Integrative and Comparative Biology Advance Access published August 4, 2016

Integrative and Comparative Biology

Integrative and Comparative Biology, pp. 1–14
doi:10.1093/icb/icw095 Society for Integrative and Comparative Biology

SYMPOSIUM

Microbiome Composition and Diversity of the Ice-Dwelling Sea

Anemone, Edwardsiella andrillae

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
Alison E. Murray,* Frank R. Rack,§ Robert Zook,§ Michael J. M. Williams,‡ Mary L. Higham,*
Michael Broe,† Ronald S. Kaufmann,ô and Marymegan Daly,1,†
*Division of Earth and Ecosystem Sciences, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, USA;
†
Department of Evolution, Ecology, and Organismal Biology, The Ohio State University, 1315 Kinnear Rd, Columbus,
OH 43212, USA; ‡National Institute of Water and Atmospheric Research, 301 Evans Bay Parade, Wellington, 6021, New
Zealand; §ANDRILL Science Management Office, 126 Bessey Hall, University of Nebraska-Lincoln, Lincoln, NE 68588-
0341, USA; ôDepartment of Environmental and Ocean Sciences, University of San Diego, 5998 Alcalá Park, San Diego,
CA 92110-2492, USA
From the symposium ‘‘Life on the Edge: the Biology of Organisms Inhabiting Extreme Environments’’ presented at the
annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2016 at Portland, Oregon.
1
E-mail: [email protected]

Synopsis Edwardsiella andrillae is a sea anemone (Cnidaria: Anthozoa: Actiniaria) only known to live embedded in the
ice at the seawater interface on the underside of the Ross Ice Shelf, Antarctica. Although the anatomy and morphological
characteristics of E. andrillae have been described, the adaptations of this species to the under-ice ecosystem have yet to
be examined. One feature that may be important to the physiology and ecology of E. andrillae is its microbiome, which
may play a role in health and survival, as has been deduced in other metazoans, including anthozoans. Here we describe
the microbiome of five specimens of E. andrillae, compare the diversity we recovered to that known for temperate
anemones and another Antarctic cnidarian, and consider the phylogenetic and functional implications of microbial
diversity for these animals. The E. andrillae microbiome was relatively low in diversity, with seven phyla detected, yet
included substantial phylogenetic novelty. Among the five anemones investigated, the distribution of microbial taxa
varied; this trait appears to be shared by many anthozoans. Most importantly, specimens either appeared to be dominated
by Proteobacteria-affiliated members or by deeply branching Tenericute sequences. There were few closely related se-
quence types that were common to temperate and Antarctic sea anemone microbiomes, the exception being an
Acinetobacter-related representative. Similar observations were made between microbes associated with E. andrillae and
an Antarctic soft coral; however, there were several closely-related, low abundance Gammaproteobacteria in both
Antarctic microbiomes, particularly from the soft coral, that are also commonly detected in Southern Ocean seawater.
Although this preliminary study leaves open many questions concerning microbiome diversity and its role in host
ecology, we identify major lineages of microbes (e.g., diverse deep-branching Alphaproteobacteria,
Epsilonproteobacteria, and divergent Tenericutes affiliates) that may play critical roles, and we highlight the current
understanding and the need for future studies of sea anemone–microbiome relationships.

Introduction microbes (e.g., Rusch et al. 2007; Sunagawa et al.

The importance of microbes to the health and ecol- 2015), the extent of diversity and interactions be-
ogy of marine animals and the ecosystems in which tween marine animals and host-associated microbes
they live is well-appreciated at a conceptual level, yet (termed the microbiome) is still poorly known across
understanding microbial diversity, interactions, and organisms and ecosystems (Bourne et al. 2009;
functions remains elusive in many systems. Although Ainsworth et al. 2010). In addition to playing a foun-
global plankton surveys are shedding light on the dational role in shaping evolution (McFall-Ngai et al.
diversity and functional capacities of free-living 2013) and development (e.g., Fraune and Bosch

ß The Author 2016. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology.
All rights reserved. For permissions please email: [email protected].
2 A. E. Murray et al.

2010), microbes can play significant roles in animal

health (e.g., Ritchie 2006; Rosenberg et al. 2007;
Krediet et al. 2013; Lokmer and Wegner 2015) and
nutrient acquisition (e.g., Hoffmann et al. 2009;
Roeselers and Newton 2012). Marine microbes that
are symbiotic can also play a critical supportive role
in the physiological acclimatization or adaptation of
their hosts to challenging habitats (e.g., Grzymski et
al. 2008; Schönknecht et al. 2013; O’Connor et al.

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
2014).
The ecosystems of the Southern Ocean and its
associated coastal and ice habitats represent a partic-
ular deficit in our understanding of microbial sym-
bioses. What is known from this large, diverse set of Fig. 1. Underside of the Ross Ice Shelf with E. andrillae imbedded
in the ice. Image captured using the ROV SCINI at Coulman
ecosystems highlights the potential importance of
High.
microbes to their host. Microbial diversity in
Antarctic sponges (Webster et al. 2004; Rodriguez-
Marconi et al. 2015) and soft corals (Webster and an extensive investigation into the microbiome of the
Bourne 2007) is host-specific and distinct from that model anemone, Nematostella vectensis. Har et al.
of lower latitude marine ecosystems. The bacterial
(2015) found a pattern of microbial association
diversity of the Antarctic ascidian, Synoicum adarea-
that suggests that, like their relatives the scleractinian
num, is limited, but bacterial symbionts are poten-
corals, actiniarians may have selective and specific
tially responsible for biosynthesis of the bioactive
microbial associations and that these may be critical
compound, palmerolide, in the holobiont
to supporting host survival. Indeed, metatranscrip-
(Riesenfeld et al. 2008). Further, microbial partners
tome and bacterial cultivar genome sequencing ef-
have been implicated in cryoadaptation of the
forts in Nematostella indicated potential microbial
Antarctic ciliate, Euplotes focardii, in which two bac-
roles in energy metabolism, nutrient acquisition
terial genome-encoded ice binding proteins were
and storage, and environmental acclimation (Har et
identified (Pucciarelli et al. 2014). These examples
are just the ‘‘tip of the iceberg’’ with respect to un- al. 2015).
covering the expansive diversity of marine microbial Here we describe the bacterial microbiome of
symbioses in the Southern Ocean. Edwardsiella andrillae, an ice-inhabiting actiniarian
Actiniarian sea anemones are among the most that lives with its body column in the ice at under-
conspicuous inhabitants of marine ecosystems, side of the Ross Ice Shelf, with the tentacle crown
where they play key roles in benthic-pelagic coupling and mouth at the ice–water interface (Fig. 1; Rack et
and primary production through their symbioses al. 2012). E. andrillae is the first actiniarian known to
with photosynthetic unicellular organisms (e.g., have an ice-associated lifestyle. Although broad salin-
Sebens 1981; Shick 1991). Actiniarians are unique ity tolerance is fairly rare across sea anemones (Shick
among anthozoans in showing broad physiological 1991), it is most common in Edwardsiidae, the
tolerance with respect to salinity, temperature, and family that includes E. andrillae (Shick 1991; Hand
oxygen levels (reviewed in Shick 1991). Excluding the and Uhlinger 1994; Daly et al. 2012). In its burrows
extensive study of the diversity and interactions be- in the dynamic basal ice shelf frontal zone, E. andril-
tween actiniarians and photosynthetic eukaryotes lae may encounter conditions that are challenging
(e.g., LaJeunesse and Trench 2000; Müller-Parker and variable. Their food sources are unknown, as is
and Davy 2001; Lewis and Müller-Parker 2004), the their role as prey for, e.g., fishes, crustaceans, or
role of the microbiome in the ecology and physiol- other organisms. Temperatures in this aphotic habi-
ogy of actiniarians, particularly those in habitats tat are close to 2 8C, although in summer, waters
devoid of photosynthetically active radiation, is vir- up to 0.4 8C above freezing can be advected under
tually unknown. Reports of actiniarian host-microbe the ice shelf from the Ross Sea Polynya (Arzeno et al.
associations were limited to a few microscopic obser- 2014). Salinity can range from near zero in meltwater
vations and studies of microbial diversity via rRNA in the ice-boundary layer to 34.6–34.9 ppt in ambi-
gene sequences (Schuett et al. 2007; Williams et al. ent sea water (Jacobs et al. 2002).
2007; Du et al. 2010; Schuett and Doepke 2010; Microbes are key to the ecosystems of the ‘‘cryo-
Meron et al. 2013) until Har et al. (2015) conducted sphere’’ (reviewed in Boetius et al. 2015) and may
Ice-anemone microbiome 3

play a role in the adaptive capability of E. andrillae Foster City, CA)] followed by phenol:chloro-
to life in the ice shelf. Our preliminary observations form:isoamyl alcohol purification following
of the E. andrillae microbiome include its diversity, Massana et al. (1997). Extracts resuspended in
community structure, and phylogenetic and func- 10 mM Tris-Cl were quantified by fluorescence with
tional context. Our study of the microbiome of E. Picogreen (Life Technologies, Carlsbad, CA) on a
andrillae benefits from its close phylogenetic rela- Spectramax Gemini (Molecular Devices, Sunnyvale,
tionship to N. vectensis, facilitating comparisons CA). The DNA extracts were screened by polymerase
with the N. vectensis microbiome (Har et al. 2015). chain reaction (PCR) using primers targeting the
We also compared the E. andrillae microbiome to three domains of life. The bacterial rRNA gene was

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
the only other Antarctic anthozoan microbiome targeted using primers 27F and 1931R (Lane 1991);
data set known to us (Alcyonia antarctica; Webster the archaeal rRNA gene was targeted using primers
and Bourne 2007). 20F (Massana et al. 1997) and 958R (DeLong 1992);
and the eukaryal rRNA gene was amplified using
Materials and Methods primers 960F and 1200R (Gast et al. 2004).
Sample collection Following successful amplification of the 16S rRNA
gene to verify amplifiability of the extracts, DNA
The Remotely Operated Vehicle Submersible was prepared for pyrosequencing (MRDNA,
Capable of Under-Ice Navigation and Imaging Lubbock, TX). Briefly, barcoded amplicon sequenc-
(ROV SCINI; Cazenave et al. 2011) was deployed ing used bacteria-targeted primers (27Fmod
through a  30 cm borehole in the Ross Ice Shelf. GRGTTTGATCMTGGCTCAG and 519R
Specimens of E. andrillae were recovered using a GWATTACCGCGGCKGCTG) in a single-step 30
makeshift suction sampler (see Daly et al. 2013 for cycle PCR with a HotStarTaq Plus Master Mix Kit
details) at dive site 3 (77.5267 S, 171.3350 E), (Qiagen, Valencia). Amplification conditions were: de-
10 km south of the ice shelf edge in late December naturation at 948C for 3 min; followed by 28 cycles at
2010 (Rack et al. 2012). At that location, the ice shelf 948C for 30 s, 538C for 40 s, and 728C for 1 min; fol-
was  260 m thick, with  220 m below mean sea lowed by a final elongation step at 728C for 5 min.
level, and the sea floor  570 m below the bottom of The amplicon product libraries were mixed in equal
the ice (Rack et al. 2012). Once recovered to the proportions and purified using Agencourt Ampure
surface, anemone specimens were initially stored in beads (Agencourt Bioscience Corp., Beverly, MA).
sterile tubes with denatured alcohol (the only preser- Libraries were sequenced using Roche 454 FLX
vative available on site) then transferred to 96% eth- titanium reagents and instrument following the
anol after  2 weeks and stored at 48C. The sample manufacturer’s guidelines at a scale targeting
numbers in the manuscript indicate the numbers 3K sequences per library. Combined SFF files re-
given to each anemone at the time of collection. ceived from MRDNA were analyzed using Mothur
Seawater physical and chemical parameters were (version 1.36.1) following the 454SOP (Schloss et
measured at the ice shelf–seawater interface using a al. 2011), which incorporates the PyroNoise algo-
Seabird Electronics SBE19þ CTD and Niskin bottle rithm to mitigate errors in amplification and se-
to collect water for chemical analysis. Results from quencing (Quince et al. 2009). The pipeline
three replicate CTD casts are presented here. utilized seed alignment and taxonomy files pro-
vided by SILVA rRNA database project (www.
Microbiome sampling arb-silva.de) Silva.seed_v119.align silva.seed_v119.-
Five specimens of E. andrillae were each rinsed in tax and Perseus (Quince et al. 2011) to detect chi-
sterile seawater and then homogenized by hand meras, which removed 225 sequences. Sequences
with a sterile pellet pestle (Sigma-Aldrich, St. Louis, with ambiguities were removed, as were all se-
MO) in 1 mL of sterile seawater in a 1.5 mL micro- quences with48 homopolymers, and sequences
centrifuge tube. Homogenates were centrifuged for shorter than 341 bases. The final data set contained
2 min at 3000 g to pellet the larger cellular debris, 12,052 sequences (1660–3919 per library; 341–386
and then the supernatant was removed and centri- bases) that fell into 771 unique Operational
fuged at maximum speed for 10 min to pellet the Taxonomic Units (OTUs), and 163 OTUs de-
microbial cells. The cells were resuspended in fined at a distance of 0.03 (97 OTUs when single-
200 L sucrose lysis buffer, and DNA was extracted tons were removed). The sff data sets and
using an enzymatic digestion [lysozyme (Roche, associated MIMARKS file have been submitted to
Branchburg, NJ), proteinase K (Thermo Fisher the National Center for Biotechnology Information
Scientific, Waltham, MA), and RNAse A (Ambion, (NCBI) under BioProject PRJNA315709, and the
4 A. E. Murray et al.

sample metadata are also available at the Microbial much more complete picture of host life, in an eco-
Antarctic Resource System (mARS.biodiversity.aq). system bathed in microbial food sources, allies, and
adversaries (Kelly et al. 2014; Ainsworth et al. 2015;
Data analysis Daniels et al. 2015).
OTUs were taxonomically classified in Mothur. The waters at the ice shelf–seawater interface (mean
Comparative analyses between the microbiomes of depth of 256 m) had a mean ( standard deviation)
anemones Ea3–7 were conducted with data that temperature of 1.949  0.0538C (n ¼ 3) and salinity
had been subsampled to the lowest number of se- of 34.573  0.017 psu, suggesting little fresh water at
quences for all five anemone microbiomes using the interface, which was close to the pressure-depen-

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
Mothur for (Chao1) and -diversity (Bray– dent freezing temperature. The inorganic nutrients
Curtis) statistics and Daisy Chopper (Gilbert et al. were replete with NO3-N (419 mg m3) and dissolved
2009) for presence/absence and relative abundance of reactive phosphorus (DRP; 65 mg m3), giving a N:P
taxonomic classes. The subsampled microbiome data molar ratio of 14.25 that is lower than the Redfield
sets were clustered based on Spearman rank correla- ratio of 16, and not indicative of waters depleted
tion (we used a nonparametric approach, given the either in P by Phaeocystis or in N by diatoms in
non-linear nature of the occurrence data) using Ross Sea waters (Arrigo et al. 2002).
Cluster 3.0 (adapted from Eisen et al. 1998), and Despite unconventional sample preservation meth-
the heatmap representation and clustering were visu- ods for molecular analysis of microorganisms, DNA
alized using Java Treeview 3.0 (https://sourceforge. yields from microbial cell-enriched preparations of E.
net/projects/jtreeview/). Rarefaction and rank abun- andrillae retrieved by SCINI in late 2010 were ade-
dance curves derived from Mothur were based on quate for downstream analysis (0.8–3.3 g), and 16S
observed samples of uneven size. Phylogenetic diver- rRNA gene PCR products were amplified using bac-
sity comparisons of microbiomes from E. andrillae terial and eukaryal primer sets. Although Archaea
and N. vectensis (Har et al. 2015) were a focus of this have been found in association with other Antarctic
study, although we also included the clone library marine invertebrates (e.g., Webster et al. 2004;
and culture-based sequence data from Alcyonium Rodriguez-Marconi et al. 2015), we did not recover
antarcticum sampled in McMurdo Sound ( any archaeal sequences in the E. andrillae micro-
118 km from Coulman High; Webster and Bourne biome. Thus far, we have not investigated the
2007). All sequence data sets were aligned using the nature of the eukaryal rRNA gene signal to deter-
SILVA Incremental Aligner (SINA) aligner and bacte- mine whether there are parasites, fungi, or other
rial variability profile (Pruesse et al. 2012). The align- commensal eukaryotes associated with the E. andril-
ment was inspected and trimmed in MEGA (version lae host.
6; Tamura et al. 2013), and analyses included gener-
ating a distance matrix, calculating evolutionary diver- Taxonomic composition of the microbiome
gence of sequence pairs between groups, and
Low coverage pyrosequencing of the V1-3 regions of
neighbor-joining phylogenetic analysis using 1000
the bacterial rRNA gene was performed to obtain an
bootstrap resampling. In all pairwise comparisons,
initial picture of the E. andrillae-associated bacterial
ambiguous positions were removed from each pair.
diversity. Sequences were clustered into OTUs at a
Subtrees were generated from the same analysis.
distance of 0.03. The complete (non-subsampled)
data set had 163 OTUs. The subsampled data set
Results and Discussion had 126 OTUs, of which 47 were represented by a
Other than the taxonomic characterization of E. single sequence and not included in the analysis here,
andrillae, virtually nothing is known of the life history given the risk of errors in sequencing (Kunin et al.
of this anemone, nor the ecosystem characteristics at 2010; Flynn et al. 2015). The 79 OTUs with two or
the ice shelf–seawater interface. One potentially critical more sequences remaining in the data set indicated a
aspect for understanding E. andrillae and its ecology is low–moderate level of diversity, with 10–36 OTUs
understanding the associations between the anemone per anemone microbiome. The OTUs could be clas-
and natural microbiota in the under-ice shelf ecosys- sified into six phyla, with Bacteroidetes and
tem. We attempt to view the E. andrillae holobiont in Proteobacteria represented by multiple lineages
a similar framework as is currently done with corals (Fig. 2A). We found two OTUs that were related
and other species for which species-specific associa- to each other but could not be placed in a known
tions, metagenome-encoded metabolic functions, and phylum and thus may represent an undescribed
even metatranscriptome-expressed responses provide a phylum. One phylum, Firmicutes, represented in
Ice-anemone microbiome 5

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016

Fig. 2. Class-level taxonomic diversity of microbial OTUs from each specimen of E. andrillae. (A) OTU presence/absence across all E.
andrillae microbiomes sampled and (B) the relative abundance of sequences assigned to taxonomic class. Singleton OTUs are not
included. Anemone samples are abbreviated with respect to their collection ID.

the initial data set by one OTU, was eliminated representatives in one microbiome (Ea5), 11 in two
through the subsampling process. others (Ea3 and Ea4), and interestingly only 1 in
The lineage with the greatest number of OTUs was Ea6, and none in Ea7 (Fig. 2A). OTUs
Gammaproteobacteria, for which there were 12 associated with Flavobacteria, Alphaproteobacteria,
6 A. E. Murray et al.

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
Fig. 3 Heatmap representation of OTU sequence depth across the five specimens of E. andrillae. (A) All OTUs with 2 or more
sequences detected are shown. (B) Heatmap representation of OTUs40.5% relative abundance across the five anemone microbiomes.
More highly represented OTUs are shown in darker colors (max 92%); white represents a value of 0%.

Epsilonproteobacteria, and Gammaproteobacteria Cyanobacteria, Deferribacteres, and Verrucomicrobia

were found in each of the five microbiomes, al- were only detected in the N. vectensis microbiome.
though there was no single OTU common to all The microbiome of Anemonia viridis, which resides
anemones analyzed (Fig. 3). Spearman rank correla- in a different superfamily than E. andrillae and N.
tion-based clustering indicated that the microbiomes vectensis and which hosts photosymbionts, had three
grouped into two pairs of two members each (Ea3 bacterial phyla in common with E. andrillae
and 4; Ea6 and 7) and one (Ea5) that was more (Proteobacteria, Bacteroidetes, and Actinobacteria)
distantly related (Fig. 3). The lack of a common or and shows broad representation of Proteobacteria clas-
‘‘core’’ set of microbial OTUs could be due to ses, similar to what we detected in E. andrillae (Meron
undersampling, as the rarefaction curves (Fig. 4A) et al. 2013). Proteobacteria, Bacteroidetes, and
suggest that the microbiomes were not sampled to Actinobacteria were also detected in the soft coral, A.
saturation; likewise, sample preservation or extrac- antarcticum (Webster and Bourne 2007), with Alpha-,
tion procedures may not have fully retained the orig- Beta-, and Gammaproteobacteria classes in common
inal microbiome diversity. with E. andrillae. Thus at a very coarse level of taxo-
In comparison, 16S rRNA gene sequence composi- nomic resolution, the groups common to all four of
tion of the N. vectensis microbiomes from anemones these Anthozoa microbiome studies include
(pooled) at four sites were characterized by bacteria Alphaproteobacteria, Gammaproteobacteria, and
associated with seven phyla (Har et al. 2015). Four of Bacteroidetes-related bacteria.
these (Proteobacteria, Bacteroidetes, Spirochaetae, and
Tenericutes) also were found in E. andrillae. Microbiome community structure
Lentispherae and sequences affiliated with the The rank abundance distributions varied across the
Betaproteobacteria class of Proteobacteria were unique anemone microbiomes (Fig. 4B), with the most dis-
to E. andrillae, and OTUs associated with Chloroflexi, tinct difference found in microbiome Ea5, which had
Ice-anemone microbiome 7

compounds that influence host selection and specifi-

city, as has been observed in corals (e.g., Mao-Jones
et al. 2010). Another cnidarian, Hydra, has been
shown to resist infection through antifungal produc-
tion by cooperating bacteria associated with the ep-
ithelium (Fraune et al. 2015).
The differences among microbiomes from individ-
ual E. andrillae hosts should be interpreted with cau-
tion, as they could reflect limitations of the

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
sequencing effort that would be reduced upon
deeper sequencing that provides access to less-well
represented sequences. However, the use of short se-
quences is likely to inflate the perception of similar-
ity rather than of difference, as differences may not
be detected in the short fragment used (but may be
present in longer sequences). The unconventional
sample preservation (storage in ethanol at 48C for
several years) might influence the outcome in ways
that would be difficult to predict. Although this is
not standard practice for microbial samples, ethanol
is widely used for biological sample preservation
(e.g., Gaither et al. 2011). Although Anthozoan
microbiome preservation approaches have not been
rigorously evaluated, host-associated bacterial and
mitochondrial DNA can persist for extremely long
periods in ethanol-preserved human tissues
(Spigelman et al. 2001). Alternatively, the observed
Fig. 4 Microbiome sequence distribution across OTUs are rep-
resented in terms of (A) rarefaction plot of observed OTUs for
differences could reflect real differences in the micro-
anemone (Ea) microbiomes and (B) rank abundance. The legend biomes between anemones, an interpretation bol-
applies to both panels. stered by the high degree of similarity within the
two pairs of E. andrillae microbiomes described
here (Ea3 and 4 and Ea6 and 7).
higher numbers of sequences distributed across more
The variable patterns we observed resemble trends
OTUs compared to the other four anemones. The
in other anemone, gorgonian, and coral micro-
others (Ea3, 4, 6, and 7) had a fewer OTUs with
biomes. Direct comparisons of structure are made
high numbers of sequences and many rare members.
difficult by different sampling approaches (e.g., pool-
Differences in structure are also reflected in the
ing individuals for DNA extraction and sequencing;
Bray–Curtis pairwise similarity values among the
Har et al. 2015). Nevertheless, between-site and
five E. andrillae microbiomes: Ea3 and 4, and inter-seasonal variation clearly occur in N. vectensis
Ea6 and 7 have values greater than 0.69, whereas (Har et al. 2015), and some samples were highly
all other comparisons have values below 0.26 skewed in representation of single phylotypes (e.g.,
(Supplementary Fig. S1) These differences are Epsilonproteobacteria) compared to others, which
driven by the strong differences in OTUs across the were much more even in terms of community struc-
microbiomes (Figs. 2B and 3), most strikingly be- ture. The microbial community structure of cold-
tween samples dominated by a single Tenericutes-re- water gorgonians was also similar to E. andrillae
lated OTU (Otu001; Ea6 and 7; 80.3 and 91.5% in that some colonies were dominated by a
relative abundance), those with substantial represen- single Tenericutes phylotype (Gray et al. 2011). The
tation of an Alphaproteobacteria OTU (Otu002; 14– Antarctic soft coral A. antarcticum, has a diverse
53% relative abundance in Ea3, 4, and 5), and those yet perhaps stable and specific microbial community
with abundant Gammaproteobacteria-affiliated dominated by Gammaproteobacteria (Webster and
Acinetobacter sequences (Otu003; 48–60.5% relative Bourne 2007). Initial studies that address global
abundance in Ea3 and 4). The difference in member- trends of coral microbiome structure suggest that
ship between these two clusters could potentially re- habitat and environmental conditions drive compo-
flect different microbiome-produced antimicrobial sition (Pantos et al. 2015; Roder et al. 2015; Zhang
8 A. E. Murray et al.

et al. 2015). Although coral microbiomes vary with al. 2009). This Alphaproteobacterium is likely to
species and environment, meta-analysis has also in- be novel at least at the family level, as it is not
dicated a small, broadly distributed ‘‘core’’ of coral- closely related to any cultivated strains. There
associated bacteria (Ainsworth et al. 2015). were other divergent E. andrillae-associated Alphap
roteobacteria sequences that also did not cluster with
Phylogenetic context for microbial associates of known taxonomic groups; for example, the N. vec-
E. andrillae tensis and A. antarcticum microbiome data sets had
very limited representation in this class. The percep-
We used SSU rRNA gene sequences of OTUs from
tion of novelty is not an artifact of the short se-
the microbiomes of E. andrillae, N. vectensis, the

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
quence length, which would tend to obscure
Antarctic octocoral A. antarcticum, and cloned se-
novelty: because there are fewer positions at which
quences and closely related neighbors from a sequence can vary, there is a lower chance of these
GenBank to contextualize the microbial diversity as- being unique.
sociated with E. andrillae (Fig. 5A and There was only one OTU (OTU003; well-repre-
Supplementary Fig. S2). Our primary objective was sented in Ea3 and 4) affiliated with Gammaproteo-
to determine the relative relationship between OTUs bacteria related to Acinetobacter sp. that was nearly
we recovered from E. andrillae to OTUs from other identical (99.7% sequence identity) between the two
Antarctic (or cold-water) species versus OTUs from anemone microbiomes (Fig. 5C). There were closely
other sea anemones The anemone microbiomes show related sequences associated with Pseudomonas pre-
a high degree of evolutionary novelty across several sent in E. andrillae and N. vectensis (distance: 0.032)
lineages. For example, six of the 10 most abundant and two cultures from A. antarcticum that were clo-
OUT’s in E. andrillae are at most 94% identical to sely related to a different E. andrillae OTU (Fig. 5C).
previously known sequences, based on Blastn Two other genera of Gammaproteobacteria also
searches of Genbank’s nr database. An initial evolu- have closely related sequences in some of these
tionary comparison among the three microbiome microbiomes: there were several highly related
groups indicated considerable distance between all Psychrobacter sp.-related sequences in both E. andril-
pairs: 0.402 between N. vectensis and E. andrillae, lae and A. antarcticum (distance: 0.015–0.026), and
0.349 between N. vectensis and A. antarcticum and sequences affiliated with Endozoicomonas were
0.278 between E. andrillae and A. antarcticum. common to N. vectensis and A. antarcticum (dis-
These coarse comparisons suggest that at the level tance: 0.044; Fig. 5D). Although Endozoicomonas
of the microbial community, habitat may be more appear frequently in anthozoan microbiomes
important than phylogenetic relatedness among (Ainsworth et al. 2015) sequences closely related to
hosts. Endozoicomonas were not detected in our survey of
Most of the sequences that are closely related to the E. andrillae microbiome.
the OTUs from E. andrillae are environmentally de- Phylogenetic novelty was also found in spiro-
rived and from cold-water, host-associated studies. chaete-related sequences, which were recovered
The phylogenetic diversity of Alphaproteobacteria in both E. andrillae and N. vectensis. These were rel-
from E. andrillae microbiome sequences is more ex- atively abundant, as in the case of OTU004, yet
tensive than in N. vectensis and A. antarcticum and only distantly related to cultivated spirochaetes
includes many of the recognized subclasses of this (85% identical to Spirochaeta sp. SR, FJ80060;
group (Rickettsiales, Pelagibacterales, Holosporales, Supplementary Fig. S2). Likewise, although both
Rhodospirillales, Sphingomonadales, Rhizobales, and anemone species harbor Epsilonproteobacteria-affili-
Rhodobacterales; Feria et al. 2013). Given this diver- ated Campylobacteriales related to known sulfur-ox-
sity and the limited length of the sequences ( 350 idizers, the most closely related sequences between
bases), the Alphaproteobacteria were not monophy- the two anemone microbiomes were quite distantly
letic in the complete data set neighbor-joining related to each other (distance: 0.172). The most
tree (Supplementary Fig. S2); thus, all sequences in common OTU, OTU009, detected in four of the
this group were run through a separate neighbor- five E. andrillae microbiomes (Fig. 3) is affiliated
joining analysis (Fig. 5B). Alphaproteobacteria with a commensal Helicobacteraceae-affiliated se-
OTU002, common to four of five E. andrillae micro- quence reported from a marine gastropod
biomes, did not have any related sequences in the N. (Supplementary Fig. S2).
vectensis microbiome but is distantly related to a mi- Tenericutes were common to both anemone
crobial sequence from a cold-water scleractinian microbiomes, though the dominance of this group
coral (92% sequence identity; FJ041553; Kellogg et in the microbiome of E. andrillae was not observed
Ice-anemone microbiome 9

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
Fig. 5. Neighbor-joining phylogenetic tree and subtrees. (A) Sequences include the E. andrillae pyrosequenced data set (blue circles;
representatives from each OTU, not including singletons), the N. vectensis clone library (pink triangles; Har et al. 2015), the A.
antarcticum clone library and cultivated bacteria (green inverted triangles; Webster and Bourne 2007) and near sequence neighbors.
Abbreviations: Pro, Proteobacteria; Unc, unclassified; Eps, Epsilonproteobacteria; Cya, Cyanobacteria; Spi, Spirochaetes; Fir, Firmicutes;
Act, Actinobacteria; Chl, Chloroflexi; Len, Lentisphaera; Ver, Verrucomicrobia. Subtrees highlighting particular aspects of the full tree
demonstrate deep diversity of E. andrillae-associated Alphaproteobacteria (B), close relationships between Anthozoan microbiome
sequences affiliated with Gammaproteobacteria genera Acinetobacter and Pseudomonas (C) and Psychrobacter and Endozoicomonas (D),
and diverse Tenerricutes-related sequences in microbiomes of Anthozoans including E. andrillae and N. vectensis, and octocorals from
cold water habitats. Subtree abbreviations: Ric, Rickettsiales; Pel, Pelagibacterales; Sph, Sphingomonadales; Rhb, Rhodobacterales; Rhs,
Rhodospiralalles; Hol, Holobacterales,? Unclassified sequences. A total of 303 16S rRNA gene sequences were included in the tree; all
ambiguous positions were removed for each sequence pair resulting in a total of 504 positions in the final data set. Numbers at the
nodes indicated percent of 1000 bootstrap resamplings for values449. The neighbor-joining subtree in (B) was recalculated with all
Alphaproteobacteria sequences using the same parameters for the complete data set, while the subtrees in (C-E) were clipped from
the large tree.
10 A. E. Murray et al.

in N. vectensis, and the Tenericutes sequences in each related OTUs (Fig. 5B), (ii) Gammaproteobacteria-
microbiome were only distantly related (75.4% se- related Acinetobacter (Fig. 5C), and (iii) Betaproteo-
quence identity). The nearest neighbor of the bacteria-affiliated Comamonadaceae (one sequence
Tenericutes OTU (OTU001) that dominated the was highly related to Acidovorax sp. and a denitrify-
Ea6 and 7 microbiomes is distantly related to a se- ing strain Comomonas sp. R-25060). The Tenericutes
quence from the microbiome of a cold-water chiton OTUs in E. andrillae also clustered into two groups,
(88% sequence identity; HE663394.1) and to se- with a total of 10 OTUs (Fig. 5E). In all but
quences from other marine invertebrates, including Holospora, the microdiverse sequences were spread
an isolate from the scleractinian coral Lophelia per- across 2–4 specimens; the Holospora sequences were

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
tusa (Fig. 5E; 81% identity, Neulinger et al. 2008). all from Ea4. Further interrogation of the data will
be necessary to tease out the nature of these micro-
The microbiome of E. andrillae: phylogenetic and diverse lineages, and to understand how environ-
functional diversification mental selection is acting on these genomes within
the same population (Shapiro and Polz 2014)
Seawater samples were not collected at the same time and whether they play important roles in host
as our samples of E. andrillae. However, based on the ecology.
biology and patterns of association of the microbial Given the large phylogenetic distances between
lineages we found in our samples, we expect that many of the members of the E. andrillae microbiome
some members of the E. andrillae microbiome may and organisms with well-characterized physiological
be free-living or particle-associated bacterioplankton. attributes, as well as the lack of information about
Of the likely free-living bacterioplankton, the most the biology and physiology of E. andrillae, our un-
noteworthy is OTU007 (detected in Ea3 and 5), derstanding of the functional roles played by the
which is 100% identical to the 16S rRNA gene of members of the E. andrillae microbiome is very lim-
Ant4D3, a marine Gammaproteobacterium that was ited. Nevertheless, there are a few members whose
initially reported from a metagenomic study in phylogenetic affiliation with lineages having known
Antarctic Peninsula waters (Grzymski et al. 2006) functions (across a range of coarse to fine distances)
and that has been shown to be active in the are provocative in light of the ecology and biology of
uptake of amino acids and proteins from Dissolved E. andrillae. The cell-wall free Tenericutes, including
Organic Matter (DOM) (Straza et al. 2010, Nikrad et the genera Mycoplasma, Spiroplasma, and
al. 2014). Other Gammaproteobacteria associated Ureaplasma, are frequently endosymbiotic, parasitic
with E. andrillae that are likely free-living and in and/or pathogenic, and are often associated with
most cases are highly related to well-characterized metazoan guts. Interestingly, this group is relatively
‘‘true’’ psychrophiles from polar marine ecosystems commonly detected in cnidarian microbiomes, in-
(e.g., sequences were affiliated with cluding those of gorgonians (Gray et al. 2011),
Gammaproteobacteria-related Colwellia psychrery- stony corals (Neulinger et al. 2008), and jellyfish
thraea, Psychrobacter, Marinobacter, and (Weiland-Bräuer et al. 2015). Although there is
Flavobacteria-related Polaribacter, Crocinitomix, and shared representation at the phylum level for
Flavobacterium frigidarium) were not detected in Tenericutes between the two anemone microbiomes
the microbiome of N. vectensis. Some of these, e.g., (and with other cnidarians), the phylogenetic dis-
those related to Psychrobacter sp. (Fig. 5D), were de- tance between the Mycoplasma-related N. vectensis
tected in A. antarcticum (Webster and Bourne 2007). sequences, those associated with E. andrillae (dis-
Likewise, several of these microbes are frequently tance ¼ 0.244–0.301), and other cultivated microor-
particle-associated and could have been ingested ganisms is too large to infer robust functional roles.
with marine snow from the water column. Another group of potentially parasitic or patho-
Several lineages represented in the E. andrillae genic microorganisms abundant in the mircrobiomes
microbiome appear to be diversified, having five of three of five of the anemones were the
or more closely related OTUs across the specimens Alphaproteobacteria-affiliated order Holosporales.
we sampled. This level of microheterogeneity at the These microorganisms are frequently found to be
16S rRNA gene level is common in environmental associated with single celled eukaryotes, and the
molecular surveys of both free-living (Thompson most closely related microbe (H. obtusa) to the
et al. 2005) and host-associated (Grzymski et al. OUT’s in E. andrillae (distance of 0.145) is an endo-
2008) systems. The lineages diversified in E. andr symbiont of Paramecium caudatum that has been
illae include several Proteobacteria: (i) Alphaproteo- shown to aid in host thermotolerance (Hori and
bacteria-affiliated Rickettsiales and Holospora obtusa- Fujishima 2003). Though the phylogenetic distance
Ice-anemone microbiome 11

prohibits detailed functional interpretation, these that E. andrillae might create hot spots of produc-
relationships points to three questions for future tivity that have a positive feedback in terms of mi-
study: (i) does the E. andrillae holobiont harbor crobial production that then can provide food
persistent relationships with unicellular eukaryotes; resources for the community. Understanding this
(ii) do members of the microbiome confer assistance in and other potential roles of E. andrillae in the
thermotolerance (in this environment-freezing resis- under-ice-shelf ecosystem, too, will be an important
tance), and (iii) what is the ecology of the host-path- focus of future studies.
ogen relationship in the under-ice-shelf ecosystem, and
what means do the anemones use to select for syner- Conclusions and Future Directions

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
gistic versus pathogenic microbial invasions? The dy-
namic habitat in which E. andrillae occurs is affected Edwardsiella andrillae lives in a newly recognized and
by ice melt and a temporally variable regime of cur- difficult to access ice shelf habitat 260 m under the
rents and food supply that could influence E. andrillae ice surface (Rack et al. 2012). The limited number of
physiology, stress levels, or susceptibility to microbial specimens provides only a glimpse into the biology
invasions. However, the relationship between these fac- of the animal and the ecosystem in which it lives
tors, the anemones and their microbiomes is presently (Daly et al. 2013). Although necessarily limited by
unknown. the material at hand, the present study of the micro-
The microbiomes of sessile marine invertebrates in bial community associated with these anemones
general and anthozoans in particular are believed to broadens our perspective on the ecosystem and
play a role in antifouling and antimicrobial defense hints at a dynamic biological system with high var-
(reviewed by Satheesh et al. 2016). One of the inter- iability among individuals and high novelty in the
esting features of marine Acinetobacter is that some microbial members.
strains produce compounds that assist the host in Our preliminary characterization of the micro-
through antifouling (Olguin-Uribe et al. 1997) and biome of E. andrillae is unusual, not only because
antimicrobial (Graça et al. 2013) activities. The of the host ecology, but also because we have exam-
Pseudomonas-related OTUs, too, may be involved ined the microbiome at the level of the individual
in production of antimicrobial compounds. host. Although pooling samples may provide a more
Understanding the functional roles between representative snapshot of diversity at the level of the
microbiome constituents and the host will be best host population, it obscures potentially important
accomplished using complimentary ’omics differences among individual hosts. Understanding
approaches, if this ecosystem is accessed in the the ecology of the holobiont requires deeper sam-
future. However, there are links between elemental pling across individuals than is found in most stud-
cycling and organisms detected in the microbiomes ies. For example, a minimum of six replicates has
of E. andrillae, N. vectensis, and even A. viridis. For been recommended to infer microbiome ecology
example, Campylobacterales-related Epsilonproteo- (Kvennefors et al. 2010), and the variation among
bacteria in all three anemone microbiomes and our five E. andrillae individuals attests to the impor-
Gammaproteobacteria-related Thiotrichales OTUs in tance of replicate sampling for recovering the
E. andrillae may be involved with sulfur oxidation breadth of microbial diversity associated with a
either through chemolithoautotrophic or mixo- host species. Inter-sample variation, at either the in-
trophic pathways. Har et al. (2015) cultivated strains dividual or population level, and differences among
(Limnobacter thiooxidans and Stappia stellulata) from sequencing approaches and depths, make compara-
N. vectensis that are capable of mixotrophic sulfur tive approaches challenging but important.
oxidation, further suggesting the potential for Although this initial characterization of the E.
sulfur metabolism and alternative energy metabo- andrillae microbiome provides some additional per-
lisms in the holobiont. The exchange of carbon spective on both the anemone and its habitat, we
also may be important between host and micro- find no microbial ‘‘smoking gun’’ capable of explain-
biome, or even between the host and the free-living ing how this sea anemone survives at the ice shelf–
plankton. Based on phylogenetic coherence, the ma- water interface as has been done, e.g., for ciliates
jority of the OTUs in the microbiome have rela- (Pucciarelli et al. 2014). Instead, we have identified
tives with heterotrophic metabolisms. Corals in a suite of microbes that could contribute to the suc-
temperate regions have recently been recognized to cess of the animal: some bacteria that are affiliated
function as suppliers of both organic carbon and with broad, often endosymbiotic lineages that may
nitrogen to the microbial loop (Fonvielle et al. span the range of commensal, parasitic, or patho-
2015). In the under-ice-shelf ecosystem, it is possible genic relationships, and others that may be involved
12 A. E. Murray et al.

in elemental cycling or chemical defense. Variation Bourne DG, Garren M, Work TM, Rosenberg E, Smith GW,
across anemones, the lack of comparative samples Harvell CD. 2009. Microbial disease and the coral holo-
from anemones in other habitats or from other biont. Trends Microbiol 17:554–62.
Cazenave F, Zook R, Carroll D, Flagg M, Kim S. 2011.
members of the ice shelf community, and the high Development of the ROV SCINI and deployment in
degree of novelty of many of the microbial OTUs McMurdo Sound, Antarctica. J Ocean Technol 6:39–58.
limit our ability to make inferences about the func- Daly M, Perissinotto R, Laird M, Dyer D, Todaro A. 2012.
tion and diversity of the microbiome of E. andrillae. Description and ecology of a new species of Edwardsia de
Future ’omics-enabled studies of host and micro- Quatrefages, 1842 Anthozoa, Actiniaria from the St Lucia
biome, in addition to a deeper survey of microbiome Estuary, South Africa. Mar Biol Res 83:233–45.
Daly M, Rack F, Zook R. 2013. Edwardsiella andrillae, a new

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
community structure, will be necessary to provide
species of sea anemone from Antarctic Ice. PLoS ONE
valuable insights into the ecology of E. andrillae
8:e83476.
and other metazoan species in this newly-recognized Daniels CA, Baumgarten S, Yum L, Michell CT, Bayer T, Arif
ice-associated habitat. C, Roder C, Weil E, Voolstra CR. 2015. Metatranscriptome
analysis of the reef-building coral Orbicella faveolata indi-
Supplementary data cates holobiont response to coral disease. Front Mar Sci
2:62.
Supplementary Data available at ICB online. DeLong EF. 1992. Archaea in coastal marine environments.
Proc Natl Acad Sci 89:5685–9.
Acknowledgments Du Z, Zhang W, Xia H, Lu G, Chen G. 2010. Isolation and
diversity analysis of heterotrophic bacteria associated with
Appreciation is extended to Antarctic Support sea anemones. Acta Oceanol Sin 29:62–9.
Contactor and Antarctica New Zealand for support Eisen MB, Spellman PT, Brown PO, Botstein D. 1998. Cluster
of field logistics. We thank Annie Lindgren for the analysis and display of genome-wide expression patterns.
invitation to participate in the symposium at the Proc Natl Acad Sci 95:14863–8.
2016 meeting of SICB on ‘‘Life at the Extremes.’’ Feria M, Thrash J, Giovannoni S, Patrick W. 2013. New rRNA
Conversations with Jason Macrander and Adam gene-based phylogenies of the Alphaproteobacteria provide
perspective on major groups, mitochondrial ancestry and
Reitzel spurred our thinking on these issues. This
phylogenetic instability. PLoS ONE 8:e83383.
article is a contribution to the SCAR AntEco SRP. Flynn JM, Brown EA, Chain FJJ, MacIsaac HJ, Cristescu ME.
2015. Toward accurate molecular identification of species
Funding in complex environmental samples: testing the performance
of sequence filtering and clustering methods. Ecol Evol
This work was supported by the National Science 5:2252–66.
Foundation [1257796 to M.D., 125009 to A.E.M., Fonvielle JA, Reynaud S, Jacquet S, LeBerre B, Ferrier-Pages
0839108 to F.R.] and the New Zealand Ministry of C. 2015. First evidence of an important organic matter
Business Innovation and Employment [Contract trophic pathway between temperate corals and pelagic mi-
C05X1001 to M.J.M.W.]. crobial communities. PLoS ONE 10:e0139175.
Fraune S, Anton-Erxleben F, Augustin R, Franzenburg S,
Knop M, Schroeder K, Willoweit-Ohl D, Bosch TCG.
References 2015. Bacteria-bacteria interactions within the microbiota
Ainsworth TD, Thurber RV, Gates RD. 2010. The future of of the ancestral metazoan Hydra contribute to fungal resis-
coral reefs: a microbial perspective. Trends Ecol Evol tance. ISME J 9:1543–56.
25:233–40. Fraune S, Bosch TCG. 2010. Why bacteria matter in animal
Ainsworth TD, Knack B, Ukani L, Seneca F, Weiss Y, Leggat development and evolution. Bioessays 32:571–80.
W. 2015. In situ hybridisation detects pro-apoptotic gene Gaither MR, Szabó Z, Crepeau MW, Bird CE, Toonen RJ.
expression of a Bcl-2 family member in white syndrome- 2011. Preservation of corals in salt-saturated DMSO
affected coral. Dis Aquat Organ 117:155–63. buffer is superior to ethanol for PCR experiments. Coral
Arrigo KR, Dunbar RB, Lizotte MP, Robinson DH. 2002. Reefs 30:329–33.
Taxon-specific differences in C/P and N/P drawdown for Gast RJ, Dennett MR, Caron DA. 2004. Characterization of
phytoplankton in the Ross Sea, Antarctica. Geophys Res protistan assemblages in the Ross Sea, Antarctica by dena-
Lett 29:441–4. turing gradient gel electrophoresis. Appl Environ Microbiol
Arzeno IB, Beardsley RC, Limeburner R, Owens B, Padman L, 70:2028–37.
Springer SR, Stewart CL, Williams MJM. 2014. Ocean var- Gilbert JA, Field D, Swift P, Newbold L, Oliver A, Smyth T,
iability contributing to basal melt rate near the ice front of Somerfield PJ, Huse S, Joint I. 2009. The seasonal structure
Ross Ice Shelf, Antarctica. J Geophys Res-Oceans 119:4214– of microbial communities in the Western English Channel.
33. Environ Microbiol 11:3132–9.
Boetius A, Anesio AM, Deming JW, Mikucki JA, Rapp JZ. Graça AP, Bondoso J, Gaspar H, Xavier JR, Monteiro MC, de
2015. Microbial ecology of the cryosphere: sea ice and gla- la Cruz M, Oves-Costales D, Vicente F, Lage OM. 2013.
cial habitats. Nat Rev Microbiol 13:677–90. Antimicrobial activity of heterotrophic bacterial
Ice-anemone microbiome 13

communities from the marine sponge Erylus discophorus Lewis LA, Müller-Parker G. 2004. Phylogenetic placement of
(Astrophorida, Geodiidae). PLoS ONE 8:e78992. ‘‘zoochlorellae’’(Chlorophyta), algal symbiont of the temperate
Gray MA, Stone RP, McLaughlin MR, Kellogg CA. 2011. sea anemone Anthopleura elegantissima. Biol Bull 20:87–92.
Microbial consortia of gorgonian corals from the Aleutian Lokmer A, Wegner KM. 2015. Hemolymph microbiome of
Islands. FEMS Microbiol Ecol 76:109–20. Pacific oysters in response to temperature, temperature
Grzymski JJ, Carter BJ, DeLong EF, Feldman RA, Ghadiri A, stress and infection. ISME J 9:670–82.
Murray AE. 2006. Comparative genomics of DNA frag- Mao-Jones J, Ritchie KB, Jones LE, Ellner SP. 2010. How
ments from six Antarctic marine planktonic bacteria. App microbial community compotision regulates coral disease
Environ Microbiol 72:1532–41. development. PLoS Biol 8:e1000345.
Grzymski JJ, Murray AE, Campbell BJ, Kaplarevic M, Gao Massana R, Murray AE, Preston CM, DeLong EF. 1997.
GR, Lee C, Daniel R, Ghadiri A, Feldman RA, Cary SC. Vertical distribution and phylogenetic characterization of

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
2008. Metagenome analysis of an extreme microbial sym- marine planktonic Archaea in the Santa Barbara Channel.
biosis reveals eurythermal adaptation and metabolic flexi- Appl Environ Microbiol 63:50–6.
bility. Proc Natl Acad Sci 105:17516–21. McFall-Ngai M, Hadfield MG, Bosch TCG, Carey HV,
Hand C, Uhlinger KR. 1994. The unique, widely distributed, Domazet-Loso T, Douglas AE, Dubilier N, Eberl G,
estuarine sea anemone, Nematostella vectensis Stephenson: a Fukami T, Gilbert SF, et al. 2013. Animals in a bacterial
review, new facts, and questions. Estuaries 17:501–8. world, a new imperative for the life sciences. Proc Natl
Har JY, Helbig T, Lim JH, Fernando SC, Reitzel AM, Penn K, Acad Sci 110:3229–36.
Thompson JR. 2015. Microbial diversity and activity in the Meron D, Buia MC, Fine M, Banin E. 2013. Changes in mi-
Nematostella vectensis holobiont: insights from 16S rRNA crobial communities associated with the sea anemone
gene sequencing, isolate genomes, and a pilot-scale survey Anemonia viridis in a natural pH gradient. Microbial Ecol
of gene expression. Front Microbiol 6:818. 65:269–76.
Hoffmann F, Radax R, Woebken D, Holtappels M, Lavik G, Müller-Parker G, Davy SK. 2001. Temperate and tropical
Rapp HT, Schlappy ML, Schleper C, Kuypers MMM. 2009. algal-sea anemone symbioses. Invert Biol 120:104–23.
Complex nitrogen cycling in the sponge Geodia barretti. Neulinger SC, Järnegren J, Ludvigsen M, Lochte K, Dullo
Environ Microbiol 11:2228–43. WC. 2008. Phenotype-specific bacterial communities in
Hori M, Fujishima M. 2003. The endosymbiotic bacterium the cold-water coral Lophelia pertusa (Scleractinia) and
Holospora obtusa enhances heat-shock gene expression of their implications for the coral’s nutrition, health, and dis-
the host Paramecium caudatum. J Eukaryot Microbiol tribution. Appl Environ Microbiol 74:7272–85.
50:293–8. Nikrad MP, Cottrell MT, Kirchman DL. 2014. Growth activity
Jacobs SS, Giulivi CF, Mele PA. 2002. Freshening of the Ross of gammaproteobacterial subgroups in waters off the west
Sea during the late 20th century. Science 297:386–9. Antarctic Peninsula in summer and fall. Environ Microbiol
Kellogg CA, Lisle JT, Galkiewicz JP. 2009. Culture-indepen- 16:1513–23.
dent characterization of bacterial communities associated O’Connor RM, Fung JM, Sharp KH, Benner JS, McClung C,
with the cold-water coral Lophelia pertusa in the northeast- Cushing S, Lamkin ER, Fomenkov AI, Henrissat B, Londer
ern Gulf of Mexico. Appl Environ Microbiol 75:2294–303. YY, et al. 2014. Gill bacteria enable a novel digestive strat-
Kelly LW, Williams GJ, Barott KL, Carlson CA, Dinsdale EA, egy in a wood-feeding mollusk. Proc Natl Acad Sci
Edwards RA, Haas AF, Haynes M, Lim YW, McDole T, 111:E5096–104.
et al. 2014. Local genomic adaptation of coral reef-associ- Olguin-Uribe G, Abou-Mansour E, Boulander A, Debard H,
ated microbiomes to gradients of natural variability and Francisco C, Combaut G. 1997. 6-Bromoindole-3-carbalde-
anthropogenic stressors. Proc Natl Acad Sci 111:10227–32. hyde, from an Acinetobacter sp. bacterium associated with
Krediet CJ, Ritchie KB, Paul VJ, Teplitski M. 2013. Coral- the ascidian Stomozoa murrayi. J Chem Ecol 23:2507–21.
associated micro-organisms and their roles in promoting Pantos O, Bongaerts P, Dennis PG, Tyson GW, Hoegh-
coral health and thwarting diseases. P Roy Soc Lond B Guldberg O. 2015. Habitat-specific environmental condi-
Bio 280:20122328. tions primarily control the microbiomes of the coral
Kunin V, Engelbrektson A, Ochman H, Hugenholtz P. 2010. Seriatopora hystrix. ISME J 9:1916–27.
Wrinkles in the rare biosphere: pyrosequencing errors can Pruesse E, Peplies J, Glöckner FO. 2012. SINA: Accurate high-
lead to artificial inflation of diversity estimates. Environ throughput multiple sequence alignment of ribosomal RNA
Microbiol 12:118–23. genes. Bioinformatics 28:1823–9.
Kvennefors ECE, Sampayo E, Ridgway T, Barnes AC, Pucciarelli S, Chiappori F, Devaraj RR, Yang G, Yu T,
Hoegh-Guldberg O. 2010. Bacterial communities of two Ballarini P, Miceli C. 2014. Identification and analysis of
ubiquitous great barrier reef corals reveals both site- and two sequences encoding ice-binding proteins obtained from
species-specificity of common bacterial associates. PLoS a putative bacterial symbiont of the psychrophilic Antarctic
ONE 5:e10401. ciliate Euplotes focardii. Antarctic Sci 26:491–501.
LaJeunesse TC, Trench RK. 2000. Biogeography of two species Quince C, Lanzen A, Curtis T, Davenport R, Hall N, Head I,
of Symbiodinium Freudenthal inhabiting the intertidal sea Read L, Sloan W. 2009. Accurate determination of micro-
anemone Anthopleura elegantissima Brandt. Biol Bull bial diversity from 454 pyrosequencing data. Nat Methods
199:126–34. 6:639–41.
Lane DJ. 1991. 16S/23S rRNA Sequencing. In: Stackebrandt E, Quince C, Lanzen A, Davenport RJ, Turnbaugh PJ. 2011.
Goodfellow M, editors. Nucleic acid techniques in bacterial Removing noise from pyrosequenced amplicons. BMC
systematics. New York, NY: Wiley, p. 115–75. Bioinformatics 12:28.
14 A. E. Murray et al.

Rack FR, Zook R, Levy RH, Limeburner R, Stewart C, Williams Schuett C, Doepke H. 2010. Endobiotic bacteria and their
MJM, Luyendyk B. and the ANDRILL Coulman High Project pathogenic potential in cnidarian tentacles. Helgol Mar
Site Survey Team 2012. What lies beneath? Interdisciplinary Res 64:205–12.
outcomes of the ANDRILL Coulman High Project site surveys Sebens KP. 1981. The allometry of feeding, energetics, and
on the Ross Ice Shelf. Oceanography 25:84–9. body size in three sea anemone species. Biol Bull 161:152–71.
Riesenfeld CS, Murray AE, Baker BJ. 2008. Characterization of Shapiro B, Polz M. 2014. Ordering microbial diversity into
the microbial community and polyketide biosynthetic ecologically and genetically cohesive units. Trends
potential in the palmerolide-producing tunicate Synoicum Microbiol 22:235–47.
adareanum. J Nat Prod 71:1812–8. Shick JM. 1991. A functional biology of sea anemones. New
Ritchie KB. 2006. Regulation of microbial populations by York, NY: Springer Science and Business Media.
coral surface mucus and mucus-associated bacteria. Mar Spigelman M, Barnes I, Holton J, Vaira D, Thomas MG.

Downloaded from http://icb.oxfordjournals.org/ at Periodicals Dept University Libraries Northern Illinois University on August 23, 2016
Ecol Prog Ser 322:1–14. 2001. Long-term DNA survival in ethanol-preserved archi-
Roder C, Bayer T, Aranda M, Kruse M, Voolstra CR. 2015. val material. Ann R Coll Surg Engl 83:283.
Microbiome structure of the fungid coral Ctenactis echinata Straza TRA, Ducklow HW, Murray AE, Kirchman DL. 2010.
aligns with environmental differences. Mol Ecol 24:3501– Abundance and single-cell activity of bacterial groups in
11. Antarctic coastal waters. Limnol Oceanog 55:2526–36.
Rodriguez-Marconi S, De la Iglesia R, Diez B, Fonseca CA, Sunagawa S, Coelho LP, Chaffron S, Kultima JR, Labadie K,
Hajdu E, Trefault N. 2015. Characterization of bacterial, Salazar G, Djahanschiri B, Zeller G, Mende DR, Alberti A,
archaeal and eukaryote symbionts from Antarctic sponges et al. 2015. Structure and function of the global ocean
reveals a high diversity at a three-domain level and a par- microbiome. Science 348:1261359.
ticular signature for this ecosystem. PLoS ONE Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013.
10:e0138837. MEGA6: Molecular Evolutionary Genetics Analysis version
Rosenberg E, Koren O, Reshef L, Efrony R, Zilber-Rosenberg 6.0. Mol Biol Evol 30:2725–9.
I. 2007. The role of microorganisms in coral health, disease Thompson JR, Pacocha S, Pharino C, Klepac-Ceraj V, Hunt
and evolution. Nature Rev Microbiol 5:355–62. DE, Benoit J, Sarma-Rupavtarm R, Distel DL, Polz MF.
Roeselers G, Newton ILG. 2012. On the evolutionary ecology 2005. Genotypic diversity within a natural coastal bacterio-
of symbioses between chemosynthetic bacteria and bivalves. plankton population. Science 307:1311–3.
Appl Microbiol Biotech 94:1–10. Webster NS, Bourne D. 2007. Bacterial community structure
Rusch DB, Halpern AL, Sutton G, Heidelberg KB, Williamson associated with the Antarctic soft coral, Alcyonium antarc-
S, Yooseph S, Wu DY, Eisen JA, Hoffman JM, Remington ticum. FEMS Microbiol Ecol 59:81–94.
K, et al. 2007. The Sorcerer II Global Ocean Sampling ex- Webster NS, Negri AP, Munro M, Battershill CN. 2004.
pedition: Northwest Atlantic through Eastern Tropical Diverse microbial communities inhabit Antarctic sponges.
Pacific. PLoS Biol 5:398–431. Environ Microbiol 6:288–300.
Satheesh S, Ba-akdah MA, Al-Sofyani AA. 2016. Natural an- Weiland-Bräuer N, Neulinger SC, Pinnow N, Künzel S, Baines
tifouling compound production by microbes associated JF, Schmitz RA. 2015. Composition of Bacterial communi-
with marine macroorganisms: A review. Elect J Biotechnol ties associated with Aurelia aurita changes with compart-
(doi:10.1016/j.ejbt.2016.02.002). ment, life stage, and population. Appl Environ Microbiol
Schloss PD, Gevers D, Westcott SL. 2011. Reducing the effects 81:6038–52.
of PCR amplification and sequencing artifacts on 16S Williams GP, Babu S, Ravikumar S, Kathiresan K, Arul
rRNA-based studies. PLoS ONE 6:e27310. Prathap S, Shinnapparaj S, Marian MP, Likath Alikhan S.
Schönknecht G, Chen WH, Ternes CM, Barbier GG, Shrestha 2007. Antimicrobial activity of tissue and associated bacte-
RP, Stanke M, Brautigam A, Baker BJ, Banfield JF, Garavito ria from benthic sea anemone Stichodactyla haddoni against
RM, et al. 2013. Gene transfer from Bacteria and Archaea microbial pathogens. J Environ Biol 28:789–93.
facilitated evolution of an extremophilic Eukaryote. Science Zhang YY, Ling J, Yang QS, Wang YS, Sun CC, Sun HY, Feng
339:1207–10. JB, Jiang YF, Zhang YZ, Wu ML, et al. 2015. The diversity
Schuett C, Doepke H, Grathoff A, Gedde M. 2007. Bacterial of coral associated bacteria and the environmental factors
aggregates in the tentacles of the sea anemone Metridium affect their community variation. Ecotoxicology
senile. Helgol Mar Res 61:211–6. 24:1467–77.