switching in reef-building corals. ISME J 10:2693–2701. https://​
doi.​org/​10.​1038/​ismej.​2016.​54

13