Temporal dynamics of prokaryotic communities in themarine sponge Sarcotragus spinosulus

CRISTIANE C. P. HARDOIM and RODRIGO COSTA

Microbial Ecology and Evolution Research Group, Centre of Marine Sciences (CCMar), University of Algarve (UAlg),

Gambelas, 8005-139 Faro, Portugal

Abstract

In spite of their putative relevance to host functioning, in-depth knowledge of sponge

microbiome stability over time is scarce. This study tackles the temporal maintenance

of bacterial and archaeal assemblages in the model host Sarcotragus spinosulus along

three successive years. Prokaryotic communities were profiled by polymerase chain

reaction-denaturing gradient gel electrophoresis (PCR-DGGE) and 454-pyrosequencing

of S. spinosulus-derived 16S rRNA gene amplicons. Prevailing bacterial phyla were

Actinobacteria, Acidobacteria, Proteobacteria, Poribacteria, PAUC34f, Chloroflexi and

Bacteroidetes, with Bacteroidetes, Chloroflexi and Poribacteria showing different abun-

dances over the years. At the approximate species level (operational taxonomic units,

OTUs, defined at 97% sequence similarity), no major changes in bacterial richness and

composition were found through time. Nearly 50% of all detected bacterial symbionts

(96 in 205 OTUs) were recovered from all sampling years, whereas a taxonomically

equivalent community of less dominant bacteria characterized the transient sponge

microbiota. Despite the evidence for temporal symbiont maintenance, an intriguing

cumulative degree of variation between individuals was unravelled, with all the sur-

veyed sponge specimens sharing only 27 bacterial OTUs. Archaeal communities were

dominated by one single symbiont of the candidate genus Nitrosopumilus (Thau-marchaeota), known for its ability to aerobically oxidize ammonia to nitrite. Only few

bacterial ammonia oxidizers consistently occurred in S. spinosulus across the years as

documented by PCR-DGGE fingerprinting. In conclusion, prokaryotic symbionts of

S. spinosulus display a state of dynamic stability shaped by the interplay between the

maintenance of dominant players and turnover of less prevalent community members,

in time and across host individuals, with no apparent consequences to holobiont

functioning.

Keywords: Ammonia oxidizers, host–microbe interactions, microbial diversity, next-generation

sequencing, symbiosis

Received 10 February 2014; revision received 30 April 2014; accepted 5 May 2014

Introduction

Marine sponges harbour complex and diverse prokary-

otic assemblages. Based on 16S rRNA gene analyses, 28

bacterial phyla have thus far been identified in close

association with these animals (Taylor et al. 2007;

Hentschel et al. 2012; Schmitt et al. 2012), of which

Proteobacteria (especially Alpha and Gamma classes),

Actinobacteria, Cyanobacteria, Bacteroidetes, Acidobacteria,

Poribacteria and Chloroflexi are among the most dominant.

Sequences from the two main archaeal phyla—

Thaumarchaeota and Euryarchaeota—have also been

recovered from marine sponges (Taylor et al. 2007; Simis-

ter et al. 2012; Webster & Taylor 2012). A comprehensive

phylogenetic analysis carried out with more than 7500

sponge-derived 16S rRNA gene sequences revealed 173

monophyletic ‘sponge-specific’ bacterial clusters as well

as five monophyletic ‘sponge-specific’ archaeal clusters

(Simister et al. 2012) according to the definition ofCorrespondence: Rodrigo Costa, Fax: (+351) 289 800 069;

E-mails: rscosta@ualg.pt; microecoevo@gmail.com

© 2014 John Wiley & Sons Ltd

Molecular Ecology (2014) doi: 10.1111/mec.12789

Hentschel et al. (2002). Recently, the application of high-

throughput sequencing technology has shown that many

of the so-called ‘sponge-specific’ microbial lineages occur

in rare numbers in the open environment, being notice-

ably enriched within the sponge body (Taylor et al. 2013).

Sponge prokaryotic symbionts probably mediate the

cycling of several chemical elements (Taylor et al. 2007).

In particular, the first nitrification step within the nitro-

gen cycle, that is, the conversion of ammonia to nitrite

performed by ammonia-oxidizing archaea and bacteria

(AOA and AOB, respectively), is of supposed relevance

for marine sponges because it could aid the host in the

process of metabolic waste removal (Taylor et al. 2007;

Steger et al. 2008). The subunit A of the ammonia mono-

oxygenase gene (amoA gene) has been extensively

applied as a genetic marker to detect prokaryotic ammo-

nia oxidizers in several ecosystems (Rotthauwe et al.

1997; Francis et al. 2005), including the marine sponge

holobiome where AOA seem to prevail over AOB (Bayer

et al. 2008; Hoffmann et al. 2009; Cardoso et al. 2013).

However, little is known about the spatial and temporal

dynamics of these functional cohorts in the sponge host.

In the last 10 years or so, we experienced a sharp

increase in knowledge concerning the diversity and func-

tion of sponge symbiont communities (Taylor et al. 2007;

Hentschel et al. 2012; Webster & Taylor 2012). However,

our perspective of their temporal dynamics remains

limited as the majority of sponge microbiology surveys

undertaken so far relied on single sampling events.

Regardless of their diversity, unveiling the spatiotempo-

ral stability of the marine sponge microbiome is central

not only to our understanding of host–symbiont co-evo-

lutionary relationships, but also to the management of

marine resources of potential biotechnological use and

relevance to global biogeochemistry. Overall, studies that

have thus far addressed the marine sponge microbiome

along a time series proposed that the bacterial commu-

nity in these hosts was temporally stable, from 6-month

to 3-year periods (Taylor et al. 2004; Thiel et al. 2007;

Erwin et al. 2012; White et al. 2012). With one exception

(White et al. 2012) these surveys used more traditional

molecular tools to characterize the communities, such as

rRNA gene clone libraries or polymerase chain reaction-

denaturing gradient gel electrophoresis (PCR-DGGE)/

terminal restriction fragment length polymorphism (T-

RFLP) fingerprinting. With ever-evolving resolving

power, next-generation sequencing technologies are

needed to overcome depth biases of previous protocols

and will soon certainly enable novel insights into sponge

microbiome dynamics across time.

In this study, the hypothesis of temporal stability of

prokaryotic communities in the temperate sponge Sar-

cotragus spinosulus Schmidt, 1862 (Demospongiae, Irci-

niidae) is tested over a 3-year period. Sarcotragus

spinosulus is a high-microbial-abundance sponge hosting

distinct bacterial communities in comparison with those

in the environmental surroundings or inhabiting other

sponge hosts (Hardoim et al. 2012). Bacteria isolated

from this model species have shown bioactivities in vi-

tro that highlight their potential use in biotechnological

applications (Esteves et al. 2013). Here, the 16S rRNA

gene was used as a phylogenetic marker in PCR-DGGE

and 454-pyrosequencing profiling of the Bacteria and

Archaea domains, and the amoA gene was used as a tar-

get to PCR-DGGE fingerprint prokaryotic ammonia oxi-

dizers associated with S. spinosulus over time.

Material and methods

Sponge sampling

Four specimens of Sarcotragus spinosulus (Schmidt, 1862;

Demospongiae, Irciniidae) were randomly collected by

scuba-diving at depths around 15 m at Gal�e Alta,

Armac�~ao de Pera (37° 040 09.6″N and 8° 190 52.1″W) off

the coast of the Algarve, South Portugal, in June 2010,

September 2011 and October 2012. Samples were placed

in plastic bags (type Ziploc�) underwater, transported

to the laboratory inside cooling boxes and processed

upon arrival. Sampling procedure was followed as

given previously (Hardoim et al. 2012).

Sponge identification

Traditional sponge classification was performed as

explained elsewhere (Hardoim et al. 2012). Sponge

molecular phylogeny was applied to aid in the identifi-

cation of the specimens. To this end, total community

DNA was extracted from the sponge samples and used

for PCR amplification of the subunit I of the cyto-

chrome oxidase gene (CO1) with the primer pair

dgLCO1490 and dgHCO2189 (Meyer et al. 2005;

Table 1). PCR conditions, amplicon sequencing and

phylogenetic inference were carried out following the

procedures of Hardoim et al. (2012).

Total community DNA extraction

Metagenomic DNA from 0.25 g of the internal sponge

body was extracted using the UltraClean� Soil DNA

isolation kit (Mo Bio, Carlsbad, CA, USA) according to

the manufacturer’s protocol.

Preparation of amplicons for PCR-DGGEfingerprinting

Bacterial 16S rRNA gene. A nested PCR-DGGE approach

was selected to fingerprint bacterial communities in

© 2014 John Wiley & Sons Ltd

2 C. C. P . HARDOIM and R. COSTA

S. spinosulus using the primer pairs F27-R1492

(first PCR, Table 1) and F984-GC-1378R (second

PCR, Table 1) as described previously (Hardoim et al.

2012).

Archaeal 16S rRNA gene. A nested PCR-DGGE approach

was also applied to profile archaeal symbionts inhabit-

ing S. spinosulus. The first PCR reaction mixture was

prepared with the Archaea-specific primer pair

ARC344f-mod–Arch958R-mod (Table 1) as described by

Pires et al. (2012), using 3.75 mM MgCl2, 0.1 mg mL�1

of bovine serum albumin, 2% (vol/vol) dimethyl

sulphoxide and 0.625 U of BioTaqTM DNA polymerase

(Bioline, London, UK). The reaction mixture was pre-

pared as explained by Pires et al. (2012) with the primer

pair 524F-10–Arch958R-mod (GC) (Table 1) using 2 lLof the previous amplicons as templates, 3.75 mM MgCl2and 0.625U of BioTaqTM DNA polymerase. Thermal

cycling in both PCRs was as described by Pires et al.

(2012).

Bacterial amoA gene. A semi-nested PCR-DGGE was

chosen to address ammonia-oxidizing bacteria associ-

ated with S. spinosulus. For the first PCR, the reaction

mixture (25 lL) contained 1.0 lL of template DNA

(~10 ng), 19 reaction buffer, 0.2 mM dNTPs, 3.75 mM

MgCl2, 4% (vol/vol) acetamide, 0.2 lM of primer pair

amoA1F–amoA2R-GG (Nicolaisen & Ramsing 2002) and

0.625 U of BioTaqTM DNA polymerase. After initial

denaturation at 92 °C for 1 min, 35 cycles of 30 s at

92 °C, 30 s at 57 °C and 60 s at 72 °C were performed.

A final extension of 5 min at 72 °C was used to finish

the reaction. The obtained amplicons [3.0–4.5 lL] were

used as templates in a second PCR-DGGE with the

primer pair amoA1F-GC–amoA2R-GG (Nicolaisen &

Ramsing 2002; Table 1) in 15 thermal cycles using the

Table 1 Primer pairs used in this study

Primer Sequence (50–30) Target Usage Reference

dgLCO1490 GGTCAACAAATCA

TAAAGAYATYGG

Cytochrome

oxidase gene

Sponge

phylogeny

Meyer et al. (2005)

dgHCO2189 TAAACTTCAGGGTGAC

CAAARAAYCA

F27 AGAGTTTGATCMTGG

CTCAG

Bacteria 16S

rRNA gene

First DGGE

PCR

Weisburg et al. (1991)

R1492 TACGGYTACCTTGTT

ACACTT

F984-GC GC Clamp-AACGCGAAG

AACCTTAC

Bacteria 16S

rRNA gene

Second DGGE

PCR

Heuer et al. (1997)

R1378 CGGTGTGTACAAGGCC

CGGGAACG

ARC344f-mod ACGGGGYGCASSAG

KCGVGA

Archaea 16S

rRNA gene

First DGGE

PCR

Pires et al. (2012)

Arch958R-mod YCCGGCGTTGAVTCCAATT

524F-10 GCCGCGGTAA Archaea 16S

rRNA gene

Second DGGE

PCR

Pires et al. (2012)

Arch958R-mod

(GC)

GC Clamp-CCGGCGTT

GAVTCCAATT

amoA1F GGGGTTTCTACTGGTGGT Bacteria

amoA gene

First DGGE

PCR

Nicolaisen & Ramsing (2002)

amoA2R-GG CCCCTCGGGAAAGC

CTTCTTC

Bacteria

amoA gene

First/second

DGGE PCR

Nicolaisen & Ramsing (2002)

amoA1F-GC GC Clamp-GGGGTTTCTA

CTGGTGGT

Bacteria

amoA gene

Second DGGE

PCR

Nicolaisen & Ramsing (2002)

Crenamo23f ATGGTCTGGCTWAGACG Archaea

amoA gene

DGGE PCR Tourna et al. (2008)

CrenamoA616r GCCATCCATCTGTATGTCCA

V4_titF AYTGGGYDTAAAGNG Bacteria 16S

rRNA gene

454-pyrosequencing http://pyro.cme.msu.edu/pyro/help.jsp#intro

V4_tit_R TACNVRRGTHTCTAATYC

524F-10-ext TGYCAGCCGCCGCGGTAA Archaea 16S

rRNA gene

454-pyrosequencing Pires et al. (2012)

Arch958R-mod YCCGGCGTTGAVTCCAATT

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 3

same reaction mixture and cycling conditions described

for the first PCR.

Archaeal amoA gene. One single step was used to

amplify archaeal amoA gene fragments for PCR-DGGE

profiling. The reaction mixture was prepared as for the

first PCR used in the amplification of the bacterial amoA

gene, except for the primer pair Crenamo23f–Crena-

moA616r (Tourna et al. 2008; Table 1) and template

DNA quantity (~30 ng). After initial denaturation at

95 °C for 5 min, 35 cycles of 30 s at 92 °C, 30 s at 55 °Cand 60 s at 72 °C were performed, followed by a final

extension of 10 min at 72 °C.

PCR-DGGE profiling and analysis

Polymerase chain reaction–denaturing gradient gel elec-

trophoresis was carried out in a PhorU-2 gradient sys-

tem (Ingeny International, Goes, the Netherlands). The

16S rRNA and amoA gene amplicons were applied in

even concentrations onto polyacrylamide gels

containing a 40–75% gradient of denaturants (100%

denaturants defined as 7M urea and 40% formamide)

and 6% of acrylamide, except for the bacterial amoA

gene, where 8% acrylamide was used. Marker constitu-

ents, electrophoresis conditions and staining procedures

were described previously (Hardoim et al. 2012). The

PCR-DGGE profiles were processed with the software

GelCompar II 5.1 (Applied Maths, Kortrijk, Belgium) as

explained by Hardoim et al. (2009). This analysis

delivered a table species vs. samples encompassing the

relative abundance of all bands in each profile, which

was further used as input for ordination analyses of

PCR-DGGE fingerprints with CANOCO for Windows 4.5

(Microcomputer Power, Ithaca, NY, USA) as described

by Costa et al. (2006), using Hellinger-transformed

abundance data.

Preparation of samples for pyrosequencing

A bar-coded pyrosequencing method was applied for in-

depth analysis of bacterial and archaeal community com-

position and diversity. A thorough description of (i) py-

rosequencing sample preparation, (ii) data processing

and (iii) analysis is provided in Appendix S1 (Supporting

information). Briefly, the V4 hypervariable region of bac-

terial 16S rRNA genes was PCR-amplified using the

Ribosomal Database Project primer set (Table 1), which

generates amplicons of around 248 bp in length. For

Archaea, the V4–V5 hypervariable region of the 16S rRNA

gene was targeted and two independent PCR amplifica-

tions (25 lL) were performed for each sample. Ampli-

cons from the first archaeal PCR (Table 1) were used as

templates (2 lL) in the pyrosequencing reaction, which

was carried out with 25 cycles using the primers 524F-10-

ext–Arch958R-mod (Table 1) and conditions as described

by Pires et al. (2012). Bacterial and archaeal amplicons

were delivered for pyrosequencing on a 454 Genome

Sequencer GS FLX Titanium platform (Roche Diagnostics

Ltd, West Sussex, UK) at Biocant (Biotechnology Innova-

tion Center, Portugal).

Pyrosequencing data processing and analysis

454-pyrosequencing raw data were subjected to quality

filtering and removal of homopolymers and chimeras

with AmpliconNoise (Quince et al. 2011). Analyses of fil-

tered sequences were carried out as explained in detail in

Appendix S1 (Supporting information), using the QUANTI-

TATIVE INSIGHTS INTO MICROBIAL ECOLOGY (QIIME) software

package (Caporaso et al. 2010). In summary, taxonomic

assignments of bacterial and archaeal sequences

were inferred with the greengenes 13_05 database

release (http://greengenes.secondgenome.com/downlo-

ads/database/13_5) within the QIIME environment. Opera-

tional taxonomic units (OTUs) were defined at ≥97% 16S

rRNA gene sequence similarity. A final OTU vs. samples

table was generated for both prokaryotic domains after

the removal of unclassified OTUs, chloroplasts and sin-

gletons and was used for downstream analyses. These

comprised (i) phylum- and class-level composition in

individual and pooled samples, (ii) determination of

specific and common symbionts across years by OTU

networking and Venn diagrams, (iii) estimates of sym-

biont richness (Chao1) and diversity (Shannon’s index)

and (iv) multivariate analysis of OTU data. The latter

was achieved by (a) unconstrained UPGMA clustering

and principal coordinate analysis (PCoA) of OTU

profiles using the Unifrac metric within QIIME and (b)

constrained Redundancy Analysis (RDA) of OTU profiles

and environmental variables (i.e. sampling years) with the

software package CANOCO 4.5 using Hellinger-transformed

OTU abundance data. Analyses were performed using

full-size and size-normalized, quality-filtered sample

libraries, hereafter called, for each prokaryotic domain,

‘full’ and ‘normalized’ data sets, respectively.

Tests of significance

Homogeneity of variance tests were used to check the

normal distribution of the richness and diversity mea-

surements from PCR-DGGE fingerprints and 454-py-

rosequencing. Analysis of variance (ANOVA) tested

whether or not the mean values obtained for all sample

groups were equal. A pairwise t-test—which analyses

the significance between groups, in our case distinct

sampling years—was then carried out. Homogeneity

of variance and ANOVA were also used to compare the

© 2014 John Wiley & Sons Ltd

4 C. C. P . HARDOIM and R. COSTA

tag-pyrosequencing relative abundances of the most

dominant bacterial phyla and classes found in S. spinosu-

lus across the years. All the above-mentioned analyses

were performed with the stat package in R programming

(R Development Core Team 2012). For both PCR-DGGE

and 454-pyrosequencing data, Monte Carlo permutations

were run to test whether the generated sponge symbiont

profiles clustered according to the sampling year.

Results

Sponge identification

Analysis of 579-bp-long CO1 gene sequences obtained

for all 12 sponge specimens inspected in this study and

the other three remaining Sarcotragus spp. sequences

available at NCBI showed a high level of gene conser-

vation within the genus, with genetic distances (p-dis-

tance) between pairs of sequences ranging from 0% to

0.003%. This well-supported group was clearly placed

apart from the phylogenetic clusters representing the

other two genera of the family Irciniidae, namely Ircinia

and Psammocinia (Fig. 1).

PCR-DGGE analyses

Regardless of the microbial cohort examined—that is,

Bacteria, Archaea or ammonia-oxidizing bacteria (AOB)

—PCR-DGGE profiling revealed that prevailing bands

were usually detected in all sampling years/sponge

specimens, while fainter bands presented a varied

pattern of occurrence across the profiles (Fig. S1,

Supporting information). Bacterial profiles consisted of

nine dominant and 11–36 fainter bands, whereas archa-

eal and AOB fingerprints were less complex, showing

two dominant bands along with few other detectable

bands across the samples. Overall, only a minor portion

of the total PCR-DGGE band data variation within each

microbial group could be attributed to the factor ‘year

of sampling’. In spite of this, ordination analysis of

PCR-DGGE fingerprints suggested a subtle transition in

community structures through time, characterized by

significant differences between sponge specimens sam-

pled in 2012 and those collected in 2010 and 2011 (Fig.

S1). For the archaeal amoA gene, no amplification was

obtained from any of the S. spinosulus replicates. A

thorough description of PCR-DGGE results is given in

Appendix S2 (Supporting information).

454-pyrosequencing analyses

Bacterial 16S rRNA gene data set. In total, 86 639 bacte-

rial 16S rRNA V4-tag sequences were obtained after

preliminary filtering on the 454 equipment. Noise

filtering and trimming with AmpliconNoise and

Galaxy, respectively, delivered 77 125 16S rRNA V4-tag

sequences further analysed with QIIME. After passing the

script that excluded unclassifiable, chloroplast and sin-

gleton OTUs from the final OTU table, 71 404 sequences

were assigned to 205 OTUs at 97% sequence similarity

(Table S1, Supporting information). In the following,

results obtained using full-size (‘full data set’) and nor-

malized sequencing libraries set at 3155 sequence reads

per sample (‘normalized data set’) are presented. One

replicate from 2012 (Alg12/83) was removed from the

normalized comparisons because it did not reach the

chosen sequence threshold (Table S1).

Bacterial community composition at high taxonomic

ranks. Actinobacteria (average relative abundance of

20.95% across all sampled specimens), Acidobacteria

(20.30%), Proteobacteria (13.58%), Poribacteria (11.45%),

PAUC34f (9.97%), Chloroflexi (9.75%), Bacteroidetes

(5.74%) and AncK6 (3.86%), together corresponding to

95.63% of all quality-filtered sequences analysed,

dominated the bacterial community associated with

S. spinosulus (Fig. 2a,b, full data set). Among these

phyla, only Chloroflexi and Bacteroidetes showed signifi-

cant differences (P < 0.05) in relative abundances over

the years (Fig. 2b). The largest variability in phylum

abundances between sponge individuals was observed

for PAUC34f and Acidobacteria within 2012 specimens,

with values ranging from 6.7% to 32.6% and from 9.4%

to 33.1%, respectively (Fig. 2a). One of the lowest levels

of variability between individuals was detected for the

phylum Acidobacteria in 2010 (from 16.5% to 19.1%,

Fig. 2a). Regarding bacterial classes within phyla, Acidi-

microbiia dominated the pool of Actinobacteria hits in the

sequence libraries. Likewise, Sva0725, Sphingobacteriia

and Gammaproteobacteria were the most abundant clas-

ses within the Acidobacteria, Bacteroidetes and Proteobacte-

ria phyla, respectively (Fig. 2c,d). Highest and lowest

variations in relative abundance across individuals were

registered for the class Sva0725 of the Acidobacteria

(from 6% to 28.7% in 2012) and the Deltaproteobacteria

within the Proteobacteria (from 0.5% to 3.5% in 2012),

respectively (Fig. 2c). The above-mentioned trends were

reproduced in the normalized data set, with the excep-

tion that Poribacteria, instead of Bacteroidetes, displayed

significant abundance changes over the years besides

Chloroflexi (Fig. S2, Supporting information). For both

full and normalized data sets, no significant shifts in

OTU richness were observed within the Bacteroidetes,

Chloroflexi and Poribacteria phyla in spite of their abun-

dance shifts within the study period.

Bacterial richness and diversity. For adequate quantitative

comparisons, here we show only results obtained

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 5

with the normalized data set (3155 sequences per

sample). Under this sequencing depth, the observed

bacterial richness found in S. spinosulus in 2010

(105 � 3.44), 2011 (94 � 12.5) and 2012 (95 � 8.33) did

not significantly differ from one another (P > 0.05,

Fig. S3a, Supporting information). The Shannon diver-

sity indices of 2010 (4.83 � 0.23) and 2011

(4.71 � 0.22) significantly differed (P < 0.05) from

the measurements observed in 2012 (4.18 � 0.37)

(Fig. S3b).

Specificity and sharedness of bacterial OTUs. The assign-

ment of all 205 bacterial OTUs detected in this study to

their source samples—that is, the 12 replicates of S. spi-

nosulus in the full data set—was depicted in an OTU

network (Fig. 3a). The majority of the bacterial OTUs

Fig. 1 Phylogenetic inference of the Irciniidae family based on cytochrome oxidase gene subunit 1 sequences. The maximum likeli-

hood tree (-ln likelihood: 1204.8294) is presented, with sequences recovered in this study highlighted in bold. Numbers at tree nodes

are bootstrap values and posterior probabilities calculated in Maximum Likelihood and MCMC Bayesian analyses, respectively,

and values above 75/0.95 are shown. Scale bar: nucleotide substitutions per site. The tree is rooted in the genus Rhopaloeides

(Dictyoceratida, Spongiidae).

© 2014 John Wiley & Sons Ltd

6 C. C. P . HARDOIM and R. COSTA

were common to two or more S. spinosulus specimens,

and thus only few were found to be exclusively associ-

ated with each sponge individual. When the biological

replicates from the same year were pooled, the bacterial

OTU network revealed that, although the majority of

OTUs detected in S. spinosulus in any given year could

be resampled in another year, still a considerable

amount of ‘year-specific’ OTUs was unveiled (Fig. 3b).

(a)

(b)

(d)

(c)

Fig. 2 Phylum- (a, b) and class-level (c, d) bacterial community composition in Sarcotragus spinosulus in three consecutive years using the

full data set. The compositions of each replicate sample (a, c) and of pooled replicate samples (b, d) are shown. Asterisks on bars denote

dominant taxa displaying significant shifts in relative abundance through time. See Fig. S2 for results with the normalized data set.

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 7

Year-to-year maintenance and variation in bacterial

OTUs in S. spinosulus were further investigated and

precisely quantified with Venn diagrams (Fig. 3c,d).

These analyses revealed that 34, 19 and 16 OTUs were

found exclusively in 2010, 2011 and 2012, respectively,

comprising 69 year-specific OTUs (Table S2, Supporting

information), whereas a greater ‘temporal core’ was

observed comprising 96 OTUs common to all sampling

years (Fig. 3c), but not necessarily to all sponge indi-

viduals. Interestingly, the bacterial OTUs exclusively

detected in each sampling year were classified into

diverse phyla (Table S2), of which several are typical

constituents of the sponge microbiome such as Acido-

bacteria, Actinobacteria, PAUC34f and Poribacteria. These

OTUs usually contained few sequences (from 2 to 49)

mirroring ‘rare’ S. spinosulus symbionts, which collec-

tively represented only 0.85% (611 in 71 404) of all

analysed sequences (Table S2) and displayed high simi-

larity with other sponge-derived sequences present in

public databases. When only OTUs containing at least

50 sequences were considered, the total number of

analysed OTUs dropped from 205 to 78, of which 70

OTUs comprised the temporal bacterial core in S. spi-

nosulus, whereas the number of year-specific OTUs

became negligible (Fig. 3d). In spite of the presumed

sensitivity of this approach to library sizes, we

observed only a slight shift in—and no difference in

the proportions of—the numbers of year-specific vs.

temporally stable OTUs when Venn diagrams and net-

works were constructed using the normalized data set

(Fig. S4, Supporting information). The ‘pan bacteriome’

associated with S. spinosulus, hereby defined as the total

number of bacterial phylotypes detected across all

analysed specimens (205 OTUs, full data set), was com-

posed of 19 formally recognized and candidate bacterial

phyla (Table 2, Fig. S5a, Supporting information). In

contrast, the ‘essential bacterial core’ of the sponge, that

is, the pool of symbionts common to all 12 S. spinosulus

specimens, consisted of only 27 OTUs belonging to

eight archetypical sponge-associated bacterial phyla

and one candidate phylum (Table 3, Fig. S5b). Remark-

ably, these OTUs comprised about 65% of the full

S. spinosulus bacterial community in terms of abun-

dance (46 209 in 71 404 sequences, Table 3). Because

we explored the full data set to depict the pan and

essential bacteriomes of S. spinosulus, the magnitude of

the former bacteriome is maximized as it represents all

OTUs detected across the 12 specimens, whereas the

dimension of the latter is minimized as OTU common-

ality data are limited by the smallest library size

(Alg12/83, Table S1).

Bacterial community ordination. Principal coordinate analy-

sis revealed no clear pattern of community composition

variation along time (Fig. 4a, normalized data set).

Indeed, several replicates from different years are

proximate from each other in the ordination diagram,

with only a few observed sample outliers. Figure 4a

(a) (b)

(d)(c)

Fig. 3 Bacterial OTU networking and

maintenance in the marine sponge Sar-

cotragus spinosulus across time. Networks

made with all OTUs found per sponge

replicate (a) and with all OTUs found in

composite samples where replicates were

pooled according to the sampling year

(b) and Venn diagrams (c) with all OTUs

and (d) with OTUs containing ≥50sequences per sampling year are shown.

In the networks, red, green and purple

lines correspond to years 2010, 2011 and

2012, respectively. See Fig. S4 for results

with the normalized data set.

© 2014 John Wiley & Sons Ltd

8 C. C. P . HARDOIM and R. COSTA

also shows the distribution of the 10 most abundant

phyla across the ordination space, whereby their central

and overlapping positions reveal the absence of taxon-

sampling year correlations. Similarly, constrained ordi-

nation applied to the full data set showed that only

23.8% of the total data set variation could be attributed

to the factor ‘year of sampling’ and that this factor did

not influence the patterns of OTU distribution across the

bacterial 16S rRNA V4-tag sequences (P > 0.05, Fig. 4b).

In spite of this, by ‘stretching’ the placement of samples

along the horizontal axis of the diagram, which repre-

sents 66% of the data set variability explained by the

sampling year, RDA ordination hints at a slight transi-

tion in community structures from 2010 to 2012 bridged

by the 2011 samples (Fig. 4b), resembling the patterns

obtained via ordination of PCR-DGGE profiles (Fig. S1).

Similar observations were made when RDA ordination

was applied to the normalized data set (Fig. S6,

Supporting information).

Archaeal 16S rRNA gene data set. Overall, 64 025 archa-

eal 16S rRNA V4–V5-tag sequences were generated by

Table 2 The pan bacteriome of Sarcotragus spinosulus

Phylum OTUs* Sequences† Class OTUs* Sequences†

Acidobacteria 17 14 501 Acidobacteria-6 6 2013

BPC102 1 44

PAUC37f 1 661

Solibacteres 5 2606

Sva0725 4 9177

Actinobacteria 11 14 381 Acidimicrobiia 8 14 374

Actinobacteria 1 2

Coriobacteriia 1 3

Rubrobacteria 1 2

AncK6 2 2706 n.a. 2 2706

Bacteroidetes 17 6419 Sphingobacteriia 3 6354

Saprospirae 1 2

Cytophagia 2 4

Flavobacteriia 11 59

Caldithrix 1 2 Caldithrixae 1 2

Chloroflexi 33 4203 Anaerolineae 10 1459

Ktedonobacteria 2 210

SAR202 16 1735

TK17 5 799

Cyanobacteria 1 3 Synechococcophycideae 1 3

Deferribacteres 1 2 Deferribacteres 1 2

Firmicutes 5 119 Bacilli 1 8

Clostridia 3 107

Erysipelotrichi 1 4

Gemmatimonadetes 7 1947 Gemm-2 7 1947

Nitrospirae 2 388 Nitrospira 2 388

PAUC34f 6 8411 n.a. 6 8411

Planctomycetes 6 29 unknown 1 2

Planctomycetia 5 27

Poribacteria 6 7167 n.a. 6 7167

Proteobacteria 84 10 374 Alphaproteobacteria 29 2479

Betaproteobacteria 2 57

Deltaproteobacteria 10 2592

Gammaproteobacteria 43 5246

SBR1093 1 130 EC124 1 130

Spirochaetes 2 584 Spirochaetes 2 584

TM6 1 3 SJA-4 1 3

Verrucomicrobia 2 35 Opitutae 1 33

Verrucomicrobiae 1 2

Total 205 71 404 205 71 404

n.a., Not applicable.

*Number of OTUs and †sequences found in each bacterial phylum and class, respectively, using full-size libraries. OTUs were

defined at ≥97% similarity between 16S rRNA gene sequences.

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 9

454-pyrosequencing. After quality filtering, trimming

and singleton removal, 60 729 16S rRNA V4–V5-tag

sequences were further analysed (Table S3, Supporting

information). Archaeal sequence libraries were highly

even in terms of depth across all sponge specimens

(Table S3), and results obtained with full and normal-

ized data sets, summarized in Fig. 5, were virtually

identical. The 60 729 inspected sequences were assigned

to 14 OTUs at 97% sequence similarity.

Archaeal community analyses. In 11 of 12 sponge speci-

mens, the archaeal community was dominated by the

candidate genus Nitrosopumilus (Thaumarchaeota; Nitroso-

pumilales), whereas the remaining specimen (Alg11/51)

had Cenarchaeum (Thaumarchaeota, Cenarchaeales) as the

dominant archaeon (Fig. 5a).

Normalized libraries with 4050 sequences were used

for the assessment of archaeal OTU diversity and rich-

ness in S. spinosulus (Fig. S3c,d, Supporting informa-

tion). Richness estimates from 2010 (2.30 � 1.35), 2011

(4.60 � 4.04) and 2012 (1.72 � 0.48) were not signifi-

cantly different (P > 0.05, Fig. S3c). Likewise, no differ-

ence (P > 0.05) was detected between Shannon

diversity indices from 2010 (0.1 � 0.16), 2011

(0.04 � 0.03) and 2012 (0.23 � 0.45) (Fig. S3d). The

OTU network demonstrated that replicates from 2011

contained more specific OTUs than 2010 and 2012 repli-

cates (Fig. 5b). It also showed that the dominant OTU

(OTU 17, Nitrosopumilus) was shared among the three

sampling years and detected in all sponge individuals,

whereas the second most dominant OTU (OTU 4, Cen-

archaeum) was only common to 2011 and 2012 replicates

(Fig. 5b).

The PCoA carried out with normalized libraries

showed that 10 sponge specimens grouped at the edge

of the first ordination axis (PC1), whereas one specimen

from 2012 (Alg12/81) was placed between this group,

and one further specimen from 2011 (Alg11/51)

was located near the PC2-PC3 extreme (Fig. 5c). The

distribution of the main archaeal taxa across the

samples is represented along the ordination diagram,

highlighting their correlation with the source specimens

in terms of relative abundance (Fig. 5c). UPGMA cluster

analysis of archaeal community profiles confirmed the

trend revealed by PCoA, as one replicate from 2011

(Alg11/51) and another from 2012 (Alg12/81) grouped

apart from the rest of the replicates (Fig. 5d). RDA

carried out for the full data set showed that 14.4% of

the whole data set variation could be attributed to the

independent variable ‘year of sampling’, which did not

significantly affect archaeal OTU composition across the

sampled specimens (P > 0.05, Fig. 5e). The horizontal

axis of the diagram accounted for 95% of the data set

variability explained by the independent variable and

discriminated Alg11/51 and Alg12/81 from the rest of

the replicates (Fig. 5e), corroborating results obtained

with normalized libraries (Fig. 5 c,d).

Discussion

In this survey, we posit that the bacterial community

associated with S. spinosulus was dynamically stable over

three successive sampling years based on oscillations in

OTU abundances that underpin the maintenance of pre-

vailing bacterial symbionts across all years along with

the sample-to-sample or year-to-year emergence of tran-

Table 3 The essential bacterial core of Sarcotragus spinosulus

Phylum OTUs* Sequences† Class OTUs* Sequences†

Acidobacteria 5 13 000 Acidobacteria-6 3 1740

Solibacteres 1 2093

Sva0725 1 9167

Actinobacteria 6 14 220 Acidimicrobiia 6 14 220

AncK6 1 2703 n.a. 1 2703

Chloroflexi 1 581 TK17 1 581

Gemmatimonadetes 1 474 Gemm-2 1 474

Nitrospirae 1 386 Nitrospira 1 386

PAUC34f 1 3809 n.a. 1 3809

Poribacteria 1 4950 n.a. 1 4950

Proteobacteria 10 6086 Alphaproteobacteria 3 1327

Deltaproteobacteria 1 1566

Gammaproteobacteria 6 3193

Total 27 46 209 27 46 209

n.a., Not applicable.

*Number of OTUs and †sequences found in each bacterial phylum and class, respectively, using full-size libraries. OTUs were

defined at ≥97% similarity between 16S rRNA gene sequences. The total number of OTUs common to all 12 surveyed sponge

specimens delineates the ‘essential bacterial core’.

© 2014 John Wiley & Sons Ltd

10 C. C. P . HARDOIM and R. COSTA

sient and less dominant symbionts. All S. spinosulus spec-

imens sampled in this study displayed a very high

degree of intra-specific CO1 gene conservation, being

thus almost identical genetically and permitting adequate

temporal analysis of symbiont temporal dynamics within

one single host genotype. Previous PCR-DGGE and

T-RFLP studies, employing either shorter or equivalent

temporal scales/sampling intervals, corroborate the

hypothesis of host-associated community maintenance,

however, with less emphasis on the extent of variability

among less abundant phylotypes across individuals or

time. For instance, a PCR-DGGE profiling study showed

that the bacterial communities associated with three Aus-

tralian sponges (Cymbastela concentrica, Callyspongia sp.

and Stylinos sp.), collected once per season, were stable

along five successive seasons (Taylor et al. 2004). Bacte-

rial community stability was also suggested for Chondrilla

nucula collected in three consecutive years from the Adri-

atic Sea, as several common PCR-DGGE bands were

observed in all duplicate individuals sampled per year

(Thiel et al. 2007). Using T-RFLP, Erwin et al. (2012)

revealed that the bacterial community associated with

each of three Ircinia spp. sampled every month was

highly stable along a period of one and a half years,

regardless of changes in temperature and light intensity

over the distinct seasons.

To circumvent the inherent limitations of fingerprint-

ing techniques (Oros-Sichler et al. 2007), we also used

454-pyrosequencing to unravel the structure of prokary-

otic communities in S. spinosulus over time. Using this

method, White et al. (2012) registered overall stability of

most bacterial symbionts associated with Axinella cor-

rugata collected in spring and fall within one single

year, but reported on slight shifts in structure between

the communities from both seasons. In line with

these observations, we further noticed that although

individual variation in community profiles did not erode

the conspicuous signal for temporal bacterial symbiont

maintenance in S. spinosulus, its cumulative representa-

tiveness to symbiont community dynamics deserves clo-

ser inspection. Importantly, among the usually rare

OTUs that were ‘exclusive’ to each sampling year or

specimen, many of them belong to typical sponge-associ-

ated phyla, for instance Acidobacteria, Actinobacteria, Chlo-

roflexi, PAUC34f and Poribacteria, and thus are less likely

to represent food or typical free-living bacteria. These

populations might therefore represent genetic variants of

functionally equivalent symbionts that could replace the

more dominant phylotypes in the face of changing host

physiological or microenvironment conditions. Func-

tional convergence of symbiont communities from diver-

gent sponge hosts has been proposed recently (Fan et al.

2012) and might likewise be supported by a high diver-

sity of microbial phylotypes (OTUs) occupying similar

niches in phylogenetically distinct sponge species. Fur-

ther, although it is not possible, on the basis of available

data, to propose a cause–effect relationship to explain

the observed shifts in Chloroflexi, Bacteroidetes and Pori-

bacteria abundances along the three successive years, we

noted for the three phyla that there was no correlation

between abundance and OTU richness changes. This

suggests that the range of potential metabolic capacities

comprised within these groups was rather conserved

throughout the study period.

The ‘essential bacteriome’ of S. spinosulus comprised

27 OTUs distributed across nine bacterial phyla, of

which the vast majority (24 OTUs) have often been

found in association with marine sponges (Taylor et al.

2007; Webster & Taylor 2012). Several putative, host

fitness-enhancing functions have been suggested for

(a) (b)

Fig. 4 Ordination analysis of bacterial OTU profiles. (a) Principal coordinate analysis (PCoA) of OTUs using the UniFrac metric for

normalized libraries. Red, green and purple circles correspond to samples from years 2010, 2011 and 2012, respectively. The 10 most

dominant bacterial taxa (at phylum or class level) are shown, and their symbol sizes represent the respective, mean relative abun-

dances across the data set. The position of bacterial taxa in the ordination space is determined by the correlation between their rela-

tive abundances and the year of sampling. (b) Redundancy analysis (RDA) of bacterial OTU profiles using the full data set. Symbols:

○ S. spinosulus collected in 2010, □ in 2011 and M in 2012. The ‘star’ symbol represents the centroid positions of the independent

variables in the diagram. Arrows represent OTUs displaying high abundance variation across samples/years. See Fig. S6 for RDA

ordination applied to the normalized data set.

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 11

these phyla such as chemical defence, in-faunal

biogeochemical cycling, and carbon and overall nutrient

provision (Taylor et al. 2007; Siegl & Hentschel 2010;

Siegl et al. 2011; Hentschel et al. 2012). Because its

composition cannot be explained by chance alone, it is

tempting to speculate that the essential and poor

bacterial core (at the phylotype level) common to all

S. spinosulus individuals could conceptually represent a

(a)

(b) (c)

(d)

(e)

Fig. 5 Structure of archaeal communities in Sarcotragus spinosulus across time. (a) Heatmap showing the distribution of the archaeal

16S rRNA gene taxonomic affiliation in each S. spinosulus specimen; (b) OTU network (97% sequence similarity) where red, green

and purple lines correspond to OTUs retrieved from years 2010, 2011 and 2012, respectively; (c) PCoA of the archaeal OTU composi-

tion; (d) UPGMA clustering and (e) RDA of archaeal OTU composition in S. spinosulus specimens collected ○ in 2010, □ in 2011 and

M in 2012. The ‘star’ symbol represents the centroid positions of the environmental variables in the RDA diagram. Graphs (a), (b)

and (d) were created using full-size sample libraries. Graphs (c) and (d) were constructed using normalized sample libraries (4050

sequence reads/sample).

© 2014 John Wiley & Sons Ltd

12 C. C. P . HARDOIM and R. COSTA

minimum repertoire of symbionts (nevertheless charac-

terized by high phylum-level richness) needed for full

holobiont functioning. The sharp reduction in phylotype

richness, but simultaneous maintenance of the archety-

pical sponge-associated phyla and classes in the

essential core corroborates the perception of higher bac-

terial taxonomic ranks as ecologically relevant units to

microecosystem functioning (Philippot et al. 2010). The

absence of the Bacteroidetes from this core suggests a

more sporadic pattern of occurrence of these symbionts

in S. spinosulus, supporting the notion of Bacteroidetes as

copiotrophic r-strategists (Fierer et al. 2007) and

opportunistic dwellers (Thomas et al. 2011) in open and

host-related microhabitats.

In summary, temporally transient bacterial symbionts

that do correspond to typical sponge-associated lineages

were unravelled. Collectively, they compete in richness

with the persistent microbiota, but are much less abun-

dant. Conversely, very few symbionts were common to

all analysed specimens, but found to rule the full

microbiome makeup—and probably its functioning—in

this host as they encompassed about 65% of all bacterial

sequences recovered with our effort. Sequence depth

largely influences observations of sponge microbiome

diversity (Moitinho-Silva et al. 2014). Hence, the slope

of the curves describing the pan and essential bacterial

core in S. spinosulus, as portrayed here, is to be chal-

lenged by analyses employing increased sequencing

power and other sponge hosts. Although it is likely that

the ‘rare’ or less frequent OTUs, as delineated in this

study, might lose their year- or specimen-dependent

‘specificity’ statuses when deeper analyses are under-

taken, gradients in relative abundances among commu-

nity members are expected to be preserved. Further, we

noticed considerable oscillations in the relative abun-

dance of several OTUs across sponge individuals or

time, this being the main factor underpinning differenti-

ation in community profiles among sponge specimens

(Appendix S3, Supporting information). Thus, the mi-

crobiome of S. spinosulus appears to be in dynamic sta-

bility as supported by the community turnover concept

in which the maintenance of the most dominant species

would be under constant assault by the emergence/

establishment of transient/rare species, which might

replace the dominant ones when conditions favour

(Gonz�alez et al. 2011). Such dynamics has been

observed in other environments, for instance the human

body (Costello et al. 2009; Caporaso et al. 2011), oceans

(Gilbert et al. 2009), lakes (Shade et al. 2007) and soils

(Costello & Schmidt 2006; Griffiths et al. 2011). In the

light of the current evidence for a tight control of the

host species in shaping its own microbiome (Erwin

et al. 2012; Hardoim et al. 2012; Pita et al. 2013), it

remains to be understood which ‘conditions’ would

likely trigger such replacements and drive symbiont

turnover, if any, within the sponge host. Finally, it is

felt that exacerbating specimen replication in diversity

surveys, as performed here through the inspection of

symbiont sharedness across 12 genetically equivalent

individuals, holds promise in enlightening our under-

standing of microbial community dynamics in marine

sponges and essentially any other host.

There is a current lack of knowledge of the temporal

persistence of archaeal symbionts in marine sponges.

With one exception (Lee et al. 2011), previous assays

have documented low archaeal diversity (usually ≤fourDGGE bands or OTUs) in these animals (Webster et al.

2001, 2004; Meyer & Kuever 2008). Here, the community

of archaeal symbionts in S. spinosulus was highly stable

and poorly diversified throughout the study period. It

was largely dominated by Nitrosopumilus in the three

consecutive years, except for one single sponge speci-

men that was dominated by Cenarchaeum sp. Candidatus

Nitrosopumilus maritimus is capable to grow chemoli-

thoautotrophically by aerobically oxidizing ammonia to

nitrite (K€onneke et al. 2005). Cenarchaeum symbiosum has

been observed in different sponge species from various

locations and even found to persist in one species kept

in aquarium during a 6-month period (Preston et al.

1996; Schleper et al. 1998; Margot et al. 2002). Genomic

analysis of C. symbiosum revealed homologues of genes

related to chemolithotrophic ammonia oxidation,

including ammonia monooxygenase-encoding genes

(Hallam et al. 2006). This indicates functional redun-

dancy between the Cenarchaeum and Nitrosopumilus spe-

cies found in association with S. spinosulus in our study.

Thus, alterations in their shared prevalence in a given

host individual, regardless of whether these changes

were caused by a deterministic factor or stochastic event

(s), do not seem to compromise the presumed roles

archaeal sponge symbionts play in ammonia oxidation.

The bacterial ammonia oxidizers (AOB) associated

with S. spinosulus comprised up to 14 amoA PCR-DGGE

bands. Previous studies are in congruence with our sur-

vey. PCR-DGGE profiling of the Caribbean Halisarca

caerulea and the deep cold-water species Higginsia thielei

and Nodastrella nodastrella (North Atlantic), generated

from a single sampling, showed few bands associated

with these hosts (Cardoso et al. 2013). Moreover, only

two AOB OTUs were retrieved from the endosome of

Astrosclera willeyana (Yang & Li 2012), and the AOB

associated with Geodia barretti was usually on the detec-

tion limit for PCR amplification (Hoffmann et al. 2009).

In S. spinosulus, the AOB community was stable

through time, presenting only small variation among

replicates and similar measurements of richness and

diversity. Although no archaeal amoA gene amplicon

could be obtained from S. spinosulus with the PCR-

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 13

DGGE system employed in this study, 454-pyrose-

quencing profiling of 16S rRNA genes hints at archaeal

ammonia oxidation potential in this host given the

documented dominance of Nitrosopumilus within these

communities. Altogether, these results represent the

first evidence for AOA and AOB stability in marine

sponges over time.

Conclusions

We reveal a predominantly stable microbiota in S. spi-

nosulus across time through the simultaneous analysis

of bacterial, archaeal and AOB community structures. A

subtle signal for bacterial community shifts within the

study period was nevertheless captured, characterized

by oscillations in OTU abundances driving lower

within-year and higher between-year variability among

replicate sponge specimens. This pattern became more

apparent when constrained ordination was applied to

the data, highlighting the usefulness in employing such

methods in microbial community profiling (Erb-Down-

ward et al. 2012). Studies covering larger time scales

(e.g. decades) are now needed to better verify the

strength of the sponge microbiome stability usually

reported in short-term surveys.

Acknowledgements

We acknowledge Joana R. Xavier and Francisco R. Pires for

their assistance in the identification of sponge specimens. We

thank Priya T. Suryakumar for the processing of DGGE images

with the GelCompar II software. This research was financed by

the Portuguese Foundation for Science and Technology (FCT)

through the research project PTDC/MAR/101431/2008,

granted to RC. It was further partially supported by the Euro-

pean Regional Development Fund (ERDF) through the Opera-

tional Competitiveness Programme (COMPETE) and by

national funds through FCT (Foundation for Science and Tech-

nology), under the project ‘PEst-C/MAR/LA0015/2011. CCPH

was the recipient of a PhD fellowship granted by FCT (Grant

No. SFRH/BD/60873/2009).

References

Bayer K, Schmitt S, Hentschel U (2008) Physiology, phylogeny

and in situ evidence for bacterial and archaeal nitrifiers in

the marine sponge Aplysina aerophoba. Environmental Microbi-

ology, 10, 2942–2955.

Caporaso JG, Kuczynski J, Stombaugh J et al. (2010) QIIME

allows analysis of high-throughput community sequencing

data. Nature Methods, 7, 335–336.Caporaso JG, Lauber CL, Costello EK et al. (2011) Moving

pictures of the human microbiome. Genome Biology, 12, R50.

Cardoso J, van Bleijswijk JDL, Witte H, van Duyl FC (2013)

Diversity and abundance of ammonia-oxidizing Archaea and

Bacteria in tropical and cold-water coral reef sponges. Aquatic

Microbial Ecology, 68, 215–230.

Costa R, Salles JF, Berg G, Smalla K (2006) Cultivation-inde-

pendent analysis of Pseudomonas species in soil and in the

rhizosphere of field-grown Verticillium dahliae host plants.

Environmental Microbiology, 8, 2136–2149.Costello EK, Schmidt SK (2006) Microbial diversity in alpine

tundra wet meadow soil: novel Chloroflexi from a cold,

water-saturated environment. Environmental Microbiology, 8,

1471–1486.Costello EK, Lauber CL, Hamady M et al. (2009) Bacterial Com-

munity variation in human body habitats across space and

time. Science, 326, 1694–1697.

Erb-Downward JR, Akha AAS, Wang J et al. (2012) Use of

direct gradient analysis to uncover biological hypotheses in

16S survey data and beyond. Scientific Reports, 2, 774.

Erwin PM, Pita L, L�opez-Legentil S, Turon X (2012) Stability of

sponge-associated bacteria over large seasonal shifts in tem-

perature and irradiance. Applied and Environmental Microbiol-

ogy, 78, 7358–7368.Esteves AIS, Hardoim CCP, Xavier JR, Gonc�alves JMS, Costa R

(2013) Molecular richness and biotechnological potential of

bacteria cultured from Irciniidae sponges in the north-east

Atlantic. FEMS Microbiology Ecology, 85, 519–536.Fan L, Reynolds D, Liu M et al. (2012) Functional equivalence

and evolutionary convergence in complex communities of

microbial sponge symbionts. Proceedings of the National Acad-

emy of Sciences, USA, 109, E1878–E1887.Fierer N, Bradford MA, Jackson RB (2007) Toward an ecologi-

cal classification of soil bacteria. Ecology, 88, 1354–1364.

Francis CA, Roberts KJ, Beman JM, Santoro AE, Oakley BB

(2005) Ubiquity and diversity of ammonia-oxidizing archaea

in water columns and sediments of the ocean. Proceedings of

the National Academy of Sciences, USA, 102, 14683–14688.

Gilbert JA, Field D, Swift P et al. (2009) The seasonal structure

of microbial communities in the Western English Channel.

Environmental Microbiology, 11, 3132–3139.Gonz�alez A, Clemente JC, Shade A et al. (2011) Our micro-

bial selves: what ecology can teach us. EMBO Reports, 12,

775–784.

Griffiths RI, Thomson BC, James P et al. (2011) The bacterial

biogeography of British soils. Environmental Microbiology, 13,

1642–1654.Hallam SJ, Konstantinidis KT, Putnam N et al. (2006) Genomic

analysis of the uncultivated marine crenarchaeote Cenarchae-

um symbiosum. Proceedings of the National Academy of Sciences,

USA, 103, 18296–18301.Hardoim CCP, Costa R, Ara�ujo FV et al. (2009) Diversity of

Bacteria in the marine sponge Aplysina fulva in Brazilian

Coastal Waters. Applied and Environmental Microbiology, 75,

3331–3343.Hardoim CCP, Esteves AIS, Pires FR et al. (2012) Phylogeneti-

cally and spatially close marine sponges harbour divergent

bacterial communities. PLoS One, 7, e53029.

Hentschel U, Hopke J, Horn M et al. (2002) Molecular evidence

for a uniform microbial community in sponges from different

oceans. Applied and Environmental Microbiology, 68, 4431–4440.Hentschel U, Piel J, Degnan SM, Taylor MW (2012) Genomic

insights into the marine sponge microbiome. Nature Reviews

Microbiology, 10, 641–654.

Heuer H, Krsek M, Baker P, Smalla K, Wellington EMH (1997)

Analysis of Actinomycete communities by specific amplifica-

tion of genes encoding 16S rRNA and gel-electrophoretic

© 2014 John Wiley & Sons Ltd

14 C. C. P . HARDOIM and R. COSTA

separation in denaturing gradients. Applied Environmental

Microbiology, 63, 3233–3241.Hoffmann F, Radax R, Woebken D et al. (2009) Complex nitro-

gen cycling in the sponge Geodia barretti. Environmental

Microbiology, 11, 2228–2243.

K€onneke M, Bernhard AE, de la Torre JR et al. (2005) Isolation

of an autotrophic ammonia-oxidizing marine archaeon. Nat-

ure, 437, 543–546.Lee OO, Wang Y, Yang JK et al. (2011) Pyrosequencing reveals

highly diverse and species-specific microbial communities in

sponges from the Red Sea. ISME Journal, 5, 650–664.

Margot H, Acebal C, Toril E, Amils R, Puentes JLF (2002) Con-

sistent association of crenarchaeal Archaea with sponges of

the genus Axinella. Marine Biology, 140, 739–745.Meyer B, Kuever J (2008) Phylogenetic diversity and spatial

distribution of the microbial community associated with the

Caribbean deep-water sponge Polymastia cf. corticata by 16S

rRNA, aprA, and amoA gene analysis. Microbial Ecology, 56,

306–321.

Meyer CP, Geller JB, Paulay G (2005) Fine scale endemism on

coral reefs: archipelagic differentiation in turbinid gastro-

pods. Evolution, 59, 113–125.Moitinho-Silva L, Bayer K, Cannistraci CV et al. (2014) Specific-

ity and transcriptional activity of microbiota associated with

low and high microbial abundance sponges from the Red

Sea. Molecular Ecology, 23, 1348–1363.Nicolaisen MH, Ramsing NB (2002) Denaturing gradient gel

electrophoresis (DGGE) approaches to study the diversity of

ammonia-oxidizing bacteria. Journal of Microbiological Meth-

ods, 50, 189–203.

Oros-Sichler M, Costa R, Heuer H, Smalla K (2007) Molecular

fingerprinting techniques of microbial communities. In: Mod-

ern Soil Microbiology II (eds van Elsas JD, Trevors JT, Jansson

JK), pp. 357–388. CRC Press, Boca Raton, Florida, USA.

Philippot L, Andersson SGE, Battin TJ et al. (2010) The ecologi-

cal coherence of high bacterial taxonomic ranks. Nature

Reviews Microbiology, 8, 523–529.Pires ACC, Cleary DFR, Almeida A et al. (2012) Denaturing

gradient gel electrophoresis and barcoded pyrosequencing

reveal unprecedented archaeal diversity in mangrove sedi-

ment and rhizosphere samples. Applied and Environmental

Microbiology, 78, 5520–5528.

Pita L, Turon X, L�opez-Legentil S, Erwin PM (2013) Host rules:

spatial stability of bacterial communities associated with

marine sponges (Ircinia spp.) in the Western Mediterranean

Sea. FEMS Microbiology Ecology, 86, 268–276.

Preston CM, Wu KY, Molinski TF, DeLong EF (1996) A psy-

chrophilic crenarchaeon inhabits a marine sponge: Cenarchae-

um symbiosum gen nov, sp, nov. Proceedings of the National

Academy of Sciences, USA, 93, 6241–6246.

Quince C, Lanzen A, Davenport RJ, Turnbaugh PJ (2011)

Removing noise from pyrosequenced amplicons. BMC Bioin-

formatics, 12, 38.

R Development Core Team (2012) R: A Language and Environ-

ment for Statistical Computing. R Foundation for Statistical

Computing, Vienna, Austria. Available at: http://www.

R-project.org.

Rotthauwe JH, Witzel KP, Liesack W (1997) The ammonia

monooxygenase structural gene amoA as a functional mar-

ker: molecular fine-scale analysis of natural ammonia-oxidiz-

ing populations. Applied and Environmental Microbiology, 63,

4704–4712.Schleper C, DeLong EF, Preston CM et al. (1998) Genomic

analysis reveals chromosomal variation in natural popula-

tions of the uncultured psychrophilic archaeon Cenarchaeum

symbiosum. Journal of Bacteriology, 180, 5003–5009.Schmitt S, Tsai P, Bell J et al. (2012) Assessing the complex

sponge microbiota: core, variable and species-specific bacte-

rial communities in marine sponges. ISME Journal, 6, 564–

576.

Shade A, Kent AD, Jones SE et al. (2007) Interannual dynamics

and phenology of bacterial communities in a eutrophic lake.

Limnology and Oceanography, 52, 487–494.

Siegl A, Hentschel U (2010) PKS and NRPS gene clusters from

microbial symbiont cells of marine sponges by whole gen-

ome amplification. Environmental Microbiology Reports, 2, 507–513.

Siegl A, Kamke J, Hochmuth T et al. (2011) Single-cell genom-

ics reveals the lifestyle of Poribacteria, a candidate phylum

symbiotically associated with marine sponges. ISME Journal,

5, 61–70.

Simister RL, Deines P, Botte ES, Webster NS, Taylor MW

(2012) Sponge-specific clusters revisited: a comprehensive

phylogeny of sponge-associated microorganisms. Environ-

mental Microbiology, 14, 517–524.

Steger D, Ettinger-Epstein P, Whalan S et al. (2008) Diversity

and mode of transmission of ammonia-oxidizing archaea in

marine sponges. Environmental Microbiology, 10, 1087–1094.

Taylor MW, Schupp PJ, Dahllof I, Kjelleberg S, Steinberg PD

(2004) Host specificity in marine sponge-associated bacteria,

and potential implications for marine microbial diversity.

Environmental Microbiology, 6, 121–130.

Taylor MW, Radax R, Steger D, Wagner M (2007) Sponge-asso-

ciated microorganisms: evolution, ecology, and biotechnolog-

ical potential. Microbiology and Molecular Biology Reviews, 71,

295–347.

Taylor MW, Tsai P, Simister RL et al. (2013) ‘Sponge-specific’

bacteria are widespread (but rare) in diverse marine environ-

ments. ISME Journal, 7, 438–443.Thiel V, Leininger S, Schmaljohann R, Bruemmer F, Imhoff JF

(2007) Sponge-specific bacterial associations of the Mediterra-

nean sponge Chondrilla nucula (Demospongiae, Tetractino-

morpha). Microbial Ecology, 54, 101–111.Thomas F, Hehemann J-H, Rebuffet E, Czjzek M, Michel G

(2011) Environmental and gut Bacteroidetes: the food connec-

tion. Frontiers in Microbiology, 2, 93.

Tourna M, Freitag TE, Nicol GW, Prosser JI (2008) Growth,

activity and temperature responses of ammonia-oxidizing ar-

chaea and bacteria in soil microcosms. Environmental Microbi-

ology, 10, 1357–1364.

Webster NS, Taylor MW (2012) Marine sponges and their

microbial symbionts: love and other relationships. Environ-

mental Microbiology, 14, 335–346.Webster NS, Watts JEM, Hill RT (2001) Detection and phyloge-

netic analysis of novel crenarchaeote and euryarchaeote 16S

ribosomal RNA gene sequences from a Great Barrier Reef

sponge. Marine Biotechnology, 3, 600–608.Webster NS, Negri AP, Munro M, Battershill CN (2004)

Diverse microbial communities inhabit Antarctic sponges.

Environmental Microbiology, 6, 288–300.

© 2014 John Wiley & Sons Ltd

TEMPORAL DYNAMICS OF SPONGE SYMBIONTS 15

Weisburg WG, Barns SM, Pelletier DA, Lane DJ (1991) 16S

ribosomal DNA amplification for phylogenetic study. Journal

of Bacteriology, 173, 697–703.

White JR, Patel J, Ottesen A et al. (2012) Pyrosequencing of bac-

terial symbionts within Axinella corrugata sponges: diversity

and seasonal variability. PLoS One, 7, e38204.

Yang ZY, Li ZY (2012) Spatial distribution of prokaryotic sym-

bionts and ammoxidation, denitrifier bacteria in marine

sponge Astrosclera willeyana. Scientific Reports, 2, 528.

C.C.P.H. and R.C. conceived and designed the study.

C.C.P.H. performed the laboratory experiments and

analysed the data. R.C. contributed reagents, materials

and analysis tools. C.C.P.H. and R.C. wrote the article.

Data accessibility

Sponge CO1 sequences were deposited under the NCBI

accession nos HE797930-HE797933, HG816008-

HG816010 and HG816023-HG816027. Raw pyrosequenc-

ing sequences were deposited in the NCBI Sequence

Read Archive (SRA) with the study accession no.

SRP033344. Scripts used in the analysis of 454-pyrose-

quencing data have been submitted as supporting infor-

mation (Appendix S1). Quality-filtered OTU tables and

taxonomic assignment of bacterial and archaeal OTUs

are provided as supporting information (Appendices

S3–S6). Alignment of CO1 sequences used in sponge

phylogeny is accessible as supporting information

(Appendix S7).

Supporting information

Additional supporting information may be found in the online

version of this article.

Appendix S1 Detailed 454-pyrosequencing methodology.

Appendix S2 Detailed description of PCR-DGGE results.

Appendix S3 Bacterial OTU table constructed from the full

quality-filtered data set.

Appendix S4 Taxonomic assignment of bacterial OTUs.

Appendix S5 Archaeal OTU table constructed from the full

quality-filtered data set.

Appendix S6 Taxonomic assignment of archaeal OTUs.

Appendix S7 Alignment of CO1 sequences used in sponge

phylogenetic inference.

Fig. S1 PCR-DGGE fingerprinting of bacterial 16S rRNA (a), ar-

chaeal 16S rRNA (b) and bacterial amoA (c) gene fragments

amplified from the marine sponge Sarcotragus spinosulus in

three consecutive years.

Fig. S2 Phylum- (a, b) and class-level (c, d) bacterial commu-

nity composition in Sarcotragus spinosulus in three consecutive

years using the normalized data set (3155 sequence reads per

sample).

Fig. S3 Observed and estimated (Chao1 index) richness and

diversity indices of bacterial (a,b) and archaeal (c,d) 454-py-

rosequencing OTUs (97% cut-off) detected in the marine

sponge Sarcotragus spinosulus along three consecutive years.

Fig. S4 Bacterial OTU networking and maintenance in the mar-

ine sponge Sarcotragus spinosulus across time using normalized

library sizes (3155 sequence reads per sample).

Fig. S5 The pan (left) and essential core (right) bacteriome of

Sarcotragus spinosulus.

Fig. S6 Redundancy analysis (RDA) of bacterial OTU profiles

using the normalized data set.

Table S1 Bacterial 16S rRNA gene pyrosequencing data set.

Table S2 Taxonomic classification and absolute abundance of

bacterial OTUs found exclusively in each sampling year.

Table S3 Archaeal 16S rRNA gene pyrosequencing data set.

© 2014 John Wiley & Sons Ltd

16 C. C. P . HARDOIM and R. COSTA