Introduction

Sponges (phylum Porifera) are arguably the earliest branching Metazoa (Philippe et al. 2009, 2011), with a fossil record dating back nearly 700 million years (Erwin et al. 2011). Sponges harbor rich and diverse microbial communities in their tissues (for a comprehensive review see Taylor et al. 2007). In the so-called ‘high microbial abundance sponges’ or ‘bacteriosponges’, microbes can make up to 40 % of the biomass of the host, whereas the ‘low microbial abundance sponges’ harbor relatively small numbers of microorganisms (Reiswig 1981). Currently, 30 bacterial phyla, two major lineages of the Archaea and several types of eukaryotic microbes associated with sponges have been identified (Hentschel et al. 2002, 2003; Taylor et al. 2007; Hardoim et al. 2009; Webster et al. 2010; Schmitt et al. 2012). Phylogenetic analyses have indicated a low temporal variability in marine sponge-associated microbial communities, which are fairly stable within individuals and through time (Taylor et al. 2004; Webster et al. 2004; Hentschel et al. 2006). Hentschel et al. (2002) established criteria to define monophyletic, sponge-specific “clusters” (SSC): a group of at least three sequences that (i) are found in different host sponge species and/or from different geographical locations, (ii) are more similar to each other than to any other sequence from a non-sponge source and (iii) are grouped together into one clade independent of the tree construction method. This and further studies have shown that taxonomically distantly related sponges with geographically non-overlapping distribution patterns and with host-specific secondary metabolite profiles contain surprisingly uniform microbial signatures (for a review see Taylor et al. 2007). However, a recent pyrosequencing study by Webster and colleagues reported the presence of Poribacteria and 17 of the other 33 currently reported sponge-specific groups in seawater (Webster et al. 2010; Taylor et al. 2012). Therefore, this study questions the hypothesis that some groups of microbes are restricted to the sponge host and distinct from those in the surrounding seawater (Taylor et al. 2004).

Coralline sponges, also known as sclerosponges, are unique members of phylum Porifera (Reitner 1992; Wörheide 2008) because they build a solid secondary calcareous skeleton (Reitner 1992; Chombard et al. 1997) in addition to a primary, often spicular, skeleton. During long periods of the Earth’s history, sclerosponges were dominant, diverse and abundant reef-building organisms (Vacelet 1985). These organisms were thought to be extinct until their rediscovery in the late 1960s (Hartman 1969). Today, only approximately 15 taxa live, mainly restricted to the cryptic niches of coral reefs with reduced light and oligothrophic conditions, such as caves and deeper fore-reef areas (Reitner 1992; Wörheide 1998). Sclerosponge genera, such as Acanthochatetes, Vaceletia or Astrosclera, are regarded as “living fossils” due to their occupation of the same ecological niches for hundreds of millions of years. In addition, these organisms display very similar morphological characteristics when compared to their fossil relatives that lived millions of years ago (Reitner et al. 2001). Therefore, coralline sponges might provide insight into the evolution of sponge-microbial symbioses. Fossil records from Silurian microbial reefs, with stromatoporoids neighbouring ubiquitous microbial laminae or less commonly encrusted by cyanobacteria, might already indicate those close associations (Soja et al. 2003).

Sponges of the genus Vaceletia, which was thought to be extinct until its rediscovery in the 1970s (Vacelet 1977), systematically belong to the Keratosa, a group of sponges devoid of a primary mineralized skeleton (Wörheide 2008). Bacteria may make up more than 50 % of the entire biomass of the sponge (Reitner and Wörheide 2002). Vaceletia species occur in two putative sister-species with different growth modes (solitary vs. colonial; Wörheide and Reitner 1996). For the detailed descriptions and definitions of the solitary and colonial forms see Vacelet (1988), Vacelet et al. (1992), Wörheide and Reitner (1996).

The microbial communities in coralline sponges have yet to be investigated in detail. Here, we aimed to perform detailed characterizations and sequenced a 16S rRNA clone library of a randomly picked Vaceletia crypta specimen, the only validly described recent species of the genus. To further investigate whether growth mode and/or putative sister-species relationships lead to differences in microbial diversity, we additionally performed denaturing gradient gel electrophoresis (DGGE). We aimed to determine whether the microbial communities of these “living fossil” sponges differ from those reported from other sponges and, by phylogenetic analyses, contribute to the question of the maintenance of sponge-microbe symbioses.

Materials and methods

Sample collection

Seventeen samples were collected by SCUBA diving at depths from 7 to 30 m at several sampling sites in the Coral Sea and Pacific Ocean. Sponges were excised from the substrate using a chisel and a hammer and transferred directly (underwater) to plastic bags. Two samples were collected by a remotely operated vehicle (ROV) at depths from 200 to 250 m. The sampling details for all samples used are listed in Table 1. After collection, sponge samples were preserved either in silica gel (Erpenbeck et al. 2004), DMSO buffer (20 % DMSO, 0.25 M EDTA and NaCl to saturation, pH 8.0; adapted from Seutin et al. 1991) or 95 % ethanol.

Table 1 Sample data of investigated Vaceletia sp. specimens, with collection sites details
Full size table

DNA extraction

Samples were rinsed with autoclaved Millipore water and the preserved living tissue was cut and crushed aseptically with a sterile scalpel on a Petri dish. Total DNA was extracted from 3 mg of tissue using a Qiagen DNeasy Tissue kit (Qiagen GmbH, Hilden, Germany) following the manufacturer’s instructions.

PCR amplification and cloning of the 16S rRNA genes of V. crypta from Yonge Reef from Great Barrier Reef (GBR), Australia (sample no. GW947)

The bacterial 16S rRNA genes were amplified from the DNA extract obtained by PCR using GoTaq polymerase (Promega GmbH, Mannheim, Germany) and universal bacterial primers (616F: 5′-AGA GTT TGA TYM TGG CTC AG-3′ and 1525R: 5′-AGA AAG GAG GTG ATC CAG CC-3′) (Lane 1991). Cycling conditions for the Biometra thermocycler were as follows: an initial denaturation step (2 min at 95 °C) followed by 35 cycles of denaturation (30 s at 94 °C), primer annealing (1 min at 55 °C), elongation (2 min + 4 s at 72 °C) and a final extension step (5 min at 72 °C). After purification using the mi-PCR Purification Kit (metabion GmbH, Martinsried, Germany), the DNA was subsequently cloned into the plasmid cloning vector using an Invitrogen TOPO® TA Cloning Kit for Sequencing according to the manufacturer’s instructions (Life Technologies GmbH, Darmstadt, Germany). The inserts of 427 clones were PCR reamplified using vector specific primers (M13) (Sambrook and Russell 2001) and Promega GoTaq polymerase. Analytical digestions of PCR products of 1,500 base pairs length were performed in single reactions using the restriction enzyme MspI (Fermentas GmbH, St. Leon-Rot, Germany) following the manufacturer’s instructions. Based on the restriction patterns, similar clones were grouped together and chosen randomly for sequencing. Clones with undefined restriction patterns were additionally taken for sequencing. Prior to sequencing, amplified inserts were purified using a silica-based protocol modified after Boyle and Lew (1995).

For the amplification of archaeal 16S rRNA gene from the DNA extract, touchdown PCR using universal archaeal primers (21F: 5′-TTC CGG TTG ATC CYG CCG GA-3′ and 915R 5′-GTG CTC CCC CGC CAA TTC CT-3′) (DeLong 1992; Raskin et al. 1994) and an annealing temperature decreasing from 60 to 50.5 °C (30 s each) in 0.5 °C increments was employed. The cycling conditions for the Biometra thermocycler using Promega GoTaq were as follows: one cycle of initial denaturation (2 min at 95 °C); 35 cycles of denaturation (30 s at 94 °C), primer annealing (30 s from 60 °C minus 0.5 °C), and elongation (2 min + 4 s at 72 °C) followed by 25 cycles of denaturation (30 s at 94 °C), primer annealing (30 s at 51 °C), and elongation (2 min + 4 s at 72 °C) and a final extension step (5 min at 72 °C). A strong band of approx. 900 base pairs was excised and the DNA was purified from an agarose gel using the E.Z.N.A Gel Extraction Kit (VWR International GmbH, Darmstadt, Germany) following the manufacturer’s instructions and subsequently taken for sequencing.

Sequencing

Sequencing was performed by the Genomics Service Unit (Ludwig-Maximilians-Universität München) using the BigDye® Terminator v3.1 on a 48-capillary sequencer (ABI 3730, Applied Biosystems). For the cloned bacterial inserts, the primers: 610RII (5′-ACC GCG/T A/GCT GCT GGC AC-3′) (Dotzauer et al. 2002), 616F and 1492R (5′-GGT TAC CTT GTT ACG ACT T-3′) (Lane 1991), or 614F (5′-GTG CAT GGC TGT CGT CAG CTC G-3′) (this study) were used. The archaeal PCR product was sequenced with AR20F primer (5′-TTC CGG TTG ATC CYG CCRG-3′) (Moyer et al. 1998). CodonCode Aligner (http://www.codoncode.com/aligner/) was used for sequence editing and assembly. Sequences were checked for chimeras using the Bellerophon web applications (Huber et al. 2004). Chimera sequences were removed before further analyses.

Phylogenetic analyses

Phylogenetic analyses were performed using ARB (Ludwig et al. 2004). The initial ARB database was constructed from the SILVA Small Subunit rRNA Database (release 96; (Pruesse et al. 2007) and from the database constructed by Taylor and colleagues, which contains sponge–derived sequences along with their closest relatives (Taylor et al. 2007). Our sequences were compared to the sequences available in public databases using BLAST (http://blast.ncbi.nlm.nih.gov/) and the nearest relative sequences obtained from different sponges, corals and non-sponge sources, were also incorporated into the ARB Database. The sequences were aligned using the ARB Integrated Aligner. The alignment was checked and corrected manually for alignment errors. The partial sequences were added to the ARB database using the ARB parsimony “quick add” tool. Initial phylogenetic trees were evaluated using the neighbour-joining algorithm (Jukes–Cantor correction) using ARB. Subsequently, the alignment was exported from the ARB database and maximum likelihood trees were constructed using RAxML (v.7.2.5; (Stamatakis 2006), 1,000 bootstrap replicates and the GTR+GAMMA model of sequence evolution. The resulting trees were visualized using the program FigTree (v.1.3.1) (http://tree.bio.ed.ac.uk/software/figtree/). Monophyletic, sponge-specific and sponge-coral clusters (SSC/SCC) were defined based on established criteria (Hentschel et al. 2002). Sequences obtained from the sponges and corals that grouped together into one clade independent of the tree reconstruction method (neighbour-joining and maximum likelihood) were regarded as SSC and/or SCC.

Estimation of microbial diversity and statistical analyses

Based on the distance matrix generated by ARB, the sequences were assigned to operational taxonomic units (OTUs) using Mothur (Schloss et al. 2009). The clones that were only analyzed by restriction digestion were assigned to corresponding OTUs based on their restriction patterns. For the analysis of an OTU, a cut-off value of 0.03 was used (Schloss and Handelsman 2005). The rarefaction curves were also calculated using Mothur. The curves were plotted using the R software package (http://www.R-project.org). The Shannon–Wiener index (Spellerberg and Fedor 2003) was calculated to determine the abundance and richness of the bacterial community associated with V. crypta. The Chao1 index (Colwell and Coddington 1994) was employed to estimate total species richness. Calculations were performed using the Mothur software. In order to determine the phylogenetic composition of the clone library constructed from the microbial community associated with V. crypta, the percentage for each phylogenetic group was calculated based on the number of clones assigned to the particular group.

Denaturing gradient gel electrophoresis (DGGE)

Nineteen samples of members of the genus Vaceletia representing colonial (9 samples) and solitary (10 samples) growth forms were used for DGGE. All of the samples used are listed in Table 1. The bacterial 16S rRNA genes were amplified from the DNA extracts using touchdown PCR, Promega GoTaq polymerase and the universal bacterial primers 341F-GC and 907RC (Muyzer and Smalla 1998; Schäfer 2001). Cycling conditions for the Biometra thermocycler were as follows: one cycle of initial denaturation (2 min at 95 °C); 15 cycles of denaturation (30 s at 94 °C), primer annealing (30 s from 58 °C minus 0.5 °C), and elongation (2 min + 4 s at 72 °C) followed by 25 cycles of denaturation (30 s at 94 °C), primer annealing (30 s from 51 °C minus 0.5 °C), and elongation (2 min + 4 s at 72 °C) and a final extension step (5 min at 72 °C).

The DGGE analysis was performed with an Ingeny phorU-2 system (Ingeny International), Power Pac 300 (BioRad) as a power supplier, and a denaturing gradient of 30–70 % (urea and formamide) in a 6 % polyacrylamide gel. Gels were run for 16 h at 180 V (60 °C), then stained for 25 min in SYBR Gold (Molecular Probes) and photographed using a RT Color SPOT camera and SPOT advance imaging software (Visitron Systems GmbH).

Gel images were analyzed using Quantity One software (version 4.69, Bio-Rad). Similarities between the DGGE banding patterns were calculated using the band-matching Dice coefficient with an optimization at 0.75 % and a tolerance level of 0.75 %. The unweighted pair-group method with arithmetic averages (UPGMA) was used for cluster analysis to obtain similarity dendrograms.

Nucleotide sequence accession numbers

The 16S rRNA gene sequences representing all OTUs generated in this study have been deposited in EMBL database under the accession numbers HE817775 to HE817870.

Results

Clone library construction and OTU assignment

427 clones were selected from the 16S rRNA clone library amplified from a solitary form specimen of V. crypta from Yonge Reef (GBR, Australia), which possessed a high microbial abundance, in parts of the sponge outnumbering sponge cells (Fig. 1). From those clones, 253 were sequenced and the remaining 174 clones were assigned to a particular OTU based on their restriction patterns. A single archaeal 16S rRNA sequence was retrieved, however the multiple genotypes cannot be entirely ruled out. Three sequences were discarded as chimeras. The remaining 250 sequences were clustered into 96 OTUs using a 97 % similarity criterion. From those 96 OTUs, 39 were singletons (Fig. 2). Only 8 OTUs grouped more than 10 clones (two OTUs with 18 and 11 clones and single OTUs with 15, 22, 23 and with 30 clones, respectively).

Fig. 1
figure 1

TEM micrograph of the choansosome of Vaceletia crypta with numerous bacterial cells and a few sponge cells (Sp) in the mesohyl, as well as in the choanocyte chamber (ChCh). Scale bar 2 μm

Full size image
Fig. 2
figure 2

Distribution of 16S rRNA gene clones among the OTUs defined at the 97 % similarity criterion. * N1 number of singletons, N2 number of doubletons, etc

Full size image

Phylogenetic analyses

Using BLAST, 88 % of the 96 OTUs (84 OTUs) were found to be closely related with other previously described sponge- or coral-derived sequences. Of the OTUs, 70 % (67 OTUs) were related to other sponge-derived 16S rRNA genes obtained from 22 different sponge species. A further 18 % of the OTUs contained sequences obtained from four different species of corals as closest relatives. Of those OTUs, 14 were related to 16S rRNA sequences obtained from Montastraea faveolata and a single OTU was related to sequences obtained from Oculina patagonica, Diploria strigosa and Erythropodium caribaeorum. Only one OTU was closely related (99 % similarity) to a validly named organism (Delftia acidovorans), and one was distantly related (91 % similarity) to a 16S rRNA gene sequence from a chloroplast of a red alga. Further, 7 % of the OTUs had sequences from marine environments as next relatives (3 OTUs from seawater, 1 OTU from basaltic glass from a seamount and 3 OTUs from sediment, with one of the 3 from deep sea). Three OTUs contained sequences from the terrestrial subsurface as next relatives. The results of the BLAST search are summarized in the supplementary Table (ESM 1 of the Online Resource).

The phylogeny obtained with ARB showed that the majority of the sponge-derived microbial clones were assigned to the Chloroflexi (39 OTUs, number of clones n = 144) and Gammaproteobacteria (13 OTUs, n = 46). Clones affiliated with the Deltaproteobacteria (9 OTUs, n = 29), Acidobacteria (6 OTUs, n = 24), Gemmatimonadetes (5 OTUs, n = 46), and Alphaproteobacteria (4 OTUs, n = 11) were also observed. Numerous clones affiliated to the Nitrospirae and Actinobacteria were associated with a single OTU (n = 30 clones) or three OTUs (n = 34), respectively. The minor components of the clone library were clones affiliated with the Poribacteria (3 OTUs, n = 4), Betaproteobacteria (2 OTUs, n = 3), Cyanobacteria (2 OTUs, n = 3), Spirochaetes (1 OTU, n = 2), Deinococcus-Thermus (1 OTU, n = 1), and Bacteroidetes (1 OTU, n = 1). The single sequence obtained by PCR using universal archaeal primers was affiliated to the Crenarchaeota. Over 9 % of the clone sequences (5 OTUs, n = 30) were not classified using ARB database to any described phylum. However, based on the EMBL phylogeny the available sequences that were most similar implied an affiliation of these sequences with the phylum Deferribacteres. The phylogenetic trees present the OTUs with nearest similar sequences assigned to the Chloroflexi (Fig. 3), Proteobacteria (Fig. 4) and to all other phyla (Fig. 5).

Fig. 3
figure 3

Maximum likelihood phylogeny of V. crypta-derived 16S rRNA sequences affiliated to the phylum Chloroflexi with next similar sequences obtained from other sponges or corals, and from the environment. Reference sequences are listed with their GenBank numbers. Bold text signifies clones obtained during this study. Shaded boxes represent sponge-specific clusters; the percentage values next to the boxes indicate the similarity between the sequences belonging to the clusters. Bootstrap analysis was based on 1,000 replicates—the support values 70–85 % are indicated by asterisk. Scale bar signifies 10 % sequence divergence

Full size image
Fig. 4
figure 4

Maximum likelihood phylogeny of V. crypta-derived 16S rRNA sequences affiliated to the phylum Proteobacteria with next similar sequences obtained from other sponges or corals, and from the environment. Reference sequences are listed with their GenBank numbers. Bold text signifies clones obtained during this study. Shaded boxes represent sponge-specific clusters; the percentage values next to the boxes indicate the similarity between the sequences belonging to the clusters. Bootstrap analysis was based on 1,000 replicates—the support values 70–85 % are indicated by asterisk. Scale bar signifies 10 % sequence divergence

Full size image
Fig. 5
figure 5

Maximum likelihood phylogeny of V. crypta-derived 16S rRNA sequences affiliated to several phyla with next similar sequences obtained from other sponges or corals, and from the environment. Reference sequences are listed with their GenBank numbers. Bold text signifies clones obtained during this study. Shaded boxes represent sponge-specific clusters; the percentage values next to the boxes indicate the similarity between the sequences belonging to the clusters. Bootstrap analysis was based on 1,000 replicates—the support values 70–85 % are indicated by asterisk. Scale bar signifies 10 % sequence divergence

Full size image

Sponge-specific and sponge-coral clusters

From the 84 OTUs that contained sequences similar to sequences obtained from other sponges or corals, 71 OTUs (85 %) were assigned to 63 SSC or SCC. The largest number of clusters belonged to the phylum Chloroflexi (27 clusters with 30 OTUs). A further 15 clusters were defined among the Proteobacteria (15 clusters with 16 OTUs). The SSC and SCC are indicated with grey-shaded boxes in the phylogenetic trees (Figs. 3, 4, 5).

Estimation of microbial diversity and statistical analyses

The microbial community composition was calculated for all clones affiliated to each phylogenetic group and revealed a high diversity with a complex composition (Fig. 6). The most abundant taxa were the Chloroflexi (35 %). Due to the complexity, variety and diversity of the phylum Proteobacteria, the proteobacterial classes were treated as separate phylogenetic groups (Alpha-, Beta-, Gamma- and Deltaproteobacteria) and as separate groups for the estimations of the microbial community composition. If, for the calculation of community composition, the Proteobacteria were regarded as one single group (phylum), it would be the second most abundant group in the community (22 %) behind the Chloroflexi.

Fig. 6
figure 6

Distribution of the 16S rRNA clones among particular phylogenetic groups in the clone library obtained from the V. crypta

Full size image

A rarefaction analysis was used to assess whether the number of clones sequenced from the library represented the full diversity of the microbial community. The rarefaction curves calculated using 97 % and 95 % cut-off criteria for grouping OTUs at the “species” and “genus” levels as well as 90 % did not reach a clear saturation (Fig. 7). However, according to the Chao1 index (Table 2), we sequenced over 70 % of the predicted number of microbial species, which provides a representative picture of the core microbial community of V. crypta. Using Sanger sequencing to cover the remaining 30 % of the community would be costly and time-consuming (Schmitt et al. 2012).

Fig. 7
figure 7

Rarefaction curves for the 16S rRNA sequences obtained from V. crypta. Operational taxonomic units (OTU) were defined at the 97, 95 and 90 % similarity criterion

Full size image
Table 2 Sample diversity
Full size table

Denaturing gradient gel electrophoresis

DGGE analysis of nine colonial and ten solitary forms of individuals from genus Vaceletia indicated that the diversity of all the bacterial communities was very high. The number of bands ranged from 27 to 35 per sample. The lowest number of bands was obtained from the bacterial community associated with a solitary form of Vaceletia from Palau, Siaes Tunnel (sample #3, ESM 2 of the Online Resource) and the largest number of bands was obtained from a colonial form from Bougainville Reef (sample #2, ESM 2 of the Online Resource). The banding patterns displayed numerous co-occurring bands; however, only four bands were found in all samples from both growth forms. 18 bands were specific to the samples from the solitary Vaceletia form and 19 to the samples from the colonial form. A cluster analysis showed that the microbial communities appear to be growth-form specific (Fig. 8); however, the solitary specimens from Palau, which clustered together, displayed a higher affiliation to the cluster of colonial samples. The bacterial profiles for the samples obtained from Norfolk Ridge (colonial form) and from Solomon Island (solitary form), both from deeper sampling zones, did not cluster with the other samples.

Fig. 8
figure 8

UPGMA dendrogram constructed from DGGE profiling of PCR-amplified bacterial 16S rRNA genes of microbial community associated with S solitary and C colonial forms of Vaceletia sponges from different locations. Sample names according to the Table 1 (column: Site)

Full size image

Discussion

To our knowledge, this is the first study assessing the phylogenetic diversity of Bacteria and Archaea in coralline sponges using molecular approaches. The 16S rRNA gene-based diversity analysis of V. crypta revealed that its associated microbial community is phylogenetically complex and diverse because it is composed of representatives of the Archaea and 13 bacterial phyla. The distribution of the sequences was relatively even between phylogenetic groups, however the largest number of sequences was affiliated with the Chloroflexi, which have frequently been reported as members of sponge-associated microbial communities and often as the predominant group (Hentschel et al. 2002; Webster et al. 2004; Thiel et al. 2007). In a recent work Schmitt and colleagues (2011) showed that high microbial abundance sponges host more diverse, abundant, and similar Chloroflexi bacteria then low microbial abundance sponges. Of the sequences belonging to the Chloroflexi, 91 % of these V. crypta-associated sequences fell into sponge- or sponge/coral clusters (Fig. 3), which is consistent with the results of Schmitt et al. (2011).

The second most abundant group of V. crypta symbionts belonged to the Proteobacteria, which are commonly found and often predominant in microbial consortia associated with different sponges from different marine sites (Friedrich et al. 1999, 2001; Schmidt et al. 2000; Burja and Hill 2001; Webster and Hill 2001; Webster et al. 2001, 2004; Hentschel et al. 2002; Li et al. 2006). D. acidovorans, an obligate aerobe able to grow in 1.5 % NaCl (Wen et al. 1999), was the only validly named next relative for one of the V. crypta betaproteobacterial clones. D. acidovorans has been found in several habitats, such as soil, sediment, activated sludge, crude oil, fresh water and various clinical samples, and Kennedy et al. have reported that it is a sponge-associated bacterium (Kennedy et al. 2008).

Actinobacteria from the microbial communities of sponges have been the focus of natural product screenings (see the review by Taylor et al. 2007), since members of this phylum display the most promising biosynthetic potential for secondary metabolite production (Schneemann et al. 2010). Approximately half of the bioactive secondary metabolites that have been currently discovered in bacteria are attributed to the Actinobacteria (Lam 2006) and many new chemical entities and bioactive metabolites have been reported from marine members of this phylum (Blunt et al. 2004; Salomon et al. 2004; Fiedler et al. 2005; Jensen et al. 2005). In this study, two OTUs belonged to the family Acidimicrobiaceae, which might be involved in secondary metabolite production; however, secondary metabolites are, at present, unexplored from Vaceletia sponges.

Investigations on prokaryotic diversity provide first hypotheses into the putative functions of the microbial communities associated with these sponges. The presence of some clades of the ammonia-oxidizing Beta- and Gammaproteobacteria or some genera of the nitrite-oxidizing Deltaproteobacteria/Nitrospina and Nitrospirae in the community suggests that pathways for nitrogen metabolism (Bayer et al. 2008) are also present in V. crypta. The ammonia-oxidizing bacteria (AOB) were represented here by four OTUs. One was associated with the Betaproteobacteria/Nitrosospira and the others with the Gammaproteobacteria/Nitrososcoccus. In addition, numerous 16S rRNA gene sequences for nitrite-oxidizing bacteria (NOB) were found in our clone library. Three OTUs, representing 5 % of the community, were affiliated with the Deltaproteobacteria/Nitrospina, and 30 sequences were recognized as belonging to the Nitrospira. All of the sequences affiliated with the phylum Nitrospira (9 % of the community) were defined as a single OTU.

The V. crypta microbial community also contains microorganisms which show high sequence homologies to known sulfur-metabolizing bacteria, indicating their possible role in the sulfur cycle. One clone indicated the presence of sulfate-reducing bacteria belonging to the Desulfurellaceae/Deltaproteobacteria. Hoffmann et al. (2005) provided evidence that anaerobic sulfate reduction occurs in Geodia barretti tissue in zones of hypoxia and anoxia, which are created by changes in sponge pumping activity. Sulfate-reducing bacteria (SRB) were also detected by fluorescence in situ hybridization (FISH) in the Mediterranean sponges Chondrosia reniformis and Petrosia ficiformis (Schumann-Kindel et al. 1997; Manz et al. 2000).

Coralline sponges of the genus Vaceletia are representatives of the keratose sponges (Wörheide 2008), which form an early-branching lineage in the Demospongiae (Philippe et al. 2009; Pick et al. 2010), with their earliest fossil record most likely in the late Proterozoic era (Reitner and Wörheide 2002). Our results demonstrate that the complex microbial communities associated with V. crypta are very similar to the microbiota found in other sponges (Taylor et al. 2007). An overwhelming majority of the OTUs were very closely related to other sponge- or coral-derived sequences and moreover fell into SSC/SCC, which underscores that this “living fossil” sponge shares features of its microbial community with other sponges. Such a relatively small divergence between the 16S rRNA gene sequences obtained from different sponges might suggest an environmental acquisition of symbionts (Hentschel et al. 2002; Taylor et al. 2007). If we assume that 50 million years of evolution corresponds to an ~1–2 % 16S rRNA sequence difference (Ochman et al. 1999), then a greater discrepancy should occur if these bacteria had been living separately within their host sponges for 600 million years (Taylor et al. 2007). Moreover, small populations of endosymbiotic microorganisms enhance the fixation of mutations and are, therefore, believed to evolve more rapidly (Ochman et al. 1999). Several studies have shown that sponges from different oceans and with distant taxonomic origins harbor specific microbial consortia (Hentschel et al. 2002; Taylor et al. 2007). Our study is consistent with that pattern because the sponges and corals that contain the microorganisms that are the closest relatives to those associated with V. crypta were collected from different, mostly tropical, geographic regions. In contrast, our DGGE analysis of the 16S rRNA genes of symbionts obtained from a further nine solitary and nine colonial specimens of Vaceletia, reveal that solitary and colonial growth forms appear to harbor distinct communities and suggest a closer relationship between the microbial communities from the same growth form (solitary vs. colonial) than from the same geographic origin (Fig. 8). This observation suggests that the bacterial community might have been achieved not through an environmental acquisition, but through a different mechanism of the transmission followed by successive bacterial speciation within the sponge hosts. Erwin et al. (2012) categorized numerous factors (environmental and host related) which could affect the structure of the microbial communities and noted that factors specific to different host species might have influenced the differences between the Ircinia-associated symbiotic communities. The number of SSC/SCC (63) and the proportion of sequences within SSC or SCC (88 %) for V. crypta appear to be the highest ever reported, indicating a particularly tight sponge-microbe association, which might be related to the evolutionary age of the host species. In addition, some symbionts were specific for V. crypta because they were absent in the microbial community analyzed from another coralline sponge, Astrosclera willeyana, which co-occurs at, and was sampled from, the same site (Karlińska-Batres and Wörheide, in review). A similar trend was observed in sympatric Ircinia species from Mediterranean Sea, which harbored different symbiont communities (Erwin et al. 2012).

This work on sclerosponges from genus Vaceletia enhances our knowledge about microbial communities in sponges and further provides initial insights into the diversity, structure, and composition of the microbiota of these unique sponges. Further research using deeper sequencing, FISH probes and/or specific primers designed for genes involved in denitrification, anammox or particular microbial groups (e.g. SRB and SOB) might reveal these processes in V. crypta providing a clearer picture of the metabolism of this sponge’s microbial community. Future studies might aim to examine if other coralline sponges harbor such diverse communities of symbionts and how much those communities differ from each other and between different geographical locations.