Spatial scales of bacterial diversity in cold-water coral reef ecosystems.

Abstract

Cold-water coral reef ecosystems are recognized as biodiversity hotspots in the deep sea, but insights into their associated bacterial communities are still limited. Deciphering principle patterns of bacterial community variation over multiple spatial scales may however prove critical for a better understanding of factors contributing to cold-water coral reef stability and functioning.
Bacterial community structure, as determined by Automated Ribosomal Intergenic Spacer Analysis (ARISA), was investigated with respect to (i) microbial habitat type and (ii) coral species and color, as well as the three spatial components (iii) geomorphologic reef zoning, (iv) reef boundary, and (v) reef location. Communities revealed fundamental differences between coral-generated (branch surface, mucus) and ambient microbial habitats (seawater, sediments). This habitat specificity appeared pivotal for determining bacterial community shifts over all other study levels investigated. Coral-derived surfaces showed species-specific patterns, differing significantly between Lophelia pertusa and Madrepora oculata, but not between L. pertusa color types. Within the reef center, no community distinction corresponded to geomorphologic reef zoning for both coral-generated and ambient microbial habitats. Beyond the reef center, however, bacterial communities varied considerably from local to regional scales, with marked shifts toward the reef periphery as well as between different in- and offshore reef sites, suggesting significant biogeographic imprinting but weak microbe-host specificity.
This study presents the first multi-scale survey of bacterial diversity in cold-water coral reefs, spanning a total of five observational levels including three spatial scales. It demonstrates that bacterial communities in cold-water coral reefs are structured by multiple factors acting at different spatial scales, which has fundamental implications for the monitoring of microbial diversity and function in those ecosystems.

Full text

Page 1

Spatial Scales of Bacterial Diversity in Cold-Water Coral
Reef Ecosystems
Sandra Scho ¨ttner1¤, Christian Wild2, Friederike Hoffmann3,4, Antje Boetius1, Alban Ramette1*
1HGF-MPG Joint Research Group on Deep-Sea Ecology and Technology, Max Planck Institute for Marine Microbiology, Bremen, Germany, 2Coral Reef Ecology Group,
Leibniz Center for Tropical Marine Ecology, Bremen, Germany, 3Center for Geobiology, University of Bergen, Bergen, Norway, 4Uni Environment, Uni Research AS,
Bergen, Norway
Abstract
Background: Cold-water coral reef ecosystems are recognized as biodiversity hotspots in the deep sea, but insights into
their associated bacterial communities are still limited. Deciphering principle patterns of bacterial community variation over
multiple spatial scales may however prove critical for a better understanding of factors contributing to cold-water coral reef
stability and functioning.
Methodology/Principal Findings: Bacterial community structure, as determined by Automated Ribosomal Intergenic
Spacer Analysis (ARISA), was investigated with respect to (i) microbial habitat type and (ii) coral species and color, as well as
the three spatial components (iii) geomorphologic reef zoning, (iv) reef boundary, and (v) reef location. Communities
revealed fundamental differences between coral-generated (branch surface, mucus) and ambient microbial habitats
(seawater, sediments). This habitat specificity appeared pivotal for determining bacterial community shifts over all other
study levels investigated. Coral-derived surfaces showed species-specific patterns, differing significantly between Lophelia
pertusa and Madrepora oculata, but not between L. pertusa color types. Within the reef center, no community distinction
corresponded to geomorphologic reef zoning for both coral-generated and ambient microbial habitats. Beyond the reef
center, however, bacterial communities varied considerably from local to regional scales, with marked shifts toward the reef
periphery as well as between different in- and offshore reef sites, suggesting significant biogeographic imprinting but weak
microbe-host specificity.
Conclusions/Significance: This study presents the first multi-scale survey of bacterial diversity in cold-water coral reefs,
spanning a total of five observational levels including three spatial scales. It demonstrates that bacterial communities in
cold-water coral reefs are structured by multiple factors acting at different spatial scales, which has fundamental
implications for the monitoring of microbial diversity and function in those ecosystems.
Citation: Scho ¨ttner S, Wild C, Hoffmann F, Boetius A, Ramette A (2012) Spatial Scales of Bacterial Diversity in Cold-Water Coral Reef Ecosystems. PLoS ONE 7(3):
e32093. doi:10.1371/journal.pone.0032093
Editor: John Murray Roberts, Heriot-Watt University, United Kingdom
Received November 1, 2011; Accepted January 23, 2012; Published March 5, 2012
Copyright: ? 2012 Scho ¨ttner et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by the German Research Foundation (DFG) (grant Wi 2677/3-1) to CW within the EuroDIVERSITY project MiCROSYSTEMS, by the
German Research Foundation to FH, and by the EU 6th FP HERMES and the EU 7th FP HERMIONE (grant agreement nu 226354) to AB. Further support was
received from the Max Planck Society and the Helmholtz Association. The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: aramette@mpi-bremen.de
¤ Current address: Marine Microbiology Group, University of Bergen, Bergen, Norway
Introduction
Cold-water coral (CWC) reef ecosystems are increasingly
portrayed as biodiversity hotspots on continental margins,
seamounts and mid-ocean ridges around the world [1]. They
appear as speciose, abundant and widespread as their warm-water
counterparts [2–5], and represent important species pools [6–8]
and speciation centers [9] in the deep sea. Their potential to foster
a high degree of local diversity and biomass is assumed to be
rooted in the ecosystem engineering capacity of scleractinian
corals [10,11], such as the cosmopolitan key species Lophelia pertusa
(L. 1758, Caryophylliidae) and Madrepora oculata (L. 1758,
Oculinidae). By forming enormous dendritic skeletal frameworks,
these corals provide complex three-dimensional living space for a
plethora of mobile and sessile organisms [7,8,12]. They also alter
flow regimes and sedimentation rates, thereby modifying the
abiotic environment in time and space ([1] and references therein).
Often, structural complexity in CWC reefs is promoted by
pronounced ecosystem heterogeneity. Unlike warm-water coral
ecosystems that constitute relatively contiguous reef environments
with clear wave and sun energy-related zoning [13], CWC
ecosystems can consist of isolated colonies, small patch accumu-
lations, large reefs, or giant carbonate mounds, and differ
substantially with respect to their spatial configuration [14–19].
Often, individual clusters of coral frameworks form entire reef
complexes which, depending on local seabed geology as well as
community history (i.e. the combined effects of past community
assembly, succession and interaction, including individual life
trajectories and trade-offs), exhibit distinctive geomorphologic and
taphonomic (i.e. seabed form- and fossilization-related) zoning,
PLoS ONE | www.plosone.org1 March 2012 | Volume 7 | Issue 3 | e32093

Page 2

with marked transitions in sediment type, faunal composition and
proliferation stage [7,19–21].
Despite mounting evidence of warm-water coral reefs as
structured landscapes of complex microbial communities [22–
24], insights into the microbial diversity of CWC ecosystems are
limited. Deciphering principle patterns of microbial community
variation may, however, prove critical for a better understanding
of factors contributing to CWC reef stability and functioning.
Especially bacteria play important ecological roles for corals and
entire reef systems by contributing substantially to biogeochemical
processes, invertebrate life cycles, host metabolism, protection and
adaptation, as well as to overall species diversity (e.g. [24] and
references therein). Hence, their spatial and temporal dynamics
are relevant features for the functioning of coral ecosystems.
So far, studies of CWC reef microbiology mainly focused on the
identification of bacteria associated with scleractinians [25–32] or
octocorals [33–37]. Community fingerprinting methods (e.g.
Automated Ribosomal Intergenic Spacer Analysis, ARISA;
Terminal Restriction Fragment Length Polymorphism, T-RFLP;
Denaturing Gradient Gel Electrophoresis, DGGE) and 16 S
rRNA gene sequencing were used to show that bacterial
assemblages colonizing living scleractinians and octocorals differ
from those of dead corals or from those of the ambient
environment like seawater or sediments [25,26,28,29,32,33]. Even
distinct coral-generated microbial habitats, such as branch surface,
mucus and tissue, were found to exhibit specific bacterial
community signatures [26,29,32]. Spatial patterns of bacterial
communities associated with different octocoral species appeared
either highly distinct [37] or conserved between different reef sites
(and across an environmental impact gradient [36]).
On L. pertusa, several bacterial sequences were shared across
geographically separate regions, such as the Gulf of Mexico and
the Trondheimsfjord in Norway [28,30]. Strict host-specificity of
L.pertusa-associated bacteria was so far not evidenced due to
significant community variations between (i) sampling locations
within the same geographic area or reef complex [26,28,29,30], (ii)
colonies of the same coral species [29], (iii) single polyps within the
same coral colony [29], and (iv) differently colored types within the
same coral species [28]. In fact, current evidence suggests that
coral-bacteria associations considerably differ with both coral-
derived microbial habitats and prevailing environmental condi-
tions. However, due to the variations in spatial scale and
methodology applied in aforementioned surveys, the relative
importance of spatial and reef-organizational factors that deter-
mine bacterial biogeography across various scales is not evident.
The aim of the present study was to identify patterns of bacterial
communities in CWC reefs (Fig. 1) using a multi-scale, hierarchical
sampling approach spanning five study levels including three spatial
scales(Fig.2). Sources of bacterialcommunityvariation wereassessed
from local (intra-reef) to regional (inter-reef) scale by considering (i)
microbial habitat type on and around coral colonies (coral branch
surface,coralmucus,ambientseawater,proximalsediments),(ii)coral
species (L. pertusa, M. oculata) and color phenotype (white, red), (iii)
geomorphologic reef zoning (ridge crest, slope, depression), (iv) reef
boundary (up-slope reef center, down-slope reef periphery), and (v)
reef site (Røst Reef, Trænadjupet Reef, Tisler Reef, Langenuen
Fjord) and proximity to shore (offshore, inshore).
For this purpose, bacterial community DNA derived from two
constructional corals L. pertusa and M. oculata, as well as from
associated seawater and surface sediments, was collected from four
CWC reef ecosystems on the Norwegian continental margin
(Fig. 1). The community fingerprinting approach ARISA then
allowed for a time- and cost-effective analysis of the large,
heterogeneous sample set. Despite the lack of information on
OTU identity, ARISA was chosen for its proven ability to provide
robust insights about bacterial community dynamics at different
spatial (and temporal) scales (e.g. [38,39]). Ultimately, the different
but not mutually exclusive sources of bacterial community
variation were disentangled by quantifying their respective effects
with multivariate statistics.
Results
Variation in bacterial OTU number and occurrence
From a pool of 440 different operational taxonomic units
(OTUs) occurring in the whole data set (104 samples), between 9–
Figure 1. Reef sites and corals targeted in this study. (A) Offshore and CWC ecosystems along the Norwegian continental margin. (B) Living
colonies of Lophelia pertusa and Madrepora oculata in their natural environment at Røst, northern mid-Norwegian continental margin. (C) Fragments
of freshly sampled white L. pertusa (left), red L. pertusa (middle), and red M. oculata (right).
doi:10.1371/journal.pone.0032093.g001
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org2 March 2012 | Volume 7 | Issue 3 | e32093

Page 3

223 OTUs were obtained per sample. OTU number was strongly
related to microbial habitat type (KW, P,0.001), and to the reef
site (P=0.05; Fig. 3). The most pronounced difference occurred
between coral-generated surfaces and the ambient environment,
with branch (34615 OTUs) and mucus (58622 OTUs) featuring
30–80% lower mean OTU numbers than water (135616 OTUs)
and sediments (192619 OTUs). At Røst, mean OTU numbers
showed a slight increase (P=0.0398; Fig. 3) between reef center
(Røst-in, 86611 OTUs) and reef periphery (Røst-out, 116662
OTUs), which was mainly related to branch and mucus variability.
When studying local trends at Røst-in and Røst-out separately,
however, neither geomorphologic reef zoning (P=0.098), nor
gradual distance away from the apparent reef margin (P=0.956)
resulted in any significant variation of OTU numbers. These were
also not significantly related to coral species (L. pertusa and M.
oculata with 36613 and 35617 OTUs, respectively; P.0.05) or to
coral color (white and red colonies with 34614 and 38614 OTUs,
respectively; P.0.05).
OTU partitioning among the four different microbial habitats
showed that, from a total of 380–390 different OTUs detected in
coral- water-, and sediment samples at Røst-in, only very few were
exclusively present on L. pertusa and M. oculata surfaces of a given
color phenotype (branch: ,1–2%, mucus: 1%; Fig. S2A). Even
with all coral samples from Røst-in combined, branch and mucus
each contained only about 2% unique OTUs, from a total of 396
different OTUs (Fig. S2B). By contrast, the surrounding water and
proximal sediments at Røst-in contained 4–5% and 19–28%
unique OTUs, respectively, and therefore more of the microbial
habitat-specific bacterial signatures (Fig. S2B). These patterns were
confirmed for all study sites, with only minor variations (data not
shown). OTU partitioning with the whole data set, i.e. with
altogether 440 different OTUs from Røst, Trænadjupet, Tisler,
and Langenuen combined (Fig. S2B), revealed clearly lower
fractions of water- and sediment-specific OTUs (,1% and 5%,
respectively), while those of branch- and mucus-specific OTUs
remained virtually unchanged (branch: 2–3%, mucus: 1–2%).
Concomitantly, the fraction of shared OTUs increased consider-
ably from local (Røst-in: 4–55 and 9%, Fig. S2A) to regional scale
(all sites: 27%; Fig. S2B), indicating a decrease in habitat-
specificity for water- and sediment-associated OTUs, due to their
partial presence in branch and mucus samples from other sites.
The detailed analyses of OTU overlap between samples
confirmed those habitat-specific trends and distinguished partic-
ularly coral-derived surfaces from the surrounding environment
(Fig. 4). Differences in OTU overlap clearly reflected variations in
OTU number, with OTU-poor habitats (branch, mucus) sharing a
much higher percentage of their OTU pool with OTU-rich
habitats (water, sediments) than reciprocally. Branch and mucus
shared at least half of their OTUs with sediments (50% and 73%,
respectively), and a comparatively lower fraction with water (25%
and 36%, respectively). Conversely, only 9–10% and 17–18% of
all water and sediment OTU were found among branch and
mucus OTUs, respectively. The number of OTUs shared solely
between both coral-associated habitats amounted to a third of
their respective OTU content (33–34%), whereas the water shared
a much higher fraction of its OTU pool with the sediments (74%)
than vice versa (34%). Between different reef sites (within each
habitat separately; Fig. 4), the mean number of OTUs overlapping
Figure 2. Multi-scale, hierarchical sampling design. Nested frames indicate the different levels of observation, increasing in scale from inside
(left) to outside (right). Boxes within frames symbolize each lower level as integral part of respective next higher level. At the main study site, Røst,
sampling was implemented on all levels (continuous line); at all other sites, it was performed only on the lowest and highest level, respectively
(dotted line). The following levels and scales of observation were considered: 1) HABITAT (mm–cm): coral branch surface (‘‘b’’), coral mucus (‘‘m’’),
ambient seawater (‘‘w’’), proximal sediments (‘‘s’’); 2) SPECIES (cm–m) – white Lophelia pertusa (‘‘wL’’), red Lophelia pertusa (‘‘rL’’), red Madrepora
oculata (‘‘rM’’); 3) ZONE (1 m–10 m): ridge top with coral terraces (‘‘crest’’), ridge slope with single coral colonies on rubble (‘‘slope’’), ridge depression
with single colonies on clay (‘‘valley’’); 4) IN-OUT (1 m–10 m–100 m): reef center (‘‘reef-in’’), reef periphery in distances of 1, 10, 100 m away from the
apparent reef margin (‘‘reef-out’’); 5) REEF (km): Røst Reef (‘‘Røst’’), Trænadjupet Reef (‘‘Trænadjupet’’), Tisler Reef (‘‘Tisler’’), Langenuen Fjord
(‘‘Langenuen’’).
doi:10.1371/journal.pone.0032093.g002
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org3 March 2012 | Volume 7 | Issue 3 | e32093

Page 4

between any two reefs was the lowest in mucus (mean: 37%) and
also the most variable (range: 11–94%). In contrast, reef-specific
OTU fractions were usually the highest in sediments (76%), with
relatively high and constant OTU overlap among reefs (63–90%).
Overall, no difference between offshore (Røst, Trænadjupet) and
inshore (Tisler, Langenuen) reef sites was detected (Fig. 4).
Differences in community structure
Bacterial community structure strongly differed between
microbial habitat types, with the greatest difference between
coral-associated surfaces and ambient environment (Fig. 5, Fig. S3)
as confirmed by PERMANOVA (Table 1). At Røst-in, as well as
for all study sites combined, distinct community patterns were
associated with branch, mucus, water or sediments. For the other
reef sites (Table 1), the microbial habitat-specific separation was
also similarly pronounced (PERMANOVA R2=0.75–0.80). In
addition, bacterial community structure associated with each of
the four reef systems clearly differed from each other, especially
between offshore (Røst, Trænadjupet) and inshore (Tisler,
Langenuen) sites (Table 1). These regional, inter-reef differences
appeared particularly pronounced for mucus communities (Fig. 5,
Fig. S4), but also for water and sediment communities.
At Røst, significant intra-reef differences were detected between
the up-slope reef center (Røst-in) and the down-slope reef
periphery (Røst-out) (Table 2). Similarly to the aforementioned
regional patterns, this local pattern was mainly evidenced in
mucus, water, marginally in sediments, but not in branch (Fig. S4).
While those intra-reef differences in the Røst area indicated some
degree of separation, the pronounced geomorphologic (including
vertical) zoning at the reef center itself (Røst-in) revealed only a
very weak trend among bacterial communities (Table 3, Fig. 3,
Fig. S4). Communities did not change significantly over distances
of 1, 10, and 100 m in the reef periphery (Røst-out; P.0.05).
L. pertusa and M. oculata harbored significantly different bacterial
assemblages (Table 3) despite some overlap (Fig. S4), and
exhibited slight differences in their response to local reef
complexity (Table S2A, C, B): While M. oculata-associated seemed
to reflect changes in in-/out-reef location and geomorphologic reef
zoning, L. pertusa-associated bacteria appeared more stable over
space. In contrast, bacterial community variation related to coral
color was, albeit overall significant, never supported, neither on
branch nor in mucus (Table 3). All described results were also
generally confirmed by ANOSIM and cluster analysis (data not
shown).
Discussion
Microbial habitat type
Bacterial communities associated with the CWCs L. pertusa and
M. oculata substantially differed according to the type of microbial
habitat sampled. Coral branch surface, coral mucus, ambient
seawater and proximal sediments each featured a specific
community structure that significantly varied both in OTU
composition and relative abundance. This habitat specificity
seemed valid at all study sites, and confirmed earlier findings
based on samples from one reef location [32]. Other studies have
already reported evidence for bacterial habitat specificity in CWC
reefs from the North-East Atlantic [26,28,29], the Central
Mediterranean [25] and the Gulf of Alaska [33], or warm-water
coral reefs (e.g. [40,41]). The most pronounced differences
concerned the distinction between bacterial communities associ-
ated with coral-generated surfaces and the ambient environment,
in both OTU number and composition.
Not surprisingly, the difference in bacterial OTU number
between the OTU-poor coral-associated and OTU-rich ambient
microbial habitats strongly determined the overall degree of
bacterial community partitioning and overlap: Irrespective of reef
site and local zoning, branch and mucus exhibited notably few
specific OTUs, as most of their respective OTU pool was shared
with water and sediments. As expected, sediments generally
Figure 3. Number of ARISA-derived OTUs in distinct microbial habitats at each reef site. Top, middle, and bottom lines of the boxes
represent the 25th, 50th (median), and 75th percentiles, respectively, while the end of the whiskers represent the 5th and 95th percentiles,
respectively; box height and symmetry around the median indicate the degree of dispersion and skewness in the data, respectively; outliers above
and below the whiskers denote extreme values.
doi:10.1371/journal.pone.0032093.g003
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org4March 2012 | Volume 7 | Issue 3 | e32093

Page 5

exhibited the highest OTU abundance and number of specific
OTUs, which was also previously observed in other ARISA-based
studies on warm-water coral reefs (e.g. [38]).
Albeit strongly reduced in overall bacterial OTU number and
specificity, bacterial communities on coral surfaces were charac-
terized by high inter-sample, intra-habitat differences, clearly
exceeding those of OTU-rich water and sediment communities.
This may result from both stochastic events during community
assembly, such as the random attachment of environmental
bacteria on coral surfaces [42], and deterministic processes, such
as the selection of few opportunists through the coral host (e.g.
[43]), or antagonistic interactions between bacterial types [44].
Furthermore, this high variation in coral-associated assemblages
may reflect local, inter-colony differences in host status, such as
genetic identity [45], physiological condition [46], or develop-
mental state ([47] and references therein). In their study of M.
oculata-associated microbes, Hansson et al. [29] also reported
significant inter-colony differences, which may even be further
enhanced by intra-colony differences between single polyps
[48,49]. In addition to passive controls, bacterial colonization
may also be actively regulated by the coral host (e.g. [50]) in
adaptation to changing environmental conditions [51].
Reef zoning, boundary and location
In general, bacterial communities mapped within the Røst reef
center (Røst-in) revealed surprisingly similar patterns, despite the
pronounced geomorphologic reef zoning (Table 3). This was
unexpected, because seabed features often strongly affect and
reflect local environmental dynamics of e.g. current regime,
sediment deposition and diagenesis as well as organic matter
quality, transport and remineralization within only few tens of
meters (e.g. [18,19]). Hence, the clearly distinguishable reef
features present at Røst-in were assumed to significantly contribute
to the structuring of bacterial assemblages, particularly so in water
and sediments, but also in coral-derived microbial habitats.
The observed local similarity of bacterial community structure
across reef-internal zones was not maintained beyond the reef
center (Røst-in), due to significant community changes towards the
reef periphery (Røst-out; Table 2). Only the branch communities
remained similar, thereby marking an intriguing partition between
the two coral-generated microbial habitats, branch and mucus.
The scale-independent similarity of communities in branch versus
mucus may be attributed to the circumstance that branch surface
samples also included traces of coenosarc tissue, which may not
only contain internal bacterial cells [31] but also exhibit external
Figure 4. Pairwise comparison of OTU overlap between microbial habitats at each reef site. Samples were grouped according to
microbial habitat type and reef site. In this asymmetrical representation, rows correspond to the reference group and columns to the group being
compared. It provides an overview of potential directional dynamics between different microbial habitat types, with the respective fraction (%) of
shared OTUs indicated by different degrees of shading.
doi:10.1371/journal.pone.0032093.g004
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org5 March 2012 | Volume 7 | Issue 3 | e32093

Page 6

biofilm formation [43]. Those tissue- and biofilm-associated cells
could be buffered more from exogenous change than mucus-
inhabiting assemblages due to their embedding in the respective
intra- and/or extra-cellular matrices, while the latter are more
exposed to water column processes and thereby more prone to
mirroring local, meso-scale spatial and environmental shifts.
Within Røst, such shifts may occur as a function of the marked
zone transition from the reef center to the periphery, with the
latter representing an interface (‘‘ecotone’’; sensu [52]) between the
structurally complex reef ecosystem and the more uniform, level
bottom down-sloping into the abyssal plain [7]. Consequently, the
underlying spatial and environmental changes that select for
certain community structures may not necessarily follow linear
distance relationships, but rather be subject to a whole interplay of
locally different, ecosystem-specific factors.
The presence of spatial and environmental imprinting became
even more evident by the finding of significant bacterial
community differences between the four reef sites (Table 1).
Figure 5. NMDS ordination of all ARISA community profiles. For each sample, consensus signals of PCR triplicates were used. Objects
represent consensus signals for all PCR triplicates per sample and share a more similar community structure when plotting closer to each other (Bray-
Curtis distance). A posteriori groupings specify microbial habitat type and reef site. The low stress value indicates appropriate representation of the
original Bray-Curtis dissimilarity matrix into a 2-dimensional space.
doi:10.1371/journal.pone.0032093.g005
Table 1. PERMANOVA of regional, inter-reef bacterial
variation at all study sites as related to reef site (REEF) and
microbial habitat type (HABITAT).
Test factora
Test groupR2b
F-ratioPc
HABITAT all samples 0.46717.222 0.001 ***d
HABITAT Røst0.777 20.947 0.001 ***d
HABITAT Trænadjupet0.794 10.273 0.001 ***d
HABITATTisler0.801 24.1350.001 ***d
HABITAT Langenuen0.748 9.8880.001 ***d
REEF all samples0.124 2.795 0.001 ***d
REEF branch 0.3932.262 0.032 *
REEFmucus 0.75411.210 0.001 ***d
REEFwater 0.83513.476 0.001 ***d
REEF sediments0.504 7.4520.001 ***d
aSource of variation.
bAmount of explained variation.
cSignificance level, assessed by 999 random permutations (*** P#0.001,
** P#0.01, * P#0.05).
dSignificance level below Bonferroni correction threshold.
doi:10.1371/journal.pone.0032093.t001
Table 2. PERMANOVA of local, intra-reef bacterial variation at
Røst as related to reef boundary, i.e. in-/out-reef location (IN-
OUT).
Test factora
Test groupR2b
F-ratioPc
IN-OUT all Røst samples 0.0483.100 0.006 **
IN-OUTbranch 0.0951.684 0.073
IN-OUTmucus 0.3197.478 0.001 ***d
IN-OUT water0.300 3.4330.01 *
IN-OUTsediments 0.1652.967 0.003 **
aSource of variation.
bAmount of explained variation.
cSignificance level, assessed by 999 random permutations (*** P#0.001,
** P#0.01, * P#0.05).
dSignificance level below Bonferroni correction threshold.
doi:10.1371/journal.pone.0032093.t002
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org6March 2012 | Volume 7 | Issue 3 | e32093

Page 7

Remarkably, observed patterns not only reflected local site
specificity (i.e. characteristic assemblages at different reefs), but
also regional specificity (i.e. marked separation between both
offshore versus each of the inshore reefs; Fig. S4). Although all reef
sites share specific geological and hydrological features that are
pivotal for local coral recruitment and proliferation (e.g. [1] and
references therein), Røst, Trænadjupet, Tisler and Langenuen
substantially differ in their geo- and hydrographical setting and
structure. This seemed clearly reflected in the pronounced site
specificity of water- and sediment-inhabiting bacteria. But also
coral-generated microbial habitats exhibited reef-specific bacterial
differences, more so in mucus than in branch samples, which
basically confirmed the aforementioned differences from the local
(meso-) to the regional (large-) scale. Those differences may involve
reef-specific variations in mucus composition [53] as governed by
environmental controls [54], or geographic fluctuations in coral
reproduction strategy and genetic variability [45] as well as local
coral food supply and quality [55,56].
Coral species and color
L. pertusa and M. oculata exhibited significantly different bacterial
community structures, largely due to differences in OTU relative
abundances. This corroborates evidence presented by Hansson et
al. [29] who found DGGE signals from both species to group
separately in NMDS and cluster analyses (yet, with .50%
similarity). Like in our study, those corals originated from the
same sampling location, where they occurred right next to each
other, hence, a mere spatial separation of bacterial assemblages
may not explain this pattern. Also, comparisons of 16S rRNA gene
sequences from different studies [28–30] suggested divergence
between L. pertusa and M. oculata-associated bacterial communities.
As both coral species differ with respect to tissue-contained acid
concentrations [57] as well as to the carbohydrate composition of
their mucus [53], the involvement of host-related traits in
structuring bacterial assemblages seems likely. Species-specific
differences in mucus composition are also known from warm-
water corals [58] and even held responsible for a close attuning of
bacterial communities to host metabolism [59]. Host-related traits
may also explain why L. pertusa and M. oculata-associated
communities slightly varied in their response to spatial heteroge-
neity (Table S2A, C, B), indicating possible coral-specific
differences in host-microbe interactions.
In contrast to the marked species-related patterns, no bacterial
community differences were significantly linked to L. pertusa color
type. This was due to the fact that none out of the 387–390 L.
pertusa-associated OTUs were exclusively attributable to either
white or red specimens. In contrast, Neulinger et al. [28] who
previously studied bacterial associates of white and red L. pertusa
from the Tautra Reef by 16S rRNA-based T-RFLP and sequence
analysis, observed color-specific associations of distinct 16S rRNA
gene phylotypes, but could not resolve community differences
among color phenotypes by fingerprinting. In principle, ARISA
offers more resolution than T-RFLP to detect OTU changes [60]
as well as intra-genomic heterogeneities within closely related gene
clusters [61]. In addition, the two studies differ with respect to
sample origin and processing: While the Tautra corals were
completely homogenized, introducing considerable amounts of
tissue and carbonate skeleton into the analysis, the Røst corals
were distinctly sampled for branch surface plaques and mucus
exudates, targeting mainly interfacial communities on the coral
surface. Hence, the consistent absence of significant phenotypic
community change in both these fingerprinting studies may indeed
indicate that bacterial associates of white and red L. pertusa are
largely indiscriminative, at least by those techniques.
Noticeably, the finding of low numbers of shared bacterial types
combined with significant community differences related to host
taxonomy, are often interpreted as signs of host specificity,
implying the selection of few beneficial associates as part of
commensalistic or mutualistic relationships (e.g. [62] and refer-
ences therein). Yet, host-microbe associations may only be called
‘‘specific’’ if the respective bacterial signatures are maintained over
time and space, which is so far unresolved for bacteria associated
with CWCs due to the lack of studies across spatial and temporal
scales. Here, habitat-specific community patterns were conserved
over all sites, including differences between low and high numbers
of unique OTUs in coral-derived versus ambient microbial habitats,
respectively. The divergence between bacterial communities
associated with L. pertusa and M. oculata also appeared consistent,
at least within a highly heterogeneous reef site such as Røst.
Further, 3 out of 8 OTUs unique to L. pertusa branch at Røst-in
also occurred in the branch samples at all other sites (i.e. Røst-out,
Trænadjupet, Tisler, Langenuen), potentially indicating special-
ized bacterial associates. By contrast, such property was not
identified for any of the mucus-contained OTUs, suggesting
variable colonization of the mucus matrix by locally occurring
communities. Most of the coral-associated OTUs were also found
in the proximal sediments, suggesting a potential source for coral-
associated bacteria. This was corroborated by the change of coral-
associated bacterial communities from local (small- and meso-) to
regional (large-) scale, resulting in biogeographic patterns
comparable to those of ambient bacterial communities. Host
specificity of bacterial types and communities should therefore be
further investigated by colonization experiments with artificial
surfaces.
Bacterial biodiversity hotspots?
By providing a high degree of structural complexity and habitat
heterogeneity, CWC reef ecosystems locally promote faunal
Table 3. PERMANOVA of local, intra-reef bacterial variation at
Røst-in as related to geomorphologic reef zoning (ZONE),
coral species (SPECIES), coral color phenotype (COLOR), and
microbial habitat type (HABITAT).
Test factora
Test groupR2b
F-ratioPc
ZONE all Røst-in samples
ZONEbranch 0.1430.7520.752
ZONEmucus0.1460.7670.803
ZONEwater0.399 1.3260.260
ZONEsediments0.2091.589 0.134
SPECIESf
all Røst-in samples0.41611.742 0.001 ***d
SPECIESf
branch 0.236 3.0860.001 ***d
SPECIESf
mucus0.1631.942 0.043 *
COLORf
all Røst-in samples 0.024 0.3980.868
COLORf
branch0.0110.8780.607
COLORf
mucus0.0820.6310.794
HABITATf
all Røst-in samples0.58723.411 0.001 ***d
aSource of variation.
bAmount of explained variation.
cSignificance level, assessed by 999 random permutations (*** P#0.001,
** P#0.01, * P#0.05).
dSignificance level below Bonferroni correction threshold.
fHierarchical sampling design, with each factor nested within the next higher
one.
doi:10.1371/journal.pone.0032093.t003
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org7 March 2012 | Volume 7 | Issue 3 | e32093

Page 8

diversity in the deep sea [7,12]. The potential of CWC reefs to also
represent biodiversity hotspots for bacterial communities seems,
however, rather questionable, given the low bacterial OTU
number (as proxy for bacterial richness) and the limited OTU
specificity (as indicator of conserved bacterial signatures) associ-
ated with coral-generated surfaces detected in this study.
Nevertheless, CWCs contributed 5% (OTUs found exclusively
on corals) to 44% (OTUs shared between corals and water/
sediments) of all bacterial OTUs detected at a single proliferating
reef center (Røst-in). Furthermore, coral-generated surfaces, such
as branch and mucus, were characterized by high bacterial
community variation. Although high variability does not auto-
matically translate into high local diversity, these small-scale
differences may, combined with the meso- and large-scale
community changes occurring within and between reef sites,
increase bacterial community variation (beta diversity) in CWC
ecosystems and contribute to specific source-sink dynamics in
bacterial dispersal. In terms of bacterial types, however, it may
rather be the reef sediments, water column or filtering inverte-
brates such as sponges, that contribute more to the overall
bacterial diversity in CWC reefs than surfaces generated by L.
pertusa or M. oculata. Those microbial habitats can feature high
levels of organic matter and nutrients ([1] and references therein;
[63]) and are usually characterized by diverse bacterial commu-
nities (e.g. [64,65]).
Given that ARISA is based on the discrimination of ITS length
among bacterial types, the technique cannot be used to precisely
delineate the taxonomic levels at which the observed community
changes operate [61]. Yet, based on our evaluation of community
shifts at multiple levels of observation, future investigations may
involve sequencing of ribosomal genes on representative samples
to explore finer taxa-environment relationships. As reported by
previous studies (e.g. [61,66,67]), bacterial community patterns
derived from ARISA and sequencing (both Sanger and high-
throughput) may, after all, be highly comparable and lead to very
similar ecological conclusions.
Conclusion
This study presents the first multi-scale survey of bacterial
communities associated with L. pertusa and M. oculata in different
CWC reefs, spanning a total of five study levels: (i) Microbial
habitat type and (ii) coral species and color, as well as the spatial
components (iii) geomorphologic reef zoning, (iv) reef boundary,
and (v) reef location. Our findings revealed fundamental
differences in bacterial specificity and community structure
between distinct coral-generated (branch surface, mucus) and
ambient microbial habitats (seawater, sediments), which appeared
pivotal for determining bacterial variation over all observational
levels investigated. Especially the high community variability
associated with coral-derived surfaces represented a consistent
feature of all four CWC reef sites under study. In addition,
bacterial communities changed markedly from local (small- and
meso-) to regional (large-) scale, suggesting significant biogeo-
graphic imprinting of seawater-, sediment- and even coral-
associated communities, but weak microbe-host specificity.
Overall, the relative effects of the different test parameters did
not reflect any linear or even hierarchical relationship of bacterial
community organization, but the deterministic effect of microbial
habitat type and the strong effect of reef location seemed to play
dominant roles in structuring bacteria in CWC reefs. The bacterial
communities were structured across different spatial scales, from
local within-reef habitats to regional across-reef systems, which
may reflect the combined effects of local community history (e.g.,
community assembly and interactions) and environmental filter-
ing. As exploring bacterial diversity in CWC reefs is but one of the
first steps to a better understanding of coral-microbe relationships
in complex deep-sea environments, further studies need to address
how changes in bacterial (and general microbial) diversity affect
the dynamics and functioning of CWC reef ecosystems – especially
in the context of global environmental changes and the protection
of CWC reefs as biodiversity hotspots.
Materials and Methods
Study sites
The four Norwegian CWC reef ecosystems sampled (Fig. 1A)
comprised two offshore sites on the northern mid-Norwegian
continental shelf (Røst, Trænadjupet), as well as two inshore sites
located in the Norwegian Skagerrak (Tisler) and on the Norwegian
South-West coast (Langenuen). A more detailed description,
including sampling times, sampling coordinates, water depths
and sample type (i.e. coral, seawater and sediment) yields, is
provided as Supporting Information (Text S1; Fig. S1, Table S1).
Hierarchical sampling design
Sampling was performed hierarchically, encompassing five
levels of observation, including three spatial scales (Fig. 2). The
first study level (HABITAT) comprised four potentially distinct
types of microbial habitats associated with and surrounding a
scleractinian coral colony in its reef environment: Coral branch
surface, coral mucus, ambient seawater, and proximal sediments.
The second level (COLOR, SPECIES) featured specific coral
species (L. pertusa, M. oculata; Fig. 1B) and coral color types (white
and red individuals of L. pertusa). Geomorphologic reef zoning, as
prevailing at Røst, determined the third level (ZONE), including
the terrace-covered ridge crest, the rubble- and sponge-dominated
ridge slope, and the clay-bearing, sparsely populated inter-ridge
depression in the reef center. At the fourth level (IN-OUT), the up-
slope reef center (Røst-in) was compared with the down-slope reef
periphery (Røst-out) in distances of 1 m, 10 m and 100 m away
from the reef margin. The fifth level (REEF) allowed a comparison
of the offshore Røst site with the nearby Trænadjupet site and the
two inshore sites, Tisler and Langenuen. Arranged in a nested
layout, each of these study levels was considered as integral part of
the respective next higher level, along an increasing gradient of
complexity ranging from level one (HABITAT) to level five
(REEF).
At Røst, the main study site, where sampling focused on intra-
reef differences (HABITAT, SPECIES/COLOR, ZONE, IN-
OUT; Fig. 2), the collection of corals (L. pertusa, M. oculata),
seawater and surface sediments was performed during two
manned submersible dives down-slope across the reef (Fig. S1,
Table S1). The first dive (ship station: PS 70/17-1) traversed two
of the uppermost ridges in the reef center, while the second dive
(ship station: PS 70/31-1) extended further down-slope to the reef
periphery at a distance of approx. 2.5 km. At the other study sites
(inter-reef differences: REEF; Fig. 2), sampling involved the
collection of L. pertusa, seawater and sediments at random locations
within the respective main reef area (Table S1).
In-situ sample collection
Specimens of living CWCs were sampled by manned submers-
ible (Røst, Trænadjupet) or video-assisted remotely operated
vehicle (Tisler, Langenuen; Table S1). After visual assessment of
each target colony in situ, one healthy looking fragment was picked
from the colony’s living outer rind using the manipulator arm, and
placed into a separate compartment of the sampling reservoir.
Onboard, each specimen was inspected for epigrowth, impurities
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org8 March 2012 | Volume 7 | Issue 3 | e32093

Page 9

or degeneration, before selecting an intact fragment (5–15 cm in
length) for sub-sampling of coral-associated microbial habitats.
Fragments needed for branch surface and mucus sampling were
maintained in flow-through tanks with in-situ water at a
temperature of 10–11uC for #30 min until subsequent processing.
Seawater was sampled with 2l-Niskin bottles attached to the
submersible (Røst, Trænadjupet), or mounted on a conductivity-
temperature-depth rosette sampler (Langenuen) or video-assisted
steel cable (Tisler; Table S1). Immediately after retrieval, 2l-water
samples were kept at 4uC, filtered in 500 ml aliquots onto sterile
polycarbonate membranes (0.22 mm pore size, Millipore, Billerica,
MA), and stored at 220uC until further treatment. Surface
sediments (approximately 0–5 cm sediment depth) were collected
by custom sampling scoops operated via the submersible and
vehicle manipulator arm (Røst, Trænadjupet, Tisler) or by Van-
Veen grab (Langenuen; Table S1). Upon retrieval, sediment
samples were immediately transferred into sterile 50-ml vials and
stored at 220uC until further processing. At Røst, sediment
sampling was not possible on the ridge crests owing to the density
of the prevailing coral framework cover.
Coral sub-sampling procedures
After gentle rinsing with sterile-filtered (Whatman, Maidstone,
UK) local seawater, branch surfaces of living corals were sampled
by scraping an area of up to 5 cm2per fragment with sterile scalpel
blades, yielding a mixture of surface plaques, coenosarc tissue, and
calcareous particles. Scraping was carried out on the primary, and
partly secondary, branches of each fragment, avoiding fragile
outer branches as well as polyp calices. All material accumulated
per fragment was directly transferred into a DNA extraction tube
(see below). Freshly produced coral mucus was sampled by gently
rinsing living coral fragments with sterile-filtered seawater and
inducing mucus exudation through 2–5 min air exposure. After
discarding exudate released during the first minute, subsequent
production of up to 0.5 ml per fragment was collected directly
from polyp surfaces by using sterile syringes. Resulting mucus-
seawater mixture was concentrated onto sterile polycarbonate
filters (Whatman), and frozen at 220uC until DNA extraction.
DNA extraction
Total community DNA was extracted and purified with the
Ultra Clean Soil DNA Kit (MoBio, Carlsbad, CA, USA) following
the manufacturer’s instructions for maximum yield, with slight
modifications. Branch samples (scrapings from up to 5 cm2surface
per fragment) and sediment samples (361 g per scoop) were
directly transferred into extraction tubes, mucus samples (up to
0.5 ml per fragment) and water samples (2–46500 ml per Niskin
bottle) on respective filter membranes. DNA yields were quantified
by NanoDrop spectrophotometry (NanoDrop, Wilmington, DE).
For a complete overview of sample units and replicates subjected
to DNA extraction and subsequent community analyses, see Table
S1.
Community fingerprinting and multivariate analyses
Universal bacterial ARISA [68] based on 3 PCR replicates per
sample, and subsequent binning into OTUs were carried out as
described previously [69]. Based on a threshold of $0.09% in
relative fluorescence intensity (individual peak areas divided by the
total peak area of the respective sample) and 50 in fluorescence
units, only ARISA fragments in the size range of 100–1000 bp
were subjected to the binning procedure with a window size of
2 bp. Total numbers of ARISA-based OTUs (i.e. the number of
bacterial types contained in each ARISA sample; used as relative
proxy for richness), were assessed for mean difference by applying
the non-parametric, omnibus Kruskal-Wallis test (KW), followed
by pairwise Wilcoxon-Mann-Whitney tests (WMW). Multivariate
patterns in community structure were analyzed based on Bray-
Curtis dissimilarity matrices which were visually inspected by Non-
metric MultiDimensional Scaling (NMDS), cluster analysis, and
heat-mapping. Differences in community overlap between a
posteriori defined sample categories were tested by Analysis of
Similarity (ANOSIM) and corrected for multiple tests according to
the Bonferroni criterion. Sources of bacterial community differ-
ences were further assessed by Permutational Multivariate Analysis
of Variance (PERMANOVA; [70]) at three levels: (i) regional,
inter-reef differences between the four different reef sites (REEF,
HABITAT; Fig. 2) (ii) local, intra-reef differences in the whole
Røst area (IN-OUT; Fig. 2), and (iii) local, intra-reef differences
(ZONE, SPECIES, COLOR, HABITAT; Fig. 2) at the reef center
of Røst. Numerical analyses were implemented in PAST v2.0
(Palaeontological Statistics) and in R v.2.9 (The R Project for
Statistical Computing) using the standard and vegan packages, as
well as custom scripts.
Supporting Information
Text S1
(DOC)
More detailed description of the study sites.
Figure S1
sampling events at Røst. (A) Røst bathymetry, including dive
transects at Røst-in (reef center) and Røst-out (reef periphery;
map: courtesy of V. Unnithan, JUB), (B) Røst transversal scheme
(not to scale) indicating topographical reef structure, geomorpho-
logical reef zoning and single sampling stations (reef center:
val=valley,slo=slope, cre=crest;
100 m=1/10/100 m beyond the apparent reef margin).
(TIF)
Geographical and topographical setting of
reef periphery:1/10/
Figure S2
distinct coral-associated and ambient microbial habi-
tats. Numbers indicate the amount of OTUs unique to each
microbial habitat, or common to any two or all microbial habitats:
(A) Bacterial OTUs associated with samples of white L. pertusa (left),
red L. pertusa (middle) or M. oculata (right) and their ambient
environment at Røst-in, (B) Bacterial OTUs associated with
samples of all coral species/colors and their ambient environment
at Røst-in (left) or at all sites combined (right).
(TIF)
Partitioning of bacterial OTUs between
Figure S3
ships between all community profiles. Samples are grouped
according to microbial habitat type, coral species and color,
geomorphologic reef zoning, reef boundary (incl. distances away
from the apparent reef margin), and reef site. Cell position
corresponds to the symmetrical pairing of single sample groups.
Cell shading indicates the magnitude of dissimilarity between
sample pairs.
(TIF)
Pairwise Bray-Curtis dissimilarity relation-
Figure S4
profiles per microbial habitat. For each microbial habitat
type, differences in bacterial community structure are plotted as
related to reef site, reef boundary, geomorphologic reef zoning,
coral species and color. Objects represent consensus signals for all
PCR triplicates per sample and share a more similar community
structure when plotting closer to each other (Bray-Curtis distance).
Stress values indicate the goodness-of-fit of the 2-dimensional
representation compared to the original multi-dimensional matrix.
(TIF)
NMDS ordinations of ARISA community
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org9March 2012 | Volume 7 | Issue 3 | e32093

Page 10

Table S1
specimens, seawater and sediments at the different study
sites. wL=white L. pertusa, rL=red L. pertusa, rM=red M. oculata.
aSample lost during processing. The full station list of ARKXXII/1a is
available via the PANGAEA database at: http://www.pangaea.
de/ddi?retr=events/HERMES/ARK-XXII_1a.retr&conf=events/
CruiseReportHTML.conf&title=Station+list+of+cruise+ARK-XXII/
1a&format=html.
(DOC)
Overview on sampling events of living coral
Table S2
variation done at Røst. A) Analyses considering A) reef
boundary, i.e. in/out-reef location (IN-OUT) and B)
geomorphologic reef zoning (ZONE) within Røst-in.
aSource of variation.
cance level, assessed by 999 random permutations (*** P#0.001,
** P#0.01, * P#0.05).
correction threshold.
(DOC)
PERMANOVA of coral-associated bacterial
bAmount of explained variation.
cSignifi-
dSignificance level below Bonferroni
Acknowledgments
Our sincere thanks go to the JAGO Team (IfM Geomar) as well as the
crew and scientific shipboard parties of RV Polarstern (ARK XXII/1a),
RV G.O. Sars, RV Lophelia, and RV Hans Brattstrøm for excellent
support during in-situ sampling and onboard sample processing. We also
greatly acknowledge Hans Tore Rapp and Christopher Schander (UiB), as
well as Tomas Lunda ¨lv, Lisbeth Jonsson, and Anne Larsson (SLCMS) for
fruitful ‘joint ventures’, helpful discussions, and logistical support. We
further thank Laura Wehrmann and Verena Witt (MPI-MM) for assistance
in the field and in the lab. Special thanks to Stefanie Gru ¨nke (MPI-MM) for
valuable comments on the manuscript.
Author Contributions
Conceived and designed the experiments: SS CW FH AB AR. Performed
the experiments: SS FH. Analyzed the data: SS AR. Contributed reagents/
materials/analysis tools: AR. Wrote the paper: SS CW FH AB AR.
References
1.Roberts M, Wheeler A, Freiwald A, Cairns S (2009) Cold-water corals: The
biology and geology of deep-sea coral habitats. New York: Cambridge
University Press.
Jensen A, Frederiksen R (1992) The fauna associated with bank-forming
deepwater coral Lophelia pertusa (Scleractinia) on the Faroe shelf. Sarsia 77:
53–69.
Freiwald A, Fossa ˚ JH, Grehan A, Koslow T, Roberts JM (2004) Cold-water
coral reefs. Cambridge: UNEP-WCMC.
Henry LA, Roberts JM (2007) Biodiversity and ecological composition of
macrobenthos on cold-water coral mounds and adjacent off-mound habitat in
the bathyal Porcupine Seabight, NE Atlantic. Deep Sea Res I 54: 654–672.
Hovland M (2008) Deep-water coral reefs: Unique biodiversity hotspots. Berlin:
Springer.
Roberts JM (2006) Reefs of the deep: The biology and geology of cold-water
coral ecosystems. Science 312: 543–547.
Buhl-Mortensen L, Vanreusel A, Gooday AJ, Levin LA, Priede IG, et al. (2010)
Biological structures as source of habitat heterogeneity and biodiversity on deep-
ocean margins. Mar Ecol 31: 21–50.
Henry LA, Davies AJ, Roberts JM (2010) Beta diversity of cold-water coral reef
communities off western Scotland. Coral Reefs 29: 427–436.
Lindner A, Cairns SD, Cunningham CW (2008) From offshore to onshore:
Multiple origins of shallow-water corals from deep-sea ancestors. PLoS ONE 3:
e2429.
10. Jones CG, Lawton JH, Shachak M (1994) Organisms as ecosystem engineers.
Oikos 69: 373–386.
11. Berke SK (2010) Functional groups of ecosystem engineers: A proposed
classification with comments on current issues. Integrat Comparat Biol 50:
147–157.
12. Bongiorni L, Mea M, Gambi C, Pusceddu A, Taviani M, et al. (2010) Deep-
water scleractinian corals promote higher biodiversity in deep-sea meiofaunal
assemblages along continental margins. Biol Conserv 143: 1687–1700.
13. Spalding M, Ravilious C, Green P (2001) UNEP-WCMC world atlas of coral
reefs. Berkeley: University of California Press.
14. Wilson JB (1979) ‘Patch’ development of the deep-water coral Lophelia pertusa (L.)
on Rockall Bank. J Mar Biol Assoc UK 59: 165–177.
15. Freiwald A, Henrich R, Pa ¨tzold J (1997) Anatomy of a deep-water coral reef
mound from Stjernsund, West-Finnmark, northern Norway. In: James NP,
Clarke JAD, editors. Cool-water carbonates. SEPM Spec Publ 56: 141–161.
16. Rogers AD (1999) The biology of Lophelia pertusa (Linnaeus 1758) and other
deep-water reef-forming corals and impacts from human activities. Intl Rev
Hydrobiol 84: 315–406.
17. Mortensen PB, Hovland MT, Fossa ˚ JH, Furevik DM (2001) Distribution,
abundance and size of Lophelia pertusa coral reefs in mid-Norway in relation to
seabed characteristics. J Mar Biol Assoc UK 81: 581–597.
18. Freiwald A (2002) Reef-forming cold-water corals. In: Wefer G, Billett D,
Hebbeln D, Jørgensen BB, Schlu ¨ter M, van Weering T, eds. Ocean margin
systems. Berlin: Springer. pp 365–385.
19. Freiwald A, Hu ¨hnerbach V, Lindberg B, Wilson B, Campbell J (2002) The Sula
Reef complex, Norwegian shelf. Facies 47: 179–200.
20. Mortensen PB, Hovland M, Brattegard T, Farestveit R (1995) Deep water
bioherms of the scleractinian coral Lophelia pertusa (L.) at 64uN on the Norwegian
Shelf: Structure and associated megafauna. Sarsia 80: 145–158.
21. Freiwald A, Wilson JB (1998) Taphonomy of modern, deep, cold-temperate
water coral reefs. Historical Biol 13: 37–52.
22. Knowlton N, Rohwer F (2003) Multispecies microbial mutualisms on coral reefs:
The host as a habitat. Am Nat 162: S51–S62.
2.
3.
4.
5.
6.
7.
8.
9.
23. Rosenberg E, Kellogg CA, Rohwer F (2008) Coral microbiology. Oceanogr 20:
146–154.
24. Ainsworth TD, Thurber RV, Gates RD (2010) The future of coral reefs: A
microbial perspective. Trends Ecol Evol 25: 233–240.
25. Yakimov MM, Cappello S, Crisafi E, Tursi A, Savini A, et al. (2006)
Phylogenetic survey of metabolically active microbial communities associated
with the deep-sea coral Lophelia pertusa from the Apulian plateau, Central
Mediterranean Sea. Deep Sea Res I 53: 62–75.
26. Großkurth AK (2007) Analysis of bacterial community composition on the cold-
water coral Lophelia pertusa and antibacterial effects of coral extracts. Diploma
thesis. Oldenburg: Carl von Ossietzky University.
27. Fo ¨rsterra G, Ha ¨ussermann V (2008) Unusual symbiotic relationships between
microendolithic phototrophic organisms and azooxanthellate cold-water corals
from Chilean fjords. Mar Ecol Prog Ser 370: 121–125.
28. Neulinger SC, Ja ¨rnegren J, Ludvigsen M, Lochte K, Dullo WC (2008)
Phenotype-specific bacterial communities in the cold-water coral Lophelia pertusa
(Scleractinia) and their implications for the host’s nutrition, health, and
distribution. Appl Environ Microbiol 74: 7272–7285.
29. Hansson L, Agis M, Maier C, Weinbauer G (2009) Community composition of
bacteria associated with cold-water coral Madrepora oculata: within and between
colony variability. Mar Ecol Prog Ser 397: 89–102.
30. Kellogg CA, Lisle JT, Galkiewicz J (2009) Culture-independent characterization
of bacterial communities associated with the cold-water coral Lophelia pertusa in
the Northeastern Gulf of Mexico. Appl Environ Microbiol 75: 2294–2303.
31. Neulinger SC, Ga ¨rtner A, Ja ¨rnegren J, Ludvigsen M, Lochte K, et al. (2009)
Tissue-associated ‘‘Candidatus Mycoplasma corallicola’’ and filamentous bacteria
on the cold-water coral Lophelia pertusa (Scleractinia). Appl Environ Microbiol 75:
1437–1444.
32. Scho ¨ttner S, Hoffmann F, Wild C, Rapp HT, Boetius A, et al. (2009) Inter- and
intra- habitat bacterial diversity associated with cold-water corals. ISME J 3:
756–759.
33. Penn K, Wu D, Eisen JA, Ward NL (2006) Characterization of bacterial
communities associated with deep-sea corals on Gulf of Alaska seamounts. Appl
Environ Microbiol 72: 1680–1683.
34. Bru ¨ck TB, Bru ¨ck WM, Santiago-Va ´zquez LZ, McCarthy PJ, Kerr RG (2007)
Diversity of the bacterial communities associated with the azooxanthellate deep
water octocorals Leptogorgia minima, Iciligorgia schrammi, and Swiftia exertia. Mar
Biotechnol 9: 561–576.
35. Hall-Spencer JM, Pike J, Munn CB (2007) Diseases affect cold-water corals too:
Eunicella verrucosa (Cnidaria: Gorgonacea) necrosis in SW England. Dis Aquat
Org 76: 87–97.
36. Webster N, Bourne D (2007) Bacterial community structure associated with the
Antarctic soft coral, Alcyonium antarcticum. FEMS Microbiol Ecol 59: 81–94.
37. Gray MA, Stone RP, McLaughlin MR, Kellogg CA (2011) Microbial consortia
of gorgonian corals from the Aleutian Islands. FEMS Microbiol Ecol 76:
109–120.
38. Hewson I, Fuhrman JA (2006) Spatial and vertical biogeography of coral
reef sediment bacteria and diazotroph communities. Mar Ecol Prog Ser 306:
79–86.
39. Bo ¨er SI, Hedtkamp SIC, van Beusekom JJE, Fuhrman JA, Boetius A, et al.
(2009) Time and sediment depth-related variations in bacterial diversity and
community structure in subtidal sands. ISME J 3: 780–791.
40. Bourne DG, Munn CB (2005) Diversity of bacteria associated with the coral
Pocillopora damicornis from the Great Barrier Reef. Environ Microbiol 7:
1162–1174.
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org10March 2012 | Volume 7 | Issue 3 | e32093

Page 11

41. Koren O, Rosenberg E (2006) Bacteria associated with mucus and tissues of the
coral Oculina patagonica in summer and winter. Appl Environ Microbiol 72:
5254–5259.
42. Ritchie KB (2006) Regulation of microbial populations by coral surface mucus
and mucus-associated bacteria. Mar Ecol Prog Ser 322: 1–14.
43. Sweet MJ, Coquer A, Bythell JC (2011) Development of bacterial biofilms in
comparison to surface-assoiated microbes of hard corals. PLoS ONE 6: e21195.
44. Rypien KL, Ward JR, Azam F (2010) Antagonistic interactions among coral-
associated bacteria. Environ Microbiol 12: 28–39.
45. Le Goff-Vitry MC, Pybus OG, Rogers AD (2004) Genetic structure of the deep-
sea coral Lophelia pertusa in the northeast Atlantic revealed by microsatellites and
internal transcribed spacer sequences. Mol Ecol 13: 537–349.
46. Dodds LA, Roberts JM, Taylor AC, Marubini F (2007) Metabolic tolerance of
the cold-water coral Lophelia pertusa to temperature and dissolved oxygen change.
J Exp Mar Biol Ecol 349: 205–214.
47. Brown BE, Bythell JC (2005) Perspectives on mucus secretion in reef corals. Mar
Ecol Prog Ser 296: 291–395.
48. Brooke S, Young CM (2009) In-situ measurement of survival and growth of
Lophelia pertusa in the northern Gulf of Mexico. Mar Ecol Prog Ser 397: 153–161.
49. Maier C, Hegemann J, Weinbauer MG, Gattuso JP (2009) Calicification of the
cold-water coral Lophelia pertusa under ambient and reduced pH. Biogeosci 6:
1671–1680.
50. Rohwer F, Kelley S (2004) Culture independent analyses of coral associated
microbes. In: Rosenberg E, Loya Y, eds. Coral health and disease. Berlin:
Springer. pp 265–278.
51. Reshef L, Koren O, Loya Y, Zilber-Rosenberg I, Rosenberg E (2006) The coral
probiotic hypothesis. Environ Microbiol 8: 2068–2073.
52. Costello MJ (2009) Distinguishing marine habitat classification concepts for
ecological data management. Mar Ecol Progr Ser 397: 253–268.
53. Wild C, Naumann M, Niggl W, Haas A (2010) Carbohydrate composition of
mucus released by scleractinian warm- and cold-water reef corals. Aquat Biol 10:
131–138.
54. Drollet JH, Teai T, Faucon M, Martin PMV (1997) Field study of compensatory
changes in UV-absorbing compounds in the mucus of the solitary coral Fungia
repanda (Scleractinia: Fungiidae) in relation to solar UV-radiation, seawater
temperature and other coincident physico-chemical parameters. Mar Freshw
Res 48: 329–333.
55. Kiriakoulakis K, Fisher L, Freiwald A, Grehan A, Roberts JM, et al. (2005)
Lipids and nitrogen isotopes of two deep-water corals from the North-East
Atlantic: Initial results and implications for their nutrition. In: Freiwald A,
Roberts JM, eds. Cold-water corals and ecosystems. Berlin: Springer. pp
715–729.
56. Dodds LA, Black KD, Orr H, Roberts JM (2009) Lipid biomarkers reveal
geographical differences in food supply to the cold-water coral Lophelia pertusa
(Scleractinia). Mar Ecol Prog Ser 397: 113–124.
57. Mancini I, Guerriero A, Guella G, Bakken T, Zimbrowius H, et al. (1999) Novel
10-hydroxydocosapolyenoic acids from deep-water scleractinian corals. Helv
Chim Acta 82: 677–684.
58. Meikle P, Richards GN, Yellowlees D (1988) Structural investigation on the
mucus from six species of coral. Mar Biol 99: 187–193.
59. Ducklow HW, Mitchell L (1979) Bacterial populations and adaptations in the
mucus layers on living corals. Limnol Oceanogr 24: 715–725.
60. Danovaro R, Luna GM, Dell’Anno A, Pietrangeli B (2006) Comparison of two
fingerprinting techniques, terminal restriction fragment length polymorphism
and automated ribosomal intergenic spacer analysis, for determination of
bacterial diversity in aquatic environments. Appl Environ Microbiol 72:
5982–5989.
61. Brown M, Fuhrman JA (2005) Marine bacterial microdiversity as revealed by
internal transcribed spacer analysis. Aquat Microb Ecol 41: 15–23.
62. Dubilier N, Bergin C, Lott C (2008) Symbiotic diversity in marine animals: The
art of harnessing chemosynthesis. Nature Rev Microbiol 6: 725–740.
63. Hentschel U, Usher KM, Taylor MW (2006) Marine sponges as microbial
fermenters. FEMS Microbiol Ecol 55: 167–177.
64. Jensen S, Neufeld JD, Birkeland DK, Hovland M, Murrell JC (2008) Insight into
the microbial community structure of a Norwegian deep-water coral reef
environment. Deep Sea Res I 55: 1554–1563.
65. Lee OO, Wang Y, Yang JK, Lafi FF, Al-Suweilem K, et al. (2011)
Pyrosequencing reveals highly diverse and species-specific communities in
sponges from the Red Sea. ISME J 5: 650–664.
66. Scho ¨ttner S, Pfitzner B, Gru ¨nke S, Rasheed M, Wild C, Ramette A (2011)
Drivers of bacterial diversity dynamics in permeable carbonate and silicate coral
reef sands from the Red Sea. Environ Microbiol 13: 1815–1826.
67. Bienhold C, Boetius A, Ramette A (2011) The energy-diversity relationship of
complex bacterial communities in Arctic deep-sea sediments. ISME J.
doi:10.1038/ismej.2011.140.
68. Fisher MM, Triplett EW (1999) Automated approach for ribosomal intergenic
spacer analysis of microbial diversity and its application to freshwater
communities. Appl Environ Microbiol 65: 4630–4636.
69. Ramette A (2009) Quantitative community fingerprinting methods for
estimating the abundance of operational taxonomic units in natural microbial
communities. Appl Environ Microbiol 75: 2495–2505.
70. Anderson MJ (2001) A new method for non-parametric multivariate analysis of
variance. Austral Ecol 26: 32–46.
Multiscale Bacterial Diversity Patterns on Corals
PLoS ONE | www.plosone.org 11March 2012 | Volume 7 | Issue 3 | e32093