Introduction

Sponges contribute significantly to the total biomass of the tropical reef fauna and may dominate the benthic community in the Caribbean and other tropical waters (Wilkinson and Cheshire, 1990). As filter feeders, sponges process large quantities of seawater removing significant amounts of suspended particles and mircoorganisms (Duckworth et al., 2006). Overall, microorganisms may constitute up to 60% of the tissue biomass in sponges (‘bacteriosponges’) (Reiswig, 1981). Sponges and their microbiota have been intensively studied for their biochemical profiles and it has been shown that species-specific microbial communities differ from those in ambient seawater with certain sponges hosting a uniform microbial population in different oceans (Vacelet and Donadey, 1977; Wilkinson, 1978; Schmidt et al., 2000; Webster and Hill, 2001; Hentschel et al., 2002).

Similarly, estuarine and coastal marine sediments and their microbial communities have been studied in detail for a number of years for their biogeochemical significance (Skyring, 1987; Takii and Fukui, 1991). These, frequently muddy, shallow water sediments may contain up to 3 × 109cells ml–1 and show a clear seasonal variability in their geochemical and microbial composition (Musat et al., 2006). The microbiota associated with sediments retrieved from deep-water environments have also been studied for some time (Kato et al., 1997; Colquhoun et al., 1998; Luna et al., 2004, 2006; Wang et al., 2004). However, it is currently uncertain if these benthic sediments, which do not undergo significant seasonal changes in their geochemical and microbial composition, host specific microbial communities. Although the sediment microbial community has been shown to be dominated by members of the Planctomycetes, the Cytophaga/Flavobacterium group, Gammaproteobacteria and bacteria of the Desulfosarcina/Desulfococcus group, the role and importance of anaerobes such as sulfate-reducing bacteria (SRB) and methanogenic archaea have been well documented (Purdy et al., 2001, 2003; Loy et al., 2002; Dalsgaard et al., 2005; Musat et al., 2006).

Both sponges and marine sediments show steep variations in redox potential. Hoffmann et al. (2005) showed that the oxygen concentrations in Geodia barretti, one of the most ancestral demosponges with origins in the early Cambrian era, is strongly dependent on pumping activity, leading to anoxia in parts of the tissue and the canal system (Gruber and Reitner, 1991; Hoffmann et al., 2005). This anoxia seemed to be a common feature of living Geodia specimens and did not influence their survival while providing an environment favorable for the growth of symbiotic anaerobes. Furthermore, it was shown that in actively pumping Geodia individuals, the cortex and the subcortical spaces were well oxygenated but that the oxygen levels were depleted 4–6 mm below the sponge surface. In non-pumping individuals, oxygen was depleted directly beneath the cortex and diffusive oxygen consumption could be observed in the overlying water.

It may be assumed that anoxia is responsible for regulating the bacterial biota in sponges and marine sediments and may thus be responsible for seasonal fluctuations in the shallow water sediment microbiota. In turn, anoxia in sponges may benefit the sponge host by providing an environment favorable for chemoautotrophic microbial processes that contribute to sponge nutrition (Taylor et al., 2007). Therefore, it would be expected that although sponges and marine sediments may provide similar environments, sponges would harbor bacteria that would specifically aid in nutrient assimilation and cycling.

In this study, we used a combination of anaerobic culture and fluorescence in situ hybridization (FISH) with a comprehensive suite of probes to compare the microbiota of marine sediments and Geodia spp. collected from the Straits of Florida to detect microbiota important in deep-water nutrient cycling and sponge symbiosis. Particular consideration was given to anaerobic microbiota that are not commonly found in seawater to determine whether marine sediments may act as a reservoir for sponge associated microorganisms.

Materials and methods

All chemicals were obtained from Sigma Chemical Co. (St Louis, MO, USA). PCR reactions were performed using IQ Master Mix from BioRad (Hercules, CA, USA). Restriction enzymes were obtained from Promega (Madison, WI, USA). Pre-formulated Nutrient Broth, Marine Agar, Brain Heart Infusion Agar, Sabouraud-Dextrose Agar and Marine Broth were prepared according to the manufacturer's instructions unless stated otherwise and supplemented with cysteine 100 μg ml–1 (DIFCO Laboratories, Detroit, MI, USA).

Sample collection and treatment

Collection of sponge specimens and sediment samples (top 10 cm of the seafloor) was performed on two separate research expeditions (April 2005 and August 2005) to the Straits of Florida using the R/V Seward Johnson and the Johnson-Sea-Link I (JSL I) research submersible. Collection sites and depths are given in Table 1. All sponge specimens and sediments were handled with nitrile examination gloves. Sponges were identified using standard spicule analysis. Approximately 10 g of each sponge specimen or 10 g of each sediment sample was homogenized under sterile conditions for 3 min (25 °C) in 100-ml sterile Artificial Seawater (pH: 7.2) using a commercial Waring blender (Waring Laboratory Science, Torrington, CT, USA). This homogenate was then used to establish bacterial cultures on solid agar plates. Intact specimens of sponge tissue were shock frozen at −80 °C in 50% (v/v) glycerol.

Table 1 Collection sites
Full size table

Fluorescence in situ hybridization

FISH with ribosomal RNA-targeted probes was carried out using a variety of probes targeting archaea, common classes of bacteria, and selected anaerobic bacteria (Table 2). Results of microscopic counting were corrected by subtracting fluorescence signals of the probes with that of nonsense probe NONEUB (NON-EUB338), which was previously shown not to hybridize with any prokaryotic cells (Amann et al., 1990; Wallner et al., 1993). All probes were synthesized and monolabelled at the 5′ end with Cy3 (Ex 552 nm, Em 568 nm) by Sigma-Genosys (Dublin, Ireland). Approximately 1 ml of sponge homogenate or 1 ml of sediment homogenate was fixed in 10 ml of either 95% (v/v) ethanol (for Gram-positive bacteria) or 4% (w/v) paraformaldehyde in phosphate-buffered saline (pH 7.2, for Gram-negative bacteria) overnight. Fixed samples in 1-ml aliquots were washed three times in phosphate-buffered saline and resuspended in a mixture of 150 μl phosphate-buffered saline and 150 μl of 96% (v/v) ethanol for storage at –20 °C until further use.

Table 2 Hybridization (FISH) probes (available at http://www.microbial-ecology.net/probebase/; Loy et al., 2002)
Full size table

For hybridization, 1 ml of fixed sample was centrifuged at 2000 × g for 2 min to remove large particulate matter. A volume of 100 μl of centrifuged sample was added to Epoxy-printed three well (14-mm diameter) microscope slides (Menzel, Braunschweig, Germany) and dried at room temperature. Slides were dehydrated in ethanol (50, 80 and 96% (v/v) for 3 min each). While slides dried, 1 ml of hybridization buffer containing 0.9 M NaCl, 0.02 M TrisHCl, and 0.01% (w/v) sodium dodecyl sulfate was prepared. The formamide concentration of the hybridization buffer was adapted according to the individual probe's requirements with high performance liquid chromatography grade water being added to bring the final volume to 1 ml. Hybridization buffer 10 μl was added to 1 μl of the appropriate probe (50 ng μl–1) and then added to the sample slide. The remaining buffer solution was poured over a piece of tissue paper that was added to a 50-ml centrifuge tube in which the sample slide was placed. The tube was then placed horizontally with the cap tightly closed in a hybridization oven at 46 °C for 3 h. After hybridization, the buffer was removed from the sample using a small amount of preheated (48 °C) washing buffer containing 0.02 M TrisHCl, 0.01 M ethylene-diaminetetraacetic acid and 5–900 mM NaCl, depending on probe requirements. For total counts, 50 nM SYTO 9 (Invitrogen, Dun Laoghaire, Ireland) was added to the washing buffer. After all liquid was removed from the slide, 500 μl of the washing buffer was added to the sample slide with a further 1 ml of washing buffer being poured over a piece of tissue paper that was added to a new 50-ml centrifuge tube. The sample slide was then placed horizontally into the new tube, which was capped tightly and placed in a hybridization oven at 48 °C for 15 min. After washing, the slide was dipped in ice-cold high performance liquid chromatography grade water for 3 s and air dried. Slides were stored at −20 °C until use. For analysis, one drop of SlowFade-Light Antifade Kit component A (Invitrogen) was added to the slide. Organisms were evaluated using an Olympus IX51 microscope (Olympus UK, London, UK) with epifluorescence attachment (Olympus U-RFL-T) and appropriate filter sets. Image analysis was performed using an Olympus DP70 camera system and Olympus CellF imaging software. Fifteen random fields with a good distribution of cells (10–100) were counted for each probe and sample.

Anaerobic culture

Aliquots (10 and 100 μl) of each sponge and sediment homogenate were spread onto standard Nutrient Agar, diluted Marine Agar (1/5 strength Marine Agar 2216 diluted with Artificial Seawater and supplemented with agar to a final concentration of 1.5%(w/v)), Brain-Heart Infusion agar, Brain-Heart Infusion agar with Artificial Seawater and Sabouraud Dextrose Agar plates in triplicate. Plates were incubated at ambient temperature in screw top 2.5-l anaerobic jars containing one sachet of anaerobic atmosphere generator (Oxoid, Basingstoke, UK). The jars were further sealed in clear plastic bags filled with N2. Once back on shore, all samples were immediately transferred to an anaerobic chamber (Coy, Grass Lake, MI, USA). Bacterial growth was monitored for 4 weeks and individual colonies with unique morphotypes were serially streaked on their respective medium until pure cultures were obtained. Pure cultures were then transferred to diluted Marine Agar plates before DNA isolation. Pure cultures on diluted Marine Agar plates were also used to check cultures for aerobic growth by aerobic incubation at ambient temperature for a maximum of 4 weeks. For long-term storage, purified bacterial isolates were grown in 10-ml Marine Broth 2216 and 1-ml aliquots were frozen in 10% (v/v, final concentration) glycerol at −80 °C.

DNA extraction and PCR of cultured bacterial isolates

DNA of bacterial isolates was extracted by touching the colony with a sterile needle, which was then placed in 10 μl sterile 5% (w/v) Chelex (BioRad) solution. Samples were boiled for 10 min before centrifugation at 14 000 × g. Supernatant containing extracted DNA was transferred to a new tube and stored at −80 °C before further use. Eubacterial universal primers FC27 (5′-AGAGTTTGATCCTGGCTCAG-3′) and RC1492 (5′-TACGGCTACCTTGTTACGACTT-3′) (Mincer et al., 2004) were used to amplify the 16S ribosomal DNA in standard PCR reactions containing 10 pmol of each primer, 12.5 μl IQ Supermix (BioRad), and sterile H2O to a final volume of 24 μl. The PCR mix was added to 1 μl of DNA template and cycled as follows: initial denaturing at 95 °C for 5 min, 35 cycles at 95 °C for 30 s, 55 °C for 30 s and 72 °C for 1.5 min. A final extension of 7 min at 72 °C was added. Each PCR product was checked by gel electrophoresis. Expected PCR amplicons of the correct size (∼1.5 kbp) were manually excised and gel purified using the Minelute gel extraction kit (Qiagen, Valencia, CA, USA). The PCR products were screened and grouped by restriction fragment length polymorphism analysis using HhaI restriction patterns (data not shown). PCR products from each restriction fragment length polymorphism group and those cultures with unique morphotypes were sequenced by Northwoods DNA (Solway, MN, USA). The 16S recombinant DNA sequences were viewed and edited in Chromas (Technelysium, Tewantin, Australia). Sequences were identified using BLAST N at the NCBI database (www.ncbi.nlm.nih.gov, Altschul et al., 1997)

Phylogenetic reconstruction

All phylogenetic reconstruction was performed using ARB (Ludwig et al., 2004). Alignments of 16S small subunit (SSU) ribosomal RNA sequences were made with the ARB fast alignment tool and were checked manually before being added to the ARB data set using the Arb_Parsimony tool to select suitable outgroups. De novo trees were constructed using the ARB neighbor-joining distance matrix with Felsenstein correction and termini (parameters: .−=0) and position variance (parameters: 123456789.−=0) filters (Felsenstein, 1988). The most appropriate DNA substitution model for distance analyses was determined by MODELTEST (Posada and Crandall, 1998). New sequences generated in this study were submitted to GenBank under the following accession numbers: EF114127–EF114207 and EU203318–EU203332.

Results

Three deep-water sponge specimens used for this study were collected in August 2005 with Harbor Branch Oceanographic Institute's JSL research submersible from the support vessel R/V Seward Johnson (Table 1). A taxonomic voucher specimen is deposited for each at the Harbor Branch Oceanographic Museum: catalog numbers 003:01040, 003:01041 and 003:01042.

Specimen 003:01040 was collected (dive number JSL I-4820) in the Straits of Florida, ∼13-nm offshore southeastern Florida at a depth of 304 m from a rocky slope of the Miami Terrace escarpment. The specimen is pear-shaped, ∼20-cm tall, with a single 5-cm apical osculum. It is light brown in color; the apical tip is lighter with alternating brown and white stripes around the osculum. The striped area of the osculum is hispid with 2- to 3-mm fringing spicules. The surface is hard, smooth and finely pitted, forming a detachable ectosomal cortex 2- to 3-mm thick. The specimen fits the description of the genus Geodia Lamarck, 1815 (Phylum- Porifera, Class- Demospongiae, Order- Astrophorida; Hooper and Van Soest, 2002).

Specimen 003:01041 was collected (dive JSL I-4826) from the Miami Terrace escarpment, ∼30 nm south of the first specimen and at a depth of 341 m. It also fits the description of the genus Geodia but is a different species than the previous. The specimen is oblong to subspherical, ∼11-cm maximum diameter, and has a gray to light brown color in situ. The hard detachable ectosomal cortex is 2- to 3-mm thick and the surface is dimpled. Three oscula are 4 mm in diameter and slightly raised on 4-mm cones.

Specimen 003:01042 was collected (dive JSL I-4829) further south in the Straits of Florida off the Florida Keys on the Pourtales Terrace at a depth of 197 m. It also fits the description of the genus Geodia but is also an unknown species that is different from the other two specimens. It is a flattened sphere ∼25 cm in diameter and 15-cm tall. The hard 2-mm ectosomal cortex is hispid, covered with 2- to 5-cm long spicules and highly sedimented.

Analysis of homogenates by FISH showed that a wide variety of microbes, including some putative anaerobes are present in both Geodia spp. and sediments (Figure 1). Archaea were present at 1 × 106 cells g–1 of sample (wet weight) while the total bacterial population was 1 × 108 cells g–1 of sample. Gammaproteobacteria and Firmicutes gave the most signals in both sediment and Geodia spp. samples. Clostridium spp. and Shewanella spp. constituted only a minor part on the total Firmicutes and gammaproteobacterial population, respectively. In contrast, SRB were shown to be a significant part of the deltaproteobacterial population (between 30 and 87%) for both sample types. Anammox organisms made up between 0.65 and 2.4% of the overall population. Betaproteobacterial ammonia-oxidizing bacteria made up between 12 and 57% of the overall betaproteobacterial population. Although Chloroflexi were present at 1 × 107 cells g–1 of sample in Geodia spp., there was a significantly (P>0.005) smaller number of Chloroflexi present in sediment samples with 1.5 × 105 cells g–1 of sample.

Figure 1
figure 1

Log10 average populations of Eubacteria associated with sediment and Geodia spp. samples estimated by means of FISH counting. Error bars are ±one s.d. Log cells g–1 is the logarithmic (base 10) count of cells (bacteria) per gram of sample (wet weight). *Indicates a significant difference (P<0.005) between Geodia spp. and sediment populations. (A) In front of probe target denotes strict anaerobe organisms.

Full size image

A total of 96 bacterial isolates from sediment and Geodia spp. were derived using standard anaerobic bacterial techniques on four different microbial media, which varied in pH, nutrient content and ionic strength (salt content) to mimic different bacterial micro-environments (Tables 3a,b, Figure 2). A majority of the microbial isolates presented here, grew exclusively on a single medium type. All bacterial isolates were analyzed by classical microbiological and genetic methods, including determination of cell morphology, gram staining and 16S ribosomal RNA gene sequence analysis.

Table 3 Microbial isolates from (a) Geodia spp. and (b) sediments
Full size table
Figure 2
figure 2

Distance-based neighbor-joining phylogeny of 16S ribosomal RNA (rRNA) gene sequences obtained from anaerobic isolates from sediment and sponge tissue. Numbers at nodes are percentages indicating levels of bootstrap support, based on neighbor-joining analysis of 1000 re-sampled data sets. Only values ⩾60% are shown. Scale bar represents 0.1 substitution per nucleotide position. SPO=Isolate from Geodia spp., SED=Isolate from sediment.

Full size image

Fifty unique organisms from three Geodia spp. specimens (Table 3a, Figure 2) were found in anaerobic culture of which Firmicutes (22 isolates, 44%) represented the largest organism cluster. Gammaproteobacteria formed the second largest bacterial group in overall microbial diversity, comprising 40% (20 isolates) of all bacterial isolates. Minor culturable components of the microbial isolates consisted of Chloroflexi (one isolate, 2%), Actinobacteria (three isolates, 6%), Betaproteobacteria (three isolates, 6%) and Bacteroidetes (one isolate, 2%). With the exception of six obligate anaerobic organisms (WMB24A-E, W060), all bacteria were facultative anaerobes. A total of 46 isolates were 96–100% homologous to GenBank sequences; four were 91–95% homologous. Bacillus spp. were the most abundant organisms found in culture comprising 11 isolates (22%). Vibrio spp. were represented with 10 isolates (20%), while Staphylococcus spp. were represented by 4 isolates (8%).

Surprisingly, the phylogenetic analysis (Figure 3) of the cultured Chloroflexi species found in this study indicated a close relationship to a cluster of unculturable marine-derived Chloroflexi, which have been suggested to form stable symbiotic relationships with sponges (Taylor et al., 2007). In contrast, the cultured Chloroflexi species only showed a distant relationship with other Chloroflexi 16S RNA sequences found in the NCBI database. This current data infers that the organism isolated in this study forms a symbiotic relationship with deep-water sponges of the genus Geodia spp. To our knowledge this is the first evidence of a culturable member of the Chloroflexi isolated from a deep-water sponge.

Figure 3
figure 3

De novo neighbor-joining phylogeny comparing the Chloroflexi isolated from Geodia spp. in this study (EU203321) with other sponge-origin reference sequences. Numbers at nodes are percentages indicating levels of bootstrap support, based on neighbor-joining analysis of 1000 re-sampled data sets. Only values ⩾60% are shown. Scale bar represents 0.1 substitution per nucleotide position. Reference sequences derived from previous studies and GenBank entries are described in the text or are written with their corresponding accession numbers.

Full size image

Sediment samples used in this study had a grayish-green to black appearance and a fine sand grain structure. Anaerobic culture of sediment samples (Tab. 3b, Figure 2) again showed a wide array of organisms able to grow anaerobically, with a total of 46 isolates. Again, Firmicutes represented the largest group (63%, 29 organisms), while Gammaproteobacteria were the second largest group (28%, 13 organisms). One species (2%) each of Actinobacteria, Alphaproteobacteria, Bacteroidetes and Deltaproteobacteria was also present. With the exception of six obligate anaerobic organisms (WMB24F-I, W002, W003), all bacteria were facultative anaerobes. Forty-four isolates were 96–100% homologous to GenBank sequences. One isolate (Bacillus spp.) showed 92% homology to GenBank sequences while another was 95% (Bacillus spp.) homologous, which may indicate that these organisms are novel variants of the species Bacillus spp., which represented the major bacterial cluster found in sediment cultures amounting to 16 specimens (34%). Vibrio spp. was represented by seven isolates (15%). Staphylococcus spp. was represented with four isolates (9%).

In both the sponge and sediment samples, most sequences obtained from obligate anaerobes were most closely associated with uncultured bacterium 08SE (accession number AF018038) from the gut of the Tyrolean iceman (Cano et al., 2000). BLAST analysis further revealed the sequence to be a member of the Clostridium spp. Although these isolates are related to the same GenBank database entry, they differed in a combination of other parameters linked to their analysis set. Each culture was either isolated from a different growth medium and/or showed marked differences in gross colony morphology and color. As different bacterial strains can have a divergent biochemical and genetic make-up, these are included in the data presented here. Extensive morphological studies identified the organisms as obligate anaerobic, Gram positive, spore forming (sub-terminal to terminal) single rods, indicating the presence of a Clostridium spp.

Discussion

Although it is assumed that sponge-bacterium symbioses have existed for millions of years, the mechanisms by which this association is established are not well understood (Taylor et al., 2007). It may be assumed that microorganisms capable of association with sponges are present in the surrounding seawater and environment, but at abundances below the detection limit of currently available methods (Müller and Müller, 2003). Therefore, as anaerobic bacteria are not major components of the bacterial communities of seawater, resuspended sediment particles are a likely source for these and other microbiota. To enable the in situ assessment of microbial consortia in the deep-water sponge Geodia spp. and in surrounding sediment samples, we have utilized a combination of FISH with a diverse oligonucelotide probe set and direct microbial culture, which offered the possibility of characterizing microbes using traditional microbiological techniques. These methodologies have been proven to give a reliable overview of in situ microbial associations in various marine organisms, including sponges (Hoffmann et al., 2006; Brück et al., 2007). Our assessment of specimens of Geodia spp. and sediment showed that a significant proportion of microbiota is shared between the two. In fact, only Chloroflexi were present in significantly larger numbers in Geodia spp. in comparison with sediment. The major bacterial constituents Gammaproteobacteria, Actinobacteria and Firmicutes as well as anaerobic microbiota such as SRB and Clostridium spp. were present in similar numbers in Geodia spp. and sediment. Similar results were observed by Hoffmann et al. (2006) who showed that G. barretti is dominated by Alphaproteobacteria and Gammaproteobacteria.

Studies with the Adriatic sponge Suberites domuncula have suggested that oxygen levels within the sponge tissue are responsible for regulating the resident microbiota (Müller et al., 2004). Bacillus strains represent approximately 20% of the total heterotrophic microbiota in seawater whereas they constitute up to 80% of the total number of the culturable heterotrophic bacteria in marine sediments (Harwood, 1989). Similarly, Bacillus subtilis and Bacillus pumilus were the most abundant species among those associated with marine sponges (Ivanova et al., 1992, 1999). Spores of Bacillus and Clostridium species are metabolically dormant and extremely resistant to acute environmental stresses such as low nutrient availability and varying oxygen concentrations (James et al., 2000). Therefore, it is possible that spores of Bacillus and Clostridium species can survive for many years in marine sediments and sponge tissues until conditions are right for germination and formation of a metabolically active cell. In our study, Bacillus spp. constituted 40% of the total culturable sediment microbiota and 22% of culturable sponge associated microorganisms. FISH probing resulted in the detection of significant numbers of Clostridium spp. and other Firmicutes. Numbers in FISH did not vary significantly between sponge tissue and sediment. This suggests that these organisms were indeed viable within the examined environments.

The facultative nature of the majority of microbes isolated from sediment and Geodia spp. suggests that an anaerobic metabolism may not be a necessary survival strategy for most of the microbiota in sponges and marine sediments, however, it may provide a mechanism for coping with periods of anaerobiosis. The efficient utilization and recycling of nutrients, as previously found in Geodia spp., may rely on anoxic cycles within sponge tissue making symbiotic microorganisms capable of anaerobic metabolism an important factor in sponge survival (Schumann-Kindel et al., 1997). Furthermore, anoxic zones in Geodia spp. may have developed as an effective buffer system that prevents sulfide toxification and overgrowth of SRB (Hoffmann et al., 2006). Members of the Alphaproteobacteria, Betaproteobacteria, Gammaproteobacteria and Bacteroidetes have also been shown to have a critical role in extending the anaerobic oxidation of ammonium and nitrite to nutrient-depleted suboxic water layers by creating anoxic, nutrient-enriched microniches (Woebken et al., 2007). This niche-formation may also occur in sponges such as Geodia spp. and hence may have an important role in the removal and bioconversion of dissolved nitrogenous waste products such as nitrite, ammonia and organic detritus (marine snow; Hoffmann et al., 2005). Therefore, Geodia spp. may rely on external organic nitrogen sources in addition nitrogen-fixing symbionts for growth and tissue remodeling. As nitrogen is the limiting growth factor for biomass formation in nutrient poor environments, sponges rely on prokaryotic symbionts with specific nitrogen recycling strategies to prevent enduring nutrient stress (Meier et al., 1994; Zehr and Ward, 2002; Lenton and Klausmeier, 2006). The annamox bacterial group is capable of converting nitrite and ammonia to nitrogen, which can be fixed to organic intermediates by subsequent microbial processes (Taylor et al., 2007). Furthermore, a consolidated formation of organic nitrogen intermediates by prokaryotic ammonia and nitrite oxidizers may lead to intermediates that can be assimilated by the sponge host leading to complete nitrogen cycling within a sponge (Taylor et al., 2007). In this study, annamox organisms were found in Geodia spp. and sediment samples suggesting that bioconversion and nutrient cycling of nitrogenous compounds may occur in these samples.

It has been suggested that the nutrient exchange between sponge host and microbial symbiont may also be driven by the sponge restricting the symbiont's access to essential nutrients and thus supplying the host with an excess of organic carbon (Hinde, 1988; Wilkinson, 1992). In shallow water sponges, cyanobacteria have been identified as the dominant symbionts responsible transfer of photosynthetically derived organic carbon compounds mainly in form of glycerol (C3) for metabolism in the sponge host (Wilkinson, 1980, 1983). In addition to cyanobacteria, a specific cluster of the bacterial class Chloroflexi has been identified as sponge-specific symbionts, which is also supported by our data (Figures 1 and 3, Hentschel et al., 2002; Taylor et al., 2007). However, a possible interaction of these microbes with the sponge host has so far not been suggested. Much like cyanobacteria, Chloroflexi are filamentous microorganisms, which are capable of permanently integrating into sponge tissue. In light suffused shallow water environments Chloroflexi are also capable photosynthetic fixation of atmospheric CO2. On the basis of the similar metabolic capacities between cyanobacteria and chloroflexi, these bacteria may also provide carbonaceous photosynthates, such as glycerol, to the sponge host in shallow water environments.

Although symbiotic cyanobacteria cannot thrive in light-deprived environments (Thacker, 2005), Chloroflexi can metabolically adjust to dark, nutrient poor environments such as the deep ocean. FISH data sets reported in this study showed significantly higher counts of Chloroflexi in the tissue of Geodia spp. when compared with marine sediment samples. This data indicating that deep-water Geodia tissue is enriched with Chloroflexi points to a potential symbiotic relationship. Previous studies of Chloroflexi sponge interactions were complicated by the fact that none of the Chloroflexi species could be recovered in culture for more detailed microbiological studies. To our knowledge, this study presents the first account of a culturable marine bacterium of the class Chloroflexi isolated from sponge tissue. Comparative phylogenetic analysis of this isolate with database DNA sequences of unculturable Chloroflexi suggested that this organism may actually form a true symbiotic relationship with deep-water sponge hosts (Hentschel et al., 2002). Chloroflexi can convert inorganic into organic carbon by way of the 3- hydroxyproprionate pathway (Brock et al., 1984). The resulting organic intermediates produced by this microbial symbiont could subsequently be transferred to the host sponge and metabolized, which would impart a clear synergistic survival advantage to the host in nutrient poor environments such as the deep ocean floor. The Chloroflexi-catalyzed conversion of inorganic CO2 to metabolizable organic intermediates certainly contributes to microbial biomass formation in deep-water environments. The inferred symbiotic relationship of Chloroflexi may also contribute to additional biomass formation in deep-water Geodia species.

In conclusion, as marine microorganisms survive under harsh environmental conditions, they can be expected to be a source of novel biogeochemical and biochemical processes (Barlet et al., 1995; Turley, 2000). Through this study, it has become evident that sponge and sediment microbiota share a striking overall similarity and that facultative and obligate anaerobic microorganisms thrive in these environments, in which they may aid in nutrient recycling and bioconversion or some other, to date unknown, function. Resuspended sediment particles may further form a realistic source, in addition to vertical transfer, of microorganisms able to associate or form a symbiotic relationship with sponges. Such microbes represent another facet of the cultivable microbiota that remains to be examined for its biotechnological potential.