The ISME Journal (2010), 1–12

& 2010 International Society for Microbial Ecology All rights reserved 1751-7362/10
www.nature.com/ismej

ORIGINAL ARTICLE
Elevated seawater temperature causes a microbial
shift on crustose coralline algae with implications
for the recruitment of coral larvae
Nicole S. Webster, Rochelle Soo, Rose Cobb and Andrew P Negri
Australian Institute of Marine Science, Townsville, Queensland, Australia

Crustose coralline algae (CCA) are key reef-building primary producers that are known to induce the
metamorphosis and recruitment of many species of coral larvae. Reef biofilms (particularly
microorganisms associated with CCA) are also important as settlement cues for a variety of marine
invertebrates, including corals. If rising sea surface temperatures (SSTs) affect CCA and/or their
associated biofilms, this may in turn affect recruitment on coral reefs. Herein, we report that the CCA
Neogoniolithon fosliei, and its associated microbial communities do not tolerate SSTs of 32 1C, only
2–4 1C above the mean maximum annual SST. After 7 days at 32 1C, the CCA exhibited clear signs of
stress, including bleaching, a reduction in maximum quantum yield (Fv/Fm) and a large shift in
microbial community structure. This shift at 32 1C involved an increase in Bacteroidetes and a
reduction in Alphaproteobacteria, including the loss of the primary strain (with high-sequence
similarity to a described coral symbiont). A recovery in Fv/Fm was observed in CCA exposed to 31 1C
following 7 days of recovery (at 27 1C); however, CCA exposed to 32 1C did not recover during this
time as evidenced by the rapid growth of endolithic green algae. A 50% reduction in the ability of
N. fosliei to induce coral larval metamorphosis at 32 1C accompanied the changes in microbiology,
pigmentation and photophysiology of the CCA. This is the first experimental evidence to
demonstrate how thermal stress influences microbial associations on CCA with subsequent
downstream impacts on coral recruitment, which is critical for reef regeneration and recovery from
climate-related mortality events.
The ISME Journal advance online publication, 14 October 2010; doi:10.1038/ismej.2010.152
Subject Category: microbial ecosystem impacts
Keywords: crustose coralline algae; biofilms; temperature; microbes; bleaching; coral settlement

Introduction isolated from CCA, is able to induce significant

levels of metamorphosis of coral larvae (Acropora
Larvae of numerous marine invertebrate species, willisea and A. millepora) (Negri et al., 2001);
including sponges, echinoderms, polychaetes, numerous bacterial strains from CCA induce settle-
barnacles, molluscs, tunicates, bryozoans and ment of the sea urchin Heliocidaris erythrogramma
cnidarians settle and metamorphose in response to (Huggett et al., 2006) and metamorphosis of crown-
bacterial biofilms (reviewed by Wieczorek and of-thorns starfish larvae (Acanthaster planci) is
Todd, 1998). Coral larvae actively select their site also triggered by microorganisms associated with
of permanent attachment using chemical cues to CCA (Johnson and Sutton, 1994).
induce metamorphosis (Morse et al., 1988, 1994, Although it is well-established that settlement
1996; Heyward and Negri, 1999). The primary and recruitment of marine invertebrates varies with
source of these chemical morphogens is various biotic and abiotic factors (hydrodynamics, light,
species of non-geniculate calcareous coralline algae temperature, gravity, surface texture and presence of
(CCA). Bacterial biofilms occur on the surface of CCA conspecifics) (Rodriguez et al., 1993), little work has
(Galbary and Veltkamp, 1980; Garland et al., 1985; been done to test whether environmental variables
Lewis et al., 1985; Johnson et al., 1991) and also affect the inductive ability of marine microbial
have a role in inducing larval metamorphosis for biofilms. Biofilm age and depth can affect the
some invertebrates. A strain of Pseudoalteromonas, inductive and inhibitory capacities of biofilms
(Wieczorek and Todd, 1997; Webster et al., 2004),
and localised anthropogenic impact can also shift
Correspondence: NS Webster, Australian Institute of Marine the biofilm composition (Meyer-Reil and Koster,
Science, Microbiology and Symbiosis, PMB 3 Townsville Mail 2000; Lawrence et al., 2004) and potentially influ-
Centre, Townsville, Queensland 4810, Australia.
E-mail: [email protected] ence subsequent invertebrate recruitment (Webster
Received 27 May 2010; revised 23 July 2010; accepted 18 August et al., 2006). However, to date no research has
2010 addressed how variables associated with a changing
 Impact of temperature on CCA and microbes
NS Webster et al
2
climate (elevated sea-surface temperature (SST), and then all temperatures were returned to 27 1C for
ocean acidification and eutrophication) will affect the final 7 days of the experiment as a recovery
the inductive ability of marine microbial biofilms. period. In each tank, there were 6  6 wells, 12-ml
The potential impacts of climate change on cell-culture plates (Iwaki, Japan) with three wells
marine microorganisms are rarely assessed, despite per plate containing reef sediment at 6 mm depth
the critical contribution of microbes to marine and the other three wells containing the CCA discs.
biomass, diversity and ecosystem function. One of An individual six-well plate was randomly removed
the primary effects of climate change is a projected per tank at each time point (nine total replicates per
1.8–4 1C increase in global air temperature by 2100, temperature at each time for each CCA and sedi-
along with similar rises in SST (range of best ment). Incoming seawater was filtered to 1 mm.
estimates from the six emissions marker scenarios All samples were maintained under a diel cycle of
of Special Report on Emission Scenarios (SRES)) (12:12 h) at 80 mmol quanta m2 s1, reflecting light
(IPCC, 2007). Elevated sea temperatures could intensity at 8 m on the reef. Initially, all aquaria were
significantly impact on marine microbes, potentially left at 27 1C for 72 h and then temperatures were
altering microbial diversity, function and commu- adjusted gradually (0.15 1C h1) until reaching the
nity dynamics (Webster and Hill, 2007). Elevated final temperatures (at T ¼ 0 day).
SST may affect recruitment in a number of ways: At each time point all CCA and sediment samples
(1) by altering the ability of microbes to produce were photographed and frozen in liquid nitrogen for
morphogenic signalling compounds, or (2) by caus- molecular analysis. Seawater samples (1 l) from each
ing a shift in the community composition and temperature at each time point were filtered through
microbial succession of the biofilm, which may a 0.2-mm sterivex filter (Durapore, Millipore, North
adversely (or positively) affect subsequent recruit- Ryde, New South Wales, Australia), which were
ment of macro-organisms, including corals. These filled with 1.8 ml of lysis buffer (40 mM EDTA (pH
outcomes could have serious implications for reef- 8.0), 50 mM Tris and 0.75 M sucrose) and frozen at
building, maintenance and recovery processes. 80 1C.
Environmental conditions that adversely affect the Thermal stress can result in the loss of pigments
distribution and abundance of microbes involved in (bleaching) in CCA (Anthony et al., 2008). Damage
settlement and metamorphosis of reef invertebrates to photosystem II (PSII) of CCA was assessed using
could therefore have large-scale impacts on ecosys- pulse amplitude modulation fluorometry. Fluores-
tem structure and the distribution and reproductive cence measurements were taken following a 30-min
fitness of some keystone species. dark adaptation period using a Mini-PAM fluoro-
Owing to the importance of both inert (sediment- meter (Walz, Effeltrich, Germany) by placing a 6-mm
associated) and CCA-associated biofilms in the fibre-optic probe perpendicular to the centre of the
recruitment of coral larvae, this study aimed to CCA discs. Initial fluorescence (F0) was determined
(1) investigate the impact of elevated SST on the by applying a weak pulse-modulated red-measuring
microbial biofilms of CCA and reef sediments and light (0.15 mmol photons m2 s1). To quantify
(2) assess whether the ability of CCA to induce maximum fluorescence (Fm), a short pulse (800 ms)
metamorphosis in coral larvae can be affected by of saturating actinic light (43000 mmol photons
exposure to elevated SST. m2 s1, 650 nm) was applied (Jones et al., 2003).
Maximum quantum yield (Fv/Fm) of photochemical
Materials and methods energy conversion in PSII was calculated based on
F0 and Fm as per Equation (1) (Genty et al., 1989).
Sample collection and temperature exposure Fv =Fm ¼ ðFm  F0 Þ=Fm ð1Þ
Fragments of the crustose coralline algae Neogonio-
lithon fosliei (8  8 cm) and 1 kg of surface sediments a reduction in Fv/Fm indicates damage to the D1
were collected between 5–8 m at Davies Reef (19109. protein in PSII (Osmond et al., 1999).
350 S; 146152.870 E), Queensland, Australia in Octo-
ber 2008. The upper 2 cm of reef sediments were Coral metamorphosis assays
scraped into glass beakers using a sterile spatula. Larvae were raised from gametes collected from live
All samples were transported in flowing seawater to colonies of the reef-building coral A. millepora
the indoor temperature-controlled aquarium at the (Ehrenberg, 1834). Corals were collected by snorkel-
Australian Institute of Marine Science, Townsville, ling from a depth of 6–8 m on a fringing reef of
Australia. CCA discs of identical size were prepared Pelorus Island, Great Barrier Reef (GBR) (18132 S,
using a 20-mm diamond hole saw. 146129 E) and transported to flow-through aquaria at
The experimental design incorporated four the Australian Institute of Marine Science. Synchro-
temperatures (27 1C, 29 1C, 31 1C and 32 1C; range nous spawning occurred on 14 November 2008, and
±0.2 1C) in three replicate 30-l flow-throughs the released gametes, in the form of buoyant
(600 ml min1) aquaria per temperature. Sampling egg–sperm bundles, were cross-fertilised as des-
times were T ¼ 0 day (all treatments at 27 1C), T ¼ 1 cribed in Negri and Heyward (2000). Larvae
day, T ¼ 2 days, T ¼ 7 days and T ¼ 14 days. were raised in a 500-l plastic tank with flowing
Temperatures were maintained for the first 7 days 1 mm-filtered seawater at 27 1C. Larval metamorphosis

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
3
assays were performed in 10-cm diameter sterile Alphaproteobacteria (Gelsomino and Cacco, 2006;
plastic petri dishes maintained at 28 1C. Coral larvae Huys et al., 2008) and CFB319 with 788R to amplify
(8 days old) (n ¼ 10–20) were introduced to each the 16S rRNA gene from the Cytophaga–Flavobacteria–
dish containing a single CCA disc that had been Bacteroidetes group (Fierer et al., 2005; Huys et al.,
exposed to each of the treatment temperatures 2008). The reverse primers rpoB2041 and 788R were
for 7 days and seawater was added to a final volume modified to incorporate the GC clamp as per Ferris
of 30 ml. Larvae were defined as metamorphosed et al. (1996). DGGE analysis was performed as
when they had changed from either free-swimming described above and band analysis was performed on
or casually-attached pear-shaped forms to squat, Fragment NT Version 1.1a (Quantity One, Bio-Rad
firmly-attached disk-shaped structures with pro- Laboratories Pty Ltd., Gladesville, New South Wales,
nounced flattening of the oral–aboral axis and Australia). Individual band numbers were assigned
typically obvious septal mesenteries radiating from based on the bands migration relative to control bands.
the central mouth region (Heyward and Negri, Bands assigned the same number had identical
1999). A time period of 24 h was chosen as the end migration end points. Bands that did not migrate to
point for scoring early-stage metamorphosis. the same point on duplicate gels were not considered
in subsequent analyses.
The 18S rRNA gene from each CCA sample was
Denaturing gradient gel electrophoresis also amplified by PCR with universal eukaryotic
Frozen CCA tissue and reef sediments (approxi- primers as follows: NS1f, 50 -GTAGTCATATGCTTG
mately 0.5 g per sample) were aseptically transferred TCTC-30 and NS2r, 50 -GGCTGCTGGCACCAGACTT
to 1.5 ml Eppendorf tubes using sterile scalpels. GC-30 (White et al., 1990). The PCR and DGGE were
Grinding buffer (0.5 ml) was added to each replicate performed as described above with the exception
sample (100 mM Tris (pH 9.0), 100 mM EDTA that the products were electrophoresed on a 35–70%
(pH 8.0), 1% SDS and 100 mM NaCl). Tubes were denaturing gradient.
immersed in liquid nitrogen and ground with plastic
pestles. Samples were incubated at 65 1C for 60 min 16D rRNA gene cloning and sequencing
before addition of 187 ml 5 M potassium acetate. For a more comprehensive phylogenetic comparison
Samples were incubated on ice for 30 min and between the different temperatures, the 16S rRNA
centrifuged at 8000 g for 15 min. The supernatants gene from each CCA disc sampled at 2- and 7 day
were transferred to fresh tubes and DNA was was amplified by PCR with primers 63f and 1387r
precipitated with 0.8 vol of isopropanol. DNA was (Marchesi et al., 1998). The PCR products were
extracted from seawater filters by addition of combined for all replicate CCA discs per tempera-
200 ml lysozyme (10 mg ml1), incubation at 37 1C ture at each time point. To assess whether biotic
for 45 min, addition of 200 ml of proteinase K (CCA) and abiotic (sediment) biofilms responded
(0.2 mg ml1) and 1% SDS and incubation at 55 1C differently to elevated temperature, the sediment
for 1 h. Lysates were recovered into fresh eppendorf samples at 27 1C and 32 1C were also collected at
tubes and DNA was extracted with a standard 2- and 7 day and assessed by 16S rRNA gene
phenol:chloroform:isoamyl alcohol procedure and sequence analysis. PCR products were ligated into
precipitated with 0.8 vol of isopropanol. the TOPO TA cloning vector (Invitrogen Australia
The 16S rRNA gene from each CCA disc, sediment Pty Ltd., Mulgrave, Victoria, Australia). Ligations
and seawater sample was amplified by PCR with were sent to the Australian Genome Research
primers 1055f and 1406r (Ferris et al., 1996). PCR Facility (University of Queensland, Brisbane) for
reactions were performed as described by Ferris transformation and sequencing of 192 clones for
et al. (1996) and products from duplicate PCR each CCA library and 96 clones for each sediment
reactions were combined and 15 ml applied to 8% library. Australian Genome Research Facility
w/v polyacrylamide (37.5:1) gels containing a transformed the ligations into Genehog electro-
50–70% denaturing gradient of formamide and competent cells (Invitrogen), grew the cells in
urea. Gels were electrophoresed at 60 1C for 17 h in 2  YT þampicillin media, performed a standard
1  TAE buffer at 50 V using the Ingeny D-Code alkaline lysis plasmid preparation and sequenced
system (Amundsenweg, The Netherlands) and the clones with ABI Bigdye v3.1 chemistry on
stained with 1  Sybr Gold for 30 min, visualised the AB3730xl (standard 50-cm array run module;
under UV illumination and photographed. The Mulgrave, Victoria, Australia). Clone sequences
denaturing gradient gel electrophoresis (DGGE) were submitted to the Genbank under the accession
banding profiles using these universal 16S rRNA numbers HM177481 to HM178920.
gene primers could not be resolved with the CCA
and sediment samples because of extreme bacterial
diversity producing a smear of bands on the gels. Phylogenetic analysis
Additional DGGE-PCR was conducted for the CCA Clone sequence quality was checked manually
samples with primers rpoB1698f and rpoB2041r using Sequencher (Genesearch, Brisbane, Queensland,
specific to the rpoB gene (Dahllöf et al., 2000), F203a Austraila). Chimeric sequences were identified using
with 788R to amplify the 16S rRNA gene from the the programs CHECK_CHIMERA (Maidak et al.,

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
4
1999) and Bellerophon (Huber et al., 2004). Se- maximum quantum yield (Fv/Fm) in CCA at 31 1C
quences were imported into the ARB software and 32 1C (Figure 2). A subsequent increase in Fv/Fm
package (http://www.arb-home.de) (Ludwig et al., was observed when CCA treated at 31 1C was
1998), automatically aligned using FastAligner and returned to 27 1C after 7 days. The CCA discs treated
manually edited. Phylogenetic trees were calculated at 32 1C for 7 days (Figure 1c) developed a green
with almost complete 16S rRNA (1400 bp) se- shade under the surface during recovery, indicating
quences for all the close relatives of target sequences colonisation by endolithic green algae. The Fv/Fm
using the neighbour-joining and maximum parsi- values were not compared with the 32 1C discs
mony methods in ARB. Partial sequences were during recovery, as these endolithic algae would
subsequently imported to the tree without changing have contributed to the fluorescence responses. Live
branch topology using the ARB parsimony-inter- CCA discs treated at 27 1C for 7 days induced
active method. The robustness of inferred tree settlement and metamorphosis of 93±2% of coral
topologies was evaluated after 1000 bootstrap larvae (Figure 3). This level of metamorphosis
resamplings of the neighbour-joining data in the was consistent for CCA treated at 29 1C and 31 1C,
Phylip program (Felsenstein, 1993). but dropped significantly (to 58±2%) for heavily
bleached CCA exposed to 32 1C for 7 days (Po0.01,
Data analyses ANOVA F3,2.05 ¼ 15.3).
Clone sequence results were visualised as phyla Screening of the CCA microbial communities by
level pie charts to assess the phylogenetic affiliation DGGE with universal 16S rRNA gene primers was
of the shifting microbes and heat maps were not possible because of the high microbial diversity,
constructed at a 97% sequence similarity level to which resulted in banding patterns too dense and
reveal the extent of operational taxonomic unit complex for adequate resolution. Conversely,
(OTU) differences between the libraries. screening of the CCA microbial communities by
The variability in percentage of coral metamor- DGGE with rpoB gene primers showed no distinct
phosis on CCA at each temperature treatment clustering because of temperature treatment
was assessed using one-way analysis of variance (Figure 4; Supplementary Figure 1). However, more
(Statistica 6.0, StatSoft, Tulsa, OK, USA). detailed community analysis using 16S rRNA clone
Principal components analysis was performed to libraries revealed a distinct shift in the CCA
analyse similarity in microbial community compo- microbial community composition at the phyla/
sition using DGGE data (Statistica 6.0, StatSoft). class level at 32 1C after 7 days (Figure 5a). Sediment
Matrices were constructed separately for the rpoB biofilms, on the other hand, exhibited no obvious
gene DGGE, the combined analysis of the Alphapro- community differences between the two tempera-
teobacteria and Cytophaga–Flavobacteria–Bacteroi- ture extremes (Figure 5b). The 32 1C CCA libraries
detes DGGE gels, the eukaryotic DGGE gel and were dominated by Bacteroidetes with a lesser
the seawater 16SrRNA DGGE gel using the presence abundance of Alpha-, Gamma- and Deltaproteo-
(1) or absence (0) of a band in each sample. bacteria (Figure 5a). All other CCA libraries were
Pie charts were constructed using the phylo- dominated by Alphaproteobacteria, with a moderate
genetic assignments produced in ARB and plotted abundance of Bacteroidetes and a low abundance of
using Sigmaplot (v7.101, SPSS). all other phyla. The sediment biofilms all contained
Distance matrices were generated in ARB by high proportions of Alpha and Gammaproteobacteria
using Jukes–Cantor correction and OTU assign- and moderate abundances of Deltaproteobacteria,
ments; rarefaction curves and diversity estimates Bacteroidetes, Cyanobacteria, Firmicutes and ‘other’
were produced in DOTUR using a distance of 0.03 comprising the Chloroflexi, Nitrospira and
(Schloss and Handelsman, 2005). Statistical analysis Actinobacteria (Figure 5b).
for CCA and sediment libraries were performed on LIBSHUFF statistical analysis of the libraries
the genetic distance matrices using LIBSHUFF revealed a highly significant difference between
(Schloss et al., 2004) to determine whether varia- the 2- and 7-day 32 1C CCA libraries and all other
bility in library composition between treatments temperature treatments (Table 1) and confirmed that
was because of chance or a real biological effect. there was no significant difference between any of
OTU heat maps were constructed using a log2 scale the sediment biofilm treatments (P40.01). Although
in MOTHUR version 1.8.1 (Schloss et al., 2009). A the 2-day CCA library at 31 1C was significantly
community Newick-formatted tree with a distance different to the 2-day CCA libraries from 27 1C and
of 0.03 was produced for the CCA libraries using 29 1C, respectively; the 7-day CCA library at 31 1C
MOTHUR version 1.8.1 according to the Yue and was only significantly different to the 27 1C library
Clayton theta structural diversity measure. at 7 days. The community composition identified in
the 27 1C libraries for 2- and 7 day were also found to
be statistically different to all other communities
Results (Table 1).
To assess the differences in community composi-
Elevated SST resulted in visible CCA bleaching tion at a higher level of phylogenetic resolution,
(Figure 1) and a corresponding reduction in heat maps were constructed based on an OTU

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
5

Figure 1 Photographic panel of CCA after 7 days at each of the experimental temperature treatments showing loss of pigment at 31 1C
and complete bleaching at 32 1C.

0.6 100

0.5
80
% Metamorphosis (+ SE)

0.4

60
0.3
Fv/ Fm

0.2 40
27°C
29°C
0.1 31°C
32°C 20
0.0 Final treatment
Heating temperatures Recovery (27°C)
0
-2 0 2 4 6 8 10 12 14 27 29 31 32
Time (days) Temperatue of CCA exposure (°C)
Figure 2 Maximum potential quantum yield (Fv/Fm) of CCA in Figure 3 Percent metamorphosis of coral larvae on CCA that had
each temperature treatment during the 7 days exposure and been treated at 27 1C, 29 1C, 31 1C and 32 1C for 7 days.
subsequent 7 days recovery period.

assignment at a species-level (97%) sequence-simi- analysis of the CCA biofilm community phyloge-
larity threshold (Figure 6). Although the heat map of netic tree in ARB revealed three major differences
OTUs revealed some variability within most of the between the 7-day 32 1C treatment and all other
CCA libraries, the 7-day 32 1C treatment was a clear libraries. First, there was a loss of an Alphaproteo-
outlier (Figure 6a). In contrast, much less variability bacteria (HM177926), which made up 30% of the
was detected between the sediment biofilm com- sequences from the 2-day 32 1C library, but only
munities exposed to the two temperature extremes 0.6% of the sequences from the 7-day 32 1C library.
after both 2 and 7 days (Figure 6b). Subsequent This strain dominated all the other CCA libraries

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
6
at both time points and had closest sequence (Sunagawa et al., 2010). The second difference
similarity (96%) to a symbiont from the coral was the appearance of a Deltaproteobacteria
Montastrea faveolata (accession number FJ202556, (HM178559), which was absent from all other
libraries but made up 8% of the sequences from
27°C the 7-day 32 1C library and had highest similarity
8 29°C
(98%) to a Desulfovibrio from a study of healthy and
diseased corals (accession number FJ425598). The
31°C
6 third major difference was a substantial increase in
32°C the abundance of a Bacteroidetes strain (HM178650)
Factor 2 (9.0%)

4 0d with 99% sequence similarity to a Flavobacterium

1d (accession number AB270586). This strain made up
2 39% of the total sequences from the 7-day 32 1C
2d
7d
library, but only 0.5% and 1.6% of the 2-day 32 1C
0 and 7-day 27 1C libraries, respectively.
Recovery
To further visualise the similarity between CCA
-2
biofilm communities at different SSTs, a community
-4 tree was constructed based on an OTU assignment at
-6 -4 -2 0 2 4 a 97% sequence-similarity threshold (Figure 7). The
Factor 1 (10.1%) tree demonstrates the clear differentiation of the
CCA exposed to 32 1C for 7 days. This analysis also
Figure 4 Principal components analysis of CCA bacterial
community composition at each of the temperature treatments.
reveals that the 7-day 31 1C and 2-day 32 1C biofilms
DGGE banding pattern data generated with rpoB gene primers was cluster together and apart from the biofilms exposed
used to construct a similarity matrix. to lower temperatures.

2 Days CCA

7 Days CCA

Bacteroidetes
Alphaproteobacteria
Gammaproteobacteria
27°C 29°C 31°C 32°C Deltaproteobacteria
Epsilonproteobacteria
Firmicutes
Tennericutes
2 Days Sediment 7 Days Sediment
Cyanobacteria / Diatoms
27°C 32°C 27°C 32°C Other

Figure 5 Phyla and class-level differences in bacterial biofilm composition on CCA (a) and reef sediment (b) at each temperature
treatment. The ‘other’ category contains sequences from the Chloroflexi, Nitrospira and Actinobacteria phyla. Graphs were constructed
using the frequency of 16s rRNA sequences belonging to each bacterial group from clone library analysis.

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
7
Table 1 LIBSHUFF analysis of variability in CCA clone library samples (Supplementary Figure 5), there was no trend
community composition associated with temperature treatment of the CCA
Library comparison Significance Notes
samples. This was further illustrated with principle
components analysis revealing no clear separation of
Time 2 days
the samples according to temperature treatment
27 1C vs ALL 2 days Po0.002 T 2 days 29 1C (Supplementary Figure 6). Similarly, 16S rRNA gene
subset of T 2 days DGGE analysis of the seawater samples from each of
27 1C (P40.01) the temperature treatments revealed a greater number
29 1C vs 31 1C P ¼ 0.002 of bands in the 7 days and recovery time points
32 1C vs ALL 2 days Po0.0002
(Supplementary Figure 7), but no clear effect of
Time 7 days temperature. Principle components analysis of this
27 1C vs ALL 7 days Po0.001 T 7 days 29 1C DGGE revealed clear clustering of the microbial
subset of T 7 days communities according to sampling time rather than
27 1C (P40.01)
32 1C vs ALL 7 days P ¼ 0.0001
treatment temperature (Supplementary Figure 8).

Between time
7 days 27 1C vs ALL 2 days P ¼ 0.0001
7 days 29 1C vs 2 days 32 1C P ¼ 0.0001 Discussion
7 days 31 1C vs 2 days 29 1C Po0.001
7 days 31 1C vs 2 days 32 1C P ¼ 0.0001 T 2 days 32 1C This study revealed that the CCA N. fosliei, and its
subset of T 7 days associated microbial communities do not tolerate
31 1C (P40.01) SSTs of 32 1C, only 2–4 1C above the mean maximum
7 days 32 1C vs ALL 2 days P ¼ 0.0001 annual SST at Davies Reef where the CCA
was collected (http://data.aims.gov.au/awsqaqc/do/
Abbreviations: ALL, all other treatments; CCA, crustose coralline
algae; T, time. start.do). After 7 days at 32 1C, the CCA exhibited
Only statistically significant comparisons are shown. clear signs of stress, including visual evidence of
**No significant differences were detected between any of the four bleaching, a reduction in maximum quantum yield
sediment libraries.
and a large shift in microbial community structure
involving a move away from Alphaproteobacteria
To verify the species-level shift detected in CCA towards a dominance by Bacteroidetes. A recovery
clone library analysis, additional DGGE was con- in photosynthetic capacity (Fv/Fm) was observed in
ducted using primers specific to the Alphaproteo- CCA exposed to 31 1C following 7 days of recovery
bacteria and Bacteroidetes groups (Supplementary (at 27 1C); however CCA exposed to 32 1C did not
Figures 3 and 4). Principle components analysis of appear to recover during this time as evidenced
these DGGE profiles revealed clear separation of the by the rapid growth of endolithic green algae. A
27 1C and 29 1C libraries from the 0- and the 1-day reduction in the ability of N. fosliei to induce coral
27 1C libraries (Figure 8). Consistent with the clone larval metamorphosis at 32 1C accompanied the
library analysis, the profile from CCA at 32 1C was changes in microbiology, pigmentation and photo-
distinct from the other samples after 7 days. However, physiology of the CCA. This is the first experimental
the samples from CCA at 32 1C after 2 days did not evidence to demonstrate how thermal stress on
generate different banding profiles to the samples coralline algae may influence coral recruitment
from the lower temperatures. Although the CCA at and this may have implications for reef regeneration
31 1C for 7 days clustered with the 2- and 7-day and recovery from climate-related mortality events.
samples from the lower temperatures, the 31 1C-CCA The thermal threshold for bleaching in N. fosliei
profile from the recovery period did clearly separate was similar to that of the GBR corals A. formosa,
from the other samples (Figure 8). A. elseyi and Pocillopora damicornis (Berkelmans
In terms of diversity and evenness of the clone and Willis, 1999). The visible bleaching at 31 1C and
libraries, there was no apparent trend with tempera- 32 1C was preceded by drops in Fv/Fm, which
ture. Rarefaction analysis indicated that between declined by 53% and 80%, respectively after 7 days
71% and 88% of the diversity was sampled in the compared with CCA maintained at 27 1C. This
various CCA biofilm treatments, but only 23–36% of reduction in Fv/Fm indicates potential chronic
the diversity was sampled in the sediment biofilm photoinactivation, caused by thermal and oxidative
libraries (Table 2; Supplementary Figures 2a and b). damage to (or loss of) the D1 reaction centre in PSII
The most diverse CCA library using both Chao1 and (Warner et al., 1999). In a recent study, the shallow
Ace indices was the 7-day 29 1C library followed by CCA species Porolithon onkodes from the GBR also
the 7-day 31 1C and 2-day 27 1C libraries (Table 2). demonstrated partial bleaching following exposure
The lowest-diversity CCA library was the 7-day to elevated SST (Anthony et al., 2008). N. fosliei is
27 1C treatment. abundant on the GBR at depths of 5–10 m, and may
To assess how CCA eukaryotic biofilms were be less adapted to cope with increases in SST. The
affected by temperature, DGGE was performed with loss of pigmentation in P. onkodes was accompanied
universal 18S rRNA gene primers. Although more by reductions in productivity (Anthony et al., 2008),
bands were detected in the 7 days and recovery possibly because of chronic photoinactivation as

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
8
CCA Heatmap at distance 0.03
2d 27°C 2d 29°C 2d 31°C 2d 32°C 7d 27°C 7d 29°C 7d 31°C 7d 32°C

0.099 0.117 0.127 0.134 0.140

Sediment Heatmap at distance 0.03

2d 27°C 2d 32°C 7d 27°C 7d 32°C

0.
062

0.062 0.073 0.079 0.083 0.087

Figure 6 OTU heat maps generated with a distance of 0.03 for CCA (a) and sediment (b) libraries. The 7-day 32 1C CCA library is a
distinct outlier and there is greater variability between the CCA derived libraries than the sediment-derived libraries.

described here. The slow recovery in Fv/Fm of CCA The primary Alphaproteobacteria strain to be lost
exposed to 31 1C in this study is consistent with the from CCA at 32 1C had highest sequence similarity to
gradual repair of PSII damaged by thermal stress; a described coral symbiont. The role of this microbe
however, this could not be confirmed in CCA discs within the CCA biofilm has not been ascertained,
exposed to 32 1C because of the presence of but there is some evidence that the loss of bacterial
endolithic green algae under the surface of the CCA. symbionts due to elevated SST can have significant
The microbial communities associated with implications on the health of the host (Webster et al.,
CCA shifted at both 31 1C and 32 1C, but the shift 2008a). The increase in Bacteroidetes is consistent
was much more rapid at the higher temperature. with previous studies of marine organisms subjected
Although the Alphaproteobacteria clearly domi- to stressors such as elevated SST or disease. In the
nates the CCA microbial community at ambient GBR sponge Rhopaloeides odorabile, the Bacteroi-
temperatures, there is a distinct shift towards the detes were absent from sponges maintained
Bacteroidetes in CCA exposed to 32 1C for 7 days. at temperatures up to 31 1C but dominated the

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
9
7 d 32ºC Webster et al., 2008b). The closest relative to the
Bacteroidetes sequence that became highly abun-
dant at 32 1C was a novel marine Flavobacteria with
anti-oxidative properties (Shindo et al., 2007).
7 d 27ºC The Bacteroidetes are diverse and abundant in the
marine ecosystem (O’Sullivan et al., 2006) and are
major utilizers of high-molecular mass dissolved
organic matter (Cottrell and Kirchman, 2000). For
7 d 31ºC this reason, they are often abundant in nutrient-rich
waters in which biomacromolecules accumulate
(Reichenbach, 1989). It is possible that heat-stressed
or decaying CCA cells provided an abundant and
varied source of nutrients for these bacteria, result-
2 d 32ºC
ing in an increased abundance of Bacteroidetes in
the 7-day 32 1C samples. The increased abundance
of a Deltaproteobacteria in CCA exposed to 32 1C for
7 days is consistent with studies showing an
7 d 29ºC increase in Delatproteobacteria in diseased corals
and sponges. The closest described relative to this
sequence was a Desulfovibrio spp. retrieved from an
unpublished study of diseased and healthy corals.
2 d 31ºC Closely related Desulfovibrio spp. have been im-
plicated in coral black band disease (AY750147) and
in black patch syndrome affecting the sponge
Alpysina aerophoba (EU267167), (Webster et al.,
2 d 29ºC 2008b). An alternative explanation for the commu-
nity differences observed in CCA at 32 1C may be
due to microbes associated with the green endolithic
algae that colonised the bleached CCA surface after
2 d 27ºC 5 days. Future research is urgently required to assess
0.1
the implications of these microbial shifts to CCA
Figure 7 Community tree constructed from an OTU assignment health and coral recruitment.
using a distance of 0.03. The cluster analysis shows similarity The bacterial biofilms on the surface of CCA
between each of the CCA-derived 16S rRNA gene clone libraries. exhibited a different response to elevated tempera-
ture than bacterial communities associated with reef
sediments, which were largely conserved at all
6 27°C temperatures. In addition, the microbial community
29°C
of the seawater from each of the CCA tanks showed
4 no correlation to the treatment temperature. This
31°C
indicates that either (1) the thermal stress response
32°C
Factor 2 (9%)

2 of the host CCA contributes to the shift in the

0d biofilm community or (2) the composition of the
0 1d biofilm on the surface of CCA is more sensitive to
2d environmental change (either because of a more
-2 specific association, as may occur in symbiosis, or
7d
due to the initial CCA bacterial community compo-
-4 Recovery sition containing more temperature-sensitive bio-
film species). For the large number of marine
-6
-6 -4 -2 0 2 4 6 8 10 12
invertebrates that settle and metamorphose in
response to sediment biofilms, this community
Factor 1 (13%)
stability suggests that these organisms may be less
Figure 8 Principal components analysis of CCA bacterial affected by increasing SST than those that rely on
community composition at each of the temperature treatments.
DGGE banding pattern data generated with primers specific to the
more sensitive bacterial biofilms associated with
Alphaproteobacteria and Bacteroidetes groups. biotic surfaces such as CCA. DGGE analysis showed
that the CCA biofilm community was unable to
recover its original composition within 7 days once
microbial community in sponges exposed to 33 1C the SST had been returned to ambient. It is possible
(Webster et al., 2008a). Similarly, diseased corals that the prolonged temperature stress (as occurred in
and sponges host a much greater abundance the 7-day experiment) caused a phase shift in the
of Bacteroidetes than their healthy counterparts microbial community to an alternate stable state.
(Frias-Lopez et al., 2002; Pantos and Bythell, 2006; Alternatively, the impact of elevated SST on CCA

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
10
Table 2 Diversity indices calculated from sequences of 16S rRNA genes using a 97% sequence similarity threshold

Treatment Total clones Unique OTUs Diversity sampled (%) Shannon–Weaver Chao1 Ace

CCA 27 1C 2 days 155 63 81 3.69 98 111

CCA 29 1C 2 days 135 53 80 3.48 76 95
CCA 31 1C 2 days 144 53 79 3.30 82 105
CCA 32 1C 2 days 171 51 88 3.21 76 77
CCA 27 1C 7 days 117 50 81 3.54 66 82
CCA 29 1C 7 days 157 70 72 3.73 174 179
CCA 31 1C 7 days 155 63 71 3.41 98 116
CCA 32 1C 7 days 142 51 76 3.23 93 126
Sed 27 1C 2 days 65 54 29 3.90 234 296
Sed 32 1C 2 days 62 55 23 3.97 196 244
Sed 27 1C 7 days 69 55 36 3.91 165 212
Sed 32 1C 7 days 64 54 29 3.93 202 215

Abbreviations: CCA, crustose coralline algae; OTU, operational taxonomic unit, Sed, sediment.

health may have adversely affected the ability of the to 32 1C for 7 days. The reduction in inductive
CCA immune system to regulate microbial colonisa- capacity at 32 1C may be due to a number of factors
tion and growth. including: (1) the loss of a critical microbial species
This study has highlighted the importance of on the surface of CCA that is directly responsible for
using multiple methods for assessing microbial the induction of metamorphosis (Johnson and
community responses. DGGE facilitated the analysis Sutton, 1994), (2) the loss of a microbial species
of large numbers of replicates, whereas clone required to enable the CCA to produce an inducer
libraries were utilised to explore the composition for larval metamorphosis (disruption of a shared
of the shifting microbes. LIBSHUFF analysis of 16S metabolic pathway), (3) the inhibition of a bio-
rRNA clone libraries found many more differences chemical process within the CCA alone to produce
than the DGGE analysis (for example, microbial an indicator for larval metamorphosis, (4) the
community differences in libraries at 27 1C and inhibition of larval metamorphosis by specific
29 1C). DGGE with the rpoB gene was unable to microbes (Dobretsov and Qian, 2004), or (5) an
distinguish the microbial shift that occurred at 32 1C interaction (for example, pathogenicity) between the
after 7 days. The results from 16S rRNA gene clone microbial community and the CCA (for instance
libraries and 16S rRNA gene DGGE with Alphapro- Deltaproteobacteria and Desulfovibrio in CCA
teobacteria and Bacteroidetes-specific primers were exposed to 32 1C have been identified in association
largely consistent, although the DGGE did not detect with diseased corals and sponges).
the shift at 32 1C after 2 days. This is possibly N. fosliei is a key reef-building primary producer
because of the fact that only the predominant on the GBR and other coral reefs (Harrington et al.,
microbes will be detected with this type of finger- 2004). The low-thermal tolerance of this species is
printing approach, whereas deep clone sequencing of concern as CCA are especially vulnerable to
can assess more subtle effects or the response of the increases in ocean acidification (pCO2), which is
‘rarer’ microbial community members. projected to intensify under continued conditions of
Interestingly, the elevated SST did not have an climate change (Kuffner et al., 2008). Combinations
impact on microbial community diversity or even- of elevated SSTs and acidification further enhance
ness. This is in contrast to other studies using CCA bleaching, productivity and reduce calcifica-
marine models in which stressors such as SST tion (Anthony et al., 2008; Martin and Gattuso,
(Webster et al., 2008a) eutrophication (Uthicke and 2009). Furthermore, CCA (including N. fosliei) are
McGuire, 2007) or disease (Cooney et al., 2002; critical to the recruitment of many coral species
Pantos et al., 2003; Pantos and Bythell, 2006; (Heyward and Negri, 1999; Harrington et al., 2004),
Webster et al., 2008b) cause a shift (either an including many Acropora species, which often
increase or decrease) in microbial diversity. A dominate reef habitats of the GBR. Any reduction
reduced microbial diversity was observed in sedi- in the capacity of CCA to induce coral larval
ments from an inshore reef with high inorganic metamorphosis by thermal stress, as described here,
carbon compared with a nearby offshore reef would clearly add to the pressures already faced by
(Uthicke and McGuire, 2007); however, a 53% corals under conditions of a changing climate.
increase in bacterial diversity was detected in a Although shifts in CCA-associated microbial com-
GBR sponge at 33 1C compared with the sponge munities correlate with reductions in coral larval
microbial community at 27 1C. metamorphosis, further work is required to deter-
A significant reduction in coral metamorphosis mine the role of microbiology in the recruitment
occurred in response to CCA that had been exposed process.

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
11
Acknowledgements transport and quenching of chlorophyll fluorescence.
Biochim Biophys Acta 990: 87–92.
We thank Dr Sven Uthicke for helpful discussions about Harrington LM, Fabricius K, De’ath G, Negri A. (2004).
data analysis and Dr Ken Wasmund for assisting with Recognition and selection of settlement substrata
sample processing. determine post-settlement survival in corals. Ecology
85: 3428–3437.
Heyward AJ, Negri AP. (1999). Natural inducers for coral
larval metamorphosis. Coral Reefs 18: 273–279.
Huber T, Faulkner G, Hugenholtz P. (2004). Bellerophon: a
References program to detect chimeric sequences in multiple
sequence alignments. Bioinformatic Applications Note
Anthony KRN, Kline DI, Diaz-Pulido G, Dove S, 20: 2317–2319.
Hoegh-Guldberg O. (2008). Ocean acidification causes Huggett MJ, Williamson JE, de Nys R, Kjelleberg S,
bleaching and productivity loss in coral reef builders. Steinberg PD. (2006). Settlement of the common
Proc Natl Acad Sci 105: 17442–17446. Australian sea urchin Heliocidaris erythrogramma in
Berkelmans R, Willis BL. (1999). Seasonal and local response to bacteria from the surface of coralline algae.
spatial patterns in the upper thermal limits of corals Oecologia 149: 604–619.
on the inshore Central Great Barrier Reef. Coral Reefs Huys G, Vanhoutte T, Vandamme P. (2008). Applications
18: 219–228. of sequence-dependent electrophoresis fingerprinting
Cooney RP, Pantos O, Le Tissier MD, Barer MR, in exploring biodiversity and population dynamics of
O’Donnell AG, Bythell JC. (2002). Characterization of human intestinal microbiota: What can be revealed?
the bacterial consortium associated with black band Interdiscipl Perspect Infect Dis 2008: 26pp.
disease in coral using molecular microbiological IPCC. (2007). Climate Change 2007: The Physical Basis.
techniques. Environ Microbiol 4: 401–413. Contribution of Working Group I to the Fourth
Cottrell MT, Kirchman DL. (2000). Natural assemblages of Assessment Report of the Intergovernmental
marine proteobacteria and members of the Cytophaga- Panel on Climate Change. Cambridge University Press
Flavobacter cluster consuming low- and high-mole- Cambridge: UK.
cular weight dissolved organic matter. Appl Environ Johnson CR, Muir DG, Reysenbach A-L. (1991). Character-
Microbiol 66: 1692–1697. istic bacteria associated with the surfaces of coralline
Dahllöf I, Baillie H, Kjelleberg S. (2000). rpoB-based algae: a hypotheses for bacterial induction of marine
microbial community analysis avoids limitations invertebrate larvae. Mar Ecol Prog Ser 74: 281–294.
inherent in 16S rRNA gene intraspecies heterogeneity. Johnson CR, Sutton DC. (1994). Bacteria on the surface of
Appl Environ Microbiol 66: 3376–3380. crustose coralline algae induce metamorphosis of the
Dobretsov S, Qian PY. (2004). The role of epibotic bacteria crown of thorns starfish Acanthaster planci. Mar Biol
from the surface of the soft coral Dendronephthya sp. 120: 305–310.
in the inhibition of larval settlement. J Exp Mar Biol Jones RJ, Muller J, Haynes D, Schreiber U. (2003). Effects
Ecol 299: 35–50. of herbicides diuron and atrazine on corals of the
Felsenstein J. (1993). PHYLIP (Phylogenetic Inference Great Barrier Reef, Australia. Mar Ecol Prog Ser 251:
Package) Version 3.5c. Department of Genetics, 153–167.
University of Washington: Seattle, WA. Kuffner IB, Andersson AJ, Jokiel PL, Rodgers K,
Ferris MJ, Muyzer G, Ward DM. (1996). Denaturing Mackenzie FT. (2008). Decreased abundance of
gradient gel electrophoresis profiles of 16S rRNA- crustose coralline algae due to ocean acidification.
defined populations inhabiting a hot spring microbial Nat Geosci 1: 114–117.
mat community. Appl Environ Microbiol 62: 340–346. Lawrence JR, Chenier MR, Roy R, Beaumier D, Fortin N,
Fierer N, Jackson JA, Bilgalys R, Jackson RB. (2005). Swerhone GD et al. (2004). Microscale and molecular
Assessment of soil microbial community structure by assessment of impacts of nickel, nutrients, and oxygen
use of taxon-specific quantitative PCR assays. Appl level on structure and function of river biofilm
Environ Microbiol 71: 4117–4120. communities. Appl Environ Microbiol 70: 4326–4339.
Frias-Lopez J, Zerkle AL, Bonheyo GT, Fouke BW. (2002). Lewis TE, Garland CD, McMeekin TA. (1985). The bacterial
Partitioning of bacterial communities between sea- biota on crustose (nonarticulated) coralline algae from
water and healthy, black-band diseased, and dead Tasmanian waters. Microb Ecol 11: 221–230.
coral surfaces. Appl Environ Microbiol 68: 2214–2228. Ludwig W, Strunk O, Klugbauer S, Klugbauer N, Weize-
Galbary D, Veltkamp CJ. (1980). Observations on Meso- negger M, Neumaier J et al. (1998). Bacterial phylo-
phyllum lichenoides (Corallinaceae, Rhodophyta) geny based on comparative sequence analysis.
with scanning electron microscope. Phycologia 19: Electrophoresis 19: 554–568.
49–53. Maidak BL, Cole JR, Parker Jr CT, Garrity GM, Larsen N, Li
Garland CD, Cooke SL, Grant JF, McMeekin TA. (1985). B et al. (1999). A new version of the RDP (Ribosomal
Ingestion of the bacteria on and the cuticle of crustose Database Project). Nucleic Acids Res 27: 171–173.
(non-articulated) coralline algae by post-larval Marchesi JR, Sato T, Weightman AJ, Martin TA, Fry JC,
and juvenile abalone (Haliotis ruber leach) from Hiom SJ et al. (1998). Design and evaluation of useful
Tasmanian waters. J Exp Mar Biol Ecol 91: 137–149. bacterium-specific PCR primers that amplify genes
Gelsomino A, Cacco G. (2006). Compositional shifts of coding for bacterial 16S rRNA [published erratum
bacterial groups in a solarised and amended soil as appears in Appl Environ Microbiol 1998 Jun; 64:
determined by denaturing gradient gel electro- 2333]. Appl Environ Microbiol 64: 795–799.
phoresis. Soil Biol Biochem 38: 91–102. Martin S, Gattuso JP. (2009). Response of Mediterranean
Genty B, Briantais JM, Baker NR. (1989). The relationship coralline algae to ocean acidification and elevated
between the quantum yield of photosynthetic electron temperature. Glob Change Biol 15: 2089–2100.

The ISME Journal

 Impact of temperature on CCA and microbes
NS Webster et al
12
Meyer-Reil L-A, Koster M. (2000). Eutrophication of Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M,
marine waters: effects on benthic microbial commu- Hollister EB et al. (2009). Introducing mothur:
nities. Mar Pollut Bull 41: 255–263. open-source, platform-independent, community-sup-
Morse ANC, Iwao K, Baba M, Shimoike K, Hayashibara T, ported software for describing and comparing
Omori M. (1996). An ancient chemosensory mec- microbial communities. Appl Environ Microbiol 75:
hanism brings new life to coral reefs. Biol Bull 191: 7537–7541.
149–154. Shindo K, Kikuta K, Suzuki A, Katsuta A, Kasai H,
Morse DE, Hooker N, Morse ANC, Jensen RA. (1988). Yasumoto-Hirose M et al. (2007). Rare carotenoids,
Control of larval metamorphosis and recruitment in (3R)-saproxanthin and (3R,20 S)-myxol, isolated from
sympatric agariciid corals. J Exp Mar Biol Ecol 116: novel marine bacteria (Flavobacteriaceae) and their
193–217. antioxidative activities. Appl Microbiol Biotech 74:
Morse DE, Morse AN, Raimondi PT, Hooker N. (1994). 1350–1357.
Morphogen-based chemical flypaper for Agaricia Sunagawa S, Woodley CM, Medina M. (2010). Threatened
humilis coral larvae. Biol Bull 186: 172–181. corals provide underexplored microbial habitats. PLoS
Negri AP, Heyward AJ. (2000). Inhibition of fertilization One 5: e9554.
and larval metamorphosis of the coral Acropora Uthicke S, McGuire K. (2007). Bacterial communities in
millepora (Ehrenberg, 1834) by petroleum products. Great Barrier Reef calcareous sediments: contrasting
Mar Pollut Bull 41: 420–427. 16S rDNA libraries from nearshore and outer shelf
Negri AP, Webster NS, Hill RT, Heyward AJ. (2001). reefs. Est Coast Shelf Sci 72: 188–200.
Metamorphosis of broadcast spawning corals in Warner ME, Fitt WK, Schmidt GW. (1999). Damage to
response to bacteria isolated from crustose algae. photosystem II in symbiotic dinoflagellates: a
Mar Ecol Prog Ser 223: 121–131. determinant of coral bleaching. P Natl Acad Sci 96:
O’Sullivan LA, Rinna J, Humphreys G, Weightman AJ, 8007–8012.
Fry JC. (2006). Culturable phylogenetic diversity of the Webster NS, Battershill CN, Negri AP. (2006). Recruitment
phylum ‘Bacteroidetes’ from river epilithon and coast- of Antarctic marine eukaryotes onto artificial surfaces.
al water and description of novel members of the Polar Biol 30: 1–10.
family Flavobacteriaceae: Epilithonimonas tenax gen. Webster NS, Cobb RE, Negri AP. (2008a). Temperature
nov., sp. nov. and Persicivirga xylanidelens gen. nov., thresholds for bacterial symbiosis with a sponge.
sp. nov. Int J Syst Evol Microbiol 56: 169–180. ISME J 2: 830–842.
Osmond CB, Anderson JM, Ball MC, Egerton JJG. (1999). Webster NS, Hill RTH. (2007). Vulnerability of marine
Compromising efficiency: the molecular ecology microbes on the Great Barrier Reef to climate change.
of light resource utilisation in terrestrial plants. In: Johnson JE, Marshall PA (eds). Climate Change and
In: Scholes C, Baker M (eds). Advances in Physiolo- the Great Barrier Reef. Great Barrier Reef Marine Park
gical Plant Ecology. Blackwell: Oxford. Authority and Australian Greenhouse Office: Austra-
Pantos O, Cooney RP, Le Tissier MD, Barer MR, O’Donnell lia, Townsville.
AG, Bythell JC. (2003). The bacterial ecology of a Webster NS, Smith LD, Heyward AJ, Watts JE, Webb RI,
plague-like disease affecting the Caribbean coral Blackall LL et al. (2004). Metamorphosis of a scler-
Montastrea annularis. Environ Microbiol 5: 370–382. actinian coral in response to microbial biofilms.
Pantos O, Bythell JC. (2006). Bacterial community struc- Appl Environ Microbiol 70: 1213–1221.
ture associated with white band disease in the elkhorn Webster NS, Xavier JR, Freckelton M, Motti CA, Cobb R.
coral Acropora palmata determined using culture (2008b). Shifts in microbial and chemical patterns
independent 16S rRNA techniques. Dis Aquat Org within the marine sponge Aplysina aerophoba during
69: 79–88. a disease outbreak. Environ Microbiol 10: 3366–3376.
Reichenbach H. (1989). Nonphotosynthetic, nonfruiting White TJ, Bruns T, Lee S, Taylor J. (1990). Amplification
gliding bacteria. Genus 1. Cytophaga Winogradsky and direct sequencing of fungal ribosomal RNA genes
1929, 577AL, emend. In: Staley JT, Bryant MP, Pfennig for phylogenetics. In: Innis MA, Gelfand DH,
N, Holt JG (eds). Bergey’s Manual of Systematic Sninsky JJ, White TJ (eds). PCR Protocols: A Guide to
Bacteriology. Baltimore. Methods and Applications. Academic Press: New
Rodriguez SR, Ojeda FP, Inestrosa C. (1993). Settlement of York, pp 315–322.
benthic marine invertebrates. Mar Ecol Prog Ser 97: Wieczorek SK, Todd CD. (1997). Inhibition and facilitation
193–207. of bryozoan and ascidian settlement by natural multi-
Schloss PD, Handelsman J. (2005). Introducing DOTUR, a species biofilms: effects of film age and the roles of
computer program for defining operational taxonomic active and passive larval attachment. Mar Biol 128:
units and estimating species richness. Appl Environ 463–473.
Microbiol 71: 1501–1506. Wieczorek SK, Todd CD. (1998). Inhibition and faci-
Schloss PD, Larget BR, Handelsman J. (2004). Integration litation of settlement of epifaunal marine invertebrate
of microbial ecology and statistics: a test to compare larvae by microbial biofilm cues. Biofouling 12:
gene libraries. Appl Environ Microbiol 70: 5485–5492. 81–118.

Supplementary Information accompanies the paper on The ISME Journal website (http://www.nature.com/ismej)

The ISME Journal